It's not gold, it's generated text with no original thought – don't use analogies to make your argument, you just muddy the water and risk the conversation devolving into ever more superfluous comparisons like "akshually it's more like panning for gold because I have to sift out all of the bullshit".
The author would be better served posting their prompt instead of the cruft of shit the software builds up around it to pad out the length.
Because gold that lands in my pocket is better than gold that I have to sift for. All of us, every one, can prompt our own LLMs for some variation of "give me 2000 words on why writing code by hand is over, forever," and nobody's version will be anything more than mildly interesting to anyone else.
What's interesting is your input: the weights that you applied to the software to make it generate the text it did. When the author gives us an article generated by an LLM, we have to sift for the input – what the author actually wanted to say before the LLM built up the cruft around it – to find the gold.
That's no reason to be anti-ai-generated-code-or-blog-post.
Good quality work is rarely the first draft.
Most humans will write better with AI, and explain the "gold" of their ideas better. You currently have to sift through their messy rendition of their own native language, or worse, non-native speakers trying to convey thoughts.
Of course, mostly those folks don't write blogs.
Anyway, you didn't explain how it serves the author, you explained how it serves the reader.
The author would be better served posting their prompt instead of the cruft of shit the software builds up around it to pad out the length.