Privacy-focused search engine DeepSearch slinks out of hiding Developers are working on a privacy-focused search engine that goes beyond the likes of DuckDuckGo. DeepSearch from TSignal is an AI-based search engine that does not collect any user information, according to the team behind the project. The crawler-based engine aims to maintain user privacy while offering unbiased …
... otherwise I can see Google pulling a Laker Airways on these guys...
Good luck to 'em!
DeepSearch from TSignal is an AI-based search engine that does not collect any user information, according to the team behind the project. The crawler-based engine aims to maintain user privacy while offering unbiased information discovery.
.. but if you dare to try reading their (one article) blog you are first required to Medium's blog system, which offers you such fantastic options as logging in with Twitter or Facebook (both of which options contain the worrying promise that "they won't post without asking"), Google (yeah, right) or, if you're weary of helping such data slurping giants to even more data because, you know, privacy, you can use another email account which merely gathers that bit of personal data on its own.
To that I add the fact that there is as yet no independent validation of what tsignal.io is alleging and that their email goes to yandex.net (read: 100% Russian) and for the moment you will end up with questions.
Mind you, this is brand new and it takes a while to get all ducks in a row so it may improve. At least the webserver itself is hosted in the EU (Netherlands) so that's a start. There is in my opinion indeed an urgent need for more search engines that deliver what Google once offered too: RELEVANT search results, I just hope there will be more substance behind it soon.
Caesar's wife time - something like this must be completely above suspicion. Using Yandex mail isn't a good start!
Good luck to them.
If it works, it'll be interesting to see how they fund it. Will advertisers be happy to display adverts without knowing who's seeing them?
For now I'll stick to BangBangDuck!
something like this must be completely above suspicion
True, but -as I observed- this is not as easy as it sounds because you need to be able to step back and see the bigger picture. It also takes a fairly peculiar skillset (pub friends would observe ".. or fairly peculiar people") to get this right - your average tech is going to be more exclusively focused on getting the code right and rely on commercialisation to put the code in the right frame.
I am thus willing to give them the benefit of the doubt, but it also means I don't see this as anywhere near Ready For Service.
If it's currently only indexed 4% of its target sites, it might not even be fast by the time it's done all of them and a lot more people are using it.
And how do you test if it's private and safe? Search for "how to make bombs" and see if your front door gets kicked down eventually?
Yes, it gives stale search results. Apparently they're using the test dataset from http://commoncrawl.org/ as mentioned in their blog because crawling is the easiest part a search engine. Nonetheless, it's a good start if it's truly decentralized and can scale up to compete with Google.
They use an interesting definition of "artificial intelligence" to mean there's no human intervention in the ranking algorithms, thus no censorship, no cadre of biased employees to feed the beast, and much lower costs. If that's workable it could open up much-needed competition in the search market. Lofty goal though.
I wouldn't worry about them getting sued. Open source search engines Lucene and Xapian have been around for years, they're widely used, and the algorithms are unlikely to be encumbered by patents.
Maybe I chose a difficult search term but 3 of the 4 search engines I have configured in firefox gave me what I was looking for as the 1st hit while DeepSearch started with 1 1/2 pages of wikipedia hits. It gave me some hits on the wrong part of the right site 3-4 pages in.
Maybe it hasn't spidered deep enough into the site, but, not looking good so far.
Why does the Reg use DuckDuckGo as a benchmark? They're a liability now. They're weren't ever transparent about their tie-in with Yahoo and we know how untrustworthy Yahoo are! What kind of corp would cover up the hack of every single account and work around its own IT team to feed the NSA when it wasn't even necessary... Try Startpage or Qwant instead.
http://securityspread.com/2016/10/24/duckduckgo-startpage-2016-update/
https://web.archive.org/web/20160724030640/https://duck.co/help/results/yahoo-technical-implementation
https://help.yahoo.com/kb/search-for-desktop/SLN27299.html
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
