Hacker News Reader

Comment by pianopatrick | Hacker News Reader

I'd like to see the results for Ada on these same measures. On the theory that the Ada type system covers more classes of errors than other languages, and so AI can self correct better.

View on HN

Unfortunately for static type weenies like me (and you, presumably), types don't seem to matter at all, or Python and Javascript wouldn't be on top. There's no reason to believe that Ada's type system is so unique that it alone can help AI self-correct, and Rust, Haskell, ML, Typescript, etc. can't.

Well the reason I'm interested in Ada is because I saw a study that showed AI did worse at functional programming. So that might explain the problems with Haskell et al. But Ada has a strong type system while still having procedural code. So it would be an interesting comparison with Haskell etc. if the problem was that Haskell is functional or if the problem was that these are not so popular.

The study didn't really show that. What it did show, is that in this particular study there was no strong effect either way and there is really no way to generalize from the results. See in particular the details on what made the agent stumble, it was things like "cargo repeatedly gets invoked with the wrong arguments" - nothing to do with functional or static, just ecosystem idiosyncrasies and trivial differences.

Well, it did show you probably shouldn't use assembly, but that's about all it showed very strongly.

And of course it also showed very strongly that you should not base any choice-of-language decisions on single studies.

You're in luck! https://news.ycombinator.com/item?id=49254706

Thanks

Reply on news.ycombinator.com