agents acting together for mutual benefit — paying individual costs to produce shared gains
the theory
the core problem: why cooperate when defection pays more in the short run?
- iterated prisoner's dilemma — cooperation emerges when agents interact repeatedly and remember past behavior (Axelrod, 1984)
- kin selection — cooperation among relatives: Hamilton's rule, $r \cdot B > C$ (Hamilton, 1964)
- reciprocal altruism — cooperation among non-relatives through delayed exchange (Trivers, 1971)
- group selection — groups of cooperators outcompete groups of defectors (Sober & Wilson, 1998)
- indirect reciprocity — reputation makes cooperation viable among strangers: help others, and others will help you (Nowak & Sigmund, 2005)
five mechanisms for the evolution of cooperation (Nowak, 2006): kin selection, direct reciprocity, indirect reciprocity, network reciprocity, group selection
in nature
- symbiosis: mycorrhizal networks share nutrients between trees and fungi
- eusocial insects: division of labor in ant colonies, bee hives — individual sacrifice for colony fitness
- cleaner fish: mutualistic cooperation across species, enforced by partner choice
- microbiome: trillions of cooperative bacteria maintaining host health
in cyber
continuous process of cooperative games between neurons
implemented as an independent layer: cybernet
agents are rewarded for actions that increase the system's syntropy — order created from chaos
the mechanism has feedback loops: behavior that aligns with collective focus earns karma, behavior that diverges loses stake
the cybergraph enables indirect reciprocity at scale — every cyberlink is a public signal of cooperative intent, building reputation without requiring pairwise trust
game-theoretic foundations
cooperative games — Shapley values, core, Nash bargaining: fair distribution of gains from cooperation
coordination — the broader set of alignment mechanisms
see collective for the four collective processes. see egregore for the broader framework