Anthropic proposes temporary pause on AI development and convenes policymakers (theguardian.com)

• Anthropic calls for a worldwide temporary pause on AI development to discuss risks.
• Claude model shows progress toward recursive self-improvement; over 80% of Anthropic's code is now AI-authored.
• Anthropic has embedded engineers in the NSA for offensive cyber operations.
• Experts say no step change in AI capabilities and note contradiction in Anthropic's actions.

"Anthropic has proposed a worldwide "temporary pause" on AI development and said it will convene policymakers to discuss risks, in a post detailing its Claude model's progress toward recursive self-improvement. The company warned that continued progress could lead to humans losing control over AI systems. Separately, the Financial Times reported that Anthropic has embedded engineers in the US National Security Agency to assist with offensive cybersecurity operations. Experts noted the contradiction and said Anthropic's definition of AI safety is narrow, with no evidence of a step change in AI capabilities. The company recently filed for an IPO."