3.5
challenging reflective

A decent exploration of the subtle and not-so-subtle ways that algorithms manipulate individuals and cultural trends in the US and abroad.

A lot of the book is devoted to calling out unoriginal, shallow creations inspired by social media or profit-squeezing analytics: endless Marvel movies, soulless coffee shops, mediocre Book Tok reads, etc. Ironically, there are no groundbreaking observations there. 

However, it has some fascinating details about how cleverly designed big tech’s algos can be. Netflix will show different thumbnails based on user content preferences. For the show Outer Banks, romcom fans see main characters about to kiss, while sports fans see surfing.

Some social media algos are deeply flawed from a user experience perspective. I don’t think the author does a great of explaining this, however, so I’ll add some my own analysis.  Twitter and Facebook (and, as of a few years ago, YouTube) don’t have dislike functions, so absurd hot takes, rage bait, and bots are free to run rampant. Reddit has downvotes, so bots are usually hidden. Controversial content is buried by default, but it’s available in a separate tab if you’re curious about it. Twitter’s bot engagement led to a controversy that nearly resulted in Elon pulling out of the deal to buy it. For Facebook and other platforms, businesses are buying ads based on inflated impressions figures. Is there an influencer bubble if a large portion of their metrics are fake? 

Chayka’s solutions for big tech’s adverse impacts on society are somewhat questionable. The FTC/antitrust angle is far easier said than done; their brilliant idea for regulating Google is spinning off the search business. I think the best path forward is for parents to ensure their children don’t spend their critical years of socialization online. This could be easily accomplished with firewalls or screen time restrictions. Sure, at first children feel excluded from their friends’ online hangouts, but collective action can gradually shift the norm to in-person interaction.