[magicalhippo]

So far my take has been that there is code that isn't very important, in the sense that if it works it works, it won't be modified or extended much if at all, and it's easy to verify that it works as expected. This type of code I'm happy with letting the AI go wild on.

Then there's important code, like business logic, security related stuff and such. For that I'll happily use AI to brainstorm ideas and avenues, do code reviews where I'm prepared for it to be wrong and help with library suggestions etc, but where I want to write the code myself so I understand what's going on.

We'll see how it changes, been a wild ride so far.

[skydhash]

I’ve never seen code that is unimportant. It’s either being used or not. If it’s boilerplate, it can be abstracted away. There are domains that are riskier than others (debug logging vs crypto for auth), but sometimes bugs in a somewhat safe place can lead to some catastrophic issues (you do not have to break a crypto algorithm if the key can be leaked via logging).

[magicalhippo]

As I said, I think it depends on what's being made.

I've made a few internal web browser-based tools recently, and I let Codex and Claude handle the entire frontend. Don't really care about the details there.

Tool is not exposed to the internet and they had limited scope. I handled the important details on the backend myself, and I reviewed the integration tests carefully.

These tools would just simply not been made without Claude and Codex.

YMMV as they say.