Google WikiSkill: Small Models Outperform Larger Ones via Skill Evolution
mrdrozdov · x · 2026-08-29
Google published the WikiSkill paper, proposing a new architecture for agent skill evolution systems. It separates three components often collapsed together: raw execution traces, a persistent wiki of accumulated knowledge, and executable skills themselves.
- Core Mechanism: Experience is consolidated into the wiki, and every subsequent skill update builds on this wiki instead of on a scattered optimization history.
- Results: Ablations confirm the wiki is crucial for performance gains. Smaller models with evolved skills beat substantially larger models without them, and skills transfer across model families.
More from coding & agent
- How to hook Claude Code or Codex up to iMessage in 3 steps — jacob_posel · 2026-10-16
- Are the fastest GPU kernels now written by AI? Kernels are hyper-verifiable, argues one dev — fleetwood___ · 2026-10-16
- Anthropic lead engineer's 4-question prompt makes agents more ambitious, now a free skill — daniel_mac8 · 2026-10-16
- ChainForge dev shows 'feature map' for orchestrating subagents on large PRs — IanArawjo · 2026-10-16
- aiDotEngineer benchmarks agent orchestration pipelines, staging humans as Foreman and workers — vikvang1 · 2026-10-16
- Developer mass-produces 200 videos in hours using Claude Code + Hyperframes — mcraddock · 2026-10-16