Can someone explain to me how PoS is not centralized? The whole point of crypto and the blockchain is decentralization, no? I feel like crypto/blockchain will end up looking nothing like how it was intended, and be regulated to the point that it's just the next iteration of the traditional banking system, where the current power holders continue holding the power. Maybe I was naive to ever think it would be any different.
How is it more centralized than PoW? I see multiple factors that hamper decentralization:
- Fixed costs that act as barrier of entry
- Economies of scale that lead to centralization
- Geographic factors (operation costs being different in different parts of the world, regulation/taxation, supply chain...)
This is how I see each factor playing out in both scenarios:
- Fixed costs: PoS runs on consumer-grade hardware, while PoW requires specific HW (ASICs or high-grade GPUs). PoS requires a minimum amount of stake but there are pooling solutions, which effectively make this minimum non existent. All in all PoS is at advantage here, unless you want to insist on solo staking in which case PoW is at advantage.
Analogy: This would be equivalent to flat fees to open a savings account or a minimum amount balance required to open it.
- Economies of scale: In PoS they are almost non-existent. You don't stake more efficiently by having a more powerful machine. You just get to reuse the same HW for more nodes but since fixed costs are low this has a very small impact. In PoW there are economies of scale, though, better/more expensive ASICs can mine more efficiently than smaller/cheaper ones. Same with GPUs. Someone with more initial capital can get ahead faster in PoW, while in PoS earns at a same rate as everyone else.
Analogy: This would be equivalent to the interest rate you get in your savings account being dependent on how much money you have. In PoW, the richer you are the higher interest rate you get from your bank, in PoS everyone gets the same.
- Geographic factors: Cheap access to energy has a large impact on PoW as it dictates most of your OpEx. In PoS this is largely irrelevant (PoS is 99.95% more energy efficient than PoW). Taxation/regulation would need its own analysis but I imagine is equally spread across both alternatives. Supply chain is again in favor of PoS as it can run on general-purpose HW, while ASICs are heavily centralized around a single manufacturer.
Analogy: This would be equivalent to different geographic locations resulting in different conditions for maintaining open your bank account or taxing your accrued interest.
In practice exchanges end up holding most of the crypto, and they will decide how to use the crypto in PoS. With PoW it looks like miners/pools will quite rarely work as exchanges, so the powers tend to be separated. With PoS, it looks like exchanges will basically run Ethereum, and it look something like Blockstream Liquid, where group of exchanges basically mint the blocks.
This is exactly right. Too many folks assume PoW is more decentralized because it's not tied to money, but what's the difference if mining requires money? As you note, PoS lowers the barrier to entry and actually creates a more egalitarian system, despite with Bitcoin maximalists think.
Can't you just stake more coins in PoS and receive more income? I though this was how PoS coins worked. How else would you even have a notion of people requried to satisfy the condition "all people get the same amount"? Surely not photo ID?
May I ask you to elaborate a bit? What is they are centralized around? Capital? Misbehaving capital can easily be destroyed in PoS by a fork. Please see Steem/Hive case.
- My money is secured by a private key and nobody can touch it ever unless they get my private key. I can simply be cautious about my private key and I know there is no other way anyone can ever get my money
- I can fill out a text box with the amount I want to send and press send
- I don't have to deal with some harebrained naming system. I can literally just send to someone's public key. This is literally how cryptography is inteded to work. It's up to _me_ how I obtain his public key.
- Monero etc exists (admittedly, not sure if it still works when someone has all the mining/stake power)
- Some guy is getting rich because he owns more miners or stake
Fiat:
- My password is 8 digits (this is not even an exaggeration, some of the biggest banks in my country do this)
- The bank might give all my money away if someone knows where I ate KFC last
- My money may be stolen for other reasons, because the bank wont tell me what data I need to keep private to avoid having someone transact as me
- The ID they use for authentication was also given to some 30 other e-commerce platforms and cannot be considered secure
- There is almost certainly a way to get into my account without the password
- I have to be paranoid and try to keep random trivia private such as how much I payed on an electricity bill
- I have to type codes from insecure SMS on a phone that I do not want in the first place, because the bank and all e-commerce platforms considers me an idiot and does not even give me an option to turn that shit off
- If I transfer from one country to another, my transaction may be blocked
- If I transfer some certain amount, my transfer may be blocked
- If I use a certain IP address, my transfer may be blocked
- If I transact at a certain time, my transfer may be blocked
- If I update Firefox too fast or too slow, my transfer may be blocked
- If I click buttons to fast or too slow, my transfer may be blocked
- Someone might hack my computer because it has a Big 4 web browser and the giant stack of software required to support that, instead of a hypothetical OS where people care about security and don't use C, at the cost of some microseconds.
- If I change my email address (which I don't want associated to banking in the first place) for some reason, my transfer may be blocked
- When I call the bank, I have to be polite and try to avoid saying anything suspicious (in their own mind) that will make them hold my money yet longer. I will have to supply them will all kinds of nonsense like where I ate KFC last, more ID, and a "phone password"
- My transactions may be permanently blocked and there's nothing I can do about it because the bank reps just talk to a black box "risk analysis" machine and at some point there's no way to override its decisions
- Money I receive can be "reversed" for all kinds of bogus, emotional, and/or "risk analysis" reasons
- The bank can just take my money and claim I was hacked. They have N pieces of my photo ID, address, phone number, email, and much more, and so they can choose a few people they don't like and do this to only them
- I have to interact with my bank through web pages that crash every 3 button clicks, and PDF files that may or may not render correctly (or snail mail, which is equally full of bad security)
- Some guy is getting rich because he's positioned a certain way with the bank
TL;DR even if the top cryptocurrencies were effectively centralized by one entity controlling all miner/staking power, I would still want to use them at least for transacting, just so I can have a sane interface to money.
The key distinction is that PoW is permissionless, whereas PoS is permissioned.
Bitcoin is secured by hashpower, which is produced by physical capital outside the network. Nobody needs to ask for permission to start hashing and trade kilowatts for sats.
PoS networks are secured by on-chain assets. This means you can't "mine" it without first buying tokens from someone who already owns them. You need permission from an existing player in order to start participating.
Another aspect of this is 51% attacks are recoverable for PoW, but are a permanent takeover condition for PoS networks. If a single entity ever accumulates more than half the tokens on a PoS network, they are unassailable.
> You need permission from an existing player in order to start participating.
This is an incorrect explanation of what a permissioned blockchain is. A permissioned blockchain is one in which the ability to add blocks is limited to a certain collection of entities whose public keys are hard coded into the blockchain's consensus mechanism. We don't say that needing to buy tokens constitutes needing "permission" any more than you need permission from a chip manufacturer to buy ASICs to mine a PoW cryptocurrency.
>Another aspect of this is 51% attacks are recoverable for PoW, but are a permanent takeover condition for PoS networks. If a single entity ever accumulates more than half the tokens on a PoS network, they are unassailable.
This is not true. PoS has many design flavours and the one Ethereum is planning on implementing includes random selection of validators and the amount staked has no influence on the inclusion or the vote "weight".
Also with PoS an attacker will always incur economic losses similar to having your mining rig burning down if you were to try to foce a bad block through. In PoW networks attackers can keep on mixing attacks with producing normal blocks and remain profitable
> Another aspect of this is 51% attacks are recoverable for PoW
You can switch the protocol once. Making ASICs useless. But you can't do it twice.
> but are a permanent takeover condition for PoS networks.
This is false in both theory and practice. It is true that PoS does not offer in-protocol solution for the problem. But there is a historical precedent of people forking away money of Justin Sun in Steem project, creating Hive. Community has followed the fork, basically destroying Justin's Sun funds.
It seems like this argument proves too much for your purposes, in the sense that it can be used to show that neither algorithm is any good as far as distributed governance is concerned.
While it’s true that you can’t buy Bitcoin (for example) unless someone else is selling, most people aren’t concerned about market liquidity for buyers due to whales being unwilling to sell. The permission to buy doesn’t seem hard to get?
Also, for the most part, people are happy when the price goes up, which is what happens when there are more buyers than sellers.
I guess in theory, money drops could distribute ownership more widely and that would be more equitable, but this sort of inequality (some people have a lot more Bitcoin than others) isn’t normally considered too much of a problem.
But if you’re going to take distributed governance seriously, neither proof-of-work nor proof-of-stake give ordinary people much of a say in how things go. In this way it’s similar to the stock market, where we’re told our votes are meaningful but in practice they aren’t unless you have a huge amount of shares. Participating in governance is usually an illusion and it’s not normally why you invest, unless you’re a corporate raider or something.
>>PoS networks are secured by on-chain assets. This means you can't "mine" it without first buying tokens from someone who already owns them. You need permission from an existing player in order to start participating.
Only in the most pedantic sense of 'permission'. There will always be thousands of disparate parties, across numerous markets, with offers to sell their ETH. It will never be harder to procure ETH than to procure hash-generation hardware.
If somebody does a 51% attack his tokens would be worth nothing the next day. Might be plausible if you target smart contracts that are worth a lot more than the coin itself.
The first stocks ever sold to the public were claims on land in a new world, an escape from servitude to kings. Turns out they were just being swindled by a bankrupt crown to keep their own oppressors in power. But there was a new world, and people did escape.
How many stakers actually have enough to win blocks though.
What is the impact of shorting on PoS - what happens when I borrow enough ETH to win blocks, deliberately mis-verify TXs, and screw up consensus.
I know Ethereum has planned recovery for situations like this, but PoS introduces risks that will never be as present in PoW bc the latter has built-in latency to how easy it is to aggregate resources which increase market power and centralize around certain miners (ie buy asics and build a data center takes time, so you just have to watch mining pools for liquid malicious hash rates for equivalent risks in PoW).
But that's not peer to peer. And people compare bitcoin/ ethereum to gold, but you can't really exchange ethereum without a centralised authority when you use POS. If these people refuse to process your payment, there is not much you can do about it.
What does it matter what the technology supports if it incentivizes the community to consolidate? Penalties for being offline and slashing (for misconfigured nodes) are a massive incentive for nodes to centralize
Capital is capital, be it in the form of money or hardware. Putting the onus on capital in a form of hardware is just making it harder for smaller guys. I can throw 32 ETH into staking. It's way beyond my capability to mine - wrong geography, electricity prices, accommodation situation etc.
PoS is decentralized as soon as you have a big number of stakers. The hard thing being “how do you know that there are in fact a big number of people and not just a bunch of guys hidden behind a huge number of aliases” (the Sybil attack problem) and this is for this exact problem that proof of work helps (contrary to popular belief, Proof of work has nothing to do with double spending).
For this reason you cannot build a trust-less distributed network with proof of stake alone: you'll need some kind of proof of waste to bootstrap it.
PoS should be less centralized than PoW as I understand it. At least in Ethereum if you stake your 32 ETH then you get transactions to validate assigned to you randomly. You don't need a purpose-built rig with a ton of hash power to compete so smaller players are incentivized to participate, leading to a far greater incentive to run a node.
Because anyone in the world can participate in validation. It may be costly to do so, or whatever other hoops, but there's no "central" authority managing participation.
This is not a rant for or against Bitcoin or Etherium. It's more to point out the persuasive and idealistic rhetoric of crypto gets confused with what may happen in the real world as people work on use cases.
for proofs of "whatever", where whatever is something that can be accrued with money, then it's proof of money, and money is centralized, or at least reallllllly spikily concentrated, so, yeah
Note that execution shards are not really on the roadmap anymore as sharding is planned to be data only and serve as a scaling factor for L2 rollups.
In just a few years the only usage of the L1 will be security and data availability for rollups and a settlement layer for international organizations and state level actors with the vast majority of normal activity happening on various L2s.
If Ethereum can hard fork like this, it stands to reason as well that the 21 Million Bitcoin limit is a lot more changeable than Bitcoin proponents would say.
Or the opposite is true - and Ethereum is under heavily centralized control.
Ethereum is so enthusiastic about hard forks (or in ETH speak, upgrades) that they have the difficulty bomb which works like the debt ceiling. It's necessary to do a hard fork/upgrade every so often.
Meanwhile bitcoin is so against hard forks/upgrades that they can basically only do soft forks and currently do that by a 90% voting agreement by miners.
Bitcoin can not even increase the blocksize, the sure as hell aren't going to change the emission. Some day far down the line the block subsidy might be low enough that lots of miners go dark. At that point they'll need to choose between 21 million bitcoins and network security.
So the personal beliefs of users decide the technical future of the project. Personal beliefs and culture can change rapidly, or become overshadowed by new users.
It just means that Bitcoin is only unchangeable because of the current whims of current users - nothing fundamental.
All blockchains are controlled by rough consensus. There's already like half a dozen Bitcoin forks in the wild (Bitcoin Cash, BSV, etc). When Bitcoin Cash forked off from the main chain in 2017, the community ultimately decided which chain was the "legitimate" one.
Of course Bitcoin could decide to increase the issuance. In doing so, however, it would inevitably lead to a fork (with the 21M holdouts in one camp and the inflationists in another) and the community would ultimately need to decide which one to support.
Imo the block size wars was the beginning of the end for Bitcoin. I chose the chain most hashed and it's only now that I realize that Bitcoin gave up on growing and improving at that point.
Bch is a joke, PoW just doesn't work unless you're the biggest game in town. BTC is shackled to staying incredibly conservative.
I agree with all of this - but the conclusion I reach is that proponents of Bitcoin are lying when they say it's unchangeable and certainly deflationary, or digital gold. It's changeable if people want it to change.
No it doesn't. You're comparing something with no advantage to the people who are invested in Bitcoin to something which is meant to objectively improve Ethereum and a majority can agree upon.
I always heard it explained that it's full-node wallet users who tend to control the future direction of a blockchain, not just those invested or mining.
I would like the 21M limit to be increased so I can stand to make some profit mining! Currently that's impossible for me. I'd like it to change.
I asked this months ago and did not get any convincing answers.
Ethereum is currently decentralized because of the initial POW distribution. Won't there be centralized aggregators of eth so some point in the future a handful of POS nodes control a disproportionate amount of power? Is it so hard to imagine that coinbase or some other exchange accumulates enough eth to sway transaction validation?
>For certain kinds of 51% attacks (particularly, reverting finalized blocks), there is a built-in "slashing" mechanism in the proof of stake consensus by which a large portion of the attacker's stake (and no one else's stake) can get automatically destroyed.
To be honest, even without the actual 51% attacks happening, it would really suck to live in a world where just one entity had more than half ownership of a currency used by the people.
As I understand it a one way to disincentivise centralization is that when your node loses access to the network you lose more money if at the same time many other nodes/stake do not have access to the network. So it would be profitable to have an independent server and internet access.
> Ethereum is currently decentralized because of the initial POW distribution
Ethereum was launched with a 100% 'premine' of 72M coins, printed out of thin air and handed to the founders to keep or sell as they decided.
Since then another 44M have been mined, though it's not clear how much of that mining was also by insiders who influenced the work function's hardware compatibility or protected their hardware investment by controlling when/if the work function was changed or when mining was discontinued.
I don't think that threat ever really materalized. MEV happened and then there was a suggestion by VB of switching to the PoS chain in response to a miner lead fork: https://notes.ethereum.org/@vbuterin/B1mUf6DXO
- Fixed costs that act as barrier of entry
- Economies of scale that lead to centralization
- Geographic factors (operation costs being different in different parts of the world, regulation/taxation, supply chain...)
This is how I see each factor playing out in both scenarios:
- Fixed costs: PoS runs on consumer-grade hardware, while PoW requires specific HW (ASICs or high-grade GPUs). PoS requires a minimum amount of stake but there are pooling solutions, which effectively make this minimum non existent. All in all PoS is at advantage here, unless you want to insist on solo staking in which case PoW is at advantage.
Analogy: This would be equivalent to flat fees to open a savings account or a minimum amount balance required to open it.
- Economies of scale: In PoS they are almost non-existent. You don't stake more efficiently by having a more powerful machine. You just get to reuse the same HW for more nodes but since fixed costs are low this has a very small impact. In PoW there are economies of scale, though, better/more expensive ASICs can mine more efficiently than smaller/cheaper ones. Same with GPUs. Someone with more initial capital can get ahead faster in PoW, while in PoS earns at a same rate as everyone else.
Analogy: This would be equivalent to the interest rate you get in your savings account being dependent on how much money you have. In PoW, the richer you are the higher interest rate you get from your bank, in PoS everyone gets the same.
- Geographic factors: Cheap access to energy has a large impact on PoW as it dictates most of your OpEx. In PoS this is largely irrelevant (PoS is 99.95% more energy efficient than PoW). Taxation/regulation would need its own analysis but I imagine is equally spread across both alternatives. Supply chain is again in favor of PoS as it can run on general-purpose HW, while ASICs are heavily centralized around a single manufacturer.
Analogy: This would be equivalent to different geographic locations resulting in different conditions for maintaining open your bank account or taxing your accrued interest.
That's to me similar, if not worse than with PoW.
They're distributed, not decentralized. And this distinction, and the refusal to acknowledge it is what put me off crypto entirely.
Bitcoin is secured by hashpower, which is produced by physical capital outside the network. Nobody needs to ask for permission to start hashing and trade kilowatts for sats.
PoS networks are secured by on-chain assets. This means you can't "mine" it without first buying tokens from someone who already owns them. You need permission from an existing player in order to start participating.
Another aspect of this is 51% attacks are recoverable for PoW, but are a permanent takeover condition for PoS networks. If a single entity ever accumulates more than half the tokens on a PoS network, they are unassailable.
This is an incorrect explanation of what a permissioned blockchain is. A permissioned blockchain is one in which the ability to add blocks is limited to a certain collection of entities whose public keys are hard coded into the blockchain's consensus mechanism. We don't say that needing to buy tokens constitutes needing "permission" any more than you need permission from a chip manufacturer to buy ASICs to mine a PoW cryptocurrency.
This is not true. PoS has many design flavours and the one Ethereum is planning on implementing includes random selection of validators and the amount staked has no influence on the inclusion or the vote "weight".
Also with PoS an attacker will always incur economic losses similar to having your mining rig burning down if you were to try to foce a bad block through. In PoW networks attackers can keep on mixing attacks with producing normal blocks and remain profitable
You can switch the protocol once. Making ASICs useless. But you can't do it twice.
> but are a permanent takeover condition for PoS networks.
This is false in both theory and practice. It is true that PoS does not offer in-protocol solution for the problem. But there is a historical precedent of people forking away money of Justin Sun in Steem project, creating Hive. Community has followed the fork, basically destroying Justin's Sun funds.
While it’s true that you can’t buy Bitcoin (for example) unless someone else is selling, most people aren’t concerned about market liquidity for buyers due to whales being unwilling to sell. The permission to buy doesn’t seem hard to get?
Also, for the most part, people are happy when the price goes up, which is what happens when there are more buyers than sellers.
I guess in theory, money drops could distribute ownership more widely and that would be more equitable, but this sort of inequality (some people have a lot more Bitcoin than others) isn’t normally considered too much of a problem.
But if you’re going to take distributed governance seriously, neither proof-of-work nor proof-of-stake give ordinary people much of a say in how things go. In this way it’s similar to the stock market, where we’re told our votes are meaningful but in practice they aren’t unless you have a huge amount of shares. Participating in governance is usually an illusion and it’s not normally why you invest, unless you’re a corporate raider or something.
Similarly for mining. It’s done to make money.
Only in the most pedantic sense of 'permission'. There will always be thousands of disparate parties, across numerous markets, with offers to sell their ETH. It will never be harder to procure ETH than to procure hash-generation hardware.
Otherwise this is just a nonissue.
From sea to shining sea, the new world's plundered resources eventually turned into Bitcoin.
What is the impact of shorting on PoS - what happens when I borrow enough ETH to win blocks, deliberately mis-verify TXs, and screw up consensus.
I know Ethereum has planned recovery for situations like this, but PoS introduces risks that will never be as present in PoW bc the latter has built-in latency to how easy it is to aggregate resources which increase market power and centralize around certain miners (ie buy asics and build a data center takes time, so you just have to watch mining pools for liquid malicious hash rates for equivalent risks in PoW).
Which means, whether you can stake is very dependent on your personal situation - especially geography - much like whether you can profitably mine.
Of course staking is still better than mining for many reasons. But it's a somewhat rich person's game.
For this reason you cannot build a trust-less distributed network with proof of stake alone: you'll need some kind of proof of waste to bootstrap it.
But you do need ETH in the first place, and the more the better. This seems like it still disincentivizes smaller players.
This is not a rant for or against Bitcoin or Etherium. It's more to point out the persuasive and idealistic rhetoric of crypto gets confused with what may happen in the real world as people work on use cases.
All these points for/against PoS/PoW are technicalities at this point given that "decentralization" is a spectrum.
Dead Comment
Pretty impressive.
I wanted to swap a bunch of coins worth 8$ into another token and the gas fees were around 28$, which is just insane.
In just a few years the only usage of the L1 will be security and data availability for rollups and a settlement layer for international organizations and state level actors with the vast majority of normal activity happening on various L2s.
Deleted Comment
Deleted Comment
Or the opposite is true - and Ethereum is under heavily centralized control.
Meanwhile bitcoin is so against hard forks/upgrades that they can basically only do soft forks and currently do that by a 90% voting agreement by miners.
Bitcoin can not even increase the blocksize, the sure as hell aren't going to change the emission. Some day far down the line the block subsidy might be low enough that lots of miners go dark. At that point they'll need to choose between 21 million bitcoins and network security.
It just means that Bitcoin is only unchangeable because of the current whims of current users - nothing fundamental.
Of course Bitcoin could decide to increase the issuance. In doing so, however, it would inevitably lead to a fork (with the 21M holdouts in one camp and the inflationists in another) and the community would ultimately need to decide which one to support.
Imo the block size wars was the beginning of the end for Bitcoin. I chose the chain most hashed and it's only now that I realize that Bitcoin gave up on growing and improving at that point.
Bch is a joke, PoW just doesn't work unless you're the biggest game in town. BTC is shackled to staying incredibly conservative.
No it doesn't. You're comparing something with no advantage to the people who are invested in Bitcoin to something which is meant to objectively improve Ethereum and a majority can agree upon.
I would like the 21M limit to be increased so I can stand to make some profit mining! Currently that's impossible for me. I'd like it to change.
The key takeaway is that yes, it's ultimately just controlled by majority agreement between users.
Ethereum is currently decentralized because of the initial POW distribution. Won't there be centralized aggregators of eth so some point in the future a handful of POS nodes control a disproportionate amount of power? Is it so hard to imagine that coinbase or some other exchange accumulates enough eth to sway transaction validation?
Seriously, please answer if this is wrong!
>For certain kinds of 51% attacks (particularly, reverting finalized blocks), there is a built-in "slashing" mechanism in the proof of stake consensus by which a large portion of the attacker's stake (and no one else's stake) can get automatically destroyed.
Ethereum was launched with a 100% 'premine' of 72M coins, printed out of thin air and handed to the founders to keep or sell as they decided.
Since then another 44M have been mined, though it's not clear how much of that mining was also by insiders who influenced the work function's hardware compatibility or protected their hardware investment by controlling when/if the work function was changed or when mining was discontinued.