Medal, a platform for uploading and sharing video game clips, has created a new artificial intelligence research lab that uses its gaming videos to train and build basic models and artificial intelligence agents that can understand how objects and entities move in space and time – a concept known as spatiotemporal logic.
Called General Intuition, the startup is betting that Medal’s dataset — 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’re essentially transferring your perception, usually through the first-person view of the camera, to different environments,” Pim de Witte, 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 edge cases for training. “You have this selection bias towards exactly what kind of data you really want to use for educational work.”
That data moat is what reportedly caught the attention of OpenAI, which late last year tried to acquire Medal for $500 million, per The Information. (Neither OpenAI nor General Intuition would comment on the report.)
It’s also what led General Intuition to raise a whopping $133.7 million in funding, led by Khosla Ventures and General Catalyst with Raine participating.
The startup plans to use the funds to grow its team of researchers and engineers focused on training a general agent that can interact with the world around it, targeting initial applications in games and search and rescue drones.
De Witte says the founding team has already made strides: General Intuition’s model can understand environments it wasn’t trained in and correctly predict actions within them. It is able to do this purely through visual input. Agents see only what a human player would see and move through space following controller inputs. This approach, the company says, can be naturally transferred to physical systems like robotic arms, drones and autonomous vehicles, which are often operated by humans using video game controllers.
Techcrunch event
San Francisco
|
27-29 October 2025
General Intuition’s next milestone is twofold: the creation of new simulated worlds to train other agents and autonomous navigation in completely unknown physical environments.
This technical approach shapes how the company plans to commercialize its technology and sets it apart from competitors who build global models.
While General Intuition also builds global models on which to train its agents, such models are not the product. Unlike other world model makers such as DeepMind and World Labs, who sell their Genie world models and Marblerespectively, for agent training and content creation, General Intuition focuses on other use cases to avoid copyright issues.
“Our goal is not to produce models that compete with game developers,” de Witte said.
Instead, the startup’s game applications focus on creating player-less bots and characters that can outperform traditional “deterministic bots,” or pre-programmed characters that produce the same output every time.
“[The bots] it can scale to any level of difficulty,” Moritz Baier-Lentz, founding member of General Intuition and partner at Lightspeed Ventures, told TechCrunch. “It’s not imperative to build a God bot that beats everyone, but if you can scale and fill fluidity for any player situation so that their win rate is always and maxes out at 50% again.
De Witte also has a background in humanitarian work, which informs the startup’s focus on powering search and rescue drones, which sometimes must navigate unfamiliar environments and extract information without GPS.
Ultimately, de Witte and Baier-Lentz see the core functionality of General Intuition—spatiotemporal reasoning—as a critical piece in the race toward artificial general intelligence (AGI). While the big AI labs are focused on building increasingly powerful large language models, General Intuition believes that true AGI requires something that LLMs essentially lack.
“As humans, we create text to describe what’s going on in our world, but in doing so, you lose a lot of information,” said de Witte. “You lose the general intuition around spatiotemporal reasoning.”
