Absolutely Zillow overpaid, at least a lot of the time. However, part of that is because they apparently had a hard time getting people to accept their offer often enough, and so tweaked the algorithm to pay higher. This suggests that, if you pay the amount you can actually make money on in six months, you don't get very many people accepting your offer, because people don't know as much as Zillow knows about what's about to happen in the housing market.

Or, maybe, Zillow is just no good at predicting housing prices. But it's at least worth considering that, if they were actually reasonably good at it, they might have come to the conclusion that the only way to make money right now is to convince house sellers that their home is about to be worth less than it seems to be now, which isn't likely to be something they wanted to try to do. Maybe they're just incompetent, but they might be competent and have nonetheless come to the same conclusion that Zillow Offers needed to be shut down.

p.s. thanks for the great links, but I would like to point out that the St Louis Fed link goes up to July 1, which leaves open the possibility that Zillow saw something in the last 1-2 months different from what came before.