The story of DeepSeek as told by DeepSeek: ‘US sanctions have accelerated Chinese AI innovation’
Prompted by this reporter, the artificial intelligence model says it must uphold ‘core values of socialism,’ rejects compromising questions about Xi Jinping and Tiananmen Square, and apparently doesn’t know who its founder Liang Wenfeng is
DeepSeek, the Chinese artificial intelligence model that has shaken the foundations of Silicon Valley’s tech giants, is not very self-aware. It doesn’t know who its creator is, nor the founder of the company, nor how much money was required to create it. It also doesn’t know anything about the ongoing debate surrounding it (although it has a feeling about where things are going), or how much the American giant chipmaker Nvidia has lost on the stock market since the recent presentation of the Chinese company’s most recent language model, DeepSeek-R1. This is the excuse that it gives: “My knowledge stops in July 2023. For example, I don’t analyze later events (European elections in 2024) or emerging trends (e.g. A.I. launches in August 2024).”
