Ben Horowitz is unconvinced by the growing chorus of warnings that artificial intelligence is on the verge of wiping out jobs at scale. Instead, the Andreessen Horowitz cofounder argues that the most confident predictions about AI-driven mass unemployment rest on a shaky assumption: that the future of work can be forecast with any real certainty.
Speaking on Tuesday on the Invest Like The Best podcast, Horowitz said the debate around AI and employment has become overly deterministic, with too many commentators treating job losses as an inevitable and linear outcome of technological progress.
“I think people are acting as though it’s very predictable when it’s not at all predictable,” Horowitz said. “Why are you so sure it’s going to happen next? And why are you so sure no jobs are going to be created? I don’t think it’s nearly as predictable as people are saying.”
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
His comments land at a moment when anxiety over AI’s economic impact is intensifying, even as adoption accelerates across industries. Governments are weighing regulation, companies are racing to integrate generative AI into workflows, and workers are trying to understand whether the technology represents an opportunity or an existential threat.
The debate has sharply divided the tech world. Some of the field’s most prominent voices have issued stark warnings. Geoffrey Hinton, often described as the “Godfather of AI,” has repeatedly cautioned that advanced AI systems could displace large numbers of workers. Similar concerns have been raised by UC Berkeley professor Stuart Russell, University of Louisville computer scientist Roman Yampolskiy, and Anthropic chief executive Dario Amodei, who has said AI could eliminate a significant share of white-collar jobs.
On the other side are executives who see AI less as a job killer and more as a force for reconfiguration. OpenAI chief executive Sam Altman and Nvidia boss Jensen Huang have both argued that while AI will disrupt existing roles, it is more likely to reshape work and create new categories of employment rather than erase work altogether.
Horowitz aligns more closely with the latter camp, but his argument goes beyond optimism about new roles. He frames AI as the latest phase in a much longer historical process, one in which automation has repeatedly transformed economies in ways that were impossible to foresee in advance.
To make his point, Horowitz drew on the most dramatic example of all: agriculture. In the early years of the US economy, he noted, farming accounted for roughly 95% of jobs. Today, those roles have almost entirely disappeared, replaced by work that would have been unrecognizable to earlier generations.
“We’ve been automating things since the agricultural days,” Horowitz said. “Almost all those jobs have been eliminated. And the jobs we have now, the people doing agriculture wouldn’t even consider jobs.”
For Horowitz, that history undercuts the confidence of today’s AI doomsayers. The error, he argues, is assuming that current job categories offer a reliable map of the future. Each major technological shift, from industrial machinery to computers and the internet, destroyed large numbers of existing roles while giving rise to entirely new forms of work that were difficult to imagine beforehand.
“The idea that we could imagine all the jobs that are going to come, sitting here now, that AI is going to enable, I think is low,” he said.
He also questioned the sense of urgency that often accompanies predictions of imminent mass displacement. Modern AI, he argued, did not appear overnight. Many of its core technologies have been under development for more than a decade.
Image recognition breakthroughs date back to around 2012, while advances in natural language processing gathered pace by 2015. Even ChatGPT, which ignited the current wave of public interest, was launched in 2022. Yet, Horowitz asked, where is the wave of job destruction that was supposed to follow?
“We’ve had AI going right — ImageNet was what, 2012 and then natural language stuff was like 2015, and then ChatGPT was 2022,” he said. “Where’s all the job destruction? Why hasn’t it happened yet?”
That skepticism does not mean Horowitz believes AI will leave the labor market untouched. He acknowledged that certain roles are likely to come under pressure, particularly jobs focused on processing or relaying information for others, tasks that AI systems are increasingly capable of performing faster and cheaper.
But he expects those losses to be offset by growing demand in other areas, especially work that leans on creativity, judgment, and the ability to define problems rather than simply execute them.
“I don’t really think that the door is going to close behind you,” Horowitz said. “I think the opportunities tend to multiply when you open up a new way of doing things.”
His view reflects a broader strand of thinking within Silicon Valley that sees AI less as an endpoint and more as a general-purpose tool, one that will alter how work is done rather than eliminate the need for people altogether. Still, the gap between that outlook and the more pessimistic warnings from parts of the AI research community underscores how unsettled the conversation remains.
Horowitz’s message is one of caution against certainty. History, he suggests, offers plenty of evidence that technology disrupts labor markets in messy, uneven, and often surprising ways. What it does not offer is a clear script for how AI’s story will unfold.