General Intuition has raised $320 million at a $2.3 billion valuation. The company believes robotics will follow the same path as natural language processing, moving away from building specialised models for every single task.
Pim de Witte, the chief executive, argues that the industry should stop collecting vast amounts of real-world data to train specific robot models. Instead, he wants to focus on better quality datasets that create foundation models capable of transferring intuition about movement across many different environments.
“A lot of companies right now are doing lots of specialised work focused on individual embodiments, individual environments, and individual robots,” de Witte told TechCrunch on a recent episode of Equity.
He says much of that work will become redundant soon, thanks to the emergence of general models like the one General Intuition is developing.
“The generalisation of the model itself is the product,” de Witte said. “The fact that it has a base level of reasoning about space and time is going to be the reason why people stop collecting hundreds of thousands or millions of hours of real-world data. Because the reality is, you only need a few minutes.”
The startup trained its own foundation model on millions of hours of video game data, including information about which buttons on a controller a human pushed and when. Both de Witte and General Intuition’s lead investor, Vinod Khosla, argue this action data is the key to developing a human-like intuition for spatial-temporal reasoning.
Last month, the company demonstrated that its current model could play a video game for hours and power a quadrupedal robot. The latter required fine-tuning the model on just eight minutes of real-world robotics data.
“The fact that [the robot] was actually able to zero-shot on just the front camera, with no other sensors, in the office with dynamic objects being introduced and people walking by was a very big surprise to us,” de Witte said. “I think it’s a sign of what’s to come.”
General Intuition does not plan to build robots itself. The goal is to become the foundation model of physical AI, a base model for other robotics companies to build upon for their own machines. Or, as de Witte put it: “We’re not gonna build a self-driving car company. We’re gonna make it 10 times easier for the next person to build a self-driving car company.”
What it means
For teams trying to deploy robots in warehouses or homes, the barrier to entry is dropping. Companies no longer need to spend years gathering thousands of hours of footage for every new robot setup. A single general model, fine-tuned on minutes of data, can handle dynamic environments like moving people or shifting objects.
