The U.S. and China race to advance AI self-improvement technology—who will break through first?

Foreign Media: China and the U.S. are now racing to develop AI "self-improvement" technology, which is regarded as the "Holy Grail" of the AI field.

Last week, American AI company Anthropic announced that its Mythos model is nearing the achievement of "recursive self-improvement" (RSI)—a scenario in which AI enters an uncontrollable, fully automated cycle of self-optimization, potentially rapidly surpassing human intelligence. The company called for a global pause in AI development to prevent such runaway scenarios, though this warning has been questioned as potential marketing hype.

Meanwhile, China is also catching up with this trend. At the Zhongguancun Forum in March, the head of Xiaomi's MiMo model stated that "self-evolution" will be the biggest trend in AI over the next year, and that viable pathways toward realization have already been identified. The nation or company that first achieves RSI will gain an unshakable competitive advantage.

Original source: toutiao.com/article/1867808194818060/

Disclaimer: The views expressed in this article are solely those of the author.