I think we're scaring people, and that's irresponsible.
As the architects of artificial intelligence prepare to offer their companies to public markets, they have begun quietly revising the warnings they once issued about the future of work — suggesting that the disruption they predicted was either premature or overstated. This recalibration arrives not in a vacuum, but against a backdrop of real layoffs at real institutions, and cautionary words from central bankers who believe the deeper reckoning may simply not have arrived yet. History has often seen those who shape transformative forces reframe their consequences when the stakes of perception grow high enough.
- AI executives who once warned of sweeping job destruction are now publicly walking back those predictions, with OpenAI's Sam Altman admitting his forecasts about entry-level job losses were simply wrong.
- The reversal is shadowed by an uncomfortable irony: Standard Chartered is cutting thousands of jobs by 2030 and Snapchat eliminated a thousand positions last month, both explicitly citing AI automation.
- With OpenAI and Anthropic eyeing high-stakes IPOs, the industry faces a credibility problem — apocalyptic warnings that once generated headlines are now eroding the investor and public confidence these companies urgently need.
- Federal Reserve Governor Lisa Cook warned this week that AI's most significant employment effects may still be ahead, cautioning that losses could arrive well before any long-term gains take hold.
- The current landscape sits in an uneasy middle ground — modest measurable disruption so far, according to major economic institutions, but with the architecture of a much larger transformation already being laid.
The leaders of the artificial intelligence industry are telling a new story about jobs — and it sounds nothing like the one they told before.
Nvidia's Jensen Huang has pushed back hard against what he calls an easy narrative, arguing that AI has only recently become genuinely useful and that companies citing it as a reason for layoffs were simply using the technology as convenient cover. "It was just a way to look smart, and I hate that deeply," he said, insisting that AI will create as many jobs as it displaces. Sam Altman, speaking in Sydney this week, went further and admitted his own predictions were wrong — he expected far more entry-level job losses by now than have occurred. He rejected the "job apocalypse" framing while stopping short of dismissing employment risk entirely. Anthropic's Dario Amodei softened his own stance, suggesting that even in a heavily automated future, the remaining human workers would simply be far more productive.
The real world, however, is not waiting for the rhetoric to settle. Standard Chartered announced plans to eliminate thousands of positions by 2030 as AI absorbs administrative work. Snapchat's parent company cut a thousand jobs last month, explicitly crediting AI efficiency. These are not projections — they are decisions already made, with consequences already felt.
The timing of the executives' reassurances is difficult to separate from financial self-interest. Both OpenAI and Anthropic are moving toward public offerings that will demand broad investor confidence, even as surveys show growing public skepticism about AI's impact on livelihoods. The warnings these same leaders once issued are now working against them.
Federal Reserve Governor Lisa Cook offered a grounding counterpoint at Stanford this week, warning that the full employment effects of AI may still lie ahead — and that losses could arrive before any long-term gains materialize. Most major economic institutions have found AI's impact modest so far, but that assessment, she suggested, may only reflect the earliest chapter of a much longer story.
The leaders of the artificial intelligence industry are now telling a different story about the technology's effect on jobs. Where they once warned of sweeping disruption, they now argue that the fears have been overblown—a shift in tone that arrives precisely as their companies prepare to go public and need broad investor support.
Jensen Huang, who runs Nvidia, has been particularly forceful in pushing back against what he calls an easy narrative. He argues that artificial intelligence has only recently become genuinely useful, so how could it already be destroying employment at the scale some claim? When companies announced layoffs over the past two years, he suggests, they were simply using AI as cover for decisions made for other reasons. "It was just a way to look smart, and I hate that deeply," he said. "I think we're scaring people, and that's irresponsible." Huang has long maintained that AI will create as many jobs as it eliminates, a position he continues to defend.
Sam Altman, the chief executive of OpenAI, has taken a different approach: he has acknowledged that his own predictions were wrong. Speaking at a conference in Sydney this week, Altman said he expected to see far more job losses among entry-level positions by now than have actually occurred. "I thought there would have been a bigger impact on eliminating early-level executive positions than there actually has been," he told the audience. "I now understand better why that hasn't happened—obviously with relief—but my intuitions in this area were simply wrong." He stopped short of saying AI poses no employment threat, but he rejected what he called the "job apocalypse" that some in his own industry have predicted.
Dario Amodei, who leads Anthropic, has similarly softened his stance. He recently suggested that even if ninety percent of jobs become automated, the remaining ten percent will go to human workers who will be vastly more productive. Amodei has long been criticized by other AI leaders for what they see as fatalism about the technology's trajectory. Huang said last year that he disagreed with "almost nothing" that Amodei says—a pointed dismissal from a rival.
Yet the real world continues to move faster than the rhetoric. Standard Chartered, a major British bank, announced this week that it plans to eliminate thousands of positions by 2030 as artificial intelligence takes over administrative roles. Snapchat's parent company cut a thousand jobs last month, explicitly citing AI's efficiency gains as it pursues profitability. These are not hypothetical scenarios; they are concrete decisions being made now, by major institutions, with real consequences for real people.
The timing of these executive statements is not accidental. Both OpenAI and Anthropic are expected to pursue initial public offerings with high-profile valuations that will require extensive investor confidence. Public opinion in the United States, however, has grown skeptical. Surveys show deep concern about the disruption that AI companies and political leaders have themselves predicted. The apocalyptic warnings of the past are now working against the industry.
Lisa Cook, a governor of the Federal Reserve, offered a sobering counterpoint this week. She warned that the full employment effects of artificial intelligence may still lie ahead. "We may be approaching the most significant reorganization of work in generations," she said during a speech at Stanford University. She noted that job losses tied to AI could arrive before any long-term gains materialize, even if the overall outlook remains positive over decades. Most major economic institutions, including the European Central Bank, have found that AI's impact on employment has been modest so far—but that assessment may reflect only the early stages of a much larger transformation.
Citas Notables
I thought there would have been a bigger impact on eliminating early-level executive positions than there actually has been. My intuitions in this area were simply wrong.— Sam Altman, OpenAI CEO
We may be approaching the most significant reorganization of work in generations. Job losses tied to AI could arrive before any long-term gains materialize.— Lisa Cook, Federal Reserve governor
La Conversación del Hearth Otra perspectiva de la historia
Why are these executives changing their tune now, specifically?
Their companies are about to ask the public to invest billions of dollars. Apocalyptic warnings about job destruction don't help that pitch. It's simpler than conspiracy—it's just incentives.
But Altman admitted he was wrong about job losses. Isn't that honest?
It is, but it's also convenient. He's saying his predictions were too pessimistic, which sounds humble. What he's not saying is whether the real losses will eventually match what he originally feared.
So the job cuts at Standard Chartered and Snapchat—those aren't evidence that AI is already doing what they warned about?
They're evidence that something is happening. But the executives can argue those cuts were driven by other factors, or that they're isolated cases. The Fed's Lisa Cook is saying something different: the big wave might still be coming.
What does Cook think is different about her position?
She's not trying to sell stock. She's trying to prepare policy makers for a possibility that the tech industry now has incentive to downplay. She's saying the losses could come before the gains, which contradicts the rosy long-term narrative.
Is there a way to know who's right?
Not yet. We're in the middle of it. But the fact that executives are reversing course while job cuts are still being announced suggests they're managing perception, not reporting reality.