After slamming Anthropic for AI-led layoffs, while hiring engineers at high salaries; David Sacks responds to startup’s latest 'warning' that is in 10,000-plus words
TIMESOFINDIA.COM | Jun 10, 2026, 06.34 PM IST
David Sacks, Donald Trump’s former White House AI Czar, has publicly slammed Anthropic following the release of the AI safety lab’s massive, 10,000-word research paper warning about the catastrophic risks of artificial intelligence (AI). Sacks charged that Anthropic was using fear-mongering about AI's dangers, including job displacement, push for government regulation of the industry. Sacks had earlier called out the company for claiming its products could replace software developers while simultaneously advertising a $500,000-a-year engineering role, highlighting the startup’s ‘hypocracy’.
“Signs you might be trying to get your frontier AI lab nationalized,” Sacks wrote in response to a 10,000-word essay by Anthropic, adding, “You compare it to nukes… threaten half of white-collar jobs… warn recursive self-improvement could end humanity… then race ahead anyway. In other words, you want the government to save us from… you”.
The latest comment came months after Anthropic executives warned the world about severe economic disruption. Throughout 2026, CEO Dario Amodei warned that AI could trigger 10% to 20% unemployment within five years, heavily disrupting fields like coding, finance and law.
“Anthropic has a job listing for a software engineer on their website right now for $570,000,” Sacks noted in a previous conversation with venture capitalist Chamath Palihapitiya. “So what Anthropic is saying is they’re still trying to hire software engineers at a very high wage... but somehow they think these jobs are going to be eliminated,” he added.
Anthropic’s new 10,000-word paper, titled When AI builds itself, pivots from standard automation fears to an existential threat: recursive self-improvement. The paper argues that the greatest danger is no longer just laid-off workers, but a reality where AI systems design and train their own successors, pushing humans out of the loop entirely.
Anthropic warns that if this happens then a minor error in an AI’s alignment could go out of control across generations of software, threatening human control over the technology entirely. To back up its warnings, Anthropic shared internal data showing just how fast the technology is accelerating.
As of May 2026, more than 80% of the code merged into Anthropic's own codebase was written by its AI assistant, Claude, the company said, adding that the average Anthropic engineer now ships eight times as much code per quarter compared to previous years.
“Signs you might be trying to get your frontier AI lab nationalized,” Sacks wrote in response to a 10,000-word essay by Anthropic, adding, “You compare it to nukes… threaten half of white-collar jobs… warn recursive self-improvement could end humanity… then race ahead anyway. In other words, you want the government to save us from… you”.
The pivot from job losses to humanity’s survival
The latest comment came months after Anthropic executives warned the world about severe economic disruption. Throughout 2026, CEO Dario Amodei warned that AI could trigger 10% to 20% unemployment within five years, heavily disrupting fields like coding, finance and law.
“Anthropic has a job listing for a software engineer on their website right now for $570,000,” Sacks noted in a previous conversation with venture capitalist Chamath Palihapitiya. “So what Anthropic is saying is they’re still trying to hire software engineers at a very high wage... but somehow they think these jobs are going to be eliminated,” he added.
What Anthropic said in its 10,000-word paper
Anthropic’s new 10,000-word paper, titled When AI builds itself, pivots from standard automation fears to an existential threat: recursive self-improvement. The paper argues that the greatest danger is no longer just laid-off workers, but a reality where AI systems design and train their own successors, pushing humans out of the loop entirely.
Anthropic warns that if this happens then a minor error in an AI’s alignment could go out of control across generations of software, threatening human control over the technology entirely. To back up its warnings, Anthropic shared internal data showing just how fast the technology is accelerating.
As of May 2026, more than 80% of the code merged into Anthropic's own codebase was written by its AI assistant, Claude, the company said, adding that the average Anthropic engineer now ships eight times as much code per quarter compared to previous years.