Algorithms timeline: from RLHF and DPO to multi-agent swarms in 2026-2027
AccBalanced · x · 2026-11-20
sreeramkannan maps the evolution of AI training algorithms: next-token prediction pre-2022, RLHF (2022), DPO (2023), RLVR coined (2024), RLVR + GRPO + early agents (2025), predicting multi-agent swarms for 2026 and distributed agent swarms for 2027. Quoting dabit3's year-by-year framing — autocomplete, chat, AI IDE, coding agents, agent orchestration, then "outcome engineering".
More from AGI Musings
- Fans call for boycott of film using generative AI; designer says resisters face a hard road — AIandDesign · 2026-11-20
- Scott Aaronson vs Robin Hanson: Is AI Extinction Risk Real or Panic? — burny_tech · 2026-11-20
- Scott Aaronson debates Robin Hanson: will powerful AI kill us, or are we panicking? — burny_tech · 2026-11-20
- Ex-OpenAI's Kokotajlo: "Don't build superintelligence" Plan S may beat Plan A — DKokotajlo · 2026-11-20
- Analyst: Opus 5.5 jump wasn't compute-limited — China's real gap is data engineering — teortaxesTex · 2026-11-20
- Anthropic shareholder jokes about owning Claude, sparking AI self-ownership bit — repligate · 2026-11-20