I don’t think you read the article, which documents many cases of him being wrong

For example saying in 2024 that LLMs had peaked. That’s not early, that is already, definitively wrong.