Replicating Picbreeder via Large Vision-Language Models
July 19, 2026
Sakana AI demonstrates that large vision-language models can replicate the open-ended evolutionary search process of Picbreeder. This approach uses VLMs to drive iterative, creative exploration within complex visual spaces.
HOW THIS AFFECTS YOU
●
researcherYou can explore how VLMs drive open-endedness in evolutionary computation.
●
designerThis may lead to new generative tools for iterative, visual exploration.