The core concept is being proven pretty well, in markets with good weather and good roads at least... whether they can achieve some sort of critical mass that allows them to recoup any of the investment, even on a single capitalized vehicle... who knows. I've heard each Jag costs ~$250k fully loaded... that's a lot of ~$14-on-average trips before breakeven, and obviously the trips aren't free.
That all being said, riding in one feels like the actual future, and if they close in on this first-mover advantage and keep on keeping on, and strengthen their Uber integration, and open in more and more markets... there's a real chance they establish themselves and make this thing actually capture a significant portion of the trip economy.
That being said, it's not generalized, so opening markets is expensive because they're all individually trained for each mile of road in the market, I expect they'll be functionally useless in bad winters (but that's not a dealbreaker, per se), and those operating expenses remain high. Though I do expect we'll begin to see those lower, now that they're so deep into the proof of the concept. (They're finally starting highway driving here in PHX, which is really the last hill to climb.)