Machine codes: DeepSeek has ripped away the veil of mystique around AI

No, it was not a "Sputnik moment".

- Kenan Malik

The launch last month of DeepSeek R1, the Chinese generative AI or chatbot, created mayhem in the tech world, with stocks plummeting and much chatter about the US losing its supremacy in AI technology.

The original Sputnik moment came on 4 October 1957 when the Soviet Union shocked the world by launching Sputnik 1, the first time humanity had sent a satellite into orbit. It was, to borrow a phrase from a later landmark, "one giant leap for mankind", in Neil Armstrong's historic words as he took a "small step" on to the surface of the moon.

It was a significant moment in the cold war, too. A confidential White House report worried that "American prestige" had "sustained a severe blow", giving the USSR "clear advantage in the cold war".

That fear spurred Washington into reshaping its space programme.

DeepSeek is a notable achievement. Technically, though, it is no advance on existing large language models (LLMs). It is neither faster nor "cleverer" than OpenAI's ChatGPT or Anthropic's Claude and just as prone to "hallucinations" -the tendency, exhibited by all LLMs, to give false answers or make up "facts" to fill gaps in its data.