General Intuition gets $134 million to teach customers spatial thinking using video game clips

Medal, a platform for uploading and sharing video game clips, has created a new AI research lab that uses its collection of gaming videos to train and build basic models and AI agents that can understand how objects and entities move across space and time — a concept known as spatio-temporal inference.

The startup is called General Intuition, and it’s betting that Medal’s data set — which consists of 2 billion videos a year from 10 million monthly active users across tens of thousands of games — outperforms alternatives like Twitch or YouTube for training agents.

“When you play video games, you basically transfer your perception, usually through a first-person camera view, to different environments,” Pim de Wit, CEO of Medal and General Intuition, told TechCrunch. He noted that players who upload clips tend to post very negative or positive examples, which serve as really useful instances for training. “You get this selection bias toward the type of data that you actually want to use in the training work.”

That data moat is what caught the attention of OpenAI, which late last year tried to get the medal for $500 million, per $. Information. (Neither OpenAI nor General Intuition commented on the report.)

This is also what led to General Intuition raising a whopping $133.7 million in seed funding, led by Khosla Ventures and General Catalyst with participation from Raine.

The startup intends to use the funds to grow its team of researchers and engineers focused on training a public agent that can interact with the world around it, with the goal of initial applications in gaming, and search-and-rescue drones.

De Witte says the founding team has already made great strides: The General Intuition model can understand environments for which it was not trained and correctly predict actions within them. It is able to do this through visual input only; Agents see only what a human player would see, and move through space following controller inputs. The company says this approach could naturally transfer to physical systems such as robotic weapons, drones, and autonomous vehicles, which are often manipulated by humans using video game consoles.

General Intuition’s next milestone is twofold: creating new simulated worlds to train other agents and independently navigate completely unfamiliar physical environments.

This technical approach shapes how the company plans to market its technology and sets it apart from competitors who are building global models.

While General Intuition also builds global models to train its agents, such models are not the product. Unlike other global model makers like DeepMind and World Labs, who sell their global models The genie and marbleRespectively, for agent training and content creation, General Intuition is focusing on other use cases to avoid copyright issues.

“Our goal is not to produce models that compete with game developers,” De Wit said.

Instead, the startup’s gaming apps focus on creating bots and non-player characters that can override traditional “deterministic bots,” or pre-programmed characters that produce the same output every time.

“Robots can reach any level of difficulty,” Moritz Bayer-Lentz, founding member of General Intuition and partner at Lightspeed Ventures, told TechCrunch. “You don’t have to create a bot that beats everyone, but if you can gradually scale and fill liquidity for any player’s position so that their win rate is always around 50%, it will increase their engagement and retention.”

De Witte also has a background in humanitarian work, which explains the startup’s focus on operating search and rescue drones, which sometimes have to navigate unfamiliar environments and extract information without GPS.

Ultimately, de Witte and Baer-Lentz argue that the core function of general intuition—spatial and temporal reasoning—is a crucial part of the race toward artificial general intelligence (AGI). While major AI labs focus on building larger, more powerful language models, General Intuition believes that true general AI requires something that MBAs fundamentally lack.

“As humans, we create text to describe what is happening in our world, but by doing that, you lose a lot of information,” De Wit said. “You lose general intuition about spatial and temporal reasoning.”