LLM predicts ambulance ETA within 0.2s but still loses to a simple queue rule in traffic sim
ViceGodIm · reddit · 2026-11-06
A developer tested an LLM as a traffic signal controller in a Godot simulation, logging its ETA forecasts before advancing the sim and scoring actions against a baseline rule.
At sim time 65s the model predicted the ambulance would arrive at 100s; at 95s it revised to 112s — actual arrival was 112.2s, a remarkably accurate late estimate. But a basic queue rule got the ambulance through in 107.6s under identical arrivals, routes and safety constraints, beating the LLM controller by 4.6s.
The takeaway: for planning ahead, the early forecast is the metric that matters, not the post-hoc correction. A recorded run is available on YouTube.
More from Embodied
- Lamport Robotics' Pincer1: a Six-Legged Robot Wallet That Signs Only When You're in the Room — R4nsack · 2026-11-06
- UCLA's FlowPilot uses human preferences to cut sidewalk robot interventions by 40% — zhoubolei · 2026-11-06
- RoboColosseum tests 5 VLA policies across 4 embodiments; environment diversity is the new bottleneck — DJiafei · 2026-11-06
- RoboColosseum: 20,000+ runs, 100+ robots, rankings flip by task — DJiafei · 2026-11-06
- Using Dance Choreography Tool Swing Choreo to Design Robot Motions — andrew_n_carr · 2026-11-06
- Cartwheel lands on founder-generated Emergent startup list for kinematic world models — andrew_n_carr · 2026-11-06