consensus-specs

Ethereum 2.0 will have a large number of validator nodes even at genesis[+] and the number will grow over time. A deposit of 32 ether will be required to become a validator — not cheap, but also not astronomically expensive. Here's a presentation from Devcon5 earlier this month demonstrating an Eth 2.0 client running on a Raspberry Pi: https://slideslive.com/38920013/eth-20-on-a-pi [+] https://github.com/ethereum/eth2.0-specs/blob/master/specs/c...

You can take a look at the eth2.0 specification here https://github.com/ethereum/eth2.0-specs. They have solved the nothing at stake problem, and they currently have 8 independent client implementations written in different languages that can 'talk' to each other on a multi-client testnet. They are much further along then people may realize.

In most PoS designs, registering a new validator account takes some fixed amount of time, just as a measure to prevent grinding attacks. If registration takes an hour, and I can't predict the random seed an hour in advance, then I can't predict how my account will be scored, so grinding would be futile. Most PoS designs don't otherwise favor old accounts. I know Peercoin does, but that's a very small project, and I'm not aware of any others. Ethereum 2.0 definitely won't have any such mechanism. In the current spec [1], staking is binary; either you have an active validator account or you don't. [1] https://github.com/ethereum/eth2.0-specs/blob/master/specs/c...