Tuesday, January 20, 2015

A Solution To Everything?

Twenty-five days ago I wrote a critique of the idea that blockchain based crypto-currencies could be a basis for peer-to-peer storage. My main argument was from an earlier post, that economies of scale meant a successful currency would not in fact be decentralized. But the occasion for the post was the observation that Bitcoin's value was collapsing, reinforcing the economies of scale argument by reducing the income of miners, and thus forcing all but the lowest-cost (largest) miners to drop out, as shown by the hash rate ceasing to grow.

Since then, there has been a flood of proposals to base other P2P storage systems, election voting, even a replacement for the Internet on blockchain technology. Every one of these proposals for using the blockchain as a Solution for Everything I've looked at appears to make three highly questionable assumptions:
About a year ago Bitcoin peaked at over $1100. When I wrote that post a Bitcoin was worth $319. Today it is worth $215. It has lost one-third of its value in less than a month, and over 80% in a year. In less than a month miner's revenue has gone from about $1.4M to less than $800K, a drop of 43%. Below the fold, more on the implications of this price implosion.

MIT's Technology Review has a good overview of the problems the price drop is causing:
The Bitcoin software has a mechanism that is designed to ensure there are always enough miners working to keep the currency operating and to regulate their output. It does that by altering the difficulty of the work that mining software has to do so that their miners’ combined rate of output is always the same. ... Now that interest in mining is softening, the mining difficulty is set to drop so as to keep the Bitcoin network running correctly. That would give miners a break on their power bills.
Harvard's Prof. Benjamin Edelman, says:
“I think that miners are pulling out and dumping Bitcoin to make up for lost revenue, which itself changes the value,” ... Those that do continue mining will contribute to the problem because they will have to immediately sell any new Bitcoins they mine, he says.
But the Bitcoin ecosystem includes more than just users and miners. There are many companies competing to produce better and better mining chips. Each generation of chip leads the pack for a short time before being rendered uneconomic by a successor - so short that the vast majority of the chips will never mine a Bitcoin. This means that the first (largest) customer for the chips has a major advantage in mining costs, so large that the FBI caught Butterfly Labs delaying customer shipments to mine with them on its own behalf.

Chips take a while to design, manufacture, test and ship, so there is a pipeline of new chips representing sunk investment by the mining chip companies that they need to recover. So new chips are going to continue to enter the market, reinforcing the cost advantage of the large miners. But with their income halved, they can't pay as much for their chips. Just as the miners need to dump Bitcoin, the chip companies need to dump chips to cover costs. Just as miners will be forced out of the market, so will chip companies be forced out.

Clearly, a technology with this much volatility is a wonderful basis for gambling - shorting Bitcoin would have been a terrific investment over the past year had it been possible. But why would anyone think that it would make a suitable basis for any important social function, such as elections, or long-term information storage? Perhaps because they never stop to ask themselves "What could possibly go wrong?"


  1. Anyone thinking that the blockchain is a Solution to Something needs to read Henry Farrell's excellent Dark Leviathan about the Dread Pirate Roberts. It isn't pointing out flaws in the blockchain itself, but in the application of pseudonymous technology to real-world problems.

  2. The New York Times weighs in on the blockchain is a Solution to Everything meme. Surprisingly, they sound some cautionary notes, such as:

    "The virtual currency must also maintain some value for the network to work. Bitcoin miners, the computers that drive the block chain, win Bitcoins if they successfully solve complex cryptographic problems. If Bitcoin’s value continues to drop, it could become too expensive to keep the computers running."

  3. Now apparently IBM is infected with the meme, thinking about an "official" blockchain of transactions in real currencies run by central banks.

  4. Charlie Shrem, founder of the Bitcoin Foundation, has a clear explanation of how the Bitcoin blockchain works to facilitate transactions as he considers how Bitcoin might work in the prison in which he is about to be confined for aiding and abetting an illegal drug buy on Silk Road via his "unlicensed money transmitting business".

  5. And Ars Technica has details of a newly-elected board member of the Bitcoin Foundation's claims that the previous board members looted the Foundation.

    In the context of this and the previous comment it is worth reading this book review.

  6. Nathaniel Popper has a book extract at the New York Times with details of the progenitors of bitcoin including Adam Back's "hashcash" and Nick Szabo's "bit gold".

  7. As well as getting looted by the bad guys and by the Feds who are investigating you, your wallet may have bugs that send the Bitcoins to the wrong address.