blog

Personal blog
git clone git://git.laack.co/blog.git
Log | Files | Refs

search-is-the-problem.md (2134B)


      1 # (Centralized) Search is the Problem
      2 
      3 Search engines killed the useful, interesting internet. All corners of the www 
      4 have been infested by companies astroturfing, people with dumbass agendas, and 
      5 AI slop (no, your slop isn't different, no you don't hold it differently, it's
      6 all fucking slop, stop kidding yourself).
      7 
      8 Why did this happen? Put simply, I believe this is because of search engines. 
      9 Search engines attempt to answer queries from a diffuse set of sources, and 
     10 since this is how most people use the internet, it is a high value target. It 
     11 has since been solved as a problem. You will never find a blog post of mine in
     12 search results, but you'll find some slop top ten list, even if you search the 
     13 exact name of my post, which are indexed by search engines mind you. I guarantee
     14 my posts are at least better than the average search result quality because I'm
     15 not a sell-out, but you won't find them. I won't bend the knee to use shady 
     16 tactics to rank better at the cost of informational quality, and so here I 
     17 stand.
     18 
     19 This leads me to this: what's the best way to find works from people like me?
     20 Distributionally, there's nothing special about my or anyone else's blog posts
     21 that couldn't be co-opted by special interests, so the solution is to not use 
     22 search engines, or at least ones that are calibrated in this way. It's possible
     23 personal search engines could work, ones that index good parts of the web in a 
     24 selective way, but this isn't well done yet. 
     25 
     26 Given this, it seems that discovery is relegated to being your own spider.
     27 To be your own spider takes effort, but it's to the point that the web is so 
     28 utterly fucking useless that I don't know what else is left.
     29 
     30 Perhaps an alternative could be some sort of self-serve common crawl to decrease
     31 diffuse server loads, but I'd imagine the kinds of entities capable of serving
     32 such an index would be the kinds to try and exploit users.
     33 
     34 I'm thirsting for knowlege; blogs like these that are actually interesting:
     35 
     36 - purplesyringa
     37 - joshua blais
     38 
     39 UNRELATED:
     40 I regret adding search to my Gemini protocol browser. It's slightly complex
     41 and utterly cancerous.