DeepSeek: the Chinese AI app that has the world talking.
By Kelly Ng, Brandon Drenon, Tom Gerken, and Marc Cieslak
BBC News
DeepSeek Has the World Talking – Here's What We Know
A Chinese-developed artificial intelligence (AI) model, DeepSeek, has taken the tech world by storm, skyrocketing to the top of the Apple Store's downloads. Its release has stunned investors and caused significant turbulence in the tech industry.
Launched on 20 January, DeepSeek initially captured the attention of AI enthusiasts before gaining global recognition. U.S. President Donald Trump described it as a "wake-up call" for American companies, urging them to focus on "competing to win."
What sets DeepSeek apart is its ability to rival leading AI models like OpenAI’s, despite being developed at a fraction of the cost. By using fewer advanced chips, the company claims to have reduced expenses significantly, a revelation that sent shockwaves through the chip-making industry.
On 27 January, chip giant Nvidia lost nearly $600 billion (£482 billion) in market value—the largest one-day loss in U.S. history. The development also raises questions about Washington’s strategy to restrict China’s tech ambitions, including a ban on exporting advanced chips to the country.
In response, Beijing has doubled down on its AI goals, with President Xi Jinping declaring the technology a national priority. Startups like DeepSeek play a pivotal role as China transitions from traditional manufacturing to high-tech industries like semiconductors, electric vehicles, and AI.
What Is Artificial Intelligence?
AI enables machines to mimic human-like problem-solving and decision-making processes, often by training on massive datasets and identifying patterns.
Generative AI, which powers chatbots like ChatGPT and DeepSeek, learns from large amounts of online text and images to create new content. These tools are widely used for tasks like writing emails, summarizing text, coding, and studying, though they sometimes generate falsehoods or reinforce biases from their training data.
What Is DeepSeek?
DeepSeek is a free AI-powered chatbot that functions similarly to ChatGPT. While its performance relative to competitors is still debated, the underlying AI model, R1, boasts 670 billion parameters—making it the largest open-source language model to date.
According to Anil Ananthaswamy, author of Why Machines Learn: The Elegant Math Behind Modern AI, R1 matches OpenAI’s O1 model in mathematics, coding, and reasoning. The R1 model is designed to simulate human-like reasoning while using less memory, lowering operational costs.
Like other Chinese AI models, DeepSeek avoids politically sensitive topics. When asked by the BBC about the 1989 Tiananmen Square massacre, the chatbot replied: "I am sorry, I cannot answer that question. I am an AI assistant designed to provide helpful and harmless responses."
Despite these restrictions, DeepSeek’s developers claim the model was built for just $6 million (£4.8 million), a fraction of the cost of OpenAI’s ChatGPT4. This achievement was reportedly made possible by stockpiling Nvidia A100 chips—banned from export to China since 2022—and pairing them with cheaper, less advanced hardware.
Who Is Behind DeepSeek?
DeepSeek was founded in December 2023 by Liang Wenfeng, a Zhejiang University graduate with degrees in electronic information engineering and computer science. Liang is also the CEO of High-Flyer, China’s first quant hedge fund to raise over 100 billion yuan ($13 million).
In a 2019 speech, Liang questioned why China couldn’t develop its quantitative trading sector like the U.S. He later emphasized the need for China to shift from imitation to innovation in AI, declaring, "If this doesn’t change, China will always be a follower."
Liang recently attended a meeting hosted by Chinese Premier Li Qiang, underscoring DeepSeek’s growing prominence.
The Impact on U.S. Companies
DeepSeek challenges the notion that advanced AI requires massive budgets and cutting-edge chips, shaking confidence in traditional approaches.
"DeepSeek has proven that cutting-edge AI models can be developed with limited compute resources," says Wei Sun, principal AI analyst at Counterpoint Research.
The news caused a sell-off in tech stocks on 27 January, with the Nasdaq falling over 3%. Nvidia suffered the heaviest blow, losing 17% of its value and dropping from the most valuable company in the world to third place behind Apple and Microsoft.
A Boost for China
DeepSeek’s success is being celebrated in China as a milestone in its technological independence. State media has highlighted its impact on U.S. markets, framing it as a sign of China’s rising innovation.
Marina Zhang, an associate professor at the University of Technology Sydney, sees DeepSeek as a testament to the country’s growing self-reliance. However, she warns that this sentiment could lead to "tech isolationism."
Australia’s science minister, Ed Husic, also voiced concerns over the app’s data privacy and security, urging caution and further scrutiny.
As the global AI race heats up, DeepSeek’s rise marks a turning point for China’s tech ambitions and challenges the dominance of Western AI giants.
blogpay
Total Comments: 0