By analogy to driverless cars, it'd make sense to avoid letting your car out on public roads; but it wouldn't make sense to avoid building a prototype and running it around a test track. A real thing lends itself to iteration much better than an on-paper idea does.

In that case Ethereum shouldn't exist at all. Solidity and the EVM itself are a good example of that philosophy. If they weren't gonna move fast they should have waited until they had a better story for static analysis and model checking of smart contracts.

Not necessarily. Keep in mind when Ethereum started, it had no value. But with bitcoin, there is huge value and so potentiality for people to loose a lot of money if there was one minor bug.