Startups that promise to predict human behaviour are having a moment, with Mirror Particle proposing a fundamentally different approach. Unlike traditional large language models (LLMs), Mirror Particle aims to build a 'world model' from scratch, capturing the changing nature of human behaviour.

Abhivyakti Ahuja, co-founder and CEO, argues that LLMs are limited because they are trained on written language, missing the visual and social aspects that define us. Mirror Particle’s engine will track motivations and triggers, providing deeper insights into consumer decisions.

The company’s strategy focuses on 'revealed behaviour,' using data from clients, current events, and social media to model demographic segments. This approach will help brands not just predict, but understand why consumers make certain choices, making strategic recommendations that go beyond self-reported surveys.

Mirror Particle has already raised significant funding, including an angel round and a $480 million seed round. They are competing in TechCrunch’s prestigious Startup Battlefield, where they will face other innovative startups. The long-term vision is to become the 'general layer for anticipating human behaviour,' moving from population-level analyses to individual insights.