I'm not a stranger to chatgpt / llms and I've also used them to do basic stuff it seems like this junior is heavily relying on it. At first I really held her hand and talked her through implementing a feature just to get to know the application and after let her loose (I'm available for questions).
The feature delivered contained a bug, which is OK, it happens and I quickly spotted it. Instead of guiding her through it I created a ticket which described exactly how to debug the feature (like the exact call which would make the bug happen) and let her have at it. I don't think debugging was taught, which is a bummer but with a bit of curiosity can be learned (stepping into/through methods and examining the data). Pointed out the exact number where the bug happened but no dice. It was a really simple thing where just reading or knowing the standard library or examining the result would've said it all.
Anyways what prompted (heh) me to write this ask hn was that I was guiding her through an issue and at some point she said wait I have an idea, pulled up chatgpt and asked how to implement what we were working on. Then started examining the code and asking me about it. I was baffled. I realize gpt's are here to stay but this isn't normal or good right? I said that I think she relies on chatgpt too much and we should look into the docs for Object X. This als explains the previous code where the fault was so silly and easily found it you actually tried to reason.
How do I deal with this? Should / can I forbid the usage of chatgpt? I'd rather quit than review half baked generated code and don't really feel like helping anymore. Is this the future? I don't think she's learning much of anything trying to cram some generated code into a bigger context instead of taking the time (which as a junior is available) and using a senior mentor to actually learn to program.