1. AI Code Generation and Technical Debt
LLMs make it easy to write garbage code that creates technical debt and must eventually be rewritten by developers.
Hear it at 0:00 →Engineering video
1:15 · 16 September 2025
How AI Creates More Work For Software Developers (and a lot of garbage code)
Key concepts in this video:
Transcript
AI actually generates a lot of technical debt and someone will have to fix that. And the point here is that LLMs made it incredibly easy to write basically garbage code that has to be rewritten. And guess what? Someone will have to fix it, someone will reel software skills. As you can see in this image that we analyzed in another video together with Bogdan, a lot of the code that's been added with AI, and this is a study from May 2024 to December 2024, but it still applies from last year.
Uh you see a lot of the lines were reworked. Uh there is a similar more recent study that came out about open source developers which actually reached the same conclusion a lot of the code that's being pushed with AI has to be rewritten in order for it to be functional. And if you multiply this by the millions of code bases that we have, worldwide enterprise startup, you understand that we are creating a huge problem and if you stay around, somebody will call you up to fix it.
Just like this bite coder, who was very famous on Twitter, because he used cursor to build his SAS, and then the whole app started doing random things and he couldn't fix it, and this is a solo developer, imagine full development teams suffering from the same.