3 ms·

This is additionally a reminder that crypto currently holds a fairly irreconcilable catch-22. "Freezeable" tokens completely defeat the purpose of a decentrali

by tevon 4y ago

This is additionally a reminder that crypto currently holds a fairly irreconcilable catch-22. "Freezeable" tokens completely defeat the purpose of a decentralized currency. Why create nodes/verifiers etc if a wallet can be unilaterally frozen, either under pressure from gov't, community, or another entity. However NOT allowing freezeability allows for massive hacks, either by insiders, poor security, or sloppiness. Seems the solve is somewhere in the middle, a pool of trusted intermediaries.

lupire 4y ago
51% vote to freeze (by validators) is decentralized.
- jl2718 4y ago
  This 51% number is not really relevant in blockchain consensus. There is Byzantine consensus, which derives from a specific set of assumptions, but those don't really apply. In reality, consensus is not at all a vote. Coinbase and Circle can accept a chain with only 1% validator support, and that will probably be the canonical chain. Apologies to the many blockchain economists and computer pseudo-scientists that spent the last decade on this, but Consensus is mostly an irrelevant concept. The good news is that new cryptographic primitives are there to fill the same role, and we're just getting started with that.
jl2718 4y ago
It seems to me that validators aren't taking their sovereignty seriously. They are either following the exact suggested rules of the blockchain, or they are following the laws, but I have not seen the idealized consensus scenario where each node is soft-forking their rule-set to align with their own values. This was explicitly suggested in early discussions of bitcoin. There are three basic scenarios where a block participant may make a values judgement: 1. Including a transaction into a block 2. Block acceptance and propagation 3. Chain acceptance by the client Everything else is cryptographically locked, as in, a full-chain audit would reveal discrepancies, but these choices maintain internal consistency. You may hate this idea, but it is coming, and the default rule set will come from the jurisdictions. However, they will only have the power of transaction censorship by unanimous consent, total power of block censorship, and almost no control over chain censorship. I just had a half-baked idea: a chain that has the ability to hard-fork only in small differences, but essentially allows an infinite number of forks and merges in such a way that each transaction is valid if not explicitly invalid. In other words, a transaction fork does not invalidate the proof of an unrelated transaction that descends from the forked root hash. This would have to be a very different kind of hash/proof. Such a chain would greatly reduce the risk of rule modifications such as censoring transactions from an attacker, while also supporting the choice to use a chain without censorship, and enable reconciliation between the two.

by tevon 4y ago

lupire 4y ago

51% vote to freeze (by validators) is decentralized.

jl2718 4y ago
This 51% number is not really relevant in blockchain consensus. There is Byzantine consensus, which derives from a specific set of assumptions, but those don't really apply. In reality, consensus is not at all a vote. Coinbase and Circle can accept a chain with only 1% validator support, and that will probably be the canonical chain. Apologies to the many blockchain economists and computer pseudo-scientists that spent the last decade on this, but Consensus is mostly an irrelevant concept. The good news is that new cryptographic primitives are there to fill the same role, and we're just getting started with that.

jl2718 4y ago

It seems to me that validators aren't taking their sovereignty seriously. They are either following the exact suggested rules of the blockchain, or they are following the laws, but I have not seen the idealized consensus scenario where each node is soft-forking their rule-set to align with their own values. This was explicitly suggested in early discussions of bitcoin. There are three basic scenarios where a block participant may make a values judgement: 1. Including a transaction into a block 2. Block acceptance and propagation 3. Chain acceptance by the client Everything else is cryptographically locked, as in, a full-chain audit would reveal discrepancies, but these choices maintain internal consistency. You may hate this idea, but it is coming, and the default rule set will come from the jurisdictions. However, they will only have the power of transaction censorship by unanimous consent, total power of block censorship, and almost no control over chain censorship. I just had a half-baked idea: a chain that has the ability to hard-fork only in small differences, but essentially allows an infinite number of forks and merges in such a way that each transaction is valid if not explicitly invalid. In other words, a transaction fork does not invalidate the proof of an unrelated transaction that descends from the forked root hash. This would have to be a very different kind of hash/proof. Such a chain would greatly reduce the risk of rule modifications such as censoring transactions from an attacker, while also supporting the choice to use a chain without censorship, and enable reconciliation between the two.