4 ms·

Proof of work = trust physics to determine what happened. Proof of stake = trust humans to determine what happened. PoW is superior.

hudon 5y ago
> Proof of work = trust physics to determine what happened. This is the kind of vapor Bitcoiners love to inhale and it's just not true. In the history of Bitcoin, it has happened many times where the "work" of a chain was irrelevant because humans disagreed with the contents of that chain. See: 2010 fork, 2013 fork, 2017 fork, etc. When two chains are mined using the same binary, yes, "physics" seems to be in control on which chain wins (ie. the one with the most work). But when one chain expresses a slightly different protocol, humans take over and decide which protocol is the valid one (even if it has less work). And if humans are ultimately deciding on what's valid Bitcoin and what's not, we have to ask ourselves which humans are the most influential in the discussion on what's valid and what's not. We need to know who these players are because with a project worth this much, you can bet that they are targets of powerful and wealthy entities.
- atweiden 5y ago
  Boring histrionics aside, you haven’t refuted GP’s point. Pure proof-of-stake consensus systems all suffer from the misfeature of not even having a basic quantitative fork ranking protocol. Jude C. Nelson critiques this better than anyone [1]: PoW requires less proactive trust and coordination between community members than PoS -- and thus is better able to recover from both liveness and safety failures -- precisely because it both (1) provides a computational method for ranking fork quality, and (2) allows anyone to participate in producing a fork at any time. If the canonical chain is 51%-attacked, and the attack eventually subsides, then the canonical chain can eventually be re-established in-band by honest miners simply continuing to work on the non-attacker chain. In PoS, block-producers have no such protocol -- such a protocol cannot exist because to the rest of the network, it looks like the honest nodes have been slashed for being dishonest. Any recovery procedure necessarily includes block-producers having to go around and convince people out-of-band that they were totally not dishonest, and were slashed due to a "hack" (and, since there's lots of money on the line, who knows if they're being honest about this?). Under adversarial conditions like those outlined by Jude C. Nelson, PoS reverts to “phone-a-friend consensus”, or — to use a euphemism coined by Vitalik Buterin — “weak subjectivity”. Centralized Git repos offer a close to identical consensus model. What’s the point in having a “blockchain” at all if under adversarial conditions, miners are necessarily demoted in favor of trusted human coordinators? Git is very “green-friendly”, too. [1]: https://news.ycombinator.com/item?id=26810619 https://news.ycombinator.com/item?id=26810619
  - hudon 5y ago
    > What’s the point in having a “blockchain” at all if under adversarial conditions, miners are necessarily demoted in favor of trusted human coordinators? If you want to see precisely how "miners are demoted in favor of trusted human coordinators" in Bitcoin, you can read this analysis of the 2013 fork: https://freedom-to-tinker.com/2015/07/28/analyzing-the-2013-bitcoin-fork-centralized-decision-making-saved-the-day/ https://freedom-to-tinker.com/2015/07/28/analyzing-the-2013-... In the quote you posted, Nelson is wrong when he states "[Bitcoin] provides a computational method for ranking fork quality". It only provides such a method when both forks agree on the underlying protocol. If two chains aren't being reconciled because the network is in disagreement on which set of rules is valid, the "computational method" is thrown out the window, and the big guns are called in (high profile developers, influencers in the space, etc.). He uses the words "canonical chain" and "attack" as if it's clear which chain is the attacker and which chain is canonical, but we need humans to help us make that determination on which is which. Chains by themselves aren't hostile or not, they're just numbers. I'm not saying Ethereum is better by the way... they're all the same.
    - atweiden 5y ago
      > In the quote you posted, Nelson is wrong when he states "[Bitcoin] provides a computational method for ranking fork quality". It only provides such a method when both forks agree on the underlying protocol. Firstly, the “underlying [PoW] protocol” assumes the “majority of CPU power is controlled by nodes that are not cooperating to attack the network” (Satoshi, 2009). IOW “Bitcoin” is defined as the chain with the highest cumulative hashing power. That assumption is baked into all PoW networks. Secondly, Jude C. Nelson — who you seem to have a professional/academic disagreement with — has a PhD in distributed systems from Princeton. I’m not saying your disagreement with Jude deserves no credibility, but I’m also not prepared to give a pseudonyous commenter on HN the same weight. (JCN runs circles around Vitalik Buterin — a highly entertaining read [1].) The 2013 bitcoin-0.7 BDB chainsplit incident you’ve propped up as “proof” PoW lacks a fork ranking protocol — which it doesn’t — was ultimately resolved with hashpower (of course). Humans only got involved there to prevent needless network downtime, read: “Eventually 24 blocks would be lost”. The incident would’ve eventually resolved itself on-chain by PoW miners, with hashing power only. > I'm not saying [PoS] is better by the way... they're all the same. Proof-of-Stake consensus lacks all notion of cumulative hashrate, and altogether lacks a quantitative fork ranking protocol. To claim PoS and PoW are “the same” in this respect is simply false equivalence. PoS systems undeniably lack an on-chain mechanism for resolving fork disputes. Hence they have little need for a blockchain — after all, if a top-down human hierarchy presiding over a blockchain is going to unilaterally decide which chainfork to favor anyway, they might as well just skip the blockchain and use Git. As with Git repos, in PoS, forking disputes have to be handled by centralized authorities who outright dictate which Git history is valid. No other dispute resolution protocols are possible here. Conversely, PoW systems make this determination by ranking cumulative hashing power amongst forked blockchains. PoW networks coldly and quantitatively come to consensus about what constitutes the longest valid chain. PoS systems use less energy than PoW systems, but they can’t even come to consensus without top-down hierarchical human intervention under adversarial conditions, hence they have little reason to use a blockchain at all. [1]: https://news.ycombinator.com/item?id=25984356 https://news.ycombinator.com/item?id=25984356

hudon 5y ago

> Proof of work = trust physics to determine what happened. This is the kind of vapor Bitcoiners love to inhale and it's just not true. In the history of Bitcoin, it has happened many times where the "work" of a chain was irrelevant because humans disagreed with the contents of that chain. See: 2010 fork, 2013 fork, 2017 fork, etc. When two chains are mined using the same binary, yes, "physics" seems to be in control on which chain wins (ie. the one with the most work). But when one chain expresses a slightly different protocol, humans take over and decide which protocol is the valid one (even if it has less work). And if humans are ultimately deciding on what's valid Bitcoin and what's not, we have to ask ourselves which humans are the most influential in the discussion on what's valid and what's not. We need to know who these players are because with a project worth this much, you can bet that they are targets of powerful and wealthy entities.

atweiden 5y ago
Boring histrionics aside, you haven’t refuted GP’s point. Pure proof-of-stake consensus systems all suffer from the misfeature of not even having a basic quantitative fork ranking protocol. Jude C. Nelson critiques this better than anyone [1]: PoW requires less proactive trust and coordination between community members than PoS -- and thus is better able to recover from both liveness and safety failures -- precisely because it both (1) provides a computational method for ranking fork quality, and (2) allows anyone to participate in producing a fork at any time. If the canonical chain is 51%-attacked, and the attack eventually subsides, then the canonical chain can eventually be re-established in-band by honest miners simply continuing to work on the non-attacker chain. In PoS, block-producers have no such protocol -- such a protocol cannot exist because to the rest of the network, it looks like the honest nodes have been slashed for being dishonest. Any recovery procedure necessarily includes block-producers having to go around and convince people out-of-band that they were totally not dishonest, and were slashed due to a "hack" (and, since there's lots of money on the line, who knows if they're being honest about this?). Under adversarial conditions like those outlined by Jude C. Nelson, PoS reverts to “phone-a-friend consensus”, or — to use a euphemism coined by Vitalik Buterin — “weak subjectivity”. Centralized Git repos offer a close to identical consensus model. What’s the point in having a “blockchain” at all if under adversarial conditions, miners are necessarily demoted in favor of trusted human coordinators? Git is very “green-friendly”, too. [1]: https://news.ycombinator.com/item?id=26810619 https://news.ycombinator.com/item?id=26810619
- hudon 5y ago
  > What’s the point in having a “blockchain” at all if under adversarial conditions, miners are necessarily demoted in favor of trusted human coordinators? If you want to see precisely how "miners are demoted in favor of trusted human coordinators" in Bitcoin, you can read this analysis of the 2013 fork: https://freedom-to-tinker.com/2015/07/28/analyzing-the-2013-bitcoin-fork-centralized-decision-making-saved-the-day/ https://freedom-to-tinker.com/2015/07/28/analyzing-the-2013-... In the quote you posted, Nelson is wrong when he states "[Bitcoin] provides a computational method for ranking fork quality". It only provides such a method when both forks agree on the underlying protocol. If two chains aren't being reconciled because the network is in disagreement on which set of rules is valid, the "computational method" is thrown out the window, and the big guns are called in (high profile developers, influencers in the space, etc.). He uses the words "canonical chain" and "attack" as if it's clear which chain is the attacker and which chain is canonical, but we need humans to help us make that determination on which is which. Chains by themselves aren't hostile or not, they're just numbers. I'm not saying Ethereum is better by the way... they're all the same.
  - atweiden 5y ago
    > In the quote you posted, Nelson is wrong when he states "[Bitcoin] provides a computational method for ranking fork quality". It only provides such a method when both forks agree on the underlying protocol. Firstly, the “underlying [PoW] protocol” assumes the “majority of CPU power is controlled by nodes that are not cooperating to attack the network” (Satoshi, 2009). IOW “Bitcoin” is defined as the chain with the highest cumulative hashing power. That assumption is baked into all PoW networks. Secondly, Jude C. Nelson — who you seem to have a professional/academic disagreement with — has a PhD in distributed systems from Princeton. I’m not saying your disagreement with Jude deserves no credibility, but I’m also not prepared to give a pseudonyous commenter on HN the same weight. (JCN runs circles around Vitalik Buterin — a highly entertaining read [1].) The 2013 bitcoin-0.7 BDB chainsplit incident you’ve propped up as “proof” PoW lacks a fork ranking protocol — which it doesn’t — was ultimately resolved with hashpower (of course). Humans only got involved there to prevent needless network downtime, read: “Eventually 24 blocks would be lost”. The incident would’ve eventually resolved itself on-chain by PoW miners, with hashing power only. > I'm not saying [PoS] is better by the way... they're all the same. Proof-of-Stake consensus lacks all notion of cumulative hashrate, and altogether lacks a quantitative fork ranking protocol. To claim PoS and PoW are “the same” in this respect is simply false equivalence. PoS systems undeniably lack an on-chain mechanism for resolving fork disputes. Hence they have little need for a blockchain — after all, if a top-down human hierarchy presiding over a blockchain is going to unilaterally decide which chainfork to favor anyway, they might as well just skip the blockchain and use Git. As with Git repos, in PoS, forking disputes have to be handled by centralized authorities who outright dictate which Git history is valid. No other dispute resolution protocols are possible here. Conversely, PoW systems make this determination by ranking cumulative hashing power amongst forked blockchains. PoW networks coldly and quantitatively come to consensus about what constitutes the longest valid chain. PoS systems use less energy than PoW systems, but they can’t even come to consensus without top-down hierarchical human intervention under adversarial conditions, hence they have little reason to use a blockchain at all. [1]: https://news.ycombinator.com/item?id=25984356 https://news.ycombinator.com/item?id=25984356