June 11, 2025
While most AI companies talk about “augmenting” workers, San Francisco-based startup Mechanize has a different message: it wants to automate all white-collar jobs — and do it fast. The founders say they’re not building digital assistants. They’re building replacements.
“Our goal is to fully automate work,” said co-founder Tamay Besiroglu at a recent event in San Francisco. “We want to get to a fully automated economy, and make that happen as fast as possible.”
(Source: The New York Times)
Mechanize was launched in 2025 by Besiroglu, Ege Erdil, and Matthew Barnett, all alumni of Epoch AI, a research firm studying AI capabilities. The company’s early focus is automating software engineering using reinforcement learning — the same AI method that powered DeepMind’s AlphaGo. But their ambitions go much further.
The team is building virtual training environments that mimic a real office desktop: email inboxes, coding tools, Slack channels, and web browsers. AI agents are dropped into the simulation and asked to complete work tasks. If they succeed, they’re rewarded. If they fail, they try again. With enough trial and error, the idea goes, they’ll learn to do what a human worker does — only faster and at scale.
“We’ll only truly know we’ve succeeded once we’ve created A.I. systems capable of taking on nearly every responsibility a human could carry out at a computer,” the company said in a recent blog post.
Mechanize has backing from influential tech leaders, including Stripe co-founder Patrick Collison and Google AI chief Jeff Dean. The company is working with leading AI labs, though it declined to name them due to confidentiality agreements.
What it hasn’t done — yet — is offer any detailed plan for how society should handle the job loss that may follow. When asked whether it’s ethical to automate all labor, co-founder Matthew Barnett replied: “If society as a whole becomes much wealthier, then I think that just outweighs the downsides of people losing their jobs.”
Mechanize says the long-term goal is “radical abundance” and a future where universal basic income could support those displaced. But for now, its efforts are squarely focused on replacing knowledge workers — not retraining them.
As white-collar unemployment ticks upward and AI models grow more capable, Mechanize represents a shift in tone for Silicon Valley — one where honesty replaces reassurance, and disruption isn’t denied, it’s promised.
