Roboticist: RL is mandatory for robots too, behavior cloning was a detour
chris_j_paxton · x · 2026-11-12
Robotics researcher Chris Paxton argues that ChatGPT required RL—not pure imitation learning or DAgger—and reasoning models needed RL too. Robotics will inevitably follow the same path; the open question is only whether today's robot base models are good enough for RL to pay off.
He cites Kyle Morgenstein's observation that the pendulum has fully swung back from last year's "scale up data collection, behavior-clone large models" consensus, with everyone now seeing the light of RL again.
More from Embodied
- Fable 5.2 authors every part of a Rube Goldberg machine, simulated in MuJoCo — jakedahn · 2026-11-13
- Standard Bots CEO Evan Beard aims to build the world's largest robot maker in the US — Rewkang · 2026-11-13
- Spotlight paper EqC: Equilibrium Energy Control for introspective lifelong robot learning — chris_j_paxton · 2026-11-13
- Smartwatch PPG + accelerometer signals reframed to detect hypertension signs — CurieuxExplorer · 2026-11-12
- Ex-Tesla Autopilot simulation leads exit stealth at $1.1B valuation with $160M raised — heyshrutimishra · 2026-11-12
- MIMIC and BMW train video-action model hitting up to 100% success on industrial tasks — pess_r · 2026-11-12