S1 robot shows physical prompt steerability, infers novel OOD behaviors at inference time
deepakpathak · x · 2026-09-10
- The S1 robot exhibits "physical prompt steerability": in the same kitchen, four different video prompts induce four distinct behaviors, including a far out-of-distribution task of putting a plate into a toaster.
- The key finding: scaling pretraining makes S1 better at acquiring entirely new physical behaviors at inference time — not just a larger skill repertoire.
- S1 can infer and execute novel OOD behaviors from context, without weight updates or task-specific post-training.
- If the trend continues, robotics starts to look like promptable general-purpose control.
Related event: Skild's robot foundation model S1: one .mp4 is the robot's .exe(7 posts)→
More from Embodied
- Skyfall AI's SKY-001 world model turns sketches into CAD builds in 3.9s, vs 199s for GPT-6 Astra — rohanpaul_ai · 2026-10-27
- Tesla FSD approved in Latvia, the 10th European country; Musk praises smaller nations for moving fastest — elonmusk · 2026-10-27
- WanGP v18.81 adds Tencent's Hy-Nomas world model, playable on 8GB VRAM — cocktailpeanut · 2026-10-27
- NVIDIA open-sources FLORA, a robotics data engine ~125x cheaper than GPT-6 Astra — nvidia · 2026-10-27
- Embodied AI startups need a different operational discipline, investor says — ___Mufasaa · 2026-10-27
- π0.5 study: raising demos from 10 to 438 lifts Place Ball in Bowl success from 7% to 83% — stingning · 2026-10-27