Mark Zuckerberg Takes Aim at Anthropic in Debate Over A.I. Slowdown - The New York Times

Open original

Mark Zuckerberg criticized Anthropic's push for a global AI slowdown, arguing that companies can manage safety risks on their own without industry-wide pauses. He positioned Meta as already doing this work internally with products like its new Muse agent.

  • Zuckerberg said labs have "responsibility and incentive" to train models safely without needing external pressure, contrasting with Amodei's call for a coordinated global slowdown
  • Meta delayed releasing Muse for several months to ensure security, which Zuckerberg offered as proof companies can self-regulate
  • Zuckerberg took a jab at competitors pursuing "recursive self-improvement" (using AI to develop itself), calling it misguided compared to serving users

Mark Zuckerberg publicly criticized Anthropic and its CEO Dario Amodei over the push to slow down AI development. Without naming them directly, Zuckerberg argued that leading AI labs should focus on safety improvements rather than racing to build more powerful systems. His post came days after Amodei called for a global AI slowdown, a position backed by Sam Altman (OpenAI), Elon Musk, and Demis Hassabis (Google DeepMind). The disagreement reflects a genuine split in how the industry's heavyweights view the risks versus the competitive pressure to advance. Zuckerberg staked out a middle position: he acknowledged AI risks exist but claimed companies already have incentives to keep users safe and can manage this internally without industry-wide pauses. He pointed to Meta's recent release of Muse, an AI agent that can operate software independently, as proof—the company delayed launch by several months for security testing. His argument boils down to: Meta handled this responsibly on its own, and other labs can too. He also took a swipe at competitors chasing "recursive self-improvement," where AI systems design themselves, saying companies should focus on serving people instead. The timing matters because Meta has fallen behind both Anthropic and OpenAI in the AI race and is preparing to release its most powerful model yet, called Watermelon, within weeks. Zuckerberg's pushback against a slowdown conveniently aligns with Meta's competitive interests while maintaining a veneer of safety-consciousness. The broader context is an Anthropic researcher's recent resignation with warnings that AI companies were "gambling with our lives," which triggered the current public debate about development speed.

Ask about this article