Reef Open-Sourced: An Inference Server That Continually Learns and Self-Improves Your Agents
pliang279 · x · 2026-09-02
Paul Pu Liang (CMU) and colleagues open-sourced Reef, infrastructure that turns inference into a continual learning loop: agents route model requests to Reef's inference endpoint instead of the provider's, and Reef continuously evaluates agent behavior and improves the served harness and model weights in the backend—so systems keep improving from deployment feedback instead of staying frozen after training.
- Agents can be downloaded via curl like codex or opencode
- Supports SAO, TTT-Discover, OpenClaw RL and more
- Open source on GitHub (Human-Agent-Society/reef) with recipes, tutorials, and artifact/checkpoint management
More from coding & agent
- Author Laurence Moroney: write your ugly first draft yourself, then let an LLM edit it — lmoroney · 2026-10-18
- Open-weight TYGER speech decision model makes 80ms turn-taking calls for full-duplex voice agents — solyarisoftware · 2026-10-18
- Devs: Opus 5.5 on medium beats Opus on max — eyishazyer · 2026-10-18
- He Stopped Telling Claude Code to Stop Making Up Timestamps—and Made It Impossible with a PreToolUse Hook — Grimmoner · 2026-10-18
- Cursor Super + Claude Pro + OpenCode: can this stack survive heavy autonomous coding? — DarioAmodeiWasTaken · 2026-10-18
- Coding AI wars: model is faster and smarter but 'dogshit at actual engineering' — MarcJSchmidt · 2026-10-18