The key here is that it success doesn't even have to be likely for VCs to jump in on something that everyone/everyone from a very large category (workers) might

The key here is that it success doesn't even have to be likely for VCs to jump in on something that everyone/everyone from a very large category (workers) might end up using e.g. Uber, WeWork, electric scooter sharing. That is, the demand is clearly already there. Rather than working from profit and trying to scale demand they are working on profitability of a huge market as the technical problem. They have more money than investment targets, so they basically fund a few moonshot projects that look like they have a 1%+ chance rather than a 0% chance. I suppose the thesis is that once one of these eventually wins, the losses on the rest will be more than recovered. You just have to do enough 1 out of 100 projects.