Are blockchains the 3D printing of trust services?
A SPECTRE is haunting the emerging blockchain community: despite all the hype surrounding this neat invention, is there really demand for the solutions blockchain technology offers? Is it the foundational technology for the next Internet revolution, as some argue — or is it an answer to a problem no one really has, as Zack Korman argues in his provocative essay, “On Blockchain, intermediaries, and hype”? We here at the ReCon research team believe that the history will show that blockchain technology turns out to be neither: instead, it might turn out to be the prototyping tool, or scaffolding, that helps us to imagine and test out novel trust services at low cost but is not the endgame in itself.
To explain our thinking, a critical look at the blockchain technology is in order. Blockchain, Mr. Korman points out, was originally developed to solve a very specific problem: how to make micropayments viable by eliminating the need for costly mediation infrastructure provided by the trusted intermediary, and therefore reducing transaction costs. In this setting, taking out the intermediary made sense (even though it didn’t really work out, as bitcoin’s increasing transaction fees and long delays hinder micropayments in bitcoin). However, as Mr. Korman argues, despite widespread belief among blockchain enthusiasts, the existence of a technology that can in certain cases even eliminate the absolute requirement for a trusted intermediary does not mean that all trusted intermediaries are now useless or outdated. Trusted intermediaries are still very useful and can do almost everything blockchain can do, and a lot more besides. Blockchains and particularly public blockchains shine in situations where there is a reason not to trust the intermediary, and if the drawbacks of not having the trusted intermediary outweigh the benefits of transferring its functions to blockchain.
BUT these situations are likely to be less common than many believe. As Mr. Korman put it, the vast majority of people are unlikely to ever think that a problem with a product or service is the existence of a trusted third party. It’s also a safe bet that most wouldn’t be willing to accept the drawbacks of blockchain (such as irreversibility of any mistakes) just for the fun of getting rid of the intermediary. Likewise, it’s not clear at all that the blockchain technology by itself is some kind of a miracle accelerant that transforms existing industries. As noted, almost everything that is now proposed “to be possible thanks to blockchain” is possible to do with existing technologies, and simply deploying blockchains is unlikely make a novel business model viable.
(As an example, since I’ve followed the energy sector fairly closely for a decade now, I’m particularly dubious about the breathless claims that blockchain will transform the energy sector and finally usher in the revolution of renewable energy. The main problems of solar power, for example, are not due to poor accounting or outdated information systems: they’re in the facts that 1) electricity is probably the most perishable good there is, 2) because of 1) the economic value of electricity is close to zero whenever there is more generation than there is demand, 3) such periods will occur with increasing frequency as the share of variable generation increases, particularly if the electricity grid simultaneously fractures to smaller microgrids, and 4) all this will, at some point probably not in a distant future, crush the profitability and therefore the economic case for new investments in solar panels that would only produce more unnecessary electricity at the same time as all the other panels do. No amount of blockchain-enabled microgrids or micropayment systems can change that, unless they can make the sun shine and the wind blow on command.)
We here at the ReCon research team believe that criticisms like Mr. Korman’s should be taken seriously. Over-hyping the promise of the new is the perennial problem in the technological society, as the famous Gartner hype cycle model nicely illustrates. (See also e.g. David Edgerton’s book The Shock of the Old for a more thorough discussion.) Trusted intermediaries emerged for a reason, and if there are no reasons not to trust the third party, why she should be eliminated? In most cases, efficiency of transactions can be increased if desired even when the intermediaries are present. As the banking industry is demonstrating with their Corda platform, novel technology combined with trusted institutions can be even more efficient than novel technology alone.
However, all these valid criticisms do not necessarily mean that blockchains cannot cause a revolution in trust services and trusted institutions. The revolution could simply happen in a slightly different manner than currently believed. Blockchains are unlikely to replace our trust in institutions with trust in code; instead, blockchains may very well promote the emergence of new trusted institutions.
In short, blockchains might turn out to roughly analogous to what 3D printing has been for product design and manufacturing: even though 3D printers have been widely touted as a revolution (any day now!) in manufacturing that enables everyone to produce whatever they want simply by downloading the proper designs, its greatest impact has been in the development and testing of new product ideas. By using a 3D printers and other rapid prototyping and manufacturing tools, any designer can these days move extremely quickly and at low cost from initial design idea to a more or less functional prototype. Not that long time ago, similar capability required a fully manned machine shop, or at least considerable crafting and tooling skills on part of the designer. Rapid manufacturing is perfectly feasible for small-scale production as well, and there is now what amounts to a cottage industry offering unique designs whose parts are largely manufactured with 3D printers and cheap laser cutters.
Advances in our capabilities also tend to result to us using those capabilities more. The increasing ease of rapid, computer-controlled manufacturing has resulted not so much to dispersion of manufacturing, but more to higher quality products. For example, I’m typing this with Macbook Air, a laptop with milled aluminium frame instead of older, clunky and less sturdy plastic injection molded parts. Just two decades ago, the thought of using very expensive computer controlled milling stations to manufacture individually major components for consumer-quality electronics would have been preposterous.
BLOCKCHAINS are likely to help us to prototype and test new trust services, and improve trust in existing services. After all, entry costs have been a major problem for innovating in the trust industry. Before blockchains, anyone who wanted to establish a novel trust service — say, a service for tracking the supply chain of a product — had to find an intermediary that 1) was reliable, 2) could do the job and 3) was acceptable to the parties involved. Existing trusted intermediaries (e.g. banks) are rarely willing to shoulder additional responsibilities that are far removed from their core competencies (at the very least, not for free) and building a trusted intermediary from scratch and establishing it as a trusted service takes time and vast amounts of money. But who would invest in such services unless there was an overwhelming need?
Establishing new trust services in this manner was possible, but difficult, and it seems reasonable to believe that trust services are therefore less common than what would be the social optimum. This may be one reason why blockchains generate such interest: we seem to instinctively know that we could use more trust services, even if we aren’t quite sure what they turn out to be. Now it’s possible, with relatively small investment of time and money, to build at least a functional trust system without having to first find the intermediary. As software engineer and scientist Ian Wright noted in his 2015 talk “Material Foundations of Algorithmic Socialism,” many key institutions of a society, such as money and banking, voting schemes, contracts, and distributed organizations are now cheap to produce. The chicken and egg problem is solved by moving directly to the omelet.
“[Blockchain and platforms such as Ethereum mean that] … any reasonable coder can prototype a simple deposit account with democratic controls in a couple of days… the production of institutional rules suddenly got a lot cheaper and a lot easier.” — Ian Wright, Material Foundations of Algorithmic Socialism talk, 2015
Most of these trust system experiments will fail. That is the nature of innovation: the availability of a new niche will inevitably cause an explosion of diversity as the niche and its limits are explored by eager optimists. But some ideas will survive, perhaps even thrive. If the systems remain small and limited, issues that would demand a trusted intermediary, such as mistakes or fraud, are going to be limited as well. (Here is one great illustration of how a small-scale blockchain system can cope with mistakes: by trusting the other people to make things right, instead of reversing the transactions.)
But if these systems gain more users, who don’t know each other outside the system and hence face less social pressure to cooperate, and as the system or the assets within it become more valuable, mistakes and abuses will become larger and larger issues. The need for a trusted intermediary that can oversee the network and make things right increases; or it may be (as is the case with bitcoin) that the technical limits of the original distributed system become obstacles to further development. However, at the same time, larger network means that some users become known to be more trusted than others — perhaps by their conduct, or just because they are more visible, or maybe because have more to lose if the network fails. The trusted intermediaries can then emerge organically, from among the most trusted (or the most involved) users. The original blockchain may continue as the underlying technology, or it may be forked into something that is better suited for the new reality, or it may even be abandoned altogether, like scaffolding that is removed once the building stands on its own.
Such development path is more likely for services that are not entirely confined to the blockchain or digital realm. It is reasonably simple to avoid trusted intermediaries when dealing with code and digital currency: but how to avoid the need for trusted someone when, say, buying goods online? Without some kind of a trusted authority, a buyer will be always screwed if the seller mails a package of stones but tells the shipping system he’s shipping a laptop — no matter how well the system preserves the claim about the laptop. Reputation systems do help in these cases, but they too can (and will) be gamed. Furthermore, a reputation system is in fact highly likely to promote the emergence of a trusted authority: the entities with higher reputation scores will almost inevitably become de facto authorities, while reputation systems by themselves form a costly barrier to entry that deters competition from emerging. Finally, as shown by the frequent scams and suspicions of scams in the marketplaces that are truly without trusted third parties — the illicit dark web markets — a world without trusted intermediaries might not be a world we’d choose to live in, if we have the choice.
We believe that despite all the hype about everything being different this time, the future will not be dominated by disintermediated trust services. Trusted intermediaries will be around; they may be different ones to what we have now, there may be more of them, and they are likely to perform services we can’t even think of right now. It’s also likely that they will make use of blockchain and blockchain-inspired technologies like Corda, and that as a result, transactions will be easier and cheaper than today. But we cannot solve all our problems with technology, and sometimes, after experimenting with new tools and using them to develop novel ideas, we may find out that old tools have their uses as well.