Yann LeCun LLMs are a dead end?

Yann LeCun just announced a billion dollar seed round to prove that LLMs are a dead end and that GPT style AI will never be truly intelligent. LeCun is one of three winners of the Turing Award, which is basically the Nobel Prize for computing. He spent 12 years leading Meta's AI research lab until Zuck turned that upside down last year. But he's consistently been one of the biggest critics in the industry, saying that large language models are just sophisticated pattern matchers. Even though his research laid the groundwork for LLMs, Dr. LeCun argued that they were not the final answer to AI development. The problem with current systems, he said, is that they do not plan ahead. Trained solely on digital data, they do not have a way of understanding difficulties in the real world. "LLMs are not a path to superintelligence or even human-level intelligence. I have said that from the beginning," he said. "The entire industry has been LLM-pilled." So four months after quitting, he's just closed Europe's largest seed round ever at a $3.5 billion dollar valuation. The company is AMI = Advanced Machine Intelligence Labs, like the French word for friend. And they're building what LeCun calls world models. The core idea is that instead of LLMs that just learn from language and text and can regurgitate it, AMI wants to build systems that learn from reality and observe how the world works by using video spatial data or real world interactions through a framework called JEPA. LeCun had spent years developing while he was at Meta. They explicitly don't want to just launch another chatbot, and they want to go after industries with complex physical systems, like cars, aerospace and pharmaceuticals. He also said that the company has had discussions with Meta about possible partnerships to use AMI's technology in its wearable devices. "There could be some products that get in the hands of consumers fairly early, depending on how the technology progresses," he said. LeCun says AMI aims to work with companies in manufacturing, biomedical, robotics, and other industries that have lots of data. For example, he says AMI could build a realistic world model of an aircraft engine and work with the manufacturer to help them optimize for efficiency, minimize emissions, or ensure reliability. Or even help catch failures beforehand. Right now the company doesn't have a product or revenue, and they only have a 12 person team. But their investors, which include Nvidia, Toyota, Samsung, Jeff Bezos, Eric Schmidt, Marc Cuban, are all really thinking this as more like a long-term scientific bet on research, as opposed to a startup that ships a monetizable product in the next six months.