What's at Stake? Why the Erasure Protocol is such a big deal (Part 2)
This is Part 2 of a two part series on the Erasure Protocol. Part 1 covers what it is and why it matters and can be found here. Part 2 covers what it’ll enable and the implications.
SO!
In aggregate from Part 1 of What’s at Stake?, we established that:
Erasure’s fixes the collapse of the internet’s information market.
it reduces the cost of bullshit discovery by pricing conviction.
In short, it makes it easier to work out if the information is true or not.
What does that actually enable in a practical sense?
By lowering costs of engagement and friction, Erasure creates markets where we've never seen them before.
To date, the first iteration of ErasureBay has been to source copies of 1930s New Yorker magazines, contact details of the lawyers who did Jeffrey Epstein’s deposition, a bunch of database building, CT scans of COVID patients lungs and more.
It’s mainly just people playing, but it’s starting to get interesting.
As the market develops, new services for ‘selling’ information will emerge.
Here are a couple of examples of things that it *could* be used for:
Troll taxes for social media comments: If someone wanted to post a comment on social media or a newspaper article, they’d have to stake to be able to post. If it becomes costly to post things that are offensive, then people are likely rethink what they post, with the result being that it becomes possible to have more ‘constructive’ discussions given that both parties have skin in the game.
Staked dating apps: On Tinder or Bumble, there’s a lot of ‘sorting’ through ill-fitting matches and dead-end chat conversations largely because daters don’t have much skin in the game. If they had the option to stake $20 which could be griefed if they were not respectful or trustworthy, it would be another signal that the dater has integrity, especially when compared against someone who doesn’t.
A crowdsourced Climate X-Prize Fund: Jeff Bezos recently announced a $10B climate fund, with a focus on funding scientists, activists and nongovernmental organizations with grants to be issued starting in summer 2020. Given the scale of the funds, they’re inundated with applications. The personnel in the Climate Fund organisation could assign $20m for a staked, crowdsourced database of potential research and technologies to invest in, paying out anyone who contributes staked intelligence of interesting projects with rewards, allowing them to find previously undiscovered potential. It’d rapidly harness the intelligence of the crowd to find the most effective areas of investment, amplifying investment impact.
Due diligence marketplace: Bernie Madoff's Ponzi scheme was the largest in history. It was also 'an open secret...one of those situations where everybody knew except the people who needed to know.' There was a market failure because those who knew weren’t incentized to say anything, and those that did weren’t trusted. Savvy investors could seek red flags about companies they’re looking to invest in quickly and rapidly.
Technical/code audits: Having skin-in-the-game will help avoid issues like the recent situation where user funds get locked up/lost.
Internal whistleblower markets at firms: a way for management to better uncover internal conflicts or abuse is by offering to pay to discover information that an employee has, using staking to incentivise skin-in-the-game. Merely the threat of malpractice being more easily discoverable is likely to improve behaviour.
An Earn.com with griefing: The original Earn.com was a conceptually good idea: allow people to send someone famous an email with the recipient being paid only if they respond. It never took off, and I posit that it’s because it missed the other side of the equation, which was ensuring that the person responding also had skin in the game, and could be ‘griefed’ if they took advantage of it. It will monetise influencers in ways that we’ve not even conceived of yet.
Put all together, we’ll be able to stitch together models of the world that we can all agree are the ‘truth’ and then build a larger single metamodel of the best information in a way that is both true and able to evolve (thanks Fred Ehrsam for this conceptual idea).
Further, as markets like the above get built out, the current model of search can be inverted. Instead of people having to search for information, information competes for people’s attention with the risk that it’ll be griefed thus increasing the ‘cost’ (and incentive) of getting it right.
Why does this matter
There are three primary reasons that I think that Erasure matters for building a better world:
Firstly, it opens up high-trust interactions to wider groups. Erasure lowers the friction to a world that already exists for a small group of people. Anyone who’s had the luxury of working in a highly networked and trusting society where you can call your very smart and well-connected friends to vet someone or provide you reliable information has had a taste of what Erasure can offer. It substantially expands the group that you can do this with and cuts through the crap using a trust-pricing mechanism.
Secondly, it flattens information marketplaces: It replaces workplace and university degree credentialism with staked conviction of ‘authenticity’ and truth. It allows the young, disadvantaged, ignored and unfairly discredited to at least be able to compete in the information economy alongside those who are more well connected in a ‘flat’ marketplace. I expect it to reduce the amount of rent-seeking those who are in positions of authority or privilege are able to demand respect.
Say for example, you're a 22 year old math graduate who lives in Kampala, Uganda and has a job working in the Treasury Department. In your spare time, you have developed really good data and models for equities predictions in the Uganda Securities Exchange that are not currently covered by any analysts.
At the moment you cannot sell that trading information easily - it’s hard to connect to hedge funds who are based overseas as they have no reason to know who you are or how to trust you. You might not have anywhere near enough capital to be able to materially trade these predictions yourself but it’s plausible you could sell that information to someone who would find that information valuable on Numerai Signals eventually substantially unlocking both new revenue for you and new investment for Uganda.
Lastly, it allows us to collectively build ‘truth’ databases: Using these tools and knowing that the information we’re gathering is accurate means we can quickly crowdsource large generalised datasets with staked information that can be used as the basis of other calculations without needing to trust any centralized authority for that information. At its core, it’ll help us agree on a common set of facts that will help us ‘weaponise truth’ to help counter disinformation and fake news.
The Case for the Numeraire (NMR) token
So, if you’re to believe the above, what is the case for the underlying token, and why would it have value?
The supply of the token is fixed - there will only be 11 million tokens created in total of which 4.5m are currently circulating with the remainder sitting with the Numerai/Erasure team to be used to pay out Numerai tournament winners and help fund the development of the protocol.
Conceptually, I think of the token as buying usable access to the community that is built around it.
Intelligently, when Numerai was started, it assembled the largest group of machine learning specialists and data scientists by giving everyone at Kaggle free Numeraire tokens. Those users supercharged the early days of the hedge fund, and they’re now leveraging this ‘supply’ over into things like Numerai Signals.
This has a natural tension tendency towards network effects in the sense that if the protocol is useful, then it’ll attract in more users who’ll use the tokens to utilise the talents of the community it has built around it. And it’s working - with 10x the number of NMR tokens staked in the last 8 months.
From Olaf Carlson-Wee of Polychain Capital (when Numeraire was only being used for Numerai):
“If the Numerai fund performs well, that should lead to more assets under management, which should lead to more revenue for Numerai’s general partner which should lead to higher payouts for the data scientists, which implies a higher Numeraire valuation.”
He says a good proxy for what a Numeraire should be worth (to either holders or speculators on secondary exchanges) is the value of all future payouts to data scientists on the Numerai platform.
Further, it will benefit from a growing scarcity over time as more and more people grief each other and tokens are burned (ie. removed from the supply). While only 3% of existing bids actually end up with a griefing process happening, a removal of 1-2% (assuming they’re not griefed at 100%) of the supply a year helps create further scarcity.
As of writing, the token is valued at around US$19. With a circulating supply of around 4.5m of their 11m total released, that gives the project a valuation of around $85m.
The total market for global market research industry turned over around $76B a year in 2018 (and this is quite narrowly defined definition of what could be done on Erasure). If 1% of that migrated to Erasure’s platform, we’d see $760 million per year in business transacted across the platform. Assuming that 25% of that gets staked as skin in the game for the information provided, we’d be looking at market demand of around $190m of NMR just for the year.
If the project continues to build new features and scale exponentially, it is not unreasonable to think that the overall Numeraire project would be valued at 5x annualized staked rate (across all use cases), or around $1 billion. If that was the case, the overall price per token would be in the range of USD$222, or 11x from current valuation.
If the Numerai hedge fund on the Erasure protocol was to grow to eventually ‘manage all the money in the world’ as was Richard Craib’s stated aim when he launched Numerai, then we’d be looking at an even larger TAM.
The world had around $75T in capital under management in 2018. If we assumed that they were generating fees of around 1% of AUM, that’s around $750B in fees paid to financial service managers a year. If Erasure/Numerai was to capture 10% of this through much superior returns to the market through their crowdsourced intelligence, and again, 25% of this was staked, the annual demand for NMR would be in the region of US$18.75B. With a valuation of 5x staked value again, we’d be at 447x in value from where we are today, or ~US$8.5k per NMR (based on full supply).
And that’s just Numerai, without also adding the other services that would also generate demand for the Numeraire token (such as the other examples above).
I’m the first to admit that the above valuation sounds outlandish. I almost didn’t include it because I know that it’ll likely attract ‘wen moon’ eyerolling that I myself direct at others in the crypto space. But this project is building a service that already has paying customers, has traction, is hard to replicate with strong network effects in a world that is looking for ways to empower new lower-cost and crypto-enabled business models.
There have been multiple calls for an more sophisticated NMR valuation model, but it’s yet to be built. I’m not sufficiently skilled to build it myself yet, but would be interested to collaborate.
I’ve got a few more points to make on Erasure/Numeraire but they’ll have to be for another day. In the meantime, would love to hear your thoughts and feedback. DM me on Twitter (@oliverbruce) if you want to talk further.