RSI models' 'algorithmic progress' is too narrow — call it software improvement
CFGeek · x · 2026-11-16
A debate on recursive self-improvement (RSI) models argues that circulating versions rest on an outdated understanding of "algorithmic progress," arguably the most important parameter.
CFGeek notes the notion is actually quite broad in most RSI models and might be better termed "software improvement": better data, RL environments, distillation, and similar all count and can even be the central driver — not just classical algorithmic efficiency gains.
Related event: Debate over RSI: real algorithmic progress or compute-driven repackaging(4 posts)→
More from AGI Musings
- Loving 'humanity' while ignoring humans: tech's forests-vs-trees problem — AccBalanced · 2026-11-16
- VraserX: AI replicating lifelong skills will spark genuine anger, not fear — VraserX · 2026-11-16
- Cohere CEO slams effective altruism's doomsday narrative as 'control' — suchenzang · 2026-11-16
- Economist on a world without work: relationships, not jobs, will define us — paulnovosad · 2026-11-16
- Palmer Luckey: a 16-year-old with a laptop will soon outproduce an entire 2015 company — Yamapama · 2026-11-16
- Ethan Mollick's Fable 5.2 test: 1,000-agent OS in 8 days, and an AI bureaucracy of 87,500 words — 新智元 · 2026-11-16