Matadisco – Decentralized Data Discovery

(matadisco.org)

58 points | by biggestfan 3 days ago

4 comments

  • mentalgear 17 hours ago
    Decentralised Web Indexes are an important avenue if we truly want to ever break free from the private indexes used and dominated by Google and chatbots that in-transparently return results based on their opaque preferences.
    • stingraycharles 16 hours ago
      Those indexes cost an incredible amount of money to build, maintain and keep improving. It’s worth it for these orgs as these are considered a competitive edge and the “secret sauce” of these orgs.

      I find it hard to believe the same amount of investment will be made into a decentralized web index, precisely for the reason that it cannot be monetized.

      • somat 8 hours ago
        No they don't. Sure, if you are trying to capture the entire web with hourly latency for billions of users, that takes significant resources. But if your goal is to catalog the subsection of the web you are interested in with weekly latency for yourself and share that catalog with like minded individuals. any body can do that. Don't make gods out of basic web indexing.
      • johnebgd 11 hours ago
        This. Unless in 2026, we suddenly see Costco sell personal quantum computers for sub $2k, with practically unlimited high speed crystal storage, all cooled/powered by unobtainium fusion reactors that are made with 3D printers at your neighborhood coop, then we either must accept private entities that do this kind of indexing at scale or we should ask government to offer it on taxpayer expense.
        • somat 8 hours ago
          On a pedantic note, that is unimportant to the discussion at hand, but I want to talk about it anyway. quantum computers have theoretical gains in some specific domains but would not do anything for this one. Anyhow I have doubts about even those specific domains, my gut says that quantum computing is an inherently analog process and we moved off analog computers for a reason. Bonus points if you know why that was.
      • gessha 7 hours ago
        Are you familiar with any systems that use collaborative filtering to distribute the work and the benefits among users?
      • dietr1ch 13 hours ago
        The keep improving part hasn't done so well in 10 years already? Maybe this year the new force-fed AI answers got a bit useful, but many times the risk of hallucination means you still have to go and read a more credible source.
  • tomnicholas1 11 hours ago
    Awesome to see this project here - it was partly inspired by my blog post (original is linked from the OP, but there's a slightly newer version on my personal site here[0]).

    [0]: https://tom-nicholas.com/blog/2025/science-needs-a-social-ne...

  • thenthenthen 15 hours ago
    The demo shows a bunch of Sentinel image tiles for me.
  • dopidopHN2 16 hours ago
    [dead]