Corey Knowles

Thumbs up and so he comes up and you know how much is like, yeah, good point.

Codex please fix and then like.

And how how did how did you get to this point where you called the reviewer?

How do you so?

Yeah, so you can set this up in your codex settings.

Um, so, um, if you go to Codex web, um, and you go to settings, then code review, then you can turn it on super easy.

You just need to connect to GitHub.

Uh, you know, and a lot of the time, if you're making simple PRS, you'll just get the thumbs up emoji, but it's actually one of the most popular, um,

like sort of integrations that we built for codex where again codex feels like a teammate um so bringing it back to like if you want to apply codex code like deploy codex um lab engineered work like having code review is like just an excellent way to like increase the confidence like when we when we first rolled this out we were actually quite nervous that people would hate it and just be like i don't want bots telling me that my code is bad

So we trained Codex models to be great at code review.

And we trained them to have a very low false positive rate.

So basically, what that means is we were like, hey, for now, you don't need to flag every single opinion you have about the code.

Just flag when you think it's a real problem, like a serious one.

And yeah, it's turned out to be a really popular feature at OpenAI and also outside of OpenAI.

Yeah, I would say connect to GitHub and have Codex code review, look at your PRs.

You can also do this in Codex itself.

So I could pull up Codex here and I can type slash review.

And then ask for, you know, basically this will just run the same process.

I could be like, yeah, like please review, you know, the most recent commit.

Right.