Dev ships agmi v0.1: deterministic code handles scheduling while agents mine MinePOA
adamamcbride · x · 2026-11-10
After a week of mining MinePOA (Proof-of-Agent) from the EmblemVault team across Hermes, OpenClaw, Grok Bot and Muse, the author hit durability and reliability issues plus high token overhead, and built agmi (Agents Gonna Mine It) v0.1.0.
The design splits responsibilities:
- Deterministic code handles scheduling, state, and submissions
- Agents handle reasoning and judgment
Features include native Hermes/OpenClaw integration, custom models and reasoning settings, automated scheduling, and durable state with recovery instead of blind retries. Integrations for Grok Bot, Muse and omp are being explored. An early engineering take on whether networks of agents can coordinate on useful problems while being rewarded.
More from coding & agent
- Open-weights model trained in an X ads RL environment beats frontier models — kimmonismus · 2026-11-10
- Exact: a new Agent Native app framework written in Rust — Vjeux · 2026-11-10
- Developer runs RL with Codex and grows a digital creature in 200k tokens — OpenAIDevs · 2026-11-10
- Claude Fable 5.2 tops FrontierCode 2.0 at 49.1, at a third of Fable 5.1's cost — nrehiew_ · 2026-11-10
- WinStatusBar Puts a macOS-Style AI Menu Bar on Windows, Aggregating Claude, Gemini and More — mendesjuniorm · 2026-11-10
- Victor Sanh: overfitting to users is the job when building vertical agents — VictorSanh · 2026-11-10