What do you all think about the idea of asking AI to generate code that has a few intentional mistakes in it, so that I can practice my debugging skills? Do you believe this approach is effective for improving debugging and problem-solving abilities?
4 Answers
This feels a bit like practicing pedaling on a stationary bike—better to just ride a real bike to really get the feel for it! By focusing on actual projects, you’ll have plenty of debugging practice.
I tried this with AI, and honestly, the results were just okay. The models generally produce decent code because that’s what they learned from! To get actual useful mistakes, it might be better to look at previous versions of software with known bugs. Tracking down bugs in real code might give you a more authentic experience.
AI can produce a lot of bugs on its own—no need to ask it for help! If you want to really learn, just create real projects again and again. You’ll encounter more than enough issues to troubleshoot.
Honestly, you’ll end up finding tons of bugs just by building stuff. Going for intentionally flawed code might not be the best idea since real projects will have bugs naturally. Why not just dive in and get your hands dirty?
That sounds interesting! I think digging through old repos for bugs could be helpful, but it might also be a hassle if you don't know where to start. Looking for open issues in projects and fixing them could be a more effective way to improve.