I'm releasing 53 slides on post-training, covering core algorithms like DPO and GRPO, as well as data quality, synthetic data pipelines, and on-policy training. I had the pleasure of presenting it yesterday as a guest lecturer in Cambridge, UK

It is becoming clearer that Jevons paradox applies to competent human software engineers. If AI makes them more efficient and more productive, demand for their work will increase.

Steve Jobs: on what separates the successful founders from the unsuccessful ones.

Kimi K2.5 tech report just dropped! Quick hits: - Joint text–vision training: pretrained with 15T vision-text tokens, zero-vision SFT (text-only) to activate visual reasoning - Agent Swarm + PARL: dynamically orchestrated parallel sub-agents, up to 4.5× lower latency, 78.4% on

Kling 3.0 Model is coming! Now in exclusive early access. Stay tuned for what’s next. 🚀

hi @togethercompute , where is kimi k2.5 @Kimi_Moonshot

Introducing the Mistral 3 family of models: Frontier intelligence at all sizes. Apache 2.0. Details in 🧵

WOOOOOOO

does @github git is having issues , now #git

Wait, why does https://t.co/niXXANrBgv redirect to Google?

Thank you for sharing this research 🙇‍♂️

Weekend is very much sorted now , Enjoy 🦙 "We are not actually building animals , we are building ghosts" https://t.co/UcUacZ5WTw

Tribute to grok-code-fast-1 I’ve used countless LLMs for coding from the popular ones to the niche. But there’s something truly special about grok-code-fast-1. When every other model stumbles on those tricky, edge-case problems… this one shines. It handles the hard stuff the

https://t.co/Z6ohVMu3PA

what the hell is this "NGHTTP2_ENHANCE_YOUR_CALM" @cursor_ai

Biomni is opensource now, Thanks to the team . @ProjectBiomni https://t.co/zWGFRwaTUq https://t.co/p83D4MsK08

https://t.co/bDPFzyJnGD

“Somewhere, something incredible is waiting to be known.” - Carl Sagan