About 10 years in, mostly consulting for smaller companies and early stage startups. The last few months something shifted in the kind of work coming my way.

Used to be people hiring me to build new things or extend existing systems. Now its cleanup, like straight up triage on codebases that are barely holding together.

The pattern is always the same. A non-technical founder pays someone to build their product. It works on the surface. Then users start hitting it and everything falls apart, slow queries, memory leaks, auth logic thats swiss cheese, error handling that catches everything and does nothing with it.

When I actually look at the code its pretty obvious what happend. AI generated top to bottom. You can tell from the comments alone, that weird overly polite explanation style that no human dev writes. Algorithms that technicaly work but make zero sense for the actual use case, data models that look like someone asked "what are all the possible fields" and the AI just listed everything.

The thing is these founders arent stupid. They saw demos, believed the hype, hired a "developer" who was really just a prompt jockey, and got something that passed a demo but crumbles under real usage.

Im not anti AI at all. I use Glm-5.1 and Claude code daily for my own work and it genuinley speeds things up. But I also know when the output is garbage cause ive written enough code by hand to smell it. Thats the part you cant shortcut.

I think we're about to see a wave of this. Companies built on AI slop that need actual engineers to come in and rebuild the foundations, job security for experienced devs honestly but depressing that it has to happen this way.