Bare coding agent hits 78% on R2R-CE navigation with zero training, no mapping or memory
jiqizhixin · x · 2026-09-06
University of Adelaide's MIP (Minimal-Interface zero-shot agents) drops a general-purpose coding agent straight into a physical environment—no mapping module, no memory, no waypoint predictor, no domain-specific code or navigation training. The agent reads raw observations and outputs raw actions via a minimal interface, relying purely on general reasoning.
Result: 78% success rate on the classic vision-and-language navigation benchmark R2R-CE, matching or beating recently released industrial-scale navigation models trained on massive specialized data (humans: 94%). The finding raises the question of whether industrial navigation models have been solving the wrong problem.
More from Embodied
- Underwater drone crab fishing: a Bilibili vlogger's wild drone skill goes viral — seanmcdonaldxyz · 2026-09-06
- Nightly training runs as pseudo continual learning: still cumbersome, costly, and power hungry — PTrubey · 2026-09-06
- MathWorks offers free 17-video robotics course: encoders to wheel-legged bots — lukas_m_ziegler · 2026-09-06
- Tesla Cybercab impresses first riders: "better than I expected" — EricETesla · 2026-09-06
- Tesla confirms you can buy a car that runs like a Cybercab while you supervise — yunta_tsai · 2026-09-06
- Could AI AR glasses make 'learn a trade' bad career advice? — Admirable_Zombie5245 · 2026-09-06