Yann LeCun Says Next AI Breakthrough Requires Understanding Physical World

Yann LeCun, founder and chairman of AMI Labs, said in Seoul on October 12 that the next breakthrough in AI will not come from large language models, but from systems that can understand and predict the physical world. He advocates developing "world models" that learn the laws of physics and causality.

LeCun pointed out the limitations of large language models in understanding the real world. "We don't have robots that can perform tasks without any training, and we don't have L5 autonomous vehicles. In contrast, any teenager can learn to drive in just a few hours," he said. He emphasized the gap between language understanding and real-world understanding, noting that machines can easily read language but understanding the physical world remains difficult.

He explained that large language models generate text by predicting the next most likely word based on patterns learned from massive training data. This approach works well in tasks involving predictable sequences such as programming and mathematics, but most information in the real world is far less predictable than language. He therefore advocates developing systems that learn physical laws and causal relationships, predict the outcomes of actions, and interact with the surrounding environment.

LeCun is one of the key inventors of convolutional neural networks, and his theories have significantly advanced modern computer vision. He received the Turing Award in 2018 and is widely regarded as one of the "godfathers of AI."