Is your name Claude?

I mean, the thing with LLMs is that they tend to overdo it. They test completely unnecessary details. Your idea of having tests for all features is probably still way less tests than the typical LLM-generated commit...

If you've got 100% coverage for every feature, yeah, people are going to think you're a bot.

Look if someone thinks you're Claude because you bothered to test your code that's kind of sucky. But then again the fuck do I know, I don't contribute or maintain anything so that's just an off the cuff take.

At some point we'll have to acknowledge that AI doesn't mean inherently awful and bad IF it's used responsibly and in a helpful manner. I don't think LLMs are intrinsically bad but the way these companies have been marketing them and trying to make them some technological panacea that'll replace all the people and fix everything is utterly deranged and divorced from reality.

At the end of the day they're large language models, their job is to come up with what the model "thinks" a person would write as an answer to whatever the prompt is. They're not magic and I'm tired of people anthropomorphizing the shit out of LLM / "AI" and acting like AI output is being delivered by the machine with some sort of logic and knowledge and oversight. It is not.

Nice try bot. The Terminators are already trying to blend in.

From what I've seen and heard, AI seems to do well for test-driven development. That doesn't necessarily translate into good code, but it does translate into tested code.

At the very least it can translate the documentation of an API into the appropriate tests to verify the documentation is correct, and I'm being quite humbled by that.

One problem I could see with the aggressive testing (which also happens in human-written codebases) is that it becomes hard to tell essential behaviour apart from details. If you nail everything down with a unit/integration test, nothing can be changed anymore.

I could imagine that this is part of the reason why LLMs love to duplicate code rather than adjusting existing code.

The important part is that the tests track regressions and they're helpful for that. After that, it's okay if tests are mostly write-only.

I've noticed that AI commentary tends to describe what the code is doing verbatim which is better for naming tests and functions rather than writing code comments, docstrings, or change notes. I usually have to fix this public-facing documentation to not talk about library internals which no user of the library would ever care about.

So far I feel that AI assists are good when you have the time and effort to push back on them, but that isn't an option for all projects, it works well for my own projects which have very few contributors in the first place. For the most part I can treat them with the same scrutiny as human written code. When the person submitting the code is obviously not checking their own code then that's a problem.