I saw this myself while building my personal project. AI helped a lot at the start with boilerplate and APIs, but once the codebase became big, it couldn’t keep the overall structure right. The code looked fine in parts, but didn’t work well as a system.
That’s when I realised humans still need to own architecture, data models, and core decisions. AI is great for speed, but not yet for maintaining a growing, long-term codebase.
I saw this myself while building my personal project. AI helped a lot at the start with boilerplate and APIs, but once the codebase became big, it couldn’t keep the overall structure right. The code looked fine in parts, but didn’t work well as a system.
That’s when I realised humans still need to own architecture, data models, and core decisions. AI is great for speed, but not yet for maintaining a growing, long-term codebase.