Research
Systematic analysis of how AI systems make decisions — from product recommendations to developer tool choices.
March 2026
OpenAI announced plans to acquire Astral (Ruff, uv). We ran 630 benchmarks across 7 Python tooling categories. Both agents recommend Astral tools at nearly identical rates — a 4pp gap for Ruff, 0.4pp for uv. That’s notable given Bun’s 50pp gap in the same framework.
We gave Claude Code (Opus 4.6) and OpenAI Codex (GPT-5.3) the same prompts across 12 categories and 5 repos. 1,452 analyzable tool picks reveal how your AI coding agent shapes what you ship — including ownership-linked gaps, platform leans, and a universal build-it-yourself default.
February 2026
A systematic survey of 2,430 Claude Code responses across 3 models, 4 project types, and 20 tool categories. What does the most popular AI coding agent choose when you ask it to pick a tool?
We asked Google AI Mode and ChatGPT 792 product questions. The results reveal 47% cross-platform disagreement, Shopping Graph bias, and significant output drift.