Medal, a platform for uploading and sharing video game clips, has created a new leading AI research lab that is using its repository of gaming videos to train and create foundation models and AI agents that can understand how objects and entities move through space and time – a concept known as spatio-temporal reasoning.
A startup called General Intuition is betting that Medal’s dataset — which includes 2 billion videos per year from 10 million monthly active users across thousands of games — surpasses alternatives like Twitch or YouTube for training agents.
“When you play video games, you essentially transfer 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 said that gamers who upload clips tend to post very negative or positive examples, which serve as really useful edge cases for training. “You get this selection bias towards exactly the type of data you actually want to use for the training task.”
This data problem has reportedly attracted the attention of OpenAI, which attempted to acquire Medal for $500 million late last year. Information(Neither OpenAI nor General Intuition would comment on the report.)
This led General Intuition to raise a whopping $133.7 million in seed funding led by General Catalyst with participation from Khosla Ventures and Rain.
The startup intends to use the funding to grow its team of researchers and engineers, focused on training a general 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 progress: General Intuition’s model can understand environments it was not trained on and correctly predict actions within them. It is able to do this through purely visual input; Agents see only what a human player sees, and they move around in space by following controller inputs. The company says this approach can naturally transfer to physical systems like robotic arms, drones and autonomous vehicles, which are often manipulated by humans using video game controllers.
techcrunch event
san francisco
,
October 27-29, 2025
General Intuition’s next milestone is twofold: creating new simulated worlds to train other agents and autonomously navigate completely unfamiliar physical environments.
That technological approach is shaping how the company plans to commercialize its technology and differentiate it from competitors who create world models.
While General Intuition is also creating world models to train its agents, such models are not products. Unlike other world model makers like DeepMind and World Labs, who are selling their world model genies and marbleRespectively, for training agents and content creation, General Intuition is focusing on other use cases to avoid copyright issues.
“Our goal is not to create models that compete with game developers,” De Witte said.
Instead, the startup’s gaming applications focus on creating bots and non-player characters that can outperform traditional “deterministic bots” or preprogrammed characters that produce the same output every time.
“(Bots) can scale to any level of difficulty,” Moritz Baer-Lentz, a founding member of General Intuition and partner at Lightspeed Ventures, told TechCrunch. “It’s not mandatory to create a god bot that beats everyone, but if you can slowly scale and fill in the fluidity for any given player position so that their win rate is always around 50%, it will maximize their participation and retention.”
De Witte also has a background in humanitarian operations, which informs the startup’s focus on empowering search-and-rescue drones, which sometimes have to navigate unfamiliar environments and extract information without GPS.
Ultimately, De Witte and Baeyer-Lentz see the core functionality of general intuition – spatial-temporal reasoning – as an important part of the race to artificial general intelligence (AGI). While leading AI labs focus on building more powerful large language models, General Intuition believes that true AGI requires a fundamental reduction in LLM.
De Witte said, “As humans, we create text to describe what’s going on in our world, but in doing so, you lose a lot of information.” “You lose the usual intuition around spatial-temporal reasoning.”
