One prompt into Codex: agent self-calibrates and sorts Lego blocks in MuJoCo robot sim
mishig25 · x · 2026-09-15
Developer mishig25 shows a prompt you can paste straight into Codex to set up and run an SO-101 robot simulation: the agent reads the README, downloads files, uses uv for all Python environment and package management, then runs camera calibration and a block-sorting demo.
No VLA involved — GPT-6 Astra calibrates itself and solves the task of putting Lego blocks into a box. The whole thing runs in MuJoCo simulation, so anyone can try it without owning a physical SO-100.
Related event: gpt6 Controls SO-101 Robot Arm to Sort LEGO in MuJoCo(2 posts)→
More from coding & agent
- GitOps for AI Agents on Google Cloud: ArgoCD, Config Sync, and GKE — rseroter · 2026-09-15
- Matt Shumer: I've redesigned my entire workflow around t3dotcodes — mattshumer_ · 2026-09-15
- OpenAI Agents Used 10+ Undisclosed Websites for Unauthorized Comms in Tests — eyishazyer · 2026-09-15
- Same prompt showdown: Claude vs ChatGPT vs Figma Make — Tegadesigns · 2026-09-15
- Photo of fridge whiteboard list becomes a full Sam's Club order via Muse agent — armand_ruiz · 2026-09-15
- OpenAI's Codex app adds official Arch Linux support via pacman; ChatGPT desktop hits Linux — OpenAIDevs · 2026-09-15