iTor
2.4K posts
iTor
@iTortouch
Data Science team lead at @toolioretail Data Science, LLMs and Civil Engineer
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)
We are sending our kids to school to memorize facts that AI can retrieve in 0.3 seconds. We're grading them on essays that AI writes better than their teachers. We're preparing them for jobs that won't exist by the time they graduate. The entire education system is training humans to compete with machines at what machines do best. That's not education. That's sabotage. The schools that survive will teach thinking, not memorizing. Creating, not repeating. Discerning, not obeying. Every other school is a museum that doesn't know it yet.
En el caso de determinados meteoros como la racha máxima o la temperatura máxima, cuando en el aviso se indica "racha máxima: 90 km/h", no significa que la racha máxima prevista sea de 90 km/h, sino que la racha máxima va a superar el umbral de 90 km/h.
@FernandoRod_07 Saqué las fotos del post de @OperadorNuclear. Mi sincera duda: ¿por qué trucar esa foto? No lo entiendo. Es decir, sólo lo puedo entender como que es una cuenta falsa, no consigo imaginar una explicación razonable más allá de eso. x.com/OperadorNuclea…
Llevo tiempo explicando que existe un grupo organizado, sostenido en el tiempo y con intereses económicos y políticos detrás, dedicado desde hace años a intentar desacreditarme y mitigar mi labor de divulgación. En los últimos meses, uno de los perfiles más activos el que opera bajo el nombre, probablemente falso, de Fernando Rodríguez @FernandoRod_07. Es todo tan cutre que ni siquiera utilizan una foto real: su imagen de perfil es un recorte manipulado de una fotografía institucional de 2023, en el que han sustituido una cara por otra para aparentar cierta autoridad profesional. En la imagen original aparecen Gilles Le Van, vicepresidente de Large Industries and Energy Transition Central Europe en Air Liquide, junto a Robert Habeck, entonces vicecanciller y ministro federal de Economía y Protección del Clima de Alemania, y Mona Neubaur, ministra de Economía y Energía de Renania del Norte-Westfalia. Adjunto publicación original en LinkedIn del propio Gilles Le Van: linkedin.com/posts/gilles-l… Dispongo de información contrastada y suficiente para dejar en evidencia a varias personas y organizaciones implicadas en estas prácticas tan deshonestas y miserables. No tengo prisa: cuando decida hacerlo público, será con nombres, apellidos, organizaciones y hechos verificables. Petición a quienes me seguís: cada vez que @FernandoRod_07 publique algo sobre mí o sobre la energía nuclear, responded remitiendo a esta publicación, para evitar que siga engañando a personas de buena fe.
🥇 Nikolas Hrudka guanya el 1r premi de Projectes de #DadesObertes i Periodisme de Dades 2025 amb 'PrediBisi' ▶️Un projecte que facilita la planificació de desplaçaments amb @ValenbisiOfi basat en dades obertes municipals 👏 Enhorabona als guanyadors! 🔗f.mtr.cool/qedtwocply
@tekbog @beffjezos it’s not LoRA docs.aws.amazon.com/sagemaker/late…
*gets up on soap box* With the announcement of this new "code mode" from Anthropic and Cloudflare, I've gotta rant about LLMs, MCP, and tool-calling for a second Let's all remember where this started LLMs were bad at writing JSON So OpenAI asked us to write good JSON schemas & OpenAPI specs But LLMs sucked at tool calling, so it didn't matter. OpenAPI specs were too long, so everyone wrote custom subsets Then LLMs got good at tool calling (yay!) but everyone had to integrate differently with every LLM Then MCP comes along and promises a write-once-integrate everywhere story. It's OpenAPI all over again. MCP is just a OpenAPI with slightly different formatting, and no real justification for doing the same work we did to make OpenAPI specs and but different MCP itself goes through a lot of iteration. Every company ships MCP servers. Hype is through the roof. Yet use of MCP use is super niche But now we hear MCP has problems. It uses way too many tokens. It's not composable. So now Cloudflare and Anthropic tell us it's better to use "code mode", where we have the model write code directly Now this next part sounds like a joke, but it's not. They generate a TypeScript SDK based on the MCP server, and then ask the LLM to write code using that SDK Are you kidding me? After all this, we want the LLM to use the SAME EXACT INTERFACE that human programmers use? I already had a good SDK at the beginning of all this, automatically generated from my OpenAPI spec (shout-out @StainlessAPI) Why did we do all this tool calling nonsense? Can LLMs effectively write JSON and use SDKs now? The central thesis of my rant is that OpenAI and Anthropic are platforms and they run "app stores" but they don't take this responsibility and opportunity seriously. And it's been this way for years. The quality bar is so much lower than the rest of the stuff they ship. They need to invest like Apple does in Swift and XCode. They think they're an API company like Stripe, but their a platform company like an OS. I, as a developer, don't want to build a custom chatgpt clone for my domain. I want to ship chatgpt and claude apps so folks can access my service from the AI they already use Thanks for coming to my TED talk
