marginalia search bets on text first web

see also: Latency Budget · Platform Risk

Marginalia launched as a search engine that favors text-heavy sites and downranks modern SEO-heavy design (Marginalia). The project is less about scale and more about taste in discovery. I read it as a reminder that search quality is partly a curatorial choice.

constraint map

• Small crawlers cannot index the full web, so curation substitutes for breadth.
• Ranking models reflect values, and text-first bias is a deliberate tradeoff.
• Spam pressure will always push results toward gaming unless the incentives shift.

counter-model

A niche search engine can be too small to matter and still rely on the same infrastructure it critiques. That is fair, but the value is in offering an alternative signal, not in replacing Google. A smaller index can still steer the conversation about what search should reward.

linkage anchor

This pairs naturally with Google Search Is Dying and Google Search Quality Declines as a practical counterpoint. It also complements RSS Brain as another bet on human-scale discovery.

my take

I want more experiments like this. A single dominant ranking model is a cultural risk, not just a technical one.

linkage

linkage tree

• tags
  • #search
  • #web
  • #indie
  • #2021
• related
  • [[Google Search Is Dying]]
  • [[Google Search Quality Declines]]
  • [[RSS Brain]]

ending questions

What ranking bias would I personally choose if I could design a search engine?

marginalia search bets on text first web