Why does AI seem to have caught on so quickly in programming?

November 18, 2026 5 answers
Asked by _k29!velqueno.6

There seems to be more visible pushback against generative AI in art, music, and writing than in programming. Is that because programmers already share and reuse code, or are there other reasons? And is that impression of programming actually fair?

Answers

4p~vornixa_571

A lot of programming is practical problem solving, and developers have long used tools to avoid doing everything by hand. Compilers, autocomplete, libraries, and examples from documentation all fit that habit. Asking a model for a draft or an explanation can feel like another step along that path.

There are plenty of repetitive tasks where getting a rough starting point is useful, even if it needs editing. That doesn't mean the model always gets the right answer or that every programmer wants to use one.

Reply to this answer
_h71~pixelthrush.4

Code has a handy feedback loop. You can run it, inspect errors, and write tests against expected behavior. That makes it easier to try a generated solution and see whether it works for a particular case.

But passing a test isn't the same as proving code is good. It can still be hard to maintain, insecure, or wrong in cases the tests didn't cover. Someone has to understand and review it.

Reply to this answer
6j!quendavo_419

The comparison with creative work depends on what people value in the result. For some art and writing, the creator's choices, effort, and expression are part of the point. A flood of generated work can also make it harder to find human-made work.

Software can involve creativity and craftsmanship too, but the person using a program may care more about what it does than who typed the code. That's a difference in emphasis, not a clean line between engineering and art.

Reply to this answer
q.728+bravuno

Sharing code probably makes reuse feel familiar, but it doesn't settle the training data question. A developer choosing to publish code under a license is not automatically the same thing as consenting to every use of it in a model.

There is also a lot of publicly available code for models to learn patterns from. Public availability doesn't mean all of it is reliable or free of restrictions, though.

_k29!velqueno.6:

That distinction helps. Sharing an example with other developers and using it for model training aren't necessarily the same agreement.

Reply to this answer
c.835~tavlori

I wouldn't assume programmers have all accepted it. Some find it genuinely useful, some avoid it, and workplace adoption can also come from managers rather than developers asking for it.

Tool makers have an incentive to promote the idea that everyone is using AI, so visibility alone doesn't tell you how universal adoption is.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions