Yann LeCun, the renowned AI scientist who spent a decade as Meta’s chief AI officer, is publicly trashing the technology he helped pioneer. In a blunt interview at the VivaTech conference in Paris this week, LeCun declared that today’s most famous AI systems—including ChatGPT, Claude, and Gemini—are fundamentally “not smart” and incapable of achieving even animal-level intelligence. His verdict comes as his new startup, Advanced Machine Intelligence Labs (AMI Labs), has already secured over $1 billion in seed funding from heavyweights like Nvidia and Jeff Bezos’s investment fund.
“We don’t have robots that are nearly as good at understanding the physical world as a rat,” LeCun said, arguing that Large Language Models (LLMs) are essentially “regurgitation engines” that lack true understanding. He demonstrated the limitation with a simple test: holding a pen upright and asking what happens when released. A toddler instinctively knows the pen will fall, but no human can predict the exact direction. An LLM, trained on statistical patterns, will guess a specific direction—and almost certainly be wrong. “They’re not a path toward human-level intelligence, or even animal-like intelligence, because they cannot deal with real-world data,” LeCun explained.
The critique strikes at the heart of the AI industry’s current hype cycle. Since ChatGPT’s launch in late 2022, tech giants have poured billions into scaling LLMs, claiming they are stepping stones to artificial general intelligence. LeCun, who left Meta in 2025 to launch AMI Labs in Paris, argues this is a dead end. His company is developing a new architecture called Joint Embedding Predictive Architecture (JEPA), which creates abstractions of the physical world—allowing AI to reason about outcomes rather than just generate statistically plausible text. This approach, he says, is essential for robots to handle tasks as mundane as household chores.
The $1 billion seed round—one of the largest in European tech history—signals that investors are betting on a paradigm shift. If LeCun is right, the next wave of AI won’t be about bigger chatbots, but about systems that can navigate the messy, unpredictable chaos of the real world. As of July 5, 2026, the race to build that smarter, more grounded intelligence is just beginning.