Interesting piece! On the point about not bothering to look below the first few returns in SERPs, I have noticed that I do that too--but not because I crave only conventional wisdom or the most popular opinions, it's because the *relevance* of information seems to fall off so steeply. SERP Page 3 is not full of unusual opinions about X; rather, it is not even about X at all. I am not entirely sure why Goog search outputs have this property, but it might suggest guarded optimism about market for alternatives or improvements.
A subset of AI/ML research in recommender systems and esp. computational creativity builds algos to reliably generate/find surprising information: novelty search, outlier detection, etc. Some startups have cropped up that want to address the issue you point out and use such algos to insert diversity into people's info consumption, but they have remained niche/artsy projects (like, *really* niche, not just DuckDuckGo niche), or turned recommender system – e.g. Foursquare was conceived as an urban serendipity engine by Dennis Crowley and has become a hipster brunch recommender system.
Interesting piece! On the point about not bothering to look below the first few returns in SERPs, I have noticed that I do that too--but not because I crave only conventional wisdom or the most popular opinions, it's because the *relevance* of information seems to fall off so steeply. SERP Page 3 is not full of unusual opinions about X; rather, it is not even about X at all. I am not entirely sure why Goog search outputs have this property, but it might suggest guarded optimism about market for alternatives or improvements.
A nice 2022 rejoinder on the Web 1.0/2.0/social media question whether the Internet is or can be a serendipity engine (https://www.bbc.co.uk/programmes/b039d4b4).
A subset of AI/ML research in recommender systems and esp. computational creativity builds algos to reliably generate/find surprising information: novelty search, outlier detection, etc. Some startups have cropped up that want to address the issue you point out and use such algos to insert diversity into people's info consumption, but they have remained niche/artsy projects (like, *really* niche, not just DuckDuckGo niche), or turned recommender system – e.g. Foursquare was conceived as an urban serendipity engine by Dennis Crowley and has become a hipster brunch recommender system.
I think broad-spectrum newspapers and (used) bookstores remain the serendipity epistemic agents in my agent zoo. But your post triggered in me – in good old Web 1.0 associative style – memories of two related posts by Clive Thompson: https://uxdesign.cc/rewilding-your-attention-d518ede18855 and https://onezero.medium.com/a-search-engine-designed-to-surprise-you-b81944ed5c06.