3 ms·
The nothing-at-stake problem was solved many years ago by introducing slashing. The earliest proof of stake systems did not have slashing (they did not actually
by AgentME 4y ago
The nothing-at-stake problem was solved many years ago by introducing slashing. The earliest proof of stake systems did not have slashing (they did not actually put validators' funds "at stake"; the use of the word "stake" wasn't so literal yet) and were vulnerable to the nothing-at-stake problem because of that. A decent write-up: https://medium.com/coinmonks/understanding-proof-of-stake-the-nothing-at-stake-theory-1f0d71bc027 https://medium.com/coinmonks/understanding-proof-of-stake-th...
- mccorrinall 4y agoAnd this slashing process does not solve the issue. Or well, it does, but it makes the whole system centralized and introduces a single point of trust. The bounty hunters, who find misbehaving validators which signed orphaned blocks, will slash the validators stake. But those bounty hunters are programs which are being run in Vitalik‘s basement. Not everyone will be able to run one like a normal validator. They are trusted entities within the network.
- AgentME 4y agoWhere did you get the idea that slashing is done by centralized/trusted bounty hunters? It's completely wrong and I'm not sure how someone could come to that idea in good faith. (Maybe you're misunderstanding some unrelated criticism made of an L2 system like Optimism from before it implemented fraud proofs?)