Scientists spend something like 20-30% of their time begging for grant money.
Worse still, in order to get the grant money, scientists have to abandon the genuine pursuit of knowledge and do some random busywork bullshit that they deep down know will produce nothing of value.
Once they have produced the bullshit research, they then spend another disproportionately large chunk of time convincing their colleagues that their research is actually valuable.
The remainder of their work time is spent pretending that they don't realise that their peers' and colleagues' work is complete bullshit.
Don't get into academia.
You've made a good insight with an interesting solution. You might already be aware of this, but for an academic perspective, Canadian academic Marshall McLuhan framed this with the famous assertion "the medium is the message," [0] in other words the method of delivering a message strongly influences how that message is drafted. The reasons align with what you've written: the incentives caused by YouTube algorithms do pressure creators to create more clickbait content.
Perhaps human curation can be an initial way to get more viewership for quality channels that don't try as much to game the algorithms. Recommendations on HN and other forums have led to the discoveries of such interesting channels, though this would have limited impact versus change on YouTube's end.
What I'm trying to say is, while the medium doesn't change core mechanics of human behavior, it distorts and accentuates – there is, for instance, a difference between tiktok and a novel. And we do have power of mediums, we can choose to engage in different types, and we can lobby for digital mediums to change. So pitting youtubers against the algorithm (just pimp your quantum physics lecture with an O-face thumbnail) isn't the only way forward. There's also "here's how YouTube could be better" or "we need a better alternative for this type of content". Unfortunately, we have almost 0 visibility and control over the algorithm so there's very little behavioral analysis that can be done.