Two stories arrived inside the same news cycle, and they appear to describe different worlds. In one, Wix cut 20 percent of its staff, more than a thousand people, and the CEO posted on X that the move reflects "the most significant shift in how companies are built since the invention of modern programming languages in the 1970s." In the other, the Federal Reserve Bank of New York published a study arguing that nearly two-thirds of the rise in unemployment among young US college graduates can be explained by something far more boring: the rise of remote work, and the fact that managers do not want to train juniors over Zoom.
The contrast captures the strange standoff in the AI labour debate. Executives keep blaming the bots. Economists keep finding that the bots have not yet done much.
Paul Osterman of MIT Sloan, author of Disposable Workers, told Fortune the executive line is mostly cover. "AI is a perfect excuse to justify big layoffs," he said. "It makes it seem as if it's not our decision, our fault, it's the technology." Cisco cut 4,000 jobs in May and the share price jumped 13 percent. Block, Snap, Atlassian and now Wix have all attached the same "smaller and flatter" framing to their reductions. Osterman calls the pattern AI washing, and he points out that the smaller-team rhetoric is twenty years old. Only the cited cause is new.
The pushback inside the AI industry itself is sharper than you might expect. Demis Hassabis of Google DeepMind told Wired that the layoff logic is a failure of imagination. If engineers become three or four times more productive, he said, the answer is to build three or four times as much, not to cut headcount. "I have a million ideas," he added. "I'd love to have some free engineers." Jensen Huang of Nvidia made a similar point more bluntly: "AI has just arrived. How is it possible they're already losing jobs?" Sam Altman of OpenAI now says he was "pretty wrong" about white-collar displacement. Anthropic's Dario Amodei, who once warned AI could erase half of entry-level office work, has softened the framing too. Critics note both companies are heading into IPOs and a calmer story plays better with the pension funds.
The macro data sits closer to Hassabis than to the layoff press releases. Apollo's Torsten Slok published a note this week titled "Zero Evidence of AI-Related Job Losses," citing ADP payroll data and rising demand for AI implementation specialists. He invokes Jevons paradox: cheaper automation tends to expand activity, not contract it. Yale's Budget Lab still finds no measurable shift in unemployment rates for AI-exposed occupations. In Australia, Monash economist Zac Gross has used nationwide payroll data to show weaker hiring in roles like accounting clerks and telemarketers but very little wage softening; software engineers' pay tracks the national average even as AI has become a major part of the job. "These results don't cancel the AI apocalypse," he said, "but they certainly postpone it." NAB's economists see a clearer signal, with employment 9 percent lower in AI-exposed occupations since 2022, but they call it crosscurrents rather than collapse.
Two things complicate the calm read. The first is geography. India's working-age population is the largest in the world, and the Indian economy's job-creation per unit of GDP is already among the lowest. AI lands hardest on labour markets that have nowhere to absorb displaced workers. China, ageing fast, can welcome the same wave that India dreads. The second is capital flow. Australian businesses spent a record 6 billion dollars on AI equipment in the first three months of this year, a 400 percent annual increase, and have a 155 billion dollar data-centre pipeline now growing faster than the 2010s mining boom. That money will eventually show up somewhere on the employment line. Whether as augmentation or substitution is the open question.
The honest read is that the macro numbers are calm, the micro experience is not, and both are true at the same time. Roughly 142,000 tech jobs have been announced as cut so far in 2026. College graduates are booing AI from the stage at commencements. Workers at Indian factories are wearing sensors to train the robots that will replace them, and one of them told the Times of India it felt like "working in your own grave, while you make your own casket." Numbers and feelings are running on different clocks. Which one bends toward the other first is the actual story.