> I see a lot of money in the future for competent engineers who can unfuck what ChatGPT has fucked.
I do not want to do that work. Cleaning up junior code is easy, because they mess in predictable ways.
LLM generated code can be extremely complex at the same time that nonsensical. It has a line of genius and then code that does nothing. It can use many different libraries mixed in inhuman ways.
Better to let that companies shut down and do something better elsewhere.
People have patterns for messing things up. That’s the reason they’re called anti patterns. They learned the wrong thing and then apply it consistently. And you’ll still have human limitations to help you. They’ll want the thing to be working at least superficially, so a lot of the glue code will be correct. And the anount of code will have an upper limit.
I have been exclusively doing this in the past year, selling my services as “hardening vibe-coded prototypes for production” or “helping early stage startups scale”.
In the best cases, they were able to reach funding or paying users. Architecture debt is one of the worst kinds of tech debt, so if you set it up right, it’s really hard to mess up.
In the worst case, after my contract ended, the CEO fired the whole US engineering team and replaced them with offshore resources. This was an example of messing up despite the architectural and procedural safeguards we built.
