For several years I’ve been using DuckDuckGo instead of Google Search, and I’ve been overall quite happy with the results. Only rarely had I to resort to Google search (!g).
During the last month or two, however, I’ve found myself using the !g switch and Google search more than half of the time. DuckDuckGo shows no or few results where Google shows more (and useful) ones.
Still I don’t want to give in. So:
- Have you also experienced this worsening of DuckDuckGo?
- Which other more privacy-respecting alternatives do you recommend?
You must log in or register to comment.
Maybe I search for weird things, but my major gripe with DDG is that its autocorrect is way too aggressive. But SearXNG public instances work for me 99% of the time.
I’ve noticed this is the case but not that it’s declining. I haven’t used it that long.
I think in the last few years DDG has been improving and google has been worsening for general searching. Because I have nearly stopped using
!gbefore I used it constantly.I still use google at work as the results there match a bit better.
Startpage uses Google’s results, similar to like DDG uses Bing’s, I find to be splendid replacement for spying engine.
I haven’t used DDG in years. I’ve been using Start Page and its fit all my needs (its basically old google before enshitification).
I’ve played with searxng which seems promising but I haven’t given it enough time. But it seems like I might eventually move there.
Tried Startpage a while ago, but was put off but some alleged iffy dealings of the company behind it. Trying SearXNG now and I’m impressed!
What iffy dealings are you referring to? Because they were acquired? They’ve been established, and restablished after being acquired, as being completely privacy focused.
My only criticism is that they don’t always play nice with VPNs.
But yes, SearXNG is a great way to go too.
I switched to Kagi and am beyond satisfied. If your goal is to strictly degoogle, it fits the bill, but it still does if you are looking for better privacy, as it now comes with an implementation of Privacy Pass. The algorithm is leagues above Google’s and DDGs, IMO, and the “lens” feature allows you to seamlessly filter the results to specialized sources, including the Fediverse. “Small web” is a fun feature for when you’re bored running unit tests at work, too
Forgot to mention that it unfortunately is a US-owned company, so it would be off the table for the full-on US boycott crowd, especially because it’s a paid service.
Though they seem to be a genuinely good company that consistenly provides good customer support and improves the product in tangible ways. Privacy Pass was implemented because of customer feedback, for example, and so were crypto payments, and both were publicly discussed on the forums with good transparency. They also actively promote the decentralization of the internet: with that Small Web feature I mentioned, with Fediverse and Usenet Archives search being implemented by default, by providing an interface to use any LLM model through their assistant… So I wouldn’t want to boycott them, and I don’tI tried Kagi for a while, but it was giving me less useful results than DDG, so I simply left it. I think it depends a lot on what kinds of searches one does, and Kagi is more useful for other users.
AFAIK the algorithm for Kagi is really alien compared to Google and Bing/DDG, so the results do look a little weird at first, the main difference being just the sheer reduction in quantity of results.
But I guess if you didn’t like it, you didn’t. Maybe it is worse and I’m biased because I already paid
I use DDG and can’t say I’ve had an uptick in the amount of !g I have to do. The only one recently was for an image search but that’s pretty normal when looking for something obscure.
Don’t have any solutions, but figured another input might be interesting.
It’s very likely to depend on the kinds of searches I do, indeed. Although I think it’s the same as in the previous years. Could also be just a subjective impression, so I’ll try to keep count of how often the “!g” really leads to better results.
I run my own instance of SearX. Very reliable.
Just setup my own instance! Happy so far!
I’ve definitely felt the enshittification of DDG. A couple of years ago they would start dropping hits related to my location into my search results, even when I had region off and private search by default. That gave me the impression that my IP address was being used and possibly passed on to Bing, but I don’t have the chops to confirm it 🤷
I have thought the same thing after about three years or so of using DDG, I’ve been using Qwant as of late and seem to get much better results
I’ll give Qwant too a try, cheers! Testing SearXNG for a couple of days first.
I wish you luck with SearXNG, I don’t know if I used the wrong instances or what but I didn’t find the results i got that great. However that is the reason people say you should host you own instance. Hopefully it works out I really enjoy the idea of hosting my own instance
Same, I was a DDG user for years and switched to Qwant a month ago. Qwant results are a step up from DDG, and Qwant takes the same approach to privacy as DDG but it’s based in France so it wins in that regard as well. I’m in the US and Qwant still does a great job of providing localized results.
Duckduckgo gives you Bing results. If you like Microsoft they are up the alley. If not tough luck.
DDG is often but not always a lot worse than Google in my experience.
I haven’t experienced any issues with search results with DDG. But if you feel it’s a problem, you could try a metasearch engine like SearXNG. You can self-host it or use one of the many public instances maintained by the community. The main advantage with it is that you can pull results from multiple search engines with a single query. It’s highly customizable too. You can configure it exactly how you want.
And for anyone who doesn’t know: SearX is pronounced “Search”, and it’s successor, SearXNG, is pronounced “Searching”.
(In many languages “x” signals a “ch” sound)
Note: But maybe don’t go around saying “Have you Searching’d it yet?”
Tried using public instances, but 1/3 searches was failing. Found it unusable. That was like a year ago when I was looking for DDG alternatives.
This is unfortunately the same type of experience I’ve had with SearXNG. 🙁
I’ve had good luck with searxng.site, but yeah, public instances aren’t going to have the uptime and reliability of Google or DDG. Think of public instances as a test drive. You get a vibe for it but it’s a much better ride when you self-host.
Trying it these past days and I’m impressed!
This shit right here!
I like SearXNG, you get all of all worlds
I’ve only used searx[ng] for several years. searx.space is pretty recommended to look for searxng working instances, as well as the ones that you might prefer depending of the country of the instance and so for. Public searx but no searxng working instances are really uncommon now a days.
Every now and then your preferred instance becomes useless (whether google finds its way to block it, or to apply an aggressive rate limiter, or the instance gets unmaintained), so one needs to look for another one.
DDG doesn’t give bad results, but when I realized the majority of its results come from bing, meaning it’s mostly a metasearch as well with a few entries of its own (that might have varied from that time), I then started to only use searx, and then when searx working instances were really hard to find I moved to searxng, and I’m happy with those instances. Again, at times I need to move to a different instance, though I’ve been using the last one I chose for more than a year now…
For me at least, searxng.world works. I just went to searx.space, opened loads of them in a list, and left the ones that weren’t blocked in my country. (the only ones left were searxng.world and two really specific local ones with alphabet soup names)
Public instances need to fight against the giants, but running your own local version is easy if you learn to use Docker. It just takes around a hundred of megabytes of memory. I have been super happy with it.
There’s only a handful of companies out there actually spidering. A lot of third party offerings are just re-scraping the existing spiders. I wouldn’t be surprised if deficiencies in quality were cat and mouse games between google/bing/et all and DuckDuckGo.
I’ve been self-hosting SearXNG. It’s fantastic for everything except local hits, business hours, stuff where Google maps data is being referenced.
I think the problem with free search is that somebody needs to pay for it. There’s more people block both ads and anonymize themselves, the more free options will eventually wither.
And while I’m perfectly willing to pay for ad-free anonymity, capitalism dictates that all services need to have exponential growth or fail, and eventually all that data can just be sold or otherwise make it into the wrong hands.
I’m kind of hoping that at some point you can purchase distilled search content in a locally hostable AI model. It could post ad free and complete anonymous access, and you just need to pay for updates to the search model.
Right on. I’m running searxng and whoogle. Whoogle is a low resource option, and it only sources Google. I like searxng for the deep results, all kinds of weird stuff pops.
I was recently recommended to check out YaCy. Haven’t done it yet.
I’ve been using SearXNG during the last day and I’m quite impressed too so far!
True what you say about the problems behind net search. It’s actually a very complex problem. In my opinion part of the problem is that there’s a lot (most?) of rubbish out there. It’s like a library with useful books of different genres all mixed together, and mixed with an even larger amount of nonsense books. Maybe a solution would be something completely different from indexing – but I have no idea what.
It’s an old problem. From the very start of the net, you had to sort the wheat from the chaff. Back then, the BS was human-generated. Now we have the addition of AI crap. But anyway, they solved it already. Its called wikipedia. (Or any other community curated data source as well.) I’m not some wiki fan, but that’s the world’s answer to encroaching bad data. An army of real, very corruptible, infighting, weird-as-hell wiki editors is our last stand against the BS.
Didn’t know about several of these, cheers!
Searx is good enough if you set up plenty of engines - I do look up quite a lot of stuff and not once in the past 3 months did I go “yeah I need to use google for this”.
I’m trying SearX today, after so many recommended it. It looks promising! Thank you for pointing out the multiple-engines setup.
One possible drawback: it seems I can’t do “verbatim” searches; or at least, quotation marks don’t seem to lead to verbatim searches – I’ll try with “+”. DDG was adamant with quotation marks, that’s something I liked a lot about it.
Interesting, verbatim searches work perfectly for me. Maybe it’s some search engine that doesn’t support them? I personally have bing/google/duckduckgo selected.
