Trust Models

I'm not OP, and the books that jumped to my mind when I read this are not blockchain-related.

They are related by questions of society, trust, and incentive design.

They seem largely aligned with Vitalik's assessment of the `0 of N`/`1 of N` assessment

Before clicking through to Goodreads, though, check out this book review of Seeing like a State by Scott Alexander:

https://slatestarcodex.com/2017/03/16/book-review-seeing-lik...

Next, check out both of these books:

- James C. Scott, Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed [0]

- Michael Huemer, The Problem of Political Authority: An Examination of the Right to Coerce and the Duty to Obey [1]

Neither book will disappoint!

[0]: https://www.goodreads.com/book/show/20186.Seeing_Like_a_Stat...

[1]: https://www.goodreads.com/book/show/15794037-the-problem-of-...

Edit: for clarity.

I heartily recommend Seeing Like a State to anyone interested in the development of the modern state, some of the largest misteps made by a slew of modernist authoritarian regimes (such as the Soviet Union), and an introduction to some very legitimate criticisms of modernism which postmodernism makes.

However, I want to caution anyone from thinking that this book says anything about Game Theory. I'm not sure why it's being recommended here, it doesn't have much to say about the current topic, except in that it often points out that authoritarian governments are free to ignore any feedback, feedback of the kind that incentive schemes are usually designed to generate.

I want to defend Vitalik here and say it's unreasonable to expect him to address everything in a single blog post, there are a lot more failure modes than just malicious networks, none of which are explicitly mentioned. [1]

Each of these kinds of trust fit relatively neatly into the framework he proposes though. Blockchains are designed to run on tens of thousands of nodes distributed around the world, which naturally insulates them from any one country deciding they don't want cryptocurrencies to be used within their borders. And even if every other country decided to ban Bitcoin, if America still allowed it then Bitcoin the protocol would happily continue to work within America's borders, in that sense it has 1-of-N trust in countries allowing it.

On the other hand, if America or China decided they did not want Bitcoin to exist any more, and they were truly willing to do whatever it takes to shut it down, they have enough tools at their disposal to shut it down. This situation doesn't quite fit into V's framework, I think because some of the N are more important than others?

[1]:

- Do you trust whoever writes your client / mining software to write code without showstopping bugs?

- Do you trust them to use a reliable build process?

- Which websites do you trust to tell you what their public keys are, so you can check the signatures of the binaries you run?

- Do you trust whoever sold you the computer you're running your client / mining software on?

- Do you trust the mining rig manufacturers?

- If your protocol relies on a notion of time, do you trust the time servers?

- Do you trust mining pool operators not to use their hash power maliciously?

- Do you trust that if someone found a way to forge signatures, the rest of the world would know?

The point I was making is that in the 1/N system, the failure of the one person or element brings the whole thing down, it's a single catastrophic failure mode, where when N/N at the other extreme, the failure is contained to that group. The partial ones mean that the number of people/parts that have to fail is greater to bring the whole system down. (the example case is storing a single shared secret, which if compromised, means re-enrolling all N people in that secret again)

You can replace a root of trust, but then you have to re-enroll all parties into it. When you have partial or federated trust, the failure is contained to the N/x group.

I'm thinking there may be a universal trade off between number of trusted parts and the consequence/cost of failure.

The concept being that a trust model is essentially a non-reversible function you iterate from its initial unique "trusted," conditions (like derived keys, or a certificate chain), and if those unique initial conditions are replicated, you have to re-compute the entire function from a new unique initial condition, or risk "fake" branches.

A root of trust is the root node of a tree (or a DAG these days), and a compromise of any of the roots will effectively isolate its downstream branches. Compromising a single N/1 root means you compromise the whole tree, where a federated multi-rooted tree means the damage can be contained - in the model I'm thinking of.

TL;DR: the cost of the recovery mode in the 1-of-N case requires re-instantiating or enrolling all of N, which basically means bootstrapping the whole scheme. Fine for a closed system, Hard for an open one.

Yes, also a good one. I'd recommend pretty much all of Lex Fridman's and Eric Weinstein's podcast episodes in general. Lots of insightful conversations with (in my opinion) some of the most interesting people in the world.

Links for anyone curious:

Eric Weinstein's The Portal Podcast - https://www.youtube.com/playlist?list=PLq9jO8fmlPee9ezOraOHA...

Lex Fridman's Artificial Intelligence Podcast - https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuK...