frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

RL on GPT-5 to write better kernels

https://arxiv.org/abs/2602.11000
4•atallahw•1h ago

Comments

atallahw•1h ago
We got access to OpenAI's RFT API with GPT5 and tried to see how good we could get it at one-shot Triton kernel generation. Some key decisions/observations: 1. tool use instead of multi-turn rl 2. skip SFT altogether 3. dataset curation was more important than dataset scale 4. reward hacks detection must be robust 5. models are getting a lot better at this