This, I totally agree with you having a call and letting them explain the changes. One of my coworker used claude to write and push PR's and the message was so cryptic, it's not really understandable. I got him on a call and let him explain the PR. He also actually couldn't really interpret the PR message. I warned him not do this again. Now he writes by hand. I think communication between humans should be written by only humans. AI is not really good at communication.

That here.

I once worked in a place where a guy would pick on me in code reviews, he wasn't really higher position, but had been with the company longer than I was.

He was just an asshole asserting his dominance and needed to be taken a notch down by a manager and it turned out he didn't understood (or more likely didn't try to understand) the PR when confronted.

I think that’s not an unreasonable expectation. Don’t share raw LLM output with other people.

No. I’m expecting everyone to answer direct questions that they are the only one who can answer by thinking up the answer with their own brain then writing or speaking that.

It’s just a fancy version of previous tools. Imagine a coworker instead does a refactor by grepping the entire codebase and doing a find-and-replace.

Do they ensure that it didn’t accidentally overwrite something it shouldn’t have?

Or do they just throw up a PR and assume their little refactoring bash command works fine, then expect their coworkers to debug it for them?

Firing somebody for it is ridiculous? If you say that, then you are also obliged as a customer to keep giving your money to companies who do not perform as expected. You have no right to choose a different lawyer if your lawyer makes an AI-generated report for your case. You have no right to walk past a restaurant which is using AI-generated billboards, you shall eat there.

Fish rot from the head, as they say.

My experience is it’s over eager engineers who we would have historically considered not strong who suddenly feel able to contribute in a different way

I doubt higher ups wanna babysit agent software engineers all day long, it's not fun and its pretty time consuming and cognitively demanding. They'll let humans handle it - but their dream is way less humans will be needed than before.

Replace claude code with your cousin or another employee as a thought exercise.

So write to a colleague asking to finish what your cousin/another employee thinks needs to be done. I think most people would say describe the problem clearly so that I can solve it and be responsible for the fix or let your cousin/other employee finish the job and own the solution.

This is offloading of accountability while denying credit.

Yes, just file a ticket and let the people with relevant expertise and context handle it - obviously. Do software engineers ask Claude for recommendations on handling network outages and bug IT with “AI” slop? Do accountants bring “AI” slop advice to electricians wiring a building? Do lawyers use Claude to try helping physicians diagnose illnesses?

By providing a bad “AI” slop solution that you can’t evaluate, you’re actually creating more work for the people responsible for doing the thing. Now they have to understand the slop you’ve given them, explain why it’s wrong/won’t fit in the existing architecture/has other issue you can’t see without deep context. Or, they have to play politics to ignore your slop without angering you and creating more problems down the line. Either way, it’s harder than just letting them do their job.

Without understanding they have no way of judging the competence of Claude and the slop it generates. So yes they should just file a ticket the old way, after doing the actual hard work, which is reproducing the bug.

[flagged]