Attention Junction, a tool I recently created that explores overlapping public interest in topics using Wikipedia page views, has a new feature: it now suggests related topics so you can create topic pairs while not knowing anything beyond the initial topic in which you’re interested. Does this new feature use AI? Nope, it’s using the Wikipedia API and intra-Wikipedia backlinks along with some filtering.

If you haven’t seen Attention Junction, a brief description: it’s a tool to analyze the historical views of two Wikipedia pages, identify spans of public interest as expressed by unusually-high page views, find overlaps, and turn them into Google / Google News searches.

Since you’re analyzing two pages at a time, you have to have a topic pair in mind. If all you have and all you know is one name (or place or whatever) then you’re either stuck or you have to gather some extra knowledge to use Attention Junction.

This problem brings to bear my favorite mental chew toy, “How do you ask for what you don’t know?” I want to make tools that you can use starting with as little as one idea that you can use as a wedge to find other concepts and topics, then combine those into more searching to inform yourself further. Wikipedia’s structure makes that easy.

Let’s use Jeff Bezos as an example. Each topic box has a little “Find Related” link beside it. Click on that and a popup box will whir for a few moments as it checks Wikipedia, then will present you with a list of related topics. Click on the topic to add it to your topic B box and finish your topic pair.

The visible related topics here include The Washington Post, Blue Origin, and Amazon. But there were others, including Allen & Company Sun Valley Conference, which I’d never heard of. So I added that to the second topic box, searched for overlapping public interest in the last year — and found billionaire summer camp?

I reset the search to find overlaps going back to 2017 — and yeah, billionaire summer camp. Several stories about it in 2021.

Once I had the related concepts feature installed I rapidly fell in love with Attention Junction as a current events / news search tool. Google News’ current search — throw in some keywords and let AI and the algorithm do the rest — has become, to me, more and more disappointing in its results. Attention Junction feels like a much more precision instrument, letting you use inherent relevance (related topic pairs) to build your query and public interest to specify limited search spaces. It’s like you’re taking the topic pair and dividing it into date-based “slices” of public interest-based search results. Further, each “slice” of search results is generally conceptually unique. Be careful, though. The ability to quickly set up so many precise, information-rich searches can drop you down a dozen rabbit holes.

With that in mind, give Attention Junction a try — it’s free to use and free of ads. I hope it’s useful to you.