General Intuition
📰 ArticleRebecca Bellan
General Intuition eyes a $2B valuation to advance its spatial-temporal AI agent training using massive gaming datasets.
General Intuition is pioneering a unique approach to embodied AI by utilizing interactive, first-person gameplay data to teach machines deep spatial-temporal reasoning. Unlike competitors focusing primarily on generative world models for video synthesis, General Intuition aims to produce highly capable AI agents that can perceive, anticipate, and interact within simulated environments in real time. By tapping into Medal’s ecosystem—which generates 2 billion video clips annually from 10 million active users—the startup creates a specialized training ground for agents to learn complex physical interactions. This proprietary dataset has made the company a target for major AI labs, including reported interest from OpenAI. The company, led by co-founder Pim de Witte and a team of world-modeling experts, plans to use its new funding to significantly expand its compute infrastructure, aiming to launch a new product by late 2026.
💡Highlights
- ├─2B annual gameplay video dataset
- ├─Focus on spatial-temporal reasoning
- └─Embodied AI agent development
🎯For
- ├─AI Researchers
- ├─Robotics Engineers
- └─Venture Capitalists