Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I tried it on a small Django app and was not impressed in the end.

It looks like it’s doing a lot, and at first I was very impressed, but after a while I realized that when it ran into a problem it kept on trying nonworking strategies even though it had tried them before and I had added to claude.md instructions to keep track of strategies and not reuse failing ones.

It was able to make a little progress, but not get to the end of the task, and some of its suggestions were completely insane. At one point there was a database issue and it suggested switching to an entirely different database than the one that was already used by the app, which was working and production.

$12 spent in a couple of hours later, it had created 1200 lines of partially working code and rather of a mess. I ended up throwing away all the changes and going back to using the web UI.



Now take your $12 and multiply it by 100k people or more trying it.

Even if you won’t use it again, that’s booked revenue for the next fundraise!


That's revenue, not profit. GPUs aren't cheap.


I use it like a brush for new apps and a scalpel for existing ones and it generally works well. If it can't solve something after 3 attempts though I just do it.


LLMs seem to work a lot better with statically typed languages where the compiler can give feedback.


LLMs are what’s finally going to make Haskell popular!


And Haskell is going to make LLMs popular! https://groq.com/




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: