Consider the possibility that the data isn’t the issue, and that YouTube is designed that way on purpose because that is what generates the most traffic.
Of course the argument can be made that they have the data and aren’t using it properly, but that’s not the simplest possible explanation.
The simplest is that they have the data, it’s working perfectly, and showing what people are familiar with is what generates the highest engagement, at the expense of new material.
I find 2022 YouTube profoundly boring for the same reasons you do and their recent excess of advertising is atrocious. Instead of producing something valuable enough to subscribe to in the form of product innovation, the genius MBAs at Google have decided that annoying people to death with insane cable-television-style advertising is the value driver to bank the future brand of the company on. Seems like maybe the team that was responsible for the smash hits like Google Wave and the destruction of Reader have found new positions in the company, except this time they’re joined by a 30 year television advertising veteran who really gets the internet. But I digress…
But it also seems to me that both behaviors aren’t based on bad decisions, but more likely on the viewing habits of the majority and that’s what the algo has learned. It’s an interesting concept, even if a little dark and sad.
I do hope I’m wrong and they’re just bad at dealing with data to make recommendations, but it seems the less likely scenario since Google has had virtually unlimited resources and these are the specialized type data problems they are most adept at.