Running Jev as a real-time robotics policy in MuJoCo, with a two-call decide-then-act split
facontidavide · x · 2026-09-19
Developer @dimentary tested Jev as a real-time robotics policy in MuJoCo. When it struggled, the author split each update into two calls: one to decide what to do next, another to decide how to move the arm and gripper. Since Jev doesn't accept images, simplified geometry and contact information were fed to the model as text. The decide-then-act split offers a useful pattern for similar VLM-based robot control experiments.
More from coding & agent
- willcb: in-context learning for personal continual learning needs neurosymbolic harnesses — willcb · 2026-09-19
- Online RL via task flywheels from live interaction traces is what actually works today — willcb · 2026-09-19
- Dev's take: Codex worth the $200 tier, Claude fine at $100, top-tier Cursor underused — lxfater · 2026-09-19
- AgentSky launches as an 'OpenRouter for agents': run 40+ coding agents in-browser and compare costs side by side — Aiden_Tech_Ai · 2026-09-19
- qwen-code SDK v0.1.13 fixes microcompaction to preserve prompt-cache reuse in long agent sessions — github-actions[bot] · 2026-09-19
- Agent engineering splits into three tiers: scripts, System One decision models, and reasoning models — arpit_bhayani · 2026-09-19