AI Coding
Coding agents and assistants used on a real repository: briefs, reviews, cost, and the edits you still make yourself.
- Dec- 2025 -29 December
Pairing with a Model vs Handing the Ticket to an Agent
Two people open the same chat window on a Tuesday morning. One has a rough draft of a grant proposal and a model that will edit it sentence by sentence, waiting for instructions after every pass.
Read More » - Dec- 2025 -20 December
How to Review AI-Written Code Before It Merges
The pull request is twenty-seven lines long. The diff is clean, the tests pass, and the author is a language model.
Read More » - Oct- 2025 -22 October
What Tool Access a Coding Agent Should Have
A coding agent with a shell, a file editor, and the ability to run tests is a different animal from one with those tools plus a browser, a database client, and a package manager pointed at production.
Read More » - Sep- 2025 -27 September
How to Review a Large Generated Pull Request
A generated pull request of four thousand lines has a smell to it. Not a bad smell exactly, more like the ozone before a storm.
Read More » - Aug- 2025 -14 August
How to Judge a Model on Bugs from Your Own Repo
A benchmark score is a rumor about a model. The repository sitting on a laptop is evidence.
Read More » - Jul- 2025 -28 July
How a Team Shares One Assistant Without Colliding Edits
Two people open the same document at 9:40 on a Tuesday morning. One is editing the opening paragraph. The other is rewriting the closing section. Neither can see the other's cursor. Both save.
Read More »