Back in 2017, when Elon Musk was under great pressure to get Tesla on firmer footing after a series of disappointing earnings, he took to Twitter, where he showed a great deal of vulnerability — highly unusual for a tech overlord like himself.
“The reality is great highs, terrible lows and unrelenting stress,” he tweeted. “Don’t think people want to hear about the last two.”
His ability to articulate the emotional challenges of high-level innovation is why it’s always been so interesting to interview Musk. He goes there, as they say, in ways that are sometimes funny, sometimes juvenile and always revealing, even as he pulls off what are clearly astonishing feats of digital wonder. Musk, through Tesla, has single-handedly pushed the electric vehicle sector into the mainstream and, with SpaceX, has moved space tech into a new age with reusable rockets.
In this way, Musk would be the perfect person to transform a sad violin of a company like Twitter, a serial underperformer that has become the Silicon Valley’s little engine that couldn’t. For Twitter, leaving the company in the hands of a creative mind with near-limitless resources and a clear enthusiasm for its product to possibly remake it as a private entity is perhaps the best outcome. (That is if the deal even goes through. The severe downturn in the stock market has made Musk’s $54.20-per-share offer far too high, and he’s indicated he’s re-evaluating the deal, potentially for less money.)
Many top entrepreneurs genuinely respect Musk’s tech chops. One told me he walked away from an hourslong conversation with Musk a few weeks ago “stoked” about how he might transform and fix Twitter. “If anyone can do it, when he is at his best, Elon can,” he said.
When he is at his best is key, because there is also a more disheartening side to the man, as the very voluble Musk appears to have descended over the past week into becoming what I can only describe as a chaos monkey.
For those not familiar, that is actually the name of a piece of software made by Netflix that it called “a resiliency tool that helps applications tolerate random instance failures.” In other words, it aims to throw stuff haphazardly into a system to test its robustness.
Speaking of which, in response to a very cogent response by Twitter’s chief executive, Parag Agrawal, to counter Musk’s claims about higher-than-advertised bot activity on the service, Musk flung a poop emoji at him in a tweet.
A man in the throes of pulling off the most unlikely deal in tech with savvy Spock-level chess moves lowered himself to presenting a digital piece of feces to the man he would be replacing. For what reason other than to amuse himself and his legion of rabid followers?
I, like others, suspect the reason for Musk’s most recent erratic behavior is to finagle a lower price, as Twitter has dropped by about 25 percent since he reached agreement on a deal late last month. Still others have posited that he is even looking to slither out of the deal and his $1 billion breakup fee altogether, as Tesla shares undergirding the acquisition have hit the skids amid the deal talks, falling to nearly $700 this week from about $1,000 when Twitter accepted Musk’s offer.
If it’s all negotiating tactics, then perhaps there’s an endgame to Musk unleashing a series of seemingly unrelated incendiary tweets, including that he was going to vote Republican since Democrats who were “(mostly) the kindness party” are now the party of “division & hate.” That was preceded by a more paranoid one that posited that “political attacks on me will escalate dramatically in coming months” and another saying that the “dirty tricks attacks will be next-level,” leaving out the pertinent fact that next-level is in fact a G.O.P. specialty.
In an exchange with a Yale medical school professor, Musk said the school was “the epicenter of the woke mind virus attempting to destroy civilization.”
OK then. Let me say that a shift in his political identity is Musk’s to make and his alone, even if you don’t agree. Even if he wants to bearhug Donald Trump and don a “Make America Great Again” hat, that’s his choice. But isn’t it a curious coincidence that it is also sure to draw attention away from that clear math problem that has affected his Twitter purchase?
The problem for Musk is that while politicians often lie, numbers don’t. Rage tweeting or no, it hardly matters since a deal is a deal. After Musk claimed the transaction could not “move forward” until the bot numbers were clearer, Twitter fired back: “The board and Mr. Musk agreed to a transaction at $54.20 per share. We believe this agreement is in the best interest of all shareholders. We intend to close the transaction and enforce the merger agreement.”
Focus on that word “enforce,” which translates to legal action, sure to create even more problems for both Twitter and Musk. Does he care? Do chaos monkeys care about creating chaos? Sigh. One would hope that such a great mind would not engage in such peanut-seeking antics.
It’s also a shame. He was impressively self-aware back in that stressful time in 2017, when someone on Twitter asked whether Musk suffered from bipolar disorder. He first said, “Yeah,” but then wrote: “Maybe not medically tho. Dunno. Bad feelings correlate to bad events, so maybe real problem is getting carried away in what I sign up for.”
“Carried away” certainly sounds familiar, as does what he tweeted one minute later when he added: “If you buy a ticket to hell, it isn’t fair to blame hell …”
No indeed, it is not. So, what’s going to happen next? Dunno — I guess you have to ask the monkey. Just remember to duck.
Renée DiResta is the technical research manager at the Stanford Internet Observatory and a well-known expert on disinformation. I’ve edited her answers.
Give us an overview of the current situation around hate speech proliferation on social media. Despite the companies’ vows, it feels like nothing has changed since the New Zealand shooting in 2019.
I think we’re all feeling discouraged. Yet again, a young man found his way to toxic online communities and content, became radicalized and committed a brutal act of violence that took the lives of innocent people and irreparably harmed a community. There are many important conversations happening about what caused this — the toxicity of racism, the prevalence of white supremacist content online, the ease of access to guns despite red flags — and yet across these conversations, there is a sense of intractability. Still, there have been some positive changes since the New Zealand shooting. For example, entities like the Global Internet Forum to Counter Terrorism, a nongovernmental organization tasked with preventing terrorists and violent extremists from exploiting digital platforms, now have significant buy-in from both the tech sector and governments worldwide. Major tech companies take steps to detect and moderate extremism. Civil society and research organizations attempt to expose and mitigate online radicalization. But the niche platforms where much of the hateful content spreads, where radicalization often occurs — 4chan and its ilk — are, in fact, largely unchanged.
What have the platforms done to improve the reaction in a crisis such as this? What needs to be done?
The Christchurch shooting happened on March 15, 2019. At the time, GIFCT existed, but it had been organized primarily to deal with ISIS — specifically, by creating a unique digital fingerprint for terrorist content, known as hashing, and putting it in a shared database to enable platforms to find and take it down more quickly. At the time of the New Zealand shooting, there was not much clarity on what kind of terrorist content fell, or should fall, within its purview; there was some concern about false positives, who should decide what constituted “terrorist content” and how to handle content related to acts by extremists who were not part of an organized terror group.
Following the tragedy, in what came to be known as the Christchurch Call, the tech industry and governments worldwide committed to “eliminate terrorist and violent extremist content online.” GIFCT created the Content Incident Protocol, which it activates when a mass violence event has happened. The process involves hashing related content and collaborating to ensure it is taken down as quickly as possible. GIFCT activated the Content Incident Protocol on May 14 after the attack in Buffalo.
But the internet is full of people, some of whom have broken moral compasses, and so the process of stopping the proliferation of terrorist content is adversarial: Those people try to put the content back up and to evade detection. In this particular case, although Twitch pulled the stream quickly, users on 4chan collaborated to archive and reupload it. Also, there have been both a proliferation and an expansion in use of smaller platforms, which may have under-resourced or deliberately lax approaches to content moderation. If you were to go looking for videos related to the Buffalo tragedy, you could find them in a few seconds on some of the alt-platforms and small websites. Google removed the manifesto content from Drive, but it was reposted to niche services. The major platforms play Whac-a-Mole with content shared from these smaller hosts.
You asked, what needs to be done. Preventing the sharing of content is reactive. We need to be proactive, as a society, at preventing these atrocities from occurring.
What are the up-and-coming threats you are most worried about?
We have a crisis of trust and a loss of confidence in institutions that is not caused by social media, though the overall information environment contributes to it. There is a feedback loop happening; distrust is continuously reinforced, including by false or misleading claims from incentivized hyperpartisan media and influencers. Any attempt to label or downrank even the most blatantly wrong posts or most committed, recurring manipulators is presently processed as “censorship.” Who should be the arbiter of truth? Who watches the watchmen? We are seemingly trapped in a crisis of legitimacy at all levels of society that no one has the moral authority to disrupt.
While we are not in a virtual world as yet, how will that be different from the current challenges facing social media companies?
Well, first of all, it seems we won’t have legs in virtual reality, because it’s too complicated to implement. But beyond that, challenges specific to real-time moderation are most likely to carry over into the virtual world. Issues common to voice-based platforms like Clubhouse or gaming platforms are probably more relevant than those of text or posting-based platforms. Giving users highly granular controls that can help them better set their own boundaries and shape experiences will be far more relevant. Nick Clegg, Meta’s global affairs chief, recently compared the challenge of moderating VR to that of deciding whether or not to intervene in a heated argument in a bar. So, we’ll be in a bar, legless, establishing norms as we go.
Tech’s Struggles Hit Jobs
Only a few short months ago it was a seller’s market when it came to talent in tech as highly sought-after employees flexed their muscles by demanding flexible work arrangements and big pay bumps, and seeking and landing as much start-up investment as needed. Now, as I wrote last week, all those doors are closing with a simultaneous downturn in the economy, the stock market and venture funding. After an incredible 13-year boom comes the inevitable hiring and pay freezes and even layoffs.
Among the bigger companies, Netflix has been the first with those, as weaker growth and a more robust set of competitors has sent its fortunes into a tailspin. The streaming pioneer laid off 150 employees and dozens of contractors as well. “As we explained on earnings, our slowing revenue growth means we are also having to slow our cost growth as a company,” Netflix said in a statement. “These changes are primarily driven by business needs rather than individual performance.”
That came after Uber put the kibosh on hiring, with its chief executive, Dara Khosrowshahi, calling it a “privilege,” in an effort to cut costs. That made sense since the car-sharing service still cannot manage to turn a profit.
More jarring were reports of similar sentiments from the Silicon Valley powerhouse Meta, formerly known as just Facebook, to which Mark Zuckerberg has pledged billions of dollars for its ambitious but costly metaverse dreams. Reality hit quickly, with Zuckerberg telling employees at an all-hands meeting there would be hiring slowdowns. So far, he said, layoffs were not happening, according to a recording obtained by the Verge. “I can’t sit here and make a permanent ongoing promise that as things shift that we won’t have to reconsider that,” he said. “Our expectation is not that we’re going to have to do that. And instead, basically what we’re doing is we’re dialing growth to the levels that we think are going to be manageable over time.”
That’s a nice way of saying what staffers I spoke to expected: that if Meta’s stock price and business continue to show signs of stress, layoffs could definitely happen.
Buckle up, Silicon Valley. It could be a rough ride.
We want to hear from you.
Tell us about your experience with this newsletter by answering this short survey.