I've even still got a streaming service I can do exploring on, since YouTube bundles one with Premium. I find it's a good thing I have my own collection though since it tracks my interests poorly.
I've gotten back into buying my own video too. I don't consume a ton of video and I dropped Netflix streaming a while ago because the delta between me marking something for the queue and actually getting to it was becoming routinely larger than the amount of time Netflix would still have the thing I wanted to see.
The problem is, I don't even see the second derivative on this trend turning, let alone the first. Metric-driven development, by its very nature, will take away every knob from you that you could conceivably use to drive their metrics lower. I think that's a reasonable approximation of the root cause of the reality observed in the OP. If you happen to agree with their metrics then hey, good times for you, but the odds of that are low since you're probably not looking to maximize the monetization they can extract from you as priority one.
Therefore, the only option is, get off metric-driven-development platforms. There is no alternative and will be even less of one as time goes on.
I suspect in the very long run this metric-driven development will eventually die off as all consumers come around to this realization one way or another and start turning to other alternatives, but it can easily be 5-10 years before there's enough of us for those "alternatives" to be able to survive in the market. Fortunately, MP3 players haven't gone anywhere. (Although it takes some searching to find ones that aren't also trying to match the streaming services and stick to old-school "play what you ask for and not anything else, unless you ask for shuffling or randomness explicitly".)