What you've read (and I've also seen though not studied in any depth) is that a statistical category of person called "Uber driver" may lose money. But the sta

What you've read (and I've also seen though not studied in any depth) is that a statistical category of person called "Uber driver" may lose money. But the statistical category of "Uber driver" is not the same as a "person who drivers for Uber". Consider the argument: "can an Uber driver save for retirement?" If we define Uber driver as: "works for uber exclusively for 40 hours per week". Our answer to the question might be yes or it might be no. Ultimately what matters is the number of people who fall into that category every day for the rest of their lives. If we find no people fit that category or that the people who do fit that category don't fit it for very long then we can safely say the answer to the question is irrelevant. After all, "can unpaid interns save for retirement?" is a meaningless question. It doesn't matter. Statistical categories don't retire. And people who work as unpaid interns do not remain so for very long. A person who drivers for Uber may just need quick, supplemental cash. Cash that may be exclusively funneled into a person's retirement account. If we compel Uber to provide a "living wage" or a retirement package, we may find that those who need quick cash go without it. And people who would have saved for retirement are now not able to.