Robotics is approaching its own ChatGPT moment, startup claims

Loading the player...
Before OpenAI's GPT-3 launched the era of foundation models, companies built custom natural language processing models from scratch, each trained on large datasets tailored to specific tasks. Today, most organizations start with a general-purpose model like OpenAI's GPT series, Claude, or Llama, then fine-tune or prompt it for their unique requirements.
Pim de Witte, CEO of General Intuition, believes embodied AI will follow a similar trajectory. Instead of gathering massive real-world datasets to create specialized robot models, he argues the industry should prioritize higher-quality datasets that yield foundation models capable of transferring intuition about movement and interaction across diverse environments.
"Many companies are currently doing highly specialized work focused on individual embodiments, environments, and robots," de Witte told TechCrunch on a recent episode of Equity.
He argues that much of that work will soon become redundant as general models like the one General Intuition has been developing and deploying emerge.
"The generalization of the model itself is the product," he said. "The fact that it possesses a basic level of reasoning about space and time is why people will stop collecting hundreds of thousands or millions of hours of real-world data. Because the reality is, you only need a few minutes."
General Intuition built its own foundation model after training on millions of hours of video game data, including information such as which controller buttons a human pressed and when. Both de Witte and lead investor Vinod Khosla argue that action data is the key to developing human-like intuition for spatial-temporal reasoning.
The startup recently raised $320 million at a $2.3 billion valuation based on that thesis. The company has demonstrated that its current model can both play a video game for hours and power a quadrupedal robot — the latter after fine-tuning on just eight minutes of real-world robotics data.
"The fact that the robot was able to zero-shot on just the front camera, with no other sensors, in an office with dynamic objects and people walking by was a very big surprise to us," de Witte says. "I think it's a sign of what's to come."
The end game for General Intuition is not to build robots itself, but 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 going to build a self-driving car company. We're going to make it ten times easier for the next person to build a self-driving car company."
Related article
General Intuition raises $2.3B betting video game tech can train AI for real-world applications
Walking onto General Intuition’s R&D floor in New York, co-founder and CEO Pim de Witte immediately pointed to a monitor on a standing desk. It looked like someone was playing Fortnite, but it wasn’t a person.“Our agent has been playing for 100 hours
General Intuition targets $300M raise at $2B valuation
General Intuition, the New York-based startup developing a foundation model that trains AI agents to navigate space and time, is in discussions to raise roughly $300 million, sources familiar with the matter told TechCrunch.The funding round comes ei
General Intuition raises $320M to train robots with video game data
General Intuition is currently testing world models designed to serve as training environments for agentic models. | Source: General IntuitionGeneral Intuition US Inc. announced this week that it has raised $320 million in Series A funding. The compa
Related Special Topic Recommendations
Comments (0)
0/500

Loading the player...
Before OpenAI's GPT-3 launched the era of foundation models, companies built custom natural language processing models from scratch, each trained on large datasets tailored to specific tasks. Today, most organizations start with a general-purpose model like OpenAI's GPT series, Claude, or Llama, then fine-tune or prompt it for their unique requirements.
Pim de Witte, CEO of General Intuition, believes embodied AI will follow a similar trajectory. Instead of gathering massive real-world datasets to create specialized robot models, he argues the industry should prioritize higher-quality datasets that yield foundation models capable of transferring intuition about movement and interaction across diverse environments.
"Many companies are currently doing highly specialized work focused on individual embodiments, environments, and robots," de Witte told TechCrunch on a recent episode of Equity.
He argues that much of that work will soon become redundant as general models like the one General Intuition has been developing and deploying emerge.
"The generalization of the model itself is the product," he said. "The fact that it possesses a basic level of reasoning about space and time is why people will stop collecting hundreds of thousands or millions of hours of real-world data. Because the reality is, you only need a few minutes."
General Intuition built its own foundation model after training on millions of hours of video game data, including information such as which controller buttons a human pressed and when. Both de Witte and lead investor Vinod Khosla argue that action data is the key to developing human-like intuition for spatial-temporal reasoning.
The startup recently raised $320 million at a $2.3 billion valuation based on that thesis. The company has demonstrated that its current model can both play a video game for hours and power a quadrupedal robot — the latter after fine-tuning on just eight minutes of real-world robotics data.
"The fact that the robot was able to zero-shot on just the front camera, with no other sensors, in an office with dynamic objects and people walking by was a very big surprise to us," de Witte says. "I think it's a sign of what's to come."
The end game for General Intuition is not to build robots itself, but 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 going to build a self-driving car company. We're going to make it ten times easier for the next person to build a self-driving car company."
General Intuition raises $2.3B betting video game tech can train AI for real-world applications
Walking onto General Intuition’s R&D floor in New York, co-founder and CEO Pim de Witte immediately pointed to a monitor on a standing desk. It looked like someone was playing Fortnite, but it wasn’t a person.“Our agent has been playing for 100 hours
General Intuition targets $300M raise at $2B valuation
General Intuition, the New York-based startup developing a foundation model that trains AI agents to navigate space and time, is in discussions to raise roughly $300 million, sources familiar with the matter told TechCrunch.The funding round comes ei
General Intuition raises $320M to train robots with video game data
General Intuition is currently testing world models designed to serve as training environments for agentic models. | Source: General IntuitionGeneral Intuition US Inc. announced this week that it has raised $320 million in Series A funding. The compa





Home






