A heartfelt acknowledgment to Vlad Zamfir for his contributions in shaping numerous concepts related to prediction markets for content curation.
For the previous six years, individuals have been on a quest to identify those hard-to-find implementations of blockchain technology that could finally emerge into the mainstream. In the realm of cryptocurrency, the uses are predominantly already recognized – although it certainly remains to be observed how effectively it can maintain its advantages as conventional payment systems evolve to become increasingly efficient. But what of smart contracts and crypto 2.0? One pathway we might consider is simply examining where legal agreements are currently utilized today, and discover how we can enhance efficiencies by “smart-ifying” as many of them as feasible. Alternatively, there is the Peter Thiel-inspired zero-to-one strategy: explore whether it is viable to employ these instruments to establish industries that are not yet in existence. Thus, let us take a diversion along that route, and explore some of the lesser-celebrated and captivating applications that smart contracts can offer.
Hashcash encounters Proof of Stake
While proof of work is primarily recognized today for its role in blockchain consensus mechanisms, its initial purpose was in fact something quite different. The first significant application of proof of work was Adam Back’s Hashcash, a protocol aimed at combating email spam by increasing the costs involved in generating spam emails. Similarly, proof of stake today is best known for its applications in blockchain consensus, with the theory stating that due to the way modern proof of stake operates – permitting users to put at stake substantial amounts of economic resources for security without, under normal circumstances, actually expending them – it could potentially achieve significantly greater efficiency. This presents an intriguing inquiry: is it feasible to harness these benefits to create a more efficient rendition of Hashcash utilizing proof of stake as well?
Let us begin by outlining the issue. The fundamental principle underlying Hashcash is based on the notion that email tends to be divided into two distinct categories: desirable email (“ham”), which individuals invest a significant amount of effort composing and from which they gain high value upon reading, and undesirable email (“spam”), which individuals typically expend much less effort writing and from which they derive a negative value from reading. Therefore, the theory postulates that by imposing a small obligatory cost on every email sent, the “ham” can arrive with minimal additional burden, whereas “spam” would be rendered completely unprofitable. This “cost” is incurred in the form of electricity and computational effort utilized to solve a mathematical puzzle that can be swiftly validated by the recipient’s client before the email is presented to the recipient.
The issues with this strategy are threefold. First, spammers might simply modify their approach to invest slightly more effort on each email (for instance, spending five seconds of human labor per message to increase the chances of capturing the reader’s attention or evading spam filters), and the threshold of computational effort necessary to effectively halt the majority of spam would be considerably high. Second, spammers are more professional and possess greater access to specialized computer hardware that can rapidly and cost-effectively solve these computational challenges, meaning a cost that could be five cents for the average user might only be a tenth of a cent for a spammer. Third, as this now-famous checklist outlining deficiencies in popular proposed solutions for email spam highlights, it is regarded as highly desirable to have a system where “sending email should be free”.
Consequently, here is a proof-of-stake alternative. When one sends an email, there is no need to solve an expensive computational puzzle; instead, one sends a transaction to the blockchain which establishes a contract containing a specific amount of funds as a security deposit. Along with the email, a private key is sent to the recipient, which they can submit into the contract to obliterate the deposit (or contribute it to a designated charity) if they choose to. If the deposit remains intact for a certain number of days, it is reimbursed to the sender. It is important to note that there is no benefit for the recipient in destroying the deposit – the only incentive to do so would be sheer spite. As a result, we achieve an asymmetry: the average expense for typical users to send an email would be minimal, as the recipient would only opt to click “Report Spam” in the rare event that they are malicious, while the average cost for spammers would indeed be quite substantial – and specialized hardware advantages would not assist spammers at all.
One might envision deposits reaching as high as a dollar in size, and a graduated system could even be adopted: senders can propose any deposit they prefer above a certain minimum, but the level of notification that the recipient observes would depend on the precise amount. If it’s 1, then several mobile alerts. Should it be $500, their smartphone would ring at full volume, overriding all other configurations – however, the sender ought to be ready to face the consequences if the recipient considers the intrusion unwarranted.
More sophisticated variations of this concept can be developed that do not necessitate sending a transaction to generate a new deposit for each email; one could envision a system where the sender distributes multiple keys to eliminate segments of the same security deposit, simultaneously providing signatures affirming that those keys are legitimate, with recipients sharing the signatures (but not the keys!) on a Whisper-like medium that permits them to quickly verify that a specific deposit is not “over-subscribed” with destruction keys (one particular approach to achieve this is by considering only signatures that include an index ranging from 1 to N as legitimate, and introducing a regulation that states two signatures sharing the same index can be submitted to annihilate the entire deposit with 10% allocated to the submitter; therefore, one can be fairly certain that no more than N signatures exist for that deposit). This would minimize the transaction burden to approximately one transaction per email sender annually. In any situation, smart contracts provide nearly limitless scope for innovation in refining the specifics.
Prediction Markets and Reddit
One of the predominant discussions in online forums like Reddit pertains to how much centralized regulation is warranted. One perspective posits that the strength of the internet largely derives from its equally distributed decentralized essence, with no single entity possessing a superior authority over others. While some individuals may exert more influence than others, (i) this represents a matter of degree rather than a binary classification of hierarchy, and (ii) the decision to be influenced ultimately rests with the audience. Conversely, the opposing viewpoint argues that without centralized moderation, communities are destined to descend into mediocrity and chaos of the undesirable variety; effectively, Eternal September, leading to the conclusion that having a limited number of users in charge is, as often is the case, a “necessary evil”.
In practice, community voting moderation is notably potent, yet the argument for centralization also appears to possess validity. Although comments that the community disallows eventually get voted down, at least on Reddit specifically, this process can be time-consuming, and there remains a window of one or two hours where such content lingers on the front page. Within a voting framework, this situation is to a degree unavoidable: if it were feasible for a deluge of downvotes to swiftly eliminate content from the front page, that would itself become a censorship mechanism for vocal minorities. However, what if a third avenue existed to tackle this dilemma through our favorite governance method: prediction markets?
Prediction markets have frequently been proposed, including by myself, as a governance mechanism capable of addressing significant large-scale decisions: whether or not we should provide a bailout for the banks, hire or dismiss a certain CEO, or enter into a particular trade pact. Nevertheless, it may be more beneficial to position prediction markets within a framework conducive to decisions that are significantly smaller and less alarming – perhaps on the scale of hundreds or even tens of dollars, or merely ten cents.
One could conceive of a structural design that operates in the following manner. Instead of merely being votes, upvotes and downvotes on a comment on a hypothetical PredictionReddit would serve as bets within a prediction market focused on that comment. This prediction market would be initialized by a compulsory bet that the individual contributing the comment must place, stating that their comment will be favored; from that point, upvotes and downvotes would adjust the “price” of the market according to public opinion. In 99% of cases, the market would remain inconsequential, with comments commanding high prices being displayed more prominently on the platform; however, in the remaining 1% of the instances, the comment would be forwarded to a meta-moderation panel, which would cast votes determining if the comment is favorable or unfavorable (or perhaps some gradation in between), with participants in the prediction market receiving compensation reflective of their accuracy in forecasting this evaluation.
In principle, the meta-moderation panel could potentially be extensive; it could include virtually every participant in the community, assuming an effective anti-sybil strategy is employed. Furthermore, even a SchellingCoin oracle could be utilized. It does not necessarily follow that 99% of markets need to be disregarded; instead, one could adopt a model where all markets are examined, but only a limited segment of the meta-moderation panel assesses each individual entry; the number of participants merely needs to be substantial enough to prevent collusion for insider trading purposes within the prediction markets. An alternative approach could be to make the scale or likelihood of meta-moderation correlate with the market volume, ensuring that the posts garnering the most focus are those where the stakes are highest. Ultimately, this method of amalgamating Reddit with cryptocurrency appears at least somewhat more encouraging than merely incorporating the capacity to extend condolences for someone’s demise by micro-tipping their relatives three cents.
In theory, either of these frameworks could be considerably expanded: envision advertisements that become more costly to sustain the more bothersome they are to viewers, or a decentralized search engine where anyone can “plug in” their ranking algorithms by engaging in the prediction markets, profiting only if those algorithms prove successful. Oleg Andreev’s 2-of-2 escrow could be enhanced with a reputation framework via a prediction market on the likelihood that the escrow deposit will be eradicated or payment postponed. Just keep in mind, security deposits and prediction markets are fundamentally interchangeable: a prediction market is a security deposit wherein anyone can contest and necessitate a larger deposit in response, and conversely, a security deposit is a prediction market compelling one specific party to make a compulsory wager.
This may embody a substantial component of the promise that crypto 2.0 technologies could provide: transitioning the internet from mere information technologies to economic information technologies that could radically enhance efficiencies across several sectors of the digital economy by utilizing incentives to more cleverly extract the intelligence we all inherently possess. In any event, let’s develop these instruments and discover the outcomes.
