DeepSeek Engineer Compares Anthropic AGI Lead to Hitler's Bomb
TL;DR
- Liu Shengyu, a DeepSeek engineer who worked on the V4.1 models, said Anthropic controlling AGI would rival 'Hitler acquiring atomic bomb technology before the Allies.'
- The WeChat post frames the AI future as a fork between a 'communist-style' shared outcome and a 'Cyberpunk 2077' concentration of frontier models.
- It is one of the sharpest Chinese-industry rejoinders yet to US safety-pacing arguments, days before an anticipated September 24 Xi-Trump meeting.
Liu Shengyu, a systems engineer at Chinese lab DeepSeek who worked on the company's V4.1 models, wrote in a viral WeChat post that Anthropic reaching AGI first would carry consequences "no less severe than Hitler acquiring atomic bomb technology before the Allies." The South China Morning Post reported the essay went out shortly after DeepSeek shipped V4.1 Flash.
Liu framed the coming decade as a fork between two outcomes: a "communist-style" future where AI productivity gains are broadly shared, and a "Cyberpunk 2077" future where a handful of tech firms "hoard the most advanced models and lock in permanent class stratification." "I especially don't want Anthropic to master the most advanced artificial intelligence or AGI," he wrote, extending the same distrust to OpenAI. Neither, in his read, can be trusted to keep frontier AI "open and affordable," the phrase he used for DeepSeek's mission and his own reason for staying.
The post landed during a stretch of US industry calls to slow frontier development on safety grounds, days before an anticipated September 24 meeting between Xi Jinping and Donald Trump. It is one of the sharpest Chinese-industry rejoinders yet to that pacing argument, arriving in a week of Anthropic coverage already thick with governance and safety stories.
Originally reported by scmp.com
Read the original article →Original headline: DeepSeek Engineer Shengyu Liu: Anthropic Reaching AGI First Would Be Like Hitler Getting the Atomic Bomb