TB: Proof of Stake (Buterin)

Core Thesis

Blockchains represent a third form of incentivizing productive activity alongside markets and institutions. For 10,000 years, those two were all we had. Crypto adds a new primitive: programmable, trustless, self-executing commitment mechanisms that can fund public goods and enforce agreements without a trusted intermediary. The fundamental problem crypto and AI safety share — controlling a more complex system with a simpler one — may be the defining challenge of the 21st century.

Key Takeaways

A third incentive mechanism

  • Markets incentivize private goods efficiently; institutions concentrate authority to produce public goods. Crypto is a third option.
  • Bitcoin's seigniorage — the entire value of the currency — flows directly into funding the security of the network. A public good funded by its own use.
  • Blockchains are "magic computers that anyone can upload programs to and leave the programs to self-execute" — Lego Mindstorms for building economic and social institutions.

Superrationality and trust

  • Superrationality: determine if the other player is virtuous; if yes, cooperate; if no, defect. Modern large organizations let people cheat others' ability to read their minds, making this cooperation harder.
  • Online, we are less "leaky" — it is easier to appear virtuous while intending to cheat.
  • Crypto attempts to make commitment credible without requiring trust in persons.

Crypto and AI safety as the same problem

  • Both are trying to regulate a very complex and smart system with a very simple and dumb system whose properties once created are inflexible.
  • AI safety: agents with IQ 150 trying to control agents with IQ 6000.
  • Crypto: agents with IQ 5 trying to control agents with IQ 150.
  • Cryptography is one of the very few fields where adversarial conflict heavily favors the defenders.

Free speech as a social principle

  • Freedom of speech is not just a law — it is a social principle whose goal is to ensure that ideas that win are good ideas, not ideas favored by those in power.
  • Bad argument gets counterargument. Does not get bullet. Never. Never ever never for ever.
  • Trying to get someone fired for holding an idea is silencing an idea without addressing it. A bad response.
  • The instability of majority games (cooperative game theory) is a simplified model for why there may be no "end of history" in politics — no system that proves fully satisfactory.

Mental Models

  • Incentives Matter — blockchains restructure incentive mechanisms at the protocol layer; seigniorage becomes a design variable
  • Second-Order Thinking — the market can arbitrage our own morality, placing unsavory tasks in the hands of those willing to suppress their conscience at lowest cost

Source note