Long Ouyang
@longouyang
human + machine learning @OpenAI. previously: cognitive scientist @Stanford, baby
You might like
just played around a little with tinker from thinking machines.. definitely makes sft / rl a lot more accessible
genie 3 (incredible, of course) feels like something that should have come out of meta when it was more focused on VR.
proposal for new term - structurable data: unstructured data that can be turned into structured with the right llm invocations
cool feature of macos26 - integrated language model support that user applications can generally call. this would be a nice wedge for homelab inference powered by a mac mini or even homepod.
random nice use case for LLMs - asking questions about tv shows / movies you havent seen and requesting no spoilers
in general, pretraining data curation often removes ads or boilerplate, but when agents browse or do research, they can get inundated with the stuff - it feels like they can get distracted / fooled by fluff content, possibly due to differences in train/test distributions
United States Trends
- 1. Lakers 41,1 B posts
- 2. Happy New Year's Eve 15,8 B posts
- 3. annabeth 10,3 B posts
- 4. Lincoln Riley 3.551 posts
- 5. percy 15,5 B posts
- 6. Pistons 9.603 posts
- 7. #MiraclesAblazeDay2 24,3 B posts
- 8. #RIZIN 15,4 B posts
- 9. #AlamoBowl 1.002 posts
- 10. #LingOrmHNYatICONSIAM 2,1 Mn posts
- 11. grover 2.359 posts
- 12. #TAXSTRIKE 3.854 posts
- 13. Whale - Buy 2.571 posts
- 14. Sasser 2.031 posts
- 15. The 4D 230 B posts
- 16. Ayton 3.195 posts
- 17. OutKast 1.170 posts
- 18. olympus 5.066 posts
- 19. 5m TXs 1.037 posts
- 20. Bangkok 118 B posts
You might like
-
Jan Leike
@janleike -
Bob McGrew
@bobmcgrewai -
Mark Chen
@markchen90 -
向阳乔木
@vista8 -
Ivan Zhang 🦇🔊
@crlfq -
Nat McAleese
@__nmca__ -
Ashish Vaswani
@ashVaswani -
Yao Fu
@Francis_YAO_ -
Weijia Shi
@WeijiaShi2 -
Weiyan Shi
@shi_weiyan -
PJ Wu
@WuPingJu -
Ted Sanders
@sandersted -
Ethan Perez
@EthanJPerez -
Lukasz Kaiser
@lukaszkaiser -
Nick Frosst
@nickfrosst
Something went wrong.
Something went wrong.