After Deepseek-R1 was released earlier this month, the company boasts “Performon on Par” as one of the latest models of Openai if used for tasks such as mathematics, coding, and natural language inference.
Deepseek’s technology is praised by famous numbers, including Openai Chief Sam Altoman, which called “impressive models, especially for prices”, but Openai “Openai is clearly better. We will provide it. ” forward.
“Despite the restricted access to advanced hardware, the ability comparable to DeepSeek’s US models indicates that the ingenuity of software and data efficiency can compensate for hardware restrictions.” According to Marina Zhang, an associate professor at Sydney Institute of Technology, which focuses on the Chinese high -tech industry.
Ion Stoica, a co -founder and executive chair of AI Software Company DataBricks, said that the BBC could spur more DEEPSEEK’s low costs to adopt AI in business.
“If it happens, this cost reduction can accelerate the progress of AI,” he said. “So as a whole, the market will expand faster and the market value will grow faster.”
Chinese companies argue that their models can train with 2,000 special chips compared to the estimated models of the main models of 16,000.
But not everyone is convinced. Some are asking some doubts about Deepseek’s claims, including Tech Mogul ELON MUSK.
He responded to a post that claimed that DeepSeek actually had about 50,000 NVIDIA chips and is now banned from exports to China.
Sudden popular explosions cause cyber security concerns.
In Australia, the Minister of Science Ed Fessick was one of the experts in Australia’s National Broadcasting Station ABC.
