Related questions
@ftkurt extending that logic. Would a compiler be considered an AI, and could one measure it's level of contribution?
@ftkurt If humans were primarily managing, I think that would resolve YES, since the question mentions "human coding".
@12c498e Well the problem is at what point you would call humans not coding? Me copy pasting. Or reviewing or dividing work into subtasks? I might still occasionally go low level and do some contributions to the code. As is I am already using chatgpt to generate test code for instance. I will still fix some pieces here and there. Improved ai will gradually reduce my involvement. But i will be still a software engineer.
@ftkurt True, that's a good concern. I'd probably consider something "coding" if you are still manually looking and thinking about literal code. So, copy-pasting ChatGPT output would still count as coding, while describing your business requirements and receiving a finished product in response would not be coding. But that's a very subjective stance, and it's unclear whether the author intended something similar.
@12c498e my point is humans will be still coding; like most of it will be done by AI but there will be still human coders. the question sounds like there will be a division similar to 10% fully human-coded projects and 90% fully AI-coded ones. What if for instance, we have a world where 100% of projects are 95% AI coded and 5% human coded?
@ftkurt I also agree that it seems strange to care about projects as a whole rather than the ratios of each project. Maybe "primarily" also refers to 50% per-project? So in your example, we'd say that 95% AI coded is "primarily created by AI" and 5% is "minimal human coding"?