I've heard talk of self driving cars at uber, but are they really "investing" in them? Self driving cars will be available for anyone to buy within 5 years -- Uber would be better off buying some of those. And so will their competitors.

I'll say this: uber will be the next groupon, zynga, webvan, ... it's laws of physics, we've seen it before -- take a simple but good profitable product with early user growth, a big market (ppl needing cabs, groceries, stuff at discount, addictive entertainment), and then completely kill it through over-investment. Then they use the money to buy the growth they need to justify their ever-higher valuations.

Market growth really looks more like a sigmoid than a hockey stick. Towards the top, near the plateau, it's possible to buy growth to feed the myth of the hockey stick. By doing this, you are incurring a sort of technical (real?) debt that is harder to pay back than people realize, particularly as margins get squeezed by inevitable competition.

In Uber's case, they will also mostly lose in court.