“Don't just have an AI agent write your patch and send it in. When people who are reviewing the patches look at it, they see that's the case, and they're going to lose interest. … You don't want to put yourself in a position where the senior developers, as soon as they see a patch come in from you, think: oh, I don't want to review something that Claude wrote.”

…interacting with those senior developers at events? And I say I'm sad only, only in the way that I wish that wasn't a requirement. It's not a requirement, but I think it helps because then you can explain in person and get somebody on the inside to be interested in the topic that you're working on.

…to think about pushing something out there on the mailing list, subject to a lot of scrutiny and a lot of potential criticism from the get-go. And they like that kind of one-on-one opportunity to get feedback before they push it out there to a broader audience. So I thought that was pretty cool.

Yeah. Which is especially true because you need those people in order to get anything committed. So you have to be earning their trust. I mean, it's amazing what the AI agents can do these days. If we could go back in time to a year ago and then fast forward to now, we'd be shocked, right? They are…

It's very difficult. And I think that's exactly the point. It's a great tool. It is amazing what it can do. But as with all tools, you've got to use it right. And you can't just abdicate from responsibility that you are the one who's doing this. It's a tool for you. It's not an entity of its own.

So back to office hours for a second. Do you hold them? Or have you thought about it? Or are you still thinking about it? Or it's not for you?