Hacker News new | ask | show | jobs
by gtirloni 1205 days ago
This look like someone got the output of $linter and asked ChatGPT to write a happy explanation in plain English. Is there more to it?
3 comments

Agreed. I was expecting the AI to uncover and offer more nuanced suggestions. The examples on the site were highly local issues, with most of them being one-liners our linters are already catching. I was excited to see what the AI can do when it has the context of the entire repo, but I was not wowed.
I wouldn't discard the capabilities of the LLM. I've been using it to do relational reasoning of Java comparators and it is surprisingly accurate. https://twitter.com/marceloabsousa/status/163289383288582963...
Apparently not. And for being its sole value-add, I find the the fake-happy chattiness to be quite grating in fact.
The examples on the site are meant to showcase the features and chatGPT capabilities being where it shines. Depending on the project's context the bot will make great reviews, chatGPT already does a good job at reviewing code if given enough context and that's what the bot is trying to achieve. In the future the bot can be onboarded to a codebase and the overall project for better review and possible code contribution.