Hacker News Reader

Comment by dist1ll | Hacker News Reader

Many of the things I mentioned are absolutely not solved problems.

> Payments require CVV

Many companies regularly violate regulations and store CVVs against your will. How do you think millions of CC CVVs get leaked in many of these major hacks?

> Solving the problem for international transfers simply requires greater international cooperation

Sounds great in theory. How long has traditional finance had to work out these problems? We're in 2023 and I still can't send money to an US account without paying absurd fees.

Now compare that to what several crypto projects have achieved a handful of years after their inception.

> Someone trying to sell a technical solution to a social problem ought to always be treated with extra skepticism.

What makes secure payments more social than secure messaging or secure browsing? Imagine trying to solve TLS and E2E messaging with social solutions.

View on HN

I had a similar opinion until I worked at a bank. Think of banks as an extension of the government, deputized to monitor all transactions for illegal activity. If banks fail they get in big trouble with the gov’t. Most bank transfers in the US are faster than any blockchain, but this is usually between known corporate entities that do a lot of business. For individuals there’s far more automated scrutiny. In addition, many smaller banks have ancient systems they refuse to upgrade, so overall speed of the banking system is held back by these weak links. Also, US banking oligopoly benefits from slower transactions and higher fees. Europe shows that regulators can force banks to operate faster and cheaper. That’s why this is not a tech issue, it’s a regulatory problem.

> I still can't send money to an US account without paying absurd fees.

You think that is a technological problem? Using crypto appears cheap (despite being inferior technology) because it simply skips 99% of what banks are doing.

Well, clearly I don't need 99% of whatever the bank is doing to send a transaction between two accounts I own. Add to that all the things that many banks continuously fail to provide, like proper OpsSec, transparent fees, full programmability via APIs.

> despite being inferior technology

Could you clarify? Are you talking about a particular protocol, or an implementation component, or mean crypto in general?

Happy to clarify.

a) Theory: In any engineering or optimisation problem, adding a constraint makes the solution not better, and, if the constraint is binding, worse. So, we would not expect distributed state machine replication (which was solved in the late 90s, with BFT-style consensus such as Paxos and Raft etc.) to get better by requiring permissionlessness. (And the latter is, in my view, the actual defining characteristic of what's now known as blockchain; though of course you have a couple of permissioned blockchains that are basically rebranding of existing technology to ride the hype-train.)

b) Practice: You could trivially run a BTC (ie, Longest Chain Rule) or ETH (ie, BFT-type) style network with extremely high reliability with some, say, 10 nodes that are permissioned and (at least 51%) trusted. That would use around 10-100 W (and you could probably achieve far higher throughput). Instead, BTC uses 20 GW, ETH around 1 MW. So, the requirement of permissionlessness increases energy usage by a factor of several thousand to billions.

Furthermore, you lose many many other features of centralised systems (such as that you can correct mistakes cheaply, replace lost keys, undo fraud, etc.)

So, basically, you employ a vastly inferior technology, just so you can circumvent the law. Really don't see why anyone should support that.

Better isn't objective. Not supporting it is fine but outlawing it ultimately means sending men with guns to stop people from interacting with a block chain. If you put bullet holes in people for their cryptographic fetishes, which is what the state will do ( or maybe if lucky they just flashbang a baby), you might be the baddie.

Sure, just as the bad bad SEC sends men with guns after people trying to do an unregistered IPO. Terrible, the last 90 years of regulated IPOs. The oppression.

> BFT-style consensus such as Paxos and Raft

Raft cannot deal with byzantine failures, so it's not a BFT protocol. Neither is the original Paxos, unless extended to Byzantine Paxos. Both are CFT by default.

> So, we would not expect distributed state machine replication to get better by requiring permissionlessness.

Right, but the banking system isn't a 600LoC Raft implementation. By your own admission, the bulk of inefficiencies come from the other 99% of stuff banks do.

So the question is: how much regulatory burden can be reduced if we were to switch to a crypto-based infrastructure.

> just so you can circumvent the law.

I can pay taxes in crypto to my local government. We also have crypto ATMs in several places here. Blockchain companies are working very closely with financial regulators. Not sure what kind of laws I'm supposedly circumventing here.

By BFT-style I meant employing rounds of voting rather than longest chain, thus being always consistent and eventually available (unlike LCR, which is always available and eventually consistent), but fair enough. Anyways, in the permissioned setting this was all solved in the 90s iirc.

Then, so, just moving stuff around in a robust and even distributed ledger is, we can all agree, while not trivial, a solved problem.

So then one can impose further constraints. Either, 1), comply with the law and rules and regulation, and sure, that will make it less efficient. But that is the 99% of the work banks do that crypto doesn’t do, and they are not being done for fun, but for good reasons. Or 2), permissionlessness, but what for? So you can avoid 1), because otherwise you wouldn’t need to incur the tremendous costs of 2).

I don’t think we should adapt the AirBnB/Uber approach (how much of the regulatory cost of fire regulations/working time restrictions/insurance cost can we avoid) to money transmission.

It is obviously trivially possible to do legitimate transactions with crypto, just as you can cut an onion with a machine gun. But that’s not the raison d’être of a machine gun. And the raison d’être of crypto is to escape regulation, as Satoshi basically said in the original emails, and as is abundantly clear not only from looking at the last 15 years of history, but also from the willingness to incur the massive overheads of permissionlessness.

ETA: appreciate the discussion, btw.

> Now compare that to what several crypto projects have achieved a handful of years after their inception.

By ignoring regulations entirely. Guess what, if you're the bank, now you're liable.

Reply on news.ycombinator.com