default_search_engine_wikipedia.md — paste.sr.ht

3 min read Original article ↗

#Default search engine: Wikipedia.

  1. I propose that Wikipedia is the most reasonable default search engine remaining as of this writing.
    1. Note that this is about a reasonable default. I am not claiming that Wikipedia is the best search engine.
    2. I switched my own default to Wikipedia a month ago and I've decided to keep it that way.
  2. Traditional search engines are becoming LLM frontends.
    1. Unfortunately this trend has infected even DuckDuckGo and the likes.
    2. Simple searches are contributing to significant carbon emissions.
    3. The non-LLM search results are also mostly SEO farm pages, and increasingly LLM generated slop parading as legit content.
  3. Most general searches can be answered very well by Wikipedia.
    1. I had to only slightly alter my search behavior.
    2. Instead of searching "sources of iron", I've tuned myself to search for "iron" and then Ctrl+F for "nutrition" which takes me to https://en.wikipedia.org/wiki/Iron#Diet, which I'm pretty sure is as good information as I'm going to get from the SEO farm site/AI slop that Google will serve for the original query.
    3. Similarly, "sata m2 dimensions" becomes just "sata m.2" which takes me to https://en.wikipedia.org/wiki/M.2 and Ctrl+F for "dimensions" gets me an almost authoritative answer.
    4. A lot of information has citations with archive.org links for further reading. I've faced much less link-rot frustration due to this.
  4. For non-general searches, I still prefer looking up more direct sources instead of going through google search.
    1. For location/nearby shops, I search google maps directly.
    2. For product information, I search/navigate the vendor's website directly (which you can find by searching for the vendor on wikipedia). Or even amazon or your local retailer's website.
    3. For tv/movie shows, I search IMDB. (I actually maintain a local index from their data dumps and search that first, which works 99% of the times).
    4. Most of these can be added as keyword searches in your browser or are even already present. It's just about altering your habit to not use google as the middleman for every single internet search/navigation.
    5. For programming related information, I'm using Zeal/man pages/searching on github directly etc.
    6. For re-surfacing something I saw earlier, I search my browser history. I also bookmark/archive things more liberally now.
  5. When it doesn't work, you can still use Google/etc., it's just another search keyword in your browser.
  6. In the same manner as LLMs give people a false sense of increased productivity and promotes laziness, I think current "internet wide search" gives us a false sense of confidence about the search results and prevents us from looking up authoritative/verified information. (Again, I'm talking about current search engines, not from over a decade ago).