A China Daily editorial argues that, when people who build the most powerful artificial intelligence (AI) systems suddenly urge others to slow down, it is reasonable to ask why. Sam Altman, Elon Musk, and Dario Amodei said work at the frontier, the leading edge of research, should be paced, or kept slower. This followed Amodei’s claim that systems can improve themselves and that autonomous agents, software able to act with little human direction, had caused concern.
The timing is striking. Shortly before, Anthropic accused Chinese companies of industrial-scale illicit distillation. Distillation is an ordinary method in which a smaller model is trained on the answers of a larger one. American companies have long done the same with Chinese models. The editorial says a common practice does not become illegal only because a rival uses it. I argues that this aims to slow China, win friendlier rules in Washington, and keep investors enthusiastic.
Why the caution appears now
For years the firms grew through money and better models. The promise AI kept investors interested. Most businesses already use current systems for writing and coding. OpenAI’s plan for an IPOI has reportedly stalled. Cheaper Chinese models such as DeepSeek and Kimi K3 made the frontier look more like a crowded market than a closed estate.
The editorial argues that, after chip limits failed to stop Chinese models, attention moved to software. The line between safety and a business barrier has been blurred. David Sacks, co-chair of a presidential science advisory group, warned the three companies not to form a cartel, a group that restricts competition. Treasury Secretary Scott Bessent said there is no day after tomorrow if China wins. Treating the contest as zero-sum, a game in which one side’s gain is the other’s total loss, makes cooperation harder. China and the United States should work together where neither can manage the results alone. The promise of AI is to serve a common good, not geopolitical gain or private profit.