AI wrote 80% of my code last week and I'm not sure how to feel about it

I was building a little API scraper for my side project, nothing crazy. Copilot just kept finishing my lines and suggesting whole functions. By friday I checked the git log and 80% of the code was AI generated, but I still had to fix a ton of logic bugs it introduced. On one hand it saved me hours, on the other I had to debug stuff I didn't fully write. Anyone else hitting this weird middle ground where AI speeds things up but also makes you question your own understanding of the code?

2 comments

2 Comments

maryh962d agoMost Upvoted

Nah, that's a good thing actually. You're not losing understanding, you're learning to debug better. That's the real skill anyway. Copying code from Stack Overflow taught us all the same way. You fix the bugs, you learn what broke. It's not some magic crutch. If anything, it's making you faster. You spent hours debugging instead of typing boilerplate. That's time well spent. You know what works now and what doesn't. That's how you get better. Nobody writes everything from scratch anymore. The gatekeeping about "real code" is pointless. Just ship the thing.

wyatt_shah852d ago

Gotta push back a little here @maryh96, because debugging code you didn't fully write hits different than fixing your own mistakes. It's easy to patch the bugs without ever really understanding why the AI wrote it that way in the first place.