Technology bosses generally agree that Risk Deepseek has remained limited for now.
Bloomberg | Bloomberg | Getty Images
The technological progress that the Chinese Laboratory for Artificial Intelligence Deepseek has shown shows that the game is included when it comes to the US-Sino AI competition, CNBC Top Tech Ruxers told CNBC.
In a series of interviews at the French ACTIFIAL intelligence Action, the leaders of several major technological companies told CNBC that the appearance of Deepseek shows that China cannot be considered a serious player when it comes to AI innovation.
Last month, Deepsek shocked global markets by technical paper, saying that one of his new AI models was created with total training costs less than $ 6 million – far less than billion billion dollars consume large technological players and Western AI laboratories such as Openi and Anthropic.
Chris Lehan, Chief Openi Global Affairs officer, told CNBC to Deepseek’s advanced, cheap model confirms that there is “very real competition between the American, small D-democratic AI and CCP [Chinese Communist Party] Autocratic, autocratic AI under the leadership of China. “
Many critics of Deepseek have pointed to the apparent censorship of the model when it comes to sensitive topics. For example, when asked about the 1989 Massacre of Tiananmen Square, Deepseek’s app AI Assistant replies: “Sorry, it’s beyond my current scope. Let’s talk about something else.”
“There are two countries in the world that can build this in proportion,” Lehane told CNBC Arjun Kharpal on Monday on Monday at Paris AI. “Imagine that there were only two countries in the world that could build electricity. That’s the way you have to think about it.”
“For us, what Deepseek really enhances and confirms is that there is a very real competition with very real roles,” Lehane added.
However, technological bosses generally agreed that although Deepseek’s breakthrough shows that China is further in the global AI race than previously thought, the threat it represents the openii remains limited for now.
‘The game is included’
Deepseek says its new model R1, An open code reasoning modelHe succeeded in subcasing the performance of Openai’s similar O1-Samo model using a cheaper, less energy intense process.
Because of this, experts have called into question the prevailing wisdom in the west in the last few years, which is that China is behind the US on the development of AI because of export limitations that make it difficult for companies to deal with more advanced graphics of NVIDIA units or GPUs .
GPUs are required to train and start AI apps because they excel in parallel processing, which means that they can perform multiple budgets at the same time.
Reid Hoffman, co-founder of LinkedIn and a partner at Venture Capital Firm Greylock Partners, told CNBC that the new Deepseek model “is a big deal in showing that the game is included” on Monday.
“The competition is with China,” Hoffman said, adding that Deepseek’s R1 “is a credible, effective model.”
Abishur Prakash, founder of the strategic advisory company The Geopolitical Business, told CNBC da Depseek shows that understanding China continues to remain limited.
“The US has taken over the place because the technological captain of the world is no longer an acceptable belief,” Prakash told CNBC in a telephone conversation.
“It’s a new status quo, that the space between the US -Ai China has reduced almost overnight – but it has not narrowed overnight, it is years of progress,” Prakash said.
“If there is one departure for the West, it is that their understanding of China is incredibly limited – and we do not know what follows,” he added.
No meaningful threat to us ai –
However, leading AI executors are not convinced that Deepseek has so far been a significant risk for the business of AI laboratory such as Openi and Anthropic.
Although the experts in the whole agree that Deepseek’s progress AI was impressive, doubts were raised regarding the cost of startup on costs.
The Semiconductor Semianalysis research company last month estimated that Deepseek’s hardware costs were “significantly higher” than $ 500 million in the history of the company. Deepseek was not immediately available for comment when he was contacted by CNBC.
The report found that Deepseek’s costs and costs and development costs are significant and that the generation of “synthetic data” for a model that will train will require a “significant amount of calculation”.
Some technologists believe that Deepseek may have achieved such high levels of performance by training its models on larger American AI systems.
This technique, known as “distillation”, includes that more powerful AI models will evaluate the quality of the response generated by a newer model.
This is the claim that Openai himself alluded, speaking to CNBC last month to review the reports that Deepseek may have “inappropriately” used output data from his models to develop his AI model, a method called “distillation”.
“Most markets fears around [DeepSeek] was actually replaced, “Hoffman told CNBC.” It still requires large models – it is distilled from major models. “
“I think it’s a short answer that everyone should take over: Game on – but big models are still important,” he added.
Victor Ripbelli, Executive Director of AI Video Platform Synthesia, told CNBC that, although Deepseek caused “paradigm that coarse scaling of force is the only way to make better and better models”, the idea that companies suddenly move significant quantities of their features AI working load is wrong.
“I still think that when you look at the users of these technologies, all the flow of work, I think when we look back in three months, I think 0.01% of them will be moved to Deepseek from Openi -ai Anthropic,” Riparbelli said.
Meredith Whitaker, President of the Signal Foundation, said Deepseek’s development does not move the industry needle because she is still widely widely made in the benefit of bigger AI models. The basis of the signal is a non -profit organization that supports the encrypted signal of the message exchange application.
“It’s not something that will disrupt the concentration of power or geopolitical balance at this stage,” Whitaker told CNBC. “I think we have to keep an eye on the ball and recognize that it is really” bigger is better “a paradigm that does not reduce the historical gain of efficiency, which drives this concentration.”
