Proof of work is the only objective datapoint that can’t be faked. It has value in itself proportional to how much energy went into the proof, so it’s wrong to say something was wasted - all the energy went into producing something of value.
With proof of stake if you get out of the bunker and are presented with multiple competing chains - you can’t decide which one is genuine and which is a malicious fork. With proof of work it’s trivial and that’s where value comes from.
Why compare PoW with PoS if there is consensus https://xrpl.org/intro-to-consensus.html which already solved the problems of both before PoS even existed.
Consensus will never have 2 chains unless someone creates one on purpose. For all others it's always clear witch chain is the correct because everyone can check for correctness.
in case the consensus fails (for example serious global network problems) the consensus would halt which is not ideal but better than uncertain chains.
> Each participant in the network chooses a set of validators, servers specifically configured to participate actively in consensus, run by different parties who are expected to behave honestly most of the time.
maybe you're fine with that, i'm not and everybody who values bitcoin aren't either. it's the kind of thing that PoW solves and PoS doesn't. every PoS system i've seen just obfuscates this glaring hole instead of admitting that there's no way around it - you can't know which chain is genuine without trusting third party. in Bitcoin you can, or at least it is trivial to detect when something malicious is happening. and i'm not talking about correctness here, correctness is trivial, you don't even need to bring it up.
Why exactly aren't you fine with that? You run a node and you decide which nodes you want to connect your node to. If such a node gets a Tx it will be relayed to your node and the other way around.
It's not like your node would copy their validation and just agree with what they agree. Every node always enforces all the rules. So your node checks if a Tx is valid anyway and if you get invalid Tx from a node you trust it wont make your node accept that Tx. But it should make you overthink whether that node is trustable.
Trusting nodes is a "reliability rating thing" it doesn't really affect the consensus decisions because Tx are either valid or aren't.
The consensus that must be found is only about the order of Tx. If the majority of the validators say Tx X was first and therefore Tx Y is invalid (attempt to double spend) but your node got Tx Y first then its totally fine to flip the order of these two since both are valid just not at the same time.
"Voting" which nodes you trust means you trust them to be reliable/fast and not controlled by a single entity or controllable by a single entity (gov.) So with a clever trust list you help decentralize the network and help that the network runs on the most reliable nodes. You don't change the rules or allow other to change the rules. You can even make mistakes. You can choose some nodes that turn out to be not reliable/fast or even actively malicious. It has no fatal effect and can be corrected as soon as it is detected. Only if everyone would select over 20% "bad" nodes it could halt the consensus. Still would not allow a single false Tx or a single Tx reverse. It would just stop until some nodes remove the bad actors form their trust list.
As your quotes says "...run by different parties who are expected to behave honestly most of the time." The "trust" you give them is very very very limited.
because i don't want to trust anyone. XRP is not trustless or decentralized enough.
> Every node always enforces all the rules. So your node checks if a Tx is valid anyway and if you get invalid Tx from a node you trust it wont make your node accept that Tx.
as i said - validity is trivial. i asked you to not bring it up but you did anyway.
> majority of the validators say Tx X was first
"majority of validators" is not something you can reliably even define for yourself because you don't know which parties have colluded with each other. this consensus model fails on every layer.
> Only if everyone would select over 20% "bad" nodes it could halt the consensus.
i don't care how small you think this problem is. i want to never rely on having to select "correct" validators to ensure my financial future isn't at risk.
i want universally objective measure by which i can compare competing chains. if your protocol doesn't provide it without having to trust third parties - it's a failure.
> because i don't want to trust anyone. XRP is not trustless or decentralized enough.
And yet you trust the Internet to get packets to HN. You trust the government to administer the roads, the schools, the army, the police, the firefighters, and so on for days. The FDA to verify your drugs, Agriculture to verify your food.
You trust so many people every single day to make it through from breakfast to dinner, and yet this is where you draw the line for some reason you can't really quantify.
Trust can not be removed it can only be spread. If you use BTC you also have to trust that most miners/pools act honest. You can't even choose which miners you think are trust able. You have to just hope (trust) that at least 50% are honest an cant be compromised at the same time.
How is that any better than (optional, yes completely optional) choosing trusted Nodes?
You can very reliably define the "majority of validators" its defined as 80% it doesn't matter if or how many validators colluded. If Tx X was actually first but Tx Y reaches majority this simply means that either Tx X was not relayed fast enough to all the nodes OR it could mean that a lot colluded nodes voted for Tx Y.
Either way the consensus can "fail" because of technical reasons (slow relaying) or "fail" trough colluded nodes both "fails" do not lead to changed rules. No invalid Tx can happen this way. No Tx can be reversed this way. The whole situation can only ever happen if 2 Tx are valid signed but try to send the same funds (double spend attempt) and are inserted into the network nearly at the same time. Only one Tx will be processed. Enough colluded validators could effect which one but its completely irrelevant. Who cares which Tx of a double spend attempt is processed?
If you try to pay 2 persons with 100 bucks each by placing just one 100 dollar bill on the table in front of them, you don't know which one is gonna pick it up but clearly you have not fooled anyone into thinking you payed both.
>i don't care how small you think this problem is. i want to never rely on having to select "correct" validators to ensure my financial future isn't at risk.
You completely misunderstood, you don't have to choose "correct" validators to make it work correctly it can only work correctly! Bad validators don't "hurt" they just don't contribute.
Assuming 50% of all validators are bad and suddenly start censoring Txs or reordering non final Txs that would have the same effect as if these validators would just shut down. It does NOT produce a wrong output. Worst case is that the network halts and that is wanted because if 50% go offline at the same time there is probably something seriously wrong like a global internet collapse. Something humans have to fix first before resuming.
>i want universally objective measure by which i can compare competing chains.
By "universally objective measure" you mean you choose the longer chain? Fully aware that this can later change? Whats point? Having just one chain that is final doesn't need "universally objective measure" doesn't need comparing and makes final a binary option instead of "final" but better wait some more blocks to be sure.
https://en.wikipedia.org/wiki/Rolling_coal