Hackathon project PMO gives AI agents permanent memory, cutting tokens by half
AIandDesign · x · 2026-11-05
A Meta Hackathon entry, Muse Dashboard + Permanent Memory Optimizer (PMO), is now open with full methodology.
- Core idea: agents pay for memory expensively by re-reading everything every step; PMO gives Muse Code project-level permanent memory — the model never forgets and picks up where it left off.
- Claimed result: same work and quality with less than half the tokens, no large cache or oversized models needed.
- Four testable skills: CHAT (everyday chat, point at a project folder), CODING (auto-reads files, makes changes, runs tests until green), REASONING (triggered by "research" + "debug", sources over guessing), and 3D modeling (attach front/side/back reference images for a game-ready, textured, rigged model).
- GitHub repo and demo page available.
More from coding & agent
- Tencent Hunyuan Hy4 lands in OpenClaw with vision, 1M-token context, agent gains — TencentHunyuan · 2026-11-05
- BenchFlow at Bench Summit: we're not ambitious enough with how we build evals — MeganRisdal · 2026-11-05
- kimo: a 15MB CPU-only open-source video editor that AI agents drive entirely over MCP — LieTop7965 · 2026-11-05
- AI UX auditor passes 4 tests then invents problems: 20/27 findings accurate, 2 false positives — noya1979 · 2026-11-05
- Anthropic merges 30,000 code changes a day: Claude reviews first, humans check evidence — dotey · 2026-11-05
- How one creator built a 113-second AI dialogue short with Seedance 2.5 and ElevenLabs — DavidmComfort · 2026-11-05