“Some missing zeros”: Sam Altman takes a search for the claim of China’s deep “lowcost” | Tendency

Openai CEO Sam Altman has excavated in the claim now now now contained to form its AI tool at a low cost of only $ 6 million in an interview with Hindustan Times. Altman spoke on Wednesday in Hindustan Times on Wednesday during his current trip to India and discussed a range of central questions of artificial intelligence, including the rise of Rival Deepseek.

“What was your first answer when the news of the Chinese model, In depthcame out? At least, the title was that they had managed to train their model at a much lower cost, although it turned out later than this was not really the case, “he asked.

Altman responded with a jibe with the Chinese competitor of his business.

“I was extremely skeptical about the number of costs. It was like, zeros lacked. But, yes, it’s a good model, and we will have to make better models, which we will do, “he replied.

A little context

The startup of Chinese AI Deepseek sent shock waves via the Silicon Valley by claiming that it has formed its AI advanced model, R1, with only $ 6 million and 2,048 GPUs, which makes it very profitable by relation to models like O1 of Openai.

However, a semiianalysis report challenged this assertion, revealing that Deepseek has in fact spent up to $ 1.6 billion on equipment and has a Fleet of 50,000 GPU Nvidia Hopper. Although the company has recognized that its reported cost only covered “official training” and not preliminary research or experience, the new results suggest that the development of Deepseek AI was much more expensive than initially presented.

While Deepseek’s profitability remains questionable, its rapid progress in AI has positioned it as a solid competitor in industry.