Vitalik Buterin Says AI Is Knowledgeable but Not Very Wise

Ethereum co-founder Vitalik Buterin said in an X discussion on October 29, 2026, that artificial intelligence can recognize references to nearly anything but makes only surface-level judgments about a text’s central argument. The exchange concerned whether chatbots rival the best human readers and followed a Wall Street Journal report that applications to Harvard’s device-free Humanities 10 course rose 40% this fall, with 180 students seeking 90 seats. Buterin distinguished knowledge from wisdom, saying wisdom may be hard to verify and that a model’s weakness increases when it reads a text for the first time. He has also warned that AI could weaken parts of crypto security within two years and has backed keeping human checks in AI systems.