Solving For Twitter
Elon Musk is going to learn the hard way that Twitter is a political and sociological problem, not an engineering one.
Whatever else you want to say about Elon Musk, there is one undeniable fact in his favor: he is the first person to become the richest person in the world by deliberately setting out to reduce existential risk for humanity. For most of the past two centuries, the people at the top of the economic ladder–the Rockefellers, Vanderbilts, Carnegies, Fords—made their money in carbon-heavy industries that exacerbated our long-term risks, even if the titans themselves weren’t fully aware of the magnitude of the damage they were doing. Bill Gates, of course, has devoted the second act of his career to reducing many important risks around the world—from global warming to pandemics—but the company that made him, for a time, the world’s richest man was at best indifferent to these kinds of long-term threats. To be fair, the fact that Musk has been able to amass such a vast fortune through a company whose primary goal is decarbonization may also be a sign that something fundamental is shifting in what public markets are valuing these days. But clearly Musk deserves an immense amount of credit for what he pulled off with Tesla—even if some of us disagree on the question of whether any one person deserves to accumulate that much wealth.
I mention all that as a preface to the second thing that should be said about Elon Musk, which is that his talents–impressive as they are—are not at all suited to solve Twitter’s problems. Musk likes to talk about how he runs all his businesses from “first principles”—starting with what the laws of physics will let you do, and building out from there—which is apparently an excellent strategy if you are making rockets for a living, or electric batteries. But the things that make Twitter such a hard nut are not “first principles” problems. The rallying cry of free speech sounds like it should be a law straight out of Newton, something fixed and undeniable, but in practice—like so many of the variables at play in social media—it turns out to be more like quantum mechanics: murky and unpredictable, more probabilities and gray areas than absolute truths. Twitter is a political and sociological problem—a problem rooted in conflicts over values—not an engineering problem. It requires a different set of skills.
The other day, I noted in a conversation on the Next Big Idea podcast with Eli Pariser and Rufus Griscom that I’d been unimpressed by Musk’s performance on the TED stage a few weeks ago. If you’re going to basically declare that Twitter is a sacred public good, the world’s town square, you’ve got to at least acknowledge—and presumably have an argument to justify—the fact that there’s something just a little troubling about the richest man in the world buying the town square. I mean, you don’t have to be a Leninist to think that might not be the best way to organize things. But the objection didn’t seem to cross Musk’s mind. To this date I don’t think he’s ever really addressed the critique. I imagine he would argue something he’s argued in the past in other contexts, which that is the super-wealthy—in some cases at least, certainly in his—have a track record of allocating capital wisely, and thus it makes sense for society to give them the latitude to continue investing outsized amounts of capital on other ventures. I’m not sure I buy that argument generally—most billionaires seem to be awfully attached to allocating their capital towards mega-yachts—but with Musk I would certainly agree that he’s outperformed on that metric so far. (As Farhad Manjoo wrote earlier this week, “As billionaires go, there are many I worry about much more.”) But does that track record mean that he’s a million times better than some random other human, or better yet, a representative group of stakeholders deliberating over how best to allocate capital, or build a social media platform? I don’t think so.
The other thing that troubles me about Musk’s “only I can save Twitter” bravado is that the guy has more important things on his plate right now. If Twitter stumbles along in its current state, underperforming on Wall Street, enraging everyone at different moments, but basically serving its function as a nexus for rapid-fire news, commentary, and goofing around—that’s not a terrible outcome. It’s probably a good outcome, given all the alternatives. But if Tesla fails, or goes sideways for the next ten years, because Musk is distracted by an endless firestorm at Twitter, that’s a big deal.
These are my main objections to the Musk acquisition, but I should mention two additional complications. A few days ago Shira Ovide wrote a smart column in the Times that laid out some of the argument above regarding the complexities of free speech and content moderation generally. But at the end of the piece, she included this line:
One of the paradoxes of the Silicon Valley revolution is that it disempowered old gatekeepers of information and persuasion such as media tycoons and political leaders but created new ones. Mr. Musk’s purchase of Twitter won’t change that. We may not want these digital media barons to have so much power, but the reality is that they do.
Now, to be clear, I’ve been on the record since at least 2012 raising alarms about the fact that Mark Zuckerberg owns more than fifty percent of a social media empire that has more users than the entire Web did just a decade before. So having another single individual controlling yet another important platform only makes things more problematic. But we should also remind ourselves that the “return of the gatekeepers” narrative is not quite as tidy as it might seem. During the heyday of mass media—say in the 1960s—every single word printed in the New York Times or uttered on all the network news broadcasts was directly overseen and authorized by no more than a few dozen people, almost all of them white men. You could fit that group of gatekeepers into a large conference room, I’d wager. But today, even with the consolidation of wealth and power of the “digital media barons,” the overwhelming majority of the information shared on Facebook or Twitter circulates without any employee of either company ever even seeing it—much less the digital media barons at the top. Yes, social media algorithms can promote or demote certain voices, and those algorithms are shaped by decisions made by executives like Zuckerberg, and potentially Musk. But in terms of the actual flow of information, we are not living through a “meet the new boss, same as the old boss” situation. The titans today aren’t really gatekeepers any more; they’re more like herders, trying to steer an unruly mass of independent agents in a particular direction, often ineffectively. Even the algorithms are limited in their impact, particularly on Twitter, where the main “decision” of what gets promoted is made collectively by the users themselves through retweets and likes.
Speaking of algorithms, Musk has said cryptically at several points that he intends to “open source the Twitter algorithm.” I’m not sure what that means in practice, but if that line of thought ultimately leads him to release the Twitter social graph–the map of all the connections you have to other Twitter users–as some kind of new open protocol, that would be a potentially positive outcome. Last week, Eli Pariser posted an excellent long thread of concerns about the Musk acquisition that included this tweet near the end of it:
An open source version of the Twitter social graph would offer a significant mitigation of the risks posed by our digital media barons. If Musk runs Twitter into the ground, we’d be free to migrate to some other platform, without having to rebuild our networks, follower by follower. Of course, as I discussed with Eli on the Next Big Idea podcast, that also creates a world where you might see more “filter bubbles” as people flock to the social network that most closely mirrors their worldviews. Just think about the beleaguered right-wing network Truth Social: it would have a much easier path to success if it could be built on top of the existing Twitter protocol—particularly if Donald Trump were to be reunited with his 80 million followers thanks to the open graph. On the whole, I suspect having a social graph that allowed this kind of interoperability would be a good thing, but undoubtedly it would create a whole new set of challenges. There are no easy fixes here, no first principles that we can fall back on.
One additional note: Almost a year ago, I started talking with the Polish company Netguru about a new publishing initiative called “Hidden Heroes” that would explore a different part of the tech world and its history, focused less on the superhero CEOs and more on the crucial tradition of folks working outside the venture-backed startup mode: in open source communities, academic research, or government-backed initiatives. The first installment launched last week, a profile of the MIT scientist Pattie Maes, who with some of her grad students in the early 1990s developed many of the core ideas about algorithmic recommendations later commercialized by companies like Netflix and Facebook. Old-timers who were around for the early days of the Web will find the profile a pleasant walk down memory lane I suspect. And it includes some thoughts at the end about Maes’ concept of “remembrance agents”—a software interface that still hasn’t been built despite its obvious utility. I’ve got some of the rest of the series mapped out, but if there are other under-appreciated trailblazers from the early days of the digital age that you think I should profile, I’d love to hear your suggestions. A new profile will go live every month for the rest of the year.