🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

This was so aura-maxxed that it will cause a bunch of people on the fence to finally move to SF.

Open AI Gym, a single-agent open source RL environment released by Open AI in 2016 (now called Gymnasium) was way ahead of its time:

Super cool! A hackable little robot from Hugging Face. Congrats Thomas Wolf, clem 🤗 and team!

Grok 4 achieves 66.6% on ARC AGI 1. So many folks ignored this benchmark for a long time and in the early days. Congrats to the xAI team

Seeing these results on ARC-AGI from Grok 4, I am so tempted to spin up different RL environments in the form of similar adjacent games and see how far the (new) small <7B models can get with ARC-AGI. Last November got to the top 1% of the ARC AGI submissions with no RL and very

Grok 4 is available in Cursor! We're curious to hear what you think.

This is a good benchmark, TritonBench. Most LLMs aren’t great at writing Triton code for GPU kernels (there isn’t much Triton code publicly available yet, but it’s definitely increasing):

NIK A Cursor engineering team is at xAI today, so integration issues are being solved in real-time

AI now writes 50% of the code at google:

1T MoE Param model, open source, 32b active:

The dream for a long time of not using tokenizers at all might be here:

Lee Kuan Yew: “Air conditioning was a most important invention for us, perhaps one of the signal inventions of history. It changed the nature of civilization by making development possible in the tropics. Without air conditioning you can work only in the cool early-morning

This is quite interesting. The tokens per parameter in LLaMA 4 seem off (to what LLMs are trained on today): We are well over the Chinchilla optimal (20 tokens per param), but Llama 4 Behemoth had (only) 104 Tokens Per Param (TPP), similar to Mistral7B (dense), while Llama 4

The improvement rate of Cursor is amazing, it keeps getting even better and better

If you be super neat if the labs producing the top OS LLMs like Kimi K2, etc. could release the smaller distilled versions of them (like DeepSeek R1 with the smaller distills they released), so that we all can run the most optimized, on pair speculative decoding

Neat! Daniel Han you guys work fast! Dynamic 1.8-bit (245GB from 1.1TB) for Kimi K2:

Can’t wait for Kimi K2 reasoning - K2 (base) seems pretty good even on creative writing. Hopefully we can get better training/sampling efficiencies with RL and can't wait to see how these models perform when RL compute is the vast majority of training (and hopefully when it's