Disagree. In my experience particularly over the last few weeks, Gemini has gotten very good at understanding what i'm asking for and giving a relevant response that 9 times out of 10 just answers my question without me having to go further into the relevant wiki link etc.
Try asking it about a sporting event or score. It literally just makes shit up. I asked it Tuesday if there was a world series game; it told me with complete confidence there wasn't and that Milwaukee was leading the series 2-1 after losing the night before.
A bit off-topic, but: How often do they re-train models on current events? Every month? Every day? Every hour?
Given how AIs work, if they weren't trained with last night's sports news, you can't get an accurate answer this morning. All you can get is a hallucination.
Agreed. I have no doubt I could throw curveballs at it, but it has been getting things mostly correct for a while. And the fact that I can follow up with a direct challenge is kind of nice.
