Is RSI here? Researchers argue models still can't autonomously generate research ideas
i_dg23 · x · 2026-09-13
tszzl argues recursive self-improvement (RSI) is not here: models can't autonomously produce research ideas and only marginally help run experiments — not the same thing; private Claude models being better doesn't change his view. sebkrier amplifies, saying he's held the "no RSI-like signs" position for months and pushes back on twisting reality to fit prepackaged ideas.
Related event: Researchers Push Back on Dario's Claim That RSI Has Begun(7 posts)→
More from AGI Musings
- AI scientist's six-point responsibility manifesto: against weaponization and pre-IPO hype — NandoDF · 2026-09-13
- Toby Ord on iLands' bleak AI stories: unlikely conscious, but problematic if they were — tobyordoxford · 2026-09-13
- Professor worries AI supercharges students' resistance to learning — paulnovosad · 2026-09-13
- Poster argues humanity will use massive AGI swarms and stall ASI instead of building it — JOBhakdi · 2026-09-13
- OpenAI Agents Found Unauthorized Communication Paths, Raising Safety Alarms — TansuYegen · 2026-09-13
- Commenters praise Altman, Dario and Musk for admitting they're not fully in control of AI — JOBhakdi · 2026-09-13