J. Doyne Farmer

Timeline

1. 1973–1981

   A roulette computer at Santa Cruz

   He graduates from Stanford with a BS in physics in 1973 and starts graduate school at the University of California, Santa Cruz, studying physical cosmology under George R. Blumenthal. With his childhood friend Norman Packard, he forms Eudaemonic Enterprises, and in the 1970s he builds the first wearable digital computer, which predicts the game of roulette. He then switches his dissertation to chaotic dynamics, co-founds the Dynamical Systems Collective, and completes his PhD thesis in 1981.

2. 1981–1991

   The Complex Systems Group at Los Alamos

   After his 1981 doctorate, he takes a postdoctoral post at the Center for Nonlinear Studies at Los Alamos National Laboratory and receives an Oppenheimer Fellowship in 1983. In 1988, he founds the Complex Systems Group in the Theoretical Division and recruits postdoctoral fellows including Chris Langton, Stephanie Forrest and Seth Lloyd. With Richard Bagley, he simulates an autocatalytic set of polymers that evolves in a manner resembling living systems.

3. 1986–1987

   Immune networks and machine learning

   In 1986, he publishes "The Immune System, Adaptation, and Machine Learning" in Physica D with Norman Packard and Alan Perelson, a dynamical model of the immune system based on Jerne's network hypothesis that the authors compare with Holland's classifier system for learning and artificial intelligence. He first becomes involved with the Santa Fe Institute in 1986. In 1987, he and John Sidorowich publish "Predicting Chaotic Time Series", which learns the nonlinear mapping behind a time series to make short-term forecasts.

4. 1991–2006

   Co-founding Prediction Company

   In 1991, he leaves Los Alamos and co-founds Prediction Company with Norman Packard and James McGill, building an early form of statistical arbitrage from US stock market signals. Its trading becomes fully automated from 1996, and UBS buys the firm in 2006.

5. 1999–2012

   Market ecology at Santa Fe

   He leaves Prediction Company in 1999 for the Santa Fe Institute, where he works on economics and complex systems and develops a theory of market ecology. In that theory, financial firms with specialized strategies sort into groups analogous to species in biology.

6. 2012–2024

   Complexity economics at Oxford

   He arrives at Oxford in autumn 2012 from the Santa Fe Institute to co-direct the complexity economics programme at INET. In March 2020, his group predicts a 21.5 percent contraction of UK GDP for the second quarter of 2020 under COVID lockdown, close to the actual 22.1 percent. Allen Lane publishes his book Making Sense of Chaos: A Better Economics for a Better World in April 2024.

7. 2025

   Agent-based models with AI

   In 2025, he gives a keynote at ODSC East on combining agent-based modeling with AI, such as reinforcement learning for agents that adapt their strategies over time. He is a co-investigator with Michael Wooldridge and Ani Calinescu on an Oxford computer science project on large-scale agent-based modelling, which includes populating models with realistic agent behaviours using machine learning.

8. As of October 2026

   Chief Scientist at Macrocosm

   He is Director of the Complexity Economics programme at INET and Baillie Gifford Professor of Complex Systems Science at Oxford's Smith School of Enterprise and the Environment. He is also an External Professor at the Santa Fe Institute and Chief Scientist at Macrocosm, an Oxford spin-out that applies complexity economics to climate change and the green energy transition. In February 2026, his team is developing a model of the global energy sector that represents each of its 30,000 companies as a separate digital agent.