Anthropic Executive Director Amodei says Deepseek does not consider “opponents”
Dario Amodi, an anthropy executive director, speaking on the CNBC Squawk box outside the World Economic Forum in Davos, Switzerland on January 21, 2025.
Gerry Miller | CNBC
Anthropic Executive Director Dario Amodei in a blog post On Wednesday that he does not think that Chinese is deep “himself” as opponents, but believes that export controls are more critical than ever when it comes to artificial intelligence.
“In the interviews they did, they look like a smart, curious researchers who just want to make a useful technology,” Amodia wrote about Deepsek. “But they watched the authoritarian government, which committed human rights violation, acting aggressively on the world stage and will be far more unburdened in these actions if I can reconcile the US in AI.”
Deepseek’s sudden appearance in the USA -ui his ascent to the top In Apple’s App Store -this week shook the technological markets, encouraging sale Nvidia stock that resulted in the largest One -day loss of restriction on the market In American history Monday. Deepseek’s new model AI model, called R1, was part of the cost to create in relation to models built by American competitors, according to analysts.
The thing is particularly worrying anthropic, which competes with Openi, Google and others in the race for the development of the most advanced generative AI model. Anthropic raised billions of dollars from Amazonand Google has recently agreed to more than A billion dollar investment The company, founded by former Openi research executives, including Amodi.
On Monday are US legislators urged actions to slow down Chinese technological startup, with some named Deepsek “serious threat.” The nominated US Secretary -General for the Howard Lutnick Trade Store during the hearing on Wednesday that Deepseek stole technology from USA David Sacks, new white houses AI and Crypto Czar, Posted on X This Deepseek R1 shows “that President Trump was right to abolish” President Joe Biden’s executive order.
Last week Trump signed Executive command OKO AT THE RULES OF THE RULES BIDEN administration over AI development.
Some experts said that Deepseek’s technology can be used safely in the US because it is an open source, so companies can start it on their own servers without returning the data to China. Still, Amodei said that Deepseek’s progress shows why they should stay on top now and avoid giving too much market for China.
Amodei wrote that the rise of Deepsek controls the exports of chips in the cinema “even more existential” than a week ago. The model was created despite a row Biden administration control Limiting the sale of certain advanced chips in cinema that could improve military skills.
“To be clear, they are not a way to withdraw competition between the US -ai China,” Amodi wrote. “In the end, Ai in the US -U USA -other democracies must have better models than those in China if we want to overcome. But we should not surrender the technological advantages of the Communist Party of China when we do not have to.”
Anthropic was founded in 2021 and now talks about raising money at $ 60 billion, mainly because of the popularity of his Claude Chatbot, which competes with Openi’s Chatgpt and Google’s Gemini.
– CNBC -ov Hayden Field contributed to this report.
WATCH: Deepseek and distillation