I disagree that it is no longer development, I see LLMs as yet another tool for generating code, and we’ve had generated “source” code since before C was standardized.

Fair. There is a difference between using LLMs to generate boilerplate code customised to your context or provide a starting point if you’re stuck on a given problem and struggle to find a different perspective for approaching it, and using it to get around having to do mental work.

My term is intended for the kind of vibe coding where there is little, if any, technical skill involved and people are just letting LLMs slop together code without meaningful code quality assurance. In those cases, I don’t think it warrants recognition as development. If it produces workable results, cool. Call it software generation.

Using it as a learning assistant would probably be the most justified use case in my opinion. I have my reservations whether it is suitable for that purpose but I don’t know enough about the specific way it is applied to comment on that. If it produces training code that isn’t directly published you dodge the legal iffyiness, and if it helps build skills, that solves the “relying on AI makes you unlearn skills” issue.