> 2. Decentralized voting on reversal of transactions (90-95%+ vote needed to reverse to avoid 51% attacks)
Couldn't you technically just 'git checkout' a previous commit from before the fraudulent transaction occurred and pretend it never happened? Isn't the real problem that you'd have to convince a majority of users to do the same?
The DAO experiment ended this way. Once an exploit started siphoning tokens to a new fund, that same exploit allowed anyone the same maneuver. Fixing an exploit is changing the rules, and the experiment would have ended in deadlock without it.
Couldn't you technically just 'git checkout' a previous commit from before the fraudulent transaction occurred and pretend it never happened? Isn't the real problem that you'd have to convince a majority of users to do the same?