Tom Tang (@nycShibe) header
Tom Tang (@nycShibe) profile picture

Looking for hidden variables.

Seattle, WA
Joined October 2008
Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)Photo by Tom Tang (@nycShibe)
Be serious about what you do. But never take yourself seriously. In mid-May, I posted about wokeism. Over 100 million views. Picked up by a ton of media outlets. And right after, Elon Musk and Javier Milei followed me. I'm not telling you this to flex. I'm telling you because it made me look at these two much more closely. And the more I watch them, the more convinced I am that they embody exactly what entrepreneurs will have to become. First, they don't take status seriously. Elon posts memes at 2am between rocket launches. Milei campaigned with a chainsaw. You can find that ridiculous. I find it deeply healthy. Because I'm so done with guys who, after a bit of business success, give themselves an air of importance that Elon doesn't have 1% of. The grave tone. The persona. The distance. As if a fundraise or an exit turned them into sages. Second, they're driven purely by infinite-sum games. A finite-sum game is a fixed-size pie: whatever you win, someone else loses. An infinite-sum game is one where the goal is to grow the pie and keep playing. And here's what I've come to understand: in finite-sum ecosystems, people almost always end up behaving badly. Not because they're bad people. Because the structure of the game pushes them there. When the pie doesn't grow, the only way to move forward is to take someone else's slice. Office politics, rent-seeking, backstabbing, envy. This is exactly what Peter Thiel understood when he wrote "Competition is for losers." It's not a VC punchline. It's a thesis about human nature. Thiel had René Girard as a mentor at Stanford, and Girard spent his life showing one thing: we desire what others desire. Mimetic desire. And when everyone wants the same thing in a limited space, rivalry turns destructive. Competitors become so obsessed with each other that they forget what they were supposed to build. Thiel's lesson: don't fight over the same pie as everyone else. Create a pie that didn't exist. And I believe that in the age of abundance that's coming, with AI, energy, robotics, this is precisely what will redefine capitalism. We're moving from a capitalism of capture to a capitalism of creation. Those who play finite games will tear each other apart over crumbs. Those who play infinite games will build what comes next. Elon doesn't take market share, he creates markets. Milei is trying to do the same thing at the scale of a country: pull Argentina out of a logic where people fight to divide up scarcity, and put it back into a game where wealth gets created. I deeply admire them both. And I think these two qualities are connected. Not taking yourself seriously isn't just a nice personality trait. It's a competitive advantage. It's what lets you keep learning. Because the day you start taking yourself seriously, you have a persona to protect. And when you have a persona to protect, you're afraid of being wrong, afraid of looking stupid, afraid of asking the naive question. You stop listening. And you stop learning. Status is a finite game: there's only one spot at the top, and everyone's fighting for it. Learning is an infinite game. Be serious about what you do. Never take yourself seriously.
38
222
9
958
22,777
“He jailed socialists, so he wasn’t a socialist.” This is the argument every single person in my replies thinks is a mic drop. It’s actually the weakest one available, and it collapses in five seconds. Lenin banned the Mensheviks and the Socialist Revolutionaries. Both were socialist parties. By your logic, Lenin wasn’t a socialist. Stalin shot Bukharin, Zinoviev, Kamenev, and sent an ice axe through Trotsky’s skull. All of them were founding fathers of the Bolshevik revolution. By your logic, Stalin wasn’t a socialist. Mao purged Liu Shaoqi, the President of the People’s Republic, and let him die in a cell. Pol Pot murdered his own cadres by the thousands. Robespierre sent Danton to the guillotine, then got the guillotine himself. There’s a reason Vergniaud said the Revolution, like Saturn, devours its own children. Purging rival socialists isn’t evidence against collectivism. It’s one of its most reliable signatures. Why? Because when the state decides everything, there can only be one plan. Two factions with two plans for the same economy can’t coexist. One has to disappear. In a free society, disagreements get settled by markets and elections. In a planned society, they get settled by the secret police. Hitler crushing the SPD and the KPD, then murdering Röhm and Strasser in his own party, is exactly what you’d predict from a regime competing for the same thing: total control of the economy and of society. And that’s the real question. Not “what did he call himself” and not “who did he arrest”. The question is: who runs the economy? Under the Nazis, independent unions were abolished and replaced by a state labor front. Prices and wages were frozen by decree. Investment was directed by a Four-Year Plan. Private property existed on paper, as long as owners did exactly what the state told them to do. Hayek saw this in real time. He dedicated The Road to Serfdom “To the socialists of all parties” precisely because he understood that the planners of the left and the planners of the right were the same species. Collectivism isn’t a color on the political spectrum. It’s the belief that a central authority can and should direct the economy. Once you accept that, the purges, the camps and the one-party state aren’t accidents. They’re the operating manual. So no, “he jailed communists” doesn’t refute anything. Communists jailed communists for seventy years. Read Hayek. Then come back.
This probably hits hard if you failed history in high school. Hitler was a far-Right fascist whose party was named the National Socialist German Workers' Party to attract working-class voters. Then he banned the Social Democratic and Communist parties, and sent socialists and communists to concentration camps.
315
1,367
109
5,814
110,489
We should teach kids to do amazing things, not prioritize empathy above all else. When Tesla demonstrated its flying Roadster, some people said, “That’s incredible!” Others asked, “What’s the point?” Still others said, “Cool, but it won’t cure world hunger.” That last reaction illustrates something deeply wrong with how we think about progress. The Roadster doesn’t need to cure world hunger. It doesn’t even need an immediate practical purpose. Doing something extraordinary is justification enough. Humanity didn’t overcome scarcity through empathy alone. A primitive tribe could care deeply about its starving members, but compassion wouldn’t save it from a failed harvest. What changed humanity’s circumstances was technology: generations of people inventing, experimenting, and pursuing possibilities without always knowing their eventual applications. We feed billions today not because we became more empathetic, but because we became more capable. Empathy matters. But empathy should not become a moral obligation to feel guilty about every problem we haven’t solved, or a demand that every ambitious project justify itself through immediate social utility. Teach children to build extraordinary things, to push boundaries, to be curious and audacious. Empathy has its place. But it should never extinguish the ambition that makes progress possible.
20
One of the most profound implications of converting code into an LLM is that software development could become a training problem rather than a programming problem. Today, every edge case requires modifying code, adding conditions, and managing growing complexity. What if instead, we simply trained the model on the desired behavior? Eventually, entire applications could be learned functions mapping inputs directly to outputs. No sprawling conditional logic. No endless refactoring. Just examples, feedback, and training. We’d move from managing code to managing behavior. That could fundamentally redefine software engineering.
Replying to @_chenglou
Here’s a little OS made with it. Giant scroll views, idiomatic text input, animation and styling, etc. You can check in the browser inspector that there’s no diffusion, no llm, and just regular DOM elements Yet if you check the GitHub source, it’s a bunch of compiled neural net numbers doing inference in real time in a browser
6
Tom Tang retweeted
My dear friends, I’d like to show you my latest toy project Introducing Handwired: turn your regular JavaScript code into a neural net! - no training - exact outputs vs the JS program - fully differentiable All runnable in a browser, without GPU
459
1,022
471
11,357
1,480,636
Programmable matter is the next great technological unlock. We’re seeing the same idea emerge at radically different scales: Atomic: Magic-angle graphene — rotating graphene layers by just 1.1° creates new electronic properties, including superconductivity. Microscopic: Atomic Machines — developing a “Matter Compiler” to manufacture functional micromachines directly from code. Macroscopic: Machina Labs — AI-driven robots form complex sheet metal parts directly from CAD designs, without specialized dies or molds. Electronic: Itera — developing reconfigurable liquid-metal circuits. The common thread: replacing specialized, fixed-function physical systems with general-purpose, programmable building. We may see a physical tech stack emerge from just a few components. Graphene/sheet metal/etc
19
What if cleanliness and allergies form a self-reinforcing feedback loop? Humans evolved immune systems adapted to parasites and microbes. Modern sanitation reduced those exposures, potentially increasing allergic susceptibility. Allergies may then motivate greater cleanliness and environmental avoidance, further reducing microbial exposure. A biological immune response reinforcing a behavioral one. An evolutionary adaptation that may have become a positive feedback loop.
8
Winter is coming for the international socialists. Hier j'ai écrit sur Gramsci. Beaucoup m'ont demandé : concrètement, c'est qui ? Alors faisons la carte. D'abord, ne vous laissez pas piéger par les étiquettes. Macron n'est pas socialiste au sens du PS. Bill Gates n'est pas marxiste. Soros est un financier. Aucun d'eux ne porte de drapeau rouge. Mais ils partagent une conviction, et c'est elle qui définit le socialisme international du XXIe siècle. Les grandes décisions sont trop importantes pour être laissées aux peuples. Il faut les confier à une élite éclairée, transnationale, qui planifie pour tout le monde. Plus de centralisation, plus de normes, moins de vote. Le Gosplan, mais en costume Davos. Voici les étages de la machine. Premier étage : la technocratie politique. Emmanuel Macron en est le produit le plus pur. Inspection des Finances, banque d'affaires, invité à la conférence Bilderberg en 2014 alors qu'il était à l'Élysée, président trois ans plus tard sans jamais avoir été élu à quoi que ce soit avant. Sa doctrine tient en une phrase : il existe un « cercle de la raison », et tout ce qui est en dehors est populiste. Deuxième étage : Bruxelles. Une Commission qui produit l'essentiel de la norme qui s'applique à 450 millions d'Européens, et dont aucun membre n'est élu directement. En 2025, la Cour des comptes européenne a publié un rapport sur l'opacité des financements de l'UE aux ONG. La même année, la Commission a dû reconnaître que certains contrats du programme LIFE avaient financé des ONG pour faire du lobbying… auprès de la Commission elle-même. L'argent public paie des militants pour faire pression sur ceux qui distribuent l'argent public. La boucle est parfaite. Troisième étage : les fondations. L'Open Society de Soros distribue de l'ordre d'un milliard et demi de dollars par an, dans plus d'une centaine de pays : ONG, médias, programmes juridiques, universités. La Fondation Gates est l'un des tout premiers financeurs de l'OMS. Un homme non élu pèse plus sur l'agenda sanitaire mondial que la plupart des États membres. Quatrième étage : les clubs. Bilderberg depuis 1954, à huis clos. Davos, où le WEF a vu son fondateur Klaus Schwab partir en 2025 dans un climat d'enquête interne. Ces lieux ne votent rien. Ils font quelque chose de plus puissant : ils fabriquent le consensus avant que le débat public commence. Cinquième étage : les relais. Les agences onusiennes, une partie de la presse subventionnée, une partie de l'université. Ce sont eux qui transforment le consensus des clubs en « ce que pense la science », « ce que pensent les experts », « ce que pense le monde civilisé ». Le dernier étage, c'était le robinet. USAID. On sait ce qu'il en est advenu. Pendant soixante ans, cette architecture a tenu grâce à un seul avantage : le monopole du récit. Ils contrôlaient la distribution de l'information. Donc ils pouvaient se tromper sans jamais payer le prix de l'erreur. Et ils se sont beaucoup trompés. L'hypothèse d'une fuite de laboratoire à Wuhan était un « complot ». Aujourd'hui le FBI, le Département de l'Énergie et la CIA la considèrent comme plausible ou probable. L'ordinateur de Hunter Biden était de la « désinformation russe », selon une lettre de 51 anciens responsables du renseignement. Il était authentique. Le nucléaire était dépassé. L'Allemagne a fermé ses dernières centrales en 2023 et redémarre ses centrales à charbon. Un récit qui doit être protégé de la contradiction n'est pas un récit fort. C'est un récit qui sait qu'il ne survivrait pas au contact du réel. Et le réel est arrivé. X a cassé le monopole de la distribution. Les podcasts ont cassé le monopole de l'interview. L'IA va casser le monopole de l'expertise. Le DOGE a cassé le monopole du financement. Quand tu ne contrôles plus le micro, il ne te reste que tes arguments. Et leurs arguments ne tiennent plus. L'Amérique latine bascule. Les États-Unis ont basculé. L'Europe est le dernier bastion, et c'est précisément pour ça qu'elle multiplie les lois sur la « désinformation » et les régulations des plateformes. On ne cherche pas à contrôler la parole quand on gagne le débat. Je ne crois pas à un grand soir. Je crois à un dégel. Lent, puis brutal. Gramsci disait que la crise, c'est quand l'ancien meurt et que le nouveau ne peut pas naître. Le nouveau est en train de naître. Et cette fois, personne ne pourra le faire taire. L'hiver arrive pour les socialistes internationaux. Et le printemps, pour tous les autres.
Dans 50 ans, les livres d'histoire diront qu'Elon Musk a gagné une guerre que personne ne savait qu'on était en train de perdre. Une guerre commencée il y a 100 ans, dans une cellule de prison italienne. En 1926, Mussolini fait arrêter un intellectuel sarde de 35 ans. Au procès, le procureur aurait dit : « Nous devons empêcher ce cerveau de fonctionner pendant vingt ans. » Il s'appelait Antonio Gramsci. Son cerveau n'a jamais cessé de fonctionner. En prison, il a rempli des milliers de pages qui sont devenues l'un des textes politiques les plus influents du XXe siècle. Je pense qu'on est en train de vivre le moment où sa théorie s'effondre. Laissez-moi vous expliquer. Gramsci se posait une question simple. Pourquoi la révolution communiste a-t-elle réussi en Russie mais échoué partout en Occident ? Sa réponse : en Occident, le pouvoir ne tient pas seulement par l'État, la police et l'armée. Il tient par la culture. Par ce que les gens trouvent normal, évident, « de bon sens ». Il appelle ça l'hégémonie culturelle. Donc, pour prendre le pouvoir, il ne faut pas d'abord prendre le palais. Il faut d'abord prendre les esprits. Les écoles. Les universités. Les journaux. Le théâtre. L'édition. Les tribunaux. C'est la « guerre de position » : une conquête lente, institution par institution. Quand la culture a basculé, la politique suit toute seule. Quarante ans plus tard, l'activiste allemand Rudi Dutschke en fera un slogan : la longue marche à travers les institutions. Pendant des décennies, on a vécu avec une énigme que personne n'arrivait vraiment à expliquer. Pourquoi presque tous les artistes sont-ils de gauche ? Les acteurs, les chanteurs, les profs, les journalistes, une bonne partie de la magistrature. Pourquoi un pays pouvait voter à droite élection après élection et voir quand même sa culture parler d'une seule voix ? On nous répondait : « c'est parce que les gens éduqués sont de gauche. » Je crois que la réponse est plus simple. Gramsci avait écrit la recette, et elle a été appliquée. Quand tu contrôles qui obtient la subvention, le poste, la bonne critique, le prix, l'invitation sur le plateau, tu contrôles ce qui devient « la culture ». La même énigme existait à l'échelle du monde. Pourquoi tant de pays en développement, notamment en Amérique latine, revotaient-ils encore et encore pour des gouvernements socialistes ou de gauche radicale ? Venezuela, Bolivie, Nicaragua, l'Argentine des Kirchner, la Colombie de Petro. L'économie s'effondrait, les cartels prospéraient, le chaos s'installait. Et pourtant le même vote revenait. Ma conviction : la pauvreté n'était pas un accident du système. C'était un électorat. Et autour, il y avait tout un écosystème d'ONG, de médias « indépendants » et de programmes de « société civile », financé en grande partie par l'argent occidental. Puis, début 2025, il s'est passé quelque chose que personne n'avait jamais osé faire. Elon et DOGE ont ouvert les livres de USAID. Une agence quasi intouchable depuis 1961, des dizaines de milliards de dollars par an. La grande majorité des programmes ont été coupés, et l'agence a été absorbée par le Département d'État. Le robinet a été fermé. Et regardez ce qui s'est passé depuis. Bolivie : le MAS perd le pouvoir après presque vingt ans. Équateur : Noboa réélu. Honduras : la droite reprend le pays. Chili : Kast élu. Argentine : Milei remporte largement les élections de mi-mandat. Et il y a deux jours, le Brésil. Jair Bolsonaro est inéligible, condamné à 27 ans de prison. La gauche pensait avoir refermé le chapitre. Lula, 80 ans, briguait un quatrième mandat face à un pays qu'on disait acquis. Dimanche, son fils Flávio est arrivé en tête du premier tour. Plus de 47 % contre près de 45 % pour Lula. Lula lui-même a reconnu un résultat « inattendu ». Le plus grand pays d'Amérique latine est à un second tour de basculer. Le plus grand virage à droite du continent depuis des décennies. Coïncidence ? Je ne crois pas. Quand le financement disparaît, le récit perd ses haut-parleurs. Il reste un dernier vrai bastion. L'Europe. Bruxelles finance ses propres réseaux d'ONG, avec une opacité que même la Cour des comptes européenne a pointée. Les grandes fondations comme l'Open Society de Soros. Les entités supranationales qui produisent de la norme sans jamais passer devant un électeur. C'est la dernière forteresse de l'hégémonie culturelle. Et elle commence à se fissurer. Je suis convaincu que tout ça va s'effondrer, et plus vite qu'on ne le pense. Gramsci avait raison sur un point essentiel : la culture précède la politique. Mais il avait oublié que ça marche dans les deux sens. Quand une culture artificielle perd ses perfusions, ce qu'il y a de vrai en dessous remonte à la surface. Ce qui vient après, je crois que c'est une renaissance. Une culture qui recommence à aimer la beauté, l'ambition, la science, la civilisation qui l'a fait naître. Et dans les livres d'histoire, on se souviendra que tout a commencé quand un homme a décidé de regarder les comptes. Il fallait quelqu'un d'assez riche pour ne dépendre de personne, d'assez fou pour s'y attaquer et d'assez libre pour encaisser la haine qui allait suivre. Merci Elon.
80
678
60
1,909
34,756
The elites of Victorian Britain got up late, drank ceaselessly, worked short hours, partied through the night, and knew nothing about science, technology or engineering. And yet, by their own lights, they were hugely effective. worksinprogress.co/issue/the… The Victorians are often held up as a model against the supposedly soft, feckless, humanities-educated elites of today. But I argue that the genius of their system may have been its drastic weighting towards soft skills, team-building and intra-elite networking: - A Victorian’s gentleman’s office hours typically began about 11am and finished about 5pm, including a multi-course lunch with wine. - Upper-class Victorians drank ceaselessly. Conduct guides recommended four or five glasses of wine at dinner *for married women*. Port was often served in pints. - Party culture was intense and nocturnal. Balls started at about 10pm, peaked after 1am, and continued till dawn. During the social season, bankers, judges and politicians might go to several balls per week. - Elite Victorians learnt almost no STEM at school or university. At least half of class time was given over to ancient languages, often as much as 80 percent. One leading educationalist remarked that it was not worth teaching science to gentlemen because it does not involve ‘continued reasoning’. - Sports and character formation were valued far above transmitting facts. A top headmaster described his school’s priorities as ‘First – character, second – physique, third – intelligence, fourth – manners, fifth – information’. This cannot have made elite Victorians more productive individually, but maybe the dense social and emotional network it created helped them to work as a sort of extended team. Perhaps we should think again about what makes groups of people effective.
263
625
288
5,407
1,044,828
In the last few months, I have been building deletes into LastDB. That's correct, the append-only db is no longer append-only! Why? Well, that's because as I've been using it, I've bloated my own DB with lots of data, many of which are telemetry data from the apps I've built and stale historical artifacts (From putting Git on LastDB, which I now regret immensely). I digress... Deletes are now possible! Now, this obviously wreaked havoc on cloud backup because with deletes, writing to the cloud and writing deletes to the cloud is 10x harder. On top of that, multi-device simultaneous writes/deletes are now the bane of my existence. I've got some major changes in the works to support it, primarily with a dual upload method (snapshots + mutation logs), but the devil is in the details, as we are all aware. I've been tempted to wipe my db many times to make things easier for myself, but everytime I ask claude, claude doesn't seem to think it's a good idea so I will keep on keeping on. So the good news is, I've been running my local db since March and no unintentional data loss yet! But performance is major issue now and I'm currently in the midst of wrestling with that beast. So through proving out the db, I'm also proving how memory hungry it currently is. Hopefully some much needed refactoring will make it better.
13
A French engineer who lives quietly in Paris has spent 30 years writing software that the entire internet now runs on without knowing his name. He wrote the code that streams every YouTube video, every Netflix show, every TikTok clip. He wrote the code that runs the virtual servers underneath AWS, Google Cloud, and Microsoft Azure. He calculated more digits of pi than anyone in history. He has no Twitter. He has no marketing. He just keeps shipping. His name is Fabrice Bellard. Here is the story, because almost nobody outside the systems programming world knows what one man has built. Fabrice was born in 1972 in Grenoble, France. He studied at École Polytechnique, the top French engineering school. He never went to Silicon Valley. He never built a startup empire. He just wrote code. In 2000 he started a project called FFmpeg, an open-source multimedia framework for encoding, decoding, and streaming video. He was 28. The project did one thing nobody else had done well. It handled every video and audio format that existed, in one library, on every operating system. He led it himself for years. Today FFmpeg is the invisible engine of the internet. YouTube uses it. Netflix uses it. VLC uses it. Chrome and Firefox use parts of it. Every Android phone, every iPhone, every smart TV, every video editing tool you have ever touched runs FFmpeg somewhere underneath. If you have watched a video on a screen in the last 20 years, Fabrice's code processed it. He was not done. In 2003 he started QEMU, a machine emulator and virtualizer. He wrote it solo until version 0.7.1 in 2005. QEMU lets you run any operating system on any other operating system. It became the foundation of modern virtualization. KVM, the Linux kernel hypervisor, runs on top of QEMU. Every major cloud provider, AWS, Google Cloud, Microsoft Azure, IBM Cloud, runs virtual machines on infrastructure built around it. The Quick Emulator is the most cited piece of cloud infrastructure code on Earth. He kept going. In 2001 he won the International Obfuscated C Code Contest with a small C compiler that grew into TCC, the Tiny C Compiler. TCC can compile and boot a Linux kernel from source in under 15 seconds. In 2004 he calculated the most digits of pi ever computed at the time, using a personal desktop computer and an algorithm he derived himself called Bellard's formula. In 2011 he wrote a complete PC emulator in pure JavaScript that runs Linux in your browser, a project called JSLinux that engineers still cannot believe is real. In 2019 he released QuickJS, a small but complete JavaScript engine that fits where V8 cannot. In 2021 he released NNCP, a neural network based lossless data compressor that immediately took the lead on the Large Text Compression Benchmark. Then he turned his attention to large language models. He built TextSynth Server, a web server with a REST API for running LLMs locally. He released ts_zip and ts_sms, compression utilities that use language models to compress text and short messages at ratios traditional algorithms cannot reach. He released TSAC, a very low bitrate audio compression system. In December 2025 he released Micro QuickJS, a new JavaScript engine for microcontrollers, separate from QuickJS, designed for environments with almost no memory. Fabrice co-founded a telecom company called Amarisoft in 2012, where he serves as CTO. Amarisoft builds 4G and 5G base station software used by carriers and labs around the world. He has been running it for over a decade while continuing to ship personal projects from his own home page at bellard dot org He has no Twitter. He has no Instagram. He gives almost no interviews. His personal website is a flat list of projects with no styling, no fonts, no marketing copy. Just titles and links. A quiet French engineer who never moved to Silicon Valley wrote the code that quietly runs the internet. He is still shipping.
113
1,367
160
7,240
1,277,161
This highlights the dangers of connecting personal accounts to AI services. It becomes a giant honeypot increasing the probability of attack the more connected it becomes across all users who are connected.
On July 25, we hacked OpenAI. Two bugs let us take over ChatGPT/Codex accounts of OpenAI employees (+some unaffiliated users) and reach connected services: Outlook, Slack, GitHub, etc. We proved it with a PR in OpenAI’s internal codebase . It took us <72h. 🧵
23
Tom Tang retweeted
🔥 Today, we are truly excited to announce our technical prototype, the Generative World Simulation system, which integrates JING(镜), an interactive experience model, with DAO(道), a computable shared-world engine. The coupled model and engine connect first-person experience with a shared world that continues to evolve beyond any individual observer. Conditioned on actions and observation history, JING enables agent navigate, manipulate, and communicate from a first-person perspective in the world. Watch our demo video to see it in action! On the official WBench leaderboard as of September 17, 2026, XGEN-JING ranked #1 on the Full split, and #2 on the Navi split. 🎉 DAO maintains shared world state and rules, computes the consequences of actions, and provides JING with only what the current observer can perceive. It also supports autonomous agent decision-making, enabling agents to act independently within an evolving shared world. Together, DAO and JING move beyond generating the next frame toward simulating the world behind it. This marks a small step towards OASIS: not just a world that responds to you, but a world—and a society—that evolves with and without you. 💪 Explore XGEN Labs~: 🔗 Website: xgenlabs.ai/ 🤗 HF: huggingface.co/XGENlabs/XGEN… 🦊 GitHub: github.com/XGEN-Labs/XGEN-JI…
357
538
301
2,386
1,509,314
I've seen people describe Jev as an "AI if statement". But what if it actually WAS an if statement? Introducing Probably: a programming language powered by Jev: probably-lang.southpolesteve… Jev baked into the language. “feels” asks a question. “match” routes between descriptions. “while” keeps going until something stops feeling true. This is obviously a toy, but it's fun to think about what something like Jev unlocks. Jev makes the decisions, an LLM does the writing, and a little program ties it together.
199
305
230
4,633
984,830
Qwen 3.8 flash is truly impressive, and llama.cpp has been rapidly getting better and better at processing it columns are: pre-existing prompt, new prompt, generated, input tok/s, output tok/s This is on my laptop (strix halo). I think we're very close to the point where you can just use local models for a large share of tasks, and for anything more advanced, workflows like "use your local model to orchestrate queries to powerful models so your queries don't leak your personal information" actually become viable.
272
190
54
2,096
384,094
Should you believe in something even if scientific evidence suggests it’s not true? Yes — if it’s shown to produce better outcomes. For example, free will has been undermined by scientific literature for decades, but the evidence for better outcomes has been unwavering for those who believe it to be true. There are other beliefs that fall into a similar category such as “The Secret”, “Affirmations”, or variations of wishful thinking. The biggest one being the belief in an omnipotent god. Despite zero evidence supporting the existence of a supernatural being, believe in him/her confers real tangible benefits that are measured in economic output, military strength, and better communities. The belief in beliefs that confer good outcomes I call beliefism. Scientific evidence for the belief being rational or not typically comes from reductionist methods. And while evidence for or against the existence of free will gets updated over the ages as technology advances, beliefism side steps those arguments. If it’s creating good outcomes, believe first.
12
Tom Tang retweeted
Introducing the US Gov Graph - a complete map of the people and positions of power in the federal government. To fix our institutions we must understand how they work, so we are using AI to model and monitor every government in America. This is Palantir for The People. Check it out here: graph.civlab.org/us We exist to set the conditions for an era of civic excellence unmatched since the Founding. We will need nothing less to navigate the age of AI. I'm looking for exceptional engineers and designers to work with me in San Francisco. If that's you, inquire here: tally.so/r/KY97RK. Self-governing civilization is counting on you.
475
1,574
271
9,553
606,294
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
4,006
8,256
6,821
76,494
40,212,158
Tom Tang retweeted
Elon just gave a much more practical explanation of the current AI safety problem....and laid out a practical framework for making frontier AI safer OpenAI and Anthropic are the two leading AI companies....and their models are close enough in capability that it’s incredibly difficult for either one to slow down without basically handing the lead to the other Elon also made an important distinction: On balance, Anthropic puts more care into safety than OpenAI....but even people inside Anthropic are publicly worried about what their own models can do So instead of asking one company to voluntarily fall behind, Elon’s solution is much more practical: Have Anthropic run its safety test harness on OpenAI models Have OpenAI test Anthropic Have SpaceXAI test both And bring the leading Chinese AI companies into the same system Everyone tries to break everyone else’s models before release “The odds that you will find issues are dramatically greater” That makes way more sense than asking frontier labs to grade their own homework
All-In Summit: Elon Musk & Gwynne Shotwell -- AI Risks and Peer Review -- Starship -- Terafab as Taiwan Insurance -- SpaceX/Tesla Merger Potential -- Management at SpaceX ++ much more! (0:00) SpaceX's @Gwynne_Shotwell joins The Besties! (1:45) Gwynne's SpaceX story, selling rockets, and working for Elon (6:45) Running SpaceX: AI, Starlink, Rockets, and X (13:43) Direct to cell with Starlink, retiring rockets, competition (18:34) Data centers in space (22:00) Management at SpaceX (27:37) @elonmusk joins: AI's real risk, model peer review, what he meant by "Dario is right" (36:53) Elon and Gwynne on their working relationship, Starship's future (46:16) Terafab, Tesla, Flying cars?, merging Tesla and SpaceX (52:51) Lying AIs, how to do AI peer review right $TSLA $SPCX
296
534
70
2,916
612,632
The "free" in "free market" doesn't mean "free from restriction". It means "free from coercion". In order to get "free from coercion", you need quite a lot of restriction. Whether that restriction is applied by laws, by religious indoctrination, by cultural mores, by a technology stack, or by tribal loyalty, doesn't really matter. In other words, the free market is not a natural state of affairs. It is a garden protected by walls, pruning shears, and vigorous weeding. In an environment without restriction, those who are able will apply coercion constantly, at the slightest hint of advantage, and almost no one except the very fortunate and well-placed will be free. If the grocer or the restaurateur is "free" to add addictive things to the food you buy, and "free" not to tell you about it, then you are not free to choose what you buy and consume. Instead, you are coerced by the chemical hijacking of the basal mechanisms your brain depends on. If your nation engages in "free" trade with other nations where slavery is legal, then the price of your labor is set not by its value to you and others, but by the price of guns, whips, cattle prods and overseers in the slave camps. You may sell your labor or not as you choose, but that choice is meaningless when you must compete with the owners of those who do not have it. And if you must compete with imported labor who can be sent back to a terrible place if they quit, strike, or even complain, then the difference is only one of degree. A market free from coercion is not only compatible with the thriving of the family and the community, it is necessary for both. Sadly, this concept has been so driven from both our public consciousness and our legal practice that the very words "free market" have been stolen by slaver and used to endorse slavery. Slavery is the worst of all possible evils, because unlike evils like murder, robbery, or rape, it harms not only the direct victim, but everyone who must share an economy with the slaver, who reaps the benefits alone. Ultimately, the word "free" in "free market" means freedom from slavery.... and not merely the direct and obvious kind, but all forms of coercion and slavery by degrees. Unfortunately, modern libertarians are mostly uninterested in this, and prefer to focus on sex and drugs.
The problem is "not left" isn't the durable basis for a coalition Opposing wokeness served as a Shelling point but "the right" contains several world view that are mutually exclusive so when they attempt to use shared language discourse breaks down Liberty to the libertarian is not liberty to the religious conservative, the propositional nation and traditional definition of nation can't co-exist, we can't prioritize totally unfettered economic transactions and the wellbeing of the family Someone has to win these positions and someone has to lose
33
174
15
1,288
31,723