If it’s as private as they claim it shouldn’t be an issue.
I hope so
how can we make sure that claim isnt bullshit?
How can we make sure their privacy claims about search aren’t bullshit?
(I still trust them though, just pointing out it is an equivalent problem)
It is neither, their AI search features only optional at this stage. It is one of the most private implementations of AI that I’ve seen though.
It is one of the most private implementations of AI that I’ve seen though.
Based on what information/criteria?
Privacy policies.
Thanks.
Try SearXNG. SearXNG is a free internet metasearch engine which aggregates results from more than 70 search services. Users are neither tracked nor profiled. Additionally, SearXNG can be used over Tor for online anonymity.
Duckduckgo is also available on Tor. https://duckduckgo.com/?q=duckduckgo+onion&t=fpas&ia=web
As is Startpage now.
Startpage is owned by an advertising company
And DDG is on record for censorship and caving to Microsoft. They also both provide results based off MS and Google results.
Neither are ideal. It’s why only use them as a fallback. Brave Search all the way.
I very, very strongly disagree.
Removed by mod
What a great argument! You definitely convinced me and everyone else! Congrats!
Removed by mod
This one was so obvious, if you really need the /s to understand that I’m not being serious then you are dumb as fuck. You are the troll who responds to legitimate criticism with “shut up”.
It’s neither. Why do you expect they calculating an answer to have any impact at privacy?
I mean DDG search is safe, but my search results and search queries are going through AI, who knows they are collecting it
DDG already sends your search query to third parties. What they don’t send is your identity.
(Or, at least they say they don’t, whether you trust them is your option. Any 3rd party can betray your trust.)
Anyway, that image implies an in-house implementation.
It only goes through their Wikipedia LLM if you push that button
Really? It mean it only use Wikipedia. Then IDC.
Personally I’ve used it a few times, I tend to get my answer significantly easier than wading through the shit results DDG has been giving me.
Is it good
no.
Oh no :( DDG also becoming bad. Can you suggest me any other search engine (Not Self Hosted one).
DDG is still the best non self host search engine imao. But Startpage and Swisscows are also great options and don’t have the AI stuff and some neat features like private preview (both) or music search (swisscows). Of those two options startpage is superior in privacy
I’m currently stuck using DuckDuckGo myself, just with the AI fluff disabled in settings. I’ve tried various other search engines but they’re either also buying in the LLM hype, or are just Not There Yet™ (https://stract.com, for example), or both.
Brave Search. At this point, I almost never need a different one. They also have optional AI but they’re doing so in a way that preserves privacy.
No, VLC is its own thing however it uses libavcodec from the FFmpeg project for a large number of the codecs included in VLC. But VLC is far from being just an FFmpeg GUI.
Why.
To hell with anything using this AI garbage. Apparently I’m going to have to completely stop using computers because all these idiot companies are using this trash technology that just steals from other sites and presents it as its own. Time to move on from DDG.
They claim not to save your chats and not to use them for improving the AI model. Using GPT-3.5 Turbo through DuckDuckGo is definitely better than using your phone number to sign up for an OpenAI account, so yes, I think it makes sense to use this. But ideally, you could just run an AI model offline on your own hardware, I recently saw a Lemmy comment that explains this very well: https://lemmy.dbzer0.com/comment/10348570
You can also use @db0@lemmy.dbzer0.com’s AI Horde as explained here: https://lemmy.dbzer0.com/comment/9295041
Honestly their results have been absolute shit for a while now. All I get are big business (mostly MS with Bing) results like Amazon links or worst yet, some BS AI built website with poorly stolen info meant to just push visitor traffic and not actually provide value.
I’ve been looking into alternates and find Yandex actually provides real results. They don’t blacklist either something all the major US engines do because DMCA.
At this point it’s like to I give up privacy and get real results or stick with DDG that seems in MS’ pocket so I have little trust they are still anonymizing your results.
If you will use it your question will be sent to OpenAI or Anthropic and possibly DuckDuckGo companies + search engines it is connected to.
If you don’t your question will be sent to the website you chose, possibly to DuckDuckGo + search engines it uses.
So you decide what is more private in this case.
As unfortunate as it is, the only way privacy search engines will survive is by adding AI. Brave recognized this early on and implemented it very well, so I’m glad to see DDG seems to be implementing it well also. Hopefully Startpage catches up because even though I almost exclusively use Brave Search, I do enjoy Startpage as a secondary source.
Honestly I see no use for the ai response in search as it is hard to trust 100% but I do see that it could become a requirement for a search engine
I agree somewhat. I know how to use a search engine, I don’t need AI. The main problem is that, sadly, AI is going to be required to stay competitive.
I will say Brave’s AI is very useful, though.
I’m not sure, but if I’m going to use AI features, I’d trust DDG more than I’d trust the big players in AI.
Yes.
Some day surely Windows and Mac will get something similar to SearXNG, then you won’t have to worry about this stuff
Nah, self hosted local search engine on Microsoft’s platform? Lol
Yep
I am a Linux user.
Then I’d suggest SearXNG. I use public instances because no one has made a way for me to natively host it myself, but I recommend it regardless
What prevents you? Lots of people run their own instance.
Unfortunately it’s broken all the time because it gets IP-banned by Google. Perhaps they could implement something like the Invidious smart-ipv6-rotator to prevent this.