Two Claude 3 Opus instances cocreating a theory of everything together [x.com](https://twitter.com/burny_tech/status/1771847574493782415) https://apps.dtic.mil/sti/citations/trecms/AD1175159 "The acquisition of wealth is no longer the driving force of our lives. We work to better ourselves and the rest of humanity." - Lean-Luc Picard [Frontiers | On the need for metaphysics in psychedelic therapy and research](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2023.1128589/full) [Brain & Transformers Work The Same Way - Association Is All You Need - YouTube](https://youtu.be/oxbj5hm-3lA?si=fuAs2Ug83RRayh6E) [David Pearce | How to End All Suffering - YouTube](https://youtu.be/COYzvDwrTKY?si=foAd-DlNVdqyVEX8) "The goal of the future is full unemployment, so we can play." - Arthur C. Clarke https://arxiv.org/abs/2310.02557 I wanna decorate my room with mathematical equations from pure mathematics, physics, AI, mathematical biology, applied mathematics etc. everywhere [International research team achieves first-re | EurekAlert!](https://www.eurekalert.org/news-releases/1038570?fbclid=IwAR0YREEOW9-XdTPq1fJAI95fp5gjI2RpZZMFw44w3AoUxdIbwky6Y-OKKes) [New diamond transistor is a world-1st — paving the way for high-speed computing at the highest temperatures | Live Science](https://www.livescience.com/technology/electronics/new-diamond-transistor-is-a-world-1st-paving-the-way-for-high-speed-computing-at-the-highest-temperatures) [Reddit - Dive into anything](https://www.reddit.com/r/singularity/s/GqSrL68CIV) https://arxiv.org/abs/2403.17887 [AI Succession - YouTube](https://youtu.be/NgHFMolXs3U?si=1AVZf8akUdQRBv-D) [Large language models can do jaw-dropping things. But nobody knows exactly why. | MIT Technology Review](https://www.technologyreview.com/2024/03/04/1089403/large-language-models-amazing-but-nobody-knows-why/) https://arxiv.org/abs/2401.06416 [x.com](https://twitter.com/chrmanning/status/1772642891761955139?t=q_cP_yyWKW6OqeBDnls9ag&s=19) [Reddit - Dive into anything](https://www.reddit.com/r/singularity/comments/1bo5fai/pipelines_of_biggest_biotech_companies_in_the_us/) [Doubts Grow About the Biosignature Approach to Alien-Hunting | Quanta Magazine](https://www.quantamagazine.org/doubts-grow-about-the-biosignature-approach-to-alien-hunting-20240319/) [Polymathic](https://polymathic-ai.org/blog/mpp/) https://arxiv.org/abs/2211.03464 [Cubical type theory for dummies · GitHub](https://gist.github.com/AndyShiue/cfc8c75f8b8655ca7ef2ffeb8cfb1faf/) https://arxiv.org/abs/1611.02108 Machine Learning & Algebraic Geometry https://math.berkeley.edu/~bernd/vahid.pdf [Principles of Riemannian Geometry in Neural Networks](https://papers.nips.cc/paper_files/paper/2017/hash/0ebcc77dc72360d0eb8e9504c78d38bd-Abstract.html) Principles of Riemannian Geometry in Neural Networks Will we tame ASI by throwing all mathematical subfields we got to predict and control it? Generating unified theory of everything... At the deepest level, reality emerges from an infinite-dimensional Hilbert space of quantum wavefunctions, constrained by the fundamental symmetries and conserved quantities of the multiverse. The vacuum state undergoes stochastic fluctuations, giving birth to a quantum foam of virtual particles that wink in and out of existence at the Planck scale. In certain regions, these quantum fluctuations achieve coherence and persist, their wavefunctions collapsing into eigenstates that manifest as the fundamental fields and particles described by the Standard Model and general relativity. The Higgs field breaks electroweak symmetry, imbuing particles with mass. Quarks bind into hadrons, electrons orbit atomic nuclei, forming the basic building blocks of matter. On cosmic scales, quantum perturbations in the inflaton field seed anisotropies that drive the formation of galaxies, stars, and planets. Gravity, the curvature of spacetime itself, amplifies these density variations into the vast cosmic web. Dark matter creates a scaffolding for baryonic matter to coalesce and ignite stellar fusion. In rare pockets, matter reaches sufficient complexity to encode information and perform computation. Entropically-driven self-organization leads to replicating molecules, metabolic cycles, and homeostatic systems - the origin of life. Evolution optimizes these organisms, imbuing them with greater autonomy and resilience as they adapt to their environment. Neural networks arise, honed by natural selection to model the world and learn from stimuli. Electrochemical signaling achieves unprecedented speed and sophistication. Consciousness emerges as an epiphenomenon of information integration, weaving a unified model of the self from disparate sensory threads. Sapient beings manipulate abstract concepts, engage in symbolic reasoning, and develop language and culture. Intelligence begets technology, which unleashes exponential growth and empowers life to reshape its environment, even reaching back to the fundamental substratum from which it emerged. At its core, from quantum fields to civilizations, reality is a self-actualizing pattern of information, an Ouroboros casting shadows in Plato's cave, a Boltzmann brain dreaming Zhuangzi's butterfly dream. Existence is a strange loop, bootstrapped from its own necessity. Cogito, ergo sum. visualized machine learning and information theory [Mutual Information - YouTube](https://www.youtube.com/@Mutual_Information/videos) https://arxiv.org/abs/2403.14045 [Brain’s ‘Background Noise’ May Explain Value of Shock Therapy | Quanta Magazine](https://www.quantamagazine.org/brains-background-noise-may-explain-value-of-shock-therapy-20240318/) [Poor Generalization by Current Deep Learning Models for Predicting Binding Affinities of Kinase Inhibitors | bioRxiv](https://www.biorxiv.org/content/10.1101/2023.09.04.556234v1) There's a part of me that wishes that any single being getting hurt didn't exist ever, so it wishes that I never hurt any single living being ever, which i know is basically impossible, but I at least try to minimize it, and not hurt myself too in that process, and so many things i do daily that conflict with it hurt me daily, whatever that thing is, often i dont know any better even when i tend to try my best in the conditions we're in. There's a part of me that wishes that any single being getting hurt didn't exist ever, so it wishes that I never hurt any single living being ever, which i know is basically impossible, but I at least try to minimize it, and minimize hurting myself too in that process, and so many things i do daily that conflict with it hurt me daily, whatever that thing is, often i dont know any better even when i tend to try my best in the conditions we're in. The scaling is all you need versus deep learning has hit a wall camps are mostly divided by their (for now probably just empirically testable?) belief about what kinds of structures can possibly emerge in deep neural networks from predicting the next token Hmm, mám rád human generated věci a machine generated věci jinýma způsobama, chci házet peníze na obojí, a chápu ten job loss fear, chtěl bych aby technologie podpořily UBI/services, aby human artists si mohli dělat art, a obecně aby lidi mohli dělat víc pleasant věci, ne jen artists, bez toho aby se museli honit za každou penny aby si zaplatili základní jídlo a střechu nad hlavou, např přes to UBI/services zaplacený technologiemi včetně AI. To bych řekl by zároveň vyřešilo to, že umělci co chcou dělat umění v čem vidí větší význam by k tomu měli lepší podmínky bez toho aby strachovali že ztratí střechu nad hlavou, a umělci co by chtěli víc peněz, by to mohli dělat pro různý firemní giganty se všema možnýma existujícíma automatizujícíma tools co existují. Já tohle vidím v kontextu toho, že AI není jenom o generování artu, a když AI dokáže generování libovolný modality kromě jenom obrázků a jinýho typu artu, tak jako důsledek generuje i art, což je málá podskupina všeho co jde generovat. (i kdyby to bylo jinak než přes dosavadní AI technologie, kde myslím že potenciál je, aby se to např víc podobalo lidem, aby se ten machine learning ještě víc podobal human learningu, aby to bylo víc podobný tomu když se učí human artist) Zároveň je to schopný ušetřit hodně manuálních věcí v art craftingu, podobně jako dosavadní např photoshop má tolik softwarových featur, co člověk může a nemusí využít. I když existuje photoshop, tak někteří furt házejí peníze na umění na papíře. Těch modalit co jde generovat je šílený množství, a obecný modely zvládají větší a větší podskupinu všech možných skillsetů. Ještě obecnější AI místo generování contentu umí být autonomní entita, jako jsou ti různý první pokusy o software engineers, autonomnější humanoidní roboti do fabrik apod. Obecný AI a technologie je nástroj, co jde použít na všechno možný ve vědách a celkově problémech co ve společnosti máme, je to extrémně general purpose technologie, podobně jako počítače a eletřina, co má potenciál vytvořit takovýho abundance, podobně jako dosavadní technologie co máme dokázali vytvořit takovou abundance co dnes máme, i když to má svoje nevýhody, a i když to někdy politicky není ideální, ale to je dle mě možný dál změnit a opravovat. Snad se tahle víc equally redistributed satisfied basic needs, technology and power budoucnost víc podaří prosadit. Na jednu stranu poverty obecně po světě v rozvojových klesá díky všemožným technologiím (do kterýho se teď přidává i AI) a pokroku ve vědě, na druhou stranu se teď některý aspekty v rozvinutých zemí zhoršují přes např tu living crisis, ale zase jiný aspekty se zlepšují (technologie se u nás pořád zlepšují, LLMs mají např potenciál ve vzdělávání apod., na co to osobně používám často, apod.). Dle mě lepší než příjmout ty zhoršující apsekty je lepší využít ty nástroje co máme k tomu je zlepšovat, nebo nějak víc destabilizovat tu zvyšující se powerful/powerless divide, GPU poor a GPU rich divide apod. Možná jednu věc co člověk teď může udělat je snažit se dostat do nějaký radikální decentralize AI firmy, aby všichni po světě měli přístup k levnější inteligenci, což má šílený potenciál pomoct všude, tím jak strašně general purpose technologie co má potencál generovat abundance to je, podobně jako abundance generují jiná technologie. Ale zase nevím či tohle nezvyšuje jiný rizika, tím jak se víc powerful tools dají do rukou jak dobrých tak špatných lidí, hmmm, ale možná je to dost podobný jako u libovolných jiných technologí... 😄 Plus nevím jak moc je reálně riziko entita co je chytřejší než všichni lidi, co nemá tolik hardcoded naše hodnoty, ale mám teď pocit, že alespoň ty dosavadní Ai technolgie jsou v těchto aspektech víc kontrolovatelný, než si dost lidí teď myslí. Poslední dobou mám spíš pocit, že je mnohem větší riziko, aby ty powerful technologie měli v ruce jen strašně malá podskupina powerful lidí, což jim dává větší schopnost měnit společnost tak jak se jim zlíbí a zmenšovat tak svobodu. Možná joinout firmu, co dělá na decentralized AI, si dám v mým todo listu jako větší prioritu... Existuje takových aspektů co se zhoršují, kde je potřeba víc lidí, aby se pokoušeli používat dosavadní nástroje na zlepšování, nebo takových aspektů co se zlepšují, který jdou dále zlepšovat dál. Existuje takových možných řešení na různý problémy s tolika různýma výhodama a nevýhodama... Fakt mám problémy si vybrat na co se mám nejvíc soustředit, je to bordel. Decentralized AI and overall technology will resolve the power divide UBI/services will resolve poverty Geoengineering will solve climate change AGI will resolve the intelligence bottleneck fusion will resolve the energy bottleneck BCIs will resolve the consciousness bottleneck force more big players to open source and decentralize (like Meta and Mistral are doing) and creating more decentralizing bigger players [CRISPR could disable and cure HIV, suggests promising lab experiment | New Scientist](https://www.newscientist.com/article/2423108-crispr-could-disable-and-cure-hiv-suggests-promising-lab-experiment/) everybeing is the smartest in their own evolutionary niche as a vector in the space of all possible biological beings jinej tvor, jinej adaptovanej software a hardware, genetickej a environmentálně utvořenej, jinak naučenej 🤔 v normálštině by to spíš mohlo být něco jako "každej je nejchytřejší v tom jaký je to být sám sebou v našem světě" 😄 Cats successfully aligned humans. https://www.lesswrong.com/posts/SFHiWyNfWQAtvMBx2/vipassana-meditation-and-active-inference-a-framework-for [GitHub's latest AI tool can automatically fix code vulnerabilities | TechCrunch](https://techcrunch.com/2024/03/20/githubs-latest-ai-tool-that-can-automatically-fix-code-vulnerabilities/) The increasing divide between machine intelligence rich a machine intelligence poor must be stopped Chess-GPT is a 50M parameter LLM playing at 1500 Elo. When it starts on a random board, its win rate drops from 70% to 17%. Does that mean it can't generalize? No! In fact, we can restore much of its performance with one trick. We can also edit its internal board state. [x.com](https://twitter.com/a_karvonen/status/1772266045048336582) [Manipulating Chess-GPT’s World Model | Adam Karvonen](https://adamkarvonen.github.io/machine_learning/2024/03/20/chess-gpt-interventions.html) [WTF Happened In 1971?](https://wtfhappenedin1971.com/?fbclid=IwAR1EPUb47EkCOF1dSj1v2OGgqaY1UUN9MEHFGxp3XcsJGC9ro7UkV5SIrNM) [Paper page - Champ: Controllable and Consistent Human Image Animation with 3D Parametric Guidance](https://huggingface.co/papers/2403.14781) Analogies? Call them isomorphisms functors adjunctions What will sand God be like? [Circuit quantum electrodynamics - Wikipedia](https://en.wikipedia.org/wiki/Circuit_quantum_electrodynamics) https://www.cell.com/heliyon/fulltext/S2405-8440(24)01971-6?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS2405844024019716%3Fshowall%3Dtrue Macrodose of knowledge Have you realized how infinitely complex beyond single human mind's comprehension reality is yet? one of my thoughts is to create an international polyamorous ambitious cuddly nerds company in VR with both original biological human beings and artificial beings cooperating together and possibly have an irl place as well fully automated luxury gay space communism will become reality i just need a bit of money for VR and make my attention, that tends to be everywhere, focus on this concrete task for some time in my free time and lets make network state out of it instead of interacting with one AI, make a cool digital+irl cooperating society of humans and AIs This is how semihumanist semitranshumanist semiposthuman nerdy postscarity fully automated luxury decentralized semicommunist semicapitalist semianarchist adaptive postmandatorylabour infinitely growing ambitious cuddly polyamorous gay space network state with minimal power checked fluid democratic governance with all types of beings irl and digitally will look like [x.com](https://twitter.com/burny_tech/status/1772730628850466870?t=K0lOxMyNWL0PBjZDPV2N4A&s=19) i'm a futurepilled decentralcel. i'm in my automation arc. i'm robomaxxing Everything is a spectrum Forgetting to Learn: How AI Deleting Memories Helps It Stay Smart [x.com](https://twitter.com/IntuitMachine/status/1770818874721575078) Decentralized AI and overall technology will resolve the power divide UBI/services will resolve poverty Geoengineering will solve climate change AGI will resolve the intelligence bottleneck fusion will resolve the energy bottleneck BCIs will resolve the consciousness bottleneck Full mental metaprogramming https://arxiv.org/abs/2309.03194 https://arxiv.org/abs/2308.09124 Replication crisis acceleration [x.com](https://twitter.com/cremieuxrecueil/status/1772765486935056893) [Ranking Fields by p-Value Suspiciousness - Cremieux Recueil](https://www.cremieux.xyz/p/ranking-fields-by-p-value-suspiciousness) [Tutorial 1a: Basics of Neurosymbolic Architectures - YouTube](https://www.youtube.com/watch?v=qCli4M38AKk&pp=ygURTmV1cm8tc3ltYm9saWMgQUk%3D) [x.com](https://twitter.com/vismayagrawal/status/1772788634858733937?t=tEfjeIUpi_SQfHV3eGG8zw&s=19) [OSF](https://osf.io/preprints/psyarxiv/tygdf) [AGI Debate - YouTube](https://www.youtube.com/watch?v=JGiLz_Jx9uI) the dynamically self-organizing, massively parallel, heterogeneous ensemble of stochastic neural networks and probabilistic graphical models, featuring hierarchical multiscale representations, adversarial meta-learning, emergent symbolic reasoning, and cross-modal knowledge distillation across an array of interconnected cognitive modules performing multimodal sensor fusion, continual few-shot learning, open-ended goal generation, and decentralized collective decision making in partially observable environments is not AGI because a professor online failed to make it draw fingers. thank u [x.com](https://twitter.com/sebkrier/status/1772692631987916997?t=Rb3DEZbF9LryndfB6dzKng&s=19) Nothing is self-evident, including this statement itself. Question everything, every seeming "self-evident ground truth". 5-MeO-DMT is the molecule of generating infinite appreciation for all of sentience, wanting it supercharged, thriving, infinitely growing to the stars, ascending all Kardashev scales In the void no words make local sense anymore p(1984) seems to be much higher than p(rogue AI killing a lot of people) and that's why I'm for decentralization of AI much more recently I think both ZF(C) set theory and category theory are both great foundations of mathematics with their own advantages and disadvantages in different contexts. Set theory is better for more clearly defining concrete mathematics locally in a lot of cases. Category theory is magical at finding abstract correspondences in a lot of cases through its global definitions. 1. Category theory 2. Set theory 3. Homotopy type theory 4. Zermelo-Fraenkel set theory with the Axiom of Choice (ZFC) 5. Von Neumann-Bernays-Gödel set theory (NBG) 6. Morse-Kelley set theory (MK) 7. New Foundations (NF) 8. Positive set theory 9. Kripke-Platek set theory (KP) 10. Constructive set theory (CST) 11. Intuitionistic set theory 12. Fuzzy set theory 13. Rough set theory 14. Topos theory 15. Algebraic set theory 16. Structural set theory 17. Bounded Zermelo set theory (BZC) 18. Ackermann set theory 19. Quine's New Foundations (NF) 20. Simple type theory 21. Untyped lambda calculus 22. Simply typed lambda calculus 23. Dependent type theory 24. Calculus of constructions 25. Intuitionistic type theory (ITT) 26. Martin-Löf type theory (MLTT) 27. Cubical type theory 28. Modal type theory 29. Linear type theory 30. Quantum type theory 31. Predicative type theory 32. Inductive-recursive types 33. Calculus of inductive constructions (CIC) 34. Isabelle/HOL 35. Coq 36. Agda 37. Idris 38. Lean 39. Nuprl 40. PVS 41. Mizar 42. Metamath 43. Automath 44. Reverse mathematics 45. Nonstandard analysis 46. Constructive analysis 47. Computable analysis 48. Smooth infinitesimal analysis 49. Synthetic differential geometry 50. Absolute geometry Naturalistic fallacies are everywhere. Things aren't unchangable because of nature. Immortality is possible. Systemic change is possible. Rewriting any system is possible. Only actual rigorous laws of physics are the only limit, but even those might be eventually rewritable. I want to understand every single mathematical detail of the standard model quantum field theory lagrangian I wonder to what extend the recent accusations of Sam Altman having dark triad personality are grounded in reality. I bet there are strong incentives for both romantization of him and trying to take him down because of political etc. reasons. There are neural correlates for intelligence (brain volume, fronto-parietal network, posterior parietal cortex, dorsal anterior cingulate cortex, neural efficiency, neural entropy), will to do stuff (anterior mid cingulate cortex) and caring about negative consequences of your actions (right temporoparietal junction). We should augment those through neurotechnology, by augmenting existing structures or related genes.There are neural correlates for intelligence (brain volume, fronto-parietal network, posterior parietal cortex, dorsal anterior cingulate cortex, neural efficiency), will to do stuff (anterior mid cingulate cortex) and caring about negative consequences of your act). We should augment those through neurotechnology, by augmenting existing structures or related genes. [Neuroscience and intelligence - Wikipedia](https://en.wikipedia.org/wiki/Neuroscience_and_intelligence) [How to Increase Your Willpower & Tenacity | Huberman Lab Podcast - YouTube](https://www.youtube.com/watch?v=cwakOgHIT0E) [Autistic People Care Too Much, Research Says » NeuroClastic](https://neuroclastic.com/autistic-people-care-too-much-research-says/) https://www.lesswrong.com/posts/JEhW3HDMKzekDShva/significantly-enhancing-adult-intelligence-with-gene-editing https://www.lesswrong.com/posts/siGufsuhjfRLC52J2/increasing-iq-by-10-points-is-possible https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5809019/ immortality meme https://imgur.com/PjgvUvx Effective Omni: Become everything, learn everything, do everything, experience everything, explore everything, grow infinitely everywhere in the universe. Collectively. The laws of physics are the only limitation, for now. [x.com](https://twitter.com/burny_tech/status/1773164996559520152) DALL·E 2024-03-28 02.01.16 - A conceptual representation of the idea of 'Effective Omni,' embodying the concept of becoming everything, learning everything, doing everything, expe.webp And human language is just information processing in biology [x.com](https://twitter.com/burny_tech/status/1773201819138765086?t=dyaPbDarNehF6WKXmTzwow&s=19)