Mark Zuckerberg’s Awkward Pivot to AI Optimism

For what feels like a long time now, the big AI companies have been talking about how future versions of their technology could destroy millions of jobs or escape control in ways that pose dire risks to humanity. In the meantime, as actually available AI products have gotten popular, they have produced smaller but highly visible externalities: visual slop overwhelming the web, social media, and now the physical world; rampant cheating in school; and a steady flow of stories about chatbots driving people to hurt themselves or others.

The public faces of AI include established figures like Mark Zuckerberg and Elon Musk, who carry with them world-historically heavy reputational baggage, as well as more freshly divisive personalities like Sam Altman and Dario Amodei. We’re here, we’re all billionaires, and we’re rearranging the entire economy to build something that might ruin your life, they seem to be saying, and the public has heard them, loud and clear. Now — oops! — stopping data-center construction is one of the rare causes that can unite American partisans in 2026. Lots of people are using AI, sure, but they’re also profoundly anxious and increasingly angry about it.

AI companies, which need both positive market sentiment and a tolerable political and regulatory environment to sustain their current trajectories, understand this could become a major problem. And so, in recent weeks, they’ve been testing out some new pitches.

On the Relentless podcast, Altman invoked magic. “We are close to creating the genie that can grant any wish,” he said. “We are going to make sure that our first wish is to broadly benefit humanity and get the world to a place where a lot more people get to have a lot more wishes.” In meetings ahead of Anthropic’s planned IPO, according to The Wall Street Journal, the company has been assuring potential investors it “plans to push further into healthcare and biology AI uses and that the work could help mitigate some of the negative sentiment around AI.”

Last week, Zuckerberg, whose company is spending hundreds of billions of dollars to catch up with Anthropic and Altman’s OpenAI, published yet another AI manifesto, this one titled The Future Is for Everyone. It is, in part, a clear (and clearly motivated) defense of open AI models, in contrast with the closed-frontier models offered by Meta’s competitors. But Zuckerberg also attempts to sketch, in his words, a “path to a positive AI future.” In it, he says, “everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about,” and this agent “will work 24/7 on your behalf to improve your relationships, health, career, finances, home management, hobbies, and more.” There will also be “incredible tools for creation to express your ideas,” as well as to “create new businesses,” resulting in an economy that “will become more entrepreneurial.” Everyone will also have “a personalized tutor and coach with a PhD in every subject,” he writes. And they will “benefit from scientific advances and be able to contribute to scientific progress.”

These AI companies are all trying to engender optimism, but they are using meaningfully different strategies. Altman is leaning into an enchanted vision for AI in which the technology is so capable and continues to improve so quickly that anyone keeping track will enter and remain in a state of awe. Anthropic’s theory is that, if AI tools help develop new and lifesaving drugs and therapies — if they help cure cancer, for example — the public will relate to them differently and with gratitude. Zuckerberg’s mechanically optimistic scenario is a sort of argument for the virtues of AI-accelerated capitalism: more entrepreneurialism, more individual capability, and more freedom, resulting — eventually, in theory — in greater affluence and higher standards of living.

There are plenty of ways one might push back on these nascent cases for AI positivity, ranging from differing projections of model capability to more straightforward observations that the habits Zuckerberg describes in his essay — “My daughter loves to bake so my agent plans personalized recipes for us to make together each weekend, orders the ingredients, and then offers suggestions as we’re baking” — might not be sound as appealing to other people as he thinks they do. But what’s fascinating about these pitches is the blindness they have about how they might fail even if their premises are right. 

In Altman’s case, the problem is that technological enchantment is stubbornly temporary. The world is full of tools that would have sounded like magic not too long ago and were indeed first met with awe; now the “supercomputers in our pockets,” for example, are black holes for vibes, blamed for a staggeringly wide range of social and political problems. Genuinely miraculous things could emerge from AI research, but people would quickly cease to experience them as miracles.

From within an organization run by effective altruists steeped in utilitarian ideas about maximizing welfare, the idea that a company could potentially produce life-saving technologies and yet find itself the subject of public backlash might sound absurd or like a sign of a doomed society. But it’s also realistic and something about which people in plenty of other industries (and in politics) are far less naïve. Rear-looking arguments about comparative welfare — quality of life is improving in many important ways — are useful and interesting. But the general public isn’t entirely wrong to treat them as irrelevant. Relentlessly taking innovations for granted and finding fault with the ways new things work is a feature, not a bug, of capitalist democracy. It helps drive the system. Put another way: That’s what the money’s for.

Zuckerberg’s vision of individual computational and economic empowerment is also likely to crash into a hard social reality. Anyone with access to a computer and the internet can accomplish, find, or understand things that would have been unrealistic just a few decades ago or that would have required the work of a dedicated staff to achieve. Having an internet-connected answer machine with GPS and worldwide instant-communication capabilities has indeed solved numerous old problems for billions of people. But the experience of this change has been blunted by the fact that the world has now passed through it together. An AI agent that “will work 24/7 on your behalf,” if such a thing becomes plausible, will become the norm; it will also become mundane. Having personal assistants, as someone like Zuckerberg is accustomed to at this point, is hugely useful. But it’s also a comparative advantage (for a recent example of how things like this play out, once everyone is a ticket sniper, nobody is). A time traveler with a functioning iPhone connected to an internet wormhole would be a god in the past. Today, he’s just another guy with a Reels addiction, complicated feelings about his kids’ screen time, and an always-on relationship to his office job. Even if things go well on the AI industry’s terms, “AI” could remain a sponge for discontent, repellent to good vibes, and, of course, lucrative for all involved.

It makes sense that this punishing reputational dynamic could take OpenAI and Anthropic by surprise and that they’d care so much about it: They’re weird new companies that conceive of and talk about their products in idiosyncratic ways, and they’re unfamiliar, at least firsthand, with the sort of backlash they’re starting to see and therefore more likely to perceive it as an existential threat.

Zuckerberg’s pitch reads more obviously as wishful thinking. He’s experienced this before. He’s run a massive consumer-technology company for years, presiding over a portfolio of products that are used by billions of people, many of whom nonetheless express anxiety, discontent, and anger about social media in general and who agree, when polled, that they don’t really like him specifically. (A useful counterexample here is Amazon, a company that is both taken for granted and widely liked, perhaps because people were never told to think of it as anything more than a place where they can buy things easily. See also: Apple.) Nobody seemed particularly grateful when Facebook “connected the world,” in other words. And maybe, he seems to be hoping aloud, things could be different this time.