One prompt: GPT-6 Astra trains a dexterous hand to spin a pen over 1.5 days
FinanceYF5 · x · 2026-09-18
A single prompt to GPT-6 Astra — train a Sharpa dexterous hand to spin a pen using Isaac Lab RL, self-create the pen's 3D mesh, deliver the trained policy and a video, with free web access — resulted in an autonomous 1.5-day run covering research, training, and visualization. The model delivered a working pen-spinning policy end to end.
Related event: GPT-6 Astra Autonomously Trains Robotic Hand to Spin a Pen(2 posts)→
More from coding & agent
- Claude Code 2.1.276 by the numbers: shipped in under 6 hours, +440 prompt tokens — ClaudeCodeLog · 2026-09-18
- Claude Code 2.1.276 fixes regression breaking all requests behind proxies — ClaudeCodeLog · 2026-09-18
- Claude Code 2.1.276 fixes 400 errors for all requests when ANTHROPIC_BASE_URL uses a proxy — ClaudeCodeLog · 2026-09-18
- Layers runs first real autonomous loop: Detail fixes bugs, Devin reviews, Claude guards deploy — saranormous · 2026-09-18
- Fine-tuned 4B model as a decision scorer with temperature-scaled confidence — Gradio · 2026-09-18
- Theo tears apart Jev's 'instant compaction': line-by-line context trimming misunderstands the fundamentals — AccBalanced · 2026-09-18