Hacker News new | ask | show | jobs
by Edd314159 691 days ago
To be clear, this is not Google getting it wrong. This is the AI-generated slop website getting it wrong, Google is just pulling an extract.
7 comments

The extract being used for a search would fall in Google's domain, one would think.
Their job is to surface useful search results at the top, and inaccurate or untrustworthy sources at the bottom. They used to be quite good at doing that.
Google's offense is not detecting and excluding it. Their spam filter is not working well.
The end result is Google is getting it wrong, we just have two layers of machine slop at work
This is wrong. It's Google getting it wrong, failing to parse/understand this Reddit comment: https://news.ycombinator.com/item?id=41142610
In the intervening 15 minutes the Google featured snippet has changed, it is now the first paragraph of that page, instead of the paragraph that says he is 14 feet tall.
Of all the sources (and their reliability) available to Google, they chose the one thinking that fourteen feet is a reasonable length for a human.