Yacine doubles down: published RL robotics research is 'basically wrong'

yacineMTB · x · 2026-09-13

Follow-up to his earlier quip: Yacine Lejeune says (半-jokingly) that everything in published RL robotics research is basically wrong and not a good way of going about things — a sweeping dismissal of the field's current methodology, continuing his criticism of LLM-concept-driven robotics papers.

Related event: Researcher Claims Most Published RL Robotics Work Is Fundamentally Flawed(2 posts)→

Original post →

More from Embodied

Embodied channel →