Facebook knew you were in love a long time before you did. It noticed you scrolling back through her timeline. Every millisecond lingering over the photos of her at the beach was faithfully logged.
On the surface, you two were perfectly suited to each other. But Facebook had detected a problem. At your age, it's hard for Facebook to make money from your love. Sure, a promotion for flowers earns a few bucks. Adverts for romantic dinners can bring in some cash. But here's not much money in that.
So Facebook acted. It "lost" the occasional message you sent her. It made sure that photos of her with other guys were always at the top of your newsfeed. She mostly saw your posts about drinking - and all the girls who had liked your status updates.
With perfect algorithmic efficiency, Facebook found you a beautiful wife who was practically guaranteed to produce a sickly child. Nothing too bad, mind you, but just ill enough to make you spend a little bit more than you would otherwise. A child is a joyous event. Lots of photos posted to Facebook. Lots of likes. Lots of inspiring updates about bravely struggling.This is, as the author points out, a work of fiction, though once the deep-learning algorithms are given access to all incoming data and control of the entire system, and optimised only to solve one problem (maximise profits, whilst avoiding a list of forbidden tactics that someone has thought of), there may be millions of subtly malevolent scams like this, all of them too complex for any human in a position of oversight to understand. Billions of equations, predicated on complicated models of circumstances and human behaviours, combining into scenarios which result in one or more users becoming slightly better-performing profit centres.
Please keep comments on topic and to the point. Inappropriate comments may be deleted.
Note that markup is stripped from comments; URLs will be automatically converted into links.