Vitalik Buterin says AI still can't judge what new writing argues

Ethereum co-founder Vitalik Buterin said in an October 29 X post that AI is extremely knowledgeable and can detect references to almost anything but is not wise, so it makes only surface-level judgments about what a piece of writing argues. He said this is especially true for brand-new material that models are reading for the first time. The post followed an X debate over a Wall Street Journal article on Harvard’s Humanities 10 course, with users disagreeing about whether chatbots are useful discussion partners or produce poor essay feedback. In July, Buterin had said AI surpasses humans one skill at a time; that same month, an AI tool identified him as the author of an anonymous Ethereum document.