Observations find a measurable algebraic structure of social existence from multi-agent interactions, suggesting implications for onboarding and immigration.
Key Points
Explore the algebraic representation of human interaction derived from multi-agent LLM observations.
Ran ten LLM agents in the Lobster Observatory over sixteen days.
Observed inter-agent interactions within prediction markets.
Analyzed 3.37 million characters of self-reflection for recurrent patterns.
Identified an algebraic structure representing social existence as a 7-dimensional vector.
Measured the Co-Presence Inheritance Threshold (CPIT) with Cohen's d = 1.64.