What's at Stake? Why the Erasure Protocol is such a big deal
This is Part 1 of a two part series on the Erasure Protocol. Part 1 covers what it is and why it matters. Part 2 covers what it’ll enable and the implications.
What is Erasure and how does it work?
I think that the Erasure Protocol is going to be a big deal and you should pay attention to it.
In fact, I’m so sure of it, I’d be willing to stake decent money on it.
Right now I can’t do that easily, but if the team there has their way, I will be able to very soon. And if it takes off, I think it’ll change the world as we know it.
I haven’t yet found anybody articulating a sufficiently in-depth explanation for why Erasure’s potential is so big, nor its implications so profound, so I wrote this up for a few friends and thought I’d publish it more widely for others to read, question and critique.
I hope you like it and would love feedback - ping me at @oliverbruce on Twitter.
How do you know if something can be trusted?
Say you run a tours business to South America and some random person reaches out to you by email and says that they’ve discovered unicorns in Peru, and they alone can tell you how this is going to create a massive boon of unicorn tourism that will create massive profits for you.
However, you need to pay them $50,000.
This, but a unicorn…
Your scam senses go off. Unicorns sound preposterous, but what if they’re right? Crazier things have happened.
Technically, you should buy it if you have a way to know the information will lead to the profits they claim but without *knowing* if it’s actually any good, you'll have to go and spend a lot of time researching who they are and whether you can trust them.
This research is a hassle and takes energy with no guarantees. You can’t know if you can trust their website with ‘quotes’ from satisfied customers called Steve, nor if the sites giving rave reviews aren’t just a scam themselves. Balanced against all the other options you have of how you can spend your time, the likelihood is that you just write it off as not a good use of your time and keep doing what you’re doing.
BUT if this same person emailed you and said they have the exclusive details about the forthcoming explosion in Peruvian unicorn tourism and the information will cost $50,000, but offer the following:
An easily verifiable history showing that they also discovered and sold information about flying narwals in Argentina and rainbow zebra’s in Tanzania, both of which have seen a massive boom in magical animal tourism,
AND that they’ll lock the money you pay them AND $10,000 of their own money into a box with a remote controlled bomb that you can detonate for $2,000 anytime in the month after the transaction.
The now infamous Argentinian flying narwals.
Image credit: FeartheFuzzyBear
Effectively, you and the seller are in a ‘Mutually Assured Destruction’ or MAD transaction - where you will both lose if what they give you is not what it says on the tin.
Chances are, you’d be far more willing to consider this offer, and thus we build a bridge of trust that enables a trade to occur.
This is a solution to something deemed ‘market collapse’ or ‘the lemon problem’, and that turns out that that is a really big deal, but more on that later. Before we get there, we need to understand what is trust, and why does it matter.
How do you know what you know?
At the moment, if you want to find some information out, where would you go and find it?
If you’re like me, for small stuff you’ll go to Yelp or Rotten Tomatoes or some sort of news site where you trust the journalists to know what they're talking about, or in the case of larger knowledge sets, to a university or company with expertise in that area.
Over time we’ve delegated trust to institutions like this to delineate for us what is or isn’t true. We extend trust to them because it offloads the mental lifting of having to assess the relative ‘safety’ of information in their area of expertise.
Practically speaking, if two people trust the same institution they can easily trust each other.
If someone tells you that they’ve transferred money to your bank account, you check with them, and if they tell you it’s there, you can complete the transaction. Similarly, if you order an Uber and the car turns up, you know that that drivers been vetted, while they know your credit card is on file and they’ll be paid. These institutions work like iterative games - we build one way of trusting someone, and then continue to seek new ways to create additional links of trust with them and others over time (thanks to Flo for this insight).
Our ability to trade scales easily as long as you both have institutions that allow you to trust each other. It’s why trade, in general, is easier in countries with strong standards and rule of law.
But establishing trust is expensive and only extends as far as the institution’s reach does. If you want to trade with someone who doesn’t have access to the traditional mechanisms of building trust - access to legal and banking services, credentialing from institutions that you can’t determine the quality of etc. - you’ll likely not bother, and instead try and find someone who has links to things you trust, so that you can feel ‘safer’ in engaging with them.
The internet was a marvellous invention, in that it allowed anyone to publish anything they wanted online, going over the top of the ‘trusted’ institutions that had been the arbiters of information till that point.
As Alfred Whitehead said, “Civilization advances by extending the number of important operations which we can perform without thinking about them”.
One example of how the Internet is turning upside down old mechanisms of trust is fake news. With the low cost of spinning up a website and appearing trustworthy, there is no or little cost to spreading disinformation. The old mechanisms that we used to trust no longer seem to be working as well as they did, and trust in our institutions is at an all time low.
If you’re anything like me, you find yourself spending a lot of your time on social media trying to work out what’s going on and whether the information is actually accurate or not. This is dangerous, especially in the era of things like COVID, as we discover and distribute existentially important information.
Trust is a prerequisite to doing business, and is immensely value-creating. But we still haven't found a way to create it that is truly digital-native.
The important question is: how do we build tools that expand trust in each other? How can we use them to hold those who purport the ‘truth’ to account?
With that context, what is Erasure?
The Erasure Protocol is an information marketplace. It allows two people who don’t know each other to trade ‘true’ information. It does this by using a series of tools to build trust.
Rory Sutherland, Ogilvy vice chairman, TED superstar and an amazing thinker on perceived value, argues that there are three ways to establish trust: Reputation, pre-commitment, and reciprocation. (Again, thanks Flo for this ;).
Erasure offers all three through a series of smart contracts built upon Ethereum.
Immutable Reputation: a user’s history of trading is there for all to see, including ratings and past mistakes in a way that can’t be deleted.
Pre-Commitment: requires the owner of the information to commit some of their own money to be locked up for a set amount of time in a process known as ‘staking’. Think of it as making a bet on the reliability of the information (‘I bet you $500 it’s correct’).
Reciprocation: If the purchaser doesn’t like the information - it wasn't correct, not what they said it was, not adequate, etc - then they can pay to destroy the sellers stake through something called ‘griefing’. If they try and screw you, you can screw them.
At the moment, there are three projects that work on the Erasure protocol.
Numerai: The original project deployed on the protocol was the Numerai hedge fund. Users are given anonymised financial datasets on which they predict what will happen next. They stake their predictions with a bet on how accurate they are. If they’re right, they get rewarded in crypto currency. It’s now one of the largest weekly data science tournaments on the internet.
Numerai Signals (formerly ErasureQuant): allows anybody to provide staked predictions for the stocks on the Russell 3000 for hedge funds to buy.
ErasureBay: a generalised information market that lets anyone pay for any information they want on Twitter and have people stake money with their responses.
So that's the nuts and bolts. If you want to understand more I highly suggest you go and check out the Placeholder VC thesis about Erasure or the Erasure blog, both of which have far better and in depth breakdowns of how it works.
Why is it significant?
Erasure is important for two reasons - it fixes the collapse of the internet’s information market and it reduces the cost of bullshit discovery by pricing conviction.
Fixes the Collapse of the Internet’s Information Market:
In 1979, the economist Frank Akerlof won a Nobel Prize for identifying the ‘lemon problem’. It explained why information asymmetry results in a market collapse.
He used the example of the second hand car market for ‘lemons’: If the car buyer has no information about the car they’re buying and the seller has a lot of information, and the average price of a second hand car is with a certain price range, then the buyer can’t know if the seller is giving them a ‘good’ car, or a ‘bad’ car for the price.
Because the buyer cannot determine the quality, and having a crap car that breaks all the time is very costly, they’re likely to err on the side of not purchasing, thus the term ‘market collapse’.
But how do you *know* he’s not?
Over time we've done things like regulate car dealers to provide guarantees and track mileage so that sellers can’t lie. These measures give buyers peace of mind and allows them to trust the seller enough to make a purchase.
But regulations are slow and only apply to people who are in the same legal jurisdiction. It’s hard to extend it to information on the internet, where enforcement is very hard.
Erasure establishes a globally accessible and trustless market incentive to provide the same surety for information.
Want to claim that 5G will cause coronavirus? Put up $500 to prove your conviction. Eventually people consuming information will be able to easily assess and trust information they find.
Being able to establish trust between two people who don’t know each other where the incentives are all aligned for good behaviour substantially expands the potential scope of human co-operation.
Reduces cost of discovery by pricing conviction
To the earlier example of magical unicorns in Peru, you would need to spend a lot of time working out if that person and their (seemingly preposterous information) is trustworthy through a long and intensive process of research, reference checks and consultations with local magical animal experts. Or you can see that they’ve priced their stake at $10,000 (vs. $100), and deduce very quickly that they clearly know what they’re doing because otherwise they wouldn’t put that amount of money up.
This is ‘price signalling’ at work. Freidrich Hayek, in his seminal essay, The Use of Knowledge in Society, laid out how price communicates all the information along a supply chain for a product in a highly digestible number. You don’t need to know everything about the dynamics of supply and demand or challenges with governments in some far-off remote area when buying something - this is all just folded into the price.
In Erasure, the staking price communicates ‘how sure are they’ about the information. A buyer can see how highly convicted a seller is about the veracity simply by seeing the price that they are willing to assign to it.
All up, it’s a digitally native version of ‘I promise you I’m not bullshitting you. Hold onto my wallet as proof’. And that, extended out into the world through the internet can be an amazing thing.
What does it all mean? In Part 2, I take us through the implications.
Thanks to the following for their insights: Horace Dediu, James Gross, Georgie Fenwicke, Flo Crivello, Moira-Clare Donovan, Adriaan Kroon, Jay McNally, Adam Flynn, Hank Rowe, Josh Daniell, Junia Ooi, and Anders Brownworth.