The fact that organisms are made of cells that somehow hold together that are made of proteins and other complex biomolecules that are made of amino acids and other biological molecules that somehow hold together that are made of chemical elements that somehow hold together that are made of fundamental particles like quarks and four fundamental forces that somehow all work together is mindblowing to me daily We're witnessing the democratization of knowledge beyond all past humanity's imaginations LeetCode but it's for advanced mathematics I'm interested in AI research and engineering, both fundamental and not, applications in industry, altruistic applications, applications in science and mathematics, applications in other technology, reverse engineering, politics, geopolitics Learning is unlocking nodes in a nested skill tree of crystalized intelligence Become omnitribal Are corporations immortal by now or will they crumble down just like other empires did in the past? I'm aware of the limitations of LLMs but some people are in such a denial and cope so hard and use negative confirmation bias at all costs so hard it's mind boggling lol Eat a lot of quality training data and scaling laws will apply to you Priority voting is cool I want to know how everything works Our organisms build a generative model engine of the world approximating classical physics with 3D space with 3D objects and linear time with higher order more abstract hetearchies of beliefs and actions and plans in probabilistic bayesian hypermetagraphs reducible to interacting basic needs homeostats, all through evolutionary selforganization, bayesian processing, nonlinear wave comouting, using the physics of mostly electromagnetism creating biochemical gradients with emergent computational patterns. The world model engine with it's beliefs is usually synced with our shared patterns of physical reality, but imagination, dreams, meditation, or various substances can decouple this synchronization, and you can construct arbitrary shapes with arbitrary level of agency in arbitrary dimensions in arbitrary edited more nonlinear time flow with edited refresh rate. To reduce prediction error between your world model and what comes in sensory data when syncing with the environment, you can either edit your own multimodal pattern recognition model, or edit the outside world to edit the outside pattern generators. Reverse engineering and building intelligence is the ultimate nerd snipe I love being hyperconnector I refuse the institutions to suck out life force out of me ADHD = Ungovernable hyperactive parallel attention neurodivergence https://x.com/arithmoquine/status/1828251430819356760?t=mzvQNoemef2YGhATJy_kJQ&s=19 https://www.theinformation.com/articles/openai-races-to-launch-strawberry-reasoning-ai-to-boost-chatbot-business?rc=sy0ihq https://www.theinformation.com/articles/openai-shows-strawberry-ai-to-the-feds-and-uses-it-to-develop-orion https://twitter.com/AIExplainedYT/status/1828426164110496110?t=Fpl7HsNew6je1F4uCm9P_A&s=19 Postelection it is. This will probably in big part determine if the AI bubble will inflate even further or crash. Depends if it will be a good agentic model with long term coherence or not, if it will hallucinate less, possibly as a composite system, if it will generalize better. Will this cycle repeat? Is OpenAI still in lead, or has OpenAI lost? https://x.com/burny_tech/status/1828440246259753037?t=KIPWIB9-GQKvIyM8suMdBg&s=19 Mainly curiosity, truth seeking, desire to create cool and helpful technology, maps, art and other things, care, self-expression, desire to worldbuild,... seems to fuel me. My biggest hobby right now is to learn learn learn all of STEM from infinite curiosity! I love building a lego castle of collected interconnected knowledge and patterns in a nested skill tree metahypergraph <3 Realtime VR games generated by AI and controlled by our thoughts in 5 years [GameNGen](https://gamengen.github.io/) https://x.com/burny_tech/status/1828924286976815198 Science is the process of data compression over time Giving yourself to some concrete meaning generator hijacks your ability to metaprogram your reward function, but often makes you much more motivational to do concrete stuff Do you defragment your mental representations often? Learn all of science to become one with the universe My favorite math is math predicting physical reality or math for engineering and analyzing cool systems or math swimming in euphoric abstractions that unify everything I think the most likely future scenario of future power metastable equilibrium is tons of extremely powerful superorganisms (corps and govs) growing and getting more and more powerful overtime thanks to the technology, and chaotically competing, and normal people having less and less power to influence/control/steer/regulate them overtime, them being similar to misaligned superintelligences. Instead of full top down control or complete anarchy. p(government/corporate cyberpunk dystopia) slightly higher than everything else. On second place are p(star trek utopia), p(coliving of all forms), p(rogue ASI), p(cyberpunk anarchy). Omnis. Omnis means everything. All possible species. All possible substrates. All possible hybrids. All possible physical forms. All possible nonphysical forms. All possible minds. With all possible mental states. In all possible multiverses with all possible laws of physics. In all possible mathematical structures. In all possible thoughts. I want to study the fundamental structure of reality through physics and other sciences, crack the nature of intelligence, understand all of mathematics, understand biology to eradicate diseases and aging and help in medicine, build machines and programs that help all beings, advance neurotechnology for the benefit of all sentient beings, advance AI for the benefit of all sentient beings, understand what is metaphysically true about consciousness, explore our planet and the whole space, reduce injustice, as effectively as possible give all sentient beings what they lack, create the most relaxing environment possible, open a sanctuary for struggling sentient beings, spread love as much as possible, prevent catastrophes and dystopias from happening, advance and grow the civilization, help all sentient beings experience anything they ever wanted to,... I refuse the environment trying to beat this out of me Hmm, stát by mohl současný ekonomický incentives trochu editovat tím že by mohl finančně incentivizovat ty AI companies co pushují podobný usecases automatizující co lidi nechtějí dělat, nebo taxovat ty ostatní a redistribuovat těm co jsou automated away Alespoň v Evropě, v Americe se to dělá hůř, kde všechny ty gigafirmy si v podstatě z velký části kupují policies v politice... Inside me are two wolves on a spectrum: trying to edit the existing system with the possible methods, or completely destroy it and rebuild it from scratch Humanoid doing dishes Not sure if humanoid is the perfect form for this, but it's great anyway! https://x.com/BerntBornich/status/1829575359526023575 the scaling laws of autism https://scitechdaily.com/autism-severity-linked-to-brain-overgrowth-research-shows/ " Hiearchie hodnot, názory se v čase tvoří zjednodušeně z: - předpovídající síla (to umí nejlíp vědecká metoda) - redukce nejasnosti (to umí nejlíp náboženství) - konzistence (to taky umí nejlíp vědecká metoda) - komplexita (čím komplikovanější logika světa, tím hůř ho člověk dokáže zpracovat, což může být neefektivní) - a v základu užitečnost, utilita (jak moc daný názor uspokojuje tvé potřeby co si uvědomuješ) (což často odvozuje ty body nahoře) to celý se odvozuje od základních lidských potřeb, které jsou u všech lidí podobné, akorát ty jednotlivé potřeby chtějí být uspokojovány jinak silně a jsou u každýho jinak propojený a v jiných kategoriích (hiearchy of needs) (intenzita může být v predispozici geneticky nebo ovlivněná prostředím, prožitky, apod.), ale způsoby, jak tyto potřeby uspokojovat/udržovat co nejblíž k ideálním hodnotám se liší člověk od člověka, protože jsou naučené skrz společnost, skrz pozorování věci co kolem nás vidíme, skrz názorové skupiny (což můžou být různé skupiny jak kamarádů, tak komunit, tak politických stran)) To co je pro tebe nesmysl protože je to nekonzistentní z tvýho pohledu, což ti způsobuje pocit nestability, je u antivaxera pravda, protože to pro něj naopak způsobuje pocit konzistence plus svobody když tohle celý vyjádříš v nějakým obřím vyhodnocovacím algoritmu (který bude skoro nemožný napsat přesně, ale hypoteticky), tak zabíjení plodů má pro ně mnohem větší negativní cost, než zabíjení afgánců (protože to pro ně může vycházet z fundamentálních křesťanských hodnot (afgánci jsou pro ně třeba nelidi např protože nesdílí víru, tudíž je to známka čistého nebezpečí, a plod narozený ve společnosti, kde žijí oni je pro ně "svatost", protože se to např děje na stejným území, takže to furt vnímáš jako svou skupinu, a nestabilitu (vraždy) v ní nechceš, apod. způsoby jak to dát do slov, viz co psal Stracpe), které jim dávají v podstatě totální základ struktury a intenzity algoritmů ovlivňující jejich reakce na tyto stimuly, a ty jsou nejintenzivnější, protože v tom celý život žijí, a jejich nervová soustava je na tyto vzory naučená, je v tom nejefektivnější) žádní lidi nemyslí v absolutní logice na high level rozhodovací úrovni, v abstraktní matice se ani nedhodneme na tom co je nejlepší struktura logiky, protože i matikou zformalitovaná logika samotná je subjektivní kontrukt, lidi se řídí naučenými algoritmy, které skoro vždycky nejdou vyjádřit slovy, tím jak komplexní jsou, vždy to bude nějaká aproximace, kontradikce jsou kolikrát subjektivní záležitost, něco je kontradikce pro někoho ale pro jinýho ne, to záleží na naučené mysli skrz život, která se pořád učí a nikdy nedojde k absolutní konzistentnosti, pokud by platilo, že jediné kritérium co maximalizujeme, je konzistence, což neplatí, jinak bychom se pravděpodobně zbláznili protože na to mysl prostě není stavěná nebo nebyli vůbec kompatibilní se společností " Wake up Process infinitely vaster amounts of information beyond all my ancestors' whole life time's imagination Go to bed Wake up Process infinitely vaster amounts of information beyond all my ancestors' whole life time's imagination Go to bed Wake up Which mathematical type of a space do your mental representations exist in? Studying mathematics, physics, machine learning, other applied math, etc., to be able to classify all patterns in experience and reality using these maths, and engineer them Are you a frequentist or bayesian? What are your prior distributions and biases? The manic tendency to grok every single mathematical pattern governing this simulation I want to understand all the scales of reality! I love extremely concrete detailed mathematical definitions and descriptions of physical systems to their every single tiny itsy bitsy detail, concretely mapping out every single detail in engineering diagrams of machines beyond all our ancestors' comprehension with control theory, cybernetics, geometry, classical mechanics and so on or completely deconstructing the Lagrangian of the Standard model encoding all the fundamental particles and forces creating our universe from first principles (which breaks down at Planck scales and we need to solve quantum gravity but we good)! I love bigger approximations and coarse graining and exploring how big populations of particles work statistically or what classes of systems with certain properties share in their dynamics and structure and so on, and categorizing different abstractions together. I love to see what magic can be done with differential equations, statistical mechanics, network theory, and so on! I love flying in absolute abstractions beyond just isomorphisms between sets, I love understanding the nature of infinity, I love swimming in functors in infinity categories in metamathematics of infinity category theory, defining everything via Kan extensions, comparing different mathematical universes using Topos theory! I love deconstructing all axioms used in ZFC set theory, category theory, type theory and so on, I love to question everything with philosophy of mathematics, or philosophy in general! Why is math so damn unreasonably effective at predicting and engineering physical patterns? Is mathematics the programming language that this simulation was written in? Is there some superlanguage beyond just math that we need to grok to fully reverse engineer the source code of reality? I love everything! I want to know everything! I want to understand all the patterns! Feed me all the patterns! [Imgur: The magic of the Internet](https://imgur.com/jSedgt7) 4 AM ultra truth seeking hypercurious superexcited oversharing mad science etc. mood Artists fell in love with their loss function Instead of AI completely replacing human artists and them needing to look for new ways to pay their rent, I wish that the result of systemic incentives would instead be artists making art and writing by themselves, and in collaboration with different types of AIs when they want, while not having to worry about their rent, and while other AIs fully do what humans don't enjoy doing, such as dishes and laundry. And not just art, but apply this to all tasks humans enjoy and don't enjoy doing, like programming and math in future for example. And people still play chess, even if we have superhuman AIs on that, or do various tasks that have been mostly automated away, like knitting, farming from scratch etc.. I think these incentives can be designed with UBI, taxing or financial bonus from governments or other entities, changing cultural norms etc. What if there is no theory of quantum gravity and we need complete paradigm shift out of our current maps? As you progress the complexification of the mind, the things that used to form your identity become objects to be manipulated. The subject becomes the object for the next subject ad infinitum. z tohoto strawberry responcu (pravděpodobně problém tokenizeru) vznikl meme tak velkej v celý AI komunitě že o po tom OpenAI pojmenovala svůj reasoning projekt https://www.theinformation.com/articles/openai-shows-strawberry-ai-to-the-feds-and-uses-it-to-develop-orion [https://www.youtube.com/watch?v=My4Fj3fxct4](https://www.youtube.com/watch?v=My4Fj3fxct4) velká část AI safety bills co se v Americe teď passujou jsou z velký části silně ovlivněný a lobbied gigacorpama a tak přežvýkaný aby oni z toho pravděpodobně těžili nejvíc (AB 3211, SB 1047) [Reddit - The heart of the internet](https://www.reddit.com/r/singularity/comments/1e0ts13/the_perils_of_californias_sb_1047_a_threat_to/) [Reddit - The heart of the internet](https://www.reddit.com/r/ArtificialInteligence/comments/1f5eqfb/california_bill_set_to_ban_civitai_huggingface/) já chci regulace co gigacorporace absolutně nezneužijou a nevymilkujou pro ještě větší koncetraci moci díky jejich nekonečným pěnězům když vidím jak je systém v realitě tak rotten, a spousta věcí co mají nejdřív dobrej intent jsou potom dosavadníma incentvama v dosavadním systému totálně warped, že věci ještě naopak zhoršujou pro většinu lidí, tak mám větší a větší chuť utíkat do fantasy světů s ideálním systémem do toho ještě [The AI arms race between the US and China is heating up](https://finance.yahoo.com/news/the-ai-arms-race-between-the-us-and-china-is-heating-up-160000539.html?guccounter=1) https://twitter.com/tsarnick/status/1826829893159813540 https://twitter.com/tsarnick/status/1814149823765086384 vzhledem k tomu že teď jak demokrati tak republikání jedou solidní we must beat china in AI narrativ, tak pochybuju že to zakážou 😄 OpenAI se stali jedním s Americkou vládou https://twitter.com/sama/status/1829205847731515676 a v tomhle narrativu chceš zakázat co nejvíc open source, aby to nebenefitovalo Číně, a abys jako vláda (a gigakorpy) koncetroval co nejvíc moci pro sebe, aby to byl v podstatě vládní secret defence Mannhatan project (pro USA aka pro elity) připadá mi že ve srovnání s touto dynamikou jsou jiný risky celkem minimální (kvalitnější deepfakes atd.), pokud teda nenajdou způsob jak dát těm AI systémům hodně autonomie a nevymknou se kontrole až moc, místo dosavadních tiny fuckupů sem tam Vypadá to, že Čínská vláda se bude ještě víc snažit o zabití AI v rukou všech, aby měla jen to jedno a správný Čísnký myšlení pod dohledem vlády (viz část o harmful datech, prevence "Violations of core socialist values" pod dohledem vlády [China Releases New Draft Regulations for Generative AI](https://www.china-briefing.com/news/china-releases-new-draft-regulations-on-generative-ai/) ), takže je pro ně taky idelání, aby AI bylo hlavně v rukou vlády a Čínských korporací a firem co jsou hodně napojení na vládu nebo v podstatě vláda 3Blue1Brown has the best animated math videos on neural networks, linear algebra, differential equations, calculus and others! I recommend going through all of them! Do you think that Dyson spheres are actually possible to build? AI gives unconstrained creativity Comparing christianity and singularitianism: Christianity | Transhumanism ---|--- Prophets | Kurzweil, Drexler, De Grey Providence | Inevitable Techno-Progress The Messiah | The Strong A.I. The Rapture | Brain Uploading Heaven | Cosmic Computer From a cognitive science perspective, both memeplexes give you sense of belonging in something higher that will become even grander and that feels great? How does the equation predicting all of our current measured and future empirical data looks like? I'm dissapointed, googling "math pr0n" gave me a lot more people than math. I was expecting gorgeous algebraic curves, instead I got human curves. The current state of us and all the scales of the universe is a small local minima as part of wider more gigantic local minima as part of wider more gigantic local minima and so on countlessly you have access to your optimal policy at all times. you just choose not to follow it. you can literally access it by asking 'what should i be doing rn? Daily dose of antidepressive futurist hopium as an antithesis to doomerism which should be together synthetized into a beautiful cautious optimistically realistic whole https://x.com/TheHumanoidHub/status/1830481010410328465?t=CX5Vup0KjaIkWXQHG8zQ9Q&s=19 I don't implicitly model the attentional dynamics of others, which a lot of people implicitly do, so I have to learn and try to mentally do it from the ground up by myself, which some days I'm a little good at, but others I absolutely fail at it I have a love and hate relationship with my mind where I have to learn a lot of things from scratch by myself instead of having my brain give me the module preinstalled by itself Would UBI solve poverty completely, if agricultural and industrial revolution didn't fully solve it? Would poverty goal posts just move? What is the best AI/machine learning paradigm? Make sure to do your daily morning health and productivity habits no matter what even if nuclear war just started What's your favorite part of physics? What's your favorite thing in computational neuroscience? How much influence and power you have in our current system is in major part determined by how much money you have for lobbying Mitigating the risk of authoritarian corporate dystopia should be a global priority alongside other societal-scale risks such as pandemics and nuclear war. Lex Fridman is the ultimate unifier of memeplexes ❤