Legit question: if users of LLMs are supposed to “verify the results”—and that disclaimer justifies the results being inaccurate—in a world where LLMs eventually taint/replace all the results, how will these users verify the results? Through what service or method is the verification of accuracy expected to occur?