> I feel like this sentence is in every article for a reason.

Breakthroughs, BY DEFINITION, come from people going against the grain. Breakthroughs are paradigm shifts. You don't shift the paradigm by following the paradigm.

I think we do a lot of disservice by dismissing the role of the dark horses. They are necessary. Like you suggest, there are many that fail, probably most do. But considering the impact, even just a small percentage succeeding warrants significant encouragement. Yet we often act in reverse, we discourage going against the grain. Often with reasons about fear of failure. In research, most things fail. But the only real failure is the ones you don't learn from (currently it is very hard to publish negative results. Resulting it not even being attempted. The system encourages "safe" research, which by its nature, can only be incremental. Fine, we want this, but it's ironic considering how many works get rejected due to "lack of novelty")