Mr. Fantastic: A (Focused) Response To Nouriel Roubini’s Senate Testimony
It was bound to happen eventually: economist and noted crypto FUD spreader Nouriel Roubini, aka “Dr. Doom”, testified to the US Senate Committee yesterday, trashing all things crypto and blockchain related.
All in all, the presence of a ruthless critic like Roubini in front of the Senate is certainly a good thing; all too often, those who have more money and/or political power than technical knowledge can be far too gullible in their willingness to devour the latest hype from the techno-utopian crowd. The Blockchain bubble may be one of the more severe examples of this the world has ever seen; in short, the industry is in desperate need of being taken down a few pegs.
And indeed, many of the barbs Roubini casts are aimed at appropriate targets: Tether is an out and out disaster who’s condemnation shouldn’t rest on the shoulders of a small cadre of relentless tweeters and bots; mining pool centralization is indeed a problem, and the common retorts to this are weak; market manipulation is rampant and won’t sort itself out unless concrete action is taken; enterprise blockchain may actually be the “most over-hyped technology ever”; smart contract bugs are no joke; the ICO space is overridden with irresponsible claims, useless products, and outright scams; And so forth.
It’s a shame, then, given all the things he was saying that indeed needed saying, that the useful shade Roubini throws is sullied by a number of misrepresentations, exaggerations, and outright falsehoods that are peppered throughout his testimony, particular as he veered into more technical matters.
I don’t intend for this to be an all-encompassing response to the testimony, which covers a very wide variety of topics. And I’ll try to avoid going into the broader issues regarding the potential value prop of crypto in principle; as someone in the crypto space, I obviously disagree with his outright conclusive dismissal, but his general arguments are fairly common talking points in the discourse, and are already widely debated elsewhere.
Instead, I’ll focus on a small handful of specific points that stuck out as particularly sloppy:
Scalability
Throughout his testimony, Nouriel avers that public blockchain scalability is simply impossible; he repeatedly refers to Vitalik Buterin’s “scalability trilema,” which states that when it comes to “decentralization, scalability, and security,” blockchain protocols have difficulty simultaneously achieving more than two of the three.
At one point, he enumerates what he claims are the three noteworthy scaling approaches the industry is taking:
- Creating more altcoins
- Increasing block size
- Utilising sidechains/merge mining
He is correct in saying that each of the above three scaling avenues come at the cost to decentralization and/or security. But the bizarre thing about that list — which anybody who even casually follows crypto will immediately notice — is that it omits the approaches that have absorbed the overwhelming majority of scaling research & development in the space.
Take Bitcoin and Ethereum (which are, by virtually any metric, the two most popular cryptocurrencies). The Bitcoin community focuses most of its scaling energy on layer 2 payment channels in the form of the Lightning Network; Ethereum has a number popular layer 2 state channel projects — most notably in the Plasma domain — and is also pursuing scaling on the base layer via sharding. I don’t think it’s an exaggeration to say that if an engineer or a technical researcher tells you that they’re working on blockchain scaling solutions, you could safely bet with 95% certainty what what they’re working on is related to either layer 2 or sharding.
Number of times layer 2 solutions or sharding gets even mentioned in the entire testimony? Zero.
Here, he seems to come close to accidentally stumbling upon the idea of sharding himself:
Just as we cannot record all of the world’s transactions in a single centralized database, nor shall we do so in a single distributed database. Indeed, the problem of “blockchain scaling” is still more or less unsolved, and is likely to remain so forever.
Which, given that he pretty clearly hasn’t even heard of it, I suppose would have been impressive.
You’d think someone so insistent that scalability can never be achieved would want to rebut — or at least even acknowledge? — the biggest current attempts at doing so, but alas.
Proof of Work vs. Proof of Stake
About Bitcoin, Roubini says: “It is not scalable given its proof of work (PoW) authentication mechanism — that allows only for 5 to 7 transactions a second.” He contrasts this with the planned implementation of Proof of Stake (via Casper) on Ethereum, which he says promises to be “vastly scalable” but “will be massively centralized and thus not secure.”
Public blockchain protocols do indeed have a low transaction throughput, but this is a cap that’s deliberately built into the protocol regardless of the POW/POS distinction. In other words, the reasons for this cap (which I won’t go into), and the optimal TPS choice are considerations that are essentially independent of the protocol’s use of POW or POS.
In short — POS isn’t a scaling solution*. Indeed, in discussions around Casper’s POS component, the main purported benefits that are cited do not include scalability, but rather are:
- No inefficient energy consumption
- Better security guarantees (via ability to implement explicit economic penalties)
- Less validator centralization
On point #3, Nouriel claims that POS will actually lead to more miner/staker centralization than POW. He may be right; it isn’t live yet, so only time will tell. But instead of countering the arguments that have been put forward by the Ethereum camp (like that POS’s reward system invites less economy of scale than POW, for which pooling hardware resources yields super-linear gains, for example), or even acknowledging that POS is seen by many as an avenue explicitly to avoid centralization, he simply states his conclusion as self evident.
*One caveat to the above: there is a subtle way in which POS could have some scalability benefits: in POW, there is, in principle, no way to predict from which addresses valid blocks will arrive. With POS, however, one has economic assurance that addresses with staked Ether will continue to produce blocks for a given time period; thus, one can better optimize consensus among producers, eliminating inefficiencies and yielding indirect scalability gains. These gains, however, are minor compared to the orders-of-mangitude scalability gains purported by the approaches discussed in the last section. Plus, given the general imprecision around this point, it’s hard to give Roubini the benefit of the doubt here about this being what he had in mind.
51% Attacks
In smaller coins with a small market capitalization you don’t even need a 51% hash power to mount a successful 51% attack.
Huh?
Sure, there are cases where exploits that are discussed as “51% attacks” actually didn’t require the full 51% hashrate; I’ve written about how this was possible for the Verge Hack (s), for example. But in that case, the low hash-rate attack was made possible by the nature of the time-warp exploit in conjunction with Verge’s multiple mining algorithm protocol, and had nothing to do with Verge’s market cap.
And sure, smaller market cap coins are easier to attack by virtue of the simple fact that they have net lower hash rate, but you still require 51% of that hashrate to attack them. Genuinely not sure what he’s talking about here.
…Okay One More Thing About Mining
“It is so energy-intensive (and thus environmentally toxic) to produce, and carries such high transaction costs, that even Bitcoin conferences do not accept it as a valid form of payment.”
Huh?
High transaction fees led some conferences to stop accepting Bitcoin, yes, but what does that have to do with the energy that goes into mining? The energy cost of block mining is borne by the miners, not by parties receiving payments. If his claim is that increasing energy cost of block production is what leads to higher transaction fees (I can’t tell if this is what he’s getting at?), he has to reckon with the fact that historically, the two have had no correlation.
Cold storage
Or spend a fortune to put your crypto assets into “cold storage”, ie a digital storage that is disconnected from anything online… But leaving aside the cost of such stone age security solutions the implication becomes that your crypto wealth — hidden in deep cold storage — cannot be easily traded or used for transactions of any sort. This is the contemporary equivalent of mining gold deep from the ground and then hiding it in the form of gold ingots back deep in the ground.
The two most popular cold storage wallets are the Nano Ledger and Trezor; they cost $100 and $80, respectively. As someone with decidedly non-whale wealth status, I still personally imagine a tad more money than that when I hear the word “fortune.” Also, having a separate device will always be less convenient, yes, but anyone who’s used cold storage knows it’s still fairly easy; you can keep your device within reach (it’s the backup of the private key that needs to be better hidden) and can broadcast a transaction in, I’d say, 30 more seconds than it takes to send one via a standard desktop wallet. Plus, increasingly, major wallets and applications (i.e., metamask) offer direct cold wallet integration. Claiming “your crypto wealth — hidden in deep cold storage — cannot be easily traded or used for transactions of any sort” is nonsense.
Speed/ Efficiency
Institutions, particularly those engaged in algorithmic trading, need fast and efficient transaction processing. For their purposes, a single globally distributed blockchain such as Ethereum would never be useful and they will never use it.
Once again, lack of scalability improvements is taken as a foregone conclusion. But even that aside, this argument in this form is pure strawmanning: of course processing transactions on a distributed blockchain is slower and more computationally intensive than doing so on a private server. The entire argument put forth by crypto proponents is that there are cases in which more cumbersome computational processes is a worthy trade-off for minimising or eliminating the human “wetware” of trustful coordination between parties, and can actually lead to greater overall efficiency (settlement time in transferring large sums of wealth via Bitcoin vs. via wire transfers being a concrete example.) Maybe we’re wrong, and ultimately such tools will never be adopted, but again, instead of actually engaging with the arguments, Nouriel simply throws out “fast and efficient transaction processing” as though that qualifies as a counter.
A Gaggle of Exaggerations
Centralized exchanges are being hacked daily
Moreover, hundreds of other cryptocurrencies are invented every day
Days oft go by without an exchange hack, and with <100 new cryptocurencies emerging.
Hundreds of stories of greedy crypto-criminals raising billions of dollars with scammy white papers
As far as I can tell, the only crypto projects whose fund raising has exceeded $1 Billion are EOS and Telegram; let me know in the comments if I’ve missed 98 others. (Pleading the 5th on those two being “greedy crypto-criminals”)
The early internet in the early 1990s saw a rapid boom of applications and explosion of user adoption: email became widespread and thousands of useful website used by millions of people for useful purpose sprang overnight.
The first email was sent in 1971; widespread email adoption was not achieved by the following morning. If what he means is “email wasn’t widespread until email was widespread,” then yes, that’s true.
R.e. Ethereum:
And there is nothing immutable in the “code is law” motto as the developers are police, prosecutors and judges: when something goes wrong in one of their buggy “smart” pseudo-contracts and massive hacking occurs, they simply change the code and “fork” a failing coin into another one by arbitrary fiat, revealing the entire “trustless” enterprise to have been untrustworthy from the start.
Go ahead and criticise Ethereum for the DAO fork — I’ve done so myself. But to clarify: this has happened literally once. To someone unfamiliar, which is most people this testimony was intended for, Nouriel gives the impression that Ethereum’s is a history of willy-nilly forking to revert valid transactions. This degree of misrepresentation is indefensible.
Furthermore, regarding scammy dapps on Ethereum, he criticises the community for:
doing nothing — literally nothing to stop or block such Ponzi games
Ethereum is apparently guilty of being both too active and not active enough in its governance; maybe Vitalik did solve quantum computing after all.
Conclusion
I could go on, but hopefully the point is made. I have no strong reason to think that Dr. Doom isn’t a fine economist, a kind friend, and a good neighbour. But his knowledge of the crypto space, and his willingness to do responsible due diligence and construct intellectually honest arguments, is clearly lacking.
This industry n̶e̶e̶d̶s̶ has better skeptics; David Gerard, Preston Byrne, and Angela Walsh come to mind. There is no shortage of folks out there worth listening to. Giving the spotlight to the figures who are best at inciting ire, hurling insults and drawing attention to themselves won’t improve the discourse in this space, discourse which unfortunately continues to have some serious room for improvement.
For inquiries: daniel@theabacus.io
