How could engineering with AI look like?
Risks of oversized Rails codebases, AI dependency, lock-in, and the high cost of expert intervention and maintenance
TRANSCRIPT
So you have this inhuman artifact, right? The 300,000 line blog. And you also now have to rely on an inhuman tool to do anything to it, right? So you're like, first of all, you could be screwed because the tool can't do it. As you were saying, it just spins and doesn't solve a very simple thing. You could just be screwed and you're locked in,
... See moreFor our code reviews or AI policy: Produce human legible artifacts. Don’t bet the company on AI being always acessible.
Simon Willison's Weblog
In retrospect, it made sense. Agents write units of changes that look good in isolation. They are consistent with themselves and your prompt.
Mo • After two years of vibecoding, I'm back to writing by hand
Very good comment on this article: Just telling an agent to do a big thing does not work. Small iterations are still better than large change. Iterate together.
Refactoring AI setup workflows by separating markdown from code and logging script output to preserve clean context
TRANSCRIPT
And when the step, when that script finishes, it just outputs three or four lines saying success, failure, this is the environment we're on. These things have been configured. These things haven't been. So just pure, clean context for Claude.
And when Claude needs to, it goes and looks at the log file to see the full error, to see the full report.
... See moreWhy CI is even more important for programming with AI