Home > Essays
There is no sufficient threshold of evidence necessary for someone to be king.
It doesn’t matter if:
What matters is that the right people believe they are king. The evidence only matters insofar as it supports the belief.
If we mistake the threshold of evidence necessary for belief with evidence of being king, The King Problem is unsolvable.
Not just with computers; it’s unsolvable in the real world.
Computers have an even harder challenge when approaching The King Problem: most of the evidence supporting belief in the real world is not cryptographically verifiable. I can’t cryptographically verify your parents.
Organizations, companies, governments, clubs, etc. are all just variations of The King Problem; each with varying thresholds of evidence necessary for proving membership to interested parties.
We should not try to solve unsolvable problems. The Matrix says it best:
When it comes to "organizations," they are an emergent social phenomenon. They are the result of people waking up, going out into the world, and peopling at scale.
We don’t need to “build organizations” in p2p systems, we just need to capture the relevant social interactions among interested parties in a provable way. If we document the social interactions, organizations will emerge as a feature of the system.
There is no such thing as a cryptographically verifiable identity.
Shallowly, possessing Alice’s private key does not cryptographically prove that you are Alice any more than possessing Alice’s car key proves that you own Alice’s car.
More fundamentally, trying to prove that I’m Alice is meaningless because “Alice” is meaningless as a designator. What we want to prove is not that Alice is Alice, but that the person I’m talking to is the person I refer to as Alice.
When it comes to Alice being Alice there is no such thing as global consensus because there isn’t global knowledge of Alice. The only people interested in establishing the authenticity of Alice is Alice’s social group.
We call these the interested parties.
If all of the interested parties agree that Alice is Alice, then Alice is Alice.
If all of the interested parties agree that Alice represents E-Corp within a scope of capabilities, then Alice represents E-Corp within a scope of capabilities.
Governments exist because people say they do. Organizations exist because people say they do. This is Fiat Identity. We extend that into our network by saying “These keys belong to this real-world person because I say they do.”
Not only is that sufficient, that’s the best we can do. We can’t cryptographically verify Alice’s I-9, we can only cryptographically verify that Bob was endowed with the authority to conduct I-9s on behalf of an organization and Bob attests that Alice successfully completed the I-9. That social interaction, Bob sitting with Alice and inspecting her government-issued IDs, needs to be captured in a data structure that can be socially and cryptographically validated. Once it is, Alice is now Alice within an organization.
This might seem like it is somewhere between stating the obvious and being pedantic, but proof-of-personhood appears to be a Holy Grail in p2p systems. I'm asserting that is not solvable.
I'm also asserting that peer-to-peer systems become easier to reason about, and easier to build, when you model identity as being fiat.