/tech/ - Tech


Mode: Reply

Max message length: 8192


Max file size: 80.00 MB

Max files: 5


(used to delete files and postings)


Remember to follow the rules

(6.32 KB 109x100 searx.png)
Searx Comrade 07/03/2020 (Fri) 11:53:59 No. 3073
I know that if I run my own instance of searx is the most private way to search things up. But what about public instances of searx like search.snopyta.org, are they any safer than just using pure duckduckgo? Because I am still trusting a 3 party with my data, the only other advantage that I see using a public instance of searx is that is completely open source. Are there any other positives?
>>3073 It's probably gonna be more useful that duckduckgo since duckduckgo's search results have become practically useless. I've been forced to use google for most shit because of how bad duckduckgo has become.
>>3074 Have you tried Startpage? I personally don't like their results but maybe it is just me, at least it is better than using google
>>3076 I have, but as it was bought out by an advertising company I stopped using it.
>>3073 I would never use my own "private" instance precisely because of privacy reasons. If all search requests from an instance are by you then you haven't really changed anything. Whereas if you use a public instance you're hidden in the crowd from the perspective of the search engines that the instance relies on. You should use it via Tor of course, like everything else that you want to keep anonymous. For similar reasons avoid using only one instance "provider" (like snopyta) for everything (invidious, nitter, bibliogram...), because that means snopyta has access to an aggregate of your browsing, similar to an ISP or a VPN. It's better to use official instances instead because they're usually isolated from each other. >I am still trusting a 3 party with my data You should never trust anybody in the first place, and instead technically reduce how much damage can potentially be done to you. Trust is unreliable. Use Tor and compartmentalize your browsing. >>3074 All of these search proxies have gone to shit in the past few years. I think the search engines feed them computationally "cheap" results in lieu of simply blocking them, which is what they did at first. It's probably a better strategy for Google to not block anything and instead make it shit. Won't cause controversy and people will still flock back to you.
>>3073 Try ya.cy
Wouldn't intelligence agencies target these privacy-oriented search engines? For example prioritize the decryption of their streams or try to backdoor them somehow.
>>3373 >decryption of their streams they can prioritize it all they like. shit won’t make a difference
>>3374 Can't HTTPS be decrypted eventually? They probably have some super powerful computers for that.
>>3375 use elliptic curves if you're afraid of teh quantums
>>3375 they don’t
>>3375 >Can't HTTPS be decrypted eventually? Every cypthertext can be cracked with enough resources (power * time). Each encryption algorithm will be hit by quantum computers differently, reducing the time it takes to crack a message. It would be good to know which encryption algorithms could be cracked in reasonable time (i.e. if some of your messages could cracked in your own lifetime).
>>3430 How hard is it to crack SSL with modern (super)computers?
>>3431 No idea. Some say it will break HTTPS "in a few years", which means that HTTPS is future-retro-actively already broken. Basically whatever you do over HTTPS can be considered only short-term encrypted. How fucked that is depends slightly on case-by-case basis, some messages are only short-term relevant anyway. However the metadata will always be useful to them, even if the message itself is completely banal.
Lemme ask you the other way around: What do you use the internet for? If you can find an offline alternative for some of the things you look for online, this can cut out some queries. You don't need to worry about queries you don't even send. If you use online dictionaries a lot, consider alternatives. A simple text file and ctrl-f might do. There is also an offline version of Wikipedia: https://github.com/kiwix/ You also don't need to go online just to look at maps.
>>3437 I like street view though, photo imagery of Earth's surface, rather than the layout?
>>3431 impossible
>>3431 Unless the NSA and friends have made a breakthrough on quantum computing and kept it secret, functionally impossible. With properly set up and non backdoored ssl crypto you're looking at average computation times longer than the heat death of the universe. Of course they could always do what they did with Dual_EC_DRBG and backdoor the encryption to make it significantly easier to break or just compromise the servers you're talking to.
>>3437 Good post. What's a good .txt dictionary? I tried looking for one once, but they were all antiquated
>>4032 Check this out: https://dumps.wikimedia.org/ Wiktionary has a lot of English words, plus etymologies, pronunciation, translations, etc.


no cookies?