opus-5.5VIDEO
Opus 5.5 designs rewards so a duck traces a butterfly with its feet
The author had Opus 5.5 design the reward function to train a duck to trace a butterfly drawn by his daughter, running 24 training rounds and sharing a video.
Rajat Singhal
@singhal009Archived 2026.10.03Originally posted 2026.10.02
Rajat Singhal @singhal009Originally posted 2026.10.02
@redp314 Same duck, different loop. My daughter drew a butterfly, Opus 5.5 designed the rewards, and the duck traced it with its feet. 24 training runs.
x.com/singhal009/status/2105886925412606032Archived 2026.10.03 10:14