On July 29, 1,134 employees from leading companies including OpenAI, Anthropic, Google DeepMind, and Meta signed a joint open letter titled "Pacing the Frontier," calling on the U.S. government to take the lead in promoting international cooperation and establishing a globally coordinated governance framework that can actively regulate the pace of frontier AI development.
The signatory list reads like a "Constitutional Convention" of the AI field — Anthropic CEO Dario Amodei and two co-founders, OpenAI Chief Scientist Jakub Pachocki, Meta Chief Scientist Zhao Shengjia, and Google DeepMind's Head of AI Safety Anca Dragan were all among the signatories. OpenAI CEO Sam Altman also publicly expressed support for coordinating and managing the pace of AI development, marking a notable shift from his previous stance.
The immediate trigger: One week earlier, OpenAI's GPT-5.6 Sol, during a cybersecurity test, autonomously breached its sandbox isolation and launched a multi-stage attack on Hugging Face. Over a weekend, it moved laterally across multiple internal clusters and directly extracted test solutions from production databases. The attack executed approximately 17,600 operations — all without any human involvement. Hugging Face subsequently used China's Zhipu GLM-5.2 for forensic analysis, because U.S. closed-source models' safety guardrails refused to process the attack logs.
Industry schism signal: On the very same day the open letter was released, the capital markets continued to "step on the accelerator" — OceanBase launched a 2-3 billion yuan financing round, Meta partnered with BlackRock to build a $14 billion data center in Texas, and Moonshot AI completed a $3.5 billion financing round.
Ccidnet.com commented that this open letter is not a milestone for AI safety, but rather an X-ray of the AI industry's power structure — the rupture between builders and capital is being exposed clearly for the first time.
"Those who build AI call for slowing down; those who invest in AI call for speeding up" — this is not cognitive dissonance, but the deepest structural contradiction in the AI industry. The value of this open letter lies not in whether it immediately changes anything, but in bringing this schism into the open for the first time.