Mirror Particle builds world model predicting human behavior

Mirror Particle, a two-year-old San Francisco startup, is building a foundation model it calls a world model to predict consumer behavior and the motivations behind it. Unlike rivals that fine-tune large language models, Mirror Particle uses proprietary data combining client customer data, social media, pop culture, and current events to simulate how human behavior evolves over time. In one early pilot, the startup told a pet food brand its packaging imagery was irrelevant and that a mass-market perception problem was hurting sales. The company has raised an angel round and is close to closing its first venture round.