AI -Assistant created by Chinese startup DeepSeek It became the number one most downloaded app in Apple’s US app store over the weekend, sending shockwaves through Silicon Valley and sending the price of major tech stocks tumbling. Nvidia has seen more than 460 billion dollars Wiped off its market capitalization on Monday, a drop Bloomberg characterized as “the largest in US stock market history.”
The shakeup stems from an open-source model developed by DeepSeek called R1, which debuted earlier this month. The company said it rivals the current industry leader: OpenAI 01. But what stunned the tech industry was that DeepSeek claimed to have built its model using only a small fraction of the specialized computer chips that AI companies typically need to develop cutting-edge systems.
On Monday, DeepSeek said it was temporarily limiting new registrations, citing “large-scale malicious attacks” on the company’s services, according to A message on its website.
DeepSeek’s R1 model “challenges the notion that Western AI companies hold a significant lead over Chinese ones,” Jack Clark, co-founder of the AI startup Anthropic, said. wrote in his newsletter. Venture capitalist Marc Andreessen called It “Ai’s Sputnik moment.”
Cheng Lu, a research scientist at OpenAI, said DeepSeek’s chat has demonstrated impressive Chinese conversation skills. “It’s the first time I can feel the beauty of Chinese language created by a chatbot,” he said in X -post on sunday
DeepSeek’s AI Assistant is currently available for free and comes with three main features. First, users can ask the chatbot questions and get direct answers. For example, when Wired asked for recipe ideas that incorporate pomegranate seeds, DeepSeek’s chatbot quickly provided a list of 15 options ranging from yogurt parfaits to a “Middle Eastern-inspired” rice pilaf, but it didn’t cite any specific chefs or recipes.
The DeepSeek app also has a search mode that surfaces answers from the internet. When Wired asked “What are some big news today?” DeepSeek’s Chatbot cited the Israel-Hamas shutdown and linked to several Western news outlets such as BBC News, but not all stories seemed relevant to the topic. Ironically, one was a New York Times story about DeepSeek’s impact on the bag.
Lastly, there’s a “deep” mode that allows users to plug in DeepSeek’s R1 model, which was built on top of the company’s existing V3 model. The difference between the two is that R1 has so-called “reasoning” capabilities that allow it to explain step by step how it reached its conclusions. For example, when asked “What are the most important historical events of the 20th century?” DeepSeek initially provided a long meandering answer that started with a few broad questions.
“That’s a hundred years, so there’s a lot going on,” read part of its response. “I should probably break it down from decades or major topics like wars, political changes, technological advances, social movements, etc.” The DeepSeek chat then went on to cite World War II, the Cold War, and the Holocaust.
But before R1 could finish his answer, the entire answer disappeared and was replaced by a message that read “Sorry, I’m still not sure how to approach this type of question. Let’s chat about math, coding, and logic problems instead!” Some experts and early adopters noticed That DeepSeek, like other technology platforms that operate in China, it seems Widely censored Subjects deemed sensitive by the Chinese Communist Party
But despite these limitations, DeepSeek’s free chat could pose a serious threat to competitors like OpenAI, which charges $20 a month to access its most powerful AI models. Unlike its Chinese counterpart, OpenAI does not disclose the underlying “weights” of its models that determine how the AI processes information. It also declined to publish the full “Thought chains” Produced by their own reasoning models.