Satia Nadella, CEO of Microsoft, said the Chinese R1 model for Deepseek is the first model of artificial intelligence he saw was able to compete like Openai.
Nadila said in an interview with Bloomberg Business Week on Thursday.
Nadila added: “Dibsic, and R1 in particular was the first model I saw to publish some points,” Nadila added.
In January, Deepseek won the investors after the free Chatbot app reached the first place in the Apple US App Store. Dibsic models are highly performance but relatively cheaper An Amnesty International Investors also asked whether this would lead to the fall of the artificial intelligence devices such as NVIDIA chips.
Microsoft has started making versions of the R1’s Deepseek on its cloud platform, Azure Ai Foundry, in January. Regardless of Deepseek, the platform includes other models of artificial intelligence from companies such as Openai, Meta and Mistral.
The use of R1 means on the Microsoft platform that the data will not be sent to Deepseek servers in China.
Asha Sharma, Vice President of Microsoft Corporate, who heads the company’s AI’s platform’s producer, wrote in a blog in January that R1 had passed “with collective and strict red safety” before making it available to customers.
In January, the CEO of Openai Sam Altman R1 described “an impressive model, especially about what they can offer for the price.” Al -Taman said that Openai “will” withdraw some versions “in response to the” Deepseek “competition.
“R1 is actually not unusual,” Ben Bokanan, a former specialist adviser to the Biden Administration, said in an episode of “The Ezra Klein Show” that was broadcast in March.
“The media noise around him was justified.”
“Where do you think they got an increase in their performance? We read their papers. They are smart people who do the same type of algorithm efficiently carried out by companies such as Google and Openai.”
Microsoft, Openai and Deepseek did not respond to suspension requests from Business Insider.
