> This is the prevailing theory but in practice I'm not sure it will work.

There are other proof-of-stake blockchains other than Ethereum, and for those it has indeed worked quite well in practice.