Robot Learns a Task From a Single Demo With Zero Fine-Tuning
E0M · x · 2026-08-20
A member of the Genesis team (@felixwyw) recounts that late on Aug 3, while experimenting with few-gradient learning on GEN-1.5, they improvised a test of "physical prompting": show the robot a task once, and it imitates the demo exactly with zero fine-tuning. More prompts for different tasks also landed successful rollouts. The author credits years of team work and community groundwork for the breakthrough. Reposter @benbarry recalls a 2018 conversation with Ilya Sutskever: AI progress can go from feeling nearly impossible to just working overnight — and he feels it again now.
Related event: Genesis robot learns from a single demo with zero fine-tuning(2 posts)→
More from Embodied
- Ditto-Bench stress-tests GPT-6 Astra for robot control: simple tasks shine, physics stalls — anand_bhattad · 2026-10-09
- NVIDIA shows frontier AI agents building Omniverse simulations from natural language — nordicinst · 2026-10-09
- NVIDIA Shows Six Projects Building Omniverse Simulations With Frontier AI Agents Like Astra — NVIDIA Blog · 2026-10-09
- Tesla Vision-Only Robotaxi Handles Nighttime Downpours, Silencing Earlier Skeptics — elonmusk · 2026-10-09
- Hark Sees Explosive 48-Hour Growth, Brett Adcock Says Team Scrambling to Fix Infra — adcock_brett · 2026-10-09
- aimotive driving dataset: train labels come from a tracker that sees the future — RexDouglass · 2026-10-09