Back to browse
I built a blind taste test for Claude and Codex designs

I built a blind taste test for Claude and Codex designs

by rubenflamshep·Jul 25, 2026·3 points·0 comments

AI Analysis

●●SolidRabbit HoleCozy

Fun blind taste test revealing distinct aesthetic biases in AI models.

Strengths
  • Simple, effective methodology to crowdsource preference data on model output.
  • Highlights subtle stylistic differences between Claude and Codex generations.
Weaknesses
  • Limited scope; more of a curiosity than a reusable tool or dataset.
  • Results are subjective and lack quantitative design metrics.
Category
Target Audience

Developers and designers interested in generative AI capabilities

Similar To

LMSYS Chatbot Arena · Human Preference Datasets

Post Description

Hey HN! A while back I asked Codex to generate a companion site for a book I was enjoying. It ended up generating a landing page for the book that was ...actually good? I went down a bit of a rabbit hole prompting different models to generate landing pages for different books and this is the result.

The site runs you through a series of blind "taste" tests to help you figure out which model's designs you prefer. Would love any feedback you have if you try it out!

Similar Projects