[lubujackson]

I had to learn the hard way that you need to know at least the structure of what the code should be doing. Vibe coding doesn't work for long if you just YOLO repeatedly, you need breadcrumbs back to sanity. If you get lost in the weeds the LLM can refactor itself back to understandability but it won't get there itself (at least not yet).

[florianherrengt]

I know what the code should be doing. But it's not doing it. Now I have to deep dive into a codebase that's far less familiar to me than if I'd written it. This isn't usually such a frequent issue because systems typically have dedicated owners. We structure it this way precisely because debugging someone else code is just harder.

As for the broader responsibilities. Things like architecting systems, shaping their structure and behavior, planning their evolution and keeping everything clean with regular refactoring. These sound exactly like what I've been doing for the past 15 years. So what's actually different here?