WSJ Article Claiming China Has Matched Anthropic Is Obvious Nonsense

LessWrong contributor Zvi is disputing a Wall Street Journal article claiming China's Zhipu AI has matched Anthropic's Mythos model in cybersecurity. According to Zvi, the headline is outright false and the reporting is misleading. He points out that Mythos is special because it autonomously identifies security vulnerabilities at scale and chains them into working exploits without being explicitly directed to specific code. Zhipu's GLM-5.2 model, by contrast, can find bugs but only if provided the correct code section and sufficient resources—a fundamentally different capability. Zvi criticized the WSJ for sloppy reporting that could mislead policymakers and the public about the true state of the AI race.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton