A Few Thoughts on Yann LeCun's Departure from Meta

A Few Thoughts on Yann LeCun's Departure from Meta You have probably heard that my friend Yann LeCun, one of the godfathers of AI and Chief AI Scientist at Meta, announced that he is leaving the company after 12 years and will only work until the end of December. As many of you have heard through rumors or recent media articles, I am planning to leave Meta after 12 years: 5 years as founding director of FAIR and 7 years as Chief AI Scientist. The impact of FAIR on the company, on the field of AI, on the tech community, and on the wider world has been spectacular. The creation of FAIR is my proudest non-technical accomplishment. I am creating a startup company to continue the Advanced Machine Intelligence research program (AMI) I have been pursuing over the last several years with colleagues at FAIR, at NYU, and beyond. The goal of the startup is to bring about the next big revolution in AI: systems that understand the physical world, have persistent memory, can reason, and can plan complex action sequences. I am extremely grateful to Mark Zuckerberg, Andrew Bosworth, Chris Cox, and Mike Schroepfer for their support of FAIR, and for their support of the AMI program over the last few years. Because of their continued interest and support, Meta will be a partner of the new company. As I envision it, AMI will have far-ranging applications in many sectors of the economy, some of which overlap with Meta’s commercial interests, but many of which do not. Pursuing the goal of AMI in an independent entity is a way to maximize its broad impact. I will give some more details about the new company when the time comes. In the meantime, I’m sticking around Meta until the end of the year. Officially, he is leaving to found his startup, where he will build World Models and continue his long-standing crusade against the current trend towards LLMs. In Yann's opinion, true intelligence is not just predicting the next word. Models need a real connection to the physical world; AI systems must understand the physical world, have persistent memory, be able to reason, and plan complex action sequences. This is his concept of World Models. A simple example: if we had a 'world model', we could teach systems to confidently predict what will happen next in the physical world – for example, a robotic arm could predict what would happen if a cup were knocked off a table. Interestingly, Fei-Fei Li is also trying to build something similar in her startup - but so far it is at a very early stage and nothing intelligent, other than generating 3D spaces, exists there. In recent months, there has been a clear conflict of visions at Meta. On one side is Mark, who is excited about the idea of AGI and demands quick product results from teams in the race against OpenAI. On the other side is Yann, who criticizes the current LLMs and believes that true AGI is still very far away. Against this backdrop, Zuckerberg has acquired product-oriented leader Alexander Wang.