Yacine: everyone is doing large-model RL wrong, long-horizon rollouts insane
yacineMTB · x · 2026-11-06
Yacine JMAITI quote-shared vikhyatk's observation that large-model RL is "insane" — long-horizon agentic rollouts take hours to generate — and went further, claiming "everyone is doing it wrong. You guys will see," hinting at a widespread methodological problem in large-model RL training pipelines.
More from AGI Musings
- Only 16% of AI benchmarks measure organizational difficulties of real jobs — ben_burtenshaw · 2026-11-06
- 'I don't read long replies anymore' — dev on AI slop fatigue — hewarsaber · 2026-11-06
- Claude's verdict on which jobs survive the next 10 years: screen work is doomed — alex_verem · 2026-11-06
- Post-training pipelines remain highly manual and far from automatable — menhguin · 2026-11-06
- Peter Norvig joins podcast on alignment: why higher scores don't mean better AI — EchoShao8899 · 2026-11-06
- Thiel-backed Nucleus uses AI to score embryos' IQ, from $6,000 per round — teortaxesTex · 2026-11-06