What might we have chosen to do with social media if we’d been blessed with foresight at the time of its invention? Its disruptive effects are comparable with those of the printing press, which created the concept of “the public,” a large body of people who don’t know each other but share common interests. Social media has given any member of that public who has an internet connection the ability to publish to the world.
If we could wind back the clock we would almost certainly use regulations to try to strike a balance between free speech and toxic speech, assuming we could agree on which was which. We would probably try to keep the main engines of social media in public hands, perhaps by funding the world’s public broadcasters to own and develop platforms, apply their usual editorial standards and invent new ones to fit the purpose.
Or perhaps we would choose to contain the harm by killing social media at birth. But that would mean losing its benefits: the flow of information across borders, the self-help support groups, the pleasant community chat on Facebook, the way that issues like domestic violence and Indigenous disadvantage have been made more visible.
Of course, we had neither foresight nor the necessary tools and powers when social media was invented.
In their extraordinary new book, Character Limit: How Elon Musk Destroyed Twitter, New York Times journalists Kate Conger and Ryan Mac chronicle the history of a key social media platform, Twitter. But their work also illuminates a broader narrative. Social media was founded and moulded by men-children. Yes, social media’s inventors were very clever, and they made many things we now take for granted. They seemed idealistic about the public conversation, but their naivety meant that their idealism didn’t survive encounters with reality. Fundamentally, they played with powerful things because it was fun, at least in the beginning, because it made them rich, and because they could.
Conger and Mac’s title is of course a reference to Twitter’s distinguishing feature in its early days — the 140-character limit on length. But it also points intentionally at the limitations of its creators and owners.
Mainly, Character Limit is about Elon Musk, who bought Twitter in 2022 for US$44 billion, larger than the GDP of Jamaica or Senegal. He then mismanaged the business so badly that by February this year it was worth just $11.8 billion — 27 per cent of the purchase price. In the process he damaged not only himself but also his other companies, Tesla and SpaceX.
But the story of character limitations begins much earlier, with Jack Dorsey, Twitter’s founder. Dorsey mismanaged Twitter partly through indolence and partly because he was preoccupied with his other invention, a payment platform called Square, which has helped even the smallest of traders to eschew cash.
Musk was an early chief executive of a rival platform, PayPal, before its founders pushed him out. “X” — the name he has imposed on Twitter — is a reference to his ambition since those days to create a social media platform like China’s WeChat, an all-encompassing “app for everything” through which users conduct every aspect of their lives. WeChat is a social media platform, a banking app, a telco, a broadcasting medium and one of the most powerful mass surveillance mechanisms ever invented.
Twitter was vulnerable to Musk because Dorsey had allowed it to grow without working out a way of making it pay. After Musk made his $44 billion offer the board briefly explored other options, but the company’s tanking revenue figures made it an unattractive proposition for investors. Having decided that Musk’s offer was the best way out of a bad situation, the board’s main challenge was making sure he stuck to the deal. Dorsey, no longer chief executive but still on the board, effectively colluded with Musk in secret, betraying the shaky trust of his colleagues.
Those who followed these events at the time will know that Musk repeatedly and slanderously attacked the management and board as he tried to back away from the deal, asserting that they had misled him about the number of “bots” on the platform and other issues. In truth, he had refused to sign a non-disclosure agreement during what should have been the due diligence process. He nevertheless signed documents which, subsequent events demonstrated, he barely understood.
Even for Musk, then the world’s richest man, the purchase price was challenging. To raise the equivalent of a fifth of his own net wealth without selling Tesla shares (and tanking the price of his best-known asset) he had to seek loans and partner investors, loading Twitter’s already shaky balance sheet with debt.
To tell its bleak yet fascinating story, Character Limit draws on more than a hundred interviews, along with a wealth of private tweets, text messages, emails and content from Twitter’s Slack channel. Some of this material comes from the many court filings in the forest of litigation surrounding Twitter, but other material was clearly leaked. The authors have gained extraordinary access, and the results are devastating to Musk’s reputation, and to Dorsey’s.
The result more than compensates for some irritating stylistic features — redundant personal descriptors, non sequiturs and breathless prose. For the most part this is a great story well written.
In essence, Musk destroyed Twitter by loving it too much. As one of its most enthusiastic users, he assumed everyone felt much as he did and would therefore be prepared to pay to use it. He could then shuck off pesky advertisers with their concerns about reputational damage.
He started by proclaiming himself a free speech absolutist but quickly began throttling speech he didn’t agree with and promoting increasingly right-wing views. In one of the saddest and most revealing stories in the book, Twitter engineers created a special piece of code, “author-is-elon,” to make sure that Musk’s tweets were promoted on the platform. Musk had been angered that a tweet about a Super Bowl event by Joe Biden had racked up more views than his own, despite Biden having fewer followers.
Musk’s bizarre behaviour since taking over Twitter has included a trip to Sacramento in the dead of night, accompanied by his cousins and armed with a pocket knife, to disconnect a data centre he was impatient to close down as a cost-saving measure. A series of rapid sackings meant that managers no longer knew which of their team members were still on staff. At times, the company was barely operating.
As the 2022 midterm US elections approached, Musk pushed his remaining staff to extraordinary lengths to launch Twitter Blue, the service that would allow people to buy the famous blue tick of verification. He was warned about the likelihood that impersonation and misinformation would affect the election. (The FBI was concerned enough to issue warnings and ask questions about how identifies would be verified.) Nevertheless, he pushed on.
At the last minute, after weeks of warnings, one brave executive asked Musk if he wanted to be blamed for the election outcome. Amazingly, he had to ask the date of the election. “It’s in two days,” he was told. Musk paused. “Oh, I didn’t realise.”
The launch was delayed, but nevertheless Twitter Blue was a disaster. A fake O.J. Simpson confessed to murder; a fake senator Ted Cruz declared a fondness for eating babies. More seriously, the pharmaceutical company Ely Lilly appeared to announce insulin was free, forcing the real company to put out an announcement to correct the record. Its stock dropped by 5 per cent.
Advertisers were horrified. If something wasn’t done, Nike warned, it would never advertise with Twitter again. So, after just twenty-four hours, Musk got the engineers to turn off Twitter Blue. (It was later relaunched, but never got the number of takers Musk had predicted.)
Executives were meanwhile leaving rather than risking criminal penalties because they were required to sign off on audits they had no way verifying thanks to the staff cuts and the resulting chaos. As Musk abandoned offices to save on rent, and at the same time insisted employees work in the office, facilities broke down. In New York, the stench of overburdened toilets overwhelmed the office and cockroaches flitted in and out of the drains. In San Francisco, employees took to running to nearby cafes to use the bathroom; if they wanted to use the ones in the office they had to bring their own toilet paper.
As Twitter lurched and sometimes broke, Musk ran a poll asking users whether he should step down as chief executive, promising to abide by the results. When 58 per cent of more than 17.5 million accounts called for him to resign, it brought on a mental health “episode” that some described as “the throes of a manic event.”
All this raises the question of how Musk was able to succeed with Tesla and SpaceX. Although the detail of his other businesses is beyond the scope of Character Limit, Conger and Mac nevertheless offer an explanation. Tesla and SpaceX were built in Musk’s image, they say, but with people and processes in place to manage the boss and his shifting demands.
Musk simply didn’t intervene in them as much, or in the same way. He loved Twitter — sometimes tweeting “like a twelve-year-old off his Ritalin.” He spent his days fixated on the company, monitoring his own tweets and those of others in real time. “There was no buffer, just him banging his head against the wall day after day to solve a business problem he had created for himself.”
The authors conclude that Musk has turned Twitter — or X, as it is now called — into “a harsher and much more cynical” social media platform. “One of the most important modes of global communications has become practically unrecognisable and now serves the interests of one man,” they write. “What was once called the digital town square is becoming Musk’s mirror.”
Describing Twitter as one of the most important modes of global communications might seem like an overstatement. After all, it has never been the world’s biggest social media platform, nor anywhere close. That achievement is Facebook’s, followed by Google’s YouTube and Facebook’s Instagram.
But Twitter was important as a hub of political speech — a place where journalists, politicians, academics and those who followed them could exchange information and argument in real time. It was also a powerful reporting medium for journalists.
Long before Musk was on the scene, Jack Dorsey was changing Twitter’s maintenance schedule to make sure it could be used by demonstrators in Iran’s Green Revolution. Then National Public Radio’s Andy Carvin used Twitter during the Arab Spring to communicate with protesters in the Middle East and verify eyewitness accounts from the front lines, mostly while he was on his iPhone in the United States. It seemed like the dawn of a new kind of reporting that could tap into the wisdom of the crowd and become self-policing and self fact-checking.
I used Twitter myself to report from Melbourne’s Flemington public housing estate when the state government locked down its 3000 residents without warning during a Covid outbreak in July 2020. Prowling the estate, I was receiving tips from the residents locked in their flats, first-person accounts from police via direct messages, and inside information from public servants. Thanks to what the media and communications academics would call the affordances of Twitter, I was able to serve all this up in 140 characters or less as a constant, community-engaged and -informed news feed.
Some things changed for the better as a result of those tweets. Within the community of locked-down people, the police and public servants charged with confining them, the neighbours and the families a conventional newspaper article would not have had anywhere near the same impact. It was the minute-by-minute reporting and community commentary that lent my work its power. Only social media could have done it. In fact, I think only Twitter.
Long before Musk came on the scene, people inside Twitter were struggling to strike a balance between enabling free speech and reducing harms. Conger and Mac detail insiders’ heroic efforts to decide when to label unreliable tweets by public figures or even to ban users altogether. Attempts were made to automate regulation, but the most important decisions had to be made by real people in real time.
On election day 2020, Twitter had a team tracking misinformation and voting results around the clock. As it became clear that Biden was going to beat Trump, nearly 40 per cent of Trumps’ tweets received labels warning that their content “might be misleading about an election or other civic process.”
Later, as rioters approached the Capitol on 6 January 2021, Dorsey took a call from Vijaya Gadde, his heroic corporate lawyer and the woman responsible for making the tough judgements. Trump was tweeting up a storm, and each tweet seemed to be manifesting in real-world violence.
Dorsey was on Tetiaroa island in the middle of the south Pacific. Gadde, who had been sleeping in the office as she and her team wrestled with events and Twitter’s role in them, was recommending that Trump’s account be suspended. But who were they to censor @realdonaldtrump, a sitting president? Could they? Should they?
Dorsey eventually agreed to allow Gadde to lock Trump’s account for twelve hours. Even world leaders, he tweeted, “are not above our rules entirely and cannot use Twitter to incite violence.” Trump was eventually banned from the platform for repeatedly violating its rules.
Dorsey signed off on these decisions but simultaneously hung his executives out to dry. “A ban is a failure of ours ultimately to promote healthy conversation,” he tweeted — but whatever made him think they could do that?
The Trump ban “broke Dorsey,” write Conger and Mac. “Twitter wasn’t an idyllic, free and open town square… It was like he was the kid who built the robot, which then destroyed the world.” Gadde, meanwhile, had to have security guards at her home as “free speech enthusiasts” blamed and doxed her.
Character Limit makes clear that is nearly impossible to patrol a social media platform for harm in real time, even with automated technologies and thousands of employees and content moderators. Regulators should know this. If they don’t, they should read this book.
Despite all the bad decisions and the chaos, some important ideas, and even possible solutions, emerge from the Twitter story.
At one stage, just before the Musk takeover, then chief executive Parag Agrawal (chosen and then betrayed by Dorsey) came up with an important concept. Speech should be free, he said. It was a human right. But “reach” was a privilege. It should be earned.
He developed “Project Saturn,” named after the planet’s concentric rings. Users who adhered to Twitter’s guidelines and standards would occupy the inner circle, their Tweets privileged and promoted by the algorithm. Poorly behaved users would be relegated to an outer circle, with less promotion. All this would be transparent: users would be told why they had received their ranking, and how that could improve it.
Under this system, Trump’s repeated breaches would have put him in the outermost circle at the time of the Capitol riots. His tweets wouldn’t have been shared or promoted. People would have had to seek him out rather than finding him in their feeds. But he wouldn’t have been banned.
Thanks to Musk’s takeover, Project Saturn never developed beyond a concept. Yet surely this idea — free speech as a right, reach as a privilege — remains an important conceptual distinction as we wrestle with the consequences of the all-pervasive capacity to publish.
The great benefit of social media is that conversations previously confined to a few individuals become visible, potentially globally. And that’s also its great downside. Toxic players, braggarts, mad people and schoolyard bullies are amplified and empowered. And that’s before you get to child exploitation material and violence, both of which litter X.
As it stands, it is not clear that social media can be controlled — even by the people like Musk who own it. They are indeed destroyed by their own robots.
But we are not determined by technology. We create it, and we should decide how to use it. In the coming decades, a firm conceptual grasp of why we value free speech, and the relationship between speech and reach, is surely part of the way forward. •
Character Limit: How Elon Musk Destroyed Twitter
By Kate Conger and Ryan Mac | Cornerstone Press | $36.99 | 480 pages
