Researcher Doubts LLMs Can Solve Robotic Dexterity, Challenges GPT-7 on Oil Pan Install
adam_karvonen · x · 2026-10-24
Robotics researcher adamkarvonen remains skeptical that LLMs alone can solve robotics: they manage basic pick-and-place tasks but fail at using fingered robot hands or anything requiring dexterity. He'd like to see GPT-7 manipulate large, awkward objects like installing a car oil pan. Citing a new paper, The Bitter Lesson Comes to Robotics, he notes LLMs' score doubling time on a robotics benchmark was 53.9 days (July–Oct 2026), while speed rose 4.4x and costs fell 6.2x.
More from Embodied
- Robotics model leaderboards are premature, argues practitioner: task-specific evals first — paigeinsf · 2026-10-24
- Claude designs a 200 lb-payload hexapod robot in one hour, with parts list — _Stocko_ · 2026-10-24
- iPhone Duo hands-on: clunky as a phone, but unfolds into a perfect iPad Mini — firstadopter · 2026-10-24
- Rice & NVIDIA's Sensorimotor Alignment predicts robot policy success without training (r=0.717) — CSProfKGD · 2026-10-24
- Generative Trajectory Planning: robots learn to plan with world models instead of RL — du_yilun · 2026-10-24
- Hugging Face adds OpenUSD support as NVIDIA drops 2,500+ free SimReady 3D assets — Promptmethus · 2026-10-24