I remember, when Ethereum first launched, that a lot of computer scientists I spoke with privately said it would never work—because it's kind of this crazy idea that the way we're going to get a trusted computer in the cloud is we will each run a copy of all the computations on all of our computers, and then we'll sink it up, and make sure it matches. That's the recipe for building the slowest computer in the world, but somehow we've gotten away with it. And so I think the big debate now about Ethereum has shifted from 'Will it work?' to 'Will it scale?'