Discover more from Foresight Institute
5. IMPROVE COOPERATION | New Information, Money, Rights, Contracts, and Privacy
Previous chapter: SKIM THE MANUAL | Intelligent Voluntary Cooperation & Paretotropism
Enter cryptocommerce; in cryptographic ledgers, blockchains, and smart contracts, we see the rise of new law-like systems. They are technologically embodied, cryptographically shielded, and let individuals engage in secure voluntary cooperation. They are the keys to unlocking the next levels of cooperation.
Legal systems emerged to establish neutral rule systems that support cooperation without violence. But they struggle to insulate us from actors within. Modern democratic governments have given us a taste of rule of law’s benefits. But they are also becoming increasingly centralized, with major superpowers breaking down the rule of law. Governments and their rules are strongly based on jurisdictions. But many of us no longer know or care where the people we cooperate with are. As meaningful interactions move onto the global network, jurisdictions become less relevant.
We are used to being citizens of a particular country and that changing citizenship is difficult and costly. We are also increasingly citizens of the internet ecosystems we subscribe to. The internet produces a permanent system of voluntary rules that is very costly to suppress. Free speech is a great example. Governments imprison people for violating various national speech codes. Edward Snowden is in political asylum because he is rightly worried the US won’t grant him a fair trial for releasing information to the public. But all the servers storing the released information, and all the routers transmitting that information, create an emergent phenomenon by which the information stays public.
It would require global totalitarianism to reliably and permanently remove information from public availability. In the absence of that, any single entity that tries to erase widely dispersed information renders itself irrelevant, or worse, achieves the opposite of their desire; the Barbara Streisand Effect.1 No matter how many people go to prison, and no matter how many jurisdictions try to suppress it, information remains public.
But becoming a citizen of the internet often means becoming a citizen of new jurisdictions, such as the Google services suite, with email on Gmail, videos on YouTube, etc. The rise of internet giants with arbitrary rules and pathologies is recreating centralizing dynamics. By having a Gmail account, Google and you engage in a voluntary interaction with each other. They offered you Gmail, and you signed up. Nevertheless, if you violate vague terms of service and they arbitrarily banish you, everything you managed for that identity becomes inaccessible. When restricting or cutting off access, internet giants aren't subject to accountability or due process. Within this digital jurisdiction, when banned you become a nonperson, exiled due to unrevealed criteria. This is a very poor form of cooperation.
If we think of these giants as an arrangement through which we cooperate, we don’t want those interactions to be placed at risk by a possibly-biased third party. Whenever we are at a giant’s mercy, not for interacting with the giant but for interacting through it, we want to remove its whims from the picture. The problem with the giants is not just power centralization, but that their power feels arbitrary. If we don’t know why our accounts could be suspended, we live in a state of vague fear, reminiscent of a lawless state of nature. We need rule of law and an understanding of what the rules are so we understand how to cooperate. In an extreme view, our internet giants act like historically oppressive regimes, dominated by their rulers’ unpredictable whims rather than by laws or rights of citizenship.
The next innovative leap of decentralized platforms and media with much better emergent properties is imminent. A rising tide of decentralized social networks and independent content producers usher in a complex ecosystem of polycentric communities. The more interoperable decentralized knowledge communities are, the more we can lower exit costs and network effects that lock us into sub-optimal centralized platforms.
As we lay the building blocks of tomorrow’s communication architecture, it’s a good time to ask which features we’d like to experiment with:
First, you can only be socially canceled if your name is attached to everything you say on social media. Future social media networks may be pseudonymous. Srinivasan points out that switching to pseudonymous systems makes it hard to acquire a reputation. If you already have a reputation under one name on one platform, switching to another means sacrificing your reputation. He suggests that using zero-knowledge proofs, one could transfer ZKarma, i.e. reputation points acquired in one domain, from one username to another.
A zero-knowledge proof is a method by which Alice can prove to Bob she knows a specific piece of information without revealing any underlying information that makes it true. Proving knowledge without revealing it and giving precise answers to precise questions opens up new cooperation opportunities. In principle, anything provable could be proven via zero-knowledge and recent innovations, such as zk-SNARKs, radically improve applicability. So, in principle, zero-knowledge proofs could let you transfer your karma acquired on a prediction platform to a different platform where you’d otherwise start from scratch. Pseudonym reputation-fluidity means lower costs to migrate across platforms. This ultimately pressures platforms to continue improving and enables higher knowledge transfer across different digital jurisdictions.
Healthy media will allow good ideas that survive debate to rise to the surface. For instance, hyperlinks - fine-grained, bi-directional links between a document and its criticism - would enable readers to easily access criticism from an original text. Parts of other texts could be embedded in one’s own, allowing a rich mashup of texts. A charge-per-read micropayment royalty encourages publication of in demand material.2 Subscription models could be used to regularly auto-tip one’s trusted favorite content creators. Many of these features, proposed 30+ years ago by early media prototypes, such as Project Xanadu, are now finally coming online. Such features, proposed 30+ years ago by early media prototypes, such as Project Xanadu, are now finally coming online.
They take inspiration from Karl Popper’s theory of scientific discovery, who observed that knowledge, much as biology, evolves by a process of variation, replication, and selection. Variation of knowledge as in tossing new ideas out there, replication of knowledge as in spreading ideas through conversation, and selection of knowledge as the discrediting of ideas through criticism.3
We will never reach a perfect sense-making system. There will always be imperfections that an omniscient outsider could improve on. But the key feature of this system is that there is nobody in charge. The overall opinion direction results from the voluntary interaction of many individuals. History showed that an alternative system of knowledge could be commanded to operate in a particular way risks extremely destructive commands.
Curious for more? Listen to this seminar on re-decentralizing social medias.
Another area for media innovation is prediction markets. Take the early market set up at Foresight’s 1999 annual gathering. People would contribute their expertise by opening a claim about the future for betting on a platform. For instance, “World product doubles 2025-30” tracks the likelihood of a big increase in world economic productivity by 2030. Prices represent a bet that a particular event will occur. If others believe the future price will be higher than the market indicates, they buy, and in so doing raise the consensus price. For instance, the odds for the World GDP doubling were at 30-38% in 1999, moving to a more conservative 25-50% by 2000. This early prototype had people send checks to Foresight's offices.
Twenty years later, a rich ecosystem of foresighted projects has sprung up. According to Robin Hanson, prediction markets can help swap our reliance on poorly incentivized experts for a market in ideas that encourages individuals to contribute their specialized knowledge: “if markets create a consensus about the value of an ownable item, such as the price, futures markets create an immediate consensus about future consensus.”
Plenty of variations are possible.4 Prediction markets could estimate the reproducibility of scientific research, turning into replication markets. Since some studies are more costly to replicate than others, the results could decide which studies to prioritize for replication, i.e. those with the lowest replication likelihood. The costs associated with irreproducible preclinical research alone have been estimated at US$28 billion a year in the United States.5
We can’t wave a magic wand and create a world free from interest groups trying to corrupt the system in their favor. Instead, we can create a robust system against those forces. While we can’t assume that all problems are solvable, we should not underestimate the cleverness of games with novel payoffs for solving our problems. After all, before causing today’s pathologies, at the time they were created, companies such as Google and Facebook were themselves such innovative leaps.
Listen to this group discussion on prediction markets on the problems they may solve.
Make Money Immutable
Thanks to the internet, we can send almost anyone information freely and cheaply. While sharing information is internet-native, sharing monetary value started in the constraints of the traditional financial system.6 The legacy monetary system comes with a plethora of pathologies. Robert Mugabe can print endless cash, inflating away the savings of Zimbabwe’s citizens. Vladimir Putin can freeze an NGO’s bank account, and refugees can get locked out of the banking system.7 Situations such as Operation Choke Point, when the US government used its control over banks to shut down unquestionably legal businesses, or when credit card companies blocked payments to WikiLeaks, show democracies aren’t insulated from these problems.8
Fortunately, Bitcoin happened. Bitcoin injects internet properties such as programmability, interoperability, and composability into monetary value exchange, while guaranteeing its own scarcity. The results are astounding, as Balaji Srinivasan summarizes: “first ignored, then mocked as an obvious failure, within five years after its invention Bitcoin was listed on CNBC and Bloomberg alongside blue chip stocks, and by 2020 it had changed the trajectory of the People's Bank of China, the IMF, Goldman Sachs, JP Morgan, and the World Bank.”9
Bitcoin is a great example of a new layer of rules built on top of the internet’s permanence. We can think of its rules as a constitutional system for the digital jurisdiction, so that the web’s information economy can now operate within it. It directly threatens what offline jurisdictions consider their prerogative; minting money. Nevertheless, it succeeded at providing the world with a currency that is very costly to corrupt, even for governmental jurisdictions.
Because crypto offers a cross-jurisdictional, censorship-resistant exchange medium, it is not only attractive for people lacking access to institutions but also for those who distrust them.10 It allows anyone to store, send, and receive money without asking permission and without proving one’s identity. In Nassim Taleb’s words, it offers an “insurance policy against an Orwellian Future”.11 Over half the world’s population live under an authoritarian regime, and could stand to benefit from this.
Pre-crypto, perhaps our best hope at creating trustworthy money was via competition and reputation feedback.12 Still, relying on reputation is insufficient; reliable operation of money matters most in emergencies when stressed people make compromises regardless of their reputation.13 Bitcoin, and a growing number of crypto alternatives, are doing something better. By examining its internal workings, we know that the money itself operates in an incorruptible manner, regardless of external pressures.
Bitcoin’s underlying technology, the blockchain, has two characteristics making it ideal for creating an incorruptible base cooperation layer. First, it operates according to its stated specifications. Bitcoin is a virtual computer built out of agreement rather than physical hardware. A tremendous number of separate machines replicate the same computation and cross-check each other. While physical hardware may have security trap doors, if a quorum of many different computers that don’t trust each other agree on a transaction, together they form a credible machine to run our computation on.
Blockchain’s second innovation is that it provides an agreed message order. A single piece of arbiter hardware can implement censorship by deciding to never see a message it doesn’t want to accept. In contrast, at least for proof of work, messages coming into the blockchain are coming into the memory pool, the transactions waiting room, and are visible by the blockchain’s miners. The miners compete to gather the messages into a block and publish them.14 This avoids the problem of double spending, which is fundamental to currency creation.
Just as mobile phones improved from being expensive, barely functional, and only available to the elite, Bitcoin is evolving and will become more accessible. While Bitcoin is a rather specialized computing device, the agreed message order also enables general purpose computation which we’ll look at soon. It is the combination of the blockchain’s ability to act according to the stated specifications and to reliably service requests, that makes it a candidate for creating incorruptible institutions. Simple money is the most obvious high-leverage institution needing public non-corruptibility and our initial success is encouraging here.
Bitcoin gives us monetary sovereignty. But Bitcoin, and almost all existing cryptocurrencies, put all transactions on an open public ledger, readable by anyone anywhere. Ethereum, and most blockchains that support smart contracts today, do all their contract execution on an equally public ledger. Although participation is said to be anonymous, in practice it is easy to do traffic analysis, a form of statistical reasoning, to correlate blockchain activity with players in the physical world. Anyone can do this traffic analysis, and the tools to do so easily will quickly become a commodity. Because blockchains are immutable, malicious actors have lots of time to learn more about our activities.
Blockchains without strong privacy, which constitutes most blockchains today, create dangerous new opportunities for real world crime. If real-world commerce moves onto such blockchains, we’ll be in a world where private agreements are impossible. If a criminal can determine who, i.e. which physical body anywhere in the real world, is associated with particular crypto assets or agreements, they can subject them to violence, or threats of violence, to acquire the assets or manipulate the agreements. So-called rubber hose attacks and kidnapping come immediately to mind. Besides individual criminals, we should also worry about corrupt governments engaged in human rights abuses, or blackmailing people to act as double agents. Moving real-world commerce onto such blockchains would be a massive and dangerous loss of privacy.
Leveraging zero-knowledge proofs, transactions can be fully encrypted, yet still be verified via consensus. Blockchains with privacy deny anyone any new information they need to figure out who to attack. Currently, this level of privacy is effectively available for cryptocurrency. It is not yet practical to provide it for general smart contracts, but the technology is advancing rapidly. Moving real-world commerce onto privacy-preserving public blockchains would increase our privacy and safety against criminal activity. Privacy-conscious currencies would especially empower those fearing prosecution. For instance, by preventing third-party analysis on which donations keep projects like Wikileaks running after the main payment providers stop transactions. This book will proceed under the assumption that privacy-preserving smart contracts on public chains can be a practical reality.
Check out this seminar on Zero-knowledge enabled Cooperation.
Create New Rights to (Do) Things
We have been able to build less corruptible institutions for quite some time now, long before blockchain. Digital notaries from the 1990s already allowed the creation of irrefutable cumulative public records. Yet, it is blockchains which have refocused our attention on incorruptibility. We want to depend on institutions to do the right thing, not due to offered incentives, but because that is what their program requires. Knowing it’s their internal logic that makes them function reliably lets the rest of us incentive-align around them.
When introducing property rights - one of the most basic institutions - we suggested that rights to things are really heuristics, letting us transfer rights to do things. Sometimes efficient rights to do things don’t correspond cleanly to tangible objects or geometric slices of the world. We can expect a future rich with property rights experimentation:
Cryptographic Property Transfer
In too many places around the world, it is hard to guarantee credible property rights because it is hard to figure out who owns the land one lives on. Hernando de Soto found that a village’s organically grown informal system of property rights is often not based on naïve spatial geometry.15 Rather, it is a complex set of understandings such as rights of action; “If you need to get from here to there, then you can go through this property, on this path I've set up.”
Many formalization attempts fail. Title companies come in, ask the local population for the physical boundaries, write those down into a property registry, and then naively interpret the property rights per those physical boundaries. Those boundaries don’t respect the complex systems of law, rights, and obligations that locals negotiated village by village over decades. They ignore the formal title registry as useless formality and the actual transfer of informal property rights immediately diverges from the property rights registry.
The absence of credible title transfer comes with a cost: the real estate holdings of many locals are ineffective as loan collateral for wealth-generating investments. De Soto estimated that more than $9 trillion of the world’s poorest people’s assets, mostly real estate, is locked up due to inadequate recognition by their governments.
Better arrangements are possible. A decentralized cryptographic system for property right transfers, agreed upon by locals, may provide the trust needed for transferability so that more abstract rights, such as mortgages, can be created from a credible local rights ledger. Crucially, the collateral is not the physical property itself, but the rights to use it in agreed-on ways.16
We’re learning that the physical world is not the only possible basis for establishing property rights. Upon discovering the radio spectrum, it would have been untenable to treat sending radio waves through another’s property as trespassing.17 As we discover more about the world, some discoveries will make our cooperative ontology obsolete. For instance, today’s property rights leave much room for externalities, such as air, water, and soil pollution.
Garrett Hardin uses lake pollution to explain the tragedy of the commons that can arise from this: “the rational man finds that his share of the cost of the wastes he discharges into the commons is less than the cost of purifying his waters before releasing them.”18 While we prefer a world with healthy air, water, and soil, it is not in our immediate interest to pay the costs to fix things if no one else does; we’re stuck in the sub-optimal Nash Equilibria described before.
As we become aware of the non-local interactions of pollution, we need to expand our property paradigm to track more than spatial geometry. Tradable carbon credits are a new form of property that treat greenhouse gasses as a commodity. Polluters wanting to increase their emissions can buy permits from others. Sustainability projects can ‘mint’ carbon credits and sell them. A blockchain-based system for tracking carbon credits may alleviate current problems such as double-spending. Just like Bitcoin has credibility that there will never be more than 21 million Bitcoins, the blockchain could guarantee that a tree has not been sold 100 times over to offset emissions.19
A new type of property rights enabled by the blockchain is time-restricted ownership of digital goods. Similar to how you rent an apartment or car for a set period of time today, you could hold the rights to digital goods for a specified while. With the rise of the internet of things, blockchain-based property can become smart property. First discussed by Nick Szabo, smart property is equipped with electronic components that transfer ownership by transferring the cryptographic keys. Instead of trusting one actor with access to our physical devices, ownership is decentralized. This can change a dispute’s default outcome; default on your car loan and the car can decide you no longer possess it.20
Space Property Rights
Our initial notion of property emerged from the game theory of trying to coordinate with each other. Because the Earth’s crust moves slowly, we could force our notion of property onto this very naïve geometric notion. It worked until we advanced to need things like pollution credits, radio spectrum, the rights of way so planes could fly over people's property, cross-negotiating those rights with the noise from planes, etc. There is no simple objective source of property boundaries in the physical world because physics itself has so many nonlocal interactions. If I move my arm so I am sending gravitons through your body, am I trespassing?
With respect to outer space, plots of land don't necessarily generalize to bodies moving through space in ballistic trajectories. In our solar system, bodies orbit around each other and occlude each other with respect to sunlight and its energy. What if someone builds a Dyson Sphere (a sphere completely surrounding the Sun) between me and the Sun? Even if it didn't intersect my orbit, does it violate my property since I cannot receive energy from the Sun? Determining rights in terms of orbits or radiation is clumsy and difficult.
Historical denotations of plots of land is not a bad starting point for outer space property rights. But by starting with something physical as an initial endowment, we face substantial transaction costs when rearranging our endowments into something adapted to the activities we want to engage in.
Imagine that if you want to use a piece of radio spectrum as it passes through many plots of land you would have to acquire it from every plot owner over that area. With property rights boundaries as vertical slices and the efficient property rights values as horizontal slices cutting through the vertical slices, this is a worst case misallocation of resources that maximizes transaction costs. Trying to get from the vertical regime to the horizontal regime by individual trades becomes extremely hard to do. For outer space property rights, we need to avoid outrageously high transaction costs.
If we discover a workable arrangement, it also needs to have enough initial legitimacy that it isn’t immediately contested. Establishing legitimate resource claims requires a secure title registry, for instance via a permissionless blockchain.20 As long as there is no competing governmental title registry for these resources, claims on this blockchain could be created without an opposing claim. There is little interest in abolishing the title registry and much interest in upholding it. Suddenly, there are property rights where previously there were none. Everyone is now either better or similarly off as in the absence of the registry.
Watch the seminar on Legal Considerations for Space Property Rights.
Legitimize New Property Rights: Gradual Emergence
Could such a system ever work in practice? Let’s look at Bitcoin’s history to understand some of the challenges ahead. Before Bitcoin, the value of every fiat currency was bootstrapped by being backed by a physical asset or by conversion to some prior currency with a history of physical backing. The U.S. Dollar was redeemable for gold, and the Hong Kong Dollar was backed by the U.S. Dollar. Even other items treated as currency, such as cigarettes or stamps in prison, have a use value that their symbolic value grows out of. Once these currencies are bootstrapped, their value is rooted in the expectation that they will continue to have trading value. This is a deep common knowledge-like expectation of expectation of expectation of expectation of future value. Historically, we rarely assigned value just ab initio.
Bitcoin was the first innovation to violate this expectation due to early participants being drawn to it by ideology rather than use value or redeemability. The desire was to seize something with abstract qualities of non-manipulability and voluntary trade that was un-disruptable by corruptible institutions. Bitcoin was an ideological attractor among early adopters, who caused enough growth to attract others based on speculation value. Some of this activity was bubble-like, but it had a history to look back on as the basis of further trade.
Space property, like Bitcoin, starts with most people discounting it as crazy blockchain and space nerds who have nothing to do with our economy. Most people will ignore it, but a few won’t. They may be attracted for ideological reasons or because they believe space might significantly matter to the current economy much faster than is expected. These people will attract other people who see there are people to sell to.
Space property can start with a very small audience but a growing frontier. The growing audience’s size is defined by yesterday’s small audience plus yesterday’s frontier. If it grows slowly enough, the existing audience size anchors the expectation of the new frontier for Bitcoin-like growth. Growth may well be steeper because space property is rooted in actual anticipated use value of a real, though currently hard-to-reach, physical resource.
Bitcoin started with no valuation and no history but now has a history of valuation. This would lead one to expect it to continue to dominate as the natural attractor. This position is called the Bitcoin maximalist position. But the emergence of Ethereum and others developing competing smart contract projects shows that there is no strong Schelling Point of a single blockchain or cryptocurrency.
Such competition is great for blockchains and cryptocurrencies, but not necessarily for space title registries. Competing title registries for overlapping plots of land with competing claims to legitimacy run into problems covered by title insurances. Buyers face the risk that the title that they pay for is not the title that people consider legitimate. We don’t need a universal registry covering all space resource titles, but for any one part to which there is a title, it should exist in only one title registry system.
The details of who owns what may create opposition to any particular title registry. Rather than setting up one instantaneous title registry, we could play with systems in which we jointly change several listings in response to a legitimate title change. The honor system of titles used prior to official title registries is reminiscent of natural language’s spontaneous order coherence. We saw how language continually negotiates word meanings into enough shared vocabulary so that we understand each other without having to mutually agree on one unitary decision point. Similar to how our shared word meanings emerged from individual word choices, the common law system of rights emerged by judicial decisions framed by prior judicial precedents.
We all benefit from meaning largely the same things by the same words even if we sometimes disagree on the precise meaning in individual cases. Similarly, for physical property such as space resources, we all have a payoff for having a similar mapping of property to people so we can coordinate on creating value from it. An initially agreed division can be perpetually renegotiated between a great diversity of players who pursue a diversity of goals. The initial title registry may be upheld as a Schelling Point that benefits us by establishing a system in which property rights are recognized.21
Multiple title registries with overlapping claims may also have their benefits. We make agreements in one state of the world thinking that the crucial division of the world will remain this way. But by the time our contracts unfold into the world, new discoveries or knowledge have disrupted the way we conceive of the world. Whether free speech applies to the internet, or how to account for radio spectrum with slices of property are all cases in point.
A too literal interpretation of rights creates transaction cost problems. If the correspondence of title to “rights to do things” resembles the fluidity of natural language, it can mirror how words evolve to reorient around new discoveries, such as the disambiguation of “heat” and “temperature” discussed earlier. Imagine an emergent system where titles are across different title registries but transfers are cross checked with each other so that they are in agreement. For the same reason that the meaning of vocabulary is coherent and fluid, new types of property, such as radiation, that cross cut existing divisions may be accommodated for in an emergent consensus. A flexible coherence of title registries, similar to natural language, may be more adaptive than one system.
The reality may be neither Bitcoin-like or natural-language-like. Degrees of polycentrism can vary from de jure single chain title registries to the more common law solutions that aim at sufficient emerging coherence of understood title legitimacy over time.
Watch this seminar on NFTs for engineering property rights.
Compose Complex Contracts
Complex play can emerge from simple rules. Through commerce we gradually figured out how to use contracts to bind ourselves to complex games. Right of contract states that voluntary arrangements between consenting adults should be allowed because it lets them find novel ways to cooperate.
To cooperate between mutually suspicious parties, we mostly rely on legal contracts. Contract enforcement is expensive and most disputed commercial relations are well below the threshold for going to court or arbitration. This limitation creates barriers. Unless we are rich enough to take everything to court, complex forms of cooperation are restricted to those we trust and have long-standing iterated relationships with.
Being members of the same jurisdiction somewhat reduces risk since operating under the same law lets us assume we will both behave well. The internet enabled the novel opportunity to create a website and cooperate with millions of total strangers. Still, on the internet, you may engage with someone who is completely anonymous and in a jurisdiction that does not inhibit interaction you consider misbehavior. Some cyber attacks are committed in jurisdictions where the attackers are not subject to punishment.
The result is that cooperation among strangers at a distance is rather simple, such as giving away information for free. Even taking payment in exchange for something, a rather simple deal in itself, is often enabled by third parties such as credit card companies taking on the transaction risk. Our ability to cooperate richly with the majority of the world whom we don’t know remains underdeveloped.
Smart contracts are here to change this. They were conceived by Nick Szabo in 1996 to automate contractual clauses, such that the contract execution is “embedded in the hardware and software we deal with”. They are contract-like arrangements expressed in program code, where the program’s behavior enforces the contract’s terms. Smart contracts let us create automated arrangements to which we bind ourselves by placing rights into escrow with the contract. These rights are only released back to the contract participants according to the agreed contract terms.
Miller and Stiegler explain Szabo’s infamous smart contract analogy, the vending machine; “by escrowing both drinks and payment before dispensing either, it also dispenses with the need for separate enforcement. Instead of enforcement, the contract creates an inescapable arrangement. It cannot prevent the customer from walking away before the game is over, but a customer who walks away from a contract in progress leaves behind any assets escrowed by the contract at that point.”22
Ethereum, a general purpose blockchain, was the first system to establish a sound smart contract architecture. With Ethereum as a precedent, we now have a user-extensible rules system in which every contract is a new set of rules that cannot be shut down. Lawrence Lessig suggests we can achieve a particular end through different means, with legal contracts being one and computer code another.23 Having only been able to build cooperative arrangements whose execution relies on human beings, we struggled to insulate them from human corruptibility. Smart contracts significantly increase the range of possibilities for what computer code can achieve. This makes cooperation reliable, credible, and trustworthy to an extent never achievable when institutions had to rely on human beings to function.
Smart contracts give us new tools to enforce rights and responsibilities in code at very low cost. This lets people across the world exchange, trade, and cooperate in new ways that have been too costly. Lowering cooperation costs by orders of magnitude is not just a quantitative difference. Whenever we have several orders of magnitude difference in a quantity, it often causes a change in the phenomenon’s character best seen as a qualitative difference. The currently emerging cryptocommerce may usher in a new era of cooperation, analogous to when the 1500s and 1600s commercial revolution in the Mediterranean world resulted in an inflection point toward today’s much richer world of exchange.
Rather than being the last invention of this new level of the game, we should think of smart contracts as a key to unlock the next levels. It was hard to imagine today’s internet before the first website. Likewise, it is hard to imagine the novel cooperative arrangements that can form in a mature cryptocommerce. But if, in Alan Kay’s words, “the best way to predict the future is to invent it,” it would be useful to be able to imagine it first.24 Let’s sketch out a few ideas for improving on today’s paper contract paradigm in increasing levels of complexity. We admit up front we may be as far from a mature cryptocommerce as the ARPANET was from the current web, but this shouldn’t stop us from dreaming big.
Check out this seminar on Blockchain Governance.
Reuse Successful Experiments: Templates
In the past, making attractive flyers, posters, or anything printed on paper required going to a print shop. They had the expertise and tools to create a custom design using metal and lead blocks. With the advent of laser and inkjet printers and the Macintosh, suddenly regular people were able to easily create and print attractive designs at home. What followed was a lot of horrible-looking newsletters with a tremendous overuse of fonts. People had yet to develop the skills to best use their new tools, so demand built to improve them. At first people created tools with defaults for guiding others in the right design directions. Early desktop publishing programs evolved with templates guiding users away from using Comic Sans for everything.
Similarly, our cooperative arrangements are currently crafted by highly specialized, highly paid lawyers. They are supposed to be good at writing contracts that don't fail in unexpected costly ways. Most of a legal contract is actually about how to deal with a variety of failure cases and pathological contingencies that a layperson never would have thought of. Early smart contracts will have effects no party would have wanted if they anticipated them. It takes time to develop tools and templates that embody expertise about unintended consequences and contingencies. Just as early, awkward design efforts evolved into today’s slick templates letting anyone create sophisticated publications, smart contracts will evolve from clumsy, costly proofs of concepts to ever more sophisticated plug-and-play templates.
Keep a Human in the Loop: Split Contracts
Not everything can or should be automated. We may never be able to flawlessly represent our intentions in legalese or code without risking an outcome not reflecting what we actually wanted. Dumb paper contracts lock in states without knowing what future participants will consider relevant. Lawyers try to freeze the next 10 years in hard-to-parse prose, only to litigate when the contract fails. When writing a contract, we are writing interaction rules for an unanticipated world. We face an alignment problem among current and future parties. An open contract is a plausible alignment strategy.
Software licenses on a zero to four year rolling window introduce different degrees of copyright protection. Mobile phone plans with monthly subscriptions instead of year-long contracts make for a more competitive telecom market. With sunset clauses and shorter time windows, contracts increasingly evolve from shackles that lock in the present toward iterated games. Rather than litigating and blaming when predetermined contracts fail, Open Contracts let parties embrace change and improve the game’s next iteration.
Some projects require many contract iterations. As a manufacturer of laptops, your company will depend on lots of parts produced by other companies, such as computer chips, displays, keyboards, etc. This requires a series of negotiations with different suppliers which depend on their negotiation with their suppliers. You wouldn’t want to commit to any of those contracts unless you can get all the needed parts, which makes the contracts conditional.26
Such conditional deals already exist, thanks to long-established trust and business relationships and stable supply chains. But there is often friction and risk. Using conditional smart contracts, start-ups could get created conditional on enough funding and talent pre-comitting,
A Chained Contract is a conditional contract that can guarantee selected terms from earlier in the contract chain while leaving some terms open until a specific condition is met. With blockchain-based supply chains, you could audit if contractors can fulfill their part of the contract, informing negotiation of any conditional contract chains. The entire chain could be automated via smart contracts. If certain conditions are met, such as availability of enough silicon to produce a chip, the contract could automatically authorize the next part of the chain.
Other contracts will want to rely on a human at crucial steps to negotiate future parts of an agreement chain. They could take into account information from previous rounds and consider new circumstances when initiating the game’s next iteration. Preserving a human in the loop can be a useful feature. Sometimes we don’t want just the automated contract to be the credible commitment. Instead we want the credible commitment to account for factors that require human judgment.
The American Information Exchange (AMIX) was an early prototype with this aspiration. Designed back in the 1980s before the web browser, it was a computer-mediated market for matching information buyers and sellers. Up to then, finding information on a topic involved fishing for relevant bits of knowledge from the sea of newspapers, TV, journals, and books. AMIX replaced this random walk with a targeted market on which you could ask questions and those with relevant expertise could sell their answers.
Here, Barbara wanted to know if it makes sense to build a co-working space in Palo Alto. She defined what would count as a satisfactory answer and uploaded the request to the exchange. AMIX is an example of a Split Contract that divides the contract into two parts: the automated part directly enforced by software, and the prose humans interpret in case of dispute. If either party decides to dispute the contract’s automated part, pre-chosen arbitrators judge the disputed terms and may overrule them.
AMIX’ particular Split Contract had two crucial design elements:
First, the contract terms were fine-grained; specifying the amount paid upon committing to the contract, the amount paid upon delivering the answer, and the amount paid when accepting the answer. If a consultant delivered an answer and thirty days go by without Barbara taking any action, the automated contract would pay him the full amount. To stop the payment, Barbara would have to explicitly state his document does not meet her terms. The contract would then take those contradictory propositions to the arbitrators. They would only get to overrule the amounts paid on accepting the document.
Meanwhile, with paper contracts, the consultant could have given Barbara a perfectly fine document. She could refuse to pay, leaving him with the expensive legal action to get what she owes him. Dumb paper contracts leave the default situation determined by the asset’s default location. By escrowing money, the contract changed the default distribution from Barbara to her consultant. From possession being 90% of the law, we move to behavior becoming 90% of the law.27 By rearranging the burden of initiating a dispute, both parties get skin in the game to comply.
The second innovative feature is the fine-grained negotiation process for the contract terms. Payment amounts, prose text, and arbitrator selection are left up to the discretion of the contractors. By co-negotiating those parameters upfront, they can make expensive dispute resolution less likely. We dispute a contract hoping we win, but it carries some expense and risk. The obvious cost is expensive human arbitrators, but a more subtle one is uncertainty. If we don't dispute, we are stuck with a disliked, but known, outcome. But once subject to the arbitrator's judgment, the result is less predictable. If we could have predicted it, we could have written a more mechanical prediction into the contract in the first place.
Humans in the loop are a useful feature for the first contracts we’ll automate. Reducing the odds of a human dispute in the first place is a great hack for reducing enforcement costs.
Preserve the Meeting of the Minds: Video Contracts
Open contracts are powerful because they are vague. This vagueness lowers the risk that the contract is overly specific in ways that turn out to be counterproductive. But open-endedness does not necessarily get resolved to better represent the parties’ original intentions. If one side abuses the ambiguity the other side may be even worse off than if the contract had been overly specific with respect to future change.
Often the most important question for cooperation is not how can we design the best plan? but who has the discretionary decision making power?28 Open Contracts don’t specify in detail everything everyone will do, but instead divide items into rights and responsibilities. They determine decision-makers instead of decisions. So who should decide and how?
Pre contracts, we had verbal agreements. Individuals talk, agree on something, and shake hands. They draw on live behavior full of conscious and subconscious cues, including tone of voice, body language, etc. to form a joint agreement on what exactly they agree to. At this moment, both parties have a much richer understanding of what they understand each other to be agreeing to than they could record in words alone. But memory drifts. If they end up in adversarial situations, each has an interest in the agreement’s essence drifting away in different directions. Instead, people turned to written agreements, which evolved into contracts. Written contracts provide a crisp permanent record, but are much less subtle than our understanding of what we agree to via a conversation.
A Video Contract, first proposed by Nick Szabo, is an audio-visual record of a conversation between contracting parties discussing their understanding of the contract. A video record can capture the unspoken parts of an agreement that a paper contract might not, such as tone of voice and body language. Contracting parties can record and store their conversation about the contract's meaning, with the contract for future reference. With this tool at hand, let’s revisit the earlier problem of creating land titles that capture the complex understanding of what it means to use it in pre-agreed ways.
Imagine locals could record a video while negotiating the rich contextual property rights distributions they agree to. They may now be much more inclined to regard the agreement as legitimate. Imagine such a video-backed electronic title listing gains traction in a country with poor default government legitimacy. At first, it could coexist with the official title registry, so property changes are registered in both. But if the systems diverge via government action locals regard as corrupt, the cryptographic system may well emerge as the new Schelling Point of legitimacy.
Legitimacy questions may become more important as humans change faster. In common law, the “meeting of the minds” defines what the parties intend to agree to. This is their understanding at the time, reinterpreted in a possibly very different world. In extreme cases, later parties might not even consider the contract’s terms meaningful. Issues around interpreting the US Constitution provide ongoing vivid examples. Does the Second Amendment give the right to have personal nukes? It will be difficult to interpret our contracts in a world with completely different technological realities. A video provides at least some context about what it was that the contracting parties consented to.
The Ancient Technology of Reputation: Mixed Contracts
Videos aren’t the only technologies at our disposal for improving our agreements. Mixed Contracts combine the best of contracts with the ancient technology of reputation. To see the value of reputation, let’s compare our consumer relationships to an iterated prisoner's dilemma in which a producer is defecting on the arrangement by selling a worthless product. To retaliate in the next round, we must notice this defection and decide to warn others away. Since this process is expensive and we frequently find ourselves in non-iterated games, reputation agencies have evolved. Reports by consumers put the producer in an iterated situation with the community as a whole while trademarking establishes valuable long-term producer reputations.
Reputation is itself a commitment technology, and, if paired with contracts, can help create entirely novel arrangements. Crucially, such a reputation system could also work if Carol maintains pseudonymity. If a negative reputation system is one where participants avoid entities with a bad reputation, a positive reputation system is one in which participants seek out those with a good reputation. Pseudonyms are problematic for negative reputation systems, because it is easy to let go of a negative reputation by switching between pseudonyms.
Positive reputation systems can still succeed as long as no one can claim another’s identity. Imagine pseudonymous Alice and Bob would like to draft up a contract that relies on legal enforcement. Let's say Alice also has a contract with Carol, so she is subject to legal enforcement that is invisible to Bob. Carol can assure Bob that she has enforcement powers, since she knows Alice’s real identity. If she is reputable enough, this may assure Bob even if the specifics are opaque to him.
If Carol wants to continue being consulted for arranging deals between the Alices and Bobs in this world, she will make sure that she can honor her part of the contract. This role is similar to the role credit card companies played in enabling today's e-commerce. The demand for reputation information may eventually provide a market for pseudonymous reputation services, analogous to credit rating agencies and consumer experiences. Reducing the time, energy, and money that it takes to find the other, so-called search costs, would be a game changer for cooperation.
Unlock A New Menu of Organizational Choices
Let’s graduate from problems in which a handful of people want to cooperate to those that rely on whole groups of players. When starting a new organization, laws constrain founders to choose from a menu of confusing options ranging from a for-profit to a classic non-profit 501(c)(3), with several organizational types in-between. The combination of court cases and legislative compromise that led to the particular combinations available in each state is so complicated that only a lawyer can untangle them. They solidified in a very different past from our modern world, dating back to before the World Wide Web.
Decentralized Autonomous Organizations
Compare this to a Decentralized Autonomous Organization, aka DAO. It’s a network-native entity with no associated central management and its own asset pool. Smart contracts between the organization stakeholders may decide how to use the assets. Through DAOs, people can participate in many new organizational structures that distribute reputation or bounties according to value contributed or make decisions via the wisdom of the crowd. It is very costly for people within the organization to prevent the relevant set of immutable smart contracts from executing as specified. Because they run on a blockchain with international nodes, these contracts are more resistant to internal and external corruption. This makes them an excellent candidate for experimenting with new cooperative structures. Thanks to smart contracting, we gain the ability to escape the fixed menu, design better arrangements for what we seek to achieve, and engage in lots of experiments. Many experiments will fail but we only need a few successes to create an entirely new menu of organizational choices. After all, current institutions are also made of contracts.
Financial Risk Reduction
One institutional sector surely ripe for repair is finance. The 2008 crisis was in part caused by insurance giant AIG, becoming a black hole for risk, losing $13B and being bailed out by the government. Risk from elsewhere in the economy kept pouring into AIG. Meanwhile it maintained a five-star credit rating completely impervious to the accumulating amount of risk. That crisis showed when firms go bankrupt, costs imposed on the rest of the economy can be immense, even if triggered by very many innocently made small loans.
We would like a robust form of accounting in which anyone who a firm tried to get a loan from could see that the firm constrained itself with respect to its overall risk position. Norm Hardy proposed that with zero-knowledge proofs, we could imagine an arrangement that also preserved the privacy of the company’s lenders. Imagine a company declares that it only accepts loans through a specific loan-making portal run on a public blockchain. With zero-knowledge proofs, a company may be able to prove a bound on the total aggregate risk without revealing anything about particular loans. If a company representative tries to accept a loan from someone outside of this arrangement, the company itself has implicitly publicly disavowed it as a legitimate loan.
Another major headache is the institutional gridlock around our healthcare system, courtesy of the FDA, whose research, development, and approval process often takes twelve or more years. Healthcare data is a crucial aspect of this process. Worldwide, 150 exabytes of healthcare data was stored in 2020. Plummeting costs in genome sequencing, new biomarkers and wearable devices will drive this number up, significantly speeding up medical innovation and improving clinical outcomes. Nevertheless, handling patient information imposes high costs and liabilities on research centers, private businesses, and especially patients.
In an alternative world, people could have a wallet with their medical data that can be used for personalized medical care, as well as to make data available for biomedical research in exchange for rewards. Based on patients’ data consent preferences, their off-chain data sets could be indexed on the blockchain via identifiers, populating a metadata catalog that describes data available in the network without revealing any identifiable information. Researchers could file requests for different data queries, and a smart contract could match the correspondence between the request and the associated consent options.
Here are just a few cryptographic tools that could be useful in building better application around sensitive health data:
Self-sovereign Identity lets users create a digital identifier via self-certified credentials or information others attest to, allowing fine-grained identity control without leaving an information trail. Via self-sovereign identity users could interact directly with physicians without correlation by third parties.
Secure multiparty computation allows joint computation over private inputs by bringing the algorithm to the data rather than revealing the data, such that every party only knows their own inputs and the results. Via secure multiparty computation users could request analytics over a variety of medical records.
Homomorphic Computing allows computing on already encrypted data. The data is encrypted before going to the computing device, computations are performed on encrypted data, and only the encrypted results are sent back and decrypted at the source. Since the computing device doesn’t have the decryption key, no personal information can be extracted. Through homomorphic computing data controllers could serve only as data providers rather than doing any computing themselves.
These tools, or a combination thereof, means researchers would be able to ask questions and only see results in the form of unidentifiable aggregate outputs. Finance and health care are merely the tip of the iceberg. Consider your favorite institutional pathology; could such a cryptography toolkit improve or sidestep it?
Solve Collective Action
If cracking group-wide cooperation is hard, cracking society-wide cooperation is the holy grail. There are many things that we all would like to see realized, but no individual has an incentive to contribute to. This is especially true if we think we can benefit without contributing. But what if, instead of worrying our small contribution won’t make a difference, you could ensure it does?
Kickstarter solved this collective action problem. We can pre-commit to fund the open source project of our choice if and only if enough others commit to do the same to reach a critical threshold. If we all prefer an alternate situation, we can move there via assured agreement. If not, no action takes place and any committed funds are returned. Such Assurance Contracts are great tools for solving large multiway group commitments.
A problem with assurance contracts is that even with minor transaction costs there is too little incentive to contribute. Imagine an open source software improvement costs $800 to fund and is worth $100 to each of 10 people. If each pays $80, they get the results and are all better off. But each person may choose not to donate because they may still benefit from the software bug fixed, even if they don’t contribute. If you think that no-one else will contribute, it’s rational for you not to contribute. But if you don’t, then why would anyone else? Not contributing can quickly turn into a self-fulfilling prophecy.
A creative solution introduced by Alex Tabarrok is the Dominant Assurance Contract.29 It’s an assurance contract with the added condition that if the funding benchmark isn’t reached, the provider pays a prize to the pledgers. Pledging now becomes a dominant strategy, or in Tabarrok’s words, “a no-lose proposition—if enough people pledge you get the public good and if not enough pledge you get the prize.”
Listen to this seminar on Dominant Assurance Contracts.
Coordinate on Hidden Preferences: Anonymous Assurance Contracts
Anthony Aguirre suggests that assurance contracts, when paired with cryptographic tools for privacy protection, can turn into powerful mechanisms for overcoming specific barriers to coordination.30 This can be handy to surface an unconventional opinion, oppressed due to fear of retaliation. In 2013 Edward Snowden initiated a stream of revelations about NSA’s telephone metadata collection program. NSA management and employees abused their technology to spy on American citizens, including their own romantic partners, a practice described internally as “LOVEINT”. The revelations resulted in the passage of the USA Freedom Act to supposedly end mass surveillance. Nevertheless, in 2018, three times as much American telephone data was being collected than before that law’s enactment. If potential whistleblowers could share the risk of speaking out, this would continue pressure on officials.
Imagine a platform that enables coordinated collective action based on hidden but shared preferences. Its core function is to collect verified signatures for a particular statement via a smart contract, only publicly revealed at a threshold number of signatories. A public interface lets users post a statement and a triggering condition, such as how many signatories satisfy some criteria. Both are recorded on a public blockchain. Other users, after verifying their identity, could sign the statement in encrypted form, which are publicly revealed only if the triggering condition is satisfied.
When paired with cryptographic tools for privacy protection, we can create such Anonymous Assurance Contracts that unlock new forms of activism by lowering the barrier for individuals to speak out. A public interface lets users post a statement and a triggering condition, such as how many signatories satisfy some criteria. Both are recorded on a public blockchain. Other users, after verifying their identity, could sign the statement in encrypted form and are publicly revealed only if the triggering condition is satisfied.
These are high-stakes situations and a failure of the privacy protection could be deathly for those using the mechanism. Experimentation is needed, starting with low-risk use cases first.
Another potential solution to help with the free-rider problem of public goods is Quadratic Funding.31 By pledging even small amounts to desired goods, participants can direct the relative funding amount of external entities which agree to match the funds. The trick is that matched amounts are calculated by using the quadratic funding formula. This is where the amount a project receives is proportional to the square of the sum of the square roots of received contributions.32 Matching sums are based on the number of supporters, not the amount donated. This creates strong incentives to give even a little, countering incentives to free-ride.
Retroactive Public Goods Funding adds to this the layer that it’s often much easier to agree on what was useful rather than what will be useful. By rewarding public goods projects that were successful, it can incent their creation.33 Such experiments are at an early stage but, if successful, they can create what Vitalik Buterin calls “a general purpose infrastructure for funding public goods in the same way that money is a general purpose infrastructure for funding private goods.”34
Various costs prevent us from cooperating creatively. It takes time, energy, and money to find each other, bargain an agreement, and enforce it. If only we could lower these costs, cooperations could be rich. Cryptographic ledgers, blockchains, and smart contracts provide new arenas to reinvent cooperation. Innovators want to experiment, and cryptocommerce lets them do so. New media, scientific incentives, and property rights, open, split, or assurance contracts, and cryptographically shielded transactions are just the tip of the iceberg. They free information, make money immutable, reinvent the rights to (do) things, compose contracts, and privatize transactions. Most of these experiments will fail but the few that succeed may usher in a new era of cooperation. It’s hard to see how to get there from here. In the next chapter, we’ll take the first steps.
Next chapter: GENETIC TAKEOVER | Cryptocommerce
Barbara Streisand initiated a lawsuit to have some published information withdrawn, which resulted in a lot of public attention being cast on the information she was trying to keep secret in the first place.
See Nick Szabo’s Micropayments and Mental Transaction Costs; Mark S. Miller, “The Open Society and its Media”, Proceedings of the First General Conference on Nanotechnology: Development, Applications, and Opportunities, 1995, https://dl.acm.org/doi/abs/10.5555/214633.214649.
Some current filter bubble pathologies could be avoided by combining hyperlinks with prediction markets; whenever hyperlinks surface disagreements across researchers, they can move to a prediction market to put their money where their mouth is.
See Anna Dreber et al Using Prediction markets to Estimate Reproducibility of Scientific Research.
According to Mirror, the publishing platform introduced before, the recent rise of NFTs show that the reverse is also true. Crypto protocols not only encode value as information but also information as value. Just like financial assets can be turned into media that can easily be shared, media can be turned into potential financial assets, such as NFTs.
See Benjamin Garfinkel’s Recent Developments in Cryptography and Possible Long-run Consequences.
Miners have opportunistically delayed a message before. We should think about better approaches to that danger, especially as mining is increasingly centralized in a few Chinese nodes. Nevertheless, blockchain still bounds the problem. A message cannot stay censored for long because of its content, given a competitive market of miners. So even though there is some corruption opportunity on the edge, such as to front-run, the messages should always get serviced reasonably promptly. Jazear Brooks points out that “in practice ordering of messages can take up to a few blocks for the chain to agree on. There is no set number of blocks for each chain but 3rd party devs tend to agree that, for Cosmos chains there's one block (instant finality), for Bitcoin you need 6 blocks, and for Ethereum you need 20-50 blocks. In some cases there are strong incentives for nodes to reorder transactions within the above number of blocks for frontrunning. Thus the entire field of "miner extractable value" which makes cryptocurrency transactions on Ethereum chains subject to tremendous value fluctuations for some traders. Even for chains like Cosmos with one block confirmation, a validator can still censor a transaction sent to its mempool and wait for the next validator to process it which can still bring about a short but meaningful time window for frontrunning. Commitment-reveal schemes or networks based on private transactions like Secret Network are needed to solve this problem but right now the majority of blockchains are exposed to it.”
See Scott Alexander’s Prospect on Próspera.
When creating a blockchain based title listing for space resources, we must avoid fights with governments as to which property listing is legitimate. While the conflict of blockchain versus existing government with regard to existing property is messy, establishing a blockchain-based system for space resources starts off seeming so unreal to people that it may grow to substantial legitimacy without competing government claims to the same resources.
As introduced earlier, property, similar to language, can be understood by Schelling points. We have title listings for property values for things like land and cars but not for much of our physical property. How do you know that you own this book? There is no external title with a record stating you own it. But there is also no particular reason for me to think you stole it. Much property we respect simply because of history and perception and visible lack of contest. Sometimes we get it wrong, but altogether it has a coherence to it. A title listing says that there exists a symbol and a specific meaning. When looking up the symbol, in this particular place, this title authority, we find a single record of what meaning corresponds to the symbol. A transfer of property becomes a transfer of title so that the record changes state from the same symbol designating one person to the symbol designating another person. We can get to a coherent mapping from symbols to meaning even at the opposite spectrum of rigidity. An example is the evolution of vocabulary in natural language. We all know what is meant by “cat”. There is a mapping from a symbol to a meaning, but no title listing and no single authority. The mapping from symbols to meaning drifts over time. We are both contributing to subtle shifts in word meanings over time as the drift of meaning adapts to our differences in using words to mean something. At the same time, we all pull on the words in ways that serve our own interests in different ways . The actual mapping of words to meanings that we all mutually recognize is this perpetually renegotiated mapping which comes from all of our need for the mapping to converge as a Schelling point.
From a talk in 1971 at the Xerox Palo Alto Research Center (Xerox PARC).
From private communication with Alan Karp.
Mark S. Miller, “Computer Security as the Future of Law”, 1997, Extro3.
Oliver Hart and Bengt Holmstron received the 2016 Nobel Prize in Economic Sciences for this insight.
See Anthony Aguirre’s “A Simple Secure Coordination Platform for Collective Action
While the funding formula sounds complicated, projects like WTFisQF let you playtest it. You can see your impact even with a small amount.