Re:Google Memory Loss
by fixermark on Monday January 15, 2018 @09:50PM
(#16155414)
I wonder if what you're seeing is a result of the Internet growing, not specifically Google degrading. Which is to say, as the indexable dataset grows, you'd expect the individual indices to turn up more and more heterogeneous results over time.
[ Parent
| on HN ]
Nice theory, but it's Google degrading. They actually removed features for more accurate searching. Or they arbitrarily ignore them--all the people in this thread saying that "word" now functions like +word used to, are mistaken (or didn't use it much in the past decade) because often it does not.
And you didn't even need to use +word for most words, because it would only give you results that hit all the keywords in the query. You only needed the + operator to include "stop words", a relatively short list of common words that Google would filter out.
Also, given that personal webpages are way less popular than they used to, the Internet might not even be growing that fast, at least not the parts that contain the type of random useful info that we used to be searching for. They can't index Facebook (not very deep), or most of the popular mainstream services that people write their thoughts into.
The fact that other engines are able to offer up results tells me this is a Google thing, not the size of the internet.
... unless there's something Google offers up reliably that those other engines miss. Then what you're observing is heterogeneity of returned results due to the dataset being too large for any one index to cover it fully.
It could be the internet (content) growing, or it could be the search engine changing, but to me, the most salient change over 20 years is the internet user base changing. At the risk of sounding elitist, the users of the internet have become a lot dumber on average. Search engines therefore grew dumber relative to the one that was tuned to people who were early adopters and used the tool for more productive purposes.
With respect: yes, that does sound elitist. I would avoid assuming early use cases were more productive than subsequent average use cases (for one thing, it risks defining "productive" as "stuff I like to do," not via some more objective metric like "revenue generated," "human needs satisfied," or "questions answered").
On the objective side, most revenue today is from advertisement and ad-generating content. Search engines make it easy to find ads, essentially. It's not about the content itself any more. So there is that.