It's not reinventing accounting. It's making sure that the accounting you do is public because that allows for trust.
I think it would be cool if we, as a society, had easy API access to everything our governments spend their money on. Wouldn't you say that would be for the greater good?
Making something public doesn't solve the problem. Deceiving the public is easy and the public is duped frequently. Making accounting public doesn't solve the problem at all.
It's not making accounting public. It's making accounting public in a way that you can't dupe the public. There's a question if that's possible, but you're arguing against a straw man here. I'm skeptical that they'll come up with something that convinces everyone that their new version of public accounting can be trusted, but maybe there's something novel they can come up with that will convince the skeptics.
The issue here is that making money by holding onto assets you can’t fundamentally make money by holding is hard.
There is a huge incentive to do things that people don’t actually want you doing, so someone can actually make money.
Accounting is just a word for tracking money.
It can be opaque, or obvious, detailed, or vague. And it can also be fake.
Knowing which is which, and if it is appropriate or not, is almost always specific to the circumstances and the goals of whoever is doing it, and that’s pretty fundamental.
GAAP is a generally acceptable list of practices for public entities, but there are still a ton of judgement calls going on there, and have to be, for what is a pretty ‘standard’ way to operate pretty standard businesses.
There is no automated way to do it that won’t allow manipulation.
Attempts at automating it are more likely to provide something like bank account statements, than a balance sheet. Which is something, but it’s not enough. That’s what is generally referred to as book keeping (aka transactions are recorded, but not necessarily characterized or organized correctly), but may even fall short of that.
“The issue here is that making money by holding onto assets you can’t fundamentally make money by holding is hard.”
They can make plenty of money through volume (fees), while holding the crypto 1:1. Sounds relatively easy to me.
That’s the baffling thing about FTX/Alameda. FTX on its own could be a nice profitable business, yet it seems they got greedy by lending customer funds to the hedge fund-y sister company.
> FTX on its own could be a nice profitable business
Not really. Their tech stack was too slow so other Market Makers weren't willing to come over (the price can change and leave them fulfilling sales at bad prices too frequently). So they had Alameda be their market maker and they were constantly burning money doing that. Maybe they believed there would be enough liquidity if they scaled more and it was an investment in their future (a very charitable view) and it was fine to borrow funds in the short term and then that got away from them. But from what we've heard of their tech stack, they weren't ever going to be viable because they couldn't support what HFT would need.
Not as easily (short term) as they can by walking away with billions in the raw money though.
And fees get competitive, it’s inevitably a race to the bottom, where firms need to spend a lot of money competing on things like long term reputation.
Meanwhile if they screw up in their security, they’re also liable for huge multiples of those fees in losses.
And if someone pops up that charges less fees, and seems ok enough, a lot of the money moves there quickly, until there is a scandal anyway.
It requires a mature organization who has strong controls and bulletproof working processes to avoid losing multiple years in fees by accident too.
Not saying you’re wrong - it’s where things inevitably end up when it’s heavily regulated and watched so all the other alternatives are hard (and result in major prison time). Usually.
But it should be noted that brick and mortar banks gave up on that model a very long time ago, favoring explicitly loaning out customer funds (fractional reserve banking) because it’s more sustainable for them.
You can only do that if you are holding exclusivly onto one asset. The relative values of different tokens change over time. So if you want 1:1, you need to have a robust mechanism for doing that.
But you won't want to because that'll burn through transaction fees for no real gain.
> FTX on its own could be a nice profitable business
That requires fees, and given the number of exchanges that do it for free, so long as you keep your wallet with them, its not going to make you money.
Sure you can do market making, or arbitrage, but that's still risky and you don't make that much money doing it. <0.1% on each transaction.
Fwiw, you can make money holding ETH, by staking it. Coinbase and Kraken both do that. They take a percentage of the return customers get by agreeing to lock and stake their coins on the exchange.
(Before that was available though, both lived entirely on fees, unless they've been committing massive fraud.)
Since Sept 15th 2022 yes. We’ll see how it goes longer term, and what the fees and economics look like when things settle out.
We don’t know what Coinbase and Kraken are doing internally.
Coinbase does seem to be mostly off fees (which are sky high), and does seem not sketchy. It doesn’t appear overly easy for them even then (hence the ‘it’s hard’) despite folks being itching to put money in on the way up to the point they’ll mostly ignore high fees.
They’re definitely playing the long game, compared to the others that have looted billions elsewhere.
As the main on/off ramp in the US, they have a pretty nice position from it too.
Actually since Dec. 1 2020. That's when the production beacon chain went live with real ETH, even though the proof-of-work chain was still running in parallel. Coinbase and Kraken started offering it sometime in early 2021.
When it isn’t the primary or only way of doing it, it isn’t a model for if it will work economically. Which is why I picked that date.
That it was possible to do it differently earlier, and some folks did start transitioning, doesn’t change the fact that we don’t have much track record with it yet as ‘the way it is done’.
It usually takes decades for these things to stabilize into anything predictable.
Not sure what you mean exactly, but the live beacon chain was the only way to stake ETH since it was launched in 2020. By the time of the merge, over 10% of all ETH had been staked on it. Total stake now is just a little higher than it was just before the merge.
The intent is not for most ETH to be staked. 12-13% is fine, and they don't expect it to go beyond 30% or so. We just need enough staked so the network is secure.
Staked ETH isn't available to be used for any other purpose, like paying transaction fees, depositing in defi protocols, using as currency, etc.
Anyway, this is getting a bit far afield of my point, which is just that for ETH, exchanges can earn money on deposits, without loaning them out to risky side ventures.
You mean whether it's a viable model for the exchange? I have no idea, never looked at their numbers. But Coinbase is public, so now I'm curious whether they break out their revenue sources.
That baked-in inflation rate is only about half a percent annually, less than a third of Bitcoin's current rate of 1.7%. But Ethereum's issuance is offset by the fee burn. Since the merge, the total ETH supply has actually decreased by 4000 ETH. You can track it here: https://ultrasound.money/
I don't see the reliance on large exchanges. There are plenty of solo stakers and the protocol gives them the same rate of return as large exchanges. That's an improvement over the economies of scale that large miners typically achieve. People without 32 ETH can use a decentralized staking pool, just like most solo miners use pools.
Total stake is under 15% of ETH so it's not a huge hit to liquidity. The real bottleneck for p2p value exchange is transaction rates, and on that front Ethereum is coming along quite well. If it were only doing ETH payments, it would be capable of 700 tx/sec today on chain. It actually does fewer since many of its transactions are more complex, but rollups multiply that significantly, and upcoming scaling improvements should get them to 100K tx/sec within the next several years, while still maintaining full trustlessness.
If you want to make accounting public, go dig giant limestone disks out of that island with cannibals on it, then bring them back and put them on the beach so everyone can see them: https://en.wikipedia.org/wiki/Rai_stones
Which probably won't convince those of us who are skeptics that it's possible. It needs to be technically possible, but easy enough to verify for the average user that someone can't just claim they're doing it the right way but doing something fraudulent instead. There's good reason to be skeptical here. But arguing that all they're trying to do is make public accounting isn't the right counter argument here.
You're right that making it public alone doesn't solve the problem since it is easy to "cook the books".
But if it's public and the results are the side effect of each transaction (e.g. the accounting is rules-driven and automatic as part of the system itself), I think it can change the game.
Budgets and financial statements for US federal and municipal governments already have to be public. If you're talking about the bill of materials and receipts for every individual transaction, there are some pretty serious obstacles to that. At minimum:
- The government directly provides medical service via the VA and DoD and disclosing every transaction risks HIPAA violations.
- It may not now be illegal, but I think there would be some privacy concerns about things like revealing every person's student loan payments.
- Counseling and employee assistance type stuff for government employees can't be made public for the same reasons.
- Criminal fines for anyone whose record is later expunged present a hurdle.
- Payments to confidential informants in criminal investigations obviously need to stay confidential or you're going to get them murdered.
- Witness protection payments same thing.
- At least some transactions are classified.
- Contract details are kept hidden from other contractors right now to avoid undercutting and collusion.
All in all, as long as you're forced into some system where maybe some, even most, transactions can be made on a public ledger, but others have to be kept private, you'll never prevent fraud and/or suspicions of fraud in the form of transactions being kept illegally private. At some point, you have to trust the auditors and Congressional oversight committees.
> I think it would be cool if we, as a society, had easy API access to everything our governments spend their money on. Wouldn't you say that would be for the greater good?
We can't do that, because then people would realize it's all a house of cards. In an unrelated note, when was the last time US gold reserves were audited?
Yeah, it more or less doesn't matter. The US isn't in danger of having to liquidate assets, and if it was you'd have a lot worse problems on your hands than how many gold bars are in Ft. Knox. Why do you think it matters?
That's a very general argument for a very specific request. Transparency is good, but nobody is coming on to HN asking citing transparency and asking for audited figures for the value of US-owned dump trucks, or how much wood pulp we could produce in Grand Tetons National Park, or how much money we could make grinding up the highway system for scrap metal. So why gold specifically?
> Why would anyone care how much gold the US government has?
was not intended to dismiss interest in any particular asset, but to ask why there isn’t an equal amount of interest in assets other than gold.
I can’t speak for the other commenters who raised the issue, but it’s presumably because the asset is far more valuable than the ones you’ve mentioned, and far easier to convert to other assets.
It’s also been used as medium of exchange or reserve currency for much of human history, so expecting people to dismiss it out of hand based on prevailing theories espoused by modern monetary institutions wouldn’t be reasonable.
I think it would be cool if we, as a society, had easy API access to everything our governments spend their money on. Wouldn't you say that would be for the greater good?
That's exactly what Vitalik is proposing.