Is the following a good analogy?
1. The Internet aspires to be neutral infrastructure for communication.
2. Permissionless blockchains aspire to be neutral infrastructure for computation.

Head of research at a16z crypto.
Professor, Columbia University.

Would suggest to substitute computation for coordination — this got to be the core value-prop in my mind, otherwise distributed computation wouldn’t be that valuable.

Research at cyber.fund | prev mathematician | artofkot.xyz

In general, have been using the following sequence as a quick “path” to blockchains:

1. Need for scaled coordination 
↓
2. Centralized systems 
↓
3. Collusion problems 
↓
4. Neutral (=>permissionless) distributed systems via consensus
↓
5. Incentives & Sybil problems 
↓
6. Crypto assets & PoW/PoS

And to add compute into the picture:

7. Need for “programmable” coordination
↓
8. Smart contracts & distributed computation

I sort of view centralized systems having collusion problems “in the limit”: the central party can collude with themselves and do what they want. Agree this is a bit inaccurate and there is probably a better way to present this.

Re 6: 100%, thats why I wrote PoW/PoS right under the Sybil problems :)

Interesting. I don't follow 2=>3, why is collusion a problem specifically for centralized systems?
Also, in 6, I would separate crypto assets (which are interesting in their own right) and sybil-resistance mechanisms (an implementation detail, in effect) as two different things