False analogies in Bitcoinland plus Alibaba

Saw two analogies used today that are inaccurate.

Writing at Forbes, Eric Mu interviewed Jake Smith, better known as “The Coinsman.”  Jake was responding to a comment I wrote last month:

Tim Swanson, the author of The Anatomy of a Money-like Informational Commodity, recently said that you missed the “unseen calculation, the economics of extracting and securing rents on this ledger unit, which consume scarce resources from the real economy.” – Do you think he is wrong?

I think attacking mining from an environmental point of view is quite silly, because pretty much everything in the modern era relies on resource consumption, and for the vast majority of those things society has decided that the trade-off is worth it. I think Bitcoin is one of the most valuable and revolutionary inventions the world has ever seen, so even if it is using a lot of electricity I don’t think that’s a valid criticism against it. The internet uses vastly more electricity than Bitcoin, no one is bashing the internet for using resources.

Swanson’s quote would also imply that Bitcoin is not part of “the real economy”. I would say that by virtue of its existence, it is.

Further, Bitcoin’s value is derived in part from the fact that it is difficult to create.

The biggest problem with the analogy above, which is commonly used by Bitcoin advocates, is that it is not an apple’s to apple’s comparison.  In this instance, Bitcoin acts as a distributed Excel workbook, a spreadsheet application that uses the internet to distribute itself.  Thus it is incorrect to equate it with the much broader umbrella that is the entirety of the internet.

This same problem happens when people claim that Bitcoin can and/or will replace the banking system.  For instance, last month Jake interviewed Nan Xiaoning, CEO of Bitocean:

I think Satoshi had a lot of foresight in this regard. He wasn’t a dummy, I’m sure he considered different ways of distributing coins.

Some people say that bitcoin wastes a lot of electricity. But the banking industry surely uses more resources than bitcoin does. But bitcoin is a peer-to-peer system. I think using resources to guarantee its security and stability is the way it should be.

Another inaccurate analogue/comparison.  Bitcoin’s protocol does not provide any of the functionality of the banking system beyond a security lock box (that should not be confused with a distinctly different term, a savings account) and a corresponding ledger of access and usage (the debate over whether or not someone “owns” a privkey corresponding to a UTXO it is still being argued over by lawyers globally).  The current protocol does not natively allow for lending, saving, notary, underwriting debt and equity or setting of interest rates (among many other services real banks actually provide).

In both cases above the examples above miss the forest from the trees.  As Robert Sams pointed out a few days ago, the proof of work mechanism used in Bitcoin was designed to make Sybil attacks expensive.  The verification process is a marginally trivial task and can be handled (and in practice actually is handled) by mining pools via small computers such as a Pi-based box.

How specialized is the hashing (not verification) process?  A good comment on reddit yesterday noted that:

Rather than taking the whole header, they mine using something called a midstate. Due to the nonce being at the end of the header, the software hashes up to just before the nonce, and then sends that (called a “midstate”) to the mining chip. The mining chip then only needs to add a nonce, do the end of a SHA256 round, and then one more, and then check if the result is good enough. Rather than returning data, they just return nonces which look to be valid.

Instead, a more accurate way to look at this issue is from the spectrum of centralized to decentralized (which was also discussed by the Hyperledger team in an interview a couple days ago).

Centralized tools and services have certain vulnerabilities (e.g., single point of failure and potential abuse) but its cost basis is different than say, a decentralized entity.  The economics of both need to be accounted for (and are) when rolling out a new system internally (this is called the Total Cost of Ownership).

On the other end, decentralized systems are less vulnerable to some of the same issues that centralized systems are, yet to make them less vulnerable in fact requires consuming scarce resources that centralized solution do not have to (because they are trusted networks).  In the case of Bitcoin, bitcoin miners (or technically hashers) effectively destroy (or “burn”) a corresponding amount of energy (technically exergy) to protect the network from Sybil attacks on an untrusted network.  This is a real cost that cannot be ignored yet as shown above, is often handwaved away.

[Note: as an aside, most miners, mining farms and mining manufactures do not pay for their capex or opex in bitcoins, nor is this likely going to change anytime soon.  Instead they must rely on and permaborrow the unit of account of fiat (typically a USD or RMB) to effectively measure and allocate resources.  This unit of account issue — wherein economic activity within the Bitcoin world is measured with the unit of account that is fiat to create this network — was also broached by Robert Sams several months ago.]

Furthermore, as I mentioned in chapter 8, if the TCP/IP analogy was correct then the marginal revenue for ISPs would split in half every 4 years.  And that through competition the marginal cost of protecting and sending packets would equal the marginal value of those packets.  This would not be an effective way to run a business let alone design a network topology.

In the real world, the marginal costs of running an ISP, which is centralized, have to be less than the marginal revenue otherwise they go bankrupt as they could not pay for overhead.  So yes, in fact, ISPs do try to actually mitigate the leakage, wastes and otherwise inefficiencies in its own internal network and they do this through a myriad of ways.

Bitcoin’s existence is on the other side of the spectrum.  Bitcoin was purposefully designed to make it cost prohibitive to spam ownership change on a public, untrusted network — the complete opposite in organization that an ISP is designed to operate as.  The average person would likely see this as inefficient, but that is because up until the past decade — with the advent of Bittorrent and other distributed systems — the public at large was unfamiliar with how these systems are designed.  And as Sams pointed out, using the word “efficient” versus “inefficient” may not be the most accurate terminology, because each model has different attack vectors they have to account for.

Thus again, it is not about being pro or anti proof-of-work.  Rather it is acknowledging that proof of work requires a certain economic model that have real costs that scale with token value and in the case of Bitcoin, is not environmentally “greener” than some centralized solutions (e.g., ApplePay).

The case of Alibaba

Over the past couple of days some Bitcoiners have recently claimed that the recent dip in market prices for bitcoins is because of the Alibaba IPO; “Alibaba’s US IPO May Have Crashed the Bitcoin Price.”

Not only does this show that several vocal Bitcoiners are unfamiliar with how real IPOs work (underwriters typically represent the lion’s share of additional equity ownership and the date is fixed weeks and months ahead of time) but that it illustrates how some Bitcoiners like to blame people and go on a witch hunt when prices decline but then reassure themselves that they are investment geniuses when prices trend northerly.

In point of fact, the Alibaba IPO was not a surprise to anyone, the investors are all large financial institutions and not hoi polloi.  The IPO was oversubscribed and not even well heeled, well connected HNWIs could get into an allotment — only banking institutions were able to because of the enormous demand.  And none of those institutions are: 1) large bitcoin holders and 2) needed to sell bitcoins to raise funds to buy Alibaba shares.

Perhaps this will change in the future, but that is not the case in this instance (be sure to also check out Izabella Kaminska’s lively twitter feed).

Send to Kindle

Leave a Reply

Your email address will not be published.