Proof-of-stake will ultimatumly fail. The majority stakeholders can change the protocol at will to their benefit, and there’s no defense (without mob theft). With proof-of-work, at least miners can defend or fork. Best of luck to them though.

Also once a majority of stakeholder hit 51% there is no way to take away the 51%

With PoW, an entity could theoretically out mine the 51% attacker by just throwing more hashing power at bitcoin

2/3rds is required for a full 51%-style attack in PoS.

https://github.com/ethereum/wiki/wiki/Proof-of-Stake-FAQs#wh...

There's a lot of misinformation regarding PoS. Vlad Zamfir has spoken a lot on cartel attacks and how they're mitigated, but the biggest, and I mean bigger than biggest, argument against all of these attacks is that it's always detectable by honest nodes shy of 2/3rds attack and it defaces the value of the coin itself to attack at that volume.

You can always buy them out. But honestly, a 51% attack seems like a much less likely way for ethereum to fail compared to it just not living up to the hype.

A buyer requires a seller. Who's to say they would?

Buisness plan: Get a 51% stake in Ether, and hold big users hostage?

That works until the big users hard fork and take away your ether on their fork. It's pretty much a failsafe that always works at the cost of possibly destroying all trust in the network.

Giving everyone a nuclear button is an interesting solution.

> Also once a majority of stakeholder hit 51% there is no way to take away the 51%

You could do a hard-fork and slash their stake. In PoW, you can't do this.

> With PoW, an entity could theoretically out mine the 51% attacker

Not in the case of "selfish mining attack", where you may never know that an attack is happening until it's too late, you'll have little chance in defending with hashpower since the attackers will have a significant head start...

> You could do a hard-fork and slash their stake. In PoW, you can't do this.

So ... theft?

Theft of what exactly? Each fork is as valid as any other fork, the challenge is in agreement.

You should really be careful with how you play with definitions, as that will lead down to a slippery slope. Remember that time when someone managed to generate 184 billion bitcoin? With your reasoning, we can now say that the succeeding hard-fork to 'fix' the issue by slashing the bitcoins was theft.

> Remember that time when someone managed to generate 184 billion bitcoin?

you're being intellectually dishonest or even outright manipulative here.

there were never 184 billion coins created on bitcoin chain, there was a bug in validation logic of bitcoin client that made it follow the chain that was invalid.

are you seriously trying to compare that to stealing coins from somebody on a valid chain?

really?

> are you seriously trying to compare that to stealing coins from somebody on a valid chain?

They're not stealing coins from someone on a valid chain, they are forking away to another chain and slashing the user's stake. The user will still have all their coins on "their" chain and it's up to everyone else to decide which chain is "the right one". If the 51% user is truly malicious, the majority of network participants will move away to the new chain, rendering the 51% user's tokens (near) worthless.

> If the 51% user is truly malicious, the majority of network participants will move away to the new chain, rendering the 51% user's tokens (near) worthless.

so being successful is punished by destroying all wealth of the richest participant in the network? nice.

No, being malicious is.

having lots of money is malicious?

No, it's not, that's why I said

> If the 51% user is truly malicious

Having 51% doesn't automatically make you a malicious user. People also wouldn't move away to another chain simply because someone or some group owns 51%. They would move if said person or group abuses the power they get from owning a majority.

Now you've really tangled yourself in a big web.

And why is this a problem? Sure, they can change it arbitrarily to their benefit, but they only benefit if the value of their holdings is high. If they do something that people don't like, that value will fall. So they are strongly constrained by behaving in such a way that they do not make the users of Ethereum unhappy.

One only needs to look at modern democratic politics to have _serious serious_ doubts about that argument...

It won't surprise me at all if a PoS blockchain ends up looking just like a two party political system, where every now and then you get to choose which fork/party has the least objectionable outcome for you, but where both choices leave you worse off than you started...

Except that in this case, if you think you have better ideas, you can code up your own fork, and try to drum up support for it. The open source ecosystem works far better than the political system, I find.

Like now, you can drum up support for your own new party.

Yes, but in the case of the US government at least, there are structural reasons why the two party system is a stable equilibrium. Those structural reasons do not exist for cryptocurrencies. The most important of which is that there can be only one president of the US government, only so many senators, judges, etc... There can be infinitely many concurrently competing forks of Ethereum.

You and I can go fork Ethereum right now and now just tell people it's great, but show them it's great. It's like if we could fork the US government, refactor all the policies and let it run to demonstrate how good it is, and then people can come on board. You can't do that in politics, but you can absolutely do it in crypto.

Half the point of a cryptocurrency is the network effect - so I don't see how your argument holds, at all.

It is not clear at all that a small political experiment will scale to a super power. It is equally unclear that a small crypto currency experiment will scale.

This is without even touching on how easy, or difficult to it is rally support for a new system. Quite difficult, I'd say.

> Half the point of a cryptocurrency is the network effect - so I don't see how your argument holds, at all.

It holds because you can demonstrate its utility on a small network.

Network effects are too powerful. If they’re smart, they’ll make small changes over a long time, only ever upsetting a minority at once, and so the users stay. That’s Facebook to me, other things to other people.

That tends to happen in PoW actually. Example: Last year, Monero forked and changed the protocol by changing the PoW algorithm. Effectively shutting off a significant section of the miners. Miners had no defence - it was a major blow to them. Same thing happened in Sia. There's been discussion with changing the PoW of Bitcoin as well.

This was a change meant to block ASIC mining and encourage GPU mining.

Whilst this is great for decentralization, I've read interesting arguments in favor of ASIC mining. People who buy ASICs are committed to a coin. This leads to a stable base of mining. Meanwhile, GPU miners tend to mine whatever is the most profitable at any given time. This leads to large fluctuations in mining rates.

The Ether switch to POS has some people worried, as it might free up a lot of GPU power, which might overwhelm other GPU based POW systems.

Then again, any argument regarding mining algorithms is filled by people who have biases to the tune of 100 000$ of hardware investments.

You're making a different point. Yeah miners can get screwed in a PoW change, but the chain's fundamental properties are not at risk. Users have assets on both chains in a fork and if one chain does something terrible (or even just different) the market will reflect. With PoS, the majority remain the majority ... unless you support mob theft, a bad precedent I would say.

Nope, if an attacker does something bad (eg. a group forms a 51% cartel and starts censoring transactions), then they can be forked out and their stake slashed in the fork that doesn't censor the transactions. I really do not see a problem here? It's even better security than PoW, because once the attackers have their stake slashed, they can't attack anymore. With PoW, the miners can keep on doing a 51% attack even after the fork, indefinitely.

PoS is way more secure.

...if you trust a minority to decide who is an attacker and who isn't.

It's pretty clear when you validate the transactions

Sure but if they change it in a negative way they lose because they own the mayority. That's why to be able to change anything you have to have a huge stake, it's like having skin in the game.

Ethereum claims to be the "World Computer" that lets you build "Unstoppable Applications". If that doesn't sound like an AWS competitor then I don't know.

Sure, it's in some sense "an AWS competitor" but something like a billion times less efficient, measured by throughput, energy efficiency, or cost. So, in practice, no it's not.

AWS actually made an interesting product, a distributed ledger that's 'trustless' as long as you trust Amazon, but with none of the overhead of blockchain so at least it's efficient

I think Ethereum is going big places, but the AWS ledger solution is probably also going to be good for smaller industries where competitors don't trust each other, they want a database that's shared, and they all might trust Amazon enough to control it (eg if they're all using AWS anyway)

(https://aws.amazon.com/qldb/)

> […] a distributed ledger that's 'trustless' as long as you trust Amazon […]

This has been possible since the 80’s due to public key encryption.

It’s not having to trust a central actor that’s the innovation behind e.g. Bitcoin.

Well put!

To put it another way, the problems that Ethereum will solve are not ones that could be solved by AWS.

Totally agree.

Despite both being energy/transport companies, Embraer and Tesla are solving two very different consumer problems.

I see it less of a computing platform and more like the spinning table from the beginning of Indiana Jones and the Temple of Doom. The goal isn't to compute faster or better than others, it's for parties with competing interests to trust the computations that have taken place.

It's a git with stored procedures that can push commits.

Notably, there are rules that will block invalid commits, and tools that prevent git branches.

It's a hammer and nails. /s

Comparing Ethereum to AWS is an apples and oranges comparison at best. One is first order, and the other is second order.

You could make an argument it’s a competitor to AWS Lambda, but it would be akin to saying lxc is a serverless competitor.

I'm not comparing them, Ethereum marketing team pitched Ethereum as a world computer not me.

You are misunderstanding what they mean with “world” computer

Think of it as one, single "world computer" and it will be much closer to reality. Everyone can access that computer for free, and use it for a price.

> ... for free ... for a price

"access" vs "use".

Everybody can see the code/transactions that are run for free. Everybody can have their code run and execute transactions for a price.

I use quite a lot of AWS - and there's precisely zero chance I'll ever migrate, say, a mobile app backend platform to an Ethereum "unstoppable application"...

There are quite probably innovative new things that can be built using Etherium's "world computer", but my guess is that almost _none_ of what's currently running on (and paying for) AWS resources will move to it.

Can we get real for a second? Ethereum has no chance of taking on AWS or Microsoft.

Those are multi billion dollar business with billion dollar requirements in infrastructure investment on a yearly basis. Not going to happen.

I agree! But that's what Ethereum was pitched as.

Seems to me you're making invalid assumptions about what "world computer" means.

There's not a lot to go in in those two words, but claiming they mean Etherium is out to capture all (or any) of AWS's customers is a _very_ long stretch.

PoS was known way before PoW. It simply doesn't solve the problem at hand - decentralized consensus between mutually distrusting parties.

This guy knows his byzantine general's problem.

If POS worked, great.

I wouldnt trust it for a half decade of use.

IOHK has many conference accepted peer reviewed research papers that show that proof of stake has equivalent security properties as proof of work.

A peer reviewed paper saying WPA2 is secure also exists. That clearly didn’t make it so.

Only if we enumerate all possible attack vectors, and prove resiliency against these, is the proof as general as you claim. The difficult part — as the WPA2 paper shows — is the former.

How do you know when you’ve enumerated all possible attack vectors?

That's a good question. It's impossible to prove that anything besides math itself is absolutely true but we can say that for all known knowledge what is in this paper is true and secure. Unless some new discovery is made the papers show that proof of stake has similar mathematical security properties as proof of work.

> conference accepted

sure.

> peer reviewed

not so sure.

Bingo.

By the by, it’s not clear that proof of stake is good for crypto, although reducing the CO2 budget is unambiguously good for the rest of us.

Most PoS systems I’ve seen basically reward the oldest money holders with the right to stamp new transactions. This reduces the cost, but is extremely centralizing, bordering on neo-feudal.

The cynic in me also notes that developers of these chains are regularly the oldest money, making their motivations suspect at best.

In most PoS designs, registering a new validator account takes some fixed amount of time, just as a measure to prevent grinding attacks. If registration takes an hour, and I can't predict the random seed an hour in advance, then I can't predict how my account will be scored, so grinding would be futile.

Most PoS designs don't otherwise favor old accounts. I know Peercoin does, but that's a very small project, and I'm not aware of any others. Ethereum 2.0 definitely won't have any such mechanism. In the current spec [1], staking is binary; either you have an active validator account or you don't.

[1] https://github.com/ethereum/eth2.0-specs/blob/master/specs/c...

That sounds like the "coin age" system of Peercoin. Ethereum's PoS ignores coin age, which probably couldn't be used anyway because Ethereum uses account balances instead of UTXOs.

PoS is actually more decentralized when PoW. PoW is controlled by a few mining pools but PoS will have tens of thousands or more validators.

In PoW only a few people have the resources to build and run mining farms. Certainly not me. In PoS anyone can stake on a desktop or laptop. This greatly increases the amount of people who can participate in the network.

What is it good for?