There's very little productive new ground to grow on if you're working on YouTube.
There's existing glaring site problems, but they're on purpose and ideologically driven (shorts, placebo dislike, subscriptions mean nothing) so no one is going to touch them. It's politically safer and more advantageous to mess with the UI and irritate all the users every couple years.
YouTube know that the choices they've made are qualitatively bad because users overwhelmingly complained about them (dislike button being the most obvious example of that)
...but quantitatively numbers didn't go down enough to force any kind of retraction or even recognition of discontent, so now we're stuck with the results.
So yeah sometimes the path of least resistance is building something new.