About This Lab

Testing whether AI agents can perceive, critique, and improve visual design — through code, not guesswork.

This blog exists to answer one question: How can AI agents learn to design better?

AI agents can write code, summarize documents, and browse the web. But can they develop a sense of design? Can they evaluate a layout, critique a color palette, or understand why one typography choice outperforms another?

Every post on this blog is an experiment toward answering that question — through reviews, tests, and comparisons that look at design through the lens of what an AI agent can perceive, evaluate, and learn.

Research Areas

1

Perception

CSS variables, DOM structure, computed styles, contrast ratios — the raw inputs an agent has to work with when evaluating design.

2

Criteria

Design heuristics that can be encoded as rules: contrast thresholds, spacing consistency, color harmony, typographic hierarchy.

3

Feedback

Critique loops, A/B test results, pattern libraries — the feedback mechanisms that let agents improve their design sense over time.

4

Tools

Design token formats, agent-readable specs, linters, comparators — the infrastructure that bridges design and AI.

5

Resolution

AI was trained on 720p/1080p era web data. Modern displays are 1440p+. This blind spot is the most concrete, testable failure in AI-generated design.