totally. just a nice reminder for myself how the reports of software engineer death are greatly exaggerated :)

I find LLMs are particularly bad at debugging code. They're pretty passable for fixing/suggesting fixes once the bug is identified, but identifying that problem largely falls onto the shoulders of whoever is doing the coding (or a Stack Overflow thread lol).

Just spent hours stuck on a react issue. Copilot nor chatGPT could fix it with so so many attempts.

Finally found a stack overflow thread from 2020 where someone had the same issue and their solution worked.

AI is great but the SO depth of shared experience remains unmatched 🫂

Posts