Agent RAG tweak: skip lossy summaries, extract answers, fetch 10 pages — 75% on validation
colinmcnamara · x · 2026-11-21
The top-scoring run in Julian Ghadially's agent optimization loop made small changes: two-hop retrieval without the lossy summary step, extracting the answer instead of generating it, and 10 pages per search instead of 7 — reaching 75% on the validation set. Full talk and AIMUG meetup info linked.
More from coding & agent
- sprite-stacking: open-source agent skill turns 4-view pixel art into spin-able fake-3D sprites — Jasonio · 2026-11-21
- Cargo adds GitHub integration: GTM agents open PRs, humans review before anything ships — ThePeterMick · 2026-11-21
- Taking AI Seriously: Borretti traces how programmer AI advice keeps retreating — zetalyrae · 2026-11-21
- Reverse-engineering Apple Watch: a single Claude Code prompt builds a physics watch face in 30 minutes — mtizard · 2026-11-21
- us-legal-mcp: an MCP server exposing comprehensive US legislation to AI agents — modelcontextprotocol · 2026-11-21
- Maindex launches Expert Memory Service, a full-control memory API for AI agents — modelcontextprotocol · 2026-11-21