Tech
Mark Zuckerberg cheers at UFC match
Mark Zuckerberg cheers during UFC 298. (Chris Unger/Zuffa LLC via Getty Images)
Platformer

Mark Zuckerberg’s 20-year mistake

Meta’s CEO says he’s done apologizing. Should we worry?

Casey Newton

Shortly after bounding on stage Tuesday night at Chase Center, where he sat for an interview with the hosts of the “Acquired” podcast, Mark Zuckerberg joked that he might have to book a second appearance to apologize for what he was about to say.

It was a gentle dig at Nvidia CEO Jensen Huang, who appeared in a video shown at the event to walk back an innocuous comment he had made during his own appearance on “Acquired.” Huang said then that he never would have started the company had he known how challenging it would be; on Tuesday, he said the remark had been taken out of context.

Zuckerberg, for his part, suggested that no second appearance would be necessary. After pausing a beat, he looked out into the crowd of 6,000 or so podcast fans and said — to laughter and some applause — that he’s done apologizing.

So began one of the most revealing interviews of Zuckerberg’s career. Over more than an hour, Zuckerberg spoke with hosts Ben Gilbert and David Rosenthal about how events of the past several years have changed his perspective on leadership, politics, and product development. In most interviews with journalists since 2016, Zuckerberg quickly finds himself in a defensive crouch, working to share product announcements in between questions about platform-related harms. 

In the cozier confines of “Acquired,” Zuckerberg seemed notably more relaxed. And that led him to do something he almost never does in public: reflect.

Zuckerberg’s reflections are of particular interest to me, since when I started writing a daily newsletter almost seven years ago, I wrote primarily about Facebook. Specifically, I wrote about Facebook in crisis: from the aftermath of the 2016 US presidential election to Cambridge Analytica and the broader backlash against tech that consumed Silicon Valley for the better part of the past decade.

Throughout those turbulent years, Zuckerberg and Facebook did apologize: for Russian interference in the election, for Cambridge Analytica, for the experiences of parents who say their children died as a result of their use of his platforms

Talk is cheap. But those apologies were accompanied by significant investments in policy, counterterrorism, content moderation, and other efforts to shore up the integrity of the platform. The Oversight Board, an imperfect but still somewhat radical experiment in building a quasi-judicial system for a social network, also emerged from this ongoing sense of regret.

In important ways, Facebook — now Meta — escaped any real accountability for its lapses around data privacy, content moderation, and child safety. After countless performative hearings, the United States Congress did not pass a single national law to regulate Meta and other big tech companies. Meta’s user base today is bigger than it has ever been, its stock is trading near all-time highs, and Zuckerberg’s net worth is estimated by Forbes to be $182 billion. When all the world asks you for is an apology, it’s relatively easy to give.

In other ways, though, Meta continues to live in the shadow of 2016. Exasperated by the lack of action at the federal level, dozens of states are pursuing laws to make it harder for young people to access social media. The vast majority of people do not trust the company to protect their data. And the Federal Trade Commission is now suing to break up the company.

I didn’t expect any of this to come up at “Acquired,” a feel-good podcast about business leadership that focuses more on competitive strategy than regulatory threats. But then, about 30 minutes into the podcast taping, co-host Ben Gilbert asked Zuckerberg: “Of all the criticisms that have happened over the years, which do you believe is the most legitimate? And why?”

Zuckerberg took the opportunity to discuss his shifting calculus around the value of an apology. He had expected that publicly accepting responsibility for problems would allow the company to move on. Instead, he said, it led to critics demanding that he and Meta take responsibility for even more problems. 

“I don't want to simplify this too much, and there are a lot of things that we did wrong,” he said, by a way of a caveat. And then: 

One of the things that I look back on and regret is, I think we accepted other people's view of some of the things that they were asserting that we were doing wrong, or were responsible for, that I don't actually think we were. There were a lot of things that we did mess up and that we needed to fix. But I think that there's this view where, when you're a company and someone says that there's an issue, … the right instinct is to take ownership. Say, maybe it’s not all our thing, but we’re going to fully own this problem, we’re going to take responsibility, we’re going to fix it.

When it’s a political problem … Sometimes there are people who are operating in good faith, who are identifying a problem and want something to be fixed, and there are people who are just looking for someone to blame. And if your view is “I’m going to take responsibility for all this stuff” — people are basically blaming social media and the tech industry for all these things in society — if we’re saying, we’re really gonna do our part to fix this stuff, I think there were a bunch of people who just took that and were like, oh, you're taking responsibility for that? Let me like, kick you for more stuff.

He then compared the crisis to the company’s rocky initial public offering, when the company’s stock price failed to pop and prompted a wave of lawsuits around its disclosures.

“Honestly, I think we should have been firmer about and clearer about which of the things we actually felt like we had a part in and which ones we didn't. And my guess is if the IPO was a year and a half mistake, I think that the political miscalculation was a 20 year mistake. [...]

We sort of found our footing on what the principles are, and where we think we need to improve stuff. But where people make allegations about the impact of the tech industry or our company which are just not founded in any fact, that I think we should push back on harder. And I think it's going to take another 10 years or so for us to fully work through that cycle before our brand is back to the place that it maybe could have been if I hadn't messed that up in the first place.”

He then turned to the audience and smiled.

“But look,” he said. “In the grand scheme of things, 20 years isn't that bad!”

What precise criticisms Zuckerberg thinks are fair, and which he thinks are made in bad faith, he did not elaborate on. You can guess, though, based on what the company actually chose to spend money on: the army of content moderators it brought on after 2017 to shore up the platform’s defenses, for example, or the Oversight Board, to which it has given $280 million so far amid complaints that Zuckerberg himself should not be the final word in difficult questions about online speech.

What he believes is unfair, I think, is the suggestion that Meta is primarily responsible for our polarized politics, the decline of democracy, or the dire state of mental health among young people in the United States.

What led Zuckerberg to lose faith in the value of contrition? Here’s a guess: ultimately, it bought him very little, at least in the court of public opinion. Meta seems equally disliked in red and blue states; this week, 42 attorneys general endorsed a plan to force Zuckerberg to add warning labels to his products saying that they put children at risk. Why apologize, you can imagine him and his lieutenants asking one another, when it gets you nowhere? 

I write all of this because in a world where tech companies are very rarely held accountable in meaningful ways, it matters what CEOs do and do not feel responsible for. It matters if they decide they no longer want to take ownership of societal-level problems, even when their companies may be exacerbating them. 

Almost no one thinks of the half-decade or so after 2016 as a golden age. But it was a time when Meta invested much more time, money, and talent into understanding how its products could be misused. And while surely the company would say it continues to make those investments today, it’s notable that Meta’s CEO is now signaling in public that he is turning his attention elsewhere. 

Zuckerberg’s public presentation has grown notably more flamboyant in recent years; on Tuesday he wore a T-shirt he designed himself bearing the slogan “learning through suffering” in Greek letters. (“A family slogan,” he explained.) 

Despite those changes, though, on Tuesday I was struck by what remains the same. Early in the conversation, he was asked about building products. Zuckerberg contrasted the company’s quick-shipping culture with Apple’s, which places a higher emphasis on polish.

“There are a lot of conversations that we have internally where you're almost at the line of being embarrassed about what you put out,” Zuckerberg said. “You want to put stuff out early enough so you can get good feedback.”

What about the damage that does to the brand, Gilbert asked.

“I would like to hope it’s damaging to the brand,” Zuckerberg said. “We don’t ship things we think are bad. But we want to make sure we’re shipping things that are early enough that we can get good feedback to see what they’re going to be most used for.” 

Facebook retired its old “move fast and break things” motto in 2014, but here it was again, in slightly updated form. It’s a slogan that explains so much of the company’s success: the way it quickly copies and improves on products from other companies, and routinely tests the boundaries of what tradeoffs people are willing to make with their privacy.

But it also explains Meta’s failures to release products that are safe, or build trust, or generate goodwill. 

By his own estimation, it will take Zuckerberg another 12 years for Meta’s brand to recover from the drubbing it began to take at the end of 2016. And I can understand why he would much rather now focus on artificial intelligence, or mixed reality, or the company’s other long-term bets on the future.

But I worry about a Meta that remains just as powerful while also becoming less sensitive to public pressure and criticism. The moment Mark Zuckerberg stops apologizing could give the rest of us a lot to be sorry about.

Casey Newton writes Platformer, a daily guide to understanding social networks and their relationships with the world. This piece was originally published on Platformer.

More Tech

See all Tech
tech

xAI’s trade secrets lawsuit against OpenAI dismissed

In September, xAI sued rival OpenAI, accusing the company of stealing trade secrets by hiring away key employees.

After a key employee was poached by OpenAI, xAI filed its lawsuit in Northern California federal court, alleging that the company was after its “secret sauce” — the methods by which xAI was able to build data centers so quickly.

Today Judge Rita Lin dismissed the case, citing a lack of any direct accusations against OpenAI itself. In the order dismissing the case, Lin wrote:

“The sole defendant in this lawsuit is OpenAI, whom xAI accuses of misappropriating its trade secrets. But xAI does not point to any misconduct by OpenAI. Instead, it points to eight former xAI employees who left for OpenAI at around the same time.”

The judge said that xAI is allowed to file an amended complaint if it wants to pursue the case, but must do so by March 17.

Today Judge Rita Lin dismissed the case, citing a lack of any direct accusations against OpenAI itself. In the order dismissing the case, Lin wrote:

“The sole defendant in this lawsuit is OpenAI, whom xAI accuses of misappropriating its trade secrets. But xAI does not point to any misconduct by OpenAI. Instead, it points to eight former xAI employees who left for OpenAI at around the same time.”

The judge said that xAI is allowed to file an amended complaint if it wants to pursue the case, but must do so by March 17.

tech

Chinese drone maker DJI sues to overturn FCC foreign drone ban

Chinese drone maker DJI has filed a lawsuit against the FCC challenging the December 2025 decision that effectively bans all foreign drones and components from the US market. DJI and all other foreign drone makers were added to the FCC’s “covered list” of equipment and services that the agency says “pose an unacceptable risk to the national security of the United States.”

In a lawsuit filed with the US Court of Appeals for the Ninth Circuit, DJI argues that the company is “severely harmed” by the ruling, and seeks review of the decision: “The FCC exceeded its statutory authority, failed to observe statutorily required procedures, and violated the Fifth Amendment when it purported to add DJI’s products to the Covered List.”

In a statement to Sherwood News, a DJI spokesperson said:

“The FCC can add products to the Covered List only when they present a national security threat, yet it has never identified any threat associated with DJI or its products. Despite repeated efforts to engage with the government, DJI has never been given the chance to provide information to address or refute any concerns. These procedural and substantive deficiencies violate the Constitution and federal law.”

The FCC decision has cleared the way for the nascent US drone industry just as the US military urgently races to catch up in the race to acquire drones.

tech

Anthropic follows OpenAI in rolling out agentic tools for enterprise

Just a day after OpenAI rolled out its agentic platform for enterprise, Anthropic has announced its own. Built from existing pieces of Anthropic tech that have already been previewed, the new platform essentially ties together plug-ins that can be tailored by enterprise customers into Claude Cowork.

Companies can customize their version of the tool to use their branding, communication style, and private data to speed up a long list of common tasks like performing financial analyses, human resources tasks, design, and engineering workflows. New connectors tie Claude Cowork into third-party platforms like Salesforce’s Slack, Google’s apps, LegalZoom, and DocuSign, among others.

The announcement gave a lift to some beaten-down software companies.

While OpenAI was releasing consumer apps like Sora, Anthropic was busy improving Claude’s ability to make spreadsheets and PowerPoint presentations — the boring but essential tools of the workplace.

The two competing solutions will be battling it out in the enterprise marketplace as both Anthropic and OpenAI seek to grow revenue streams to power their ambitious AI infrastructure projects.

Companies can customize their version of the tool to use their branding, communication style, and private data to speed up a long list of common tasks like performing financial analyses, human resources tasks, design, and engineering workflows. New connectors tie Claude Cowork into third-party platforms like Salesforce’s Slack, Google’s apps, LegalZoom, and DocuSign, among others.

The announcement gave a lift to some beaten-down software companies.

While OpenAI was releasing consumer apps like Sora, Anthropic was busy improving Claude’s ability to make spreadsheets and PowerPoint presentations — the boring but essential tools of the workplace.

The two competing solutions will be battling it out in the enterprise marketplace as both Anthropic and OpenAI seek to grow revenue streams to power their ambitious AI infrastructure projects.

tech

Alphabet’s Waymo is now available in 10 cities

Today, Alphabet subsidiary Waymo announced it’s now welcoming public riders to its driverless car service in four additional US cities: Houston, Dallas, San Antonio, and Orlando.

The company said the service will be available first to “select riders” who’ve downloaded the app in those cities, and it will invite new riders on a rolling basis before opening the service to everyone “later this year.”

The latest announcement brings Waymo’s total service area to 10 cities, mostly located in California and across the Sun Belt, and doubles its footprint from a few months ago.

The latest announcement brings Waymo’s total service area to 10 cities, mostly located in California and across the Sun Belt, and doubles its footprint from a few months ago.

tech

TSMC trades near record high as Apple plans to purchase “well over 100 million” chips from its Arizona facility this year

Last year Apple said it would invest $600 billion to expand in the US over four years, in an effort to reduce its reliance on overseas suppliers and avoid tariffs.

We’re now getting more detail on what that involves. According to a new announcement from Apple, it includes expanding AI server production and moving future production of suddenly trendy Mac Minis to Houston.

It also involves the purchase of “well over 100 million” chips from TSMC’s Arizona facility this year, “a significant increase from 2025.”

“We’re buying as much of the output of this fab as we can,” Apple’s global head of procurement, David Tom, told The Wall Street Journal.

Apple has long been one of TSMC’s largest customers, even as the iPhone maker shifted to designing its own processors in-house — chips that TSMC overwhelmingly manufactures. Apple’s demand is helping fund TSMC’s massive multibillion-dollar chip plant expansion in Arizona.

Shares of TSMC were up 3% and near record highs in early trading Tuesday, as the Apple announcement underlines huge demand for TSMC’s AI chips

It also involves the purchase of “well over 100 million” chips from TSMC’s Arizona facility this year, “a significant increase from 2025.”

“We’re buying as much of the output of this fab as we can,” Apple’s global head of procurement, David Tom, told The Wall Street Journal.

Apple has long been one of TSMC’s largest customers, even as the iPhone maker shifted to designing its own processors in-house — chips that TSMC overwhelmingly manufactures. Apple’s demand is helping fund TSMC’s massive multibillion-dollar chip plant expansion in Arizona.

Shares of TSMC were up 3% and near record highs in early trading Tuesday, as the Apple announcement underlines huge demand for TSMC’s AI chips

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.