Bertrand Russell, the 20th century philosohper and mathematician, once said:
The fundamental cause of the trouble is that in the modern world the stupid are cocksure while the intelligent are full of doubt.
LLMs are introducing something like this dynamic into programming. The programming version of the statement is:
The fundamental cause of the trouble is that in the modern world the less experienced programmers are writing much more code than the more experienced ones.
I’m a part of the problem. I’ve only been programming professionally for a decade and I’m writing a lot more code with LLMs. Future me will hate the AI-slop that I’m slinging.
Until I get future me’s wisdom, all I can do is pass on what I’ve learned so far. We have learn to say “no” to some of the code LLMs try to sell us on.
No. We don’t need 5k lines of code solving a problem better handled by a library.
No. We don’t need a third implementation of the same-ish React component.
No. We don’t need that tautological test.
No. We don’t need 1000 lines of test code with 0 abstraction.
No. We don’t need this code stuffed into this already massive object/function.
No. We don’t need domain logic inside our http route handler.
Saying “no” is slower than hitting TAB. When we say “no,” we’ll write less code, but we have do it anyway. If we don’t learn how to say “no,” we’ll wind up with a code base so voluminous that we won’t be able to make sense of it, even with the help of LLMs.
Many code bases get to the point where it’s so difficult to make progress that feature work stops and teams work on tidying, but if LLMs get much better than us at navigating a messy codebase, when our mess maxes out their capacity, no one — human or otherwise — will be able to save the codebase.