undefined | Better HN

0 pointsRobotbeat5y ago0 comments

That’s an interesting point. Makes me think of the loops that are required to be jumped through to enable a trustless digital currency like Bitcoin. The entire network, with transactions of about 1kB in size, can do at most 7 transactions per second (3-4 orders of magnitude lower than Visanet, which itself is lower throughput than all the world’s physical cash transactions could theoretically do if everyone bought one or two things a day on average). That’s the same data throughput of a mid-1990s modem, but with power consumption of about 10GW to secure the transactions.

Pure trustlessness indeed has a massive computational and mental overhead. Some compromise (like the Lightning network, or even a central trusted banking institution) is essential to get anything done. (Even with Bitcoin, you’re trusting open source software, often an actual intermediary like Coinbase, etc... Try to do it without any crypto libraries whatsoever—writing it really from scratch—and it is an enormous project. How many people teach themselves Finite Field theory before doing Bitcoin because they take the trustlessness seriously?)

0 comments

3 comments · 1 top-level

WJW5y ago· 2 in thread

There are quite a few places in bitcoin where trust is required. For example, the software is developed by a relatively small group of core devs and 99%+ of bitcoin users will never read that code (and its dependencies) to see if it contains backdoors. We just have to trust the devs to "do it right" and not get compromised by either greed or blackmailed by external parties. If you presuppose a sophisticated enough attacker, the distribution methods could be MITMed so that the distributed binaries are not what is in the source code. The various miner pools could be compromised as well so that they run different software than the "main" bitcoin branch (but with a similar interface). If you accept that crypto exchanges are essential to bring down the barrier to entry, they need to be trusted as well. For any transaction also involving an exchange of goods or services not on the blockchain, you need to either trust the counterparty to actually fulfill their obligations, you have to involve a (trusted) clearing house or you have to trust the legal system to enforce the contract.

Tbh, trustlessness in payment systems is pretty much unachievable at scale. And if you are going to need trusted parties anyway, why have a blockchain in the first place? You can get a much better solution with a centralized database managed by the trusted party.

RobotbeatOP5y ago

I agree. You have to write your own software (offline, perhaps), not use a mining pool, plus not use an exchange to be trustless. And you’re still having to trust the other end to actually deliver on what they’re paying for, although they don’t necessarily have to trust you. That last sentence is one of the few changes that Bitcoin makes: it moves power from the consumer (who has ability to charge back if the goods are not received) to the one accepting payment (who now cannot be charged back). I’m not convinced this is a good move.

And I agree with your overall point. The only issue is when you end up having unnecessarily high transaction fees, as we do now. Satoshi was trying to fix that part. But unfortunately, Bitcoin isn’t better on that front.

Bitcoin has high transaction fees because of the lack of scalability of the blockchain concept at scale while Visa has high transaction fees (although for most daily transactions, less than Bitcoin) due to network effects allowing them to charge near-monopoly rents.

What we need is government action to enable instant ACH/payments. And to make that common. Visa ought to be unnecessary.

Layke11235y ago

So new taxes to support the infrastructure that allows feeless ACH/payments? Not saying I'm for or against, I'm just trying to understand your argument in full. It seems like the logical way that your idea would be funded.

1 more reply

j / k navigate · click thread line to collapse