Live data from Hacker News

Proof of stake is incapable of producing a consensus

yanmaani.github.io

261–270 of 822 posts

Re: Proof of stake is incapable of producing a consensus

#261

This is a silly article. Only working in a weaker security model does not, a priori, mean that proof of stake is a scam; it just means you need to convince yourself that the weaker security model holds. You can read the post linked ( https://blog.ethereum.org/2014/11/25/proof-stake-learned-lov... ) and decide for yourself. Personally, I think this kind of "quiescent" knowledge, letting you differentiate the real chai…

> Essentially, VDFs take a certain amount of steps to compute and cannot be parallelized. However, the correctness of their output can be verified efficiently. this...sounds exactly like proof of work? Indeed, you can probably fix plutocracy with some PoW.

> Indeed, you can probably fix plutocracy with some PoW.

Uh... great! Glad you agree that this is a solution?

Let's switch to this new "some PoW" system as soon as possible, because it uses less than a megawatt to calculate proofs. Total, for the entire network.

Re: Proof of stake is incapable of producing a consensus

#262

Earlier quoted context omitted.

> Proof of Work does not get you any votes at all Within protocol, no. But when weighing a fork, those with the gold choose the rules.

The fork-wars of old will never happen again, IMO, on any chain of significance. These days, fork winners will be decided by the stablecoin operations they host. If ETH forks, for instance, hash power will follow whichever chain Jeremy Allaire deems the winner.

So literally one guy has control of Ethereum?

Re: Proof of stake is incapable of producing a consensus

#263
post #34

Is this FUD from Bitcoin maximalists? > That key is valid to sign any number of versions of, let’s say, block #200, and there is no objective, system-internal standard for which version is legitimate, other than “the one that was published first”. The real block #200 will have hundreds of attestations courtesy of randomly-selected validators, each of those signatures attesting to its validity and finality.

Attestations don't have a real world economic cost that can be validated trustlessly within the system. If you compromise or coerce enough validators, you can rewrite the history for no cost. That's what PoW provides that PoS just doesn't. Immutability. In fact, I would argue that one of the most important products of bitcoin, is providing the hardest, most immutable database human civilization has ever created. We c…

"once a piece of data gets confirmed and buried under a few days worth of bitcoin's PoW, it can never be changed or removed from the blockchain."

Could someone put child porn in the blockchain? Would that have legal consequences for anyone using Bitcoin in places that have laws about such porn?

Re: Proof of stake is incapable of producing a consensus

#264
post #246

Earlier quoted context omitted.

And Proof of Work is better because the haves buy equipment the have-nots can't buy to vote? Unless you have a citizenship based voting of some short where a single person gets a single vote and they actually vote (automatically I guess and assuming without delegating to the big whale because "I am bored") what do you think agreement via resource scarcity implies? P.S. Also lobbying...

PoW at least provides a decent and somewhat fair coin distribution mechanism. With PoS, the coin creators can assign themselves an arbitrary fraction of the coins, concentrating the wealth. Even if there is a public record of all the funds raised in a public sale and all expenditures made (which is rarely the case), it's possible for the creators to participate in the public sale and recover large parts of funds used…

Do you believe PoW can be used as bootstrap mechanism, and then some combo of PoS/Pow in the maintenance mode?

Who is the top researcher on this subject these days?

Re: Proof of stake is incapable of producing a consensus

#265

Earlier quoted context omitted.

This is a common misunderstanding. Proof of Work does not get you any votes at all. PoW is a service to the network, to create an immutable ledger. It comes with a very real nuclear option that will bankrupt miners if you misbehave and get fired by hashing algo change. It's just a boring industrial business, like smelting aluminum or iron.

At the limit, proof of work and proof of stake converge - they are the same thing, except for the Kazakh coal that is burned in the former along the way. PoS staking is simply committing a portion of your capital to the task of validating transactions. You benefit by receiving a reward in the form of additional tokens. In a PoW system, the same exact thing can be accomplished by using your tokens to purchase a stake…

> PoW in this context is a bit worse because PoW miners can rent out their hash power maliciously without being slashed.

Sounds wrong. Slashing is a means to prevent people from staking on multiple chains. In PoW, computing power is scarce, so if you allocate some compute time to one chain then you have less of it on another chain. You automatically get slashed. The difficulty in designing a PoS chain is in artificially re-creating this slashing and thereby solving the "nothing at stake" problem.

Re: Proof of stake is incapable of producing a consensus

#266
This article is complete BS. Proof of Stake is more secure than Proof of Work for a simple reason. The cost of doing a 51% attack (to stop the blockchain or to start censoring specific transactions) on a PoS blockchain is exponential, whereas the cost of doing such an attack on a PoW network is linear. This is because as an attacker acquires more tokens, the price of remaining tokens increases exponentially as the attacker approaches the 50% mark. If the network is well decentralized in terms of token ownership, it may not ever be possible for the attacker to acquire 50% of tokens; also, their incentive to continue with the attack decreases as their stake in the blockchain increases. Unlike with PoS which requires the attacker to keep buying more (limited-supply) tokens, with PoW, ASIC miners don't become more expensive as the attacker gets closer to having 51% of the hash power; this is because the market will produce more ASIC miners to compensate for any increase in demand. The global supply of ASIC miners has no upper bound.

The article is also misleading in inferring that there is a very narrow range of ways to implement PoS; in reality, there are many ways and all of the 'drawbacks' mentioned only apply to certain (poorly designed) implementations which no modern PoS blockchain would ever use.

> What happens if you’re presented with two identical blocks, and have to decide which one to pick?

Easy, you can just have a vote on one of the block and choose the one with the majority votes; it can be chosen on the basis of any attribute of the block (E.g. commonly you can look at block IDs). This is what PoS blockchains like COSMOS do with the Tendermint protocol. Other blockchains like Lisk have a delayed voting so that consensus is reached after a certain number of blocks.

> The entire point of the consensus mechanism was to allow us to tell which transaction was first, without personally having seen it take place.

Anyone who understands distributed systems knows that the exact order of transactions (down to a few hundreds of millisecond) cannot be physically determined due to latency between the nodes and the unpredictable geography of participants. This is as true for PoW as it is for PoS. The most important thing (for certain use cases such as DeFi) is that transactions cannot be predictably front-run; using block ID ordering with voting as the basis for selecting between two valid blocks guarantees this. If the forger tried to cheat the system by producing multiple blocks, the network may not be able to reach consensus on the block vote and the forger would not receive any block rewards.

Re: Proof of stake is incapable of producing a consensus

#267
post #102

Earlier quoted context omitted.

Attestations don't have a real world economic cost that can be validated trustlessly within the system. If you compromise or coerce enough validators, you can rewrite the history for no cost. That's what PoW provides that PoS just doesn't. Immutability. In fact, I would argue that one of the most important products of bitcoin, is providing the hardest, most immutable database human civilization has ever created. We c…

I have to say I appreciate your transparency in directly Bitcoin shilling. > If you compromise or coerce enough validators, you can rewrite the history for no cost. If you can. You would need to compromise thousands of randomly selected validators just to forge one block. That impossible task nonwithstanding, the validators are selected with maybe five minutes’ warning. PoW doesn’t even offer absolute immutability, i…

Exactly. These PoS FUDers fail to realize that largely the same, even if not stronger, economical factors protect the PoS as does the PoW blockchain. Both fundamentally rely on that 50% of validators need to be honest. With PoW, anyone with sufficient computational resources can execute a 50% double spending attack. I am sure that if few of the largest BTC miners conspired, they could right now execute a 50% attack. That's not in their best interest however, as it would permanently ruin the reputation of BTC and they'd be left with whatever proceeds the attack yielded, but also millions of USD worth of mining equipment that'd be rendered largely useless.

With PoS exactly the same works. Most of the holders should be honest, and a successful attack would require spending at least 50% of the market cap of the coin to successfully execute - and then losing that stake as the reputation of the coin is soiled.

Re: Proof of stake is incapable of producing a consensus

#268
Whether PoS will work, I don't know. But the author didn't realize that PoW is certainly doomed.

PoW miners tend to spend more and more resources on finding blocks, until the cost approaches the rewards. But the rewards go up as the cryptocurrency becomes more popular, because the price and transaction fees go up. Therefore, a PoW cryptocurrency tends to "eat the world" as it becomes bigger.

That's why Bitcoin is already approaching 1% of global electricity consumption, if it hasn't passed that point already. If the price were to go up tenfold, then so would electricity usage (roughly). That's not sustainable, both technically for grids and economically because electricity prices go up.

Because of that, I foresee two possible futures for PoW cryptocurrencies:

1. The resource usage overshoots and PoW collapses because it gets banned everywhere. (This seems to be playing out now with China having banned crypto mining, Kazakhstan running into grid issues because of the miner influx, and Sweden arguing for a ban in the EU.)

2. The popularity of these currencies stops growing and only some niche applications remain. Speculators leave because there's no more money to be made. Prices go down.

Re: Proof of stake is incapable of producing a consensus

#270

Earlier quoted context omitted.

There is a difference between each of those uses of electricity and PoW. PoW is throwing away electricity for the sake of it, and resists getting more efficient. If the goal is for the bitcoin network to cost $1M to do a single double spend, then PoW has to use $1M worth of electricity every 10 minutes. Let's say we live in a future where we suddenly have 10x as much electricity. Due to supply and demand, electricity…

bitcoin incentivizes the search for cheaper energy, and allows for the instant monetization of once wasted energy. That's a LOT. this whole argument of energy use... i bet most of these people would have fallen for the same energy propaganda when the internet was just getting started... "what's email?? why would you waste electricity to send an electronic mail??"

There's a trend now of Bitcoin mining businesses buying fossil fuel power plants in the US, in many cases plants that had already been shut down due to being unprofitable.

https://futurism.com/bitcoin-mining-company-buys-entire-coal... https://grist.org/technology/bitcoin-greenidge-seneca-lake-c... https://arstechnica.com/tech-policy/2021/09/old-coal-plant-i...

Post reply on HN