Comment by hx8
9 months ago
Another bad aspect is that it's becoming harder to google for general information without being hit with production recommendations. For example, if you google an ambiguous term like "Indoor light" the entire front page is products. There is no information about how indoor light impacts sleep health, or comparing different technologies of light bulbs, or different styles of lighting a room, or showing how much energy we spend on indoor lighting. It's literally all products, on a topic which has a lot of nuisances to explore.
Some search terms seem to trigger "medical information", "scholarly journals" or "technical documentation" subroutines and avoid products all together.
That's because your search query isn't specific enough. Google doesn't read your mind to figure out what about "indoor light" you seek. So it defaults to commerce.
Even Kagi is like that, since "lighting" is a product class.
If I search "Indoor lighting science" on DDG almost all the product stuff is gone. I even get an NIH paper on lighting and health.
If type "indoor light sleep health" I get the results you would expect.
Same for "indoor lights technology comparison"
Do you really think it's unreasonable that "indoor lights" goes to e-commerce? What do you think most people who enter that term are looking for?
I mean seriously, do you expect Google to read your mind somehow? Everyone complains that search results are deteriorating. What I think is really happening is that "the internet" isn't only targeted at nerds anymore.
(And of course people are also strongly opposed to behavioral profiling which would be like reading your mind...)
> I mean seriously, do you expect Google to read your mind somehow?
If there is any ambiguity Google defaults to commerce. I think the front page for such general queries should have room for science, history, art, culture. Google's default is to sell you something, and it is often the only thing it shows.
A Wikipedia style "disambiguator" might help.
If Google silos queries into different top-level categories (i. e. research vs commerce) being able to clearly pick one of them, even if once you see the first results, solves the problem.
Giving users the feeling of control and the ability to improve their skills seem to be completely beyond modern software companies though. "It's all magic and algorithms beyond your understanding, mortal!"