This is the current iteration of the Prisoner's Dilemma evolutionary agentic simulation. It now aspires to model the human condition as accurately as, well, humanly possible. Hence the project name, τὰ ἀνθρώπινα.

It uses the client-server approach tested in the Predator-Prey simulation, the server running in python, the client in javascript. Prisoners' Dilemma strategy is now downstream of the "Big Five" personality types, openness, conscientiousness, extraversion, agreeability, neuroticism, which are genes propagated by sexual procreation. Agents have biographies, familial relations and mental states (short-term happiness and longer-term "world-trust"). They have cultural-linguistic and racial-ethnic identities, and genes modifying their ingroup/outgroup behavior based on cultural, genetic and tribal similarity.

The aim here would be to simulate thousands of years of evolution on populations of thousands or tens of thousands of individuals. So there is a tradeoff between simulation complexity and simulation scope. I can currently run this on a laptop in a reasonable timeframe for a couple of centuries and a couple of thousand individuals, but there is potential for optimization.