This is a wonderfully insightful comment, I’d not heard of many of the points you raised.
As someone coming from the PoW worlds of Bitcoin and Monero there has always been something about PoS that gave me an ick, but I could never quite place my finger on it. Especially when Ethereum clearly seems to have PoS working well in practice. I’ve always had a suspicion that Ethereum could have never gotten off the ground, at least to the success level it currently has, if it had been PoS initially.
The mechanics of slashing not working are interesting. Do we know of any abuse attempts by validators and have we seen any validators get slashed yet?
Your point about PoS not being truly permissionless is a big one for me. Aside from it being fundamentally opposed to the core promises of cryptocurrencies, what’s to stop TPTB from slowly accumulating enough of a stake and then bam, we’re back to the legacy permissioned finance ecosystem but with MUCH more fine-grained tracking and telemetry. That seems like an extremely pernicious attack.
The point about trust is another big one. Cryptocurrencies are meant to be trustless. That means a user doesn’t need to trust anyone else as they can fully validate their view of the world for themselves. Of course they may trust other participants for efficiency, but they don’t need to. And caveat emptor, as with all distributed systems you must assume (trust) that you are not currency inside a partitioned network, since AFAIK there is no way a system can internally verify that it is not.
How much these things matter in practice only time will tell. Ethereum appears to be working well at the moment and just keeps chugging along.
I’d like to see more work put into finding ways to utilize the work from PoW. For example I have an idea to use Monero’s CPU-favoring PoW for PoW based DDoS protection as seen in Tor [0]. When a user accesses a website they are given a PoW challenge to complete. This challenge is actually for a share of mining rewards as in P2Pool. The mining reward share would go to the website operator. This would harmoniously improve several things about the web.
First, it would help protect websites against layer-7 DDoS attacks.
Second, this L7 DDoS protection reduces the webs dependence on companies such as Cloudflare, the internets biggest man-in-the-middle.
Third, it provides a way to pay website owners costing the users a small amount of their computers time and energy in much the same way as ads do currently.
Fourth, it reduces the webs dependence on advertising as the way to fund your website.
Fifth and finally, it helps secure the web-native currency in which website operators would be paid and which others can use for whatever they want.
There's nothing wrong with PoS as long as it is understood as a mechanism for diffusion of power and trust, like a multisig or a corporation. Probably PoS protocols do not need to have in-protocol slashing. Solana has already thought of this and does not have slashing IIRC.
PoW has its own problems. It is likely to be unstable for various reasons in the absence of inflation. I don't have time to find all the citations but there have been many papers about this subject. It's also not great for the environment, and there is something unfortunate about the fact that the best trustless protocol we can come up with as a species requires massive waste. Maybe it is currently the best we have but trying to come up with reasons that the waste is actually OK is a waste of time IMO.
As someone coming from the PoW worlds of Bitcoin and Monero there has always been something about PoS that gave me an ick, but I could never quite place my finger on it. Especially when Ethereum clearly seems to have PoS working well in practice. I’ve always had a suspicion that Ethereum could have never gotten off the ground, at least to the success level it currently has, if it had been PoS initially.
The mechanics of slashing not working are interesting. Do we know of any abuse attempts by validators and have we seen any validators get slashed yet?
Your point about PoS not being truly permissionless is a big one for me. Aside from it being fundamentally opposed to the core promises of cryptocurrencies, what’s to stop TPTB from slowly accumulating enough of a stake and then bam, we’re back to the legacy permissioned finance ecosystem but with MUCH more fine-grained tracking and telemetry. That seems like an extremely pernicious attack.
The point about trust is another big one. Cryptocurrencies are meant to be trustless. That means a user doesn’t need to trust anyone else as they can fully validate their view of the world for themselves. Of course they may trust other participants for efficiency, but they don’t need to. And caveat emptor, as with all distributed systems you must assume (trust) that you are not currency inside a partitioned network, since AFAIK there is no way a system can internally verify that it is not.
How much these things matter in practice only time will tell. Ethereum appears to be working well at the moment and just keeps chugging along.
I’d like to see more work put into finding ways to utilize the work from PoW. For example I have an idea to use Monero’s CPU-favoring PoW for PoW based DDoS protection as seen in Tor [0]. When a user accesses a website they are given a PoW challenge to complete. This challenge is actually for a share of mining rewards as in P2Pool. The mining reward share would go to the website operator. This would harmoniously improve several things about the web. First, it would help protect websites against layer-7 DDoS attacks. Second, this L7 DDoS protection reduces the webs dependence on companies such as Cloudflare, the internets biggest man-in-the-middle. Third, it provides a way to pay website owners costing the users a small amount of their computers time and energy in much the same way as ads do currently. Fourth, it reduces the webs dependence on advertising as the way to fund your website. Fifth and finally, it helps secure the web-native currency in which website operators would be paid and which others can use for whatever they want.
I think such a solution would be truly beautiful.
0: https://blog.torproject.org/introducing-proof-of-work-defens...