I have a confession.
I dislike reviewing AI-generated pull requests pushed by non-tech people.
Many years I've been mentoring developers in our team, with code reviews being the most constant and stable feedback loop so far.
In the last few weeks we've been experimenting with non-tech team members using Claude to make changes to the application. For small fixes it's nice - an absolute no-brainer. But for bigger changesets - honestly - I hate it.
Leaving remarks and improvements suggestions in a pull request would mean I'd be just talking to Claude through a human proxy. Slower, less efficient, with no developer to mentor, no actual learning needed or asked for.
What's the point then? I just feel I'm wasting my and the other person's time. It's faster to push fixes directly, skipping Github comments.
To keep it healthy and sustainable: behind every (non-trivial) pull request, disregarding if it's AI-generated or not, should always be an engineer eager to learn, willing to become a better professional.
Only then it makes sense to me.
