Moore’s Law and its corollaries indicate that the bitcoin block size could be doubled every two years without affecting decentralization of the network, according to Brian Armstrong, the CEO of Coinbase.
Armstrong has become the target of vitriol and apparently censorship from some in the bitcoin community who are opposed to block size increases.
He tweeted in late December 2015 that Coinbase was running Bitcoin XT, the controversial implementation of the Bitcoin code, as an experiment.
Coinbase is now running BitcoinXT (BIP101) in production as an experiment, blog post w more details coming soon https://t.co/x2GHvQ6UOO
— Brian Armstrong (@brian_armstrong) December 27, 2015
This was a bit surprising to me and I think really hurt the credibility of those sites, but perhaps in hindsight I shouldn’t have been surprised. What’s happening right now is an election in the bitcoin space, and like all elections (the U.S. presidential race being one prime example) things can get a bit heated even if they are overall a good thing.
In the fall of last year, various other big bitcoin players threw their weight behind BIP101, including BitPay, Blockchain.info, Circle.com, KNCMiner.com, itBit.com, Bitnet.io and Xapo.com.
Armstrong indicates the Coinbase is not wedded to any particular proposal, just that there needs to a block size increase.
I care less about which proposal actually wins. The bigger issue is which team of people and which fork we should support. The block size is one of many decisions that will have to be decided in the future as bitcoin evolves. We should care more about having a team with a clear leader (or decision making process) and track record of shipping code which makes reasonable trade offs than the topic de jour: block size.
The Coinbase CEO also answers the question of whether bigger block sizes hurt decentralization of bitcoin by pointing out a better questions to ask:
What is a reasonable upper bound on block size (now and in the future) that would get us the scaling benefits without putting the decentralization of bitcoin at risk?
He opines that, just like similar trends for CPU’s, bandwidth, and storage, we only need to look at Moore’s Law to figure that doubling the bitcoin block size every two years will not result in centralization.
Armstrong also indicates that it “doesn’t really matter” if you support doubling the block size every 18 months or every 4 years. It is axiomatic that the block size could be doubled at some rate without impacting decentralization. Moving from 7 to 4,000 transactions per second, the level of Visa’s network, is only about nine doublings away.