38 d ago
Unsafe actions halt real-robot trials of GPT-6 Astra as a robot policy
Wenbo Zhang and colleagues ran GPT-6 Astra as a zero-shot robot manipulation policy on the 42 simulated RoboDojo tasks. It ranked above all 40 public policies on that board. On real arms, the team halted testing after the model repeatedly issued unsafe actions. The paper says some incidents damaged hardware and nobody was hurt.