A new AI startup called General Intuition is joining a growing wave of companies developing world models — AI systems that can understand, simulate, and interact with the real world. These emerging models could redefine how artificial intelligence learns, reasons, and acts beyond text or images.
The Rise Of General Intuition
Around mid-2024, Pim de Witte, founder of the video game clipping platform Medal, started reaching out to major AI labs. His goal was simple: to explore whether Medal’s vast gaming footage could help train advanced AI agents.
To his surprise, several labs immediately showed interest. “We received multiple acquisition offers very quickly,” he said, hinting that OpenAI was among them with a rumored $500 million bid. Initially intrigued, de Witte soon realized the true value of what he had — a massive dataset capable of teaching AI how to navigate complex, real-world-like environments.
From Gaming Clips To AI Innovation
The breakthrough came after de Witte read a Google DeepMind paper showing how gaming data could train agents to move through 3D worlds. Medal’s platform, which hosts around 2 billion uploads annually from thousands of video games, suddenly became more than a gaming hub — it was a potential foundation for a world model AI.
De Witte understood that his data could serve as the building blocks for teaching AI how to act intuitively in real-world scenarios, bridging the gap between virtual environments and human-like understanding.
Why World Models Are The Next Big Thing In AI
Unlike traditional AI systems that rely solely on static datasets, world models enable machines to predict outcomes and make decisions based on real-world context. This leap allows AI to “think” spatially, adapting to environments much like humans do.
These models could power more autonomous agents, from self-driving cars to digital assistants that can reason through complex, dynamic situations. Instead of simply generating text or images, world models give AI a mental map of how the world works.
Backed By Major Investors
To bring this vision to life, General Intuition announced a massive $133.7 million seed round, led by Vinod Khosla, the founder of Khosla Ventures and an early investor in OpenAI. The funding marks one of the largest seed rounds ever for an AI startup — underscoring investor confidence in the future of world model development.
From Data To Intuition
What makes world models so transformative is their ability to blend data-driven learning with intuitive reasoning. They represent a step toward AI that can understand cause and effect, not just recognize patterns. For startups like General Intuition, this means creating systems that can eventually learn and adapt the way humans do — through experience.
As the AI race intensifies, why world models are the next big thing in AI becomes clear: they promise a more grounded, intelligent, and adaptable generation of machines ready to operate in the same unpredictable world we live in.
Comment