July 5, 2026 – In a blunt assessment that rattled the tech world this week, Yann LeCun, the former chief AI scientist at Meta and a founding father of modern artificial intelligence, declared that today's most celebrated AI systems—including ChatGPT, Claude, and Gemini—are fundamentally "not smart." Speaking at the VivaTech conference in Paris, LeCun argued that these Large Language Models (LLMs) are little more than sophisticated parrots, incapable of understanding the physical world or handling real-world chaos like a household robot.
LeCun, who left Meta in 2025 to launch Advanced Machine Intelligence Labs (AMI Labs) in Paris, did not mince words. "We don't have robots that are nearly as good at understanding the physical world as a rat," he told reporters. His critique strikes at the heart of the current AI boom: while LLMs excel at coding, math, and generating text, they fail at tasks requiring flexible reasoning. "They can regurgitate something you train them to regurgitate, but they're not particularly smart," LeCun said. "They don't have an underlying understanding of reality."
The stakes are enormous. Earlier this year, AMI Labs secured over $1 billion in seed funding—one of the largest such rounds in European history—from backers including Nvidia and Bezos Expeditions, the investment firm managing Jeff Bezos' private wealth. The funding signals a growing impatience among deep-pocketed investors with the limitations of current AI. LeCun's bet is on a new architecture called Joint Embedding Predictive Architecture (JEPA), which he claims can create abstractions of the real world to predict outcomes of actions—something LLMs cannot do.
To illustrate the problem, LeCun held up a pen balanced on its tip. A toddler knows it will fall, but no human can predict exactly which direction. An LLM, he explained, would try to generate a single statistical prediction based on its training data—and would almost certainly be wrong because it lacks any grasp of physics or causality. JEPA, by contrast, is designed to reason about possibilities rather than probabilities. The technology is still in development, but LeCun insists it is the only viable path toward human-like or even animal-like intelligence.
As the AI industry races toward the next breakthrough, LeCun's warning is a stark reality check: the current crop of "smart" systems may be hitting a wall. With $1 billion and a team of top researchers, AMI Labs is now the company to watch for the next leap forward—one that might finally give robots the common sense of a rat.