Having driven for Uber and Lyft for a year and a half (even during the "good years") - the rate of new hires was most definitely a major concern and there were definitely rocky times. I was lucky to be able to mitigate this by moving to San Francisco from San Diego. While in San Diego I saw my take go down from ~$1200/wk to something like $800/wk, then back up to $1200-1600/wk when I moved to SF, and then back down to about $800 by the time I finished.
Personally I approached it as an exercise in entrepreneurship, a deep study in the dynamics of real-time economics, and an opportunity to vastly improve my "people skills" in addition to making more money than a PhD postdoc in biochemistry... So a success in all fronts, personally, but also I knew when to quit.
My suspicion is that many people who begin driving for Uber overestimate their net earnings and underestimate their deferred expenses e.g. vehicle depreciation and maintenance. If Uber knows this and is taking advantage (with new-driver bonuses, etc...), they're essentially a multi-level marketing company.
If Uber drivers are, in fact, independent contractors, then they have some responsibility to determine whether the business they're engaged in is a profitable one, and to judge the risk involved. It seems like it would be very easy to underestimate both the costs and the risk.