Cracking the human simulation code: Aaru co-founders on refining the science of prediction

Summary

• Aaru is a simulation startup that uses AI to predict human behavior with higher accuracy and speed than traditional methods like surveys and prediction markets.
• Founded about two years ago by teenagers, now valued at $1 billion with over 100 customers including Ernst & Young, Accenture, Bayer, A24, Boston Beer Company, and Spindrift.
• Contrarian prediction on alcohol consumption trends: GLP-1 drugs initially reduce alcohol use, but over 2-5 years, net drinking occasions may increase due to enhanced social confidence, while alcohol per occasion decreases, with a shift from beers/seltzers to spirits.
• Traditional prediction methods are criticized for sampling bias, especially in sensitive areas like alcohol behavior where people are unwilling to disclose usage, limiting their effectiveness.
• Business model involves proving technology by simulating past client decisions without using their data, predicting outcomes more accurately to build trust and drive adoption.
• Company has 24 employees and is growing quickly; technology applications range from ad testing, packaging, pricing, distribution to policy reception.
• Long-term vision is to not only predict but also shape human behavior for societal good, with pro bono work impacting hundreds of millions.
• Key challenge: Initial sell is difficult because claiming to predict human behavior better than clients themselves is hard to believe.
• Market implication: Potential disruption in market research, advertising, product development, and public policy industries through more accurate behavioral predictions.
• Uncertainty: Predictions are based on simulation models and require long-term validation; effects like GLP-1 drug impacts may evolve over years.