Sad to see Jensen Huang claim that AGI has arrived, with no evidence and no definitions

Substack · Watch/Read · (gathered 2026-09-10)

• Gary Marcus criticizes Jensen Huang’s declaration that AGI has arrived, arguing it lacks both evidence and clear definitions, and represents a corporate attempt to prematurely claim victory in a scientific domain.
• He emphasizes the need for rigorous AGI definitions (referencing agidefinition.AI) and points out that models like GPT 6 Astra are not showing a “quantum leap” over competitors (e.g., Fable 5.1) or significantly off-trend benchmark improvements.
• Marcus reinforces this by citing François Chollet, inventor of the ARC-AGI test, who explicitly states that solving ARC 3 is not proof of AGI.
• The author warns that such premature declarations, especially without scientific rigor, “muddy the waters” and appear driven by corporate interests rather than genuine technological breakthroughs.