81 Comments
Sep 2, 2022Liked by Noah Smith

Sounds very interesting but looks like I need to add another 100 points of IQ to better understand it ;-) Perhaps you can write a few articles on this kind of deep post!

Expand full comment

Trying to understand crypto is like trying to understand the bible: A pointless exercise in futility that not's worth your time, or anyone else's.

Expand full comment

The part about ‘startup community’ sounds more alternate values, I think, sounds interesting using crypto structure. Kind of fun to think about at least!

Expand full comment

Yeah, I got nothing out of this interview. Noah usually does a great job helping readers understand what's going on, but failed at it here. Maybe it's the topic, but still. I'm more confused about crypto than ever.

Expand full comment

Loved the conversation and Vitalik for sure is a thoughtful intelligent individual. Questions arise first on a practical standpoint. PoS: how can they decide which one of the blockchains is the falsified one? How the holder of such eth has any recourse if one morning finds no more Eth in her/his account because destroyed? It is a simple question of course but goes to the core of the philosophical distributed system because if someone can have an actual power on others with no recourse possible then the equality is quite stalinesque… isn’t it?

Expand full comment

Just a quick comment - I think the link to Twitter account of Polynya is wrong - it should be this account as far as I know - @apolynya

Anyways, it was a great read :)

Expand full comment

Interesting interview, but I would have loved to get his thoughts on some of the basic scaling and cost issues that the ethereum (and all other) blockchains suffer from. Anybody who has worked in distributed databases will see the problem immediately. Let's look at the very simple example of writing data to the chain. After all, the blockchain is essentially a database and the primary purpose of a database is to able to write data and read data. So how much does it cost to store data on the ethereum blockchain? Well the problem is that this depends on something called "gas" and the price of this fluctuates. So already, if designing a system that uses ethereum as your database, you have non-deterministic pricing. I've worked in distributed systems and cloud computing for several decades and I can tell you right now, that alone is a non-starter for most enterprises. They do not like "lumpiness" or volatility in their IT spend. But let's put that aside for right now. The cost of storing 1 byte in ethereum is 600 "gas." 1 gas = 10^-9 ETH, which is the ethereum token. 1 ETH = ~$1500. So to store 1MB of data on ethereum right now would cost 10^-9 * 1024^2 * 600 * $1500 = $943.71. Compare that to Amazon's DynamoDB NoSQL service, which can scale to millions of transactions per second. Writing 1MB of data consumes 1000 write operations in DynamoDB. Write operations are 1KB in size and cost $1.25 per million. So for 1000 concurrent write operations necessary to write 1MB of data, it would cost $0.00125. This is literally orders of magnitude lower (.0001% of the cost) than writing that same 1MB to the ethereum blockchain, and is also a fixed price. The same is not true for ethereum, as ETH can (and does) fluctuate wildly in value. Now I know that the web3 developers and enthusiasts out there will say "sure, this is true but that's why most web3 apps do a lot of their work off-chain." Fine, but then remind me again what specific technical capability the blockchain is providing? If its generally accepted that a blockchain will never be nearly as cost effective as a distributed database, and that this means any decentralized app (i.e DeFi, etc) will have to do any work requiring high amounts of compute, transaction throughput, or storage "off chain" to keep costs from spiraling out of control, I fail to see what the blockchain itself is providing. It appears mostly to be an unnecessary, expensive, and complex component of a system architecture. Sure, the social and community building dynamics of a pseudonymous public ledger may be interesting, but the fact is as a scalable technology, it's horrendous from a cost standpoint (not to mention the environmental impact of "proof of work" which is pretty much a crime against humanity). Throughout my career I've seen a lot of over-hyped technologies but nothing comes close to web3 in terms of being indisputably inferior to other database technologies that already exist.

Expand full comment

"It appears mostly to be an unnecessary, expensive, and complex component of a system architecture."

Exactly. So the question is: why do people keep pushing the idea? I see two answers.

One: ideology, in that it allows people to pretend that the new world is distributed and trustless and therefore frees them from real world considerations like the law - while in practice it is neither and when it connects to the real world those real world considerations apply.

Two: greed, in that it allows some people to make a great deal of money by selling buzzwords.

Expand full comment

I didn't understand a single word this guy said.

If his target audience is experts and well-informed crypto aficionados, then my incomprehension is no problem for him. If his goal is to educate, and inspire, a larger audience, then he has some real problems.

Expand full comment

A smart guy wasting his smarts on sociopathic bullshit. Pitiful. He needs to be slapped silly.

Expand full comment

"The math nerd way of putting it would be: the price of crypto is stuck in a bounded range (between zero and all the world's wealth), and crypto can only stay highly volatile within that range for so long until repeatedly buying high and selling low becomes a mathematically almost-surely-guaranteed winning arbitrage strategy."

I'm no math nerd, but something tells me buying low and selling high might be a better strategy...

Expand full comment

Probably a mistake

Expand full comment

You linked to the wrong polynya twitter account. It should be https://twitter.com/apolynya

Expand full comment

Ethereum has most of the problems of Bitcoin, with fun new ones. Smart contracts require humans to solve an as-yet unsolvable problem: write software the first time correctly where almost any bug costs someone millions of dollars. It's not even clear a solution exists, let alone available now. Just like it's an unsolved problem how to store your wallet info securely. Anyone stating "crypto is clearly the future" is lying or selling you something. It's not clear Cryptocurrency is good at anything, other than separating a lot of gullible people from their money with the expectation that more gullible people will buy in and allow the early folks to make a profit. I'm still waiting for just one positive thing for which cryptocurrency is a good solution--they are good for money laundering, ransomware, and NFTs (which exist as just about the only thing you can buy with cryptocurrency), but those are negatives. Cryptocurrency is like leaving your money in a safe on your front doorstep, where anyone can pick up the safe and try to break in in the privacy of their home, as long as they return the safe to your front yard each morning. And you need to open the safe every day to see if you've been robbed. And you have no recourse if you've been robbed, and people will blame you, "Why didn't you put your safe in a tar pit in your front yard like we all do, you moron!".

Expand full comment

A bit late, but David Gerard has a rundown on Buterin's blog essays.

"Buterin has spent his life being told he’s smart and has important ideas. But I am deeply unconvinced any of the ideas in Buterin’s essays are worth anything — because they’re concocted without reference to anything that people who aren’t weird cryptocurrency anarchocapitalists want.

People read Buterin’s social essays, then they have to bridge the gap between the assumption that all of this must make coherent sense because it was written by an acknowledged genius wonder boy, and the undeniable fact that the stuff in front of you looks … stupid."

https://davidgerard.co.uk/blockchain/2022/09/16/vitalik-buterins-philosophical-essays-theyre-not-good/

Expand full comment

A bit late, but David Gerard has a rundown on Buterin's blog essays.

"Buterin has spent his life being told he’s smart and has important ideas. But I am deeply unconvinced any of the ideas in Buterin’s essays are worth anything — because they’re concocted without reference to anything that people who aren’t weird cryptocurrency anarchocapitalists want.

People read Buterin’s social essays, then they have to bridge the gap between the assumption that all of this must make coherent sense because it was written by an acknowledged genius wonder boy, and the undeniable fact that the stuff in front of you looks … stupid."

https://davidgerard.co.uk/blockchain/2022/09/16/vitalik-buterins-philosophical-essays-theyre-not-good/

Expand full comment

Two points for the lickmeister...

1. POS still uses difficulty and math and can be sped up with hardware...

2. Bitcoin doesn't care what you think... the final bitcoin is not expected to be generated until the year 2140, so if you live until then you can place a bet at that time.

Expand full comment

Small push back on the energy/stock vs energy/BTC dependency being different. ...Both BTC and stock prices are both currencies of energy prices......bankers can print money but they can't print energy (mostly oil)....

Expand full comment

A common position on the thread is “blockchains are interesting but inferior to other approaches and thus have no real use cases”. This argument works only by waving away applications you don’t care about as useless fads (NFTs, digital self-custodied money, etc). I find most mobile games completely worthless too but massive App Store revenues says a ton of people disagree.

As for the argument that other fintechs solutions are 100x better - maybe so, but to look at just one application, ask why, with ecommerce for 20+ years and credit cards for 70+, merchants still pay 1.5%-3.5% to visa, and similar to PayPal? For reference, avg profit margins for restaurant industry are like 5-10% so it’s not an insignificant number. If it’s as easy as spinning up some infra on AWS, given the massive VC funding for Fintech why haven’t these “orders of magnitude cheaper” solutions driven costs to near zero? The reason is bc it’s not so simple to replace a trusted network.

While Ethereum tech seems clunky, the clunkiness and expense vs other approaches is for one purpose - creating a base layer that lets 3rd parties build trusted networks for transacting with large groups at a very low startup cost. Think of it as a innovation enabler for a large swathe of use cases, much as cloud computing helped level the playing field between incumbents and challengers in a variety of industries.

Will smart contracts be the same kind of broad enabler as cloud computing or web apis? Maybe not - but people who don’t see a real use case are underplaying the value unlocked in simply offloading a ton of very hard work to the network. “Be your own bank” might be somewhat aspirational, but Uniswap with < 100 people is right now running a credible exchange settling hundreds of millions of $ in transactions a day. Can it replace nasdaq? Of course it can’t, today. Is it mostly enabling token-flipping, absolutely (though speculation has always been a big part of how exchanges work). The point is doing this at all wouldn’t have been possible without the ability to outsource a large part of their business building to the blockchain network.

It is easy to look at the rampant ponzis and overhyped products and just conclude “there’s no there there”. But look at how fast the eco system companies can move and start asking why and the picture changes.

Expand full comment

"If it’s as easy as spinning up some infra on AWS, given the massive VC funding for Fintech why haven’t these “orders of magnitude cheaper” solutions driven costs to near zero? The reason is bc it’s not so simple to replace a trusted network."

No, the reason is that the costs aren't zero. With a controlled digital payment system, such as a "debit card" in the USA, the costs can be very low. I don't follow things in the US, but a Dutch 'pinpas' transaction costs a few cents. The reason the cost is low in this case is that the transactions are all verified directly, and there is only a small cost per transaction once the verification network exists. But for things like credit cards, the costs are higher, because the users are paying for things like guaranteed payment (even if the buyer turns out not to have the funds) or the ability to charge back (if the buyer disputes the transaction).

The problem is not so much that "it’s not so simple to replace a trusted network", but that it is much MORE COSTLY to replace a trusted network with an untrusted one.

And if the "base layer" is "clunky" and "expensive", then there is no getting around this for things built upon it. One can't be successful with the model: "we lose money on every transaction, but we make it up in volume" (one may be able to do it for a short time if VCs are pumping in money, but in the end one has to recoup one's costs).

Yes, people "underplaying the value unlocked in simply offloading a ton of very hard work to the network" - if that "offloading" doesn't itself add value. If that "offloading" costs MORE than the alternatives, then doing it is a money-losing proposition.

You can talk about token exchanges, but what is their added value? It won't do to say, on the one hand "settling hundreds of millions of $ in transactions a day" (even recognizing that those are not actually "$", but some notional value of the tokens) and on the other that it is "mostly enabling token-flipping". You seem to want to argue that "well, yes, it may not have any value in itself, but it enables a bunch of other things" - where these 'other things' also don't have any actual value.

The bottom line for "blockchain" and "crypto" comes down to: they can do things that are worth doing, but only at a higher cost than other ways of doing those things. And they can do certain things better than other ways, but none of these are actually worth doing. (Leaving aside fraud, money laundering, and so on, that are arguably "worth doing" for some.)

Expand full comment

I'm not sure that the PoS chain can be rewritten more easily than a PoW chain in the case of a "many years rewrite" attack. Even for PoW, if you want to rewrite many years, you can reduce the cost a lot by bringing the difficulty way down: simulating more than 10 minutes between blocks, the protocol will accept a lower difficulty. You can do that for many cycles, and after some adjustments, it should be quite easy to rewrite lots of blocks.

Another problem that will hunt bitcoin if it is to survive for longer, are the lost coins: over time, less and less coins will be available. So they'll have to add a way to renew some coins.

Expand full comment