Today is a big day Today is Silksong day But most importantly, today is the day I finally got HF socks!!! https://t.co/pvzDhpaSL7

🚀 Big news: we just added Reachy 2 to LeRobot! Huge thanks to our friends at @PollenRobotics 💛🤗 Reachy 2 is also available in simulation, so you can try it out right away. 🎥 Check out the teleop & autonomous demo below! https://t.co/NeZKBshox3

Fuck it. Today, we open source FineVision: the finest curation of datasets for VLMs, over 200 sources! > 20% improvement across 10 benchmarks > 17M unique images > 10B answer tokens > New capabilities: GUI navigation, pointing, counting FineVision 10x’s open-source VLMs. https://t.co/S9LaWM2iVO

Hermes-4-14B has been released! https://t.co/mX8Bt9apBH Our most compact LLM from the Hermes 4 series is locally usable and optimized for consumer hardware, providing at-home access to its powerful hybrid reasoning and tool calling. The model is now available on Nous Chat for online use and downloadable on HuggingFace.

Hermes-4-14B has been released! https://t.co/mX8Bt9apBH Our most compact LLM from the Hermes 4 series is locally usable and optimized for consumer hardware, providing at-home access to its powerful hybrid reasoning and tool calling. The model is now available on Nous Chat for online use and downloadable on HuggingFace.

Honestly FineVision is a pretty impressive work of aggregation 200 training sets condensed in a dataset of 18B images, segmented in 9 different subcategories, multi-turn, with quality rating and very documented ablation studies? As always, @huggingface delivers in open data https://t.co/5DmP1ZaG6J

“It’s a good dog, @ClementDelangue” vibes https://t.co/Jw5ckhIMNJ

“It’s a good dog, @ClementDelangue” vibes https://t.co/Jw5ckhIMNJ

Cannot wait to try the new Kimi K2! @Kimi_Moonshot https://t.co/qx62R7YPvN

Cannot wait to try the new Kimi K2! @Kimi_Moonshot https://t.co/qx62R7YPvN

Introducing EmbeddingGemma🎉 🔥With only 308M params, this is the top open model under 500M 🌏Trained on 100+ languages 🪆Flexible embeddings (768 to 128 dims) with Matryoshka 🤗Works with your favorite open tools 🤏Runs with as little as 200MB https://t.co/AXPqV4aXr1 https://t.co/zMytzO0ani

Brand new, fresh out of a French printer https://t.co/kok8vfbMjF

Brand new, fresh out of a French printer https://t.co/kok8vfbMjF

Quietly landed on the hub you can try ROSE on @huggingface 🤗 —> https://t.co/iOFg3LdWUz

Quietly landed on the hub you can try ROSE on @huggingface 🤗 —> https://t.co/iOFg3LdWUz

There is significant unmet demand for developers who understand AI. At the same time, because most universities have not yet adapted their curricula to the new reality of programming jobs being much more productive with AI tools, there is also an uptick in unemployment of recent CS graduates. When I interview AI engineers — people skilled at building AI applications — I look for people who can: - Use AI assistance to rapidly engineer software systems - Use AI building blocks like prompting, RAG, evals, agentic workflows, and machine learning to build applications - Prototype and iterate rapidly Someone with these skills can get a massively greater amount done than someone who writes code the way we did in 2022, before the advent of Generative AI. I talk to large businesses every week that would love to hire hundreds or more people with these skills, as well as startups that have great ideas but not enough engineers to build them. As more businesses adopt AI, I expect this talent shortage only to grow! At the same time, recent CS graduates face an increased unemployment rate, though the underemployment rate — of graduates doing work that doesn’t require a degree — is still lower than for most other majors. This is why we hear simultaneously anecdotes of unemployed CS graduates and also of rising salaries for in-demand AI engineers. When programming evolved from punchcards to keyboard and terminal, employers continued to hire punchcard programmers for a while. But eventually, all developers had to switch to the new way of coding. AI engineering is similarly creating a huge wave of change. There is a stereotype of “AI Native” fresh college graduates who outperform experienced developers. There is some truth to this. Multiple times, I have hired, for full-stack software engineering, a new grad who really knows AI over an experienced developer who still works 2022-style. But the best developers I know aren’t recent graduates (no offense to the fresh grads!). They are experienced developers who have been on top of changes in AI. The most productive programmers today deeply understand computers, how to architect software, and how to make complex tradeoffs — and who additionally are familiar with cutting-edge AI tools. Sure, some skills from 2022 are becoming obsolete. For example, a lot of coding syntax that we had to memorize back then is no longer important, since we no longer need to code by hand as much. But even if, say, 30% of CS knowledge is obsolete, the remaining 70% — complemented with modern AI knowledge — is what makes really productive developers. (Even after punch cards became obsolete, a fundamental understanding of programming was very helpful for typing code into a keyboard.) Without understanding how computers work, you can’t just “vibe code” your way to greatness. Fundamentals are still important, and for those who additionally understand AI, job opportunities are numerous! [Original text: https://t.co/nqzPC6eUpR ]

@sama Smarter! I can go for a nice walk or something it's okay. https://t.co/1RiTqSFsKv

@SunnySanyal9 They are not entirely bad. Can't speak about proprietary models, but e.g. there is a state space model ranked similar to gpt-5-mini-high on the LM Leaderboard: https://t.co/s3x20fqE9t

Really liking the chainlit open source lib for building a quick but nice chat interface for any LLM. Here are some quick single and multi-turn examples for my Qwen3 from-scratch models: https://t.co/XnVaGVHPk8 https://t.co/Y0LlVJz0yK

@TragicHero628 @ollama @OpenWebUI Ollama is my go to for running any existing LLMs listed on the ollama website, but I think if you want to add custom LLMs, you have to convert them yourself and they also have to follow a specific format (https://t.co/LbXWqrsoPI). So, I'd say it's a less general solution and more for the post-prototyping phase.

@gpjt Nice! This reminds me of my linalg and PyTorch slides (https://t.co/R48OgiBXHy) that you may or may not like, haha https://t.co/E6EDSNcXl5

@adven_raj @Krishna70284154 @karpathy Glad you are liking it! Btw if it has been a while since you read it, I also added lots of bonus material on GitHub, including from-scratch implementations of more recent LLM like Qwen3 and Gemma 3: https://t.co/j3Knzj7TGW

daily reminder that one of the most important tech discussion sites on the web uses tables for layouts & basically your stupid https://t.co/qDM1udYtcj

The weirdest VC subsidizing of our time, 10% of the Anthropic series F goes to writers https://t.co/M7JC2Aca8y

The weirdest VC subsidizing of our time, 10% of the Anthropic series F goes to writers https://t.co/M7JC2Aca8y

@catehall Eh whatevs https://t.co/sS0z0V8wqK

Hi friends. 👋 Unfortunately I am struggling financially right now. I'm still looking for ML/DL roles in Python, LLMs, data science, automation (n8n) workflows. If you know of any leads please let me know. 🙏 Some of my work: https://t.co/JfxFVRqKu5 https://t.co/WJgJ4UtliA

If folate deficiencies really did cause autism, then you would expect to see a sharp drop in the autism-by-birth-year curve in 1997, when folic acid fortification was mandated. Instead, the curve continued its upward trend unabated. https://t.co/4zLmErJSDY

Amp's CLI is now a proper, full-on TUI. Best part for me personally? No more Ink, no more flicker. https://t.co/k6D0rc4UiQ

Had such a blast today at the Gemini event thanks to @agihippo for organising it. Hoping for more of these with a growing deepmind presence https://t.co/hTWr8IVVLB

And I got to meet @jeffdean ! It was so cool to meet him in person haha https://t.co/ZLYxTZwE4d

@quocleix shared too about his journey with LLMs from seq2seq to the present day! https://t.co/xllr9Kzvzr