No surprise... I do! So Solana didn't catch my interest when it first came on the scene. Another cheaper/faster, blah blah. Easy to promise when the chain is new and barely used. But then I read about Proof of History and got /very/ interested.
Proof of History was going to be the MEV killer we needed. Egalitarian. Your timecode is your transaction order. FIFO. Beautiful. Solana, though, would then have moments of spam and a lot of downtime. Did they get clever and figure out a way to avoid the spam without losing the beauty of FIFO?
No. They just scrapped what made Solana unique and went straight back to bribery. About a year or two later, after reading about PoH and getting excited, I read about "Priority Fees" and losing all that excitement, as Solana then becomes just like all the others. I honestly don't see what it offers that is unique anymore.
The solution for How Solana Could Get Its Groove Back is to fork solana and remove Priority Fees, or possibly make an EVM L2 that has Proof of History (the core value of Original Solana) for transactions on the L2 prior to finalizing on Ethereum.
What about the underlying problem, though? Spam to the point of network congestion? Because PoH with spam is not an effective solution to MEV.
When there's an imbalance in the difference of supply vs demand and an uneven allocation of resources, the general idea is to have pricing smooth things out. It works, but leads to MEV, and taken to its logical conclusion, game theory shows the one with the most resources, even in “fair” random games of chance, after enough rounds, will win all the resources. Pricing allows the rich to pay for, say, an arbitrage opportunity that is outside the possibility of the rest, giving the edge that wins over all possible opportunities. Thus the rich wins more "rounds." Thus one person controls everything. Thus a trend to, and eventually complete, centralization (barring an event outside of a closed system, of course).
The solution is generally two fold: taxes and socialized services. Taxes reallocate from the richest to the poorest, erasing the benefit of having more resources for every contest/wager. Socialized services keep at least some services that are absolutely needed from being at whim of ruthless game theory.
What's an absolutely needed service? Uptime. So some taxes can be used to make sure the chain (e.g. PoH Solana, let's call it PoHS) never crashes. Spam happens? Spend tax money on increased infrastructure. Whatever is left, goes to the poorest (e.g. those who have and made the least on PoHS). Presumably those spammers spam to make money, so they'll get taxed if they do. Takes away the incentive. It's not like using a stick instead of a carrot. It's just using the stick later. Not upfront as price, but later as taxes.
Of course, now you have the bigger problem of determining value for taxes. And getting those taxes. Easiest way to prevent tax dodging is to sanction or deport them if they don't pay. E.g. Don't pay, don't get to put transactions on PoHS, at least until you've both paid taxes and the tax dodging penalty. This could be extremely effective as a way to prevent removing any gains. For price, I'm afraid it's gotta be oracles, but it should be acceptable.
The logical question is how to prevent bridging off the chain, prior to paying taxes. The best way is only use official bridges, and only very simple ones. Like you can bridge SOL, BTC, or ETH. You lock the token, and then it appears on PoHS. You can bridge exactly the same tokens that you bridge there during any tax season. But any extras only after you calculate your taxes. Spammers thus will be stuck.
Of course a system of taxes and socialized essential services could possibly work even with current models of MEV and bribe-for-order-flow. But I still like the beauty of FIFO, and the fact that PoH seemed to offer that, and penalizing spam abusers after the fact. Because it gets rid of sandwich attacks, you are just left with new problems to figure out like accomplishing taxes, which may require some human involvement, without your “government” getting destroyed by rich humans. But, that took like 200 years to happen in the US. You may have some time.
This definitely sounds insightful to dig into for understanding MEV, it does seem like MEV is an inevitable factor of the same tools you need to prevent spam ... I need to do some historical research, sounds like!
I have a few other ideas. One is a different kind of PoH, Proof of Humanity--but not in the wacky Worldcoin orb way.
I mean, if they did the orb a completely different way. Where it was a device just about anyone could build using off the shelf components. Or better yet as easy as you can take a 10 year old Android (or IPhone, ideally both) phone, format it completely, and that's your orb. Or best of all, you could authenticate with any old webcam.
Besides the hardware, you need software for the device that is completely free and open source. Finally you need to have completely/safely distributed zero-knowledge proofs. E.g. zk-proofs encoded via NIST quantum-proof cryptography and then the proof is put on Bitcoin--the most distributed network–barring that Ethereum. Requiring a special device and a special token (in WRLD case, for payment–and payment via the ecosystem token is the shakiest/weakest pin in all of DePIN)--is just BS complexity that makes it less secure/efficient and is only there to make more money for the founders.
Once you authenticate as human some users, you probably still want to allow bots to do stuff. But they will be driving on all the other lanes, while PoH transactions go in the car-pool lane. And the speed limit is strict. A bot, or rather non-PoH, transaction can never go faster than any PoH transaction.
That idea might work, but I, too, don't want to stare into any device. It just feels off and surveillance state-y. I'm pretty sure even my bio data isn't safe, but I expect extreme assurances before I part with it. Hence would need: hardware that could exist before the technology and thus specific backdoors could be made, FOSS that would be completely transparent again because of backdoors, and a ledger that is both extremely reliable against censorship and private with sensitive data contained in it.
Okay, it really just needs a database, not a ledger, but a ledger can accomplish a fair amount of what a database can, just ask any one of the Fortune 500 companies that are probably running on Excel. And there's no spreadsheet more distributed and censorship proof than Bitcoin. However, what's more important is the security of the data, which might need to be large to be quantum-proof and a zk-proof (zk is generally smaller, but I don't know about quantum). Possibly too big for Bitcoin.
An adjacent idea that I've also had and I've liked more is not using a binary human-or-not designation, nor a device. But instead a more fluid account reputation for non-automation, using your transaction history and socials. Using transaction history and socials to see how human one is, is not a new idea. However, it is generally considered in the context of airdrop determination, not in transaction ordering.
The efforts I've seen are just connect with a wallet (usually just one at that), then connect your Twitter, Discord, and Github. Or slight improvements there on.
At that point, I think the only calculation is how old is the account, how many transactions total, how much gas, and maybe did it receive certain key airdrops. It should actually check to see how human the transactions are. What's the minimum amount of time between some transactions (less than humanly possible, for example)? Are there periods where a normal human might be off because of sleep? What's nice is if a metric is found that is pretty accurate for humanness, old accounts will be impossible to retrofit around it. The only possible hack is then buying/stealing old accounts. Newer accounts are at a disadvantage, but presumably there are still newer metrics to be found that will be unknown until the newer accounts become older accounts.
Per all this, you get a score, the more non-automated you seem (which likely means active as well as non-automated, because generally activity increases trust/reliability of data), the higher your priority in transactions. It's extremely elegant, and yes, requires the cat and mouse against sybillism. But, again, people have to deal with that for other reasons, so why not leverage it for transaction ordering as well?
MEV " if you have some insight or a take"
No surprise... I do! So Solana didn't catch my interest when it first came on the scene. Another cheaper/faster, blah blah. Easy to promise when the chain is new and barely used. But then I read about Proof of History and got /very/ interested.
Proof of History was going to be the MEV killer we needed. Egalitarian. Your timecode is your transaction order. FIFO. Beautiful. Solana, though, would then have moments of spam and a lot of downtime. Did they get clever and figure out a way to avoid the spam without losing the beauty of FIFO?
No. They just scrapped what made Solana unique and went straight back to bribery. About a year or two later, after reading about PoH and getting excited, I read about "Priority Fees" and losing all that excitement, as Solana then becomes just like all the others. I honestly don't see what it offers that is unique anymore.
The solution for How Solana Could Get Its Groove Back is to fork solana and remove Priority Fees, or possibly make an EVM L2 that has Proof of History (the core value of Original Solana) for transactions on the L2 prior to finalizing on Ethereum.
What about the underlying problem, though? Spam to the point of network congestion? Because PoH with spam is not an effective solution to MEV.
When there's an imbalance in the difference of supply vs demand and an uneven allocation of resources, the general idea is to have pricing smooth things out. It works, but leads to MEV, and taken to its logical conclusion, game theory shows the one with the most resources, even in “fair” random games of chance, after enough rounds, will win all the resources. Pricing allows the rich to pay for, say, an arbitrage opportunity that is outside the possibility of the rest, giving the edge that wins over all possible opportunities. Thus the rich wins more "rounds." Thus one person controls everything. Thus a trend to, and eventually complete, centralization (barring an event outside of a closed system, of course).
The solution is generally two fold: taxes and socialized services. Taxes reallocate from the richest to the poorest, erasing the benefit of having more resources for every contest/wager. Socialized services keep at least some services that are absolutely needed from being at whim of ruthless game theory.
What's an absolutely needed service? Uptime. So some taxes can be used to make sure the chain (e.g. PoH Solana, let's call it PoHS) never crashes. Spam happens? Spend tax money on increased infrastructure. Whatever is left, goes to the poorest (e.g. those who have and made the least on PoHS). Presumably those spammers spam to make money, so they'll get taxed if they do. Takes away the incentive. It's not like using a stick instead of a carrot. It's just using the stick later. Not upfront as price, but later as taxes.
Of course, now you have the bigger problem of determining value for taxes. And getting those taxes. Easiest way to prevent tax dodging is to sanction or deport them if they don't pay. E.g. Don't pay, don't get to put transactions on PoHS, at least until you've both paid taxes and the tax dodging penalty. This could be extremely effective as a way to prevent removing any gains. For price, I'm afraid it's gotta be oracles, but it should be acceptable.
The logical question is how to prevent bridging off the chain, prior to paying taxes. The best way is only use official bridges, and only very simple ones. Like you can bridge SOL, BTC, or ETH. You lock the token, and then it appears on PoHS. You can bridge exactly the same tokens that you bridge there during any tax season. But any extras only after you calculate your taxes. Spammers thus will be stuck.
Of course a system of taxes and socialized essential services could possibly work even with current models of MEV and bribe-for-order-flow. But I still like the beauty of FIFO, and the fact that PoH seemed to offer that, and penalizing spam abusers after the fact. Because it gets rid of sandwich attacks, you are just left with new problems to figure out like accomplishing taxes, which may require some human involvement, without your “government” getting destroyed by rich humans. But, that took like 200 years to happen in the US. You may have some time.
This definitely sounds insightful to dig into for understanding MEV, it does seem like MEV is an inevitable factor of the same tools you need to prevent spam ... I need to do some historical research, sounds like!
I have a few other ideas. One is a different kind of PoH, Proof of Humanity--but not in the wacky Worldcoin orb way.
I mean, if they did the orb a completely different way. Where it was a device just about anyone could build using off the shelf components. Or better yet as easy as you can take a 10 year old Android (or IPhone, ideally both) phone, format it completely, and that's your orb. Or best of all, you could authenticate with any old webcam.
Besides the hardware, you need software for the device that is completely free and open source. Finally you need to have completely/safely distributed zero-knowledge proofs. E.g. zk-proofs encoded via NIST quantum-proof cryptography and then the proof is put on Bitcoin--the most distributed network–barring that Ethereum. Requiring a special device and a special token (in WRLD case, for payment–and payment via the ecosystem token is the shakiest/weakest pin in all of DePIN)--is just BS complexity that makes it less secure/efficient and is only there to make more money for the founders.
Once you authenticate as human some users, you probably still want to allow bots to do stuff. But they will be driving on all the other lanes, while PoH transactions go in the car-pool lane. And the speed limit is strict. A bot, or rather non-PoH, transaction can never go faster than any PoH transaction.
That idea might work, but I, too, don't want to stare into any device. It just feels off and surveillance state-y. I'm pretty sure even my bio data isn't safe, but I expect extreme assurances before I part with it. Hence would need: hardware that could exist before the technology and thus specific backdoors could be made, FOSS that would be completely transparent again because of backdoors, and a ledger that is both extremely reliable against censorship and private with sensitive data contained in it.
Okay, it really just needs a database, not a ledger, but a ledger can accomplish a fair amount of what a database can, just ask any one of the Fortune 500 companies that are probably running on Excel. And there's no spreadsheet more distributed and censorship proof than Bitcoin. However, what's more important is the security of the data, which might need to be large to be quantum-proof and a zk-proof (zk is generally smaller, but I don't know about quantum). Possibly too big for Bitcoin.
An adjacent idea that I've also had and I've liked more is not using a binary human-or-not designation, nor a device. But instead a more fluid account reputation for non-automation, using your transaction history and socials. Using transaction history and socials to see how human one is, is not a new idea. However, it is generally considered in the context of airdrop determination, not in transaction ordering.
The efforts I've seen are just connect with a wallet (usually just one at that), then connect your Twitter, Discord, and Github. Or slight improvements there on.
At that point, I think the only calculation is how old is the account, how many transactions total, how much gas, and maybe did it receive certain key airdrops. It should actually check to see how human the transactions are. What's the minimum amount of time between some transactions (less than humanly possible, for example)? Are there periods where a normal human might be off because of sleep? What's nice is if a metric is found that is pretty accurate for humanness, old accounts will be impossible to retrofit around it. The only possible hack is then buying/stealing old accounts. Newer accounts are at a disadvantage, but presumably there are still newer metrics to be found that will be unknown until the newer accounts become older accounts.
Per all this, you get a score, the more non-automated you seem (which likely means active as well as non-automated, because generally activity increases trust/reliability of data), the higher your priority in transactions. It's extremely elegant, and yes, requires the cat and mouse against sybillism. But, again, people have to deal with that for other reasons, so why not leverage it for transaction ordering as well?