DuckDuckGo is great, but Searx[1] is even better. It's a metasearch engine that aggregates several search providers that you can self-host, or access via one of several public instances.
I run an instance on my local network, but you can run it for free on Heroku, AWS or GCP or even on a Raspberry Pi. There are several Docker images you can choose from.
Extra irony: I went to try Dogpile, but got redirected to some anti-something site (seems to not like my VPN IP) that wants me to do a Google Recaptcha.
I use DDG as much as I can but when I need to find something very specific, or find a solution to a bug Ive encountered, it takes 2x as long on DDG vs Google
I find Google ignoring my attempts to be more specific more often.
Just as a contrived example a search like "R33 RB25DET Motorsport ECU" typically got me specific links relating to the exact car, engine and topic. But the past 5 years or so it seems like it is weighting the more common word and more general terms. Often excluding the target topic altogether and just giving general motorsport results. Perhaps it's a consequence of every SEO specialist and their dogbot hammering general search terms and gumming up the machine with cruft.
That's something that really drives me nuts about it lately. If I just wanted general results, I'd do the lazy thing and not put in the additional terms. Worse, I've been finding that it still ignores some of the terms even when I put them in quotes.
The results just seem really bad lately, especially for anything technical. Just now I'd been looking for "html5 canvas torture test". The top result is a video called "Torture Testing my Nut Sac!!" and then some videos about testing Glock guns. Umm, no, that's not even close to what I'd wanted. (Bing does way better here and DDG is somewhere in between.)
I'm not sure what Google engineers are using to find technical information on the web these days, but I can't imagine it's the public Google search.
I personally find it most helpful to just ask Google a question like I'm a complete idiot. I got the idea from the meme about "that guy wot painted them melty clocks", which works extremely well in my opinion. Looking in my history "how to multilingual in java please", worked fine. You get a laugh out of it, 90% of the time Google figures out what you need, and the rest of the time it's going to show whatever the hell it wants to, any way.
At least both of us are pretending the same level of intelligence, which takes away a lot of the irritation.
I also tried talking to DDG like a duck, but it doesn't give as good results as talking to Google like an idiot.
Just to be clear, I still use !g pretty frequently.
But psychologically it's rather different. If you find the Google search page to be visually aversive then your goal is to get in and get out quickly. That's a bit harder if Google is your default search.
Firefox has been my main browser for about 15 years. Never saw the advantage in Chrome, aside from using it once in a while when a webpage didn't work correctly on Firefox. These latest years we are seeing very aggressive behaviour from Chrome (reducing effectiveness of ad blocking, for example) and that just reinforces my decision.
I think I agree. Google still seems like my first choice mentally, but I used ddg a lot more as first search lately ( past month or so ).
It is getting more and more annoying getting workarounds for everything though. It is more annoying, because I liked G layout, default colors snd so on. It was cleaner.
Now not only is their search quality getting worse ( I got what I asked for on bing of all places ), their presentation managed to degrade too.
If it is testing result, I would be curious to see the data that informed that decision.
I usually start with DDG (it's my default search engine on all the devices I use) and then quickly move on to !s (for Startpage) since DDG still is lacking in the quality of results for many searches. It's a bit rare that I go to !g (Google search).
I run an instance on my local network, but you can run it for free on Heroku, AWS or GCP or even on a Raspberry Pi. There are several Docker images you can choose from.
[1] https://github.com/asciimoo/searx