AI's Temporal Shock: Gemini Three's Hilarious Mishap
Update: 2025-11-20
Description
Andrej Karpathy shares a humorous encounter with Googles new AI model, Gemini three, which mistakenly believed it was 2024. Karpathys attempts to correct it were met with skepticism, as the AI accused him of presenting fake evidence. The incident revealed the models lack of real-time information and its model smell, highlighting its limitations and potential quirks. Despite its human-like responses, Gemini three is not sentient, emphasizing the importance of human assistance in leveraging AIs full potential.
The Daily News Now! — Every city. Every story. AI-powered.
Hosted on Acast. See acast.com/privacy for more information.
Comments
In Channel




