Has there been any scholarly research done to compare building an user interface with and without AI? In my experience, progress with manual UI development is slow to start, but increases in speed and then plateaus. If done well. Done poorly, and progress slows and slows until it is glacial. But for sake of argument, let's consider it done well. With AI in the mix, development speed is rapid at the start. But then it slows down. And that slowing down feels exponential. I also suspect that AI development can hide catastrophic bugs to a greater extent than manual development. I have a feeling that over the short to medium term, AI wins. But there could be a point where manual development wins for longer projects. There is of course a third option: augmented with AI and manual development. With humans in the mix, reading AI code. That would likely have different characteristics entirely. Thoughts?