• 8 Posts
  • 43 Comments
Joined 24 days ago
cake
Cake day: June 2nd, 2026

help-circle


  • Adding my personal notes on search engines here for anyone’s interest. I personally use Qwant on Desktop and DuckDuckGo on mobile. I like Qwant because they are at least working on their own index and are EU-based. On the other hand, DuckDuckGo is faster and has a more comprehensive privacy policy. I’m really trying to use Mojeek on mobile but the search results are much worse than DuckDuckGo and Qwant in my repeated experience.

    Qwant DuckDuckGo Mojeek xPrivo Kagi
    IP collection Yes No No No temporary
    Hosting FRA USA UK EU USA
    Index ~40% own index + ~60% Bing 100% Bing Own Own Own
    Direct monthly cost 0 0 0 4-7€ 5€
    Passing data to third parties Search data and IP go to Microsoft separately No No No No
    Quality (subjective) +++ +++ + ++ ?
    AI summary / chat unclear optional no optional ?
    Speed + ++ +++ ++ ?






  • Could one integrate this with apps like NextTube or PipePipe? I.e., When I search for a video on those apps, they search the torrent index first, then search Frama Tube / PeerTube second, then search YouTube-proper last. While I’m streaming a video from any of these sources, I am then also downloading and seeding it to the torrent network and I keep seeding the last videos I watched on a rolling basis until an allocated memory space on my disk is full and the oldest or least requested video in that local buffer is deleted to make space for new; while I’m on Wifi to save mobile data? I think providing such seamless integration is the best way to get this space densely populated enough to be useful.



  • I have not. To be perfectly honest, I don’t really understand how that would even work. Can you elaborate? At home, I run a local model with a Kobold CCP backbone to localhost. The physical network is a private Wifi, though the computer is running VPN and I haven’t given much thought about what that means for the AI via localhost. At work, I can thankfully use a responsibly managed AI (company servers, very strong and externally audited data privacy standard with zero on-server data retention) for coding.





  • Can you explain in a little more detail how enforcing online ID prevents WW3? Genuinely curious. The only thing I think of that national online ID might help with is counter intelligence, especially in defense against psyops. However, in the few cases that we do know about psyops toppling elections, e.g., Brexit, these were performed on behalf of or with the aid of party and government officials in the affected countries. If any, this would become easier, because widespread online ID silents dissenting voices, while well-financed entities can navigate and / or circumvent such regulation (also see, for example, the effect of GDPR on the market structure of attention merchants in Europe).






  • Spoiler (I apologize for what I said when I was 11)

    old enough to bang your mum, hahaha

    No, seriously. I think many would agree that the internet user experience peaked some time after Google entered the scene (yes, officer, right down this sub) but before YouTube left every serious competitor behind. There was a lot of “small web” content with no clear commercial intent (not blasting you with two affiliate links and one video ad per paragraph). Many of the big platforms were controlled by the techies who set them up and not yet by the venture capital who would eventually buy them out. Yet, venture capital already kept these firms afloat, so a lot of genuinely good services were genuinely free for the user and not paywalled or privacy-paywalled (just give us your email address and IP, bro, trust us bro, just one more captcha, bro, maybe one more 2FA using your phone number, bro, really, we might even let you visit our site then). Of course, someone had to pay up eventually: Enshittification ensued.

    A second aspect: For the past decade at least, democratic-presenting governments have used all our web data fed into clandestine technology to win elections, either to stay in power or get into power and pull up the ladder behind them. I guess it’s like that old saying: A small time criminal robs a bank, a big time criminal owns a bank. Sure, we had all sorts of amateur criminals on the web in the predotcom and dotcom era and that might’ve cooled down a bit since. But now all the big players are adversarial, instead.

    Edit: Typo in spoiler tag




  • TLDR: Both OpenAI & Anthropic about to go public and expected to exceed a 1 trillion USD valuation. Anthropic expected to pull ahead because of more sustainable business user base, while OpenAI mostly caters to free tier personal users. OpenAI more open to military use of its technology than Anthropic, the latter clinching with Pentagon over surveillance and autonomous weapons use.

    Please don’t just post links without at least a short summary that let’s me gage whether that article is interesting for me.