I did not set out to write a bot that writes near-future late-capitalist dystopian microfiction. I set out to write a bot that automates a particular kind of joke. But if we pause to consider the bot’s algorithm, it’s obvious where this tendency toward a very specific fiction genre originates.
The Google News sidebar described in the email thread above is Google’s attempt to parse out the subject of a bunch of related news headlines. For example, if there is a bombing in Iraq and there are a lot of news headlines about it, it will probably generate the subject “Iraq.” This is a very specific choice: it could have equally chosen “bombing” or “terrorism” or “chaos”, but Google’s algorithm tends to favor named entities over abstract concepts. What this means is that the subject of the news, as Google sees it, is almost always a corporation, a sports team, a celebrity, a nation, or a brand.
My algorithm builds its jokes by harvesting these subjects that Google has picked, and swapping them indiscriminately between headlines.
What is near-future late-capitalist dystopian fiction but a world where there is no discernible difference between corporations, nations, sports teams, brands, and celebrities?
So Adam was partly right in our original email thread. @TwoHeadlines is not generating jokes about current events. It is generating jokes about the future: a very specific future dictated by what a Google algorithm believes is important about humans and our affairs.