ZK applications enable use cases constrained by verification costs. My mental model for thinking about promising ZK app is to look for places where people are likely to lie because verification is too expensive or unpractical.
Technology and marginal costs
To make sense of new technology, you need to understand the costs they decrease, which leads to understanding the new design space it opens.
For instance, the internet brings marginal costs to zero. It costs nothing for Google to serve one more user, and it makes no difference to an author whether their blog is read by 27 people or a billion. The internet is infinitely scalable on the backend. This property upended industries, created products we use daily, and generated billions in shareholder value.
The key to creating new internet businesses was to capitalise on the absence of marginal costs. Media is a prime example. Before the internet, media companies had non-negligible marginal costs for each additional customer, involving printing and delivery. Within these parameters, your advantage lies in local implementation, and printing/distribution economies of scale, more so than the content itself.
On the other hand, online media like blogging has zero marginal costs. An author can become a leading voice in a niche and attract readers worldwide. The absence of marginal costs leads to both a radically changed cost structure and a massively expanded market. One-man businesses making millions in revenue like Stratechery or Byrne Hobart’s The Diff are great examples of ventures that were previously impossible.
The cost of verification
Building trust online is challenging. Trust is based on understanding, and physical interactions provide more signals than online. These signals are also easier to verify.
Claims are easier to verify in person because of the additional context. This allows for a quick assessment of trustworthiness by evaluating the coherence of the claims taken as a whole. Age, jobs, income, and sports performance can be easily referenced in person, but not online.
My mental model for where zero-knowledge proofs (ZKPs) make sense is in online services with high costs of verification i.e. that require personal information, where individuals are likely to lie. For example, dating apps where people lie about gender, age, salary, or height.
If you wanted to do this in a dating app right now, you would need to use a third-party verification provider, essentially asking your users to KYC, which is not only a catastrophe for conversion, but also a significant additional cost per user. By contrast, very simple ZK products like proof of passport or zkpassport, can verify a use age with a quick passport chip scan, without revealing any other information, and without having to rely on a costly trusted third-party.
There are many such cases where information is valuable but the cost of verification narrows the design space. ZK makes sense for these applications where you want to keep private information hidden but verified.