[philipp-gayret]

I very often hear from developers at clients I work with that code they (not me) generate with AI is not of enough "quality".

So I ask them what quality means. So far, I only get the most basic feedback: it should be in X style, pass Y linter, have N% coverage, have documentation...

At the same time, most, if not all manually written repositories do not pass the newfound quality metrics that must apply to AI code to be quality. I'm glad people are thinking about it at least, but let's not pretend like we cared before when it took manual labour. I'm even more glad we are in an age where quality standards can be fully automated.

[OtherShrezzing]

>So I ask them what quality means. So far, I only get the most basic feedback: it should be in X style, pass Y linter, have N% coverage, have documentation... I'm glad people are thinking about it at least, but let's not pretend like we cared before when it took manual labour

I don't think I've ever worked anywhere that'd accept a PR which introduced untested & undocumented features, in which the code doesn't even pass the company's internal linter.

Those feel like very low bars for your colleagues & clients to set.

[hannasanarion]

> At the same time, most, if not all manually written repositories do not pass the newfound quality metrics that must apply to AI code to be quality.

That's because I trust the code that I write to have quality, because I wrote it, which means I understand it. I may choose not to test something because I am certain that nothing can go wrong with it.

When your repository is thousands of lines of code written by an AI with tendencies to forget critical components, duplicate work, make bonkers editing errors that shuffles everything around to all the wrong places, and invent packages out of thin air, you need a system of accountability.

[k__]

" I only get the most basic feedback: it should be in X style, pass Y linter, have N% coverage, have documentation"

Those are things that AIs can check by themselves.

What AIs are lacking is common sense.

They have no problem to inline everything they do which makes the codebase unmaintainable for humans

If you tell them to refactor, you get useless abstractions, like functions that get called in random places with no sense of structure.

[StilesCrisis]

Most manually written repositories are hobby projects where 0% test coverage is fine because it doesn’t matter.