Can a language model ever understand the physical world?

A post frames a disagreement over whether a language model can ever understand the physical world. It says Yann’s answer is no and quotes Chris: “I think that is wrong and has been proven to be wrong.”