Thrilled to finally share what we've been working on for months at @huggingface 🤝@pollenrobotics. Our first robot: Reachy Mini. A dream come true: cute and low priced, hackable yet easy to use, powered by open-source and the infinite community. Tiny price, small size, huge

Big progress in visual reasoning for OSS models. I find it very exciting as it's key for VLMs even more than for text models: direct image Q&A is much less useful than brainstorming on a picture/doc. Also high-taste folks see Kimi/Qwen comparisons and know it's serious.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

RT @arithmoquine: new post. there's a lot in it. i suggest you check it out

RT @Hesamation: a psycho voice in my head begs me to just buy this book and go all-in on ultra-scale LLM training. hugging face cooked big….

RT @matistanis: Trying to do the same for London! If you’re around and want to join the invite list for the next event with other builders….

RT @_arohan_: Beautifully illustrated book on large scale training on gpus!.

RT @TheZachMueller: In order to celebrate the release of the print version for the Ultra-Scale Playbook (of which I have no affiliation wit….

Look what just arrived in the mail 📬 So excited to hold the final printed version in my hands after many months in the making. And now it’s out for everyone! (shortlink in the video. )

Host applications for the Kids AI Vibe Coding event this October are exploding 🚀. We’re already on 4 continents🌍 with amazing partners like LinkedIn, Bayer, Stanford, Cognizant & more. Wanna bring this free 9–12 yo AI coding event to your city? Join →

RT @NiJinjie: Token crisis: solved. ✅. We pre-trained diffusion language models (DLMs) vs. autoregressive (AR) models from scratch — up to….

I’m serious about this, though. LLM agents are (gullible) yes-men deeply integrated behind companies’ auth access. It’s a matter of time before we get a first serious data breach where hackers simply emailed a text file or fill a form convincingly asking agents for confidential.

and yes I dared use an em dash —no shame.

We have virus scanners, so why aren’t jailbreak file scanners a thing yet?. Most AI threats will come from text files—not hacked binaries like before.

If you wanna get addicted to an AI model, choose an open-source one. Nobody can snatch it away. I’m serious.

RT @ludwigABAP: The "Circuit Analysis Research Landscape" for August 2025 is out and is an interesting read on "the landscape of interpreta….

RT @rohanpaul_ai: OpenAI's GPT 5 System card (a 59 page doc). Key Takawsys 👇. 🚦 The model router. The chat system holds two main models.….

RT @mervenoyann: GPT-4.1-mini level model right in your iPhone 🤯. @OpenBMB released MiniCPM-V4, only 4B model surpassing GPT-4.1-mini in vi….

On Genie 3 potentialities

The OpenAI team has been cooking! gpt-oss is a beast. Welcome back folks, it’s great to have you in the open-source community again. Excited to see what people will build on top of gpt-oss. On the personal side, it’s been such a pleasure to work with the OpenAI team and friends

RT @dylan_ebert_: OpenAI just released GPT-OSS: An Open Source Language Model on Hugging Face. Open source meaning:.💸 Free.🔒 Private.🔧 Cust….

RT @mekkcyber: The new GPT-OSS models are Mixture of Experts (MoEs), with 20B and 120B parameters. Since expert weights make up ~90% of th….