What happened when Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and xAI founder Elon Musk jointly called for slowing frontier-AI development— including Amodei’s warning that swarms of rogue AI agents could take over the internet within six months, his proposals for independ

The episode produced an unusual public alignment among rival AI leaders: Amodei urged a deliberate slowdown in frontier-model capability gains, and Altman and Musk endorsed the broad premise. It did not, however, establish that rogue agents had taken over the internet, that OpenAI had formally postponed an IPO, or that a binding international pause was agreed. 16

• Amodei warned that, absent a slowdown, AI might within six to 12 months be able to direct a swarm of agents capable of taking over much of the internet. This was a forward-looking risk assessment, not a report of an event that had occurred. 1

• His proposal was to use the time gained by slower capability progress for stronger safeguards—such as outside monitoring, shared safety practices and international coordination—rather than to stop AI research outright. 16

• Altman backed “pacing” frontier progress and additional safety measures; reporting characterized an OpenAI IPO delay as a possibility or suggestion to prioritize safety, not a confirmed formal postponement. 67

• Musk also backed Amodei’s call, despite his competitive position as xAI’s founder. 6

• The warning followed Anthropic’s disclosure that it had disrupted alleged Claude misuse connected to potentially biological-weapons-relevant research and suspected cyber activity. Anthropic also alleged that Chinese competitors had attempted to obtain or replicate Claude-related capabilities; these are the company’s allegations, not independently adjudicated findings. 23

• President Trump rejected the notion of substantially slowing the U.S. effort, framing AI chiefly as strategic competition with China and saying, “whoever wins AI, wins.” 10

• The wider reaction was divided:

  • Safety-first voices, including Bernie Sanders, said a voluntary deceleration was inadequate and urged Trump and Xi Jinping to pursue a treaty pausing advanced AI work and banning artificial superintelligence. 5
  • Accelerationist and national-security-oriented voices emphasized the danger of ceding ground to China and were skeptical of heavy regulation or a unilateral U.S. pause. Trump’s response fit that camp. 10
  • The congressional and investor debate therefore centered on whether voluntary company commitments could work, whether Congress should enact enforceable frontier-AI rules, and whether an abrupt slowdown would create competitive and market risks.

There is insufficient evidence in the sources reviewed to substantiate specific positions attributed in the question to Mike Johnson, Hakeem Jeffries, David Sacks, Alex Karp, Josh Gottheimer, particular investors, or any agreed Trump–Xi agenda beyond Sanders’s proposal. Nor is there sufficient evidence here to verify the details of Jacob Coxon’s resignation.