Alexandre LeBrun, CEO of AMI Labs, led by Yann LeCun, dismisses the terms 'AGI' and 'superintelligence', focusing on practical applications of world models in AI.
LeBrun's Stance on AI Terminology
In an interview with TechCrunch, LeBrun stated, “We never used the word AGI. And I just noticed that nobody is using it anymore; they switched to superintelligence.” He questioned the usefulness of these terms, saying, “There’s no good definition. What is superintelligence? I don’t know. It’s not a very useful word.”
The Role of World Models
LeBrun spoke at The International Conference on Machine Learning in Seoul, seeking industrial partners for AMI Labs. He emphasized the importance of world models in robotics, noting that current robots run on fixed routines and are “completely static.” He added that context-aware AI could prevent mishaps, such as a robot inadvertently kicking a child at a public event. “The hardware is very advanced,” LeBrun said, “but there’s no brain.”
World models are designed to predict the next state of the world, akin to human intuition, while large language models (LLMs) predict the next word or text. LeBrun clarified that world models are complementary to LLMs, not replacements, providing context to AI systems.
Applications and Challenges
LeBrun highlighted the potential of world models in various industries, particularly those interacting with the physical world. He pointed out that while factory robots perform repetitive tasks efficiently, challenges arise in open environments like homes or streets where safety is a concern. “Robots are not safe right now,” he noted.
Healthcare is another area LeBrun is familiar with, having founded AI health startup Nabla. He compared current AI systems to a doctor without practical experience, emphasizing that real-world application is crucial for AI development.
Focus on Asia
LeBrun sees Asia, particularly South Korea, as key to AMI Labs' strategy. Korea’s advanced industries in robotics and semiconductors and its rapid adoption of AI technology make it an attractive partner. “Korea was the fastest adopter of the internet 25 years ago,” LeBrun said, emphasizing the unique combination of industrial capability and willingness to embrace AI.
JP Lee, CEO of SBVA and an investor in AMI, echoed this sentiment, highlighting Korea’s commitment to AI funding and innovation. Seoul's plan to invest $880 billion in AI and related technologies underscores this focus.
Despite raising $1.03 billion in March at a $3.5 billion valuation, AMI Labs remains pre-product, with LeBrun stating, “We’ll make a surprise when we’re ready.”
Source: https://techcrunch.com/2026/07/16/why-ami-labs-alexandre-lebrun-wont-call-his-ai-agi-or-superintelligence/
