Hacker News Reader

> The security model of Bitcoin comes from proof of work

This is not the security model, this is the consensus model. It's based on probabilistic finality, meaning that the probability that a transaction won't be reversed increases as more blocks are added on top. One major advantage of PoS is that it has "Absolute Finality" - after a certain amount of blocks, it's absolutely impossible to do a 51% attack. (See https://medium.com/mechanism-labs/finality-in-blockchain-con...)

Note that a higher hashrate does not mean more secure, it's a common falsehood. The security of bitcoin depends on the percentage of miners that are honest - this is mentioned the bitcoin whitepaper. Fortunately, the incentives align for the majority of miners to stay honest, and this is what the whitepaper predicted.

Exactly.

In addition, the Bitcoin proof of work is pretty much a proof of stake scheme in practice, where the stake are physical (specialized hardware and access to cheap electric power). So most of the GP's complaints about proof of stakes can be applied to bitcoin as well (“With proof of stake, the stakers only need to make an investment once in the beginning, so it results in rich get richer”).

And the biggest difference between BTC and ETH when it comes to “security” isn't “proof of work“ vs “proof of stake”, it's the fact that Vitali Buterin is alive and that there is an official centralized stewardship of ethereum, which in practice have led them to actually hard-fork the ethereum blockchain.

not only hard-fork numerous times, but also adjust monetary arbitrary policy.

It's just like any other fiat, and of course for such applications PoW is irrelevant.

Proof of work is not a consensus model, it's a method to elect a block author, an incentive mechanism designed to keep miners honest.

Consensus is the process through which the network agrees on state. Examples are Nakamoto consensus (e.g. Bitcoin), BFT (e.g. Tendermind) or GRANDPA+BABE (Polkadot).

I think the GP meant: “the security comes from a consensus, not from proof of work”.

I'm not sure of the argument -

  so it results in rich get richer

In PoW, the same 'rich get richer' applies - miners can buy more mining rigs and thus it compounds the same?

  miners invest heavily in constantly innovating and investing back into the network

In PoS, is this not exactly the same, if not even more true? Stakers are by definition highly invested into the network

  The whole system could collapse overnight like a domino if something goes wrong

Could you elaborate on this, why is this more likely in PoS over PoW?

PoW had decreased decentralization substantially to only happen in regions where electricity is the cheapest. PoS at least prevents this problem.

> so it results in rich get richer In PoW, the same 'rich get richer' applies - miners can buy more mining rigs and thus it compounds the same?

In theory, if and when proof of mining actually becomes the mainstream vehicle for global financial transactions, simply buying more mining rigs won't be enough. They will need to institutionalize and reinvest heavily into their infra. The mining rig that could find ten blocks yesterday may only be able to find one block tomorrow if the competition becomes fierce. Since it's a permissionless system, if one miner starts doing it and makes a lot of money AND Bitcoin gets accepted as a legitimate payment network, other companies will enter the space and the competition will accelerate. That can't happen with PoS.

> miners invest heavily in constantly innovating and investing back into the network In PoS, is this not exactly the same, if not even more true? Stakers are by definition highly invested into the network

The difference is in continuous investment vs. one time investment. With proof of work (again, assuming when these models actually work as designed), you will need to invest more and more into the infrastructure as a miner to stay profitable. As a result the entire network becomes more and more secure. But with proof of stake, the stakers do not have incentive to compete in this manner. With proof of stake, you can literally throw around money to gain influence over a network. With Proof of Work that's not enough because you have to "keep investing", so you need actual commitment to the future. That's much stronger security than a network made up of people who only care about the present.

> PoW had decreased decentralization substantially to only happen in regions where electricity is the cheapest. PoS prevents this problem.

This is like saying "Riding car is dangerous because people can get hit by a car and die. Walking prevents this problem." There's always a solution for every problem. Even the decentralization. The only reason why that hasn't happened is because the Bitcoin network is not worth that much when it comes to its value as a payment system. That doesn't mean there is no solution. For example, using certain clever hashing algorithm, multiple miners can specialize and only do things that each location is optimized for. Mining doesn't just involve hashing. You can generate revenue through including more transactions in a block, for example.

> With proof of stake, you can literally throw around money to gain influence over a network.

But can't you do the same with PoW. You just throw money at minners and buy their rigs or rent them for a good premium.

This guy is a clear bitcoin maximalist and is making exceptions saying that anything is good for PoW and taking that same thing and saying it's bad for PoS.

A Bitcoin miner that can 51% attack the network will need a huge factory full of mining rigs. A proof of stake miner who wants to 51% attack doesn't need that.

... and now you get to my point why I argued ETH2 is likely better on the security front.

Bitcoin needs a huge factory. ETH2 needs a billion dollars. I feel like a billion dollars could probably buy the factory you're talking about.

ETH2 also takes time to add new stakers to the network, but perhaps not as long as building the factory.

If you did this with BTC you'd cause a huge drop it the value of the network, and therefore your investment in miners

If you did this with ETH2 your investment would be destroyed by slashing, and the network might survive

... also keep in mind the Bitcoin factories already exist. You just need to buy one or two, or perhaps hack some of the mining pools.

I prefer the ETH2 semantics.

> Bitcoin needs a huge factory. ETH2 needs a billion dollars. I feel like a billion dollars could probably buy the factory you're talking about.

What you don't get is what comes after that.

In PoS, once you invest a billion dollars, you're set for life. In PoW, you invest a billion dollars but still have to keep investing more and more in order to survive. If you're complacent, a new upstart miner with innovative technology could swoop in one day with 10x more efficient mining method (it could take many forms, such as a closed source bitcoin node that does things much more efficiently yet still fully follows the Bitcoin protocol, or it could be a closed source hardware that they don't sell outside of their own mining operations, like Apple does with their own chips) This way the PoW network as a whole will keep evolving because of competition. In a PoS network there is no competition once you invest the billion dollars and it's a complacent network compared to a PoW network where everyone is constantly competing and comes and goes (those who can't keep up will get left out and leave). If you join a PoS network where someone has a huge stake collectively, you're basically joining a feudalist society where you're being ruled by a king. If you join a PoW network you're being ruled by entities that keep competing to show their competency. it's like the presidential election is going on for every block.

.. that factory full of rigs and a power plant with the capacity of Arkansas Nuclear One

That or maybe just a few colluding mining pools. The network is only 'secure' because mining cartels own most of the specialised hardware, increasing difficultly. ETH 2.0 starts to pool all that power back abit, by allowing others to compete for staking rewards.

All fine in theory, but the rich will find a way. BTC hasn't really been 51% attacked yet, so the attack vector was buy up all the commit devs.

I think you're missing the fact that with PoS, you can use positive _and_ negative incentives to incentivize desired behavior while with PoW, you only have positive incentives. Once you can slash bad actors for enabling a double spend attack, you can significantly change their calculus.

> the stakers only need to make an investment once in the beginning

> The whole system could collapse overnight like a domino if something goes wrong

IMO, these two things are one of the main points of tension which PoS uses to operate. The stakers have put up a significant amount of value, and it's in their interest to maintain that value by preventing things from going wrong.

Not only that, but they are actively pitted against each other -- it's in their interest to find any node that tries to violate the rules, and submit an attestation slashing the offender's stake for bad behavior.

There are also a number of incentives (such as the upcoming EIP-1559 upgrade) which are designed to align incentives so that cartel members are encouraged to break away for their own advantage.

see my comment in sibling thread. with proof of work, it's not just about money. proof of work mining forces miners to invest through future commitment, which means you can't just bring some one time money and gain influence. For proof of stake it's only about the present.

When you buy mining power you are bringing money in to gain influence. There's literally no difference so I'm led to believe there's a bias making you think there is.

what influence do miners have?

> it's in their interest to maintain that value by preventing things from going wrong

It's also in their interest to never sell anyone enough tokens that they could become rival stakers, which makes the system vulnerable to node or network failures. Fewer stakers means fewer nodes to disrupt to trigger consensus failure.