None of Bitcoin's recent troubles -- technological glitches and the collapse of the biggest cryptocurrency exchange, MtGox, the departure of key figures from the bitcoin Foundation -- have struck me as terminal. Now, though, I'm starting to think bitcoin needs trust to survive -- a potentially fatal flaw for the virtual currency.
To cryptocurrency enthusiasts, "trust" is a dirty word. "We have proposed a system for electronic transactions without relying on trust," Satoshi Nakamoto wrote in the original manifesto proposing the decentralized currency. That's what made the system better than traditional fiat money: It purported to make sure, as a matter of technology, that transactions would be carried out as intended and without the costs associated with the presence of a trusted middleman, such as a bank, in any traditional payment system.
Then, last month, a mining pool -- an association of people who own specialized hardware that produces Bitcoins by running specific mathematical operations -- gained a 51 percent share of computing power used in currency production. That gave it the power to reverse transactions, to pay twice with the same Bitcoins, and to delay or even cancel other people's transactions. The pool was GHash.io, run by a secretive U.K.-incorporated outfit called CEX.io, which came close to topping the 51 percent mark in January. Back then, it promised not to do that: "GHash.io will take all necessary precautions to prevent reaching 51% of all hashing power, in order to maintain stability of the bitcoin network," the pool said in a statement.
Bitcoin transactions are recorded in the block chain, a kind of public register. Transactions are broadcast and validated by other users. A group with more than half of the system's mining power can change, duplicate and stop transactions; even if every other user joined forces, they'd have insufficient resources to prevent the abuses.
Now, weeks after that promise was shattered, GHash is making new ones. "If GHash.io approaches the respective border, it will be actively asking miners to take their hardware away from Ghash.io and mine on other pools," it said in a fresh communication.
In other words, the stability of the electronic currency and all the infrastructure that has grown around it in the past five years -- at least $240 million worth of it in venture capital investment alone, most of it made last year -- now depends on the goodwill of a few people whose names nobody knows. "Jeffrey Smith" -- the name used by GHash.io's only spokesman -- is likely a pseudonym.
There's no guarantee that a certain group of people doesn't control more than 51 percent of bitcoin emission even now: GHash.io has 39 percent, but another 17 percent is ascribed to "unknown" pools by Blockchain.info, the best source on mining-pool shares.
These unknowns are unlikely to undermine the ecosystem they inhabit; cryptocurrency enthusiasts' self-interest is probably what has sustained bitcoin through all its tribulations. I'm not willing, however, to ignore the risk that a group of people who won't even tell me who they are might take the money and run.
Trusting them would be akin to acknowledging them as the system's central bank, something Nakamoto emphatically didn't want it to have. Paradoxically, it's easier to trust central banks that print fiat money, because the people in charge and their interests and motives are more or less known, or at least are the subject of much study and speculation.
Gavin Andressen, bitcoin Foundation's chief scientist, who typically downplays the currency's technological shortcomings, wrote last month that GHash.io's dominance wasn't "disastrous" because its advantage wasn't easy to misuse. He added, however, that "bitcoin is still a work in progress, and you should only risk time or money on it that you can afford to lose."
To anyone who sees value in the technology suggested by Nakamoto, and expects it to eventually be adopted by governments and firms to create frictionless payment systems, this means bitcoin should not be treated as a currency but rather as a complex real-world experiment that may lead to different outcomes. Those who participate with the goal of making money are only providing data for researchers and funding their work. They may or may not be rewarded for their interactions: It's a matter of trust.