< back to notes

How to find stuff on the web

A description of how I navigate the web, the tools I use to do so, and my motivations.

Why am I writing this?

Mainstream search engines have been changing in recent years to try to maximise "engagement" (and by extension, profit), while making the seach results less useful. So-called "AI" chatbots have been added to the engines to supposèdly provide direct and accurate answers, but they frequently hallucinate things that don't exist, and absolutely should not be trusted to give accurate answers to questions.

Meanwhile, in comparative obscurity, there have been great developments in alternative search engines, and Wikipedia has become extremely accurate. The "small web" movement of creating personal websites like in the old days™ has become quite popular while completely flying under the radar of mainstream search engines.

Answering really simple questions, or getting general information

If you're looking for the answer two a really simple question, e.g. "when was Tom Scott born?", don't search for the question. Instead, directly search Wikipedia for "Tom Scott" and (if the answer is known), you'll find the information right there.

If your question is "What is a [word]" or "What does [word] mean", use Wiktionary and you'll get a dictionary definition, and often a link to the Wikipedia page in case you want to know more.

If you're looking for an official homepage of something well known, first search for it in Wikipedia and then click the link to the official website. Alternatively, use a general search engine and click on "official site" in the knowledge panel. (this information is often provided by Wikipedia or Wikidata so is typically more accurate than the search results.)

If you're looking for a site you visit a lot, you don't need to use a search engine! Just start typing the name and it should be suggested from your browsing history (you may need to configure your browser a bit to get this behaviour to happen.)

If you're looking for general information about a subject, search for the subject's name directly in Wikipedia!

If you're playin a particular video game and find yourself referring to its wiki (by the way, many games have independent wikis that are not hosted on big sites like Fandom.com), you can add the wiki's search as a search engine in Firefox, and access it directly from the search bar.

Answering more complicated questions

If your question is more complicated, such as "how to disable the disappearing scrollbars in GTK3", use a general search engine.

Finding recipes and reviews

If you use a general search engine, you will find a lot of heavily search-engine-optimised crud. (you may even get entirely "AI"-generated recipes that have never been tested by a human!) To avoid this, you can use an alternative search engine such as Marginalia.

For example, when searching Marginalia for "lentil pate", the first three articles (1) (2) (3) are no-nonsense pages where the bulk of the page is taken up by the recipe, whereas with more mainstream search (even using SearXNG!), two of the first three articles (1) (2) (3) started with a pointless sensationalist story about the origin of the recipe (or whatever it was - I just scrolled right past it all).

If you're looking for reviews and opinion articles, Marinalia search is quite good at finding articles on small websites that are sometimes really thought-provoking!

Recommended tools

General purpose search:

• DuckDuckGo (no AI) is a privacy-respecting front-end for Bing. It attempts to remove "AI"-generated images from search results, and removes "AI"-assisted answers. When I can't get a response from my Searx instance, this is what I use.
• SearXNG is a volunteer-hosted search aggregator, that takes results from various search engines, including Google, Bing, Brave, Quant, and many more. There are many instances run by different people, and they vary a bit in the quality of the results, so it may take a few tries before you find a good one. It is also highly configurable - there are many different search engines you can enable or disable. If you're good enough with tech to switch to something like DuckDuckGo, it doesn't take much more skill to switch to SearXNG, if you're willing to put in a little bit of work to find a good instance.
• Startpage is a privacy-respecting front-end for Google and is pretty decent.
• Kagi and MetaGer are paid-for aggregators. I can't vouch for their quality.

Alternative search (when you need to avoid the SEO slop):

• Marginalia search uses a home-grown algorithm, has an open-source backend, and is funded by donations. (as of 2025, it operates at a cost of around $200 per month.)
• Wiby is a search engine that indexes primarily personal websites; all search results are pages that have been submitted by humans. It's good for finding random cool internet people.
• Eighty Eight Thirty One provides a graph of websites that link to each other using 88x31 web buttons . It can show all the personal websites that link to a particular site, and you can find internet people who are interested in particular things. It's also good if you just want a web button.
• Kagi claims to be good at this, but again, I can't vouch for its quality.

Configuring your browser:

• By the way, I would recommend that you use Firefox or one of its many derivatives (LibreWolf, IronFox, Zen browser). Also, you should install uBlock Origin (not the same as uBlock); it is the best adblocker.
• On a wiki, you can add a custom search engine by right-clicking the URL bar and selecting the option to add a search engine. If this doesn't work, you can add a search engine manually by visiting the "search" section of the Firefox settings (to get the search URL, search for something unintelligible and copy the URL of the search results page). Once you've added it, find it in the table in the search settings and double click to add a keyword (e.g. "@hollowknight" for the Hollow Knight wiki).
• I prefer turning off search engine suggestions as well; I find having mainly the browsing history is more helpful.