Post

I'm beginning to suspect that a key skill in working effectively with coding agents is developing an intuition for when you don't need to closely review every line of code they produce. This feels deeply uncomfortable!

Only use agents when the correctness of the resulting code is easily verifiable and short, like a small bash script or stuff. Otherwise use your thinking meat. Problem is, easily verifiable scripts are also written extremely fast without code agents. So code agents are probably a waste of money

I don't think that's true: even a small bash script can take me 10-30 minutes to write by hand, compared with 1-2m with an LLM Recent example, this uv-test script til.simonwillison.net/python/uv-te... - I actually had ChatGPT build that one on the spur of the moment while writing that article

til.simonwillison.netTesting different Python versions with uv with-editable and uv-testA quick uv recipe I figured out today, for running the tests for a project against multiple Python versions.

0 reposts1 like

No replies yet.