I'm not too sure but few things come to mind:

1. Upgrade protocol to include protections for well known cold wallets held by exchanges (ex: API call has to be made to the exchange's security endpoint to validate each transaction out of the wallet. Exchange staff would need to manually allowlist large transactions before they are transmitted).

2. Decentralized voting on reversal of transactions (90-95%+ vote needed to reverse to avoid 51% attacks)

This is getting pretty close to the banking system, at which point one needs to ask - maybe just improve existing protocols?

> 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.

Ethereum is programmable, such a protocol can be implemented as a smart contract.

Not going to work, otherwise it would already have been done.

People who control or take advantage of cryptocurrency don't want this to happen.

Good luck getting 90% of a large group of people to vote the sky is blue.

I think the move is less having a central thing and more advancing wallet and multisig technology. ByBit was pretty reckless by using a simple majority multisig to hold $1.5b. At that level you should probably have a few speed bumps. Like, maybe a majority of signatures allows you to make a proposal, but you can only accept the proposal after a couple hours, which would give you the chance to see the malicious transaction and bail on it.

Something like that would probably be overkill for individuals, but most people would definitely benefit from some added on chain bureaucracy regarding how their accounts are managed. And yes, for many this would lead to a system that isn't notably less centralized than the traditional banking system. But people would at least have a choice as to where their wallets gets to sit on the bureaucracy <> complete freedom spectrum. And even if they end up closer to the bureaucracy end, they'd have a lot more flexibility and lower administrative fees than what they currently have.