Sure, that's very obvious and all of that goes without saying. This is experimental, unreleased tech, and things could go very, very wrong at release. Neither is it possible to to solve the trilemma a 100%. It's all about probabilities. Testnets have been running for months, it's pretty cool to see the attack attempts being slashed. Audits are being done, bug bounties are available - there's a ton of expert eyes on the project. The phased rollout is prudent, will offer ample time to evaluate how things work in the real world. It could all still come crashing down to a halt, but the various teams developing Ethereum 2.0 must be commended for taking every precaution (and then some) to ensure it all goes smoothly, even if it takes years.