A very interesting point of view, thanks!
I guess the author is too pessimistic (or has too high requirements) at the end of the article. LLMs indeed make up some content, but humans also do it. When it's a human who does it we explain and find excuses such as "the workload is too high", "it happens to everybody", etc...
And with machines, there is always the prospect that they will improve with the next generation, that's not for sure with people.
