Should I get rid of juniors that generate AI slop?

Posted by Leopatto@reddit | ExperiencedDevs | View on Reddit | 432 comments

We’re dealing with a problem at my company that’s annoying, but honestly, it's not that surprising. Out of ~70 devs, about 10% are juniors. Lately I asked a few of my mids/seniors to pair program with them to see where they’re at — and what we have noticed is that a lot of them, when they hit something hard or challenging, they instantly open ChatGPT (or similar) and start writing out the problem and copying the solution; without understanding the problem or how to solve it. Now I get that AI tools are part of the workflow these days, but the point of pair programming is to learn from the more experienced dev, ask questions, think through problems together — not to just offload the thinking to an LLM. These juniors are working on client projects, and the stuff they generate is often low-quality, barely-understood code that my more experienced devs then have to clean up, which makes them annoyed - circle keeps expanding. That wastes time, burns money, and adds no real value - I'm running a business, and I do expect value in some form being generated. So now I’m stuck. Is this a hiring problem? (Our process is: HR screen > technical test [not live coding] > culture fit stage > decision.) Should I put more structure into mentorship and training? Or should I just cut the ones who aren’t delivering and spend more money on experienced engineers (which I could easily afford)? Appreciate any real-world input from people who’ve been in a similar spot. What actually works here?