"A mathematician is a person who can find analogies between theorems; a better mathematician is one who can see analogies between proofs and the best mathematician can notice analogies between theories. One can imagine that the ultimate mathematician is one who can see analogies between analogies." – Stefan Banach A proof and a theorem are easy. A statement is a program P, (and a definition is a subroutine), proof is a witness that P always halts (cf. complexity class NP). Theory is more fuzzy. One could say it is a conglomerate of "related" statements, definition and proofs. [AI for Good - YouTube](https://www.youtube.com/@AIforGood) [The next wave of AI for Good - towards 2030 - YouTube](https://www.youtube.com/watch?v=D6SVoEhM1Ks) The next wave of AI for Good - towards 2030 Gary Marcus https://www.sciencedirect.com/science/article/pii/S0303264722002015 Quantum aspects of the brain-mind relationship: A hypothesis with supporting evidence put poverty under economy put political compass into social and economic parts put centralized decentralized into givernance first map out everything then summary then omni/acc call [AGI Revolution: How Businesses, Governments, and Individuals can Prepare - YouTube](https://www.youtube.com/watch?v=zwLsnsp7FF4) AGI Revolution: How Businesses, Governments, and Individuals can Prepare Joscha bach on political compass Auth Enforcing top down control in the social sphere Dictators elected by violence or economic power or manipulative promises of power or status games Decentralized autonomous swarm of multimodal GPTs Solving and scaling lowcost flexibile general autonomous embodied robotics connected to symbolic-nonsymbolic hybrid multimodal GPT reasoning engine will be the next giant intelligence breakthrough Bayesianism vs frequentism (meme) AI like electricity Resonance model of consciousness [Frontiers | The Easy Part of the Hard Problem: A Resonance Theory of Consciousness](https://www.frontiersin.org/articles/10.3389/fnhum.2019.00378/full) https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/ai-timelines [GitHub - mwcvitkovic/neuromodulation_modalities: Every way to change your mind](https://github.com/mwcvitkovic/neuromodulation_modalities) LLM OS https://twitter.com/karpathy/status/1723140519554105733?t=yaYBIg3BMdKkuzWapgvfYQ&s=19 We can't predict capabilities https://twitter.com/liron/status/1724116055503835585?t=8KhLglL2keW0dWR2tllvGQ&s=19 List of current wars Does consciousness equation fundamentally reside on substrate independent emergent layer like software or on fundamental ontological layer like fields of physics or both are different lenses on the same conscious system on different levels All memes Making LLMs more symbolic https://twitter.com/IntuitMachine/status/1724104506185580589?t=oIogsXtIe74eH_zPu6CcAg&s=19 Meditation trains attention for investigation of buddhist priors that sensations are impermanent, empty, cause suffering, causally interdependent, and don't have a separate self (including "yourself" qua phenomon). Do nothing lets go of attention and thinking machinery https://twitter.com/nosilverv/status/1724104952467841532?t=K9iYjHXFS__piDV10NABjA&s=19 [The Simplest Thing - YouTube](https://www.youtube.com/live/_HG02hDl5R0?si=w0lpvgx7SALgxEcn) "So you could construe rationalists as people who can 'think, but not do'. The generalization will be wrong in many cases but not be totally unfair regarding the mass. This makes me think that in some ways e/accs are people who can 'do, but not think'. Rationality's shadow" https://twitter.com/michaelcurzi/status/1723856896015118508?t=I2aoJxYsC2Gs52bowGpegg&s=19 https://twitter.com/michaelcurzi/status/1723876408030642652?t=9z-i6mVUS0fpehP1I1VhDw&s=19 Solving quantum gravity might enable us extreme advancements in technology as it might allow us manipulate reality on the smallest scales much more efficiently, potentially creating our own subatomic computers: hardware computation effeciency, compression, accuracy, realibility, transferability of information, better brain enhancements? better AIs? Better connection between brains and AI? [KARL FRISTON - INTELLIGENCE 3.0 - YouTube](https://youtu.be/V_VXOdf1NMw?si=8oqS9xeJKpxpbT42) Symmetries in LLM mental models https://twitter.com/IntuitMachine/status/1724390864678498584?t=LfcaBCljEse9s-9c65N0uA&s=19 Theres no such thing as philosophy free science https://twitter.com/SpeedWatkins/status/1724148236527542738?t=c-Y5gBTLUNzwBXk3YP6cSQ&s=19 Is killing the biggest extremists and then acting from kindness and peace to the rest to escape the life for life feedback loop the solution to wars? [Elon Musk: War, AI, Aliens, Politics, Physics, Video Games, and Humanity | Lex Fridman Podcast #400 - YouTube](https://youtu.be/JN3KPFbWCy8?si=rYrjG8oEalBX22zo) Imperialism I'm surprised how California, with its Sillicon Valley, the incubation of the most inluencal techcapital startups, the home of the biggest techcapital giants, has overall authoritarian left characteristics according to many The wars happening right now, or WW2, are pretty small in compasion to lots of other past wars, like internal chinese wars [Elon Musk: War, AI, Aliens, Politics, Physics, Video Games, and Humanity | Lex Fridman Podcast #400 - YouTube](https://youtu.be/JN3KPFbWCy8?si=rYrjG8oEalBX22zo) Different cosmological models Fermi paradox Great filter Grabby aliens https://www.sciencedirect.com/topics/neuroscience/memory-reconsolidation Screen of consciousness is not fundamental [The Screen of Consciousness is Not Fundamental - YouTube](https://youtu.be/-fTtcOZdsy4?si=LjEWfc6405Tae4oF) [David Deutsch - AI, America, Fun, & Bayes - YouTube](https://www.youtube.com/watch?v=EVwjofV5TgU) 37:00 Brain chip SotA https://twitter.com/neurosock/status/1724401933446983804?t=Odoji_LvYDlNKvoOJ2iQDw&s=19 Shlow brain hypothesis [How deep is the brain? The shallow brain hypothesis | Nature Reviews Neuroscience](https://www.nature.com/articles/s41583-023-00756-z) Towards Universal brain decoder https://twitter.com/mehdiazabou/status/1717271279575970256?t=pXWpSQntcyEYynrPC2rClQ&s=19 You are an agent. Your goal is building a python snake game application. You have an operating system Linux Ubuntu. You see it's state in the state.png image. Actions: "rightmouseclick *xpos*, *ypos*", "leftmouseclick *xpos*, *ypos*", "mousescrollup", "mousescrolldown", "press *key*", "wait *miliseconds*", "finish" *xpos* is a number in the interval 0-1920 *ypos* is a number in the interval 0-1080 *key* is any of these keyboard keys: a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, q, r, s, t, u, v, w, x, y, z, A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, X, Y, Z, 1, 2, 3, 4, 5, 6, 7, 8, 9, {, }, [ ,] , (, ), /, ", ', ., -, +, =, :, ;, space, esc, shift, ctrl, alt, capslock, enter, backspace, delete, printscreen, arrowright, arrowleft, arrowup, arrowdown *miliseconds* is number in the interval 0-10000000 You can use visual recognition to infer what actions to initialize. You have a memory: Known installed programs: terminal, python, google chrome with internet What will be your next step? You can only use one action listed under "Actions:" Respond only with a an action. Don't write the quotation marks. Repond only with that, don't add any other text. Example: leftmouseclick 700, 200 What will be your next steps? You can only use actions listed under "Actions:" Respond only with a sequence of action separated by newlines. Don't write the quotation marks. Repond only with that, don't add any other text. Example: leftmouseclick 700, 200 press n press o Additional memory: Past states: paststates.zip Current executing action: empty Past actions: empty Current plan: nothing Past plans: empty Working memory: empty Actions: edit (will edit any above character list variable in memory) Particles are properties of the field Most people dont have big enough dreams. System should teach people to have big dreams more. Should consciousness be frame invariant I want aliens that give us solution to quantum gravity Or an explanation for how large language models work internally Or how consciousness is implemented physically Or how to maximize the total amount of matter spaciotemporally optimized for wellbeing Interpretations of qm Wiki table https://twitter.com/burny_tech/status/1724789509672563176?t=Kv_CK_Ka_D-W4pL1MNH1cw&s=19 Do waves bounce around and interact nonlinearly in EM field's brain's topological pockets giving raise to holistic field information processing [Electromagnetic Field Topology as a Solution to the Boundary Problem of Consciousness - YouTube](https://youtu.be/tX8b3ng37Nw?si=B8EubgpfZWHN-coy) Is peaks of highly interconnected clustered integration of information basis of binding in consciousness Is degree of modularity in the brain basis for determining binding of conscious experience Is lowest shared resonance solution to binding problem Is binding/unity of consciousness an illusion like computer screens Solution to neuropsychological problem of neurophenomenology. Why do some people with ontologically seeing consciousness as information processing software, field with wavecomputing, "standard" self roles, nonphysical substrate, or ieffable transcendental unknown, think consciousness or qualia is real or doesnt exist? Psychedelics and meditation increase brain entropy, minimize hiearchical structure Software is higher level of abstraction Wave computing is Closer to hardware (neural Field theory vs electromagnetism) Nonphysical soul can ve structure next to classifiers undestanding things Aš embodied physics Raw qualia can ve seen as bittom of predictive hiearchy that some people dont have access to Different models can be seen as different through hebbian learning installed predictive priors on the software level, or as different selforganized harmonic modes on the hardware level, Ieffability is chaos without linguistic classifiers attached to it Combining QM and GR https://twitter.com/getjonwithit/status/1724828998184718426?t=sPmc_z-gbm-LxqjPNJpb3A&s=19 https://twitter.com/BrianRoemmele/status/1724480388985508272 AI music Machine learning is comfy field. No hardcore math like in advanced physics. Just statistics, probability, linear algebra, calculus, discrete mathematics, algorithms and data structures. ChatGPT list symbolic AI approaches, And hybrid Potential paper steering embodied robot with monosrmanticity or topdown engineering [5-HT2AR and NMDAR psychedelics induce similar hyper-synchronous states in the rat cognitive-limbic cortex-basal ganglia system | Communications Biology](https://www.nature.com/articles/s42003-023-05093-6) 5-HT2AR and NMDAR psychedelics induce similar hyper-synchronous states in the rat cognitive-limbic cortex-basal ganglia system [Huberman Protocols](https://www.hubermanprotocols.org/) list of wellbeing, health, agency, productivity, cognition etc. sciencebased tools [Suffering-Focused Ethics (SFE) FAQ — EA Forum](https://forum.effectivealtruism.org/posts/uEKJhuQSXmTNDBq5o/suffering-focused-ethics-sfe-faq) I think consciousness ontologies are guessing that cannot be proved empirically literally ever. But some ontologies are more pragmatic than others. Hardcore disconnected dualism in practice seems to really go against the fact that we have some neural correlates to target in people for example with depression. My assumption is so radically nondualist to advance neural correlate research as far as it will be possible for as much predicting and manipulation of brains and other conscious systems as possible. [Causal analysis - Wikipedia](https://en.m.wikipedia.org/wiki/Causal_analysis) Preparing for singularity [AGI Revolution: How Businesses, Governments, and Individuals can Prepare - YouTube](https://youtu.be/zwLsnsp7FF4?si=DTKmo0uYWvHf2PEw) AI: Paperwork, email sending, planning Universal Basic services [The Danger Of AI Is Real In 2024: Economic Crisis, War, Elon Musk & The Singularity | Raoul Pal - YouTube](https://www.youtube.com/watch?v=S0DHI0DGOIw) AI REVOLUTION: The #1 MEGATHREAT to Our Economy & How To PREPARE NOW | Raoul Pal adapt or die are we at the cambiran explosion of intelligence Wellbeing, agency, cognition, intelligence, productivity, enlightenment and lots of other functions, baseline or temporary mental states and qualia, is controlled and affected by conscious software (psychotheraphy, motivating thoughts, selhelp frameworks building life, meditation), subconscious software (intuition, deep priors), and hardware (genes, molecules, structure, strength and activity of various brain circuitry), that can by directly (consciously manipulating thoughts, sending message to subconscious...) or indirectly (...which affects hardware as well (psychotheraphy), or influence propagating from hardware to subconscious to conscious (nutrition, excercise, cold showers, being outside, substances, electrical stimulation of brain)) affect eachother in nontrivial ways, by internal (psychotheraphy, meditation) or external means (people and overall environment, substances, neurotechnology (electrical stimulation of hardware, editing mental representations and logic in software, adding computational power and memory directly (brain implants, connecting brains to machines) or indirectly (external memory (notes), computers, phone, google (tons of current knowledge), ChatGPT, other (AI) technology)), under constrains of hardwired and adapted biological architecture (that might be eventually editable by technology too) and resources evolving under laws of physics USA is in economic recession. Will inflation come off very fast because of political involvment? [The Danger Of AI Is Real In 2024: Economic Crisis, War, Elon Musk & The Singularity | Raoul Pal - YouTube](https://youtu.be/S0DHI0DGOIw?si=DdMpcKXnNecFc8UD&t=3889) [Independent, economic & financial data in real time on-chain | Truflation](https://truflation.com/) unblackboxing RRNs https://twitter.com/jakub_smekal/status/1725260735058698725 Spontaneous symmetry breaking in diffusion models [Spontaneous symmetry breaking in generative diffusion models - YouTube](https://youtu.be/IG0NNE9bpWU?si=ASHCcUhplZqZPsOi) Transformer interpretability https://www.lesswrong.com/posts/hnzHrdqn3nrjveayv/how-to-transformer-mechanistic-interpretability-in-50-lines https://www.lesswrong.com/posts/nuJFTS5iiJKT5G5yh/polysemantic-attention-head-in-a-4-layer-transformer Singularity timeline [SINGULARITY AI: Ray Kurzweil Reveals Future Tech Timeline To 2100 - YouTube](https://youtu.be/wYVktxbf6t0?si=s7MSFC0QTZEyBP8N) Historie AI (inspirace neurovedou, neuro a AI se navzájem ovlivňuji, symbolic nonsymbolic, Lin regrese, NN, DNN, MLP, CNN, ngrams, Word vectors, RNN, memory RNN) SotA (alphago, language Transformers, ChatGPT, multimodal Transformers, image generators, vision, robotics, Boston dynamics) Živá ukázka videa Co je ve výzkumu a v plenkách co možná bude brzo (Yann lecun, roboti) Levels of AGI Rychlost vývoje, boj tech gigantů (Google, OpenAI, Antropic, Musk) Like scissors, internet, electricity, nuclear weapon, dual use, omni use Benefity a risky, sociální technologický ekonomický kulturní vládnoucí implikace, od nejvíc AI specific po obecný (věda, technologie pro nás nebo proti nám nebo zbraně, klíma pokazit zlepšit, equality/oprese pokazit zlepšit, loneliness/mental health pokazit zlepšit, biorisk bioimmortality, stabilita, health, security, komunikace, vzdělání, inteligence lidí dependence undevepment vs internetlike enhancing, jobs, stabilita pokazit zlepšit, governance pokazit zlepšit) (biorisk, deepfakes, polarizace, sociál support, mentoring, copilot, aligning us, drones, stabilita systému, skynet) (věda, healthcare, climate change, jobloss (new jobs), UBI, privacy, copyright, regulatory capture, inequality vs equality on all levels) (oscilovat mezi výhodám a nevýhodáma/risky change it each slide) Řešení (víc alignment) Open source, Regulovat nebo nechat? (stop, adaptace, accelerate) Možný budoucnosti (extinction, dystopie -> normal adaptace, platou, nonshattering progress-> transhumanist utopie x nature utopie x merging with singularity universe God x všechno zaraz) AI Autotranalate záznam do angličtiny How to steer artificial intelligence technological revolution correctly on all levels, technology, science (AI, physics, neurotech), culture, economy, governance,... Singularity reddit [Statement on AI Risk | CAIS](https://www.safe.ai/statement-on-ai-risk) [resource recommendations - What is a complete book for introductory quantum field theory? - Physics Stack Exchange](https://physics.stackexchange.com/questions/8441/what-is-a-complete-book-for-introductory-quantum-field-theory) Grokking, memorization vs generalization, reverseengineered https://twitter.com/VikrantVarma_/status/1699823229307699305?t=eB-Jqplyy3UPzyAJuZRO5A&s=19 Illya vs Sam on are Transformers enough https://twitter.com/umm_rit_/status/1725520795236278604?t=yyhfysZSlU4nTsSyyDTO7A&s=19 Artificial general superintelligence that is sustainable and loves all of sentience AGi meme https://media.discordapp.net/attachments/991777568007131187/1174929606940315678/20231117_053024.jpg?ex=65696177&is=6556ec77&hm=67b426ffe4973bcf67c4e7f31d11d43f7427e25ffe85bf364240a1a45f657366& https://www.reddit.com/r/singularity/comments/17xfl30/sam_altman_on_ai_in_2024_the_model_capability/ Theory will get you only so far [Conquer Aging Or Die Trying! - YouTube](https://www.youtube.com/channel/UCT1UMLpZ_CrQ_8I431K0b-g)