AI researcher Gary Marcus has publicly dismissed a viral essay predicting widespread job displacement due to artificial intelligence as “alarmist hype,” arguing it lacks supporting data and presents an exaggerated timeline for AI’s capabilities.
The essay, titled “Something Massive is Happening,” was written by entrepreneur and investor Matt Shumer and has been viewed more than 80 million times on X, formerly known as Twitter. Shumer warned that AI would disrupt not only software engineering but also most jobs performed “on a screen.”
Marcus, in an interview with Business Insider, criticized the essay’s lack of factual basis. “Hyped-up views have gotten us into a bad place, possibly one that’s going to lead to a serious economic recession or something like that,” he said. “And I guess I suppose that one should work from the facts rather than just trying to cause an alarm.”
While acknowledging AI will impact the labor market, Marcus contends the changes will occur more gradually than Shumer suggests. He stated that AI can augment human work by automating specific tasks, but it is not yet capable of fully replacing human employees across entire domains. “AI can do a small subset of the tasks, and that sometimes speeds up human beings and things like that, but it rarely does all of what a human being can do in any particular domain,” Marcus explained.
He cautioned against companies rushing to replace workers with AI, citing the experience of Klarna, a Swedish fintech company. In 2024, Klarna promoted an AI assistant intended to handle the workload of 700 employees. However, by May 2025, Klarna’s CEO Sebastian Siemiatkowski acknowledged the system’s lower quality, attributing it to prioritizing cost over effectiveness. Siemiatkowski stated that investing in human support was the more viable path forward.
“Six months or a year later, they come back with their tails between their legs because it turns out that the AI systems don’t do things as well as the human,” Marcus said. He believes companies are often too quick to adopt AI solutions without fully assessing their limitations.
Marcus suggested the immediate concern isn’t AI replacing junior employees, but rather executives misjudging AI’s capabilities and making potentially costly decisions based on that misapprehension. “The biggest thing I think junior people have to worry about right now is a misapprehension by the C-suite that these techniques work better than they actually do,” he said.
In a Substack post, Marcus further elaborated on his criticisms, labeling Shumer’s essay “weaponized hype.” He argued the essay failed to account for current data and research demonstrating the ongoing limitations of AI, and misrepresented the context of a Model Evaluation & Threat Research graph assessing AI progress.
Marcus also pointed to studies, including a June 2025 paper from Apple’s Machine Learning Research Group, that highlight the challenges facing current AI models. He further noted a pattern of overestimation from prominent figures in the AI field, citing Elon Musk’s projections for Tesla’s robotaxi rollout and Geoffrey Hinton’s earlier predictions about the future of radiology.
“What they have all learned to do is to sell the rosiest picture possible, and the media rarely calls them out,” Marcus stated.
Recent predictions from Microsoft AI CEO Mustafa Suleyman, suggesting most white-collar tasks could be automated within a year and a half, also drew skepticism from Marcus. He questioned the feasibility of automating tasks in fields requiring high accuracy, such as accounting, where even minor errors can have significant consequences.
