I think most HN people know Google is essentially a dumb search engine.
We're not surprised when image results have no correlation with the search term. The problem is when other people fail to notice this disparity.
They'll go on to write articles and blog post based on google's results. Later someone else will make a wikipedia page based on the blog post. Google's flawed results will now be confirmed by the wikipedia page.
Tangentially related, I had a surprising experience using DDG yesterday. It was... “smart” for lack of a better word.
I was watching an episode of Star Trek TNG and a character was introduced that was a little over the top and vaguely resembled Jim Carrey so I searched “Star Trek Jim Carrey” and one of the top results (after a bunch of YouTube clips of some sort of Jim Carrey Star Trek skit) was the Wikipedia page for the specific episode I was watching. And here’s the interesting part: Jim Carrey wasn’t in it.
My search results took my misguided query and still returned the one correct episode out of who knows how many hundreds of Star Trek episodes exist. I can only guess that enough people have searched “Star Trek Jim Carrey” and stopped searching once they got to information about that particular episode.
IIRC search engines also use the text of the incoming link and surrounding prose. So it is quite possible that enough people on Reddit (or any other discussion site) said "The person in this episode looks like Jim Carrey $link" the the search engine picked up on it.
Citation needed. Sounding plausible doesn't mean it happens or happens frequently. Most likely there are small number of cases that fit into this description - after all, wikipedia has 54 million articles. But I won't believe there's meaningful number of such cases, unless I see an evidence to the contrary.
Not really. I have a draft page in Wikipedia about a motorcycle model stuck for months to be approved because it does not have enough references even if it provides links to the manufacturer, dealers and motorcycle review sites. No, blog posts are not enough to put something in Wikipedia.
We're not surprised when image results have no correlation with the search term. The problem is when other people fail to notice this disparity.
They'll go on to write articles and blog post based on google's results. Later someone else will make a wikipedia page based on the blog post. Google's flawed results will now be confirmed by the wikipedia page.
A virtuous cycle indeed.