What search engine is currently showing the most useful results? What other tricks do we have aside of adding “reddit” or whatever internet community to the results?
Don’t even care about SEO fuckery, if the damn things would respect the search query.
Quotes, operands & other modifiers seem to have been straight up jettisoned.
Yep, Google decided it was too complicated and removed it all. Dont know how it was too complicated, people just wouldn’t use it if they didn’t know about it. They felt “natural language” would be more useful. Bullshit, I search for “foo and bar” it’ll return me results for foo and ignore the rest
That’s not why they ignore them. They ignore them because it is profitable.
So google has reverted to late 90s search behavior
- argv_minus_one ( @argv_minus_one@beehaw.org ) 6•1 year ago
Considering what happened to late-'90s search engines, that seems like a rather dangerous idea…
They got killed by Google. I was a Dogpile man myself, until someone showed me the google search.
Does anyone know the AOL keyword for Lemmy?
WDYM? Quotes work for me. Can you give an example that is broken?
I frequently have to look up whether a term is a misspelling/mistranslation or an actual technical term (or a term in British English, or a British spelling for a technical word). For me, quotes do nothing. It will frequently refuse to look up the term I’m specifically hunting for, just the term it thinks I should be hunting for. Sometimes that means it’s a mistranslation… but not always.
Next time it comes up for me, I’ll keep a note of it and get back to you.
I have an even bigger problem trying to exclude terms from a search. The example I always use is try to look up “Dolphins -football”, and use any version of “-” you’d like (NOT, etc). The first results will always be the latest scores for the Miami Dolphins.
This still works if you use double quotes!
Too complicated for them and their optimisations, maybe?
Yeah I really miss those days of logical operands. Back in the Alta Vista days I could do Boolean searches, but yeah that’s been replaced with speech recognition which doesn’t work as well. To this day I still like the Boolean search better. Newer does not always mean better. Most of the time it only means dumbed down to the lowest common denominator.
For my job and work. I use Kagi. Its not free, but the search returns are very good, you can filter domains out from your returns, it supports custom “bangs” ala duck duck go and theres no tracking of queries. There are also specific filters for things like programming, or recipes for cooking etc. Theres also no ads, you are paying and are the customer. They are trying to establish a sustainable model to run on that allows for privacy.
I find it quite refreshing. It isnt free and I generally hate subscription stuff, but this is easily one I dont mind as it pays dividends often when searching for work.
Wow. I don’t mind paying for stuff if it’s good. But seriously $5/month seems pretty expensive, and you only get 300 searches. $25 for unlimited searches, which seems like an insane amount of money.
The problem here is so many people are used to tech running at a loss on the books and/subsiding operating costs by selling customer data and analytics.
The reality is running tech companies is hard and expensive. The money here goes straight back into development. It’s just out of beta since march, and they have increased their quotas since I have been a customer.
But people are spoiled by free where you aren’t a customer. You are the product. If you are cool with that it’s fine. This isn’t the product for you.
For me, I like the idea and the searches are better than DDG/bing and startpage/google. So it’s worth the cost personally. I would rather pay that than say…Amazon prime where I’m both the customer and the product.
I mean yes I agree with all your points. But I stand by the assertion that it’s too expensive. I could handle $5/month, perhaps, but 300 searches is waaaay too few. That’s 10 per day. I did 10 searches this morning before I got out of bed.
For unlimited searches it’s twice the cost of a streaming service. Yet it has negligible bandwidth costs, and significantly less storage cost, probably less development cost. Sure a small user base too, but at that price they’re really going to struggle to grow it!
It’s really just too expensive.
At $10 it’s 1000 unique searches. I search a ton and have it on my phone etc. haven’t exceeded the limit. I am at 600 searches right now, with a renewal due on the 24th.
They are writing a search engine from scratch. They don’t just randomize bing or google searches. So I think you may be underestimating the operating and especially development costs, probably hosting costs too.
But to each his own. Also those streaming services you mention. They don’t really turn a profit, and definitely don’t on subscriptions.
1000 is more reasonable but it’s still only 33 per day. I’ve done 52 searches today. $10 is still way too much.
How much better would a search engine have to be to make it worth the cost of a streaming service? For me, quite a lot…
But yeah I don’t mean to say your choice to pay for it isn’t valid. As you say, to each their own.
Understandable.
I think my point is for me and in my specific use case, I actually search less.
For example if I am debugging a process or working through some setup, I will often have to iterate through a series of searches with tweaks in DDg and sometimes even google. Using tweaks like site:some site.com, quoted portions of queries to reduce useless returns etc.
Kagi, again for me, had helped reduce that. I can’t often find a very quality source in the first query or two.
So the limit wasnt hugely a problem. I was actually VERY concerned like you because above 10 dollars is pretty steep. I initially signed up at 10, set limits not to exceed 15 and figured I would cancel and either submit a request at work for an annual or just ditch it.
Luckily two things happened that retained me. The first I already mentioned. The second was they bumped the quota to 1000.
Again I may still jsut see if I can get work to pay it out. But at 10 bucks it’s digestible, for me, for the value add. I also do no filtering. Just search whatever random shit I think of n the shitter in addition to curated work searches.
I’m not trying to sway you. Idgaf if you use it or not. Just trying to help provide useful information because for me, it was more “ehhh let’s see how it works out”
Finally, I have reached out to Vlad about suggestions and even corrections on things, both in the product and ancillaries (like their documentation). He’s responded each time and even corrected some of the issues. Which is really nice.
I could actually see myself paying for the $25/mo option and leveraging that into a “free” alt-google that slurps up all your data for me to monitize however I can. Be sure to keep an eye out for it! :D
They are writing a search engine from scratch
They are using Google and a few other engines, but unlike Searx, they are using the official API instead of scraping, which is a big part of costs
But the problem is that this is what it costs for a search that doesn’t sell your data or advertise to you. Search is expensive.
Fortunately you do get into the habit of just searching sites directly, like wikipedia, MDN, archwiki, etc., rather than using up your general purpose searches.
It’s this, or sell your data to Google for free searches.
And maybe you’re right. Maybe it’s just not sustainable for searched to be paid, but Kagi is really transparent about their pricing. It’s just expensive unless it’s subsidized by ads or data collection.
I will personally always be against any paywalls on information but to each their own I guess.
The free trial with a 100 searches makes it pretty easy to figure out how much you actually search online and if you’re not a power user, that 300 searches plan is pretty OK. If you work in tech, that 10$ plan is definitely enough - in searching pretty much constantly and never got above the 800 searches the 10$ plan used to offer (now that plan has 1000 searches in it).
Not sure where you are, but there’s practically no place in the US you get a lunch for that. In flat terms it’s quite cheep. It’s only expensive relative to free.
And when you think about it, your search service really is your internet. It shapes your whole internet experience. If that’s not worth $5/month to make sure it’s good and not polluted with ads, I don’t know what to tell you.
Problem is, 300 searches is 10 per day. I’ve done 52 today. To cover that I’d be paying $25 per month.
I you could have Spotify and Netflix for that.
If I’d paid their $5 rate and done 52 searches every day they’d have billed me $63 in overage charges.
Their pricing model seems insane to me.
((52x30)-1000)0.015 is $8.40 over the $10 plan. You wouldn’t need the $25 plan yet.
And 52 is a huge number. I’d bet you could cut that in half easily.
$5 is fine. 300 seems low. I wonder how many searches I actually do in a month.
I wasn’t sure ethor. My first month (last month), I used just over 180. This month might break 200, I have 5 days left. So I’m good.
300 searches per day??
No. Per month.
- argv_minus_one ( @argv_minus_one@beehaw.org ) 1•1 year ago
Those prices don’t seem super horrible, but I don’t see any reason to trust that this company isn’t mining and selling my data in addition to collecting my money.
Seconding Kagi. I like the ability to pin/raise/lower domains as well as just block them. I tend to surface websites like the NHS.
- argv_minus_one ( @argv_minus_one@beehaw.org ) 6•1 year ago
Theres also no ads, you are paying and are the customer.
This is a fallacy. Just because you’re paying doesn’t mean you’re the customer. Whoever pays the most money is the customer; everyone else is the product and is merely paying for the privilege of being the product. Examples: Microsoft Windows, most Android phones, cable TV.
The difference being it’s literally part of their mission statement and core purposes for creating the product…
Hey, thanks for the recommendation. I had no idea a service like this existed. I’ve been frustrated with all of the search engines for years now. I just signed up. Hopefully it turns out to be rad.
i’m also a Kagi user/fan. it looks good, is fast, doesn’t have ads, & the results appear to be better than i get using other engines. the lenses are also nice.
Here’s my experience with some search engines:
A Tier – Gives me the closest results.
- Google: A classic and oftentimes, it gets what I want. A lot of the links are redirects which is annoying.
- Kagi: It’s paid but it has a lot of features like “lenses” and “quick answer”. The results are pretty good. It gives me good articles and PDFs instead of a blogspot post.
- You.com: The WORST UI EVER but the results are surprisingly decent. It’s pretty close to Kagi. It might actually be the same thing. It also has an AI chatbot but I don’t think it’s as good as Bing’s or OpenAI’s.
B Tier – Gives me decent results.
- Startpage:
It used to use Google search results but they switched to Bing. It is worse than Google.EDIT: Search results are still closer to Google but they “incorporate Microsoft Bing results”. From my experience, it filters out some of Google results that were very useful for me. Their widgets (particularly the Wikipedia one) sometimes displays irrelavant information. - DuckDuckGo: Results are worse than Google. One time a referral link came up in one of my searches.
- Bing: There’s no dark mode. The AI chat tool is pretty nice and is comparable to the OpenAI one (significantly better than Google’s Bard). Search results are worse than Google.
- Yandex: Search results are similar to DuckDuckGo.
- Ecosia: Search results are similar to the ones above.
C Tier – Gives me poor results.
- Brave: Search results feel so inconsistent and out of place. Maybe worse than the ones above.
- Mojeek: Independent search engine. Results aren’t very good.
Open Source Front Ends - Results quality varies.
- SearXNG: It depends on which instance you’re using. Sometimes search results error out due to rate limiting but you still get results anyway. It has a lot of options and configs so it fits to your liking so you can choose which search engines you want to include.
- LibreX: Actually one of my favorites since I’ve never encountered errors due to rate limiting but using it to search for images is terribly slow. It has a cool feature where you can add front ends like Libreddit and Wikiless. It also has a built-in torrent search engine.
- Whoogle: The UI isn’t very good and it performs poorly on most public instances. A smaller or private instance might be worth looking into. It uses Google search results.
F Tier – It sucks.
- Qwant: Not available in my country.
If anyone knows of any other search engine not in this list, let me know so I can try it out.
It used to use Google search results but they switched to Bing. It is worse than Google.
That’d be news to me and an ad hoc comparison I just did shows results much closer to Google than Bing with results usually just locally having switched places while on Bing it’s an entirely different order.
They do(did?) use Bing for mobile search results because daddy Google forced them to not be competitive on the platform they’re most interested in.
Now that I’m trying it again, it actually is similar to Google’s results but it filters some of the more useful results I get from Google based on some things I’m searching up.
We have more broadly incorporated Microsoft Bing into our results, using a unique solution specifically fitted to our privacy promise. Our collaboration with Microsoft also has enabled us to provide a superior mobile experience. You will see benefits like better search suggestions, fewer ads, and greatly improved similar image results, among others.
I don’t know of this only applies to the widges or actual search results.
I can agree with the google placement if you’re assuming the searcher has experience with search operators, most of the time if I’m not wasting time crafting my search results to exclude all the SEO spam sites and Q&A sites written with the same amount of padding as a middle school book report, DuckDuckGo will give me better results than Google.
Is you.com having issues currently? Tried test searches and every content element just keeps loading
It works for me right now.
Thanks for testing. Pinpointed the issue to ad blocking on DNS. I have no idea which domains it needs whitelisted
Where can I find soemthing about startpage changing to bing? Can’t find anything about that in my search engine 😅
We have more broadly incorporated Microsoft Bing into our results, using a unique solution specifically fitted to our privacy promise. Our collaboration with Microsoft also has enabled us to provide a superior mobile experience. You will see benefits like better search suggestions, fewer ads, and greatly improved similar image results, among others.
I don’t know if this applies to the actual search results or the widgets. But upon checking, the results are actually still closer to Google but it filters out some results that I find useful. It also pulls up the first Wikipedia article it can find in the first page and displays it as a little widget on the side even if it’s far from relevant.
what about yep.com ?
Thanks for the suggestion! I’ll try it out this week when I work on my papers.
How about yahoo.com?
I’ll try it too this week.
More and more I have been using the Bing “chat” search. It does a search, filters through the results and summarizes the answer with links to the sites it found them on.
For certain types of search it is a huge time saver of scrolling through results to find answers on various pages.
Over all bing search it self isn’t bad.
Dunno why you’re getting down voted. It’s literally a search engine that can read all the bullshit faster than you, so that you don’t have to.
If it isn’t open / free / private there is a % of the community that will not even try it.
Just like on Reddit lots of negative energy in some subs.
Hardly saying bing is amazing only that lately I have been drawn to trying it more since the chat based search that allows follow ups in natural language.
Google bards equivalent is only available in the US and just this last week the UK so I can’t try it out.
However over all I agree that more and more google search results have more adds and the good results pushed further and further down.
hosting a local LLM to summarize search results could be very cool though
I downvoted because I have literally no idea what that guy is talking about.
Bing has never been a good search engine. The results are always so terrible, plus you have to wade through all the Microsoft click-baity crap they put everywhere.
I do like Bing for porn tho…
the clickbait alone is enough to turn me away from Bing and Edge
cool that people don’t mind it but it shouldn’t be controversial to dislike Bing for bad UX
Have you tried using the Chat feature (GPT-4) to do searching? I just tried it, and it surprisingly works really well for some inquiries.
Like, use their chat AI, but as a natural language search engine. It’s integrated to Bing’s index so it can peruse it itself, so you don’t have to wade through all the Microsoft click-baits crap they put everywhere.
You can’t use it on mobile without downloading the app and granting it god knows what permissions. Hell nah…
AI’s problems in my experience is that they can’t always distinguish what is bullshit or even understand my question fully. It can be a great time saver, but it’s also more fallible than if I did the search myself.
perplexity is also good
I recently switched to Bing after years of disappointment from Google and months of disappointment from DDG. Bing is pretty disappointing too, but less so, so far. I tried to use the chat feature a couple of days ago, but it said I have to download the app. Nah… fuck these tech companies and their apps.
The “preview” for the chat feature requires the app or edge on desktop currently but I do find myself turning to it every time I get frustrated with a google search these days.
Less disappointing is probably the best discrimination as you said.
I signed up for Kagi.com right after posting the above. I saw someone’s recommended here in this thread and said what the heck!
I use the ChatGPT feature from desktop Firefox with no problems. Maybe it specifically denies Chrome, in which case I bet you could change the user agent string and get it to work.
It denies mobile. I was using Firefox. Same as Reddit, anything to try to force an app download.
I just tried it again on desktop and it worked, but the reason was that I downloaded an extension a while ago and forgot about it. When I disabled the extension, it stopped working.
There used to be a way to enable installing any extension on mobile FFx Dev, but I’m not sure if that still works. The desktop extension just changes the user agent string, so that might be another route to enabling it.
For a brief moment in time search engines were perfected. Then they veered off course. All of them did. Why though.
Remember when you could list vaguely some words related an obscure movie to Google. Then it would tell you the movie you’re thinking of. That’s been nerfed.
Tangentially related. What’s the deal with search engines of online stores. It’s like they aren’t even search engines at all. They’re doing nothing more than showing me products/sellers they want me to buy from. Digikey lets you drill down to precise specification filters. I wish all search engines could be like that.
Actually the vague movie description thing was due to imdb’s movie tags that users had set on the movie and search engine was doing the simples things it could i.e. “all this one word links point to this movie, perhaps it is this?”
Why’d they stop?
Have you tried https://search.brave.com?
Disclaimer: I work at Brave, but I think it’s pretty good
Brave Search shadily relicenses the content of others: https://stackdiary.com/brave-selling-copyrighted-data-for-ai-training/
Oh shoot. I actually use it. That’s bad. But then again, what isn’t? What do I switch to? It’s all shit, but at least with Brave search, I actually get results for <insert obscure Linux issue here>.
I’m using Kagi, which aggregates search results from several search engines (including their own), but without the ads, with less crap and with features like searching for literal strings and promoting/demoting certain websites. It’s a paid service, though, but I like it enough that I’m ok with that.
I’ve been using https://www.ecosia.org/ because they plow some of their profits into planting trees. They use bing results and I generally find what I need quickly.
I use Swisscows and Metager, and usually find what I need, if I don’t I retry the query with Startpage.
I am not sure whether DDG or SearX results are optimised by someone, but it is different from what I would see on Google for a given topic.
DDG seems to get most of their results from Bing, but they do tweak it a bit themselves. From the link:
Most of our search result pages feature one or more Instant Answers. To deliver Instant Answers on specific topics, DuckDuckGo leverages many sources, including specialized sources like Sportradar and crowd-sourced sites like Wikipedia. We also maintain our own crawler (DuckDuckBot) and many indexes to support our results. Of course, we have more traditional links and images in our search results too, which we largely source from Bing. Our focus is synthesizing all these sources to create a superior search experience.
Duckduckgo uses search results from Bing, combined with other search engines and their own bot if Wikipedia is to be believed.
No Google search results are used.
This explains why their results are so much worse than Google’s.
Palpable 2015 energy in this take right here.
5 years ago sure. Not sure I really agree nowadays.
Granted, that has as much to do with Google getting much worse as it does DDG getting better.
For some things. For other things they’re better. Google has really jumped the shark in the past two years.
I mostly just use it for when I want corporate style results, like shopping.
Where can I find soemthing about startpage changing to bing?
yandex.com if you don’t mind ru gov spying on you. Though you can use it with tor or VPN for privacy.
Your own whonix
