When it comes to achieving artificial general intelligence (AGI), large language models just don’t have what it takes. Models like ChatGPT and Claude are great at processing and generating text, but they’re significantly less skilled at understanding how things actually move through space and time. This spatial and temporal awareness, however, is an essential skill for producing intelligence that can generalize across real-world scenarios. That critical gap, it turns out, might be filled by an unlikely source: your gaming data.
That is the bold bet behind General Intuition, a startup recently backed by Jeff Bezos, which posits that the path to AGI requires more than just feeding algorithms endless amounts of text. While large language models have dominated the AI conversation over the past few years, their fundamental architecture limits them to a two-dimensional understanding of the world. They can describe a car crash vividly, but they cannot intuitively grasp the physics, velocity, or spatial relationships involved in the collision.
Video games, on the other hand, are rich simulations governed by strict physical rules. Every time a player navigates a complex 3D environment, dodges an obstacle, or manipulates an object, they generate highly structured data about spatial reasoning, cause and effect, and temporal progression. General Intuition aims to harness this vast, largely untapped reservoir of human-computer interaction to train a new breed of AI models.
By analyzing how players interact with dynamic virtual worlds, the startup hopes to imbue AI systems with a foundational understanding of physics and spatial awareness. This approach mirrors the way human infants learn about the world not through language, but through physical interaction and observation. If successful, General Intuition could bridge the gap between narrow, text-bound AI and true AGI capable of navigating and understanding the physical world.
The involvement of high-profile investors like Bezos signals growing industry consensus that the current paradigm of scaling up text-based models may soon hit a wall. As the race toward AGI accelerates, the key to unlocking the next leap in machine intelligence might not be found in libraries or Wikipedia, but in the countless virtual worlds where billions of gamers already spend their time. General Intuition's approach suggests that the future of AI is not just about reading, but about doing.