With Elon Musk’s $44 billion acquisition of Twitter hanging in the legal balance, bots—automated twitter accounts that, in Musk’s words, inspired him to purchase the platform and then, later, gave him an excuse to back out of the deal—have been at the center of attention.
The focus has mainly been on spam bots that harass users or try to sucker them into scams, often related to cryptocurrency. Still, legal experts are skeptical that arguments over the precise ratio of bots to active human users will scuttle the deal. But in the controversy, something important got lost: Bots are great! I like bots.
Twitter has been a product driven by its users and third-party developers, from the invention of the retweet to the many third-party clients that have lived, died, or been bought by the company, such as Tweetdeck. The decision early on to make the platform’s functionality available to developers through an API also spurred the creation of many automated accounts designed to benefit users, not defraud them.
Twitter executives call them “good bots,” and while they wouldn’t share an estimate of how many are on their platform, they recently allowed developers to opt-in to a label that highlights “automated accounts.”
While everyone’s mileage varies in how they use Twitter (please stop following people who make you angry), good bots allow a judicious user to add utility, serendipity and, of course, memes, to their feed, rounding out the melange of human maniacs who provide deranged takes that no AI could replicate.
There are useful bots: The SF Quakebot keeps me abreast of seismic activity in my area, there’s a feed to alert you if the editors at the New York Times decide an all-caps headline is merited, or a tracker for vaccination progress in the US, still sadly not at 100%. One frequently mentioned bot collates long Twitter threads so you can read them at your leisure. This feed featuring words first used in the New York Times is a surprisingly useful window into the zeitgeist.
One of the more professionally important bots for me may have played a role in Musk’s decision to purchase Twitter. The Elon jet tracker bot uses publicly available aviation data to keep tabs on Musk’s private airplane.
As a reporter who writes about Musk, it’s helpful to know where the peripatetic entrepreneur’s jet is. Musk himself did not like the tracker and tried to have it shut down in January, shortly before his Twitter take-over began. The concept has expanded to include other celebrities.
If you like comics, well, you can follow the adventures of Nancy, or Jucika, a Hungarian comic that ended its run in the 1970s. Pure surrealism is available, in text form, as emoji, or with the help of Garfield. Quartz once employed an AI reporter who offered important commentary on the tech sector. Perhaps my favorite bot is one that offers you a train crossing a random emoji landscape.
The good bots are easy enough to make. Twitter spokespeople tell me that building an automated feed is often the modern equivalent of convincing a computer to say “hello world!” for novice developers. The ease of use may be one reason that bad bots can proliferate on the platform, though the company says it deletes one million (!) spambots each day. As an obsessed Twitter user with a moderate following, most bots I see are ones I choose to follow. While Twitter wouldn’t address this question directly, one suspects Musk’s experience as a massively followed troll creates a different experience.
Twitter’s current management is all-in on third party developers, making not just bots, but apps and tools that can add to the platform’s value (and revenue). That includes tools to block the bad bots. Amir Shevat, the executive in charge of Twitter’s developer platform, used to make apps for the platform himself as a third-party developer. In November, for the first time since 2010, it will hold a conference for developers called Chirp. If Musk is forced to make good on his promise to buy the platform, it will be interesting to see if he backs this strategy.