Comment by brookst

I haven’t found that to be the case. Both LLMs and humans produce outputs that cannot be blindly trusted to be accurate.