Dario Amodei has had one favorite topic lately. Whether he is speaking at the World Economic Forum in Davos, at an AI summit in India, or on a podcast, the message is the same: AI will soon swallow up a huge share of office jobs around the world. Programmers will be first. Then lawyers, finance professionals, and consultants will follow. And unemployment? According to him, it could climb to 10 to 20 percent within five years.
Apply that figure to the global workforce, and we are immediately talking about hundreds of millions of people out of work.
This is not the first time Amodei has faced accusations of hypocrisy. After all, Anthropic itself is developing Claude, the very same AI model that he says will swallow up entire professions. When directly asked how the two fit together, he has a straightforward answer: "Most of them have no idea what is coming. It sounds crazy, and people simply do not believe it," he told Axios. He is convinced that other tech CEOs are keeping quiet about the risks. He claims to be speaking frankly.
900 billion reasons to talk about the apocalypse
Anthropic is approaching an IPO. The company’s target IPO valuation is estimated at $900 billion. That would put it above OpenAI, which closed a record funding round this year at a valuation of $852 billion. Anthropic’s annual revenue has reportedly exceeded $40 billion, but the company is still burning through enormous sums on computing power, research, and employees.
For the numbers to make sense when it goes public, Anthropic needs to convince major institutional investors such as BlackRock, Fidelity, and sovereign wealth funds. And they do not want to hear that Claude is a capable assistant. They want to hear that Claude will replace people.
If AI truly displaces half of all office workers, the addressable market is not the enterprise software category. It is the entire global wage bill for knowledge workers. "If AI takes over all jobs, it is better to own a piece of that AI," said Ben Goertzel, who coined the term AGI. Amodei may be sincere in his warning. But Wall Street reads it as an invitation to invest.
What the actual figures say
Anthropic’s own researchers published a study in March 2026 that complicates Amodei’s public statements.
Computer programmers do indeed rank at the top of their AI exposure index, with roughly 75 percent task coverage. The broader computer and mathematical category has a theoretical ceiling of as much as 94 percent. Yet Claude currently covers only 33 percent. Legal, administrative, and financial roles show similar gaps between theory and practice.
And the main finding? The study found no systematic increase in unemployment among the occupations most at risk. What the data does show is a decline in the hiring of recent graduates—specifically people aged 22 to 25—into exposed roles, by roughly 14 percent since the launch of ChatGPT. Troubling, but far from what Amodei describes in the media.
The Yale Budget Lab economic institute examined macroeconomic data through the end of 2025 and found no significant impact of AI on the labor market. "No matter how you look at the data, at this point we simply do not see any significant macroeconomic effects," the institute’s co-founder Martha Gimbel told Fortune.
Yann LeCun, Sam Altman, and others
Amodei is not alone in his apocalyptic vision. But a growing number of industry voices are pushing back against him.
Yann LeCun, one of the fathers of modern artificial intelligence, responded sharply to Amodei’s warning: "Dario is wrong. He knows absolutely nothing about the impact of technological change on the labor market." LeCun added that no technology CEO, himself included, is qualified to make such predictions. Economists who have studied this topic throughout their careers tell a different story.
An interesting shift occurred with Sam Altman. At the beginning of 2026, he was still echoing Amodei’s tone and warning that the impact of AI on jobs would soon be "tangible." Then, on May 1, he wrote something that sounds almost like a correction: "We want to build tools that augment and elevate people, not entities that replace them." He added that "pessimism about jobs is probably wrong in the long term."
AWS CEO Matt Garman announced at an Amazon company event that the firm plans to hire 11,000 software engineering interns in 2026, in line with previous years. According to him, demand for developers is actually "growing." Work is evolving, but it is not disappearing.
Anthropic itself has not eliminated jobs. Quite the opposite
And here is the most striking part. While Amodei warns about the end of programming careers, Anthropic has posted more than 400 open engineering positions, some paying as much as $405,000 a year.
Head of Growth Amol Avasare says the product management department is "completely overloaded" and that the company urgently needs to hire product managers. This is despite the fact that Claude Code has reportedly tripled the engineering team’s productivity internally.
Boris Cherny, who built Claude Code and predicted that the title "software engineer" could disappear by the end of the year, clarified his position: "I do not think we are at a point where you can be completely hands-off." He himself has not written code manually since November. But he reads every line that Claude generates. The tool’s creator says you still cannot trust it fully. The company’s CEO says it will take your job. The truth lies somewhere between those two statements and is quietly being negotiated every day, without major headlines.
Source: timesofindia.indiatimes.com
