Dev Blogs
A model can fail on features released before its knowledge cutoff, then succeed on ones released after it. We tested hundreds of product changes and found that the date tells you far less than the work does.
Your AI coding agent passed the eval. But did the model know the answer, or did it find it somewhere on your machine? A correct answer can still invalidate your measurement.
The smartest model can cost five times more and deliver the same result, or even a worse one. See how evaluating your own work helps you get more value from your agent budget.
Most changes you think will improve AI agent behavior won't. We tested a dozen hypotheses on a real project upgrade scenario and the majority failed. Learn how to emulate documentation, API, and MCP server changes locally so you can validate what works before shipping anything to production.
Your agent skill calls an API. The moment you start evaluating it, every run either costs money or mutates production data. Learn how to mock APIs transparently so you can run evals without changing your skill or hitting real endpoints.
This is the eighth and final article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better outcomes. You've read seven a...
This is the seventh article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better outcomes. You build an eval. You run i...
There's advice making the rounds: replace your CLI args with a single payload so agents can use your tool more effectively. The thinking being, that agents already think in structured formats, and nested data maps cleanly to JSON. Flat args on the other hand, force awkward conventions like repeating to delimit multi-value groups, which is inheren...
A new model drops with lower per-token pricing and better benchmarks. You switch. A week later someone asks why the agent is burning 12x more tokens on the same task while producing worse output. We ran 150 agent tasks across 15 scenarios on two models, Claude Sonnet 4.6 and Claude Sonnet 5, using GitHub Copilot Chat in VS Code on Windows. The sce...