LeCun Says ChatGPT-Style AI Is Hopeless for Robotics as AMI Labs Raised $1 Billion
Updated
Updated · Storyboard18 · Jul 4
LeCun Says ChatGPT-Style AI Is Hopeless for Robotics as AMI Labs Raised $1 Billion
1 articles · Updated · Storyboard18 · Jul 4
Summary
Yann LeCun said current models like ChatGPT, Claude and Gemini are "largely hopeless for robotics" because they cannot reason about the physical world well enough to act in it.
Speaking at VivaTech in France, he argued large language models can code and generate text but still understand real-world environments worse than a rat, making simple scale-ups a dead end for human-like intelligence.
AMI Labs is instead building a Joint Embedding Predictive Architecture, or JEPA, which LeCun said forms abstract world models rather than guessing a single likely outcome from text-like patterns.
More than $1 billion in seed funding has backed that approach, with Nvidia and a fund managing Jeff Bezos' private wealth among investors; AMI plans industrial deployments next year after refining the system this year.
The debate reflects a wider robotics push toward "world models," with groups including Google DeepMind, Wayve and Fei-Fei Li's World Labs pursuing systems that can explain cause and effect in physical settings.
Is the tech industry's massive bet on language models a dead end for creating human-like intelligence?
Will a new AI that understands physics, not just words, finally unlock truly intelligent robots?
With $1 billion, can LeCun's anti-LLM approach beat the tech giants at their own AI game?
$1.03 Billion for World Models: How AMI Labs and Yann LeCun Are Shifting AI Beyond Language
Overview
In 2026, Advanced Machine Intelligence (AMI) Labs was launched in Paris by Yann LeCun, marking a major shift in the AI landscape. LeCun, known for his skepticism of large language models (LLMs), founded AMI Labs to challenge the dominance of LLMs and promote 'world models' as a new direction for AI. Backed by over $1 billion in seed funding, AMI Labs aims to build AI systems that understand and interact with the real world, moving beyond text prediction. This bold move signals a strategic pivot in AI development and positions Europe as a key player in the global AI race.