David Sacks warns US AI policy risks falling behind China

The rise of Kimi K3, a model built with 2.8 trillion parameters and a one-million-token context window, highlights a growing gap in development speed. Beyond its coding dominance, the model achieved an Elo score of 1,668 on GDPval v2 and outperformed competitors in agent-led software workflows. Moonshot AI, the firm behind the system, plans to release open weights by July 27, further challenging the closed-model ecosystems favored by some US tech giants.

Sacks contends that the US must pivot away from the current trajectory of state-level restrictions and data center limitations. He draws a parallel to the early internet era, where minimal government interference allowed for rapid innovation. While acknowledging the necessity of managing safety risks, he warns that overly bureaucratic federal reviews will inevitably cede ground to foreign rivals who operate under fewer constraints. This competitive urgency mirrors the broader technology agenda pushed by the Trump administration, which frequently cites the need to outpace Beijing in critical sectors like artificial intelligence and digital assets.