[mouzogu]

I think part of this is due to the algorithmic nature of so many of these platforms.

Majority of people will only watch what is trending, it creates a feedback cycle that perpetuates that kind of content which is popular with the mainstream. And whatever is mainstream is usually homogenised and boring.

As someone who watches quite a lot of movies, I always felt that they peaked sometime in the 70s.

[randcraw]

Yes, I agree that the heavy reliance on algorithms by media sources today are a big part of the problem in finding new fare that's worthwhile. This is one of the major demerits of AI and machine learning -- pattern recognizers learn and highlight only the major signals they already know -- the peak of the bell curve dominates search results. If the mainstream is not what you want, you hired the wrong travel guide.

But creativity and innovation are minor signals, outliers. Probabilistic search as used by AI and ML detects not outliers, but the peak of the bell curve, the patterns that are recognizable and most popular. If you hoped to find something unusual or unrecognizable or rare, AI/ML/statistics are the wrong tools to do that. To a naive AI algorithm, good outliers look no different from bad ones, so both will be overlooked in your search results.