News Summary

In a growing concern among American AI experts, notable figures like Nathan Lambert, founder of the ATOM Project, and Andy Konwinski, co-founder of Databricks, are advocating for a shift towards open source AI models to prevent the US from falling behind China in the AI race. The US has traditionally led AI advancements with companies like OpenAI and Google. However, Chinese companies such as Kimi, Z.ai, Alibaba, and DeepSeek are rapidly gaining traction by promoting open-weight AI models that foster innovation and collaboration. Lambert argues that open models are essential for experimentation and innovation, allowing startups to run them on their own hardware. Meanwhile, Konwinski highlights the increasing number of breakthrough AI papers from China being read by American students, warning that without embracing open source, the US risks losing its innovative edge. This call for openness is echoed by the ATOM Project, which emphasizes the strategic need for the US to lead in open source to maintain its AI dominance.