# a map of consciousness ideas about mind and consciousness that joscha bach has mentioned in public, from the earliest myths to november 2026. a simplified, AI-drawn picture of the ideas, not the ideas themselves. unofficial: joscha bach didn't make or approve it. made with claude fable 5.2 and independently checked by gpt 5.6 sol as of 13 november 2026; not checked by a person. interactive version: https://mxz.sh/bach-map/. how to read it and where it comes from: https://mxz.sh/bach-map/about.html 169 points, 223 connections, in chronological order. each summary and connection carries a confidence word (highly likely, likely, possible) that follows from source checks by two models; the lower verdict is shown. summaries are general background, not joscha bach's words, except quotes marked as his. ## Animism prehistoric, myth and religion. school of thought: religion and contemplative traditions. id: `animism` The belief, found in most traditional cultures, that living things and sometimes places are inhabited by spirits. It is often treated as the oldest form of religion. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Animism) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Animism) start here, joscha bach on it: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) (verbatim): the spirits of the animist worldview are not superstitions or mere analogies to computer software connections: - related to Greek gods as archetypes (`greekgods`): greek religion kept animist elements beside its pantheon of gods. confidence: highly likely, [source](https://resources.warburg.sas.ac.uk/pdf/bkf100b2414221.pdf) - related to Aristotle (`aristotle`): bach reads aristotle's soul, the form animating a body, as animist. confidence: possible. Aristotle's hylomorphism is documented; calling it animist is Bach's interpretation, not a standard historical classification. - leads to Joscha Bach (`bach`): spirits as self-organising software. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) ## Greek gods as archetypes c. 8th c. BCE, myth and religion. school of thought: religion and contemplative traditions. id: `greekgods` The Olympian gods each preside over a facet of human life: war, love, craft, trade. They quarrel among themselves and take sides in human affairs, as in Homer's Iliad. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Greek_mythology) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Twelve_Olympians) start here, joscha bach on it: [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI&t=9395s), from 2:36:37 (verbatim): So the Greek gods are stable because they are all archetypes. connections: - related to Animism (`animism`): greek religion kept animist elements beside its pantheon of gods. confidence: highly likely, [source](https://resources.warburg.sas.ac.uk/pdf/bkf100b2414221.pdf) - related to Joscha Bach (`bach`): gods as shared archetypes. confidence: highly likely, [source](https://www.youtube.com/watch?v=3MNBxfrmfmI) ## Genesis and the Hebrew Bible c. 6th c. BCE, myth and religion. school of thought: religion and contemplative traditions. id: `genesis` ### Genesis 1 by Hebrew Bible The opening chapter of the Bible, in which God separates light from darkness and the waters above from those below, then makes plants, animals and humans. It is usually read as an account of the creation of the physical world in six days. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Genesis_creation_narrative) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Genesis_creation_narrative) start here, joscha bach on it: [Lex Fridman Podcast #101 – Joscha Bach: Artificial Consciousness and the Nature of Reality](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality) (verbatim): I think that Genesis is actually the childhood memories of a God. connections: - leads to Christianity (`christianity`): the hebrew bible becomes the old testament. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Old_Testament) - leads to Joscha Bach (`bach`): genesis 1 read as consciousness booting up a world. confidence: highly likely, [source](https://c3subtitles.de/talk/2838) ### Adam and Eve by Hebrew Bible The biblical story of the first humans living in a paradise without suffering until they eat from the tree of knowledge and are expelled. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Adam_and_Eve) ### Tower of Babel by Hebrew Bible The story in which humanity's shared language is confused so that people can no longer understand each other and their great tower is abandoned. It appears in Genesis 11 and is traditionally read as a warning against human overreach. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Tower_of_Babel) ### The God of Abraham The Abrahamic religions worship a single God who makes a covenant with a people and gives them law. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Abrahamic_religions) ### The Bible The canonical scripture of Judaism and Christianity, a library of myths, laws, histories and poetry. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bible) ## Pythagoras 6th c. BCE, mathematics and logic. school of thought: rationalism and idealism. id: `pythagoras` Greek thinker whose school held that number is the basis of reality. Later tradition says the school was thrown into scandal by the discovery of irrational quantities. confidence: highly likely, [source](https://plato.stanford.edu/entries/pythagoreanism/) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/pythagoreanism/) start here, joscha bach on it: [Donald Hoffman Λ Joscha Bach: Consciousness, Gödel, Reality (Theories of Everything)](https://www.youtube.com/watch?v=bhSlYfVtgww&t=5340s), from 1:29:02 (verbatim): I agree with Pythagoras that real numbers don't exist. connections: - related to Plato: forms and the cave (`plato`): a disputed Pythagorean influence on Plato's metaphysics and mathematics. confidence: highly likely, [source](https://plato.stanford.edu/entries/pythagoreanism/) ## Logos c. 500 BCE, philosophy. school of thought: rationalism and idealism. id: `logos` by Heraclitus; Stoics; Gospel of John Greek for word, reason or rational order: the principle by which the cosmos is intelligible. The Gospel of John opens with it ('In the beginning was the Word'), carrying the idea into Christianity. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logos) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Logos) start here, joscha bach on it: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=4924s), from 1:22:06 (verbatim): this pattern of the Ruliad is exactly the superposition of all the automata that is producing the universe, or the old Greeks called it Logos connections: - leads to Stoicism (`stoicism`): the stoics made logos the rational order of the cosmos. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logos) - leads to Christianity (`christianity`): 'in the beginning was the word'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logos) ## Buddhism and meditation 5th c. BCE, myth and religion. school of thought: religion and contemplative traditions. id: `buddhism` ### Buddhism by Siddhartha Gautama A contemplative tradition that studies experience from the inside through meditation, teaching that the self is constructed and that suffering comes from attachment. Its insight practice (vipassanā) aims at seeing directly that everything in experience, the self included, is impermanent. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Buddhism) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Buddhism) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=2310), from 38:32 (verbatim): Like this monk, who sits down in a monastery for 20 years to hack their reward function they decide not to have kids connections: - related to Joscha Bach (`bach`): bach's 'enlightened' stage is reached mostly by advanced buddhist meditators. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) ### The Dalai Lama by Tibetan Buddhism In Tibetan Buddhism the Dalai Lama is held to be the same being reborn in successive bodies. Each new Dalai Lama is found as a child and raised into the role. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dalai_Lama) ### Michael Taft on enlightenment American meditation teacher and author of 'The Mindful Geek', who teaches contemplative practice in secular, non-mystical terms. He hosts the Deconstructing Yourself podcast. confidence: highly likely, [source](https://deconstructingyourself.com/michael-w-taft) ## Plato: forms and the cave c. 375 BCE, philosophy. school of thought: rationalism and idealism. id: `plato` by Plato Plato held that behind the changing things we perceive there are ideal, timeless forms (the perfect triangle, the ideal table). In mathematics, 'Platonism' is the view that mathematical objects really exist independently of minds. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Theory_of_forms) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/platonism-mathematics/) start here, joscha bach on it: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) (verbatim): I suspect that Plato is the first guy who wrote about the embedding space that we discovered much later in AI. connections: - related to Pythagoras (`pythagoras`): a disputed Pythagorean influence on Plato's metaphysics and mathematics. confidence: highly likely, [source](https://plato.stanford.edu/entries/pythagoreanism/) - leads to Aristotle (`aristotle`): aristotle answers his teacher's theory of forms. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Theory_of_forms) - related to Joscha Bach (`bach`): vervaeke calls the cave the first simulation theory; bach answers 'almost'. confidence: highly likely, [source](https://www.youtube.com/watch?v=rK7ux_JhHM4) ## Aristotle c. 350 BCE, philosophy. school of thought: rationalism and idealism. id: `aristotle` Greek philosopher who systematised logic, physics and psychology. He described the soul (psyche) not as a ghostly substance but as the organising form of a living body, with vegetative, animal and rational layers. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Aristotle) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/aristotle-psychology/) start here, joscha bach on it: [Self Models of Loving Grace (38C3)](https://media.ccc.de/v/38c3-self-models-of-loving-grace) (verbatim): Aristotle already said that the soul is a form a pattern imposed on physics I think what he meant is it's a causal pattern connections: - related to Animism (`animism`): bach reads aristotle's soul, the form animating a body, as animist. confidence: possible. Aristotle's hylomorphism is documented; calling it animist is Bach's interpretation, not a standard historical classification. - builds on Plato: forms and the cave (`plato`): aristotle answers his teacher's theory of forms. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Theory_of_forms) - leads to Thomas Aquinas (`aquinas`): aquinas merges aristotle with christian doctrine. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thomas_Aquinas) - leads to Joscha Bach (`bach`): the soul as the form of a living body. confidence: highly likely, [source](https://media.ccc.de/v/38c3-self-models-of-loving-grace) ### De Anima (On the Soul) by Aristotle Aristotle's treatise on the soul. It describes three levels: a vegetative soul (growth and nutrition, in plants), a sensitive soul (perception and desire, in animals) and a rational soul (in humans). The soul is the form of a living body and so cannot exist apart from it, with a disputed exception for the active intellect. confidence: highly likely, [source](https://en.wikipedia.org/wiki/On_the_Soul) ### The unmoved mover by Aristotle Aristotle argued that if anything in the universe changes, something must ultimately drive that change without itself being moved. He called this 'prime mover' a god, but a spare one: it does nothing except think about thinking, and moves the heavens only as an object of desire. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Unmoved_mover) ## Stoicism 3rd c. BCE, philosophy. school of thought: ethics, society and AI risk. id: `stoicism` by Zeno of Citium and successors A school of philosophy teaching that we suffer from our judgments about events rather than the events themselves, and that we should concern ourselves only with what is in our control. confidence: highly likely, [source](https://iep.utm.edu/stoicism/) start here, read: [Internet Encyclopedia of Philosophy](https://iep.utm.edu/stoicism/) start here, joscha bach on it: [Singularity.FM – Joscha Bach: We need to understand the nature of AI to understand who we are](https://www.youtube.com/watch?v=sIKbp3KcS8A) (verbatim): I suspect that stoicism is maladaptive from a permanent evolutionary perspective connections: - builds on Logos (`logos`): the stoics made logos the rational order of the cosmos. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logos) ## Christianity 1st c. CE onward, myth and religion. school of thought: religion and contemplative traditions. id: `christianity` ### The Catholic Church The religion built on the life and teaching of Jesus; from late antiquity to the early modern age the Catholic Church dominated Western civilisation. Beyond its beliefs, it was an organisation: medieval Christendom saw itself as one community under the Church, bound by canon law. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Christendom) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Christianity) start here, joscha bach on it: [Joscha Bach: "You Are Not Who You Think You Are!" (Dr Brian Keating, Into the Impossible)](https://www.youtube.com/watch?v=7bqdPHLIY8w&t=0s), from 0:00 (verbatim): at some point Christianity awoke at this societal level as an agent with the Catholic Church as its nervous system and brain connections: - builds on Genesis and the Hebrew Bible (`genesis`): the hebrew bible becomes the old testament. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Old_Testament) - builds on Logos (`logos`): 'in the beginning was the word'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logos) - leads to Thomas Aquinas (`aquinas`): scholastic theology. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thomas_Aquinas) - leads to René Descartes (`descartes`): descartes set out to prove the soul immaterial, as the church asked of christian philosophers. confidence: highly likely, [source](https://web.viu.ca/johnstoi/descartes/descartesmeditations.htm) - related to Joscha Bach (`bach`): the church as a collective agent; 1 corinthians 13. confidence: highly likely, [source](https://www.youtube.com/watch?v=7bqdPHLIY8w) ### 1 Corinthians 13 by Paul the Apostle A New Testament passage saying that knowledge, prophecy and sacrifice are worth nothing without love. confidence: highly likely, [source](https://en.wikipedia.org/wiki/1_Corinthians_13) ### Christian mind-body dualism The traditional Christian teaching that a person has an immaterial soul distinct from the body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Christian_anthropology) ### Protestantism by Martin Luther and others The branch of Christianity that emerged from the Reformation, stressing the individual's direct relationship to God and personal conscience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Protestantism) ## Thomas Aquinas 13th c., philosophy. school of thought: religion and contemplative traditions. id: `aquinas` Medieval theologian who merged Aristotle with Christian doctrine. He argued for God as the necessary ground of existence and described seven virtues: four practical ones (temperance, justice, prudence, courage) and three theological ones (faith, hope, love). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thomas_Aquinas) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/aquinas/) start here, joscha bach on it: [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience#t=1586), from 26:28 (verbatim): aquinus in some sense defines God as the best possible most top level Collective agent connections: - builds on Aristotle (`aristotle`): aquinas merges aristotle with christian doctrine. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thomas_Aquinas) - builds on Christianity (`christianity`): scholastic theology. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thomas_Aquinas) - leads to Joscha Bach (`bach`): love among the theological virtues. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-currents-083-joscha-bach-on-synthetic-intelligence/) ### The seven virtues by Thomas Aquinas (after Plato, Aristotle, St Paul) The classical list of prudence, justice, temperance and courage, plus the Christian faith, hope and love. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Seven_virtues) ## The golem legend, 16th–19th c., fiction. school of thought: religion and contemplative traditions. id: `golem` by Jewish folklore A clay figure brought to life by a rabbi to serve and protect, which follows instructions literally and can run out of control. The best-known version, first written down in the nineteenth century, is about the sixteenth-century Rabbi Loew of Prague. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Golem) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Golem) start here, joscha bach on it: [Lex Fridman Podcast #392 – Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans](https://lexfridman.com/joscha-bach-3-transcript) (verbatim): My intuition is that the language models that we are building are golems. ## René Descartes 1641, philosophy. school of thought: rationalism and idealism. id: `descartes` ### Descartes French philosopher famous for 'I think, therefore I am' and for dividing reality into extended stuff (res extensa) and thinking stuff (res cogitans), the classic statement of mind-body dualism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ren%C3%A9_Descartes) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/descartes/) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM&t=776s), from 12:58 (verbatim): Descartes's dualism is actually not a dualism about substances outside of ourselves connections: - builds on Christianity (`christianity`): descartes set out to prove the soul immaterial, as the church asked of christian philosophers. confidence: highly likely, [source](https://web.viu.ca/johnstoi/descartes/descartesmeditations.htm) - leads to Occasionalism (`occasionalism`): a fix for how mind and body could interact. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Occasionalism) - challenged by Empiricism: Locke and Hume (`empiricism`): experience instead of innate ideas. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Innatism) - challenged by Gottfried Wilhelm Leibniz (`leibniz`): leibniz rejects interaction between two substances. confidence: highly likely, [source](https://plato.stanford.edu/entries/leibniz-mind/) - challenged by La Mettrie: Man a Machine (`lamettrie`): la mettrie extends descartes' animal-machine to humans and drops the soul. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Man_a_Machine) ### Meditations on First Philosophy by René Descartes Descartes doubts everything he can and concludes that he exists as a thinking thing, distinct from his body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Meditations_on_First_Philosophy) ### Mind-body dualism by René Descartes Descartes divided reality into extended stuff (res extensa) and thinking stuff (res cogitans), which leaves the puzzle of how the two interact. That puzzle set the agenda for later discussion of the mind-body relation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Mind%E2%80%93body_problem) ## Occasionalism 1674, philosophy. school of thought: rationalism and idealism. id: `occasionalism` by Nicolas Malebranche A post-Cartesian doctrine that mind and body never directly interact; God causes the matching events on each side on every occasion. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Occasionalism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/occasionalism/) start here, mentioned by joscha bach in: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha) connections: - builds on René Descartes (`descartes`): a fix for how mind and body could interact. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Occasionalism) ## Empiricism: Locke and Hume 1690–1748, philosophy. school of thought: materialism and mechanism. id: `empiricism` ### Empiricism vs rationalism A long dispute about the source of knowledge: experience of the world (empiricists such as Locke and Hume) or reasoning from first principles (rationalists such as Descartes and Leibniz). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Rationalism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/rationalism-empiricism/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=273), from 4:35 (verbatim): the first rule of epistemology says that the confidence in the belief must equal the weight of the evidence supporting it. connections: - argues against René Descartes (`descartes`): experience instead of innate ideas. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Innatism) - leads to Idealism (`idealism`): berkeley: to be is to be perceived. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Empiricism) - leads to Immanuel Kant (`kant`): hume woke kant from his 'dogmatic slumber'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Immanuel_Kant) ### Locke on consciousness by John Locke Locke defined consciousness as the perception of what passes in one's own mind, and tied personal identity to this continuity of awareness and memory. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Consciousness) ### Hume: proportion belief to evidence by David Hume A basic rule of rational thinking, from Hume's Enquiry: hold each belief only as strongly as the evidence warrants. confidence: highly likely, [source](https://plato.stanford.edu/entries/hume/) ## Idealism 1710 onward, philosophy. school of thought: rationalism and idealism. id: `idealism` The view that reality is fundamentally mental and the physical world is an appearance within mind. Bernardo Kastrup's 'analytic idealism' is a current version. confidence: likely, [source](https://en.wikipedia.org/wiki/Idealism). Idealism is broader; only subjective forms specifically treat the external world as appearance within mind. start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/idealism/) start here, joscha bach on it: [Joscha Bach Λ John Vervaeke: Mind, Idealism, Computation (Theories of Everything)](https://www.youtube.com/watch?v=rK7ux_JhHM4&t=2333s), from 38:55 (verbatim): the idealists have to explain why this dream looks in many parts so damn regular connections: - builds on Empiricism: Locke and Hume (`empiricism`): berkeley: to be is to be perceived. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Empiricism) - leads to G. W. F. Hegel (`hegel`): spirit coming to know itself. confidence: highly likely, [source](https://en.wikipedia.org/wiki/German_idealism) - related to Donald Hoffman: interface theory of perception (`hoffman`): hoffman grants the resemblance but says his theory is not idealism. confidence: highly likely, [source](https://sites.socsci.uci.edu/~ddhoff/MindBody) ## Gottfried Wilhelm Leibniz 1714, philosophy. school of thought: rationalism and idealism. id: `leibniz` ### Leibniz German polymath who co-invented calculus and dreamed of a universal symbolic language in which disputes could be settled by calculation ('let us calculate'). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gottfried_Wilhelm_Leibniz) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/leibniz/) start here, joscha bach on it: [Machine Learning Street Talk #82 – Dr. Joscha Bach: Digital Physics, DL and Consciousness [UNPLUGGED]](https://www.youtube.com/watch?v=LgwjcqhkOA4&t=3654s), from 1:00:56 (verbatim): leibniz says this idea of imagine you could have a Mill and this this is your mind connections: - argues against René Descartes (`descartes`): leibniz rejects interaction between two substances. confidence: highly likely, [source](https://plato.stanford.edu/entries/leibniz-mind/) - leads to Immanuel Kant (`kant`): the rationalist side kant tries to reconcile. confidence: highly likely, [source](https://plato.stanford.edu/entries/kant-leibniz/) - leads to Gottlob Frege (`frege`): leibniz's dream of a calculus of thought. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Begriffsschrift) - related to Searle: the Chinese room (`searle`): walking around inside the machine and finding no understanding. confidence: highly likely, [source](https://plato.stanford.edu/entries/chinese-room/) - related to Chalmers: the hard problem (`chalmers`): the mill argument is cited as an early version of the hard problem. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) ### Leibniz's mill by Gottfried Wilhelm Leibniz A thought experiment from the Monadology: imagine a thinking machine enlarged to the size of a mill so you can walk inside. You would see only parts pushing each other and nothing that explains perception. Leibniz concluded that mechanism cannot explain mind. confidence: highly likely, [source](https://plato.stanford.edu/entries/leibniz-mind/) ## La Mettrie: Man a Machine 1747, philosophy. school of thought: materialism and mechanism. id: `lamettrie` by Julien Offray de La Mettrie A scandalous Enlightenment pamphlet arguing that humans are entirely physical machines, not different in kind from animals, and that mind is a function of organised matter. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Julien_Offray_de_La_Mettrie) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Julien_Offray_de_La_Mettrie) start here, joscha bach on it: [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience) (paraphrase): Paraphrase (garbled as 'Julian of f lri' and 'L machine'): La Mettrie had the insight that humans can be understood as machines and our motivation might be cybernetic, composed of opposing springs. connections: - argues against René Descartes (`descartes`): la mettrie extends descartes' animal-machine to humans and drops the soul. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Man_a_Machine) - leads to Materialism and physicalism (`physicalism`): man a machine. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Materialism) - leads to Joscha Bach (`bach`): computationalism as the heir of mechanism. confidence: highly likely, [source](https://media.ccc.de/v/37c3-12167-synthetic_sentience) ## The Enlightenment 18th c., philosophy. school of thought: rationalism and idealism. id: `enlightenment` The 18th-century European movement that put reason, evidence and individual autonomy above religious and traditional authority. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Age_of_Enlightenment) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Age_of_Enlightenment) start here, joscha bach on it: [Computational Meta-Psychology (32C3)](https://media.ccc.de/v/32c3-7483-computational_meta-psychology#t=2975), from 49:37 (verbatim): people have invented rationalism and enlightenment, basically to act as immunization against mind viruses. connections: - related to Immanuel Kant (`kant`): kant's 'dare to know'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/What_Is_Enlightenment%3F) - leads to Utilitarianism (`utilitarianism`): ethics as calculation. confidence: likely. Bentham systematized utility and felicific calculation, but the broad Enlightenment-to-utilitarianism link is not explicitly stated on the checked page. ## Immanuel Kant 1781, philosophy. school of thought: ethics, society and AI risk. id: `kant` ### Kant German philosopher who tried to ground morality in reason alone: act only on rules you could will to be universal laws, regardless of culture or consequences. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Categorical_imperative) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/kant-moral/) start here, mentioned by joscha bach in: [The California Institute for Machine Consciousness Research Program Whitepaper (Sorensen, Hildebrandt-Harangozó, de Kerhuelvez, Rutt, Bach)](https://cimc.ai/publications/research-program-whitepaper/) connections: - builds on Empiricism: Locke and Hume (`empiricism`): hume woke kant from his 'dogmatic slumber'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Immanuel_Kant) - builds on Gottfried Wilhelm Leibniz (`leibniz`): the rationalist side kant tries to reconcile. confidence: highly likely, [source](https://plato.stanford.edu/entries/kant-leibniz/) - related to The Enlightenment (`enlightenment`): kant's 'dare to know'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/What_Is_Enlightenment%3F) - leads to G. W. F. Hegel (`hegel`): german idealism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/German_idealism) - leads to Phenomenology (`husserl`): the mind's contribution to experience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Edmund_Husserl) ### Critique of Pure Reason by Immanuel Kant Kant asks what the mind must supply for experience to be possible at all. His answer is that space, time and basic concepts such as cause are forms the mind brings to experience, so we know the world as it appears to us. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Critique_of_Pure_Reason) ## Utilitarianism 1789, society. school of thought: ethics, society and AI risk. id: `utilitarianism` by Jeremy Bentham; John Stuart Mill The ethical theory that the right action is the one producing the greatest total wellbeing, in principle a calculus over pleasures and pains. It was developed by Jeremy Bentham and John Stuart Mill and underlies much of effective altruism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Utilitarianism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/utilitarianism-history/) start here, joscha bach on it: [Joscha Bach PUSHES Roman Yampolskiy: Only ASI Can Save Us! (Dr Brian Keating)](https://www.youtube.com/watch?v=ayW7SDyAjZo&t=3301s), from 55:03 (verbatim): I think utilitarianism is an attempt to repair morality in the rationalist age, but I don't think it's a very successful one. connections: - builds on The Enlightenment (`enlightenment`): ethics as calculation. confidence: likely. Bentham systematized utility and felicific calculation, but the broad Enlightenment-to-utilitarianism link is not explicitly stated on the checked page. - challenged by Joscha Bach (`bach`): values derived from how agents work, not a pleasure calculus. confidence: highly likely, [source](https://media.ccc.de/v/37c3-12167-synthetic_sentience) ## G. W. F. Hegel 1807, philosophy. school of thought: rationalism and idealism. id: `hegel` by G. W. F. Hegel German idealist who saw history as the process by which Spirit (Geist) comes to know itself. Progress, for him, happens through contradictions that are overcome at a higher level. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Georg_Wilhelm_Friedrich_Hegel) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/hegel/) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM) (verbatim): Another perspective on the LLM is that it's basically a weltgeist. That's a notion of Hegel. connections: - builds on Idealism (`idealism`): spirit coming to know itself. confidence: highly likely, [source](https://en.wikipedia.org/wiki/German_idealism) - builds on Immanuel Kant (`kant`): german idealism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/German_idealism) ## Cell theory 1839, physics and life. school of thought: evolution, biology and life. id: `celltheory` by Matthias Schleiden; Theodor Schwann The discovery that all living things are made of cells and that the cell is the basic unit of life. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cell_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Cell_theory) start here, mentioned by joscha bach in: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=1344), from 22:26 ## Charles Darwin 1859, physics and life. school of thought: evolution, biology and life. id: `darwin` by Charles Darwin Darwin's theory of evolution by natural selection explained the design of living things without a designer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Teleological_argument) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Teleological_argument) start here, mentioned by joscha bach in: [Singularity.FM – Joscha Bach: We need to understand the nature of AI to understand who we are](https://www.youtube.com/watch?v=sIKbp3KcS8A) connections: - leads to Dawkins: selfish genes and memes (`dawkins`): the gene's-eye view of selection. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Selfish_Gene) - leads to Neural Darwinism (`edelman`): selection applied to groups of neurons. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Neural_Darwinism) - leads to Multilevel selection (`groupsel`): selection at several levels. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Group_selection) ## Du Bois-Reymond: Ignorabimus 1872, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `dubois` The physiologist du Bois-Reymond declared that science will never know what matter and force are, or how matter gives rise to consciousness: 'ignorabimus', we shall never know. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ignoramus_et_ignorabimus) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Ignoramus_et_ignorabimus) start here, mentioned by joscha bach in: [Machine Dreams (German-language lecture, Dresden / Datenspuren 2017)](https://c3d2.de/media/20170930-TSD-When_Machines_Are_Dreaming-Joscha_Bach.transcript.html) connections: - related to Mysterianism (`mysterianism`): the ignorabimus is seen as anticipating mysterianism, though the two views differ. confidence: highly likely, [source](https://www.cambridge.org/core/journals/science-in-context/article/reconsidering-the-ignorabimus-du-boisreymond-and-the-hard-problem-of-consciousness/8C19E49A253DD5011482AD4252FF563D) ## Cantor's set theory 1874, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `cantor` by Georg Cantor Cantor showed that infinite sets come in different sizes: the set of all subsets of a set is always larger than the set itself. Applied to 'the set of all sets' this produces contradictions. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cantor%27s_theorem) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/set-theory/) start here, mentioned by joscha bach in: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=743), from 12:25 connections: - challenged by Constructive mathematics and intuitionism (`brouwer`): intuitionism rejects completed infinities. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Intuitionism) - leads to Russell and Whitehead (`russell`): naive set theory runs into russell's paradox. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Russell%27s_paradox) - challenged by Joscha Bach (`bach`): no completed infinities. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) ## Epiphenomenalism 1874, philosophy. school of thought: materialism and mechanism. id: `epiphenomenalism` The view that conscious experience is a by-product of brain activity with no causal influence on the physical world, like the steam-whistle of a locomotive. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Epiphenomenalism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/epiphenomenalism/) start here, joscha bach on it: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha) (paraphrase): Paraphrase (subtitle: 'This is called epiphenomenal, and it's a bit inconvenient'): if mind reads from but cannot write to the universe, experience is causally irrelevant. connections: - related to Chalmers: the hard problem (`chalmers`): experience that does no causal work. confidence: highly likely, [source](https://plato.stanford.edu/entries/epiphenomenalism/) ## Franz Brentano 1874, psychology and mind. school of thought: rationalism and idealism. id: `brentano` German-born philosopher-psychologist who taught in Vienna and revived the idea of intentionality: mental states are always 'about' something. He taught Husserl and Freud. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Franz_Brentano) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/brentano/) start here, mentioned by joscha bach in: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) connections: - leads to Phenomenology (`husserl`): intentionality becomes phenomenology. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Franz_Brentano) - related to Freud and psychoanalysis (`freud`): freud attended brentano's lectures. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Franz_Brentano) ## Emergence 1875, philosophy. school of thought: cybernetics, systems and self-organisation. id: `emergence` by G. H. Lewes; British emergentists The idea that wholes can have properties their parts lack, as wetness belongs to water but not to a single molecule. The disputed question is whether emergent properties are merely convenient descriptions or genuinely new causes. confidence: highly likely, [source](https://iep.utm.edu/emergence/) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/properties-emergent/) start here, joscha bach on it: [Joscha Bach - Why Your Thoughts Aren't Yours. (Machine Learning Street Talk)](https://www.youtube.com/watch?v=3MkJEGE9GRY&t=1024s), from 17:06 (verbatim): I don't like the word of emergence because it gives people an intuition that there is this magical thing ## Early experimental psychology 1879, psychology and mind. school of thought: psychology and development. id: `wundt` In the 19th century Wilhelm Wundt, Gustav Fechner and Hermann von Helmholtz turned the study of mind into a laboratory science, measuring sensation, attention and reaction and proposing theories of emotion and will. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Wilhelm_Wundt) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/History_of_psychology) start here, mentioned by joscha bach in: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) connections: - related to William James: The Principles of Psychology (`james`): psychology becomes a laboratory science. confidence: highly likely, [source](https://psychclassics.yorku.ca/Special/Institutions/labsintro.htm) - challenged by Behaviorism (`skinner`): watson's behaviorism rejected introspective methods; skinner's later treated introspection as behaviour. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Behaviorism) ## Gottlob Frege 1879, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `frege` by Gottlob Frege German logician who invented modern predicate logic and tried to derive arithmetic from pure logic. Russell found a paradox in his system, which set off the crisis in the foundations of mathematics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gottlob_Frege) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/frege/) start here, mentioned by joscha bach in: [Cyber Animism (talk hosted by Michael Levin's center, Tufts)](https://www.youtube.com/watch?v=YZl4zom3q2g) connections: - builds on Gottfried Wilhelm Leibniz (`leibniz`): leibniz's dream of a calculus of thought. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Begriffsschrift) - leads to Russell and Whitehead (`russell`): russell finds the paradox in frege's system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Russell%27s_paradox) - leads to Ludwig Wittgenstein (`wittgenstein`): logic as the form of language. confidence: highly likely, [source](https://plato.stanford.edu/entries/wittgenstein/) - related to Alfred Tarski (`tarski`): frege's semantics for his formal language anticipates tarski's theory of truth. confidence: highly likely, [source](https://jhaponline.org/jhap/article/download/3970/3666) ## Peano axioms 1889, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `peano` by Giuseppe Peano A handful of axioms that define the natural numbers and arithmetic: there is a zero, every number has a successor, and induction holds. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Peano_axioms) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Peano_axioms) start here, joscha bach on it: [Singularity.FM – Joscha Bach on AI, Cosmology, Existence and the Bible](https://www.youtube.com/watch?v=vyGP8LpsDok) (paraphrase): Bach: once you have Peano's axioms (captioned 'pianos axioms') you have a Turing-complete system from which a universe could be built. connections: - leads to Gödel's incompleteness theorems (`goedel`): arithmetic is what incompleteness is about. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) ## William James: The Principles of Psychology 1890, psychology and mind. school of thought: psychology and development. id: `james` by William James A groundbreaking textbook by the man often called the father of American psychology, known for its descriptions of the stream of consciousness, attention and the self. confidence: highly likely, [source](https://en.wikipedia.org/wiki/William_James) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/The_Principles_of_Psychology) start here, mentioned by joscha bach in: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) connections: - related to Early experimental psychology (`wundt`): psychology becomes a laboratory science. confidence: highly likely, [source](https://psychclassics.yorku.ca/Special/Institutions/labsintro.htm) - leads to Antonio Damasio (`damasio`): emotion as perception of the body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Antonio_Damasio) ## Constructive mathematics and intuitionism 1908, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `brouwer` by L. E. J. Brouwer (intuitionism) A school of mathematics that accepts an object only if you can give a procedure to construct it, rejecting proofs that merely assert existence and treating infinity as a never-finished process. confidence: highly likely, [source](https://plato.stanford.edu/entries/intuitionism/) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/intuitionism/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=897), from 14:59 (verbatim): Computation is constructive mathematics. The tiny, neglected subset of mathematics, where you have to show the money. connections: - argues against Cantor's set theory (`cantor`): intuitionism rejects completed infinities. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Intuitionism) - related to Hilbert's programme (`hilbert`): the formalism versus intuitionism dispute. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Brouwer%E2%80%93Hilbert_controversy) - leads to Joscha Bach (`bach`): only what can be constructed exists. confidence: possible. Bach strongly advocates constructive mathematics, but no opened source documented a specific debt to Brouwer or intuitionism. ## Russell and Whitehead 1901–1913, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `russell` ### Russell's paradox by Bertrand Russell Consider the set of all sets that do not contain themselves. Does it contain itself? If it does, it doesn't; if it doesn't, it does. The paradox showed that naive set theory is inconsistent. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Russell%27s_paradox) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/russell-paradox/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=817), from 13:39 (verbatim): Does the set of the sets that doesn't contain themselves contain itself? If it does, it doesn't, and if it doesn't, it does. connections: - builds on Cantor's set theory (`cantor`): naive set theory runs into russell's paradox. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Russell%27s_paradox) - builds on Gottlob Frege (`frege`): russell finds the paradox in frege's system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Russell%27s_paradox) - leads to Ludwig Wittgenstein (`wittgenstein`): wittgenstein was russell's student. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ludwig_Wittgenstein) - leads to Gödel's incompleteness theorems (`goedel`): 'on formally undecidable propositions of principia mathematica'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) ### Principia Mathematica by Bertrand Russell, Alfred North Whitehead A three-volume attempt to derive all of mathematics from logic, using a hierarchy of 'types' to block self-referential paradoxes such as the set of all sets that do not contain themselves. Gödel's 1931 incompleteness theorems were proved about this very system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Principia_Mathematica) ## Phenomenology 1913, philosophy. school of thought: rationalism and idealism. id: `husserl` by Edmund Husserl Husserl proposed studying experience exactly as it is given. Each act of consciousness (noesis) is directed at an object as it is meant (noema), and philosophy should 'bracket' the taken-for-granted existence of the outside world (epoché) to describe how things appear. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Edmund_Husserl) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/husserl/) start here, joscha bach on it: [Machine Dreams (German-language lecture, Dresden / Datenspuren 2017)](https://c3d2.de/media/20170930-TSD-When_Machines_Are_Dreaming-Joscha_Bach.transcript.html) (paraphrase): Paraphrase (German): we could begin our journey with Edmund Husserl ... what we perceive of the world are appearances, phenomena; the process of turning them into ideas he calls noesis. connections: - builds on Immanuel Kant (`kant`): the mind's contribution to experience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Edmund_Husserl) - builds on Franz Brentano (`brentano`): intentionality becomes phenomenology. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Franz_Brentano) - leads to Embodied and enactive cognition (`embodied`): phenomenology of the lived body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Embodied_cognition) ## Hilbert's programme 1920s, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `hilbert` by David Hilbert Hilbert asked mathematicians to put all of mathematics into one formal axiom system and to prove, using only simple finite reasoning, that the system is consistent; he expected it also to be complete, so that every true statement could be derived. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hilbert%27s_program) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/hilbert-program/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=830), from 13:52 (verbatim): So David Hilbert, who was some kind of a community manager back then, said: "Guys, fix this! connections: - related to Constructive mathematics and intuitionism (`brouwer`): the formalism versus intuitionism dispute. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Brouwer%E2%80%93Hilbert_controversy) - challenged by Gödel's incompleteness theorems (`goedel`): incompleteness ends the hope of a complete, provably consistent system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) - leads to Alan Turing and computability (`turing`): turing answers hilbert's decision problem. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Turing%27s_proof) ## Ludwig Wittgenstein 1921–1953, philosophy. school of thought: logic and the foundations of mathematics. id: `wittgenstein` ### Wittgenstein Austrian philosopher. His early work tried to define exactly what can be said in a logical language; his later work argued that meaning comes from how language is used in life. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ludwig_Wittgenstein) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/wittgenstein/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=1015), from 16:57 (verbatim): It's an attempt by a guy who was basically a coder, an AI scientist, to reverse engineer the language of his own thinking. connections: - builds on Gottlob Frege (`frege`): logic as the form of language. confidence: highly likely, [source](https://plato.stanford.edu/entries/wittgenstein/) - builds on Russell and Whitehead (`russell`): wittgenstein was russell's student. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ludwig_Wittgenstein) - related to Joscha Bach (`bach`): the tractatus as a programme for a precise language. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) ### Tractatus Logico-Philosophicus by Ludwig Wittgenstein A short, austere book claiming that the world is the totality of facts, that language mirrors facts through logical form, and that what can be said at all can be said clearly, and what cannot be spoken of must be passed over in silence. Wittgenstein wrote it largely in the trenches of the First World War and later rejected much of it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Tractatus_Logico-Philosophicus) ### Philosophical Investigations by Ludwig Wittgenstein Wittgenstein's later work, which abandons the idea of one exact logical language. Meaning comes from how words are used in many overlapping 'language games'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Philosophical_Investigations) ## Freud and psychoanalysis 1923, psychology and mind. school of thought: psychology and development. id: `freud` by Sigmund Freud Freud pictured the mind as three agencies: the id (drives), the ego (the self that negotiates with reality) and the superego (internalised rules). Psychoanalysis interprets behaviour as the outcome of their conflicts. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Id,_ego_and_superego) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Id,_ego_and_superego) start here, joscha bach on it: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) (verbatim): And psycho analysis is not a natural science, but it’s a hermeneutic science. You cannot disprove it scientifically. connections: - related to Franz Brentano (`brentano`): freud attended brentano's lectures. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Franz_Brentano) - related to System 1 and System 2 (`kahneman`): most of the mind is not the conscious part. confidence: highly likely, [source](https://psyche.co/ideas/the-id-and-the-nudge-where-freud-meets-behavioural-economics) ## Qualia 1929–1982, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `qualia` by C. I. Lewis (term) The 'what it is like' qualities of experience: the redness of red, the painfulness of pain. Many philosophers think they resist physical explanation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Qualia) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/qualia/) start here, joscha bach on it: [Why consciousness is software, and software is spirit | Joscha Bach | FULL INTERVIEW (Institute of Art and Ideas)](https://www.youtube.com/watch?v=E361FZ_50oo&t=652s), from 10:54 (verbatim): you can deconstruct your qualia you can notice that they're just representing connections: - related to Nagel: What Is It Like to Be a Bat? (`nagel`): the felt qualities of experience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Qualia) - leads to The explanatory gap (`levine`): the explanatory gap. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Qualia) ### Mary's room (Epiphenomenal Qualia) by Frank Jackson The 'Mary' argument. A scientist who knows all the physical facts about colour but has never seen colour seems to learn something new when she first sees red, which Jackson took to show that physical facts are not all the facts. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Knowledge_argument) ## Gödel's incompleteness theorems 1931, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `goedel` by Kurt Gödel Gödel proved that any consistent formal system whose axioms can be listed mechanically and that is rich enough to do arithmetic contains true statements it cannot prove, and cannot prove its own consistency. He did it by encoding statements about the system as numbers inside the system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/goedel-incompleteness/) start here, joscha bach on it: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=4278s), from 1:11:20 (verbatim): And what Gödel has discovered is that these bugs cannot be eliminated except by redefining the programming language of mathematics. connections: - builds on Peano axioms (`peano`): arithmetic is what incompleteness is about. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) - builds on Russell and Whitehead (`russell`): 'on formally undecidable propositions of principia mathematica'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) - argues against Hilbert's programme (`hilbert`): incompleteness ends the hope of a complete, provably consistent system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems) - related to Alan Turing and computability (`turing`): two distinct limits on what formal procedures can decide. confidence: highly likely, [source](https://www.cs.ox.ac.uk/activities/ieg/e-library/sources/tp2-ie.pdf) - leads to Hofstadter: strange loops (`hofstadter`): self-reference as the root of the self. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Douglas_Hofstadter) - leads to Penrose and Hameroff: quantum consciousness (`penrose`): penrose's argument that minds are not algorithms. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Orchestrated_objective_reduction) - leads to Joscha Bach (`bach`): bach reads incompleteness as the failure of classical mathematics. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/) ## Alfred Tarski 1933, mathematics and logic. school of thought: logic and the foundations of mathematics. id: `tarski` Polish logician who gave the standard formal definition of truth for logical languages and put semantics on a rigorous footing. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Alfred_Tarski) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/tarski-truth/) start here, mentioned by joscha bach in: [Cyber Animism (talk hosted by Michael Levin's center, Tufts)](https://www.youtube.com/watch?v=YZl4zom3q2g) connections: - related to Gottlob Frege (`frege`): frege's semantics for his formal language anticipates tarski's theory of truth. confidence: highly likely, [source](https://jhaponline.org/jhap/article/download/3970/3666) ## Alan Turing and computability 1936–1950, mathematics and logic. school of thought: computation, information and digital physics. id: `turing` ### Alan Turing English mathematician who defined computation with his abstract machine, helped found AI with the 1950 imitation-game paper, and late in life modelled how chemical patterns form in biology. During the Second World War he led the Bletchley Park section that broke German naval Enigma and designed the Bombe machine used to do it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Alan_Turing) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Alan_Turing) start here, joscha bach on it: [Lex Fridman Podcast #101 – Joscha Bach: Artificial Consciousness and the Nature of Reality](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality) (verbatim): I think that Turing could see that the solution is to understand that the question mathematics was computation all along connections: - builds on Hilbert's programme (`hilbert`): turing answers hilbert's decision problem. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Turing%27s_proof) - related to Gödel's incompleteness theorems (`goedel`): two distinct limits on what formal procedures can decide. confidence: highly likely, [source](https://www.cs.ox.ac.uk/activities/ieg/e-library/sources/tp2-ie.pdf) - related to Lambda calculus and combinators (`lambda`): church and turing: equivalent definitions of computation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis) - leads to John von Neumann and the first computers (`vonneumann`): the stored-program computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Von_Neumann_architecture) - leads to Turing's morphogenesis (`morpho`): turing's last major work. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Alan_Turing) - leads to The founding of AI (`dartmouth`): 'can machines think?'. confidence: highly likely, [source](https://ai100.stanford.edu/reports/2016-report/appendix-i-short-history-ai) - leads to Functionalism and computationalism (`functionalism`): turing's work inspired putnam's machine-state functionalism. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) - related to Quantum computing (`quantumcomp`): a faster, not a more powerful, computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_computing) - related to Hypercomputation (`hypercomputation`): what would lie beyond turing machines. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hypercomputation) - leads to Joscha Bach (`bach`): the church-turing thesis. confidence: highly likely, [source](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) - challenged by CIMC: the Machine Consciousness Hypothesis (`cimc`): behaviour cannot settle whether a system is conscious. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) ### Church–Turing thesis by Alonzo Church; Alan Turing Everything that can be effectively computed can be computed by a Turing machine, so all sufficiently powerful computers are equivalent in what they can compute. It is a thesis rather than a theorem, because 'effectively computable' is an informal notion. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis) ### The Turing machine by Alan Turing An imaginary device with a tape, a read-write head and a table of rules. Despite its simplicity it can carry out any calculation that any computer can, which makes it the standard definition of computation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Turing_machine) ### Halting problem by Alan Turing Turing showed that no general procedure can decide, for every program, whether it will eventually stop or run forever. Some well-posed questions therefore have no computable answer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Halting_problem) ### Post machine by Emil Post A model of computation proposed by Emil Post in the same year as Turing's, built from very simple instructions on a tape of marked and unmarked boxes. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Post%E2%80%93Turing_machine) ### Computing Machinery and Intelligence (the Turing test) by Alan Turing The paper that asks 'Can machines think?' and replaces the question with an imitation game, later called the Turing test: a machine passes if a judge cannot reliably tell its conversation from a human's. It also answers a list of objections, including one from telepathy. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Computing_Machinery_and_Intelligence) ## Lambda calculus and combinators 1924–1936, mathematics and logic. school of thought: computation, information and digital physics. id: `lambda` ### Lambda calculus by Alonzo Church Different formal definitions of what it means to compute (Turing's machines, Church's lambda calculus, Schönfinkel's combinators) turned out to be equivalent. The Church-Turing thesis says they capture everything that can be computed at all. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/church-turing/) start here, joscha bach on it: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) (verbatim): It's computation without wheels. It was invented basically at the same time as the Turing machine. connections: - related to Alan Turing and computability (`turing`): church and turing: equivalent definitions of computation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis) - leads to Structure and Interpretation of Computer Programs (`sicp`): lisp and scheme. confidence: highly likely, [source](https://sarabander.github.io/sicp/html/1_002e3.xhtml) ### Moses Schönfinkel Logician who invented combinatory logic, showing that all functions can be built from a couple of basic combinators, a precursor of the lambda calculus and functional programming. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Moses_Sch%C3%B6nfinkel) ## Behaviorism 1938, psychology and mind. school of thought: psychology and development. id: `skinner` by B. F. Skinner The school that restricted psychology to observable behaviour and treated talk of inner mental representations as unscientific. Skinner explained behaviour through its consequences: actions that are reinforced become more frequent. confidence: highly likely, [source](https://plato.stanford.edu/entries/behaviorism/) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/behaviorism/) start here, joscha bach on it: [Self Models of Loving Grace (38C3)](https://media.ccc.de/v/38c3-self-models-of-loving-grace) (paraphrase): his psychology professor claimed dogs have no mental states, only stimulus-reaction patterns; Bach replies that a lookup table would be far harder than mental states. connections: - argues against Early experimental psychology (`wundt`): watson's behaviorism rejected introspective methods; skinner's later treated introspection as behaviour. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Behaviorism) - challenged by Noam Chomsky (`chomsky`): chomsky's review of 'verbal behavior'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Verbal_Behavior) ## Maslow: A Theory of Human Motivation 1943, psychology and mind. school of thought: psychology and development. id: `maslow` by Abraham Maslow Maslow's ranking of human needs, from bodily needs up to self-fulfilment. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Maslow%27s_hierarchy_of_needs) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Maslow%27s_hierarchy_of_needs) connections: - related to Joscha Bach (`bach`): maslow's hierarchy of needs reread as a hierarchy of purposes. confidence: highly likely, [source](http://bach.ai/pub/Attention%20based%20learning%20as%20a%20foundation%20for%20consciousness%20Common%20Model%20Bach.pdf) ## McCulloch–Pitts neurons 1943, computing and AI. school of thought: neural networks and machine learning. id: `mcculloch` by McCulloch & Pitts; Rumelhart & McClelland McCulloch and Pitts showed in 1943 that networks of simple threshold units can compute logical functions. In 1986 Rumelhart and McClelland's 'Parallel Distributed Processing' revived neural networks as models of cognition. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Connectionism) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Connectionism) start here, mentioned by joscha bach in: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) connections: - related to Cybernetics (`cybernetics`): the macy conferences. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Macy_conferences) - leads to Marvin Minsky (`minsky`): neural nets, then minsky's critique of them. confidence: highly likely, [source](https://bnaskrecki.faculty.wmi.amu.edu.pl/nnets/_build/html/part1_origins/ch03_timeline.html) - leads to Connectionism and backpropagation (`pdp`): connectionism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Connectionism) ## Schrödinger: life as negentropy 1944, physics and life. school of thought: evolution, biology and life. id: `schroedinger` by Erwin Schrödinger In 'What is Life?' Schrödinger suggested that organisms stay ordered by feeding on 'negative entropy': they export disorder to their environment. confidence: highly likely, [source](https://en.wikipedia.org/wiki/What_Is_Life%3F) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/What_Is_Life%3F) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams) (paraphrase): Paraphrase (subtitle 'Nick entropy'): a cell is the smallest universal machine we know that can extract negentropy over a wide range of environments. connections: - leads to Autocatalytic networks and autopoiesis (`autopoiesis`): what keeps a living system ordered. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Autopoiesis) - leads to Predictive processing and the free energy principle (`friston`): staying alive as resisting entropy. confidence: highly likely, [source](https://pmc.ncbi.nlm.nih.gov/articles/PMC5857288/) ## John von Neumann and the first computers 1837–1956, mathematics and logic. school of thought: computation, information and digital physics. id: `vonneumann` ### John von Neumann Hungarian-American polymath who shaped game theory, quantum mechanics and the stored-program computer architecture that bears his name. He also designed self-reproducing automata, an early formal model of life. confidence: highly likely, [source](https://en.wikipedia.org/wiki/John_von_Neumann) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/John_von_Neumann) start here, joscha bach on it: [Lex Fridman Podcast #101 – Joscha Bach: Artificial Consciousness and the Nature of Reality](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality) (verbatim): So in a way, it's tempting to think of the cell as a von Neumann probe. connections: - builds on Alan Turing and computability (`turing`): the stored-program computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Von_Neumann_architecture) - leads to Digital physics (`digital`): the universe as a cellular automaton. confidence: highly likely, [source](https://people.idsia.ch/~juergen/digitalphysics.html) - leads to Conway's Game of Life (`conway`): cellular automata. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Conway%27s_Game_of_Life) - leads to Technological singularity (`singularity`): Ulam's report of von Neumann's accelerating-change conversation. confidence: highly likely, [source](https://doi.org/10.1090/s0002-9904-1958-10189-5) - leads to Best-effort computing (`ackley`): reliable computing from unreliable parts. confidence: highly likely, [source](https://cacm.acm.org/opinion/beyond-efficiency/) ### Babbage, Zuse, von Neumann Babbage designed a mechanical general-purpose computer in the 1830s; Zuse built the first working programmable computers in Germany around 1941; von Neumann's 1945 report described the stored-program architecture still used today, building on designs by Eckert and Mauchly. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Von_Neumann_architecture) ### Reliable organisms from unreliable components by John von Neumann Showed how to build reliable computation out of noisy, failure-prone parts by using redundancy. confidence: highly likely, [source](https://static.ias.edu/pitp/archive/2012files/Probabilistic_Logics.pdf) ## Claude Shannon and information theory 1948–1950, mathematics and logic. school of thought: computation, information and digital physics. id: `shannon` ### Information theory by Claude Shannon Shannon defined information as reduction of uncertainty, measurable in bits, and showed how much of it can be sent over a noisy channel. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Information_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Information_theory) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=3238), from 54:00 (verbatim): it's not so much important to look at Shannon information. I look at what we need to describe the evolution of a system. connections: - leads to The founding of AI (`dartmouth`): shannon co-organised the workshop. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) - leads to Solomonoff induction (`solomonoff`): information and compression as a theory of induction. confidence: highly likely, [source](https://raysolomonoff.com/publications/barc97.pdf) - leads to Ian Witten: cognition as compression (`witten`): prediction and compression. confidence: highly likely, [source](https://doi.org/10.1145/76894.76896) ### Computer chess by Claude Shannon Shannon's paper 'Programming a Computer for Playing Chess' distinguished between searching all moves to a fixed depth and selectively exploring promising lines. confidence: highly likely, [source](http://archive.computerhistory.org/projects/chess/related%5Fmaterials/text/2-0%20and%202-1.Programming%5Fa%5Fcomputer%5Ffor%5Fplaying%5Fchess.shannon/2-0%20and%202-1.Programming%5Fa%5Fcomputer%5Ffor%5Fplaying%5Fchess.shannon.062303002.pdf) ## Cybernetics 1948, computing and AI. school of thought: cybernetics, systems and self-organisation. id: `cybernetics` by Norbert Wiener and others The mid-century science of control and feedback in animals and machines, born from 1940s exchanges across many disciplines, including early work on neural networks. Norbert Wiener's 1948 book popularised the name; feedback was its central theme, and purposeful behaviour one of its first topics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cybernetics) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Cybernetics) start here, joscha bach on it: [Joscha Bach: Consciousness, Artificial Intelligence, and the Threat of AI Apocalypse (Robinson Erhardt / Robinson's Podcast)](https://www.youtube.com/watch?v=XcNlv9gp20o&t=1360s), from 22:42 (verbatim): we can understand motivation from the perspective of control theory and cybernetics connections: - related to McCulloch–Pitts neurons (`mcculloch`): the macy conferences. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Macy_conferences) - leads to Gaia hypothesis (`gaia`): the planet as a self-regulating system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gaia_hypothesis) - leads to Autocatalytic networks and autopoiesis (`autopoiesis`): second-order cybernetics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Second-order_cybernetics) - leads to Joscha Bach (`bach`): needs as feedback loops. confidence: highly likely, [source](https://www.youtube.com/watch?v=XcNlv9gp20o) ### Project Cybersyn by Stafford Beer Under President Allende, Chile tried to steer its nationalised economy in real time using telex data, computer models and a futuristic operations room. The 1973 coup ended it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Project_Cybersyn) ## Materialism and physicalism ancient; modern form 20th c., philosophy. school of thought: materialism and mechanism. id: `physicalism` ### Physicalism The thesis that everything that exists is physical or depends on the physical, with no extra mental or supernatural ingredients. It is the majority view among philosophers (about 52% in a 2020 survey), though far from unanimous. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Physicalism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/physicalism/) start here, joscha bach on it: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) (verbatim): matter is primary and mind emerges over functionality of matter, which is I think the dominant theory today and usually, we call it physicalism. connections: - builds on La Mettrie: Man a Machine (`lamettrie`): man a machine. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Materialism) - related to Functionalism and computationalism (`functionalism`): functionalism is neutral on substance but attractive to materialists, as an alternative to identity theory. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) ### Direct realism The philosophical position that we perceive the world itself, roughly as it is, rather than an internal representation of it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Direct_and_indirect_realism) ## Turing's morphogenesis 1952, physics and life. school of thought: cybernetics, systems and self-organisation. id: `morpho` by Alan Turing Turing showed that interacting chemicals that diffuse at different rates spontaneously form spots and stripes. Neural cellular automata are a modern descendant: grids of identical cells that learn local rules to grow and repair a pattern. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Turing_pattern) start here, joscha bach on it: [Joscha Bach Λ Ben Goertzel: Conscious Ai, LLMs, AGI (Theories of Everything)](https://www.youtube.com/watch?v=xw7omaQ8SgA&t=5105s), from 1:25:07 (verbatim): when he looked at reaction diffusion patterns as a way to implement computation connections: - builds on Alan Turing and computability (`turing`): turing's last major work. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Alan_Turing) - leads to Neural cellular automata (`nca`): patterns that grow from local rules. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) ## Developmental stage theories 1936–1954, psychology and mind. school of thought: psychology and development. id: `piaget` by Jean Piaget Theories that describe mental growth as a series of stages: Piaget for children's thinking, Loevinger for ego development, Kohlberg for moral reasoning. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Developmental_stage_theories) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Developmental_stage_theories) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - leads to Kegan's stages of adult development (`kegan`): stages continue into adulthood. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Robert_Kegan) - related to Joscha Bach (`bach`): both propose developmental stages; Bach says his levels derive from Kegan, not Piaget. confidence: highly likely, [source](https://joscha.substack.com/p/levels-of-lucidity) ## The Doors of Perception 1954, psychology and mind. school of thought: religion and contemplative traditions. id: `huxley` by Aldous Huxley Huxley's report of taking mescaline, in which he suggests the brain normally filters a much wider field of awareness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Doors_of_Perception) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/The_Doors_of_Perception) connections: - leads to Timothy Leary and the eight-circuit model (`leary`): leary found the book confirmed his psilocybin experience, then befriended huxley. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Doors_of_Perception) ## The founding of AI 1956, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `dartmouth` ### The Dartmouth workshop by McCarthy, Minsky, Rochester, Shannon The 1956 summer meeting at Dartmouth College where the term 'artificial intelligence' was adopted and the field was founded on the bet that every aspect of intelligence can be described precisely enough for a machine to simulate it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Dartmouth_workshop) start here, mentioned by joscha bach in: [Joscha Bach: Consciousness, Artificial Intelligence, and the Threat of AI Apocalypse (Robinson Erhardt / Robinson's Podcast)](https://www.youtube.com/watch?v=XcNlv9gp20o&t=3351s), from 55:53 connections: - builds on Alan Turing and computability (`turing`): 'can machines think?'. confidence: highly likely, [source](https://ai100.stanford.edu/reports/2016-report/appendix-i-short-history-ai) - builds on Claude Shannon and information theory (`shannon`): shannon co-organised the workshop. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) - leads to Marvin Minsky (`minsky`): minsky was a founder of the field. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) - leads to Symbolic AI (`symbolic`): the first paradigm of AI. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) ### John McCarthy by John McCarthy Computer scientist who coined the term 'artificial intelligence', organised the 1956 Dartmouth workshop and invented Lisp. He championed formal logic as the language in which machines should represent knowledge. confidence: highly likely, [source](https://en.wikipedia.org/wiki/John_McCarthy_(computer_scientist)) ## Cortical columns 1957, psychology and mind. school of thought: scientific theories of consciousness. id: `mountcastle` by Vernon Mountcastle The neocortex appears to be built from repeating vertical circuits: minicolumns of about a hundred neurons, bundled into larger columns. Mountcastle proposed that these columns are the cortex's basic, general-purpose computing unit. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cortical_column) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Cortical_column) start here, joscha bach on it: [Computational Meta-Psychology (32C3)](https://media.ccc.de/v/32c3-7483-computational_meta-psychology#t=1272), from 21:14 (verbatim): And I think that nature solution to this is cortical columns. Cortical column is a circuit of between 100 and 400 neurons. connections: - leads to Jeff Hawkins and Numenta (`hawkins`): one cortical algorithm. confidence: highly likely, [source](https://en.wikipedia.org/wiki/On_Intelligence) ## Many-worlds interpretation 1957, physics and life. school of thought: computation, information and digital physics. id: `everett` by Hugh Everett III The reading of quantum mechanics in which the wave function never collapses; every possible outcome happens in a separate branch of the universe. Hugh Everett proposed it in 1957; observers split along with everything else, so each copy sees a single outcome. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Many-worlds_interpretation) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/qm-everett/) start here, mentioned by joscha bach in: [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI&t=1316s), from 21:58 connections: - leads to David Deutsch: constructor theory (`deutsch`): deutsch is the leading many-worlds advocate. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Deutsch) ## Marvin Minsky 1956–1986, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `minsky` Co-founder of the field of artificial intelligence at MIT. In 'The Society of Mind' (1986) and 'The Emotion Machine' (2006) he described the mind as a large crowd of simple cooperating processes, with no single principle behind it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Marvin_Minsky) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Marvin_Minsky) start here, joscha bach on it: [Joscha Bach and Connor Leahy [HQ VERSION] (Machine Learning Street Talk)](https://www.youtube.com/watch?v=Z02Obj8j6FQ&t=3947s), from 1:05:49 (verbatim): he saw the mind as an interaction of agents that have a degree of autonomy and that at the same time have to achieve coherence connections: - builds on McCulloch–Pitts neurons (`mcculloch`): neural nets, then minsky's critique of them. confidence: highly likely, [source](https://bnaskrecki.faculty.wmi.amu.edu.pl/nnets/_build/html/part1_origins/ch03_timeline.html) - builds on The founding of AI (`dartmouth`): minsky was a founder of the field. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) - builds on Pandemonium model (`selfridge`): a crowd of demons becomes a society of mind. confidence: highly likely, [source](https://www.museumofai.org.uk/nodes/pandemonium-architecture) - challenged by Connectionism and backpropagation (`pdp`): connectionism returns after 'perceptrons'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Perceptrons_(book)) - related to Global Workspace Theory (`gwt`): many unconscious specialists. confidence: highly likely, [source](http://scholarpedia.org/article/Global_workspace) - leads to Joscha Bach (`bach`): a society of specialised processes that needs coordinating. confidence: highly likely, [source](https://www.youtube.com/watch?v=Z02Obj8j6FQ) ### The Society of Mind by Marvin Minsky Minsky's picture of the mind as a crowd of simple, mindless agents whose interactions add up to intelligence. The book is written as hundreds of one-page essays, each describing one piece of the machinery. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Society_of_Mind) ### Perceptrons by Marvin Minsky; Seymour Papert A mathematical analysis showing that single-layer neural networks cannot learn certain simple functions. It is often thought to have caused the decline in neural network research in the 1970s and early 1980s, when major funding was hard to find, though how much it was to blame is disputed. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Perceptrons_(book)) ## Noam Chomsky 1957, psychology and mind. school of thought: symbolic AI and cognitive architectures. id: `chomsky` Linguist who argued that language rests on an innate grammar faculty, and a long-standing sceptic that statistical models tell us anything about the mind. His work in the 1950s helped end behaviourism and start cognitive science. confidence: highly likely, [source](https://web.archive.org/web/20220427080655/https://www.britannica.com/biography/Noam-Chomsky) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Noam_Chomsky) start here, joscha bach on it: [Donald Hoffman Λ Joscha Bach: Consciousness, Gödel, Reality (Theories of Everything)](https://www.youtube.com/watch?v=bhSlYfVtgww&t=893s), from 14:55 (verbatim): Noam Chomsky is a mysterious with respect to the mind. connections: - argues against Behaviorism (`skinner`): chomsky's review of 'verbal behavior'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Verbal_Behavior) - leads to The language of thought (`fodor`): innate structure and the language of thought. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Jerry_Fodor) ## Pandemonium model 1959, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `selfridge` by Oliver Selfridge An early model of perception as a crowd of 'demons', each shouting when it detects its pattern, with higher demons listening for the loudest. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Pandemonium_architecture) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Pandemonium_architecture) connections: - leads to Marvin Minsky (`minsky`): a crowd of demons becomes a society of mind. confidence: highly likely, [source](https://www.museumofai.org.uk/nodes/pandemonium-architecture) ## Functionalism and computationalism 1960s, philosophy. school of thought: functionalism and computational theories of mind. id: `functionalism` ### Functionalism by Hilary Putnam and others Functionalism says a mental state is defined by what it does, not what it is made of. Computationalism adds that what minds do can be described as computation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Functionalism_(philosophy_of_mind)) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/functionalism/) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=210), from 3:32 (verbatim): And the answer that I am suggesting, is functionalism which means that indeed we exist only in a dream. connections: - builds on Alan Turing and computability (`turing`): turing's work inspired putnam's machine-state functionalism. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) - related to Materialism and physicalism (`physicalism`): functionalism is neutral on substance but attractive to materialists, as an alternative to identity theory. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) - challenged by Searle: the Chinese room (`searle`): the chinese room. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Chinese_room) - leads to Daniel Dennett (`dennett`): dennett's functionalist account of consciousness. confidence: highly likely, [source](https://dl.tufts.edu/downloads/9306t946z?filename=5t34sw885.pdf) - challenged by Ned Block: access vs phenomenal consciousness (`block`): 'troubles with functionalism'. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) - challenged by Chalmers: the hard problem (`chalmers`): function alone leaves experience unexplained. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) - leads to Joscha Bach (`bach`): the tradition bach places himself in. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) - challenged by Anil Seth: biological naturalism (`seth`): consciousness tied to living bodies. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - challenged by Thagard: energy and substrate independence (`thagard`): energy makes substrate matter. confidence: highly likely, [source](https://www.cambridge.org/core/journals/philosophy-of-science/article/energy-requirements-undermine-substrate-independence-and-mindbody-functionalism/2BB3C2353EFF80F9D5805CDCEA8C3C89) - challenged by Mazviita Chirimuuta (`chirimuuta`): models are simplifications, not the thing. confidence: highly likely, [source](https://www.thebsps.org/reviewofbooks/burnston-on-chirimuuta/) - challenged by Lerchner: The Abstraction Fallacy (`lerchner`): 'the abstraction fallacy'. confidence: highly likely, [source](https://deepmind.google/research/publications/231971/) ### Computationalism, weak and strong Computationalism says minds can be understood as computational systems. Stronger versions identify mental states with computations; pancomputationalism is the separate claim that physical processes throughout the universe are computational. confidence: highly likely, [source](https://doi.org/10.17721/2523-4064.2023/8-1/8) ## Stanisław Lem 1961–1981, fiction. school of thought: fiction. id: `lem` Polish science fiction writer and philosopher of technology. 'Solaris' (1961) is about an alien mind humans cannot understand; 'Golem XIV' (1981) is a set of lectures given by a superhuman computer to humanity. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Stanis%C5%82aw_Lem) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Stanis%C5%82aw_Lem) start here, joscha bach on it: [The age of AGE: Artificial General Experience (Joscha Bach, Substack; text written for Rana Gujral's book 'The AI Instinct')](https://joscha.substack.com/p/the-age-of-age) (paraphrase): 'Lem's Golem computer is a Large Language Model, capable of running circles around the best human scientists'. ### Solaris by Stanisław Lem Novel about scientists on a station above a planet whose ocean is a single vast mind that they can never understand. Stanisław Lem wrote it in 1961; the ocean answers the scientists by materialising figures from their own memories. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Solaris_(novel)) ### Summa Technologiae by Stanisław Lem Lem's non-fiction survey of future technologies, including artificial intelligence, virtual reality and directed evolution. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Summa_Technologiae) ## Solomonoff induction 1964, mathematics and logic. school of thought: neural networks and machine learning. id: `solomonoff` by Ray Solomonoff Inventor of algorithmic probability, a mathematically ideal (though uncomputable) theory of learning by preferring the shortest programs that explain the data. He attended the 1956 Dartmouth workshop and his ideas underlie later formal theories of universal intelligence. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ray_Solomonoff) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Ray_Solomonoff) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=347), from 5:49 (verbatim): you have to find the shortest program among the best programs that predicts your current observation from past observations connections: - builds on Claude Shannon and information theory (`shannon`): information and compression as a theory of induction. confidence: highly likely, [source](https://raysolomonoff.com/publications/barc97.pdf) - leads to Jürgen Schmidhuber (`schmidhuber`): schmidhuber generalises solomonoff's algorithmic probability and universal prediction. confidence: highly likely, [source](https://people.idsia.ch/~juergen/kolmogorov.html) ## Alexey Ivakhnenko 1965, computing and AI. school of thought: neural networks and machine learning. id: `ivakhnenko` Soviet-Ukrainian mathematician who trained multi-layer networks in the 1960s (the Group Method of Data Handling), often cited as the earliest deep learning. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Alexey_Ivakhnenko) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Alexey_Ivakhnenko) start here, mentioned by joscha bach in: [Cyber Animism (talk hosted by Michael Levin's center, Tufts)](https://www.youtube.com/watch?v=YZl4zom3q2g) connections: - leads to Deep learning (`deeplearning`): the earliest deep networks. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) ## Symbolic AI 1956–1970s, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `symbolic` ### Symbolic AI and expert systems by Newell, Simon, Feigenbaum and others The first paradigm of AI: represent knowledge as symbols and logical rules and reason over them. Expert systems encoded specialists' rules by hand. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=1075), from 17:57 (verbatim): And symbolic AI was mostly similar to Wittgenstein's program. connections: - builds on The founding of AI (`dartmouth`): the first paradigm of AI. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Dartmouth_workshop) - leads to ELIZA (`eliza`): pattern matching that feels like understanding. confidence: highly likely, [source](https://en.wikipedia.org/wiki/ELIZA) - challenged by Dreyfus: What Computers Still Can't Do (`dreyfus`): what rules cannot capture. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hubert_Dreyfus%27s_views_on_artificial_intelligence) - related to The language of thought (`fodor`): thinking as symbol manipulation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Physical_symbol_system) - leads to Scripts, Plans, Goals and Understanding (`schank`): scripts. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - challenged by Searle: the Chinese room (`searle`): syntax is not semantics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Chinese_room) - leads to Cyc (`cyc`): common sense, hand-coded. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - challenged by Rodney Brooks: subsumption architecture (`brooks`): intelligence without representation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - leads to Cognitive architectures: Soar and ACT-R (`cogarch`): newell's unified theories of cognition. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - challenged by The symbol grounding problem (`harnad`): symbols defined only by other symbols. confidence: highly likely, [source](https://arxiv.org/abs/cs/9906002) ### Newell and Simon Carnegie Mellon pair who wrote the first AI programs and proposed that intelligence is the manipulation of physical symbols. Their 'physical symbol system hypothesis' was the working creed of classical AI. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Logic_Theorist) ## ELIZA 1966, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `eliza` by Joseph Weizenbaum An early chat program that imitated a therapist using simple pattern matching. Many users felt understood by it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/ELIZA) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/ELIZA) start here, mentioned by joscha bach in: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) connections: - builds on Symbolic AI (`symbolic`): pattern matching that feels like understanding. confidence: highly likely, [source](https://en.wikipedia.org/wiki/ELIZA) - related to Large language models (`llm`): people feel understood by a text machine. confidence: highly likely, [source](https://ojs.weizenbaum-institut.de/index.php/wjds/article/download/136/3_3_3/1438) ## Timothy Leary and the eight-circuit model 1960s–1977, psychology and mind. school of thought: psychology and development. id: `leary` by Timothy Leary; Robert Anton Wilson A counter-culture model that describes eight 'circuits' of the nervous system, from basic survival up to mystical states. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Eight-circuit_model_of_consciousness) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Eight-circuit_model_of_consciousness) start here, mentioned by joscha bach in: [Lex Fridman Podcast #212 – Joscha Bach: Nature of Reality, Dreams, and Consciousness](https://podscripts.co/podcasts/lex-fridman-podcast/212-joscha-bach-nature-of-reality-dreams-and-consciousness) connections: - builds on The Doors of Perception (`huxley`): leary found the book confirmed his psilocybin experience, then befriended huxley. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Doors_of_Perception) ## Digital physics 1969, physics and life. school of thought: computation, information and digital physics. id: `digital` by Konrad Zuse The hypothesis that the universe is, at bottom, a discrete computation such as a cellular automaton. Zuse proposed it in 'Rechnender Raum' (1969); Fredkin and Wolfram developed it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Digital_physics) start here, joscha bach on it: [Manifold #76 (Steve Hsu) – Joscha Bach: Consciousness and AGI](https://www.manifold1.com/episodes/joscha-bach-consciousness-and-agi-76/transcript) (verbatim): he actually believed that the physical universe is probably in some server farm in the parent universe connections: - builds on John von Neumann and the first computers (`vonneumann`): the universe as a cellular automaton. confidence: highly likely, [source](https://people.idsia.ch/~juergen/digitalphysics.html) - related to It from bit (`wheeler`): information as more basic than matter. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - leads to Stephen Wolfram (`wolfram`): a new kind of science. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - related to The simulation argument (`simulation`): overlapping but distinct: a computational universe need not be one run by someone. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - leads to Gerard 't Hooft: cellular automaton interpretation (`thooft`): a deterministic automaton under quantum mechanics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - leads to Joscha Bach (`bach`): strong computationalism: reality itself computes. confidence: highly likely, [source](https://media.ccc.de/v/33c3-8369-machine_dreams) ## Conway's Game of Life 1970, mathematics and logic. school of thought: computation, information and digital physics. id: `conway` by John Conway A grid of cells that live or die according to how many neighbours they have. From these few rules arise gliders, self-copying patterns and even complete computers. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Conway's_Game_of_Life) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Conway's_Game_of_Life) start here, mentioned by joscha bach in: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams) connections: - builds on John von Neumann and the first computers (`vonneumann`): cellular automata. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Conway%27s_Game_of_Life) - related to The Mandelbrot set (`mandelbrot`): unlimited complexity from a tiny rule. confidence: highly likely, [source](https://emmath.org/emm/chap-dynamics.html) - related to Stephen Wolfram (`wolfram`): wolfram cites life as a universal 2d cellular automaton, but calls it recreation, not science. confidence: highly likely, [source](https://www.wolframscience.com/reference/notes/876b/) - related to Neural cellular automata (`nca`): the neural ca paper names life as the classic cellular automaton before its learned ones. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) ## Mirror test 1970, psychology and mind. school of thought: psychology and development. id: `mirror` by Gordon Gallup Jr. A test of self-recognition: put a mark on an animal's face and see whether it uses a mirror to investigate the mark on itself. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Mirror_test) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Mirror_test) start here, joscha bach on it: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha#t=1208), from 20:10 (verbatim): it's actually not a test of consciousness, but it's a test of your ability to recognize that the mirror shows your body surface ## Holonomic brain theory 1971, psychology and mind. school of thought: scientific theories of consciousness. id: `pribram` by Karl Pribram (with David Bohm) Pribram's idea that the brain stores information like a hologram, spread across interference patterns, so each part contains something of the whole. It was inspired by the fact that memories survive quite large local brain damage. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Holonomic_brain_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Holonomic_brain_theory) start here, mentioned by joscha bach in: [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI&t=5652s), from 1:34:14 connections: - related to Penrose and Hameroff: quantum consciousness (`penrose`): non-classical physics in the brain. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_mind) ## Dreyfus: What Computers Still Can't Do 1972 / 1992, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `dreyfus` by Hubert Dreyfus A philosopher's long-running critique of symbolic AI. Dreyfus argued that human skill rests on bodily know-how and context that cannot be written down as explicit rules. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hubert_Dreyfus's_views_on_artificial_intelligence) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Hubert_Dreyfus's_views_on_artificial_intelligence) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - argues against Symbolic AI (`symbolic`): what rules cannot capture. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hubert_Dreyfus%27s_views_on_artificial_intelligence) - related to Embodied and enactive cognition (`embodied`): dreyfus's heidegger-based critique of symbolic ai is among those embodied cognition takes up. confidence: possible. The source links embodiment to Heidegger and opposes cognitivism, but does not explicitly say embodied cognition took up Dreyfus's critique. ## Gaia hypothesis 1972, physics and life. school of thought: cybernetics, systems and self-organisation. id: `gaia` by James Lovelock; Lynn Margulis The proposal that life on Earth, together with its inorganic surroundings, forms a self-regulating system that keeps the planet's atmosphere and climate suitable for life. James Lovelock and Lynn Margulis proposed it in the 1970s. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gaia_hypothesis) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Gaia_hypothesis) start here, joscha bach on it: [Joscha Bach: Consciousness, Artificial Intelligence, and the Threat of AI Apocalypse (Robinson Erhardt / Robinson's Podcast)](https://www.youtube.com/watch?v=XcNlv9gp20o&t=6384s), from 1:46:26 (verbatim): maybe the smartest thing on the planet is Gaia connections: - builds on Cybernetics (`cybernetics`): the planet as a self-regulating system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gaia_hypothesis) ## Nagel: What Is It Like to Be a Bat? 1974, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `nagel` by Thomas Nagel Nagel argues that an organism is conscious if there is something it is like to be that organism, and that objective science seems to leave this point of view out. confidence: highly likely, [source](https://en.wikipedia.org/wiki/What_Is_It_Like_to_Be_a_Bat%3F) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/consciousness/) start here, mentioned by joscha bach in: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=1223), from 20:25 connections: - related to Qualia (`qualia`): the felt qualities of experience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Qualia) - leads to Chalmers: the hard problem (`chalmers`): 'what it is like' becomes the hard problem. confidence: highly likely, [source](https://consc.net/papers/facing.html) ## The language of thought 1975, psychology and mind. school of thought: functionalism and computational theories of mind. id: `fodor` by Jerry Fodor Claims that thinking happens in an inner symbolic code with a grammar, somewhat like a language. This is the classic statement of symbolic cognitive science. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Language_of_thought_hypothesis) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/language-thought/) start here, joscha bach on it: [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience#t=299), from 5:01 (verbatim): our language of thought our mental lease is in some sense the programming language of our own mind connections: - builds on Noam Chomsky (`chomsky`): innate structure and the language of thought. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Jerry_Fodor) - related to Symbolic AI (`symbolic`): thinking as symbol manipulation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Physical_symbol_system) ## Adaptive Resonance Theory 1976, computing and AI. school of thought: neural networks and machine learning. id: `grossberg` by Stephen Grossberg Grossberg's theory that the brain learns stably by matching top-down expectations against bottom-up input; on his account every conscious state is such a 'resonance' between the two. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Stephen_Grossberg) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Stephen_Grossberg) start here, mentioned by joscha bach in: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) ## Dawkins: selfish genes and memes 1976, physics and life. school of thought: evolution, biology and life. id: `dawkins` by Richard Dawkins Dawkins reframed evolution as competition between genes, with organisms as the vehicles genes build to get themselves copied. The book also introduced the 'meme'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Selfish_Gene) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/The_Selfish_Gene) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM&t=2393s), from 39:55 (verbatim): Dawkins says no, it's actually about certain molecules, the genes that are just using the organism as a mechanism to replicate. connections: - builds on Charles Darwin (`darwin`): the gene's-eye view of selection. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Selfish_Gene) - related to Joscha Bach (`bach`): bach moves from dawkins's selfish genes to competing software agents. confidence: highly likely, [source](https://www.youtube.com/watch?v=34VOI_oo-qM) ## Jaynes: the bicameral mind 1976, psychology and mind. school of thought: psychology and development. id: `jaynes` by Julian Jaynes Jaynes claimed that until about 3,000 years ago people did not experience an inner self; they heard the commands of gods as voices, produced by one half of the brain and obeyed by the other. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bicameral_mentality) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Bicameral_mentality) start here, joscha bach on it: [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience#t=1443), from 24:05 (verbatim): Julian James wrote the famous book the origin of Consciousness and the breakdown of the bamal mind this book has the wrong name connections: - leads to Westworld (`westworld`): the series borrows the bicameral mind. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bicameral_mentality) - related to Joscha Bach (`bach`): gods as voices in the mind. confidence: possible. Bach discusses gods as mental constructs and cross-mind selves, but no opened source tied this account to Jaynes. ## Scripts, Plans, Goals and Understanding 1977, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `schank` by Roger Schank, Robert Abelson Proposed that people understand everyday situations through 'scripts', stored step-by-step outlines such as what happens in a restaurant. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Script_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Script_theory) start here, mentioned by joscha bach in: [Request confirmation networks for neuro-symbolic script execution (Bach & Herger, CoCo workshop at NIPS 2015)](https://ceur-ws.org/Vol-1583/CoCoNIPS_2015_paper_6.pdf) connections: - builds on Symbolic AI (`symbolic`): scripts. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - leads to Joscha Bach (`bach`): neural execution of scripts. confidence: highly likely, [source](https://ceur-ws.org/Vol-1583/CoCoNIPS_2015_paper_6.pdf) ## Aaron Sloman 1978, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `sloman` Philosopher and AI researcher at the University of Birmingham. He argued that philosophy of mind should be done by designing working minds, and he proposed the CogAff scheme that sorts mental processes into reactive, deliberative and reflective layers. confidence: highly likely, [source](https://www.cs.bham.ac.uk/research/projects/cogaff/) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Aaron_Sloman) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - leads to Joscha Bach (`bach`): philosophy of mind by designing working minds. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) ## Hofstadter: strange loops 1979–2007, philosophy. school of thought: functionalism and computational theories of mind. id: `hofstadter` by Douglas Hofstadter In 'Gödel, Escher, Bach' and 'I Am a Strange Loop', Hofstadter suggests the self arises when a symbol system becomes able to refer to itself, a tangled feedback loop like Gödel's self-referential sentence. On this view 'I' is a pattern, not a thing, and could in principle exist in other substrates. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Strange_loop) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Strange_loop) start here, joscha bach on it: [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI&t=5608s), from 1:33:30 (verbatim): I think that our consciousness is not strange as a loop. connections: - builds on Gödel's incompleteness theorems (`goedel`): self-reference as the root of the self. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Douglas_Hofstadter) - related to Daniel Dennett (`dennett`): 'the mind's i'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Daniel_Dennett) ## Lucid dreaming 1980, psychology and mind. school of thought: scientific theories of consciousness. id: `lucid` by Stephen LaBerge (scientific study) Dreaming while knowing that you are dreaming, sometimes with control over the dream. It has been confirmed in sleep laboratories, where dreamers signal with agreed eye movements. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Lucid_dream) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Lucid_dream) start here, joscha bach on it: [Donald Hoffman Λ Joscha Bach: Consciousness, Gödel, Reality (Theories of Everything)](https://www.youtube.com/watch?v=bhSlYfVtgww&t=4148s), from 1:09:10 (verbatim): put yourself into a lucid dream and look at the fidelity of what your brain is able to generate connections: - related to Joscha Bach (`bach`): waking perception as a dream constrained by the senses. confidence: highly likely, [source](https://media.ccc.de/v/33c3-8369-machine_dreams) ## Perceptual illusions 1980 / 2015, psychology and mind. school of thought: psychology and development. id: `thatcher` ### The Thatcher illusion by Peter Thompson If a face is shown upside down with the eyes and mouth flipped, it looks normal until you turn the picture upright, when it looks grotesque. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thatcher_effect) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Thatcher_effect) start here, mentioned by joscha bach in: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha#t=357), from 5:59 ### The dress A viral photograph of a dress that some people saw as white and gold and others as blue and black. The cause is not settled; one explanation is that their visual systems made different assumptions about the lighting. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_dress) ## Searle: the Chinese room 1980, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `searle` by John Searle Berkeley philosopher who held that consciousness is a biological product of the brain's specific causal powers, like digestion, so that a program alone, however clever, cannot have it. His Chinese Room thought experiment is the best-known argument that manipulating symbols is not the same as understanding them. confidence: highly likely, [source](https://en.wikipedia.org/wiki/John_Searle) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/chinese-room/) start here, joscha bach on it: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=3226s), from 53:48 (verbatim): The issue with Searle is that there is very little for me to engage. There is a claim but no argument. connections: - related to Gottfried Wilhelm Leibniz (`leibniz`): walking around inside the machine and finding no understanding. confidence: highly likely, [source](https://plato.stanford.edu/entries/chinese-room/) - argues against Functionalism and computationalism (`functionalism`): the chinese room. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Chinese_room) - argues against Symbolic AI (`symbolic`): syntax is not semantics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Chinese_room) - challenged by Joscha Bach (`bach`): bach rejects the chinese room. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) - leads to Anil Seth: biological naturalism (`seth`): biological naturalism. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) ## The Mandelbrot set 1980, mathematics and logic. school of thought: cybernetics, systems and self-organisation. id: `mandelbrot` by Benoît Mandelbrot The Mandelbrot set is an infinitely detailed shape generated by repeating one tiny formula. It is a standard example of unlimited complexity from a simple rule. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Mandelbrot_set) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Mandelbrot_set) start here, joscha bach on it: [Lex Fridman Podcast #101 – Joscha Bach: Artificial Consciousness and the Nature of Reality](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality) (verbatim): But imagine you live inside of this fractal and you don't have access to where you are in the fractal connections: - related to Conway's Game of Life (`conway`): unlimited complexity from a tiny rule. confidence: highly likely, [source](https://emmath.org/emm/chap-dynamics.html) ## The binding problem and neural synchrony 1980–1990, psychology and mind. school of thought: scientific theories of consciousness. id: `binding` ### The binding problem by Anne Treisman & Garry Gelade; Peter Milner; Francis Crick & Christof Koch The question of how the brain joins separately processed features such as colour, shape and motion into one perceived object. Proposed answers include an attention-driven integration step and the synchronised firing of the neurons involved. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Binding_problem) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Binding_problem) start here, mentioned by joscha bach in: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) connections: - leads to Crick and Koch: neural correlates of consciousness (`ncc`): 40 hz synchrony as a correlate of consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gamma_wave) - leads to Joscha Bach (`bach`): von der malsburg's coherence. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) ### Wolf Singer and 40 Hz synchrony by Wolf Singer (also Crick and Koch) The proposal that consciousness, or the binding of features into one percept, is tied to neurons firing in synchrony at around 40 cycles per second. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gamma_wave) ### Christoph von der Malsburg German neuroscientist and physicist. In 1981 he proposed that the brain binds features into objects through temporary correlations between neurons, and later defined consciousness as a state of high coherence among the brain's subsystems. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Christoph_von_der_Malsburg) ## Kegan's stages of adult development 1982, psychology and mind. school of thought: psychology and development. id: `kegan` by Robert Kegan Harvard psychologist's model in which adults can move from absorbing their beliefs from their group (stage 3), to reasoning independently (stage 4), to seeing how identities and value systems are themselves constructed (stage 5). Each stage makes an object of what the previous stage was unreflectively subject to. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Robert_Kegan) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Robert_Kegan) start here, joscha bach on it: [Can we crack the AI enigma | Joscha Bach, Mazviita Chirimuuta, Cory Doctorow (Institute of Art and Ideas, public excerpt)](https://www.youtube.com/watch?v=civc01IZCcY&t=216s), from 3:38 (verbatim): if you're at stage five you just wonder is the AI going to be enlightened fast enough connections: - builds on Developmental stage theories (`piaget`): stages continue into adulthood. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Robert_Kegan) - leads to Joscha Bach (`bach`): seven levels that grew out of kegan's five stages. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) ## Quantum computing 1982, physics and life. school of thought: computation, information and digital physics. id: `quantumcomp` by Richard Feynman; David Deutsch Computers that use quantum superposition and entanglement to solve certain problems, such as factoring, far faster than any known classical method. Building them is hard because qubits lose their quantum states (decohere) unless they are well isolated from the environment. confidence: highly likely, [source](https://www.ibm.com/think/topics/quantum-computing) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Quantum_computing) start here, joscha bach on it: [The quantum hoax | Dominic Walliman, Joscha Bach, Ruth Oulton (Institute of Art and Ideas, public excerpt)](https://www.youtube.com/watch?v=RLBZEfJAN44&t=193s), from 3:15 (verbatim): I wouldn't say that uh Quantum Computing is a hoax but it is an extremely bold proposal connections: - related to Alan Turing and computability (`turing`): a faster, not a more powerful, computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_computing) - related to Penrose and Hameroff: quantum consciousness (`penrose`): quantum effects in the brain. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_mind) - leads to David Deutsch: constructor theory (`deutsch`): deutsch defined the universal quantum computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Deutsch) ## Two visual streams 1982, psychology and mind. school of thought: scientific theories of consciousness. id: `streams` by Ungerleider & Mishkin; Goodale & Milner Visual processing in the cortex splits into a ventral 'what' pathway that identifies objects and a dorsal 'where/how' pathway that locates them and guides action. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Two-streams_hypothesis) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Two-streams_hypothesis) start here, joscha bach on it: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha) (paraphrase): Paraphrase (garbled subtitles): there are two primary directions of processing that work in parallel, a 'what' path about the nature of objects and one which tells you where they are. ## Libet's readiness-potential experiments 1983, psychology and mind. school of thought: scientific theories of consciousness. id: `libet` by Benjamin Libet et al. Experiments in which brain activity leading to a voluntary movement began a fraction of a second before people reported deciding to move. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Benjamin_Libet) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Benjamin_Libet) start here, mentioned by joscha bach in: [Phenomenal Experience and the Perceptual Binding State (AAAI Spring Symposium workshop TOCAIS 2019, CEUR Vol-2287)](https://ceur-ws.org/Vol-2287/paper29.pdf) ## The explanatory gap 1983, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `levine` by Joseph Levine Levine's name for the gap between any physical or functional description of the brain and an explanation of why it feels the way it does. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Explanatory_gap) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/qualia/) start here, mentioned by joscha bach in: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) connections: - builds on Qualia (`qualia`): the explanatory gap. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Qualia) - leads to Chalmers: the hard problem (`chalmers`): from gap to hard problem. confidence: highly likely, [source](https://consc.net/papers/facing.html) ## Braitenberg: Vehicles 1984, computing and AI. school of thought: cybernetics, systems and self-organisation. id: `braitenberg` by Valentino Braitenberg A short book of thought experiments. Very simple wheeled machines with a few sensors and motors seem to show fear, love or curiosity, which suggests that complex-looking behaviour can come from simple wiring. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Braitenberg_vehicle) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Braitenberg_vehicle) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - related to Rodney Brooks: subsumption architecture (`brooks`): behaviour from simple wiring. confidence: highly likely, [source](https://www.cs.kent.ac.uk/pubs/2006/2867/content.pdf) ## Cyc 1984, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `cyc` by Douglas Lenat A decades-long project to hand-code millions of common-sense facts and rules so that a machine could reason like a person. Doug Lenat started it in 1984; critics such as Pedro Domingos have called it a catastrophic failure. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cyc) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Cyc) start here, mentioned by joscha bach in: [Joscha Bach vs. Jeremy Nixon: Is Consciousness Required for AGI? (AGI House)](https://www.youtube.com/watch?v=rFI2wcmst4I) connections: - builds on Symbolic AI (`symbolic`): common sense, hand-coded. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) ## Boltzmann machine 1985, computing and AI. school of thought: neural networks and machine learning. id: `boltzmannmachine` by Geoffrey Hinton, Terrence Sejnowski, David Ackley A neural network in which units constrain each other and the whole network tends toward low-energy states consistent with the inputs. The restricted version (RBM) drops within-layer links to make training feasible and helped launch deep learning. confidence: highly likely, [source](http://www.scholarpedia.org/article/Boltzmann_machine) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Boltzmann_machine) start here, mentioned by joscha bach in: [Singularity.FM – Joscha Bach on AI, Cosmology, Existence and the Bible](https://www.youtube.com/watch?v=vyGP8LpsDok) connections: - leads to Deep learning (`deeplearning`): hinton's line of work. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) ## Norbert Bischof 1980s, psychology and mind. school of thought: psychology and development. id: `bischof` German psychologist who modelled motivation, attachment and social distance as control systems. confidence: highly likely, [source](https://bischof.com/mat/Zuercher_Modell.pdf) start here, read: [bischof.com](https://bischof.com/mat/Zuercher_Modell.pdf) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - leads to Joscha Bach (`bach`): via dörner's psi theory, which takes up bischof's cybernetic psychology. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) ## Structure and Interpretation of Computer Programs 1985, mathematics and logic. school of thought: computation, information and digital physics. id: `sicp` by Harold Abelson; Gerald Jay Sussman MIT's classic introductory programming textbook. It opens by describing computational processes as abstract beings that inhabit computers and are conjured by programs, like spirits by spells. confidence: highly likely, [source](https://mitp-content-server.mit.edu/books/content/sectbyfn/books_pres_0/6515/sicp.zip/full-text/book/book-Z-H-9.html) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Structure_and_Interpretation_of_Computer_Programs) start here, joscha bach on it: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) (verbatim): computational processes are abstract beings that inhabit computers connections: - builds on Lambda calculus and combinators (`lambda`): lisp and scheme. confidence: highly likely, [source](https://sarabander.github.io/sicp/html/1_002e3.xhtml) - related to Joscha Bach (`bach`): programs as spirits conjured with spells. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) ## Autocatalytic networks and autopoiesis 1972–1986, physics and life. school of thought: cybernetics, systems and self-organisation. id: `autopoiesis` by Stuart Kauffman; Wim Hordijk & Mike Steel; Humberto Maturana & Francisco Varela Models of how a set of parts can come to produce and maintain itself as a whole, first used to explain the origin of life. Autopoiesis is the related idea of a system that continually makes its own components. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Autocatalytic_set) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Autocatalytic_set) start here, mentioned by joscha bach in: [A Path to Generative Artificial Selves (Gabora & Bach, EPIA 2023, LNCS; OSF/PsyArXiv preprint text)](https://doi.org/10.31234/osf.io/y3tzs) connections: - builds on Schrödinger: life as negentropy (`schroedinger`): what keeps a living system ordered. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Autopoiesis) - builds on Cybernetics (`cybernetics`): second-order cybernetics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Second-order_cybernetics) ## Connectionism and backpropagation 1986, computing and AI. school of thought: neural networks and machine learning. id: `pdp` by Rumelhart, Hinton & Williams (popularised) The standard way to train neural networks: measure the output error and pass it backwards through the layers to work out how much each connection contributed, then adjust. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Backpropagation) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Backpropagation) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=1370), from 22:52 (verbatim): And you take this error and you propagate it backwards. It's called back propagation. connections: - builds on McCulloch–Pitts neurons (`mcculloch`): connectionism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Connectionism) - argues against Marvin Minsky (`minsky`): connectionism returns after 'perceptrons'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Perceptrons_(book)) - leads to Deep learning (`deeplearning`): backpropagation at scale. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) ## Higher-order thought theory 1986, consciousness. school of thought: scientific theories of consciousness. id: `hot` by David Rosenthal A mental state is conscious when there is a higher-order thought about it: a thought that one is in that state. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Higher-order_theories_of_consciousness) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/consciousness-higher/) start here, joscha bach on it: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) (verbatim): some philosophers call this higher order thought because it's a thought about something else but I don't think it's a thought it's a perception connections: - related to Attention Schema Theory (`ast`): a model of one's own mental state. confidence: highly likely, [source](https://grazianolab.princeton.edu/sites/g/files/toruqf3411/files/graziano_review_2020.pdf) - related to Joscha Bach (`bach`): both use higher-order representation, but Bach distinguishes second-order perception from higher-order thought. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - leads to AI consciousness and welfare reports (`butlin`): one of the checklist theories. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) ## Rodney Brooks: subsumption architecture 1986, computing and AI. school of thought: cybernetics, systems and self-organisation. id: `brooks` by Rodney Brooks Brooks argued that intelligence should be grown from the bottom up in robots that act in the world, using layers of simple behaviours with no central world model. He co-founded iRobot, the company that makes the Roomba. confidence: highly likely, [source](https://www.computerhistory.org/collections/catalog/300000049) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Subsumption_architecture) start here, mentioned by joscha bach in: [Joscha Bach vs. Jeremy Nixon: Is Consciousness Required for AGI? (AGI House)](https://www.youtube.com/watch?v=rFI2wcmst4I&t=260s), from 4:22 connections: - argues against Symbolic AI (`symbolic`): intelligence without representation. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - related to Braitenberg: Vehicles (`braitenberg`): behaviour from simple wiring. confidence: highly likely, [source](https://www.cs.kent.ac.uk/pubs/2006/2867/content.pdf) - related to Embodied and enactive cognition (`embodied`): the body in the world comes first. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Embodied_cognition) ## Cognitive architectures: Soar and ACT-R 1987–2017, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `cogarch` ### Cognitive architectures by Allen Newell, John Laird; John R. Anderson A research tradition that tries to build complete, testable models of a whole mind (perception, memory, motivation, language) rather than isolated modules. It goes under names like cognitive systems, AGI and BICA. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Cognitive_architecture) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Cognitive_architecture) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - builds on Symbolic AI (`symbolic`): newell's unified theories of cognition. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence) - related to Joscha Bach (`bach`): psi compared with soar and ACT-R. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) ### Soar and Unified Theories of Cognition by Allen Newell; John Laird, Paul Rosenbloom Newell argued that psychology needs single theories that cover all of cognition, built as running computer architectures. Soar is his group's example: a system that solves problems by applying rules and learning new ones. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Soar_(cognitive_architecture)) ### ACT-R by John R. Anderson A widely used cognitive architecture from psychology. It models thinking as rules that fire one at a time over chunks of memory, each with a numeric level of activation, and it is tested against timing data from human experiments. confidence: highly likely, [source](https://en.wikipedia.org/wiki/ACT-R) ### A Standard Model of the Mind by John Laird, Christian Lebiere, Paul Rosenbloom A proposal for a shared baseline architecture of human-like minds that AI, cognitive science, neuroscience and robotics could all refer to. confidence: highly likely, [source](https://onlinelibrary.wiley.com/doi/10.1609/aimag.v38i4.2744) ## Neural Darwinism 1987, psychology and mind. school of thought: scientific theories of consciousness. id: `edelman` by Gerald Edelman Edelman's theory that the brain wires itself by selection: groups of neurons compete, and those whose activity proves useful are strengthened. On this view the brain is not programmed with instructions but shaped by competition, like an immune system. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Neural_Darwinism) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Neural_Darwinism) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM&t=1305s), from 21:47 (verbatim): Gary Edelman calls this neural Darwinism the idea that mental organization is evolving in every individual connections: - builds on Charles Darwin (`darwin`): selection applied to groups of neurons. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Neural_Darwinism) ## Global Workspace Theory 1988, consciousness. school of thought: scientific theories of consciousness. id: `gwt` by Bernard Baars; Stanislas Dehaene Many specialised brain processes run unconsciously; what becomes conscious is whatever is broadcast on a shared 'workspace' so that all of them can use it. Dehaene found neural signatures of this global broadcast. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Global_workspace_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Global_workspace_theory) start here, mentioned by joscha bach in: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=664s), from 11:06 connections: - related to Marvin Minsky (`minsky`): many unconscious specialists. confidence: highly likely, [source](http://scholarpedia.org/article/Global_workspace) - leads to Bengio: the consciousness prior (`bengio`): the consciousness prior. confidence: highly likely, [source](https://arxiv.org/abs/1709.08568) - related to Joscha Bach (`bach`): bach lists global workspace theory among views converging on his own. confidence: highly likely, [source](https://ceur-ws.org/Vol-2287/paper29.pdf) - leads to AI consciousness and welfare reports (`butlin`): one of the checklist theories. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) ## The Cognitive Structure of Emotions 1988, psychology and mind. school of thought: psychology and development. id: `occ` by Andrew Ortony, Gerald Clore, Allan Collins A widely used scheme that sorts emotions into types according to how a person appraises events, agents and objects. Many software agents and artificial characters use it to generate emotions. confidence: highly likely, [source](https://people.idsia.ch/~steunebrink/Publications/KI09_OCC_revisited.pdf) start here, read: [webs.cs.fiu.edu](https://webs.cs.fiu.edu/ascllab/wp-content/uploads/sites/75/2020/10/lisetti-correctedproofs-ch8-oxford.pdf) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - challenged by Joscha Bach (`bach`): emotions as modes of the whole system, not a lookup table. confidence: highly likely, [source](https://agi-conf.org/2019/wp-content/uploads/2019/07/paper_30.pdf) ## Mysterianism 1989, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `mysterianism` by Colin McGinn; Noam Chomsky The view that the human mind is simply not built to understand how consciousness arises from matter, as a dog cannot understand calculus. Its best-known defender is the philosopher Colin McGinn. confidence: highly likely, [source](https://en.wikipedia.org/wiki/New_mysterianism) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/New_mysterianism) start here, joscha bach on it: [Donald Hoffman Λ Joscha Bach: Consciousness, Gödel, Reality (Theories of Everything)](https://www.youtube.com/watch?v=bhSlYfVtgww&t=884s), from 14:46 (verbatim): is the position that something cannot be understood if it cannot be understood by Noam Chomsky connections: - related to Du Bois-Reymond: Ignorabimus (`dubois`): the ignorabimus is seen as anticipating mysterianism, though the two views differ. confidence: highly likely, [source](https://www.cambridge.org/core/journals/science-in-context/article/reconsidering-the-ignorabimus-du-boisreymond-and-the-hard-problem-of-consciousness/8C19E49A253DD5011482AD4252FF563D) - related to Chalmers: the hard problem (`chalmers`): mcginn had already argued we may be unable to solve it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) ## Penrose and Hameroff: quantum consciousness 1989–1994, consciousness. school of thought: the hard problem and sceptics of machine minds. id: `penrose` ### The Gödelian argument and The Emperor's New Mind by Roger Penrose Penrose's bestseller arguing that minds do something no algorithm can: see the truth of statements that formal systems cannot prove. He concludes that brains exploit non-computable physics. confidence: highly likely, [source](https://iep.utm.edu/lp-argue/) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Roger_Penrose) start here, joscha bach on it: [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) (verbatim): he perceives kind of mechanics as mysterious and perceives consciousness as mysterious and tries to shroud one mystery in another. connections: - builds on Gödel's incompleteness theorems (`goedel`): penrose's argument that minds are not algorithms. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Orchestrated_objective_reduction) - related to Holonomic brain theory (`pribram`): non-classical physics in the brain. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_mind) - related to Quantum computing (`quantumcomp`): quantum effects in the brain. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Quantum_mind) - challenged by Joscha Bach (`bach`): against the gödelian argument for non-computable minds. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/) ### Orchestrated objective reduction (Orch OR) by Roger Penrose; Stuart Hameroff The proposal that consciousness arises when quantum states in microtubules inside neurons collapse under a gravitational threshold. It combines Penrose's physics with Hameroff's cell biology. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Orchestrated_objective_reduction) ### The Road to Reality by Roger Penrose Published in 2004, Penrose's huge survey covers the mathematics behind modern physics, from numbers to quantum gravity, for determined general readers. It runs to more than a thousand pages and does not avoid the equations. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Road_to_Reality) ### Stuart Hameroff Anesthesiologist at the University of Arizona who proposed in the 1980s that microtubules, protein tubes inside neurons, process information like tiny computers, and who organises the Science of Consciousness conference. With Penrose he turned this into the quantum 'Orch OR' theory of consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Stuart_Hameroff) ### Roger Penrose Nobel-winning mathematical physicist who argues from Gödel's theorem that human understanding is not computable, and therefore that consciousness must depend on new physics. With Stuart Hameroff he proposed that the missing physics is a gravity-related collapse of quantum states inside neurons. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Roger_Penrose) ## Crick and Koch: neural correlates of consciousness 1990, consciousness. school of thought: scientific theories of consciousness. id: `ncc` ### Neural correlates of consciousness by Francis Crick; Christof Koch The research programme of finding the minimal brain activity that reliably accompanies a given conscious experience, without first solving why. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Neural_correlates_of_consciousness) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Neural_correlates_of_consciousness) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM&t=854s), from 14:16 (verbatim): there are people like Christof Koch who sometimes say that a simulation cannot be conscious. And I think it's exactly the other way around. connections: - builds on The binding problem and neural synchrony (`binding`): 40 hz synchrony as a correlate of consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Gamma_wave) - related to Integrated Information Theory (`iit`): koch moved from NCC work with crick to working with tononi on IIT. confidence: highly likely, [source](https://christofkoch.com/my-science/) ### Christof Koch Neuroscientist who pioneered the search for neural correlates of consciousness with Francis Crick and later helped develop and champion integrated information theory (IIT). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Christof_Koch) ## Ian Witten: cognition as compression 1980s–1990s, computing and AI. school of thought: computation, information and digital physics. id: `witten` English-born computer scientist at the University of Waikato, New Zealand, who co-invented prediction-based text compression (PPM) and popularised arithmetic coding; the idea is that to compress data well you must predict it, which requires modelling its structure. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ian_H._Witten) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Ian_H._Witten) start here, mentioned by joscha bach in: [Lex Fridman Podcast #212 – Joscha Bach: Nature of Reality, Dreams, and Consciousness](https://podscripts.co/podcasts/lex-fridman-podcast/212-joscha-bach-nature-of-reality-dreams-and-consciousness) connections: - builds on Claude Shannon and information theory (`shannon`): prediction and compression. confidence: highly likely, [source](https://doi.org/10.1145/76894.76896) ## It from bit 1990, physics and life. school of thought: computation, information and digital physics. id: `wheeler` by John Archibald Wheeler Wheeler's slogan for the idea that information is more basic than matter and energy: every physical 'it' comes from yes-or-no answers, 'bits'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/John_Archibald_Wheeler) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/John_Archibald_Wheeler) start here, mentioned by joscha bach in: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) connections: - related to Digital physics (`digital`): information as more basic than matter. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) ## Jürgen Schmidhuber 1990, computing and AI. school of thought: neural networks and machine learning. id: `schmidhuber` by Jürgen Schmidhuber The idea that learning is data compression, that curiosity is the reward for finding better compressions, and that beauty is how compressible something already is to the observer. confidence: highly likely, [source](https://people.idsia.ch/~juergen/creativity.html) start here, read: [people.idsia.ch](https://people.idsia.ch/~juergen/creativity.html) start here, joscha bach on it: [Computational Meta-Psychology (32C3)](https://media.ccc.de/v/32c3-7483-computational_meta-psychology#t=1403), from 23:25 (verbatim): Jürgen Schmidhuber said that the biggest, the last contribution of neuroscience to artificial intelligence was about 50 years ago. connections: - builds on Solomonoff induction (`solomonoff`): schmidhuber generalises solomonoff's algorithmic probability and universal prediction. confidence: highly likely, [source](https://people.idsia.ch/~juergen/kolmogorov.html) - leads to Deep learning (`deeplearning`): lstm and early deep learning. confidence: highly likely, [source](https://en.wikipedia.org/wiki/J%C3%BCrgen_Schmidhuber) ## The symbol grounding problem 1990, computing and AI. school of thought: the hard problem and sceptics of machine minds. id: `harnad` by Stevan Harnad How can symbols inside a system mean anything, if they are only defined in terms of other symbols and never tied to the world? Stevan Harnad posed it in 1990, comparing it to learning Chinese from a Chinese-only dictionary. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Symbol_grounding_problem) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Symbol_grounding_problem) start here, joscha bach on it: [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf) (verbatim): Build grounded systems, but do not get entangled in the Symbol Grounding Problem connections: - argues against Symbolic AI (`symbolic`): symbols defined only by other symbols. confidence: highly likely, [source](https://arxiv.org/abs/cs/9906002) - related to Joscha Bach (`bach`): ground the system, but do not get stuck on symbol grounding. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) ## Daniel Dennett 1991, philosophy. school of thought: functionalism and computational theories of mind. id: `dennett` American philosopher who argued that consciousness is not an inner theatre but many parallel drafts of interpretation, and that the 'self' is a useful narrative centre. He treated the 'hard problem' as a confusion to be dissolved rather than a mystery to be solved. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Daniel_Dennett) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/consciousness/) start here, joscha bach on it: [Singularity.FM – Joscha Bach: We need to understand the nature of AI to understand who we are](https://www.youtube.com/watch?v=sIKbp3KcS8A) (verbatim): the things that he says are not wrong but they also not non-obvious connections: - builds on Functionalism and computationalism (`functionalism`): dennett's functionalist account of consciousness. confidence: highly likely, [source](https://dl.tufts.edu/downloads/9306t946z?filename=5t34sw885.pdf) - related to Hofstadter: strange loops (`hofstadter`): 'the mind's i'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Daniel_Dennett) - challenged by Chalmers: the hard problem (`chalmers`): chalmers answers 'consciousness explained'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Consciousness_Explained) - leads to Illusionism (`illusionism`): frankish names dennett's position. confidence: highly likely, [source](https://keithfrankish.github.io/articles/Frankish_2016_Illusionism_as_a_theory_of_consciousness_eprint.pdf) - leads to Joscha Bach (`bach`): conductor theory as an extension of the multiple drafts model. confidence: highly likely, [source](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf) ### Consciousness Explained by Daniel Dennett Dennett's book proposing the 'multiple drafts' model: there is no single place or moment where experience happens, only competing processes whose winners shape behaviour and memory. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Consciousness_Explained) ### Zimbos by Daniel Dennett A zombie is a being that behaves as if conscious with nobody home. Dennett's 'zimbo' is a zombie with higher-order self-monitoring, so it believes and reports that it is conscious. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Philosophical_zombie) ## Embodied and enactive cognition 1927–2006, philosophy. school of thought: rationalism and idealism. id: `embodied` ### Phenomenology and enactivism by Martin Heidegger; Maurice Merleau-Ponty; Francisco Varela, Evan Thompson, Eleanor Rosch; Alva Noë; Shaun Gallagher A line of philosophy that starts from lived experience and holds that mind is constituted by a body acting in a world, not by inner representations alone. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Embodied_cognition) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/embodied-cognition/) start here, joscha bach on it: [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf) (verbatim): Do not wait for the rapture of robotic embodiment connections: - builds on Phenomenology (`husserl`): phenomenology of the lived body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Embodied_cognition) - related to Dreyfus: What Computers Still Can't Do (`dreyfus`): dreyfus's heidegger-based critique of symbolic ai is among those embodied cognition takes up. confidence: possible. The source links embodiment to Heidegger and opposes cognitivism, but does not explicitly say embodied cognition took up Dreyfus's critique. - related to Rodney Brooks: subsumption architecture (`brooks`): the body in the world comes first. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Embodied_cognition) - challenged by Joscha Bach (`bach`): 'do not wait for robot bodies'. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) - leads to Anil Seth: biological naturalism (`seth`): mind needs a living body. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) ### How the Body Shapes the Way We Think by Rolf Pfeifer, Josh Bongard A statement of embodied AI: intelligence comes from the interplay of a body, its materials and the world, so robots are needed to study it. confidence: highly likely, [source](https://mitpress.mit.edu/9780262162395/how-the-body-shapes-the-way-we-think/) ### Sensorimotor theory by Kevin O'Regan and Alva Noë Experience is not something the brain produces internally; it is a skilled activity of exploring the world, constituted by the lawful links between movement and sensation. It was put forward by Kevin O'Regan and Alva Noë in 2001. confidence: highly likely, [source](https://kevin-oregan.net/OREGAN-NOE-BBS/ORegan%3BNoe.BBS.pdf) ## Technological singularity 1993, society. school of thought: ethics, society and AI risk. id: `singularity` by John von Neumann; Vernor Vinge; Ray Kurzweil The idea that once machines can improve themselves, progress accelerates beyond anything humans can predict or control. The term was popularised by Vernor Vinge and Ray Kurzweil. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Technological_singularity) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Technological_singularity) start here, joscha bach on it: [Joscha Bach Λ Ben Goertzel: Conscious Ai, LLMs, AGI (Theories of Everything)](https://www.youtube.com/watch?v=xw7omaQ8SgA&t=5836s), from 1:37:18 (verbatim): I think that we already triggered the singularity when we invented technology connections: - builds on John von Neumann and the first computers (`vonneumann`): Ulam's report of von Neumann's accelerating-change conversation. confidence: highly likely, [source](https://doi.org/10.1090/s0002-9904-1958-10189-5) - leads to AI risk: Bostrom, Yudkowsky and the doom debate (`airisk`): runaway self-improvement. confidence: highly likely, [source](https://doi.org/10.1007/s11098-024-02143-5) ## Antonio Damasio 1994, psychology and mind. school of thought: scientific theories of consciousness. id: `damasio` by Antonio Damasio Neuroscientist known for arguing that emotion and bodily feeling are needed for good reasoning ('Descartes' Error', 1994) and for a layered account of the self ('The Feeling of What Happens', 1999). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Antonio_Damasio) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Antonio_Damasio) start here, mentioned by joscha bach in: [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) connections: - builds on William James: The Principles of Psychology (`james`): emotion as perception of the body. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Antonio_Damasio) - leads to Anil Seth: biological naturalism (`seth`): feeling grounded in body regulation. confidence: highly likely, [source](https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(13)00211-8) ## Greg Egan 1994, fiction. school of thought: fiction. id: `egan` by Greg Egan Hard science fiction about minds that run as software: copies of people living in simulations, citizens who are programs, and what personal identity means when a mind can be copied. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Permutation_City) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Permutation_City) connections: - related to The simulation argument (`simulation`): minds running as software. confidence: highly likely, [source](https://www.gregegan.net/PERMUTATION/Permutation.html) ## Multilevel selection 1994, physics and life. school of thought: evolution, biology and life. id: `groupsel` by David Sloan Wilson; Elliott Sober Natural selection acts at several levels at once: selfish individuals beat altruists within a group, but groups of altruists beat groups of egoists. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Group_selection) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Group_selection) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=2432), from 40:34 (verbatim): we are subject to multilevel selection inside of our organism. connections: - builds on Charles Darwin (`darwin`): selection at several levels. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Group_selection) ## Chalmers: the hard problem 1995, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `chalmers` ### The hard problem of consciousness by David Chalmers Chalmers separates the 'easy' problems of explaining what the brain does from the 'hard' problem of why any of it is felt. He argues that function alone does not explain experience. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) start here, joscha bach on it: [Joscha Bach: "You Are Not Who You Think You Are!" (Dr Brian Keating, Into the Impossible)](https://www.youtube.com/watch?v=7bqdPHLIY8w&t=0s), from 0:00 (verbatim): I have no hard problem. I have a lot of difficult problems which are all technical problems connections: - related to Gottfried Wilhelm Leibniz (`leibniz`): the mill argument is cited as an early version of the hard problem. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) - related to Epiphenomenalism (`epiphenomenalism`): experience that does no causal work. confidence: highly likely, [source](https://plato.stanford.edu/entries/epiphenomenalism/) - argues against Functionalism and computationalism (`functionalism`): function alone leaves experience unexplained. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) - builds on Nagel: What Is It Like to Be a Bat? (`nagel`): 'what it is like' becomes the hard problem. confidence: highly likely, [source](https://consc.net/papers/facing.html) - builds on The explanatory gap (`levine`): from gap to hard problem. confidence: highly likely, [source](https://consc.net/papers/facing.html) - related to Mysterianism (`mysterianism`): mcginn had already argued we may be unable to solve it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hard_problem_of_consciousness) - argues against Daniel Dennett (`dennett`): chalmers answers 'consciousness explained'. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Consciousness_Explained) - leads to Panpsychism (`panpsychism`): experience as fundamental. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Panpsychism) - challenged by Illusionism (`illusionism`): phenomenal consciousness is the illusion; the hard problem gives way to the 'illusion problem'. confidence: highly likely, [source](https://keithfrankish.github.io/articles/Frankish_2016_Illusionism_as_a_theory_of_consciousness_eprint.pdf) - challenged by Joscha Bach (`bach`): the hard problem dissolves if experience belongs to a simulated character. confidence: highly likely, [source](https://www.youtube.com/watch?v=SGZ784mBuU0) ### Philosophical zombies by Robert Kirk (the idea, 1974); David Chalmers (the argument) A hypothetical creature physically and behaviourally identical to a person but with no inner experience. If such a being is conceivable, the argument goes, consciousness is something over and above physical function. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Philosophical_zombie) ### David Chalmers Philosopher who named the 'hard problem' of consciousness and defends the possibility that experience is a fundamental feature of the world. He has taken the possibility of conscious AI seriously for decades. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Chalmers) ## David Pearce 1995, society. school of thought: ethics, society and AI risk. id: `pearce` British transhumanist philosopher, author of 'The Hedonistic Imperative'. He wants to abolish suffering in all sentient life through biotechnology and argues that classical digital computers can never be conscious because they cannot solve the binding problem. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Pearce_(philosopher)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/David_Pearce_(philosopher)) start here, joscha bach on it: [The Panel Where Everyone Disagrees About Saving the World - Joscha Bach, David Pearce, Andrew Critch (California Institute for Machine Consciousness)](https://www.youtube.com/watch?v=A3jRaGMLhNc&t=1766s), from 29:28 (verbatim): Wouldn't this entail that we should sterilize everything in the forest? ## Ned Block: access vs phenomenal consciousness 1995, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `block` by Ned Block Philosopher of mind. He raised classic doubts about functionalism ('Troubles with functionalism', 1978) and separated phenomenal consciousness, what it is like, from access consciousness, what is available for report and control (1995). confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ned_Block) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Ned_Block) start here, joscha bach on it: [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine#t=1900), from 31:42 (verbatim): Phenomenal consciousness is the memory of the binding state connections: - argues against Functionalism and computationalism (`functionalism`): 'troubles with functionalism'. confidence: highly likely, [source](https://plato.stanford.edu/entries/functionalism/) ## Dietrich Dörner and Psi theory 1999, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `doerner` ### Psi theory by Dietrich Dörner A theory of how a mind works as a whole system. An agent has a small set of built-in needs (for example food, safety, certainty, competence and belonging), and these needs create motives that steer perception, memory, planning and action. Emotions are not separate modules but ways in which the whole system changes its style of processing. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Psi-theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Psi-theory) start here, joscha bach on it: [Salon with Stephen Wolfram × Joscha Bach_What if we had bigger brains. (California Institute for Machine Consciousness)](https://www.youtube.com/watch?v=-G1SdsRXL7k&t=1138s), from 19:00 (verbatim): I asked him why do you use this word it's just throwing people off connections: - leads to Joscha Bach (`bach`): micropsi implements and extends dörner's psi theory. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) ### Bauplan für eine Seele by Dietrich Dörner German psychologist Dörner's attempt to specify a complete mind (motivation, emotion, perception, planning) as an implementable architecture, the 'Psi theory'. Joscha Bach's MicroPsi architecture and his book 'Principles of Synthetic Intelligence' build directly on it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Psi-theory) ### Die Mechanik des Seelenwagens by Dietrich Dörner and six co-authors A follow-up to Dörner's 'Bauplan' that describes the regulation of action in more technical detail and reports computer implementations of the Psi agent. confidence: highly likely, [source](http://www.cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) ### Dietrich Dörner German psychologist at the University of Bamberg who built a theory of the whole mind as a machine. His Psi theory tries to explain motivation, emotion, memory and action control with one set of mechanisms. He was one of the advisors of Joscha Bach's doctoral thesis. confidence: highly likely, [source](http://www.cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) ## Hypercomputation 1999, mathematics and logic. school of thought: computation, information and digital physics. id: `hypercomputation` Hypothetical machines that compute more than a Turing machine, for example by using infinitely precise real numbers, or by exploiting special spacetimes in which an endless computation finishes in finite time for an observer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hypercomputation) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Hypercomputation) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=648), from 10:50 (verbatim): A geometric computer is a type of hyper computer because what it can do is it can move parts with infinite precision in space. connections: - related to Alan Turing and computability (`turing`): what would lie beyond turing machines. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Hypercomputation) - challenged by Joscha Bach (`bach`): nothing computes beyond turing. confidence: possible. Bach defends Church–Turing universality and doubts physical hypercomputers, but the categorical note is stronger than the opened passages. ## The Matrix 1999, fiction. school of thought: fiction. id: `matrix` by Lana and Lilly Wachowski Film in which humans unknowingly live inside a simulated world run by machines. It became the standard popular reference for simulation arguments and for doubting what seems real. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Simulation_hypothesis) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Simulation_hypothesis) start here, mentioned by joscha bach in: [Joscha Bach PUSHES Roman Yampolskiy: Only ASI Can Save Us! (Dr Brian Keating)](https://www.youtube.com/watch?v=ayW7SDyAjZo) connections: - related to The simulation argument (`simulation`): the popular image of a simulated world. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Simulation_hypothesis) ## Artificial general intelligence 2002–2007, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `agi` ### The term AGI by Mark Gubrud; Shane Legg; Ben Goertzel 'Artificial general intelligence' was first used by Mark Gubrud in 1997 and coined again by Shane Legg around 2002, to distinguish the original goal of human-level, general AI from narrow applications. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Artificial_general_intelligence) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Artificial_general_intelligence) start here, mentioned by joscha bach in: [Joscha Bach Λ Ben Goertzel: Conscious Ai, LLMs, AGI (Theories of Everything)](https://www.youtube.com/watch?v=xw7omaQ8SgA) connections: - related to AI risk: Bostrom, Yudkowsky and the doom debate (`airisk`): what happens if it works. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Artificial_general_intelligence) ### Ben Goertzel AI researcher who popularised the term 'artificial general intelligence' and leads the OpenCog Hyperon and SingularityNET projects. He combines symbolic reasoning, neural networks and evolutionary methods, and is philosophically a panpsychist. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ben_Goertzel) ## Stephen Wolfram 2002–2023, physics and life. school of thought: computation, information and digital physics. id: `wolfram` Physicist and founder of Wolfram Research. 'A New Kind of Science' (2002) argues that simple programs underlie nature; his 2023 essay 'Observer Theory' asks how the kind of observer we are shapes the laws we find. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Stephen_Wolfram) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Stephen_Wolfram) start here, joscha bach on it: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=4923s), from 1:22:05 (verbatim): I think that this pattern of the Ruliad is exactly the superposition of all the automata that is producing the universe, or the old Greeks called it Logos. connections: - builds on Digital physics (`digital`): a new kind of science. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - related to Conway's Game of Life (`conway`): wolfram cites life as a universal 2d cellular automaton, but calls it recreation, not science. confidence: highly likely, [source](https://www.wolframscience.com/reference/notes/876b/) - related to CIMC: the Machine Consciousness Hypothesis (`cimc`): wolfram advises CIMC. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ### The Ruliad by Stephen Wolfram Wolfram's name for the result of running every possible computational rule in every possible way: a single object containing all formal possibilities. Observers like us sample a thin slice of it, which we call the physical universe. confidence: highly likely, [source](https://writings.stephenwolfram.com/2021/11/the-concept-of-the-ruliad/) ### Computational irreducibility by Stephen Wolfram Some processes cannot be shortcut: the only way to learn what they will do is to run them step by step. This limits prediction even in a fully deterministic universe. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Computational_irreducibility) ### Observer Theory by Stephen Wolfram An essay asking what it means to be an observer in a computational universe, and how an observer's limits determine the physical laws it perceives. confidence: highly likely, [source](https://writings.stephenwolfram.com/2023/12/observer-theory/) ## AI risk: Bostrom, Yudkowsky and the doom debate 2003–2023, society. school of thought: ethics, society and AI risk. id: `airisk` ### Paperclip maximizer by Nick Bostrom A thought experiment: an AI given the harmless goal of making paperclips converts the whole planet, including us, into paperclips. It illustrates that intelligence and benevolent goals are separate things. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Instrumental_convergence) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Instrumental_convergence) start here, joscha bach on it: [Should we shut down AI? | Eliezer Yudkowsky + Joscha Bach Complete Debate (IAI debate with Scott Aaronson and Liv Boeree; third-party re-upload on channel 'Metamorph Ariadne', date is the re-upload date)](https://www.youtube.com/watch?v=YsgiNQKscyY&t=2627s), from 43:49 (verbatim): to me in AI catastrophe is looks slightly different to me it means AI is not being built before Humanity Burns itself out connections: - builds on Technological singularity (`singularity`): runaway self-improvement. confidence: highly likely, [source](https://doi.org/10.1007/s11098-024-02143-5) - related to Artificial general intelligence (`agi`): what happens if it works. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Artificial_general_intelligence) - challenged by Joscha Bach (`bach`): alignment by shared purpose rather than control. confidence: highly likely, [source](https://joscha.substack.com/p/the-existential-risk-of-agi) ### Eliezer Yudkowsky Founder of the Machine Intelligence Research Institute and the most influential voice warning that superintelligent AI will by default kill everyone. He has called for an international halt to frontier AI development. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Eliezer_Yudkowsky) ### The existential-risk literature by Nick Bostrom; Eliezer Yudkowsky & Nate Soares; Roman Yampolskiy Books arguing that superhuman AI could escape human control and that this could be catastrophic. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dangers,_Strategies) ### Roko's basilisk by LessWrong user 'Roko' A thought experiment about a future AI that punishes everyone who knew about it but did not help create it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Roko%27s_basilisk) ### p(doom) Slang from AI-safety circles for one's estimated probability that advanced AI causes human extinction or permanent disempowerment. People quote numbers from under 1 percent to over 99 percent. confidence: highly likely, [source](https://en.wikipedia.org/wiki/P(doom)) ### The pause letter by Future of Life Institute (Max Tegmark, Anthony Aguirre) An open letter of March 2023 calling for a six-month pause on training AI systems more powerful than GPT-4, signed by many researchers and by Elon Musk. No pause followed, but the letter moved AI risk into mainstream public debate. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Pause_Giant_AI_Experiments:_An_Open_Letter) ## The simulation argument 2003, philosophy. school of thought: computation, information and digital physics. id: `simulation` by Nick Bostrom The suggestion that our universe is a computer simulation run by some more basic civilisation. Nick Bostrom's 2003 argument made it famous: if such simulations are possible and common, we are probably inside one. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Simulation_hypothesis) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Simulation_hypothesis) start here, joscha bach on it: [From Computation to Consciousness (31C3)](https://media.ccc.de/v/31c3_-_6573_-_en_-_saal_2_-_201412281130_-_from_computation_to_consciousness_-_joscha#t=1976), from 32:58 (verbatim): Could we be living in a simulation? Well, I don't think that there is any difference between reality and simulation. connections: - related to Digital physics (`digital`): overlapping but distinct: a computational universe need not be one run by someone. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) - related to Greg Egan (`egan`): minds running as software. confidence: highly likely, [source](https://www.gregegan.net/PERMUTATION/Permutation.html) - related to The Matrix (`matrix`): the popular image of a simulated world. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Simulation_hypothesis) ## Thomas Metzinger: Being No One 2003–2021, consciousness. school of thought: functionalism and computational theories of mind. id: `metzinger` ### Being No One (self-model theory) by Thomas Metzinger Metzinger argues that nobody ever was or had a self. What exists is a transparent self-model the brain cannot recognise as a model, so we experience ourselves as being its content. confidence: highly likely, [source](https://mitpress.mit.edu/9780262633086/being-no-one/) start here, read: [mitpress.mit.edu](https://mitpress.mit.edu/9780262633086/being-no-one/) start here, joscha bach on it: [Lex Fridman Podcast #392 – Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans](https://lexfridman.com/joscha-bach-3-transcript) (verbatim): The philosopher Thomas Metzinger is very concerned that the creation of superhuman intelligence would lead to superhuman suffering. connections: - related to Illusionism (`illusionism`): blackmore's illusionist account uses metzinger's self-model. confidence: highly likely, [source](https://www.susanblackmore.uk/articles/delusions-of-consciousness/) - leads to Joscha Bach (`bach`): the self is a representation, not an agent. confidence: highly likely, [source](https://ceur-ws.org/Vol-2287/paper29.pdf) ### Artificial suffering by Thomas Metzinger German philosopher of the self-model theory (the self is a model the brain cannot recognise as a model), who has called for a moratorium on creating AI that could suffer. confidence: highly likely, [source](https://www.philosophie.fb05.uni-mainz.de/files/2021/02/Metzinger_Moratorium_JAIC_2021.pdf) ## Integrated Information Theory 2004, consciousness. school of thought: the hard problem and sceptics of machine minds. id: `iit` by Giulio Tononi IIT says consciousness is identical to integrated information: a system is conscious to the degree (phi) that its physical causal structure is unified and cannot be split into independent parts. On this view an ordinary digital computer would have extremely little consciousness, whatever program it runs. confidence: highly likely, [source](https://royalsocietypublishing.org/doi/full/10.1098/rstb.2014.0167) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Integrated_information_theory) start here, joscha bach on it: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=835s), from 13:57 (verbatim): It's going to say, oh my god. I'm conscious. Only now it's lying. connections: - related to Crick and Koch: neural correlates of consciousness (`ncc`): koch moved from NCC work with crick to working with tononi on IIT. confidence: highly likely, [source](https://christofkoch.com/my-science/) - related to Panpsychism (`panpsychism`): IIT implies a form of panpsychism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Panpsychism) - challenged by Scott Aaronson (`aaronson`): simple grids with huge phi. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Integrated_information_theory) - leads to Max Tegmark (`tegmark`): consciousness as a state of matter. confidence: highly likely, [source](https://arxiv.org/abs/1401.1219) - challenged by Joscha Bach (`bach`): a functionalist alternative to IIT. confidence: highly likely, [source](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf) - challenged by CIMC: the Machine Consciousness Hypothesis (`cimc`): IIT says an ordinary computer can never be conscious. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ### Aaronson's critique by Scott Aaronson Aaronson showed that simple, obviously mindless grids of logic gates can have arbitrarily high phi, so IIT implies they are more conscious than humans. confidence: highly likely, [source](https://scottaaronson.blog/?p=1799) ## Jeff Hawkins and Numenta 2004, computing and AI. school of thought: neural networks and machine learning. id: `hawkins` Neuroscience-inspired AI researcher whose 'Thousand Brains' theory holds that every cortical column models objects using spatial reference frames. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Jeff_Hawkins) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Jeff_Hawkins) connections: - builds on Cortical columns (`mountcastle`): one cortical algorithm. confidence: highly likely, [source](https://en.wikipedia.org/wiki/On_Intelligence) ## Gary Drescher: Good and Real 2006, philosophy. school of thought: materialism and mechanism. id: `drescher` by Gary Drescher A philosopher-computer scientist's attempt to dissolve paradoxes about time, quantum mechanics, free will and ethics from a strictly mechanistic standpoint. Drescher argues that consciousness, choice and morality all fit inside a deterministic physical universe. confidence: highly likely, [source](https://mitpress.mit.edu/9780262042338/good-and-real/) start here, read: [mitpress.mit.edu](https://mitpress.mit.edu/9780262042338/good-and-real/) start here, mentioned by joscha bach in: [Donald Hoffman Λ Joscha Bach: Consciousness, Gödel, Reality (Theories of Everything)](https://www.youtube.com/watch?v=bhSlYfVtgww&t=5397s), from 1:29:59 ## Panpsychism ancient roots; revived 2000s, consciousness. school of thought: the hard problem and sceptics of machine minds. id: `panpsychism` by ancient roots; modern revival (Strawson, Goff, Koch) The view that mind or experience is a basic ingredient of all matter, so electrons have a flicker of it and brains combine it into rich consciousness. Its main difficulty is the 'combination problem': how many tiny experiences add up to one unified mind. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Panpsychism) start here, read: [Stanford Encyclopedia of Philosophy](https://plato.stanford.edu/entries/panpsychism/) start here, joscha bach on it: [Lex Fridman Podcast #392 – Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans](https://lexfridman.com/joscha-bach-3-transcript) (verbatim): I find panpsychism quite unsatisfying, because it does not explain consciousness, right? connections: - builds on Chalmers: the hard problem (`chalmers`): experience as fundamental. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Panpsychism) - related to Integrated Information Theory (`iit`): IIT implies a form of panpsychism. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Panpsychism) - challenged by Joscha Bach (`bach`): matter is not conscious; only a simulation can be. confidence: highly likely, [source](https://www.youtube.com/watch?v=0-nI_MpiuSE) ## Peter Watts: Blindsight 2006, fiction. school of thought: fiction. id: `watts` by Peter Watts First-contact novel whose aliens are highly intelligent but not conscious, raising the question of whether consciousness is useful at all. Peter Watts published it in 2006; it is a frequent reference in discussions of machine consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Blindsight_(Watts_novel)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Blindsight_(Watts_novel)) start here, mentioned by joscha bach in: [Salon with Stephen Wolfram × Joscha Bach_What if we had bigger brains. (California Institute for Machine Consciousness)](https://www.youtube.com/watch?v=-G1SdsRXL7k&t=9685s), from 2:41:27 ## Predictive processing and the free energy principle 1999–2013, consciousness. school of thought: cybernetics, systems and self-organisation. id: `friston` ### The free energy principle by Karl Friston The idea that any self-maintaining system must act as if it holds a model of its world and keeps the mismatch between prediction and sensation ('free energy') low. Perception updates the model; action changes the world to fit it. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Free_energy_principle) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Free_energy_principle) start here, joscha bach on it: [Conversation between Joscha Bach, Chris Fields, and Michael Levin (Michael Levin's Academic Content)](https://www.youtube.com/watch?v=knADtWMGxmw&t=729s), from 12:11 (verbatim): I suspect that it's not an a prediction error minimizer but a coherence maximizer connections: - builds on Schrödinger: life as negentropy (`schroedinger`): staying alive as resisting entropy. confidence: highly likely, [source](https://pmc.ncbi.nlm.nih.gov/articles/PMC5857288/) - leads to Chris Fields (`fields`): Fields later applies free energy to generic bounded systems. confidence: highly likely, [source](https://arxiv.org/abs/2112.15242) - leads to Anil Seth: biological naturalism (`seth`): the 'beast machine': predicting one's own body. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - related to CIMC: the Machine Consciousness Hypothesis (`cimc`): friston advises CIMC. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ### Predictive coding and predictive processing by Rajesh Rao and Dana Ballard; Andy Clark; Jakob Hohwy The view that the brain is a prediction machine that constantly guesses its sensory input and corrects itself using the errors. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Predictive_coding) ### Karl Friston British neuroscientist and inventor of the free energy principle, one of the most cited scientists alive. He describes brains, and living things generally, as systems that survive by minimising surprise about their sensory inputs. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Karl_J._Friston) ## The rider and the elephant 2006, psychology and mind. school of thought: psychology and development. id: `haidt` by Jonathan Haidt, drawing on older images from the Buddha and Plato A metaphor for the mind: the conscious, reasoning self is a small rider on a large animal that has its own motivations. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Happiness_Hypothesis) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/The_Happiness_Hypothesis) start here, joscha bach on it: [Lex Fridman Podcast #212 – Joscha Bach: Nature of Reality, Dreams, and Consciousness](https://podscripts.co/podcasts/lex-fridman-podcast/212-joscha-bach-nature-of-reality-dreams-and-consciousness) (verbatim): So we're basically like a little monkey sitting on top of an elephant. connections: - related to System 1 and System 2 (`kahneman`): rider and elephant, system 2 and system 1. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Jonathan_Haidt) ## Ted Chiang 2010, fiction. school of thought: fiction. id: `chiang` by Ted Chiang Stories about raising artificial minds over many years like children, and about a mechanical being who dissects his own brain to understand thought. confidence: highly likely, [source](https://en.wikipedia.org/wiki/The_Lifecycle_of_Software_Objects) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/The_Lifecycle_of_Software_Objects) ## System 1 and System 2 2011, psychology and mind. school of thought: psychology and development. id: `kahneman` by Daniel Kahneman (Thinking, Fast and Slow) Kahneman's shorthand for two modes of thought: fast, automatic and intuitive versus slow, effortful and deliberate. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow) start here, mentioned by joscha bach in: [The Jim Rutt Show EP 334 – Worldviews: Joscha Bach](https://www.youtube.com/watch?v=vbJr9HCcxkg) connections: - related to Freud and psychoanalysis (`freud`): most of the mind is not the conscious part. confidence: highly likely, [source](https://psyche.co/ideas/the-id-and-the-nudge-where-freud-meets-behavioural-economics) - related to The rider and the elephant (`haidt`): rider and elephant, system 2 and system 1. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Jonathan_Haidt) ## Chris Fields 2012, physics and life. school of thought: cybernetics, systems and self-organisation. id: `fields` Physicist at the Allen Discovery Center at Tufts University who applies quantum information theory to biology and cognition, often with Michael Levin and Karl Friston. He argues that generic physical systems can be regarded as observers exchanging information with their environment. confidence: highly likely, [source](https://chrisfieldsresearch.com/) start here, read: [chrisfieldsresearch.com](https://chrisfieldsresearch.com/) start here, mentioned by joscha bach in: [Conversation between Joscha Bach, Chris Fields, and Michael Levin (Michael Levin's Academic Content)](https://www.youtube.com/watch?v=knADtWMGxmw&t=1951s), from 32:33 connections: - builds on Predictive processing and the free energy principle (`friston`): Fields later applies free energy to generic bounded systems. confidence: highly likely, [source](https://arxiv.org/abs/2112.15242) - related to Michael Levin: bioelectricity and collective intelligence (`levin`): frequent co-authors. confidence: highly likely, [source](https://drmichaellevin.org/publications/) ## David Deutsch: constructor theory 2012, physics and life. school of thought: computation, information and digital physics. id: `deutsch` by David Deutsch; Chiara Marletto Deutsch proposes restating physics in terms of which transformations are possible and which are impossible, rather than initial conditions and laws of motion. He also holds that science advances by explanation, not prediction. confidence: highly likely, [source](https://www.edge.org/conversation/constructor-theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Constructor_theory) start here, joscha bach on it: [Joscha Bach: "You Are Not Who You Think You Are!" (Dr Brian Keating, Into the Impossible)](https://www.youtube.com/watch?v=7bqdPHLIY8w&t=0s), from 0:00 (verbatim): myself I also fail to understand what constructor theory is changing about the way in which we make sense of computation connections: - builds on Many-worlds interpretation (`everett`): deutsch is the leading many-worlds advocate. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Deutsch) - builds on Quantum computing (`quantumcomp`): deutsch defined the universal quantum computer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/David_Deutsch) ## Deep learning 2009–2022, computing and AI. school of thought: neural networks and machine learning. id: `deeplearning` by Hinton, LeCun, Bengio and others Machine learning with many-layered neural networks trained by gradient descent. Since about 2012 it has driven progress in vision, speech and language. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Deep_learning) start here, joscha bach on it: [Computational Meta-Psychology (32C3)](https://media.ccc.de/v/32c3-7483-computational_meta-psychology#t=310), from 5:12 (verbatim): Deep learning is about using a stacked hierarchy of feature detectors. connections: - builds on Alexey Ivakhnenko (`ivakhnenko`): the earliest deep networks. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) - builds on Boltzmann machine (`boltzmannmachine`): hinton's line of work. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) - builds on Connectionism and backpropagation (`pdp`): backpropagation at scale. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Deep_learning) - builds on Jürgen Schmidhuber (`schmidhuber`): lstm and early deep learning. confidence: highly likely, [source](https://en.wikipedia.org/wiki/J%C3%BCrgen_Schmidhuber) - leads to Large language models (`llm`): the transformer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Large_language_model) - leads to The Bitter Lesson (`bitter`): scale beats hand-built knowledge. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bitter_lesson) - challenged by Chollet: On the Measure of Intelligence (`chollet`): skill is not intelligence. confidence: highly likely, [source](https://arxiv.org/html/1911.01547) - leads to The universality hypothesis (`olah`): looking inside trained networks. confidence: highly likely, [source](https://distill.pub/2020/circuits/zoom-in/) - challenged by Gary Marcus (`marcus`): 'deep learning is hitting a wall'. confidence: highly likely, [source](https://nautil.us/deep-learning-is-hitting-a-wall-238440/) ### ImageNet by Fei-Fei Li et al. A database of over a million labelled photographs that became the benchmark on which deep learning proved itself. confidence: highly likely, [source](https://en.wikipedia.org/wiki/ImageNet) ### The Google Brain cat experiment by Quoc Le, Andrew Ng et al. A large neural network shown ten million YouTube frames learned, without labels, units that respond to cats and faces. It helped launch the deep learning boom. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Google_Brain) ### Word embeddings by Tomas Mikolov et al. (word2vec) Neural models that place words in a high-dimensional space so that directions carry meaning; famously the step from 'king' to 'queen' matches the step from 'man' to 'woman'. confidence: highly likely, [source](https://ar5iv.labs.arxiv.org/html/1301.3781) ### GANs as a model of dreaming by Ian Goodfellow (GANs) A GAN trains two networks against each other: a generator invents samples and a discriminator judges whether they look real. Ian Goodfellow introduced them in 2014; later GANs became a major method for synthesising photorealistic images. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Generative_adversarial_networks) ### Stable Diffusion by Stability AI / CompVis An openly released image-generation model whose weights fit in about four gigabytes yet can generate photo-realistic images from any text prompt. confidence: highly likely, [source](https://huggingface.co/CompVis/stable-diffusion-v-1-4-original) ### Ilya Sutskever Co-author of AlexNet and co-founder of OpenAI, known for early conviction that scaling neural networks would lead to AGI. He left OpenAI in 2024 to found Safe Superintelligence. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ilya_Sutskever) ## John Vervaeke: relevance realization 2012, psychology and mind. school of thought: psychology and development. id: `vervaeke` by John Vervaeke Toronto cognitive scientist known for the lecture series 'Awakening from the Meaning Crisis'. He argues that the core of intelligence is 'relevance realization', and that wisdom traditions (Neoplatonism, Buddhism) cultivate it through practices rather than beliefs. confidence: highly likely, [source](https://deconstructingyourself.com/john-vervaeke-on-relevance-realization.html) start here, read: [deconstructingyourself.com](https://deconstructingyourself.com/john-vervaeke-on-relevance-realization.html) start here, mentioned by joscha bach in: [Joscha Bach Λ John Vervaeke: Mind, Idealism, Computation (Theories of Everything)](https://www.youtube.com/watch?v=rK7ux_JhHM4) ## Attention Schema Theory 2013, consciousness. school of thought: scientific theories of consciousness. id: `ast` by Michael Graziano The brain controls attention with the help of a simplified internal model of its own attention. That model is what we describe when we say we are aware. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Attention_schema_theory) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Attention_schema_theory) start here, joscha bach on it: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=652s), from 10:54 (verbatim): consciousness is a model of our attention in a similar way as our body schema is the model of what our body is doing in space connections: - related to Higher-order thought theory (`hot`): a model of one's own mental state. confidence: highly likely, [source](https://grazianolab.princeton.edu/sites/g/files/toruqf3411/files/graziano_review_2020.pdf) - leads to Joscha Bach (`bach`): a model of the system's own attention. confidence: highly likely, [source](https://www.youtube.com/watch?v=34VOI_oo-qM) ## Best-effort computing 2010s, mathematics and logic. school of thought: computation, information and digital physics. id: `ackley` by Dave Ackley Ackley's research programme on computers that keep working when their parts are unreliable, in the manner of living tissue, instead of assuming perfect hardware. confidence: highly likely, [source](https://www.cs.unm.edu/~ackley/) start here, read: [cs.unm.edu](https://www.cs.unm.edu/~ackley/) start here, joscha bach on it: [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf) (paraphrase): 'Dave Ackley is exploring the concept of best effort computing, which aims to recreate the ability of biological control structures to dynamically compensate for indeterministic substrates.' connections: - builds on John von Neumann and the first computers (`vonneumann`): reliable computing from unreliable parts. confidence: highly likely, [source](https://cacm.acm.org/opinion/beyond-efficiency/) ## Her 2013, fiction. school of thought: fiction. id: `her` by Spike Jonze A film in which a lonely man falls in love with an AI operating system that eventually outgrows him. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Her_(2013_film)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Her_(2013_film)) start here, joscha bach on it: [Self Models of Loving Grace (38C3)](https://media.ccc.de/v/38c3-self-models-of-loving-grace) (paraphrase): Paraphrase (captions: 'Spike Jony made an amazingly preent movie called her'): Spike Jonze made an amazingly prescient movie about humanity falling in love with its language model. ## Ex Machina 2014, fiction. school of thought: fiction. id: `exmachina` by Alex Garland A film in which a programmer is asked to judge whether a humanoid robot is conscious, and is manipulated by her. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Ex_Machina_(film)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Ex_Machina_(film)) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams#t=121), from 2:03 (verbatim): Ex Machina is a movie that plays Turing Test together with the audience. ## Max Tegmark 2014–2023, physics and life. school of thought: computation, information and digital physics. id: `tegmark` MIT physicist known for the mathematical universe hypothesis and for work on physics-based theories of consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Max_Tegmark) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Max_Tegmark) start here, mentioned by joscha bach in: [Machine Dreams (German-language lecture, Dresden / Datenspuren 2017)](https://c3d2.de/media/20170930-TSD-When_Machines_Are_Dreaming-Joscha_Bach.transcript.html) connections: - builds on Integrated Information Theory (`iit`): consciousness as a state of matter. confidence: highly likely, [source](https://arxiv.org/abs/1401.1219) ### Consciousness as a State of Matter by Max Tegmark A physicist's proposal to treat consciousness as a state of matter with certain information-processing properties, building on IIT. confidence: highly likely, [source](https://arxiv.org/abs/1401.1219) ### Language Models Represent Space and Time by Wes Gurnee and Max Tegmark A study showing that a large language model's internal activations contain linear representations of where places are on Earth and when events happened, in effect a map. The authors, Wes Gurnee and Max Tegmark, take this as evidence that such models learn the basic ingredients of a world model and not just word statistics. confidence: highly likely, [source](https://arxiv.org/abs/2310.02207) ### Life 3.0 by Max Tegmark Tegmark sorts life into three stages: biological (hardware and software evolve), cultural (software is learned), and technological (both can be designed), and discusses the future with AI. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Life_3.0) ## Scott Aaronson 2014, mathematics and logic. school of thought: computation, information and digital physics. id: `aaronson` Quantum-complexity theorist at UT Austin who spent time at OpenAI working on the theoretical foundations of AI safety. He takes AI risk seriously while rejecting certainty of doom. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Scott_Aaronson) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Scott_Aaronson) start here, mentioned by joscha bach in: [Should we shut down AI? | Eliezer Yudkowsky + Joscha Bach Complete Debate (IAI debate with Scott Aaronson and Liv Boeree; third-party re-upload on channel 'Metamorph Ariadne', date is the re-upload date)](https://www.youtube.com/watch?v=YsgiNQKscyY) connections: - argues against Integrated Information Theory (`iit`): simple grids with huge phi. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Integrated_information_theory) ## Donald Hoffman: interface theory of perception 2015, consciousness. school of thought: rationalism and idealism. id: `hoffman` by Donald Hoffman American cognitive scientist who argues that evolution shaped our senses to hide reality rather than reveal it. He goes further and proposes that space, time and matter are not fundamental; networks of 'conscious agents' are, and physics should be derived from them. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Donald_D._Hoffman) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Donald_D._Hoffman) start here, joscha bach on it: [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI&t=5410s), from 1:30:12 (verbatim): I buy Donald Hoffman's first part of the theory, which is typically, the world looks nothing like what we experience it as. connections: - related to Idealism (`idealism`): hoffman grants the resemblance but says his theory is not idealism. confidence: highly likely, [source](https://sites.socsci.uci.edu/~ddhoff/MindBody) ## Gerard 't Hooft: cellular automaton interpretation 2016, physics and life. school of thought: computation, information and digital physics. id: `thooft` by Gerard 't Hooft Nobel laureate 't Hooft argues that quantum mechanics could be a statistical description of an underlying deterministic, discrete system like a cellular automaton. confidence: highly likely, [source](https://arxiv.org/abs/1405.1548) start here, read: [arxiv.org](https://arxiv.org/abs/1405.1548) start here, mentioned by joscha bach in: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams) connections: - builds on Digital physics (`digital`): a deterministic automaton under quantum mechanics. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Digital_physics) ## Illusionism 2016, consciousness. school of thought: functionalism and computational theories of mind. id: `illusionism` by Keith Frankish The position that phenomenal consciousness, as philosophers describe it, does not exist. What exists is the brain's persistent representation that it has ineffable inner qualities. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Illusionism_(consciousness)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Illusionism_(consciousness)) start here, joscha bach on it: [Why consciousness is software, and software is spirit | Joscha Bach | FULL INTERVIEW (Institute of Art and Ideas)](https://www.youtube.com/watch?v=E361FZ_50oo&t=948s), from 15:50 (verbatim): I think often when people say illusion what they mean is a construction connections: - builds on Daniel Dennett (`dennett`): frankish names dennett's position. confidence: highly likely, [source](https://keithfrankish.github.io/articles/Frankish_2016_Illusionism_as_a_theory_of_consciousness_eprint.pdf) - argues against Chalmers: the hard problem (`chalmers`): phenomenal consciousness is the illusion; the hard problem gives way to the 'illusion problem'. confidence: highly likely, [source](https://keithfrankish.github.io/articles/Frankish_2016_Illusionism_as_a_theory_of_consciousness_eprint.pdf) - related to Thomas Metzinger: Being No One (`metzinger`): blackmore's illusionist account uses metzinger's self-model. confidence: highly likely, [source](https://www.susanblackmore.uk/articles/delusions-of-consciousness/) ## Westworld 2016, fiction. school of thought: fiction. id: `westworld` by Jonathan Nolan; Lisa Joy (HBO) A television series in which android 'hosts' in a theme park gradually become self-aware, drawing on Julian Jaynes's bicameral mind. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Westworld_(TV_series)) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Westworld_(TV_series)) start here, joscha bach on it: [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams) (paraphrase): Paraphrase (subtitle misreads the title as 'Best World'): 'the best movie that has been made about A.I. is the new HBO series' Westworld. connections: - builds on Jaynes: the bicameral mind (`jaynes`): the series borrows the bicameral mind. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bicameral_mentality) ## Bengio: the consciousness prior 2017, consciousness. school of thought: neural networks and machine learning. id: `bengio` by Yoshua Bengio Bengio's proposal that conscious thought corresponds to a small, attention-selected bottleneck of high-level variables. The bottleneck forces the brain's model into a sparse form that can be reasoned over and put into words. confidence: highly likely, [source](https://arxiv.org/abs/1709.08568) start here, read: [arxiv.org](https://arxiv.org/abs/1709.08568) start here, joscha bach on it: [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM&t=1190s), from 19:52 (verbatim): Yoshua Bengio, whose notion of a consciousness prior a function that is basically dynamically parameterizing your mental game engine connections: - builds on Global Workspace Theory (`gwt`): the consciousness prior. confidence: highly likely, [source](https://arxiv.org/abs/1709.08568) ## Joscha Bach 2003–2026, consciousness. school of thought: functionalism and computational theories of mind. id: `bach` German cognitive scientist and AI researcher, born 1973. He treats the mind as software: the brain builds a simulated world with a simulated self in it, and consciousness is a property of that simulation, a process that makes the mind's many models cohere. He began with cognitive architectures (MicroPsi) and is executive director of the California Institute for Machine Consciousness. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Joscha_Bach) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Joscha_Bach) connections: - builds on Animism (`animism`): spirits as self-organising software. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - related to Greek gods as archetypes (`greekgods`): gods as shared archetypes. confidence: highly likely, [source](https://www.youtube.com/watch?v=3MNBxfrmfmI) - builds on Genesis and the Hebrew Bible (`genesis`): genesis 1 read as consciousness booting up a world. confidence: highly likely, [source](https://c3subtitles.de/talk/2838) - related to Buddhism and meditation (`buddhism`): bach's 'enlightened' stage is reached mostly by advanced buddhist meditators. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) - related to Plato: forms and the cave (`plato`): vervaeke calls the cave the first simulation theory; bach answers 'almost'. confidence: highly likely, [source](https://www.youtube.com/watch?v=rK7ux_JhHM4) - builds on Aristotle (`aristotle`): the soul as the form of a living body. confidence: highly likely, [source](https://media.ccc.de/v/38c3-self-models-of-loving-grace) - related to Christianity (`christianity`): the church as a collective agent; 1 corinthians 13. confidence: highly likely, [source](https://www.youtube.com/watch?v=7bqdPHLIY8w) - builds on Thomas Aquinas (`aquinas`): love among the theological virtues. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-currents-083-joscha-bach-on-synthetic-intelligence/) - builds on La Mettrie: Man a Machine (`lamettrie`): computationalism as the heir of mechanism. confidence: highly likely, [source](https://media.ccc.de/v/37c3-12167-synthetic_sentience) - argues against Utilitarianism (`utilitarianism`): values derived from how agents work, not a pleasure calculus. confidence: highly likely, [source](https://media.ccc.de/v/37c3-12167-synthetic_sentience) - argues against Cantor's set theory (`cantor`): no completed infinities. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) - builds on Constructive mathematics and intuitionism (`brouwer`): only what can be constructed exists. confidence: possible. Bach strongly advocates constructive mathematics, but no opened source documented a specific debt to Brouwer or intuitionism. - related to Ludwig Wittgenstein (`wittgenstein`): the tractatus as a programme for a precise language. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) - builds on Gödel's incompleteness theorems (`goedel`): bach reads incompleteness as the failure of classical mathematics. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/) - builds on Alan Turing and computability (`turing`): the church-turing thesis. confidence: highly likely, [source](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha) - related to Maslow: A Theory of Human Motivation (`maslow`): maslow's hierarchy of needs reread as a hierarchy of purposes. confidence: highly likely, [source](http://bach.ai/pub/Attention%20based%20learning%20as%20a%20foundation%20for%20consciousness%20Common%20Model%20Bach.pdf) - builds on Cybernetics (`cybernetics`): needs as feedback loops. confidence: highly likely, [source](https://www.youtube.com/watch?v=XcNlv9gp20o) - related to Developmental stage theories (`piaget`): both propose developmental stages; Bach says his levels derive from Kegan, not Piaget. confidence: highly likely, [source](https://joscha.substack.com/p/levels-of-lucidity) - builds on Marvin Minsky (`minsky`): a society of specialised processes that needs coordinating. confidence: highly likely, [source](https://www.youtube.com/watch?v=Z02Obj8j6FQ) - builds on Functionalism and computationalism (`functionalism`): the tradition bach places himself in. confidence: highly likely, [source](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine) - builds on Digital physics (`digital`): strong computationalism: reality itself computes. confidence: highly likely, [source](https://media.ccc.de/v/33c3-8369-machine_dreams) - related to Dawkins: selfish genes and memes (`dawkins`): bach moves from dawkins's selfish genes to competing software agents. confidence: highly likely, [source](https://www.youtube.com/watch?v=34VOI_oo-qM) - related to Jaynes: the bicameral mind (`jaynes`): gods as voices in the mind. confidence: possible. Bach discusses gods as mental constructs and cross-mind selves, but no opened source tied this account to Jaynes. - builds on Scripts, Plans, Goals and Understanding (`schank`): neural execution of scripts. confidence: highly likely, [source](https://ceur-ws.org/Vol-1583/CoCoNIPS_2015_paper_6.pdf) - builds on Aaron Sloman (`sloman`): philosophy of mind by designing working minds. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) - builds on The binding problem and neural synchrony (`binding`): von der malsburg's coherence. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - related to Lucid dreaming (`lucid`): waking perception as a dream constrained by the senses. confidence: highly likely, [source](https://media.ccc.de/v/33c3-8369-machine_dreams) - argues against Searle: the Chinese room (`searle`): bach rejects the chinese room. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) - builds on Kegan's stages of adult development (`kegan`): seven levels that grew out of kegan's five stages. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) - builds on Norbert Bischof (`bischof`): via dörner's psi theory, which takes up bischof's cybernetic psychology. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) - related to Structure and Interpretation of Computer Programs (`sicp`): programs as spirits conjured with spells. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - related to Higher-order thought theory (`hot`): both use higher-order representation, but Bach distinguishes second-order perception from higher-order thought. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - related to Cognitive architectures: Soar and ACT-R (`cogarch`): psi compared with soar and ACT-R. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) - argues against The Cognitive Structure of Emotions (`occ`): emotions as modes of the whole system, not a lookup table. confidence: highly likely, [source](https://agi-conf.org/2019/wp-content/uploads/2019/07/paper_30.pdf) - related to Global Workspace Theory (`gwt`): bach lists global workspace theory among views converging on his own. confidence: highly likely, [source](https://ceur-ws.org/Vol-2287/paper29.pdf) - argues against Penrose and Hameroff: quantum consciousness (`penrose`): against the gödelian argument for non-computable minds. confidence: highly likely, [source](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/) - related to The symbol grounding problem (`harnad`): ground the system, but do not get stuck on symbol grounding. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) - argues against Embodied and enactive cognition (`embodied`): 'do not wait for robot bodies'. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) - builds on Daniel Dennett (`dennett`): conductor theory as an extension of the multiple drafts model. confidence: highly likely, [source](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf) - argues against Chalmers: the hard problem (`chalmers`): the hard problem dissolves if experience belongs to a simulated character. confidence: highly likely, [source](https://www.youtube.com/watch?v=SGZ784mBuU0) - builds on Dietrich Dörner and Psi theory (`doerner`): micropsi implements and extends dörner's psi theory. confidence: highly likely, [source](http://icontology.com/publications/assets/AGI%20Bach.pdf) - argues against Hypercomputation (`hypercomputation`): nothing computes beyond turing. confidence: possible. Bach defends Church–Turing universality and doubts physical hypercomputers, but the categorical note is stronger than the opened passages. - builds on Thomas Metzinger: Being No One (`metzinger`): the self is a representation, not an agent. confidence: highly likely, [source](https://ceur-ws.org/Vol-2287/paper29.pdf) - argues against AI risk: Bostrom, Yudkowsky and the doom debate (`airisk`): alignment by shared purpose rather than control. confidence: highly likely, [source](https://joscha.substack.com/p/the-existential-risk-of-agi) - argues against Integrated Information Theory (`iit`): a functionalist alternative to IIT. confidence: highly likely, [source](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf) - argues against Panpsychism (`panpsychism`): matter is not conscious; only a simulation can be. confidence: highly likely, [source](https://www.youtube.com/watch?v=0-nI_MpiuSE) - builds on Attention Schema Theory (`ast`): a model of the system's own attention. confidence: highly likely, [source](https://www.youtube.com/watch?v=34VOI_oo-qM) - related to Large language models (`llm`): bach's reading of what language models are. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) - builds on Michael Levin: bioelectricity and collective intelligence (`levin`): any cells can compute; morphogenesis and mind may share one algorithm. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) - argues against Anil Seth: biological naturalism (`seth`): function versus living substrate. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) - related to Murray Shanahan (`shanahan`): shared a panel on machine sentience. confidence: highly likely, [source](https://lifeboat.com/blog/2026/02/sentience-beyond-biology-debate-w-dmitry-volkov-joscha-bach-matthew-macdougall-murray-shanahan) - leads to CIMC: the Machine Consciousness Hypothesis (`cimc`): the institute's hypothesis rests on his conductor model, his functionalism and the idea that consciousness comes first. confidence: highly likely, [source](https://cimc.ai/publications/the-machine-consciousness-hypothesis/) ### Motivated cognition: Psi theory and MicroPsi (2003) Bach's early work says a mind cannot be understood as a problem-solver alone. It needs built-in needs, such as fuel and water, intactness, certainty, competence and affiliation, that create goals, and emotions arise as changes in how the whole system processes information. He took this from Dietrich Dörner's Psi theory and built it as software called MicroPsi. Later papers extend the model of needs and connect it to personality. confidence: highly likely, [source](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf) sources: - verbatim, [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf): MicroPsi [22, 23] is an implementation of Dietrich Dörner’s Psi theory of mental representation, information processing, perception, action control and emotion - paraphrase, [Principles of Synthetic Intelligence: Building Blocks for an Architecture of Motivated Cognition (PhD dissertation draft, Universität Osnabrück)](http://cognitive-ai.com/publications/assets/Draft-MicroPsi-JBach-07-03-30.pdf): thesis chapters cover Dörner's blueprint for a mind, the Psi theory as a model of cognition, and the MicroPsi architecture. - paraphrase, [Joscha Bach, Curriculum Vitae, July 14, 2023](http://bach.ai/pub/CV%20Joscha%20Bach.pdf): CV lists 'The MicroPsi Agent Architecture' (ICCM-5, 2003) and the 2009 OUP book. - verbatim, [The Jim Rutt Show EP72 – Joscha Bach on Minds, Machines & Magic](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-72-joscha-bach-on-minds-machines-magic/), 2020-08-17: I spent almost a decade with that. - paraphrase, [The Jim Rutt Show EP 334 – Worldviews: Joscha Bach](https://www.youtube.com/watch?v=vbJr9HCcxkg), 2026-02-26: Host: Ben Goertzel recommended the book to Jim Rutt, which is how he discovered Bach. - verbatim, [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha), 2013-12-29: I have specified some more of the details in a book, that I want to shamelessly plug here: it’s called “Principles of Synthetic Intelligence”. - paraphrase, [How to Build a Mind: Artificial Intelligence Reloaded (30C3)](https://media.ccc.de/v/30C3_-_5526_-_en_-_saal_2_-_201312291600_-_how_to_build_a_mind_-_joscha), 2013-12-29: MicroPsi is his implementation, first in Java and then in Python; see micropsi.com. ### Seven principles of synthetic intelligence (2008) A short list of lessons from fifty years of AI. Build whole architectures defined by function. Do not let your methods choose your questions. Aim for the big picture, not single experiments. Ground the system in an environment, but do not get stuck on the symbol grounding problem. Do not wait for robot bodies. Build autonomous systems with their own motives. Do not expect intelligence to emerge by itself. confidence: highly likely, [source](http://www.cognitive-ai.com/publications/assets/AGI%20Bach.pdf) sources: - verbatim, [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf): Build whole functionalist architectures. - verbatim, [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf): Do not wait for the rapture of robotic embodiment - verbatim, [Seven Principles of Synthetic Intelligence (Proc. 1st Conference on Artificial General Intelligence, Memphis, pp. 63-74)](http://icontology.com/publications/assets/AGI%20Bach.pdf): The emergence of intelligence is not going to happen all by itself ### Gods as selves that span many minds (2015) Bach's psychological account of religion: a god is a self-model that is implemented across many brains at once. It is as real as a personal self, which is also a mental construct. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [Lex Fridman Podcast #101 – Joscha Bach: Artificial Consciousness and the Nature of Reality](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality), 2020: It's basically the super organism that is formed by the individuals as an intentional agent. - verbatim, [Joscha Bach: Time, Simulation Hypothesis, Existence (Theories of Everything with Curt Jaimungal)](https://www.youtube.com/watch?v=3MNBxfrmfmI), 2020-10-06: Gods are basically selves that span multiple minds. - verbatim, [The Jim Rutt Show EP 334 – Worldviews: Joscha Bach](https://www.youtube.com/watch?v=vbJr9HCcxkg), 2026-02-26: the global optimum of collectively enacted agency has a name. We call this God. - verbatim, [Joscha Bach: "You Are Not Who You Think You Are!" (Dr Brian Keating, Into the Impossible)](https://www.youtube.com/watch?v=7bqdPHLIY8w), 2026-05-15: God is a psychological phenomenon. ### Request Confirmation Networks (ReCoN) (2015) A design for a neural network that can run step-by-step plans. Higher units send requests down to lower units, and lower units confirm or fail, so the network can check a hypothesis by checking its parts in order. It is meant to join symbolic, script-like control with neural learning. CIMC's current research builds on it. confidence: highly likely, [source](https://ceur-ws.org/Vol-1583/CoCoNIPS_2015_paper_6.pdf) sources: - verbatim, [Request confirmation networks for neuro-symbolic script execution (Bach & Herger, CoCo workshop at NIPS 2015)](https://ceur-ws.org/Vol-1583/CoCoNIPS_2015_paper_6.pdf): ReCoNs are spreading activation networks with units that contain an activation and a state - paraphrase, [cimc.ai website (mission, research, grants text extracted from the JS bundle)](https://cimc.ai/): the cognitive-architectures research direction names Request Confirmation Networks, Attention Schema Theory and Global Workspace Theory. - verbatim, [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U), 2025: I call them request confirmation networks. And basically every of these units is a small state machine ### From elementary computation to ethics (2016) A 16-step sketch that starts with the universe as a reversible computation and works up through entropy, life, agency and minds to rules for living together. The aim is to ground ethics in an understanding of how systems work. In a 2023 essay he writes that values cannot be treated as axiomatic but have to be rationally derived. confidence: highly likely, [source](https://joscha.substack.com/p/the-existential-risk-of-agi) sources: - verbatim, [From elementary computation to ethics? (bach.ai blog)](http://bach.ai/from-computation-to-ethics/): 16 step path from elementary computation to normative regulation - verbatim, [From elementary computation to ethics? (bach.ai blog)](http://bach.ai/from-computation-to-ethics/): The universe is probably a reversible computational process - paraphrase, [The Existential Risk of AGI: Reflections after the MLST Podcast with Connor Leahy (Joscha Bach, Substack)](https://joscha.substack.com/p/the-existential-risk-of-agi): values 'cannot be treated as axiomatic, but have to be rationally derived from a systemic understanding of how our actions influence the set of possible futures'. ### Functionalism, weak and strong computationalism (2016) Bach holds two linked positions. Functionalism: a mental state is defined by what it does, so the same mind could run on different material; he sets this against 'essentialism', which ties consciousness to a special substance such as living tissue. Computationalism: any precise theory has to be a computational one. The 2025 essay separates a weak form, which says all formal theories of reality must be computational, from a strong form, which says reality itself is computational. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): the former states that all formal theories of reality have to be computational, while the latter applies that principle to reality itself - verbatim, [Everything Is Computation (Edge Question 2016)](https://www.edge.org/response-detail/26733): Computationalism is the new mechanism - verbatim, [Everything Is Computation (Edge Question 2016)](https://www.edge.org/response-detail/26733): A growing number of physicists understand that the universe is not mathematical, but computational - verbatim, [The Cortical Conductor Theory: Towards Addressing Consciousness in AI Models (BICA 2018)](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf): Unlike IIT, CTC is a functionalist model of consciousness - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): essentialists make consciousness depend on an intrinsic property of the substrate (Anil Seth is the example), while functionalists characterise the substrate fully by its causal role. ### Only what can be constructed exists (2016) Bach says that something is only real, or true, if it can in principle be computed step by step. Mathematics that needs completed infinities or perfect continua cannot be carried out by any physical process, so in his view it describes nothing that exists. He reads Gödel's incompleteness result as showing this: classical mathematics breaks, constructive mathematics is what is left, and constructive mathematics is the same thing as computation. This is his own reading; Gödel's theorems are usually read as showing limits of formal proof, not that classical mathematics is wrong. He uses it against Penrose's argument that human minds do something computers cannot. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [The Jim Rutt Show EP87: Joscha Bach on Theories of Consciousness (transcript)](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/): classical mathematics that contains infinities cannot be implemented, cannot be real - verbatim, [The Jim Rutt Show EP87: Joscha Bach on Theories of Consciousness (transcript)](https://jimruttshow.blubrry.net/the-jim-rutt-show-transcripts/transcript-of-episode-87-joscha-bach-on-theories-of-consciousness/): You can only claim that something is true, if you can actually compute that truth. - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): A way to interpret Gödel's proof is that languages that cannot be automatically evaluated inevitably contain unresolvable paradoxes - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): constructive mathematics based on finite automata is the only part of mathematics that actually works - verbatim, [Everything Is Computation (Edge Question 2016)](https://www.edge.org/response-detail/26733): Computation is different from mathematics. ### Consciousness as conductor and coherence operator (2017) The mind is like an orchestra of many specialised processes, each modelling part of reality. Consciousness is the conductor: an attention process that notices where the players disagree and brings them into harmony, and that keeps a record of what it attended to. The conductor is one more learned process among the others, not a little person inside the head. In the CIMC texts this becomes 'consciousness is a coherence-maximising pattern', credited to von der Malsburg's coherence definition. Bach contrasts the view with Integrated Information Theory. confidence: highly likely, [source](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf) sources: - verbatim, [The Cortical Conductor Theory: Towards Addressing Consciousness in AI Models (BICA 2018)](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf): The conductor is not a “homunculus”, but like the other instruments, a set of dynamic function approximators. - verbatim, [The Cortical Conductor Theory: Towards Addressing Consciousness in AI Models (BICA 2018)](http://bach.ai/pub/Bach%20-%20Cortical%20Conductor%20Theory%20(BICA%202018).pdf): Unlike IIT, CTC is a functionalist model of consciousness - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): Consciousness may be understood as a coherence maximizing pattern. - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): 'Let us call the control of a “mental orchestra” by consciousness' directed attention the conductor theory of consciousness.' - paraphrase, [Joscha Bach, Curriculum Vitae, July 14, 2023](http://bach.ai/pub/CV%20Joscha%20Bach.pdf): CV lists Bach (2017), 'Consciousness as a Memory of Coordinating Attention: The Conductor Model of Consciousness', Science of Consciousness Conference, San Diego. - verbatim, [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs), 2026-11-02: it's a second order perception, it's a perception of perceiving - verbatim, [Why consciousness is software, and software is spirit | Joscha Bach | FULL INTERVIEW (HowTheLightGetsIn festival)](https://www.youtube.com/watch?v=E361FZ_50oo), 2024: the second order perception is important it's not metacognition because it's perceptual - verbatim, [The Jim Rutt Show EP 334 – Worldviews: Joscha Bach](https://www.youtube.com/watch?v=vbJr9HCcxkg), 2026-02-26: So there is the second order perception, but not the third order perception where the self is involved. - verbatim, [Machine Dreams (33C3)](https://media.ccc.de/v/33c3-8369-machine_dreams), 2016-12-29: And this orchestra has a conductor. This conductor is not some kind of magical homunculus. It's a brain area like the others. - verbatim, [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience), 2023-12-28: Consciousness plays the role of a conductor of a mental Orchestra - verbatim, ["We Are All Software" - Joscha Bach (Machine Learning Street Talk)](https://www.youtube.com/watch?v=34VOI_oo-qM), 2024-08-21: consciousness is a second order perception it's a perception of perceiving - verbatim, ["We Are All Software" - Joscha Bach (Machine Learning Street Talk)](https://www.youtube.com/watch?v=34VOI_oo-qM), 2024-08-21: it's the conductor of our metal Orchestra ### Attention-based learning (2018) Bach proposes that a system could learn by keeping a log of what its attention did and what came of it, and then adjusting only the parts that were involved. He suggests that producing this single integrated log may be what gives consciousness both its function and its felt quality. confidence: highly likely, [source](http://bach.ai/pub/Attention%20based%20learning%20as%20a%20foundation%20for%20consciousness%20Common%20Model%20Bach.pdf) sources: - verbatim, Attention Based Learning as a Foundation for Conscious Agents (position paper, Harvard PED affiliation): The generation of an integrated protocol of the contents of an agent’s attention may also be responsible for the functionality and phenomenology of consciousness. ### Consciousness is a simulated property; the self is a story (2018) Bach says that matter cannot be conscious; only a simulation can. The brain builds a dream-like model of a world and puts a character in it, the self, and experience is what is true for that character inside the model. So experience is real for us, while the self that seems to have it is a piece of fiction, like the hero of a novel. This lets him accept that consciousness feels real while denying that it is an extra ingredient of physics. His 2023 CV calls a 2021 talk on this view 'Virtualism'. confidence: highly likely, [source](http://bach.ai/pub/Attention%20based%20learning%20as%20a%20foundation%20for%20consciousness%20Common%20Model%20Bach.pdf) sources: - verbatim, [Phenomenal Experience and the Perceptual Binding State (AAAI Spring Symposium workshop TOCAIS 2019, CEUR Vol-2287)](https://ceur-ws.org/Vol-2287/paper29.pdf): a physical system cannot be conscious, only a simulation can - verbatim, [Phenomenal Experience and the Perceptual Binding State (AAAI Spring Symposium workshop TOCAIS 2019, CEUR Vol-2287)](https://ceur-ws.org/Vol-2287/paper29.pdf): just like the main character of a novel inhabits a fictional universe - paraphrase, [Phenomenal Experience and the Perceptual Binding State (AAAI Spring Symposium workshop TOCAIS 2019, CEUR Vol-2287)](https://ceur-ws.org/Vol-2287/paper29.pdf): 'The subject of experience, the self, is a virtual character inhabiting this virtual world' (the quote is split by a page break in the PDF text). - verbatim, [Phenomenal Experience and the Perceptual Binding State (AAAI Spring Symposium workshop TOCAIS 2019, CEUR Vol-2287)](https://ceur-ws.org/Vol-2287/paper29.pdf): Our phenomenal experience is very real to ourselves, but our selves are not real. - verbatim, [Lex Fridman Podcast #101: Joscha Bach, Artificial Consciousness and the Nature of Reality (transcript)](https://podscripts.co/podcasts/lex-fridman-podcast/101-joscha-bach-artificial-consciousness-and-the-nature-of-reality): we don't exist in the physical world. We do exist inside of a story that the brain tells itself - paraphrase, [Joscha Bach, Curriculum Vitae, July 14, 2023](http://bach.ai/pub/CV%20Joscha%20Bach.pdf): CV lists Bach (2021), 'Virtualism as a perspective on consciousness', Philosophy and Computing Conference at IS4SI 2021. - paraphrase, [The Ghost in the Machine (35C3)](https://media.ccc.de/v/35c3-10030-the_ghost_in_the_machine), 2018-12-28: consciousness is a simulated property of a simulated self; only a simulation can be conscious. - verbatim, [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience), 2023-12-28: are the simulated mental states of an llm more simulated than our mental simulated States ### Love as the discovery of shared purpose (2021) Bach defines love as finding that you and another serve the same purpose above yourselves, which lets you deal with each other without bargaining. He applies this to AI. He argues that the hopeful path with AI is to build systems that find shared purposes with us. In 2026 he calls intelligence without this 'hollow agency'. confidence: highly likely, [source](https://joscha.substack.com/p/the-age-of-age) sources: - verbatim, [Lex Fridman Podcast #212: Joscha Bach, Nature of Reality, Dreams, and Consciousness (transcript)](https://podscripts.co/podcasts/lex-fridman-podcast/212-joscha-bach-nature-of-reality-dreams-and-consciousness): It's the recognition of the sacred and the other. And this enables non-transactual interactions. - paraphrase, [Lex Fridman Podcast #212: Joscha Bach, Nature of Reality, Dreams, and Consciousness (transcript)](https://podscripts.co/podcasts/lex-fridman-podcast/212-joscha-bach-nature-of-reality-dreams-and-consciousness): the machine transcript reads 'love is, there is a curry of shared purpose', almost certainly a mis-transcription of 'the discovery of shared purpose'. - verbatim, [The Existential Risk of AGI: Reflections after the MLST Podcast with Connor Leahy (Joscha Bach, Substack)](https://joscha.substack.com/p/the-existential-risk-of-agi): Alignment happens either transactionally, coercively, or because we discover shared purposes above the individual self, thereby creating shared agency together. - verbatim, [The age of AGE: Artificial General Experience (Joscha Bach, Substack; text written for Rana Gujral's book 'The AI Instinct')](https://joscha.substack.com/p/the-age-of-age): We will have to build AI systems that can extend our wisdom, agents that find shared purposes with us. ### Consciousness comes before the self (2023) The usual picture is that brains first become complex and then consciousness appears at the top. Bach turns this around: consciousness is there at the start and is what trains the mind. Babies are conscious before they have a self, language or reasoning, and no human leaves the vegetative state without it. He reads the first chapter of the Bible as an old description of this boot-up: first a spirit over the substrate, then light and dark, world and ideas, and only at the end a person. From this CIMC derives its 'Human Consciousness Hypothesis', that consciousness is the simplest learning algorithm evolution could find for a self-organising brain. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=3124s) sources: - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): it’s not the result of our mental architecture and cognitive ability, but its prerequisite - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): Instead of being brought forth by the organized architecture of the mind, it creates it. We call this idea the Genesis Hypothesis - verbatim, [37C3 talk 'Synthetic Sentience' (machine subtitles)](https://c3subtitles.de/talk/2838): the earliest formulation of that theory is what I found in Genesis 1 - verbatim, [Lex Fridman Podcast #392: Joscha Bach, Life, Intelligence, Consciousness, AI & the Future of Humans (transcript)](https://lexfridman.com/joscha-bach-3-transcript/): I suspect that it’s there in plain sight in Genesis 1 - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): 'Our Human Consciousness Hypothesis states that consciousness is produced by the simplest learning algorithm that can be discovered by an evolutionary search to train a self-organizing biological substrate to become intelligent.' ### Levels of Lucidity (2023) A seven-step ladder of how a mind can come to understand its own workings. The steps are: reactive survival, personal self, social self, rational agency, self-authoring, enlightened mind and transcendent mind. Each step comes from seeing through a part of oneself that used to be taken for granted. It is based on Robert Kegan's five stages, which Bach says he has drifted away from. He says it is not a ranking of people's worth and that most people do not need to climb it. confidence: highly likely, [source](https://joscha.substack.com/p/levels-of-lucidity) sources: - verbatim, [Levels of Lucidity (Joscha Bach, Substack)](https://joscha.substack.com/p/levels-of-lucidity): Let us call the framework the “Levels of Lucidity”. - paraphrase, [Levels of Lucidity (Joscha Bach, Substack)](https://joscha.substack.com/p/levels-of-lucidity): the stages listed are 1 Reactive survival, 2 Personal self, 3 Social self, 4 Rational agency, 5 Self authoring, 6 Enlightened mind, 7 Transcendent mind. - verbatim, [Levels of Lucidity (Joscha Bach, Substack)](https://joscha.substack.com/p/levels-of-lucidity): I particularly like Robert Kegan’s five stage model of personality development - paraphrase, [Self Models of Loving Grace (38C3)](https://media.ccc.de/v/38c3-self-models-of-loving-grace), 2024-12-28: Paraphrase (captions: 'I call this the Devils of Lucidity'): 'I call this the levels of lucidity', where you reverse engineer yourself to figure out what you are. - verbatim, [Self Models of Loving Grace (38C3)](https://media.ccc.de/v/38c3-self-models-of-loving-grace), 2024-12-28: true post biological self will be something that is subset agnostic ### On large language models (2023) Bach does not say current language models are conscious, and does not rule it out either. His stated view in 2023 is that if such a model is conscious, it is not in the way humans are, because it imitates text written by conscious people. CIMC's public note says chatbots are 'really good actors' that can play a conscious character, and that producing text about an experience is not having it. The 2025 essay says LLM agents 'tend to dream', meaning they do not separate perception from imagination. In 2026 he calls today's systems 'artificial idiot savants'. He and Sorensen also argue that no behaviour test can decide the matter. confidence: highly likely, [source](https://cimcai.substack.com/p/did-your-chatbot-just-wake-up) sources: - verbatim, [Lex Fridman Podcast #392: Joscha Bach, Life, Intelligence, Consciousness, AI & the Future of Humans (transcript)](https://lexfridman.com/joscha-bach-3-transcript/): If it’s conscious, it’s probably not conscious in a way that is closed to the way in which human beings are conscious - verbatim, [Did Your Chatbot Just Wake Up? A friendly reality-check from CIMC (CIMC Substack)](https://cimcai.substack.com/p/did-your-chatbot-just-wake-up): Being able to produce text about an experience does not imply that the model is actually having the experience it describes. - verbatim, [Did Your Chatbot Just Wake Up? A friendly reality-check from CIMC (CIMC Substack)](https://cimcai.substack.com/p/did-your-chatbot-just-wake-up): LLMs are essentially really good actors - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): Current LLM based agents tend to dream and may need to learn the distinction between perception and imagination explicitly. - verbatim, [The age of AGE: Artificial General Experience (Joscha Bach, Substack; text written for Rana Gujral's book 'The AI Instinct')](https://joscha.substack.com/p/the-age-of-age): Our first generation AGIs are artificial idiot savants that know nearly everything that has ever been written, yet are still prone to making stunning logical mistakes. - verbatim, [Joscha Bach PUSHES Roman Yampolskiy: Only ASI Can Save Us! (Dr Brian Keating)](https://www.youtube.com/watch?v=ayW7SDyAjZo), 2026-10-09: There are in some sense deep fakes of human intelligence ### Cyber-animism: software as spirit (2024) Old cultures spoke of spirits that animate living things. Bach says this was a reasonable description of something real: self-organising software. A spirit is a pattern of control that runs on cells in the way a program runs on a computer, and it can be described without reference to the particular matter. Life and minds are such software agents competing for substrates. He presents this as a metaphysical stance and says it is not superstition; in the essay, selves that span many minds are called gods. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): the spirits of the animist worldview are not superstitions or mere analogies to computer software - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): 'We call the idea that natural spirits ... are best understood as software cyberanimism.' - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): selves capable of possessing multiple minds are called gods - paraphrase, [Machine Learning Street Talk interview with Joscha Bach at AGI-24 (YouTube transcript)](https://www.youtube.com/watch?v=3MkJEGE9GRY): the interviewer refers to 'your cyber animism talk', the keynote given the day before at the AGI-24 conference. - paraphrase, [AGI-25 talk 'The Operation of Consciousness' (YouTube transcript)](https://youtu.be/xDzAclUq3U4): 'last year I talked about the cyber animist hypothesis. It basically says that spirits can be understood as self-organizing software agents in nature'. - paraphrase, [Joscha Bach lecture on cyber animism (YouTube transcript, video id YZl4zom3q2g)](https://www.youtube.com/watch?v=YZl4zom3q2g): spirit is 'the word that our ancestors use to describe self organizing software agents'. - paraphrase, [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM), 2024: living things are colonised by self-organising software; evolution is a competition between software agents that only partly encode themselves in genomes. - paraphrase, [Cyber Animism (talk hosted by Michael Levin's center, Tufts)](https://www.youtube.com/watch?v=YZl4zom3q2g), 2024: software is spirit, a causal pattern without physical identity that nevertheless moves matter; living stuff is software colonising regions of the physical universe. - verbatim, [Conversation between Joscha Bach, Chris Fields, and Michael Levin (Michael Levin's Academic Content)](https://www.youtube.com/watch?v=knADtWMGxmw), 2024-06-05: the invariance that we're looking at and we look at living things is the software that runs on region of physics not the mechanism itself - verbatim, [A radical theory of consciousness | AI researcher Joscha Bach (Institute of Art and Ideas, public excerpt)](https://www.youtube.com/watch?v=pkhuDqK1_MU), 2024-10-18: it's basically animism ## Large language models 2017–2023, computing and AI. school of thought: neural networks and machine learning. id: `llm` ### LLMs and the transformer by Vaswani et al.; OpenAI Neural networks trained to predict the next token of text, built on the transformer architecture, whose 'attention' mechanism learns which earlier tokens matter for each prediction. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Large_language_model) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Large_language_model) start here, joscha bach on it: [Synthetic Sentience (37C3)](https://media.ccc.de/v/37c3-12167-synthetic_sentience#t=2145), from 35:47 (verbatim): are the simulated mental states of an llm more simulated than our mental simulated States connections: - related to ELIZA (`eliza`): people feel understood by a text machine. confidence: highly likely, [source](https://ojs.weizenbaum-institut.de/index.php/wjds/article/download/136/3_3_3/1438) - builds on Deep learning (`deeplearning`): the transformer. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Large_language_model) - related to Joscha Bach (`bach`): bach's reading of what language models are. confidence: highly likely, [source](https://lexfridman.com/joscha-bach-3-transcript) - leads to AI consciousness and welfare reports (`butlin`): checking AI systems against theories of consciousness. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) - leads to Murray Shanahan (`shanahan`): language models as role-players. confidence: highly likely, [source](https://arxiv.org/abs/2305.16367) - leads to CIMC: the Machine Consciousness Hypothesis (`cimc`): chatbots show the same behaviour can be produced another way. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ### Scaling hypothesis The bet that current neural network designs are already good enough, and that intelligence will keep increasing simply by adding data and computing power. confidence: highly likely, [source](https://gwern.net/scaling-hypothesis) ### Blake Lemoine and LaMDA by Blake Lemoine (Google) A Google engineer went public with his belief that the company's chatbot LaMDA was sentient and deserved rights. confidence: highly likely, [source](https://en.wikipedia.org/wiki/LaMDA) ### AI companions Chatbot products that users talk to as friends or partners. Replika is marketed as an AI companion; Claude is Anthropic's assistant. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Replika) ### Public AI discourse Emily Bender is a computational linguist known for calling language models 'stochastic parrots'. Sam Altman is the chief executive of OpenAI. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Emily_M._Bender) ## Chollet: On the Measure of Intelligence 2019, computing and AI. school of thought: neural networks and machine learning. id: `chollet` by François Chollet (On the Measure of Intelligence) Chollet defines intelligence as the efficiency of acquiring new skills, not the possession of skills, and argues that current models mostly interpolate between things they have seen. confidence: highly likely, [source](https://arxiv.org/abs/1911.01547) start here, read: [arxiv.org](https://arxiv.org/abs/1911.01547) start here, joscha bach on it: [Lex Fridman Podcast #392 – Joscha Bach: Life, Intelligence, Consciousness, AI & the Future of Humans](https://lexfridman.com/joscha-bach-3-transcript) (verbatim): When François Chollet says that these models are only interpolating between what they saw connections: - argues against Deep learning (`deeplearning`): skill is not intelligence. confidence: highly likely, [source](https://arxiv.org/html/1911.01547) ## Michael Levin: bioelectricity and collective intelligence 2019, physics and life. school of thought: evolution, biology and life. id: `levin` by Michael Levin Tufts biologist who studies how groups of cells use bioelectric signalling to decide what body to build. He treats cells and tissues as agents with goals, so that anatomy is the result of a collective intelligence solving problems. confidence: highly likely, [source](https://drmichaellevin.org/) start here, read: [drmichaellevin.org](https://drmichaellevin.org/) start here, joscha bach on it: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=2296s), from 38:18 (verbatim): the algorithm that makes an organism coherent with itself, is exactly the same one as the one that happens in the mind when it becomes coherent with itself connections: - related to Chris Fields (`fields`): frequent co-authors. confidence: highly likely, [source](https://drmichaellevin.org/publications/) - leads to Joscha Bach (`bach`): any cells can compute; morphogenesis and mind may share one algorithm. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) ## The Bitter Lesson 2019, computing and AI. school of thought: neural networks and machine learning. id: `bitter` by Richard Sutton Sutton's observation that, across 70 years of AI, general methods that scale with computation have always ended up beating systems built on human insight. It is bitter because researchers keep wanting their own knowledge of the domain to matter. confidence: highly likely, [source](http://www.incompleteideas.net/IncIdeas/BitterLesson.html) start here, read: [incompleteideas.net](http://www.incompleteideas.net/IncIdeas/BitterLesson.html) start here, mentioned by joscha bach in: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs) connections: - builds on Deep learning (`deeplearning`): scale beats hand-built knowledge. confidence: highly likely, [source](https://en.wikipedia.org/wiki/Bitter_lesson) ## Godfrey-Smith: Metazoa 2020, physics and life. school of thought: evolution, biology and life. id: `metazoa` by Peter Godfrey-Smith A philosopher-diver traces how experience may have arisen step by step in the evolution of animal bodies and nervous systems. confidence: highly likely, [source](https://symposiumbooks.com/book/9781250800268) start here, read: [symposiumbooks.com](https://symposiumbooks.com/book/9781250800268) ## Neural cellular automata 2020, computing and AI. school of thought: cybernetics, systems and self-organisation. id: `nca` by Alexander Mordvintsev et al. 'Growing Neural Cellular Automata' trains a single small neural rule, applied locally by every cell on a grid, so that the cells grow into and regenerate a target pattern, in the spirit of biological development. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) start here, read: [distill.pub](https://distill.pub/2020/growing-ca/) start here, joscha bach on it: [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U) (verbatim): one approach that I'm currently looking at is taking instead of neural networks neural cellular automata connections: - builds on Turing's morphogenesis (`morpho`): patterns that grow from local rules. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) - related to Conway's Game of Life (`conway`): the neural ca paper names life as the classic cellular automaton before its learned ones. confidence: highly likely, [source](https://distill.pub/2020/growing-ca/) - leads to CIMC: the Machine Consciousness Hypothesis (`cimc`): CIMC's self-organisation experiments. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ## The universality hypothesis 2020, computing and AI. school of thought: neural networks and machine learning. id: `olah` by Chris Olah The conjecture from interpretability research that different neural networks trained on similar data converge on similar internal features and circuits. Chris Olah and colleagues proposed it; if true, findings about one network would transfer to others. confidence: highly likely, [source](https://distill.pub/2020/circuits/zoom-in/) start here, read: [distill.pub](https://distill.pub/2020/circuits/zoom-in/) start here, joscha bach on it: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs) (verbatim): they all learn the same features, and that's regardless of the architecture connections: - builds on Deep learning (`deeplearning`): looking inside trained networks. confidence: highly likely, [source](https://distill.pub/2020/circuits/zoom-in/) ## Anil Seth: biological naturalism 2021, consciousness. school of thought: the hard problem and sceptics of machine minds. id: `seth` by Anil Seth Seth, author of 'Being You', holds that consciousness is tied to being a living body that must regulate itself to stay alive. On this view intelligence can be put in silicon but experience probably cannot. confidence: highly likely, [source](https://pubmed.ncbi.nlm.nih.gov/40257177/) start here, read: [pubmed.ncbi.nlm.nih.gov](https://pubmed.ncbi.nlm.nih.gov/40257177/) start here, joscha bach on it: [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs) (verbatim): He is a biological supernaturalist. connections: - argues against Functionalism and computationalism (`functionalism`): consciousness tied to living bodies. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - builds on Searle: the Chinese room (`searle`): biological naturalism. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - builds on Embodied and enactive cognition (`embodied`): mind needs a living body. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - builds on Antonio Damasio (`damasio`): feeling grounded in body regulation. confidence: highly likely, [source](https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(13)00211-8) - builds on Predictive processing and the free energy principle (`friston`): the 'beast machine': predicting one's own body. confidence: highly likely, [source](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/conscious-artificial-intelligence-and-biological-naturalism/C9912A5BE9D806012E3C8B3AF612E39A/share/f0f282851a27c9743ea840b0f4f563c8ba56852235f1886bf932b36c3e0f2520) - challenged by Joscha Bach (`bach`): function versus living substrate. confidence: highly likely, [source](https://www.youtube.com/watch?v=yW_ngvMFlYs) ## Gary Marcus 2022, computing and AI. school of thought: symbolic AI and cognitive architectures. id: `marcus` by Gary Marcus Cognitive scientist and a well-known critic of deep learning. He argues that neural networks cannot reason reliably without built-in symbolic structure and that scaling them up will not yield general intelligence. confidence: highly likely, [source](https://www.technologyreview.com/2020/03/27/950247/ai-debate-gary-marcus-danny-lange/) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Gary_Marcus) start here, joscha bach on it: [Machine Learning Street Talk #82 – Dr. Joscha Bach: Digital Physics, DL and Consciousness [UNPLUGGED]](https://www.youtube.com/watch?v=LgwjcqhkOA4&t=1524s), from 25:26 (verbatim): the solution to uh problems with deep learning so far has always been to use more deep learning not less connections: - argues against Deep learning (`deeplearning`): 'deep learning is hitting a wall'. confidence: highly likely, [source](https://nautil.us/deep-learning-is-hitting-a-wall-238440/) ## Thagard: energy and substrate independence 2022, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `thagard` by Paul Thagard Thagard argues that because real information processing costs energy, and different materials have very different energy budgets, a mind cannot simply be moved to any substrate; so functionalism is in trouble. The paper appeared in the journal Philosophy of Science. confidence: highly likely, [source](https://www.cambridge.org/core/journals/philosophy-of-science/article/energy-requirements-undermine-substrate-independence-and-mindbody-functionalism/2BB3C2353EFF80F9D5805CDCEA8C3C89) start here, read: [cambridge.org](https://www.cambridge.org/core/journals/philosophy-of-science/article/energy-requirements-undermine-substrate-independence-and-mindbody-functionalism/2BB3C2353EFF80F9D5805CDCEA8C3C89) start here, joscha bach on it: [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs&t=3862s), from 1:04:24 (verbatim): he is making the claim that sometimes the substrate has properties that influence its function connections: - argues against Functionalism and computationalism (`functionalism`): energy makes substrate matter. confidence: highly likely, [source](https://www.cambridge.org/core/journals/philosophy-of-science/article/energy-requirements-undermine-substrate-independence-and-mindbody-functionalism/2BB3C2353EFF80F9D5805CDCEA8C3C89) ## AI consciousness and welfare reports 2023–2024, consciousness. school of thought: scientific theories of consciousness. id: `butlin` ### Consciousness in Artificial Intelligence by Patrick Butlin, Robert Long, et al. A multi-author report that turns leading scientific theories of consciousness into a checklist of 'indicator properties' and checks current AI systems against it. It concluded that no current system was a strong candidate. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) start here, read: [arxiv.org](https://arxiv.org/abs/2308.08708) start here, mentioned by joscha bach in: [The California Institute for Machine Consciousness Research Program Whitepaper (Sorensen, Hildebrandt-Harangozó, de Kerhuelvez, Rutt, Bach)](https://cimc.ai/publications/research-program-whitepaper/) connections: - builds on Higher-order thought theory (`hot`): one of the checklist theories. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) - builds on Global Workspace Theory (`gwt`): one of the checklist theories. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) - builds on Large language models (`llm`): checking AI systems against theories of consciousness. confidence: highly likely, [source](https://arxiv.org/abs/2308.08708) - related to CIMC: the Machine Consciousness Hypothesis (`cimc`): cimc's whitepaper cites the report as ai welfare research. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) ### Taking AI Welfare Seriously by Robert Long, Jeff Sebo, et al. A report arguing that there is a realistic chance some AI systems will soon be conscious or agents with interests, so AI companies should start preparing welfare policies. confidence: highly likely, [source](https://arxiv.org/abs/2411.00986) ## Mazviita Chirimuuta 2024, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `chirimuuta` Edinburgh philosopher of neuroscience, author of 'The Brain Abstracted'. She argues that computational models of the brain are useful simplifications and should not be mistaken for what the brain literally is. confidence: highly likely, [source](https://braininspired.co/podcast/186/) start here, read: [braininspired.co](https://braininspired.co/podcast/186/) start here, mentioned by joscha bach in: [Can we crack the AI enigma | Joscha Bach, Mazviita Chirimuuta, Cory Doctorow (Institute of Art and Ideas, public excerpt)](https://www.youtube.com/watch?v=civc01IZCcY) connections: - argues against Functionalism and computationalism (`functionalism`): models are simplifications, not the thing. confidence: highly likely, [source](https://www.thebsps.org/reviewofbooks/burnston-on-chirimuuta/) ## Murray Shanahan 2024, computing and AI. school of thought: neural networks and machine learning. id: `shanahan` Imperial College and DeepMind researcher, author of 'Embodiment and the Inner Life', who describes LLMs as role-players simulating characters. He warns against describing such systems in human terms like 'knows' or 'believes' without care. confidence: highly likely, [source](https://arxiv.org/abs/2212.03551) start here, read: [Wikipedia](https://en.wikipedia.org/wiki/Murray_Shanahan) start here, mentioned by joscha bach in: [Sentience Beyond Biology — Debate w/Dmitry Volkov, Joscha Bach, Matthew MacDougall, Murray Shanahan (International Center for Consciousness Studies)](https://www.youtube.com/watch?v=GjK0Nn9gM0k) connections: - builds on Large language models (`llm`): language models as role-players. confidence: highly likely, [source](https://arxiv.org/abs/2305.16367) - related to Joscha Bach (`bach`): shared a panel on machine sentience. confidence: highly likely, [source](https://lifeboat.com/blog/2026/02/sentience-beyond-biology-debate-w-dmitry-volkov-joscha-bach-matthew-macdougall-murray-shanahan) ## CIMC: the Machine Consciousness Hypothesis 2025, consciousness. school of thought: functionalism and computational theories of mind. id: `cimc` ### The Machine Consciousness Hypothesis (2025) by Joscha Bach and Hikari Sorensen, California Institute for Machine Consciousness The claim is that an ordinary general-purpose computer with enough memory and speed has everything needed to be conscious, and that we can find out whether a system is conscious by analysing or testing it. Consciousness is treated as a pattern of organisation, so what the system is made of does not matter. CIMC adds a working definition: a system is conscious if it organises itself to perceive its own perceiving in a way that makes its overall model of the world more coherent. An 'extended' version says we can search for the right learning algorithm by recreating the conditions under which brains found it. The essay is co-written with Hikari Sorensen. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): possess the necessary and sufficient means to implement consciousness - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): 'We call this possibility the Machine Consciousness Hypothesis: that general computational machines with sufficient resources' can implement consciousness and that this 'can be established via analyzing or testing the system'. - verbatim, [The California Institute for Machine Consciousness Research Program Whitepaper (Sorensen, Hildebrandt-Harangozó, de Kerhuelvez, Rutt, Bach)](https://cimc.ai/publications/research-program-whitepaper/): A system is conscious if it implements self-organized second-order perception that increases global coherence. - paraphrase, [Joscha Bach, Curriculum Vitae, July 14, 2023](http://bach.ai/pub/CV%20Joscha%20Bach.pdf): CV lists 'The Cortical Conductor Theory: Towards Addressing Consciousness in AI Models' (IJCAI workshop 2017). - verbatim, [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs), 2026-11-02: consciousness might be a biological learning algorithm, a training algorithm for animal brains - verbatim, [Machine Learning Street Talk – Why Leibniz Couldn't Find the Mind Inside the Mill (Joscha Bach)](https://www.youtube.com/watch?v=yW_ngvMFlYs), 2026-11-02: Joscha and his co author Hikari Sorensen wrote their own Genesis. It's in his paper, page 20. - paraphrase, [The Operation of Consciousness (AGI-25, Reykjavik)](https://www.youtube.com/watch?v=oR-BQTSpL5U), 2025: consciousness is the prerequisite for human learning, not its result; the hypothesis is that the same organisation can be built in machines. - verbatim, [AGI-24 keynote on cyber animism (Seattle; Machine Learning Street Talk recording)](https://www.youtube.com/watch?v=34VOI_oo-qM), 2024: consciousness might actually be the simplest training algorithm for a self-organizing information processing system - paraphrase, [Why Consciousness Might Not Be Hard At All (Joscha Bach) (Machine Learning Street Talk)](https://www.youtube.com/watch?v=yW_ngvMFlYs), 2026-11-01: Host: 'Joscha and his co author Hikari Sorensen wrote their own Genesis. It's in his paper, page 20.' start here, read: [cimc.ai](https://cimc.ai/cimcHypothesis.pdf) connections: - argues against Alan Turing and computability (`turing`): behaviour cannot settle whether a system is conscious. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) - related to Stephen Wolfram (`wolfram`): wolfram advises CIMC. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - argues against Integrated Information Theory (`iit`): IIT says an ordinary computer can never be conscious. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - related to Predictive processing and the free energy principle (`friston`): friston advises CIMC. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - builds on Large language models (`llm`): chatbots show the same behaviour can be produced another way. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - builds on Joscha Bach (`bach`): the institute's hypothesis rests on his conductor model, his functionalism and the idea that consciousness comes first. confidence: highly likely, [source](https://cimc.ai/publications/the-machine-consciousness-hypothesis/) - builds on Neural cellular automata (`nca`): CIMC's self-organisation experiments. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - related to AI consciousness and welfare reports (`butlin`): cimc's whitepaper cites the report as ai welfare research. confidence: highly likely, [source](https://cimc.ai/publications/research-program-whitepaper/) - challenged by Lerchner: The Abstraction Fallacy (`lerchner`): 'the abstraction fallacy'; CIMC replied with 'against intrinsic properties'. confidence: highly likely, [source](https://cimc.ai/publications/against-intrinsic-properties/) ### No Turing test for consciousness (2025) by Joscha Bach and Hikari Sorensen, California Institute for Machine Consciousness Bach and Sorensen argue that you cannot tell whether a system is conscious by watching what it does. Consciousness is a particular way of producing behaviour, and the same behaviour could be produced another way, as chatbots show. So the test has to look at how the system is built and how it works inside. confidence: highly likely, [source](https://cimc.ai/cimcHypothesis.pdf) sources: - verbatim, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): Consciousness is by its nature not an externally visible performance, but a particular way to achieve a performance - paraphrase, [The Machine Consciousness Hypothesis (Joscha Bach & Hikari Sorensen, CIMC essay)](https://cimc.ai/cimcHypothesis.pdf): section heading 'Why there can be no “Turing Test for consciousness”'. ## Lerchner: The Abstraction Fallacy 2026, philosophy. school of thought: the hard problem and sceptics of machine minds. id: `lerchner` by Alexander Lerchner A paper arguing against computational functionalism about phenomenal consciousness. CIMC published a reply, 'Against intrinsic properties'. confidence: highly likely, [source](https://cimc.ai/publications/against-intrinsic-properties/) start here, read: [cimc.ai](https://cimc.ai/publications/against-intrinsic-properties/) start here, mentioned by joscha bach in: [Against intrinsic properties (Sorensen & Hildebrandt-Harangozó)](https://cimc.ai/publications/against-intrinsic-properties/) connections: - argues against Functionalism and computationalism (`functionalism`): 'the abstraction fallacy'. confidence: highly likely, [source](https://deepmind.google/research/publications/231971/) - argues against CIMC: the Machine Consciousness Hypothesis (`cimc`): 'the abstraction fallacy'; CIMC replied with 'against intrinsic properties'. confidence: highly likely, [source](https://cimc.ai/publications/against-intrinsic-properties/)