Microsoft president Brad Smith has spent four decades around computer scientists, and he says they keep making the same two mistakes. They overestimate how fast a new technology will spread, and they underestimate what people are capable of. The Microsoft president put both ideas at the center of a 3,000-word essay this week, written in response to a wave of graduating students who booed every mention of AI during their commencement speeches this spring. His read on those boos isn’t that the kids got it wrong. It’s that the tech sector did.“The reactions of this year’s graduates are a powerful wake-up call for the tech sector,” Smith wrote in the post, titled “AI, jobs, and the next generation,” published on Microsoft’s blog on June 10. “Hopefully, leaders across our industry will listen and seek to learn from this reaction.” He started drafting it during a return visit to Princeton, his alma mater, over Memorial Day weekend, where graduating seniors had swapped out a beer jacket design after discovering it was made with AI. The replacement jackets read “100 percent cotton” and “100 percent human.”That detail tells you what Smith is responding to. Across US campuses this season, students didn’t just sit quietly through AI cheerleading. Former Google CEO Eric Schmidt got booed at the University of Arizona. A real estate executive got the same treatment at the University of Central Florida. The pattern held often enough that Smith decided it was worth a long, careful answer rather than a one-line dismissal.
Why a Microsoft executive is suddenly listening to college students
Smith’s argument leans on a piece of history. In 1838, the camera arrived, and the French painter Paul Delaroche reportedly took one look at an early photograph and declared, “From today, painting is dead!” Painting, of course, did not die. It bent toward Impressionism, then Cubism, then Surrealism, as artists chased what a camera lens couldn’t capture. Smith uses that arc to make his case: technology disrupts a field, people adapt, and new kinds of work appear that nobody saw coming.He stretches the same logic across his own industry’s past. When word processors arrived, typists worried their work would vanish, and instead an entire knowledge economy grew up around the computer. When spreadsheets automated arithmetic, accountants didn’t do less math; they built more elaborate financial models. “When technology increases supply, human ambition often generates more demand,” Smith wrote. “As humans, we don’t plateau. We expand.”There’s a business reason he wants you to believe that, and he says so plainly. “Workers have been Microsoft’s lifeblood from the start,” he wrote. “If the world’s people don’t have jobs, then neither do we.” It’s an unusually direct admission that Microsoft’s fortunes depend on people staying employed and able to buy software. Machines, as Smith puts it, don’t buy products. People do.
The part where the essay gets uncomfortable
What makes the post land harder than the usual executive optimism is that Smith doesn’t pretend the job market is healthy. He calls the situation facing the class of 2026 a “perfect storm,” with the wind blowing from several directions at once. Graduates face “AI automation of tasks in current entry-level positions,” he wrote, plus “corporate pressure to reduce headcount to help pay for AI’s enormous capital expenditures.” Layer on geopolitical uncertainty, trade tensions, and a hangover from pandemic-era over-hiring, and you get the storm.Those aren’t abstract worries. The tech industry shed more than 38,000 jobs in May alone, the worst month since 2024. The last six months brought heavy cuts at Oracle, Meta, and AWS. Meta laid off 8,000 people the same month Smith’s essay went up, part of an AI restructuring. Standard Chartered said it would cut 7,800 back-office roles by 2030, the kind of entry-level banking jobs graduates have long used as a first rung. Goldman Sachs estimated in April that roughly 16,000 US jobs are vanishing to AI every month.And the timing of Smith’s own company complicates the message. The same week the essay published, Microsoft CFO Amy Hood told investors that headcount had fallen year-over-year in the fiscal third quarter, and that she expects the trend to continue. Microsoft plans to spend around $80 billion on AI infrastructure in 2026. Critics have noticed the shape of that math: payroll shrinking while capital spending balloons, with the savings from one helping fund the other. Smith’s essay, for all its length, came with no commitment to slow deployment, protect entry-level roles, or fund retraining at scale. That gap is exactly what some readers have seized on.
What Smith tells graduates to actually do
Smith’s prescription is built around a simple reframe borrowed from a new book by LinkedIn’s Ryan Roslansky and Aneesh Raman, “Open to Work.” Stop thinking of your job as a title, he argues, and start thinking of it as a bundle of tasks. Sort those tasks into three buckets: what AI can do, what you can do with AI, and what only a human can do. If nearly everything falls into the first bucket, find a different line of work. For most people, he says, the bulk sits in the second, which is where AI becomes a tool rather than a threat.The skills he says will hold their value are the human ones, which he lists as five C’s: curiosity, creativity, compassion, communications, and courage. Even when AI handles a stack of tasks, someone has to watch over its output, he argues, and that keeps human judgment in demand. His advice to students fretting over what to study is almost old-fashioned: pursue a field you’re passionate about, master it, then add AI fluency on top so you can apply that expertise better than anyone could before.Smith isn’t the only one softening the pitch. Across the industry, leaders who once warned of a white-collar wipeout are reaching for gentler language. OpenAI’s Sam Altman recently said he no longer expects a “jobs apocalypse,” and admitted his thinking shifted after he let AI draft his own emails and Slack messages, an experience he called “dehumanizing” to watch. Microsoft AI chief Mustafa Suleyman, who in February told the Financial Times that most white-collar tasks would be “fully automated by an AI within the next 12 to 18 months,” now says he was misunderstood. He claims he meant tasks, like writing an email or building a slide deck, not whole jobs. “That does not necessarily mean that the role goes away at all,” Suleyman said on the “Decoder” podcast.The shift in tone is hard to separate from the moment. OpenAI and Anthropic have both confirmed they’re moving toward IPOs, and SpaceX made its Nasdaq debut last week. Public opinion has curdled, too: a May Economist-YouGov poll found 71% of Americans think AI is moving too fast, and cities like Seattle have started slapping moratoriums on new data centers. Suddenly the sunny framing has commercial value.The loudest voice on the other side remains Anthropic CEO Dario Amodei, who warned last year that AI could erase up to half of all entry-level white-collar jobs within one to five years. He hasn’t backed off. “I have warned about job displacement in interviews and essays because I want both policymakers and the private sector to have the best chance to adapt and respond, not because I am trying to be a ‘prophet of doom,'” Amodei wrote this month, while also predicting AI will let individuals build billion-dollar companies.Smith lands somewhere between the alarm and the cheerleading, and he closes by handing the moment back to the graduates. “While it may feel unfair that the job market is so uncertain, you were made for this moment,” he wrote. Technology, he says, is second nature to their generation, and they don’t have decades of habits to unlearn the way older workers do. He urges them to stand firmly for values he calls timeless: agency, ambition, dignity. Whether a generation that already did the math finds that convincing is another question.
