“Tinder and similar apps were responsible for essentially engineering about 10% of the dates that took place in the countries where the app was used. The idea of a dating app that works ‘too well’ and ends up in a long term relationship was a serious risk for the apps’ business models, or so Gupta argues. Apparently companies decided that for their purposes it was better to use machine learning and AI to engineer the algorithm to bring people together who were likely to feel attraction, but whose relationships were also likely to be volatile. I started thinking about what might happen if a child was conceived in one of these short and sharp relationships. That child’s biology would be, in part, algorithmically determined.”