Proof of Work is the only viable mechanism for decentralized state | Hacker News Reader

Proof of Work is the only viable mechanism for decentralized state

lists.w3.org

24 pointsVinnl22 comments

In a Proof of Stake system, money directly, literally translates to power over the network. There is virtually zero risk to mining, and there is zero overhead cost. As such, the centralization of Proof of Stake is inevtiable: coin staked begets coin, which begets more coin.

This is principally no different to Proof of Work in practical terms. More resources gets you more power, and this is inescapable in any structure. More money begets more mining hardware, which begets more coin, which begets more mining hardware.

With PoW there is the element of specialized chipset advantages through R&D. I.e. I will always be disadvantaged because the SHA-256 ASICs available to me will always be subpar to whatever the big players have.

With PoS 1 of my coins carries same vote power as 1 of your coins. You may have the ability to stake more coins, which will reward you more, but our returns (ROI) will always be proportionally equal.

The argument has nothing to do with more resources resulting in more power, of course that is the case. The question is how does one acquire those resources and maintain them. In proof of work, resources need to be derived and maintained external to the system... who can produce the fastest processors, who can produce the cheapest electricity, who can maintain the more efficient data centers, etc etc... This comes with big risk and require constant upkeep.

In proof of stake, the person with the most resources is the person with the most coin and maintaining that not only costs nothing... maintaining it produces more of the very resource.

Are we optimizing for a fair playing field? Or a playing field where one can be advantaged?

If my authoritarian government bans mining, my energy usage is easy to track down, and I am disadvantaged. If my nation state has an abundance of cheap hydro energy, I am advantaged (and miners flock here to take advantage, centralizing mining into this geographic area).

One other thing: for Bitcoin specifically, the energy used by proof of work is guaranteed to drop every 4 years at the block reward halving, until it reaches a cost which is no more than the value users derive from their use of the network.

After the block reward dwindles to nothing, the only money made by miners is paid directly by users in the form of transaction fees. Users will not pay more than the value they derive from their use of the network, and miners cannot buy more electricity than is funded by their income.

I think it is still an open question whether users will pay enough to make the chain secure from attack in this inevitable scenario. So we may end up with miners using less energy than they should. Using too much is unlikely. The current state of excess is temporary.

Don't we track transaction security by chain length as well as new block generation speed? Those should let you determine the probability of your transaction being rejected no matter the mining incentive.

The probability you care about is the probability that your transaction will be reversed by chain rewriting. The cost of chain rewriting is the amount of work it would take to make a longer chain starting from the block before your transaction. However, chain length is simply a proxy for the amount of work done, because each block has the same difficulty (except at a difficulty adjustment).

The number of blocks or speed of blocks is not the relevant measure. The number that matters is the total accumulated work proved in the blocks. Dogecoin has faster blocks and more blocks than Bitcoin but it is not more secure. Its difficulty is much lower, so each block proves much less work, and rewriting its chain is therefore easier.

I didn't mean to compare speeds for comparing difficulty for two different currencies, I agree that's largely meaningless.

Rather, the key probabilities I would like to compute would the probability of chain rewriting (I don't like this term tbh; if someone forks off a new chain they haven't rewritten the one your block is in, rather they have "convinced" others to abandon it) within the next T days given an adversary with more compute than the Bitcoin network by a fraction f, for various T and f.

Can't I approximate that with 1) the lead that my current chain has over the next competing chain, 2) the average time it's taking to mine a new block given the current compute?

Ill confess I'm not super familiar with the intricate details of crypto, I may just be missing something obvious.

You measure the "lead" of a chain by the amount of work it proves, not by the number of blocks or speed of blocks. If you based the lead solely on block number then an attacker's chain could manipulate the difficulty adjustment to make the attack easier.

> Can't I approximate that with 1) the lead that my current chain has over the next competing chain

An attacker could keep their chain secret until the attack is sprung. You cannot know how far ahead or behind you are relative to a secret chain. What you know is the amount of work an attacker would have to do, by summing the work proved by all the blocks above and including yours. Once that value gets high enough, you can be pretty confident that nobody is going to rewrite your transaction. The probability that you would have to calculate is the probability that an attacker who has that much compute exists.

Thanks! This is good food for further reading, I had no idea miners could influence difficulty this way. I had thought, for Bitcoin, difficulty depends on how many coins mined till now so I (implicitly) assumed that small parts of the chain would have similar difficulty.

Someone on Reddit recently explained PoW vs PoS:

Besides POW there are other ways to delegate control in a decentralized system. Actually, for a smaller network, POW is not a good consensus mechanism at all, because it is not prohibitively expensive to dominate the hashpower of a small network. Many altcoins have been successfully attacked like this. So how to decide who controls the network?

Proof of Stake (POS) is probably the most common alternative. In POS there is no electricity cost to mining, there is no hash puzzle to solve. Actually there is no mining and no miners, only validators. The users who own the most coins are allowed to validate the most blocks. To attack the network you have to buy 51% of the coins and the theory is that you wouldn’t want to attack your own asset once you own that much. The “staking” refers to the fact that the validator has to lock up some of his coins and promise not to use it while he remains a validator. If you own 3% of the total staked coins, you are given 3% of the blocks to validate. Because validating a block has a reward, this means you also get 3% of all new coins.

Part of the criticism of POS is that the rich will get richer and the poor will get poorer because the more coins you own, the more blocks you will validate and the more coins you will get as reward. This means you can stake even more coins and get even more rewards, and so on. This POS mechanism gradually consolidates most coins into only a few hands and leads to centralization. You could ask how is this different with POW. The miners with the most mining rigs validate the most blocks and gets the most rewards and accumulates more and more coins?

The truth is that most miners have to sell their Bitcoins. Mining is not a passive investment like POS because there are ongoing running costs. A miner must not only pay cold, hard cash for the electricity to make a block, but also invest in expensive equipment, maintenance, salaries and so on, put in actual work. Would you rather pay someone who can prove he did some work, or someone who can prove he is wealthy?

Source: https://www.reddit.com/r/Bitcoin/comments/pfryce/pow_vs_pos/...

That's well argued and gives me pause. Anyone have a good counterargument?

> Proof of Stake, by its very design, centralizes over time.

I don't see how. If I have 10% of total supply and you have 20%, means I get 10% of mined coins an you get 20%. This means we will stay at the same percentage over time, no?

I think you're roughly correct. 49% doesn't automatically become 51% purely through the PoS process. But if the owner of 49% of the coin goes out and buys an additional 2% on the market, then they have a majority of the coins and it's all over.

It's not guaranteed to happen, but does seem plausible that some single entity could end up with 51% sometime in the future. Or a group could band together to form a majority and steal everyone else's money.

It does become harder to prevent a coup over time. As more coins are issued the cost to buy a 1% stake increases but the cost to maintain a 1% stake does not.

Isn't this true of every "Proof of X" algorithm?

I think the point is that the with PoS, you're proving you have the coin itself, which costs nothing once you have the coin. Changes things a bit.

People with less money tend to spend a higher percentage of it than people with more money. Given that, there is a threshold where people under it will have their wealth decrease over time, and people over it will have their wealth increase over time.

But the barriers to entry are much higher with proof of work and then I'm still at a disadvantage as the economics of scales heavily favor the big players.

With proof of stake I can start with almost any amount and I'm not at a disadvantage.

The important difference is that proof of work is a competition. In a steady state, competition will push costs up until they meet revenue, driving profit to zero. In contrast, proof of stake is not competitive and has very little risk or cost. It's practically pure profit, and the wealthier you are the more you make.

Proof of work miners make profits today, but this is a temporary situation caused by the large block reward which will fall, and large price increases which cannot last forever. Proof of work mining is not a risk-free way for the rich to get richer in the long term, the way proof of stake is. It is fundamentally a low margin competitive business. No better than any other business the rich might pursue with their capital, and worse than many.

> Proof of work miners make profits today, but this is a temporary situation caused by the large block reward which will fall, and large price increases which cannot last forever.

But that is the problem I talked about. Already today the entry into the mining business is costly. Not everybody can do that. There are already centralisation concerns with Bitcoin miners and the pressure will only increase in the future due to economies of scale.

> In contrast, proof of stake is not competitive and has very little risk or cost. It's practically pure profit, and the wealthier you are the more you make.

I disagree. But the competition lies elsewhere than with PoW. As with PoS you are purely securing the blockchain with capital the competition lies with other investments.

With Ethereum 2.0 the more ETH is staked the less interest is earned by the stakers (https://docs.ethhub.io/ethereum-basics/monetary-policy/#proo...). This means for big players (the wealthy stakers) their stakes need to yield as much ROI as other financial instruments. If their stakes falls below that they will take their ETH elsewhere.

For less wealthy stakers who have less easy access to financial instruments they can continue staking with less yield. Also staking is pretty easy, it needs to be, so that enough small stakers can do it and we have enough decentralisation.

Work=Force(*distance) measured in energy units is usually concentrated (pressure). The probability distribution of default may be correlated, but is not same as, the distribution of energy.

Well written and totally agree with these statements never heard of proof of capacity before will need to look into it

Look at chia.net

Reply on news.ycombinator.com