Artificial intelligence: Yann LeCun works on more flexible AI

July 3, 2026 - 09:35

Yann LeCun, a prominent figure in the field of artificial intelligence, is now leading a startup focused on creating more flexible AI systems. LeCun, who is also the chief AI scientist at Meta, has long argued that current models, including large language models, are too rigid. His new venture aims to address these limitations by developing technology that can adapt and learn in ways closer to human cognition.

The core issue, according to LeCun, is that today's AI excels at narrow tasks but fails when faced with new situations or incomplete information. Most systems rely on massive datasets and pattern matching, which makes them brittle. LeCun's approach involves building models that can understand the world through a more abstract and predictive framework. This would allow the AI to reason, plan, and adjust its behavior without needing to be retrained from scratch for every new problem.

The startup is still in its early stages, but the goal is clear: move beyond the current hype around chatbots and image generators. LeCun believes that true progress will come from systems that can learn efficiently and operate with common sense. While details remain scarce, the project has attracted significant attention from investors and researchers who are eager to see if this new architecture can overcome the obstacles that have stumped the industry for years. The work represents a major bet on a different path forward for artificial intelligence.