Adopted · Manual
AI Coding
Coding assistants are a daily tool. The Lab uses them; it has not published one.
Why this is here
Coding agents (in-editor assistants, review, scaffolding) are the most adopted use case right now. The Lab uses them to ship evaluable systems.
What I am watching
Where the tool speeds up an evaluable system, versus where it produces plausible unread code. Same rule as for agents: visible failure, human review, no silent side effects.
Current read
Adopted in the craft (mainstream tools, daily use). No Lab “coding agent” repo. No diff-acceptance metric. Evaluation of assisted code is still rarer than the usage.