When I read commit messages and comments Claude makes a large fraction of the time they don't seem terribly related to the code, nor necessarily relevant to the problem at hand. The commits don't generally make a great deal of logical, linear sense either. Reviewing Claude PRs commit-by-commit is kind of nauseating. I only have 12yrs of experience though so I must be wrong. I'll defer to you guys. Sure would be nice to know wtf everyone's so excited about though I just don't get it.
Subjective or not, there wasn't a single example of a "good" AI generated commit message or refactoring or whatever in the whole thing. If they're so good why not just show it? It's just weird.
"My new car is much better than my old car" is not something you'd reference your 27 years of car experience as evidence for--you'd just say what's good about your new car and what was bad about your old one, right?
Why do these posts always state things like this with maximum confidence and then not even make any attempt to show evidence? "I've got x years of experience trust me bro" is pretty weak.
It would be extremely valuable to be able to reliably identify this kind of behavioral tendency in an interview setting. Some kind of reliable test that screens for likelihood to go AI psychotic.
Yeah I'm not defending the blog post as a great way to build the biggest bestest community. Clearly it isn't. But it's also pretty clear that's not the poster's priority. And that is defensible! They don't owe anyone perfection in community building or any other thing.
If they want to write a maybe slightly unhinged blog post excoriating someone they had an unpleasant time with... that's a choice! It's their choice. I don't think we get to demand a different one. We can agree or disagree with what they said, but we can't demand they say something else.
If you want that kind of power over someone you need to ensnare them in a binding contract. No such contract exists here.
Does it really matter? I don't like much of the overbearing "rust culture" but I love the language and many libraries and projects are exceptionally high quality. I also like zig, people are building cool things with it. I have far less experience with it than rust but it looks interesting. I don't really get all the culture stuff, it seems like misspent energy.
Just the other day at the NATO summit he seemingly claimed Iran's nuclear program is so damaged they'll never be able to build a bomb. So, mission accomplished? Nothing the guy says is credible..
The real test is years in production. Over time your test suite grows when bugs are found and fixed, but not every bugfix necessarily gets a test, and it's very rare that a bugfix is exhaustively tested. Relying on the test suite as a directional indicator that your vibecoded rewrite functions something like the original is probably sensible. But it isn't "done" until you've run it in production for at least as long as the original. And that's where it all falls apart, because maintenance will be a nightmare. Nobody knows how the new thing works.
> Sooo.... "be more professional in your writing"?
No, just more precise. It seems absurd to brand an entire community as "assholes" based on some disagreement with a blog post by one member. I suspect great GP didn't mean it that way, but it would help to have some clarification.
I used to think like that, but I don't anymore. The most important thing about a project is the people who are working on it. Successful projects are built by small teams with deep expertise pulling together on a shared mission. If they want to use Chez Scheme, Ocaml, Haskell, or whatever else I trust their judgment.
I can recommend proptest. What you're describing is a common pattern in property-based testing which basically boils down to "comparing against an oracle". In this case, postgres would be the oracle, pgrust is the system under test, and the idea is to generate strategies comprised of sequences of valid (and invalid) SQL statements and ensure the system under test behaves the same as the oracle in every case.
uhh what..