The Use of Digital Knowledge in Society

In 1945, Austrian economist Friedrich Hayek wrote a seminal essay titled The Use of Knowledge in Society. In this essay, Hayek invites us to consider a world where all information is known to a single mind. In this world, the problem of allocating resources efficiently in society would simply be an optimization problem to the single mind/computer who possesses all the information. If this is true, should we focus on making information more transparent and available to a central planner, like a government? Hayek argues that this is futile because knowledge is fundamentally decentralized - it consists of "dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess". For example, it is an impossible task to constantly report our preferences for all the goods that we consume to a central planner, when we don't even know our preferences for tomorrow's lunch. Hence, the price system becomes an effective tool for economic coordination as individuals are able to signal what they want and how much they want it.

The price system leads to a "self-organizing system of voluntary co-operation", dubbed a Catallaxy by Hayek and his predecessors. Hayek's fixation on the idea of decentralized knowledge was for its superiority over centralized planning of societies - he even won a Nobel Prize in 1974 for these ideas. However, while Hayek was obsessed with figuring out how to allocate resources efficiently in society, I am obsessed with figuring out how to unlock the full value of local knowledge for the individual.

Local Knowledge = $

Every single human on Earth possesses knowledge which is valuable to someone else. A Google employee knows things about the business that a hedge fund would pay hundreds of thousands of dollars for. A terminally ill patient has valuable health data that could strengthen a predictive model. Your cinema booking for Toy Story 4 could be valuable to $DIS shareholders or a crazy ex-girlfriend. Fundamentally, by simply existing and making decisions in your daily life, you are generating local knowledge (location, purchases, health, consumption, ideas) that is valuable to someone else in the world.

Thanks to computers, we can express almost all types of knowledge as 0s and 1s. We can describe anything from insider secrets to scientific breakthroughs in text files, DNA sequences in .csv, and any real-world experience in video format. This effectively lets us communicate knowledge to anyone in the world - it is trivial to send or receive digital files on the Internet. This leads to the obvious question - can we create markets for all types of digital knowledge?

File-Sharing Protocols Suck

File-sharing protocols such as BitTorrent have created networks of anonymous users sharing all types of files; it ranks 5th place for the amount of traffic generated on the web. Although it suffers from free-rider problems, there are sufficient altruistic seeders in the world who keep popular content spreading across the Internet.

However, the BitTorrent model will fail to achieve our goal of building a market for digital knowledge. Firstly, there are many types of knowledge that are not valuable to a large group of people, but extremely valuable to a small group of people. An example includes solutions to homework of a specific college course - majority of people will have no use for these files, while desperate students will pay good money to acquire the knowledge. Secondly, torrents have no concept of "value" - there are no good ways for consumers to signal what they want, which leads to suppliers (seeders) essentially guessing. Naturally, this leads to a convergence towards content that appeals to the many - blockbuster movies, top40 songs, and hit games.

If we believe that every individual has local knowledge that is valuable to someone but not everyone, we need to invent a new model to match these buyers and sellers more effectively. Thankfully, we can easily build digital markets using digital money.

Blockchain-based Information Markets

Instead of regular file-sharing protocols, we can build a new protocol that allows people to attach a price tag (in cryptocurrency) to share any digital file. We can use smart contracts to facilitate the payment in exchange for digital files trustlessly. We can rely on decentralized file storage infrastructure such as IPFS & Filecoin, or use p2p communication protocols to send files directly from sellers to buyers. We can use encryption to ensure that only the person who paid for the digital file is able to access it.

For the blockchain skeptics, building this information market as a traditional web 2.0 product does not make sense. Would you sell an insider secret for a million dollars if a central party has to facilitate the trade, and hence have the power to see the contents of your message? We need to decentralize.

Building this as a decentralized, permissionless protocol also allows anyone to build applications on top of this platform. We will first see reputation calculators, Craigslist-like marketplace websites, and buyer/seller profile pages. We can further imagine insurance businesses, Netflix-like aggregation & curation websites, and mobile applications which make uploading as easy as posting on Instagram. May the best user experience win.

One of the biggest difficulties in making this work is the question of ensuring the validity of data. A seller could say he is selling a naked picture of himself, but instead send a picture of a cat. Numerai's Erasure project uses the concept of "griefing", where sellers have to stake cryptocurrency, and buyers can slash their stake if they are unsatisfied with the purchase. Myself and Ken have a different design - we propose a scarce reputation token combined with oracle/verifier services which endorse the validity of data. We also propose a suite of applications which let sellers reveal part of their data as proof while keeping the rest secret. For example, a seller could generate a "video summary", a series of screenshots of the video to show that they actually own it. In essence, we can strive to recreate the trust model that eBay buyers and sellers rely on.


We are all probably sitting on local knowledge that someone in the world would pay you money for. This will initially start with digital files you own which are already explicitly valuable but cannot be found on torrent networks - rare ebooks, esoteric digital courses, access to paid forums & private Telegram group chats. There is already a market for these types of fringe content, but it is difficult to facilitate exchange using the existing platforms we have today.

We will then see personal knowledge which is difficult to value being bought and sold - insider secrets, sex tapes, personal texts & emails. Someone somewhere will find a use for this knowledge.

Finally, the long-term goal is to foster a flourishing marketplace for all types of exchange of local knowledge. Entrepreneurial individuals can buy medical history data from 100,000 people and repackage it into a data set for a researcher. Regular people can take photographs of stores they are shopping at for a hedge fund to forecast next quarter sales. Teenagers can sell their iPhone Screen Time to venture capitalists who want to find the next hot app. There will be many interesting and unique ways for people to make money based on what they know and their daily experience of the world.