The ceiling of usage of an ecosystem is tied to the expected value of the worst case downside.

A small problem that is very likely could have a high expected value, as could a game-over but rare problem.

Even if 99% of people in the ecosystem aren't savvy enough to understand the downside risk, it doesn't matter–the system will limit itself naturally to avoid that worst case downside risk.

Technically the expected value of the worst case downside is balanced out by the expected value of the normal case upside, so something that allows something radically better than before can get usage before hitting that ceiling.

Greasemonkey back in the day had this problem.