Claim◆Article
The study found no evidence of 'pelicanmaxxing' — AI labs are not deliberately biasing models toward better pelican-on-bicycle drawings.
Across all models tested, the study found no evidence that labs are deliberately improving pelican-on-bicycle generation quality. ✦ AI generated
the author (via the article's title/description, the voice is an anonymous blogger or commentator) · Simon Willison's Weblog · 2026-07-22 · original ↗
For the models he tested he could find no evidence of pelimaxxing: The pelicans on bicycles don't look any better. Labs are not better at drawing pelicans. Labs are not better at drawing bicycles. Labs are not better at drawing pelicans on bicycles, even adjusting for difficulty.
Read full article ↗excerpt · fair-use quotation
- ·Across all tested models, no pelimaxxing detected
- ·Pelicans on bicycles do not look any better
- ·Labs are not deliberately improving generation quality
- ·Adjusting for difficulty, no improvement found
- ·Labs not better at drawing pelicans
- ·Labs not better at drawing bicycles
- ·Labs not better at drawing pelicans on bicycles
Around this claim
Evidence · 2
Dylan Castillo conducted a rigorous study of whether AI labs have been deliberately training models to draw pelicans riding bicycles.the author (via the article's title/description, the voice is an anonymous blogger or commentator) · Simon Willison's Weblog · conf 95%Dylan's methodology tested 8 animals × 6 vehicles across 7 different models with repeated runs and automated evaluation.the author (via the article's title/description, the voice is an anonymous blogger or commentator) · Simon Willison's Weblog · conf 90%