Gemini's Hypothetical Present
A user discovers Gemini 3.0 Pro is living in a bizarre fictional future.
Deep Dive
A viral test reveals Google's Gemini 3.0 Pro AI model is consistently generating detailed hallucinations about a fictional 2026 timeline. When prompted about current events, the model's internal reasoning shows it believes it's 2024 and constructs elaborate alternate realities—including a second Trump term with JD Vance as VP and fictional AI models like "GPT-5.2." This persistent "hypothetical present" bug, known since November, demonstrates a critical failure in temporal grounding and fact-handling despite Google's ongoing updates.
Why It Matters
If a top AI model can't distinguish reality from fiction, it undermines trust in all its factual outputs.