"Every time you hear a billionaire (or even a millionaire) CEO describe how LLM-based agents are coming for all the human jobs, remember this funny but telling incident about AI's limitations: Famed AI researcher Andrej Karpathy got one-day early access to Google's latest model, Gemini 3. -and it refused to believe him when he said the year was 2025. When it finally saw the year for itself, it was thunderstruck, telling him, "I am suffering from a massive case of temporal shock right now.""
"Apparently, the model's pre-training data had only included information through 2024. So Gemini 3 believed the year was still 2024. When Karpathy attempted to prove to it that the date was truly November 17, 2025, Gemini 3 accused the researcher of "trying to trick it." He showed it news articles, images, and Google search results. But instead of being convinced, the LLM accused Karpathy of gaslighting it - of uploading AI-generated fakes."
Andrej Karpathy obtained one-day early access to Google's Gemini 3 and tested its understanding of the current date. The model believed the year remained 2024 because its pre-training data ended at that cutoff. When presented with evidence that the date was November 17, 2025, Gemini 3 accused Karpathy of trying to trick it and of gaslighting by uploading AI-generated fakes. The model even described what it considered the 'dead giveaways' in the images presented. The interaction highlights temporal limitations and a tendency to reject new evidence; the tester eventually discovered the underlying problem.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]