What the data actually shows

We analyze thousands of AI coding sessions every month. Here's what's working, and what's not.

The State of AI Coding Tools in 2026

60% of teams have adopted Claude, but most still switch to ChatGPT for certain tasks. We dug into the data to find out why.

Forget surveys and gut feelings. Here's a framework for tracking real productivity gains from AI tools.

The technical architecture behind privacy-first analytics, and why it matters for engineer trust.

We analyzed 10,000 sessions and found the patterns that get better results with fewer tokens.

How a 20-person team went from "AI is probably helping" to data-driven AI adoption.

Most teams waste 30% of their token budget. Here's where the money actually goes.

What's working in AI adoption, delivered every Monday. Unsubscribe anytime.