When I was at apple, we did this thing people called "keybote-driven development" where everything was made to look great in and keynote presentation to an exec. If it looked great you got the go ahead to finish it and ship it, and if not it got skipped. There are a lot of things to like about this approach. However, Time machine's UI to me feels like a quintessential blind spot of this incentive structure, where it looks visually striking and cool but actually using it for real use cases it often falls flat quickly. To update it you would need to fit the motivation and payoff in a keynote presentation and it's hard to do so.