They used to recognize that long term it's better to make the decision that the users would want even if it's worse short term. The company was founded on this principle when Larry and Sergey tried to sell pagerank to Overature search (iirc) - the Overature ceo at the time said good results were bad because users would leave the 'portal' page too soon and not see enough ads. That's partly what lead to them creating their own company. [0]
I'd guess Google is making an Overature like mistake today. I guess it could also be failure to operate on stuff like this at scale, which also isn't great.
[0]: I'm pretty sure I'm remembering this from Levy's In The Plex, but I read it years ago so not 100% sure this is accurate.
They'd have to disambiguate that from the obvious motivated reasoning that's likely to exist there too.
Even then, sometimes you know better than users when making long term bets about what you do, faster horses and all - it's part art and part science.
Average users are also not necessarily the best to optimize on. 99% of wikipedia users never write anything. Does that mean the users that write are relatively worthless?
!w <search> to search Wikipedia directly.
(semi-joking... but complaining about typing 3-5 letters is definitely a 1st world problem ;)
That is to say that Google image search is the real problem and not Pinterest.
Google was able to break out because they managed to retain great talent and was able to stay funded for a few critical years until their bets paid off. And they did so rather quickly.
I can't help but feel that we're back to a variant of the portal mindset. Only now it is more about B2B and spying on users. And it doesn't feel like Google does anything out of passion anymore. There's really no race to perfect technology - just to be good enough to stay ahead of competition.
The Google of old might have invested in image search to fulfil its potential. The Google of today probably sees it as waste of effort.
The effect of this is it's in fact muuuuch easier to get at the original image in image search. Just right click the image and hit "view image" or "view image in new tab" and, that's the original.
What's annoying though is that Google doesn't show you the image sizes anymore. That's also relatively recent.
>A Pinterest spokesperson tells Input that the problem doesn’t lie with the company, but with search engines. “As a search engine, Google controls how results appear based on engagement and how useful searchers find the content.”
Google: "We allow shitty things because money."
Pinterest: "also money."
EDIT: strange, i tried the same process again on the same image and this time it worked! i guess it really is a cache-related issue
Yandex image search is also great, you can do semantic search by image and crop, much more advanced than Google. It's so good it feels like Pinterest without the fluff and user accounts.
Google at some point replaced identifying the exact image and finding different sizes/versions into some sort of machine learning problem where it tries (poorly) to identify the content of the image then shows you images that look similar but are usually not what you actually want.
Most of the time Google gives me a correct result ONLY if I know exactly what I'm searching for. But that's usually not the case.
Even for technical searches, I notice that many times, Bing gives a more varied search result which helps me.
Conversely, on Google after the first few links, the rest seem to skip some words from my search entirely. It feels as if Google is just trying to give me links to fill up the page and then I have to go back and put double-quotes against each term. Even after all that, I only get what I want more and more rarely.
Of late, I notice myself relying on Bing more frequently.
And that's yet another reason why I mostly use DDG/Bing now.
Except when images are scrubbed from results, such as "Tank Man."
https://news.ycombinator.com/item?id=27394925
Microsoft blamed "accidental human error": https://news.ycombinator.com/item?id=27400774
[0]: https://www.reddit.com/r/sysadmin/comments/p2gzi9/bing_searc... and https://www.reddit.com/r/sysadmin/comments/p2q4kw/rebing_sea...
https://www.youtube.com/watch?v=hHt98WE5FxU
My summation of the idea is that data is very valuable and will be more valuable in the future, but that because the value of a dataset goes up nonlinearly with its size, only the wealthiest will have the means to extract (significant) value from datasets in the future. This will cause massive societal inequality unless the people who create the data are guaranteed a piece of the value created by companies with the server farms necessary to actually create value out of the data we're constantly producing.
Even the most basic website would have at least some information about the image.
To clarify, this isn't an issue if you get multiple versions of the image, in that case one of them is hopefully not from Pinterest.
That's a bit reductive. I agree Google shares some blame, but Pinterest engages in some pretty serious black hat SEO techniques. This is well documented (and sadly not covered by this article, despite its title).
For each result you can see multiple pages that include the image. It also shows the resolution and file size of each version. I did a test search for "bear" and clicked on the first result. It showed 14 different pages containing the same image with different sizes and possibly slightly different crops. You can also select a subsection of the image to search for that.
It makes Google image search feel like an old school internet tool. I know Bing isn't cool or popular, but their image search is undeniably superior.
I don't think we need to blame Pinterest for having indexable content. This is a failing of Google image search for not showing alternate matches for the same image.
Bing has it figured out. I can't tell if Google just doesn't care enough about image search to implement it, or if there's some obscure patent preventing them from doing so.