@brent Totally agree.
Imagine you're out there in the marketplace selling a Fact Machine, and this Fact Machine is constantly lying. And when you're challenged, when you're asked: "What good is a Fact Machine that lies all the time?, your response is literally the following:
"Factual accuracy in large language models remains an area of active research.”
That's like saying:
You being able to sit on this chair we just sold you is an area of active research.