Ethereum service composition, gold rushes, and a better Wikipedia

A while back I had an idea of how to piece together a handful of Ethereum services to create something highly automated that produces value for the general public. This composition of services is meant to illustrate what is possible with this technology.

The services involved are:

  • Ethereum Alarm Clock (currently kind of sad and dormant, hopefully not for too much longer)
  • Oraclize.it
  • A prediction market which crowd-sources “truth”

The idea is as follows.

Pick a market who’s outcome can be queried via some web API query, or more ideally, many different web APIs. Lets use something like “The annual rainfall in Townsville Colorado during 2018 will be greater than 40 inches”. This this market we should be able to find a number of APIs which will provide us with an objective answer.

Using the Alarm service, schedule a transaction to occur during the phase of the market where the result is being crowd-sourced. That transaction would initiate an Oraclize.it query to the chosen web API such that the response from Oraclize.it will be reported into the market.

This is admittedly an over-simplification of how the reporting mechanism for a quality prediction market would work, but it should illustrates the fundamentals of this idea. Accounting for any complexity in the actual reporting mechanism should be a solvable software problem.


If we assume that the prediction market provides a monetary reward for those who correctly report the “truth”, then you now have a magic money machine. The only input necessary is choosing the appropriate API query. Everything else is automated.

The term I like for situations like this is a “Gold Rush”. Something with a relatively low barrier to entry which has perception that anyone can join the part and start making money. A modern example of this is dedicated bitcoin mining hardware.

Modern tech gold rushes seem to quickly become a race to the bottom. The early adopters do well for themselves. Their success prompts a large influx of new players. Since there is a limit to the amount of money to be made everyone’s profits trend towards zero.


Gold rushes tend to have two primary groups who benefit.

  1. The people who are selling “mining picks”
  2. The people who reap the secondary rewards of the mining.

An observant reader will notice that the people actually mining are not listed here. That isn’t meant to say that you can’t get rich mining a “gold rush”. Only that getting rich in a gold rush is the exception.

In the case of the early Americas, those selling actual mining picks and other mining hardware did well for themselves, while the country as a whole benefited from the settling of new territories as well as global economic benefit from the increased gold supply and circulation.

In the case of Bitcoin mining, those selling dedicated mining rigs seem to be doing well for themselves and the broader Bitcoin network benefits from increased security from the increase in hashrate (admittedly these benefits are not entirely black and white).

In the case of prediction markets I’m not yet sure who the “mining pick” vendors are, but the value being produced is “truth” and everyone benefits from a reliable source of truth.


There is another thing I find fascinating about this idea. Join me in a journey further down the “future of smart contracts” rabbit hole.

Lets imagine that rather than an individual supplying the API queries needed, you instead create a special smart contract with its own token. Token holders can vote on what API query should be used using their tokens. The smart contract could either chooses the API query with the most votes, or uses all of the submitted API queries across some number of accounts it controls.

  • Those who committed their tokens towards API queries which yielded answers which are rejected lose their tokens.
  • Those who committed their tokens towards API queries which yielded accepted answers have their tokens returned plus some reward from the pool of loser tokens.

The best term I can come up for this is a meta prediction market. It’s markets predict which API queries will produce the “correct” answer for specific markets on a generic prediction market.


There is a lot of “hand waving” going on here. There are non-trivial complexities that would surface the moment you started building this system. There are going to be difficult game theoretical problems to solve. I’m of the opinion that all problems are solvable with sufficient intellect and resources, so from my perspective this feels like something we’re very likely to see built.

But what I find exceptionally cool about this type of automation is that it produces value for the general public. The value is actually being generated by the prediction markets, but these meta-prediction markets should drive the cost down.

This could be the foundation for a new type of Wikipedia requiring less human input and with a more objective dispute and fact-checking mechanism. Wikipedia is one of the more valuable public goods generated by the web. But Wikipedia has required a monumental effort to both create and maintain. This technology could create the same body of knowledge for a fraction of the cost. If we are able to get enough things right in how we build these systems it should also be more accurate and objective than the current Wikipedia is likely to ever be.

This is why I’ve focused everything I have on Ethereum. Previously, the internet and society in general felt like a choose-your-own-adventure story where all choices led to a different dystopian future. Ethereum provided me with the first viable path that could lead to a future I wanted to be apart of. I don’t claim it’s guaranteed to get us there but rather that it is my best hope for the future I want to live in.

And for that… I’m all in.