The Full Node Religion Will Fail (Or Why Lighthouses are Better Than Tiki Torches)
“Lobbying the government will never go away. So long as the government has the power to redistribute wealth and legislate prohibitions on human action, corporations will always try to control this power for their own benefit.”
You could find these words coming out of any Bitcoin supporter’s mouth… Especially the ones who consistently advocate for decentralization. The reason lobbying will never end is because of the government’s ability to disperse costs, and concentrate benefits.
This is a key phrase for anyone interested in public choice theory. It explains why governments are able to subsidize corporations, and why voters are rationally ignorant of these policies. Why would they invest time to learn about and vote on a subject which has so little impact on their own lives? They face the mirror opposite situation of the lobbyist. Dispersed benefits, and concentrated costs. Voting, however, is not the only time dispersed benefits and concentrated costs shows up.
Imagine an archipelago who’s inhabitants are waiting for shipments of goods. The ships have to navigate through these islands, sometimes at night. Initially their leaders taught each person on these islands that it was their religious duty to help light the way for these ships to pass through. So they all lit up their tiki torches on the coastline of their islands. Over the course of time however, people noticed that even though they forgot to light up their torches some nights, the other torches on the island were able guide the ships anyway, so some of them decided to free ride on their neighbor’s torches. Eventually it got to the point where there wasn’t enough light, accidents started happening, and one ship crashed, with the entire shipment lost. This is what is known as a public goods problem.
To solve this problem, somebody clever proposed putting up lighthouses. Each person who sent a shipment was required to pay a fee to their local island’s lighthouse keeper in order to ensure that such an accident wouldn’t happen again. Some decried this solution as being centralized, and that they’d have to trust the light keepers to keep the lights on and not collude to steer the ships away toward a different port where they could steal the shipment… but the initial failure of the tiki torch solution was enough to sway most of the community to crowd fund these lighthouses.
If it’s not obvious right now, tiki torches are full nodes, lighthouses are the miners, and shipments are the current consensus protocol chain. As I articulated in my last post, running a full node is nearly useless for most users. Not only that, but these tiki torches do not add to anybody’s knowledge about the current consensus because Bitcoin has an added problem. It is trivially easy for adversaries to put up tiki torches in the middle of the water way in order to redirect ships toward crashing.
It is not trivially easy to put up a lighthouse however. So how was the public goods problem of creating a lighthouse to give a strong signal for what is actually believed to be consensus solved?
By making it expensive for the dishonest, and profitable for the honest. By giving miners tokens for putting together blocks, the network rewarded the hash power they supplied as a signal to everybody what the consensus is. By validating transactions, they received fees for including transactions in those blocks. If a miner creates a block with invalid transactions, it’s overwhelmingly likely that other miners will simply ignore these blocks, return to their existing block in which they can claim greater fees from the current pool of transactions, and mine the reward that the invalid block would have received.
So really, it’s only when then when the miners disagree about consensus that trusting in hash power isn’t necessarily enough. This includes both hard forks and 51% attacks, when a malicious actor with 51% or more of the networks hashing power decides to fork consensus rules. The theoretical possibility of the lighthouses failing in their incentivized duties does not, however, change the fact that users can free ride on others to monitor the network and then, only when necessary, make a conscious decision on which chain to use.
Ultimately, the full node priests have acknowledged and lamented that the protocol failed to reward running full nodes. They have turned toward religious duty and “education” (quoting people equally unable to bestow decentralization into their systems) in order to decrease trust in the lighthouses, and increase trust in the validations written into their implementations of the protocol.
Ironically, they’ve even shot themselves in the foot with the incentives to run full nodes, and looked only at one side of the equation: costs. By focusing solely on ensuring costs of running full nodes are low, they’ve neglected, and even harmed, the gains from running one. Ideally you don’t even want to make it inexpensive for all. You want to make it inexpensive for the honest and expensive for the dishonest.
In my next article, I will discuss how the network has decreased the utility of running non-mining full nodes for the honest, and how this can possibly be fixed.