🔥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

great thread, summarizes well why we're particularly excited about RL higher inference to training compute ratio -> less inter node communication -> better suited for globally distributed training infra with slow connection speeds

We are so back btw (got submitted in final but hey)

First tweet is a banger, I will watch his career with great interest

Highest leverage thing unskilled engineers can do rn to contribute to frontier AI research is vibecoding RL environments

SYNTHETIC-2 Datasets are now on Huggingface! We’re releasing an SFT dataset collected from the new R1-0528 as well as an RL Dataset with difficulty annotations from various smaller models. Go train some models 🫡

I have acquired Windsurf

Day 1 of asking for this at Prime Intellect HQ

Toploc Poster session tomorrow (Wed) at 4:30 PM East Hall E-1106 I’ll be around through Saturday; if you’re into decentralized training & inference, lets chat!

At #ICML2025 in Vancouver 🇨🇦 this week, presenting some work from my first year at Stanford! Come find me at posters or just around the conference! Thursday: KernelBench: Can LLMs Write Efficient GPU Kernels? 11AM East E-2010 Saturday: Kevin: Multi-Turn RL for Generating

my piquant quantization kernels are almost 50 times faster than pytorch's on the CPU. pytorch’s sub‑byte quantization (torch.quint4x2, torch.quint2x4) is quite slow. 1/2

RL with predefined tools does not matter in the long term, the most bitter lesson pilled approach is giving the model a single universal tool (a computer)

this year alone, I've met hundreds of the world's elite AI researchers + engineers steering the future of intelligence, and they ultimately want to do it here, in the US 🇺🇸 If a visa or green card are holding you back, join us on 7/31 in SF and hear real stories from

How to train the best non-reasoning model: 1. Gather a reasoning dataset 2. remove <think> tokens 3. Train the model

While LLMs are good at generating functionally correct frontend code, it’s stunning how bad AI-generated UIs are; I’m certain that this can become better with appropriate evals and reward signals Really excited about this leaderboard and the very hard-working team behind it!

Just landed in SF, I'm now stranded here without a desk while the rest of my team is still in Europe. If anyone can host me at their office for (one of) the next few days, please let me know (DMs open 🥹👉👈)

Noticed some curiosity about the specific score comparison between GSPO and GRPO. From our perspective, we’re more focused on scalability — can we achieve better performance by increasing compute (e.g., training with more steps, extending generation length, regularly updating