I'll say the thing a lot of devs think but don't post: using AI to write parts of your Unity code isn't cheating. It's a tool, the same category as autocomplete, code snippets, or the asset store before it.
What separates a good developer from a bad one hasn't actually changed. Can you read the code it generates and tell when it's wrong? Can you debug it once it breaks in a way the model didn't anticipate? Can you architect a system well enough that AI-generated pieces slot into it cleanly instead of turning into spaghetti?
If yes, AI-assisted code is a faster first draft. If no, that gap was going to show up eventually anyway, with or without AI in the picture.
Where I draw my own line: boilerplate, utility functions, repetitive setup, sure. Game feel, balance, core architecture decisions, no. Those need a human who understands why the game should feel a certain way, not just what code compiles.
Curious how other devs split this. Where's your line?
Top comments (0)