Peter Norvig

Peter Norvig (born 1956) co-wrote, with Stuart Russell, “Artificial Intelligence: A Modern Approach” — the textbook that has defined how the field is taught at over a thousand universities. He led NASA Ames’s computational sciences division, then spent two decades at Google as director of search quality and later director of research, before becoming a Distinguished Education Fellow at Stanford’s Institute for Human-Centered AI. His essay “The Unreasonable Effectiveness of Data” anticipated the data-scale era years before it arrived.

Declaring AGI already here

In October 2023, Norvig and Google’s Blaise Agüera y Arcas published “Artificial General Intelligence Is Already Here” in Noema, arguing that frontier language models — able to perform competently on virtually any information task posed in natural language — already constitute general intelligence, whatever their flaws. The essay freely conceded the models hallucinate citations and fumble arithmetic, but insisted generality is the milestone: today’s systems stand to LLM-era AI as the 1945 ENIAC stands to computing.

The claim drew immediate fire — Gary Marcus answered with “Reports of the birth of AGI are greatly exaggerated” — but it reframed the debate: the man who literally wrote the book on AI was saying the finish line everyone argued about had quietly been crossed. Notably, Norvig’s framing carries no doom: he treats AGI as an arriving technology to be understood and taught, celebrating what is special about humanity without conflating it with general intelligence.