In most situations, doing something more often makes it easier for us. When this is the case, we’re dealing with a positive cost-learning curve. In some cases however, something becomes more difficult for us as time goes by. In those cases, we’re dealing with a negative cost-learning curve. There are sometimes exceptions to this rule, in that some things become more difficult as we do them more often. There are different reasons this may happen. If you’re engaged in a subversive activity, others may implement precautions that make your activity more difficult and costly over time. If your activity depends on the utilization of scarce resources, eventually you’ll find yourself scraping the bottom of the barrel, at high cost. In other cases, you discover vulnerabilities in your project that require increasingly costlier measures to protect yourself against them. Nuclear energy is one of the few technologies found to have a negative learning curve, for this exact reason. As we use nuclear energy, we discover new ways in which the power plant can fail, requiring more expensive designs to protect against such failure.
It seems to me that Bitcoin is another rare example of a technology with a negative learning curve. As Bitcoin is used more often, it becomes costlier to use. There are different reasons for this, some of which can be fixed easily, some that can’t be fixed and others somewhere in between. To start with, the most important thing to understand about Bitcoin is that it’s a giant attack vector. Because it’s decentralized and transactions are designed to be anonymous and irreversible, the world’s criminals and crooks are given a massive incentive to abuse the project in various ways as it grows in value. One effect this has it to make it more difficult to purchase Bitcoins. If you wanted to get some in 2009, you simply left your computer to waste some electricity for a while. In the years since then, you could buy some through shady online platforms. Today you’ll have to pass strict verification requirements and submit a long list of documents, due to the abundance of fraud. This prohibits Bitcoin from being a cheap and easy payment system. Similarly, securely storing your Bitcoins now requires you to purchase specialized machinery.
Another interesting aspect where the negative learning curve is seen is in transaction fees. If more people make use of a product, costs can be saved due to economies of scale. The product becomes less expensive. In Bitcoin, an increase in use of transaction capacity is not met by an increase in supply. Bitcoin thus becomes more expensive to use, as more people want to use it. It can no longer serve as a cheap payment method, due to the high costs to get your transaction confirmed. Worse, it’s now impossible to predict in advance how long it will take for your transaction to be validated, so it can’t be considered fast anymore either. It’s original purpose is thus retconned: Bitcoin was never about cheap or fast transactions! This problem is relatively easy to solve, but its designers decided not to solve this issue, because of a much bigger issue that’s far more difficult to solve.
If transactions are expensive, the network pays for itself. When transactions are not expensive, the network depends on its inflation to function, but the inflation eventually comes to a stop. If we observe Bitcoin’s inflation, we notice the process of producing new Bitcoin suffers from another negative learning curve. As Bitcoin is used by more people, producing a new Bitcoin starts to demand more resources, due to a zero sum game that awards the right to generate new coins to someone who solves a particular computer puzzle. As people come to the conclusion that Bitcoin is here to stay, they grow willing to devote more and more resources to producing new Bitcoins. For this reason, the amount of electricity spent per individual transaction has grown over the past few years, from an absurd amount to an even more absurd amount. It continues to grow, until producing a new Bitcoin requires as much money as the Bitcoin is worth.
Once this happens however, a new problem emerges: The people who mine the coins, have no real incentive to play along with the network’s rules. If you’re earning a profit of zero dollar when mining Bitcoins, you play by the rules because you expect to earn more from your endeavor in the future. However, because the network’s inflation eventually drops, you can expect your business activities won’t be profitable in the future either. Once this happens, you’ll have an incentive to cheat the system. One easy way to cheat the system is to stop confirming transactions that pay very little fee. You’ll create transactions for yourself and include those in your blocks, while not bothering with transactions that don’t pay you a lot of money. A more daring way to cheat the system is through a 51% attack.
Once you have no clear financial incentive anymore to play by the rules, the network’s users eventually become forced to band together in one way or another, to pay you to play by the(ir) rules. The system’s participants need to convince you that you stand to gain more by working in their interest, than you do by abusing the rights awarded to you. One way this could take place is by simply paying high fees. Alternatively, the big Bitcoin investors or exchanges might start to subsidize the miners. Once this happens, the argument that Bitcoin is decentralized ceases to make increasingly little sense of course.
So, a very important factor in Bitcoin’s negative learning curve is the zero sum game underlying the project. The zero sum game awards money to users, which gives them an incentive to participate. As more people start to participate, the system becomes more expensive to run, while the incentive the participants have to play by the rules starts to shrink. The participants in the game need to start to cooperate to win the game, which makes its gradually easier for participants to play against the rules. Because most people don’t overnight step into an obscure potentially profitable endeavor, Bitcoin suffers from a negative learning curve.
Another clear way in which the system suffers from a negative learning curve is the way in which it becomes less anonymous as time goes by. Its anonymity was one of the main perks according to its proponents. However, anonymity is a subversive activity. You seek to hide yourself from others. The people you seek to hide yourself from learn to see through your disguises, which requires you to purchase increasingly more elaborate invisibility cloaks. This is effectively what happened to Bitcoin. If you want to buy or sell Bitcoins, you could originally just trust the system’s anonymity. Then blockchain analysis came along, revealing your identity. This requires the use of mixing pools, but exchanges were increasingly scrutinized and refused to accept Bitcoins originating from mixing pools. The number of layers you’re forced to push your coins through thus starts to grow, as observers learn to recognize your various disguises.
So to summarize, there are three main reasons Bitcoin becomes worse over time. The first main reason is because it was designed to be relatively easy to use for subversive activities. Those tasked with eliminating such subversive activities respond over time to Bitcoin’s innovations, by erecting new barriers. The second main reason Bitcoin becomes worse over time, is because its developers made an active choice to keep the supply of a scarce good (transaction capacity) inflexible, in response to growing demand. The third main reason, is because Bitcoin’s security depends on a zero-sum game that supplies its participants with a gradually shrinking reward divided over a growing number of participants forced to invest growing amounts of resources into claiming the reward. These participants are left with a growing incentive to subvert the system.