http://greenwire.us/articles/it-security/cryptocurrency/dr-bitcoin-e02-the-unproven-hypothesis/Andrew puts forwards the argument about if the Bitcoin network ends up being rationally driven rather than the altruistic p2p model, then you could reach a point where some nodes start fucking over other nodes for self profit. If the incentives model is not correctly designed, Bitcoin could implode and that this is more of a risk than the 51% attack.
This reminds me of the Hawk and Dove model in evolutionary biology:
http://en.wikipedia.org/wiki/Chicken_%28game%29#Hawk-DoveBitcoin might have a possible built in solution. When Hawk and Dove models is used to simulate an ecological population, you often get a stabilising point. If you take a population of Doves and a few Hawks invade, then the invader species will have a big initial advantage. But after some time of swinging the population will swing less and settle on a proportion. This is called an Evolutionary Stable Strategy.
http://en.wikipedia.org/wiki/Evolutionarily_stable_strategyIn the podcast, they use the example of miners on the blockchain. If a block contains a huge transaction fee of like 1000 BTC, then rational self interested miners will fight over the block and the blockchain will never have a chance to progress. It may be then that the best strategy is actually taking only a small slice and paying the rest to everybody else (pay off the hoard), so that others will build off your block instead of fighting for it.
When two valid blocks are created at the same time, they are in a race. They compete to have the next block built off them. Part of the problem is that miners build a new block before they discover the newly built block. This leads to a forking of the blockchain with the newly built block rejected.
Lets start by thinking how to solve this problem.
Ideally you'd want miners to discover new blocks before they create their own blocks. This step can be helped by clearing specialised paths between miners of high connectivity. The only use case for these paths is the rapid pulsing of new blocks throughout the network of miners (broadcasting).
Cutthroat efficiency is the key focus. Rapid deliverance of blocks to the targets (other miners) at low cost. The network's sole traffic is one node telling every other node about new blocks. Based on cost and topology, it's then conceivable to imagine hubs emerging. As the hubs grow, they condense outwards forming a backbone. We've seen this behaviour in other large scale networks with high traffic requirements like the internet. Attached nodes to this backbone are then given broadcast prioritisation based on how much they pay.
Our hypothetical miner loses 1% of mined blocks per month because they become orphaned. He calculates his average monthly loss to be $100. Then a broadcast service offers to cut his 1% to 0.1% for $20 per month. The miner now experiences a monthly loss of $10 + the $20 for the service. He has gained $70.
It doesn't matter what the actual numbers are. As the market expands, there will be a tipping point where the loss in expectation for a miner overcomes the cost of this service. As that industry grows and the services have to expand to meet capacity, then we see the gradual emergence of a backbone amongst miners.
When competing blocks are created, miners want other miners to accept and build off them. They might start paying other miners to accept their blocks. In effect saying “build off my block and I'll give you a share of the reward”, with a tit-for-tat relationship built up. Miners then must evaluate the likelihood of blocks becoming orphaned versus what share of the reward they will get from that block.
Since time is crucial, established agreements would become normal to codify this behaviour. You don't want to lose time negotiating. This behaviour becomes paying miners to prioritise your blocks in general. It's conceivable that a whole industry will arise of companies evaluating and scoring the mining capacity of various mining operations.
This has economic effects and could flatten the reward structure while decreasing variance among large scale operations. Mining could be a natural monopoly whereby miners experience increasing rates of return per unit as they scale their operations upwards. A simple example is the complex cooling arrangements increasing employed by mining farms. Custom setups or infrastructure might offer a far better return than small scale operations.
Against this backdrop, it's not hard to imagine a gentlemen's club forming. Everyone involved in that industry knows one another, and they have their own set of unwritten rules and agreements. The internet is like this.
This private club creates mutual infrastructure to improve their operations and establishes a backbone. Access to the backbone is prioritised either based on available mining capacity or how much you pay. There would need to be a mechanism for newly mined blocks to escape this backbone to the wider Bitcoin world. I imagine a second tier caching blocks and feeding them outwards on a slower network.
I don't want mining to be controlled by a cartel, but it might be the kind of industry which naturally tends towards that state because of economies of scale. The Ripple concept essentially depends on this concept more than Bitcoin does, and if it's a problem with Bitcoin then it'll be a huge problem with Ripple.
https://ripple.com/wiki/ConsensusI'm not proposing any change. I'm asking what will happen in the future.
