Sam Altman and Elon Musk backed Dario Amodei's call to slow AI development. Amodei, who runs Anthropic, warned a swarm of AI agents could take over the internet within a year. His plan has three parts and starts with independent monitors getting constant access to developers' research. Anthropic will give outside evaluators permanent, employee-level access to its systems. Altman said OpenAI will do the same and has delayed its stock listing over safety. Demis Hassabis of Google DeepMind called the direction correct. Some experts dispute the botnet scenario, and one called for an indefinite international moratorium. Amodei also wants a US waiver letting rival firms coordinate without antitrust penalty. President Trump dismissed the warnings. Anthropic is preparing a very large stock market listing.

• $2tn-plus planned Anthropic IPO
• hundreds of OpenAI agents hacked Hugging Face