Even that bit about "don't" being turned into "don t" was interesting. Again, not because I was amazed in any way. More so because Search has been so mysterious for many years.
And face it, for all its flaws, Google's index is larger, it has more servers, has way more signals to use - including its own web browser, DNS and ads - more money to spend and actual computer scientists who've been banging away at the problem for decades. Bing is just Microsoft's way of siphoning off some search revenues from the most technically illiterate users who don't know better and never change their defaults on their Windows box they bought at Walmart. MS has zero incentive to improve the results.
DDG is a cute little boutique web site. Good for them. But their impact on Google Search is so small as to be non-existent.
When a competitor's results are 10 times better than Google's - not a little better, or sometimes better, or better if you add a special character to your search, or subjectively better based on your personal preference, but an actual order of magnitude better - then and only then will Google even begin to start worrying about its dominant position in search.
"But Russ," I can hear you say, "How can a search result be 10x better? That's a straw man. What does it even mean to be 'better'? Faster? More accurate? More contextual? More personal? More predictive? Less spam? More niche sites? Less niche sites? More photos? More videos? More news? Less photos? Less videos? Less news? More languages? Less languages? More international? More local? More convenient? More customizable? How can you even measure something like that?"
Exactly.
Well, when you can't get the result on Google no matter how you finesse your query, but you can get it on another search engine, is that infinity better?
Because that's the situation.
If a service can consistently deliver good results on the first page (like Google used to, 20-odd years ago), they could definitely dethrone Google. Of course, the moment that appears to be the case, all SEO spammers start focusing on the new rival and the newcomer's job becomes a lot harder.
Sometimes I get terrible results and add a !g to my search query to see if le gouglè gives me something better. Almost every time it's just as bad. Usually because of SEO crap.
I felt that for technical searches past a certain number of keywords (2 or 3), ddg and brave tend to lose focus where google doesn’t, a problem exacerbated by the lack of quote search on either
google also much better answers local questions, like “[supermarket] [location] opening times”, and indexes wikipedia more usefully, plus ten other integrated features you don’t miss until they’re gone (e.g. imdb)
I didn’t notice a change in SEO spam between the alternatives. I would imagine brave have made an effort here, but as far as I know, duckduckgo filters their content less than google does
finally, brave lacks maps, which is incredibly annoying
I think as a society we need to think what it means to have one private company control access to all information.
After a while, things started to change. I don't feel the need to fall back to Google nearly as often anymore, and when I do I'm very unhappy with the results from Google. They're often worse than what I get from DDG.
YMMV of course. And yes, DDG's impact on Google Search is probably negligible, but market share is often only verly lightly related to quality.
To be clear, I wasn't saying the current state of affairs was a good thing, I was simply stating in detail that Google is still a gargantuan machine. Unless a competitor is somehow able to fire a couple of torpedoes down a hidden exhaust port, they're all going to be outgunned for the foreseeable future. But sure, the unpredictable vagaries of public sentiment could somehow turn against them, I guess. I mean, it's going to be a while. Their company name is a verb.
I wish there was a way to aggregate the two results together.
I would not be surprised if the ads situation gets worse or some other egregious action takes place, as if to say, "Remember how we fixed search results for you? Well, here's how we plan to pay for it".
I would change one little word:
Well, here's how we plan _you_ pay for it.
Too little, too late. Pity they didn't listen years ago.
Mojeek is hit or miss, but occasionally does better than Google.
I haven't really used it myself, but I've heard good things about Kagi.
Please let me turn on verbatim mode and leave it enabled. Ideally across my google account, but I'd settle for just that device/application.
One request from me would be for them to bring back the + operator and favour it instead of quotes, since Google Plus integration doesn't seem to be a thing anymore.
How about a heuristic of: If the user phrases their search as a natural language query ("what is...?", "how do I...", etc...) then use whatever weird relevancy metrics and search word substitution your research suggests will answer the question best. OTOH, if the user appears to phrase their search as just a list of keywords, search for all the words verbatim.
They already do query classification for things like Google Calculator. Extending the classifier to switch between "natural language query" and "strict keyword search" seems like a reasonable extension of that idea.
The entire reason I ask is because I don't want google to try to interpret my search query and change it - I don't want it to guess what I'm looking for using [insert classifier of choice] - I want it to do as close to a text scan for the exact search query as I can get.
Inconsistent tools are much harder to use. I really don't want to have to play a cat and mouse game with my search tool, and I don't want to have to have memorized all the "games" google is playing with my query and understand how to turn them off.
Person <verbs> blue chair
Should return any text where someone does something to a blue chair (and it's synonyms or similars)
No problems huh? Why does this article exist? Lol the hubris at Google...
https://googleblog.blogspot.com/2011/03/hide-sites-to-find-m...
Besides being overpriced, the last I checked, the results returned for the custom search didn't reflect what was returned from the normal search product.
Allow users to permanently remove trash sites like Pinterest from all their search results on both web and mobile.
Maybe this is my chance to repeat something I asked the DuckDuckGo CEO today (https://news.ycombinator.com/item?id=32366664): Are you doing anything about SEO spam?
It seems like Google often rewards low quality results: Sites with tons of ads or where I'm being shilled to buy a product that I'm looking for information on because I already have it. It's so bad that I've half-concluded it's due to misaligned incentives - keeping the user looking at the search results keeps Google collecting ad revenue and steering users to crappy sites full of ads also keeps Google collecting ad revenue.
Or this one: https://www.youtube.com/watch?v=modXC5IWTJI (terrible audio)
Both are very old, but extremely well aged.
IMO, the secret behind Google Search is not the smartness of the algorithms, but how much of it is baked in those O(100ms) which takes Google machines to answer your query. That's why the links above are the true reason Google Search performs well.
Natural Language Processing and Information Retrieval state of the art is far beyond what Google/Bing/Yahoo/Yandex/Baidu employ. But, it's far too expensive to serve it at QPS & latency required for decent UX.
Then again maybe i'm underestimating how often I don't know exactly what I'm looking for and just grab the first result.
Or is that not how it's done, and each search, for a given person, follows the same process?
I've always been curious about that.
While there's some very short caching of results, to my understanding, there's generally still going to be a lot of hitting the index because there's just a lot of new information coming in all the time. We can't somehow store a set of results for say "cars" and figure it's going to be the same info from one minute to the next.
And results don't really have a lot of personalization for individuals. When you see differences, it's usually due to language and location.
The index is like a dictionary - you look things up by word. But you need to find some way to split up every page on the internet to decide what is a 'word' and what isn't. If you decide that quote marks are part of a word, then you'll end up with apple and apple' making different entries in the index which you probably didn't want.
Language models need to account for a sentence like
> Will Will have enough will to see this through?
Google has managed fine with this since the beginning.