DeepSeek, a groundbreaking artificial intelligence model developed in China, has taken the tech world by storm. Released on January 20, 2023, coinciding with the inauguration of US President Donald Trump, DeepSeek quickly gained prominence in China. Created by AI expert Liang Wenfeng, the model operates on an open-source platform and is hailed as a "game-changer" in the Chinese AI industry. Its launch has prompted astonishment in Silicon Valley and beyond, raising questions about the United States' ability to maintain its lead in AI development.
Liang Wenfeng, a 40-year-old alumnus of Zhejiang University, founded DeepSeek. He deliberately chose an open-source platform for the model, which has contributed to its widespread accessibility and appeal. DeepSeek's development team comprises fewer than 140 individuals, most of whom are "home-grown talent" from elite Chinese universities. This team has leveraged their expertise to create a model that surpasses expectations and competes directly with American counterparts.
DeepSeek's capabilities have stunned experts and users alike. The model can perform a wide range of tasks, from answering basic queries to solving complex math problems. It also provides personalized horoscopes, or ba-zi, based on a user's date and time of birth. This versatility has led to comparisons with models like ChatGPT, but experts argue that DeepSeek is more advanced and capable.
In terms of hardware, DeepSeek boasts a "substantial stockpile" of Nvidia A100 chips, estimated to be between 10,000 to 50,000 units. The model was trained using 2,000 advanced chips alongside thousands of lower-grade ones. This strategic acquisition has enabled DeepSeek to develop a powerful AI model at a lower cost than similar models in the United States.
"The timing and the way it's being messaged – that's exactly what the Chinese government wants everybody to think – that export controls don't work and that America is not the global leader in AI." – Gregory C Allen
The success of DeepSeek has compelled Chinese companies to innovate and maximize their resources. Marina Zhang, an associate professor at the University of Technology Sydney, notes that this drive for self-reliance is deeply rooted in China's rapid technological ascent.
"Having grown up during China's rapid technological ascent, they are deeply motivated by a drive for self-reliance in innovation." – Marina Zhang
Liang Wenfeng acknowledges the challenges posed by US-imposed chip bans. Despite these hurdles, DeepSeek's success demonstrates China's ability to adapt and thrive in the face of adversity.
"The main challenge is the chip ban." – Liang Wenfeng
DeepSeek's release has generated a mix of shock and excitement within the tech community. Wei Sun describes the reaction as particularly pronounced within the open-source community.
"A blend of shock and excitement, particularly within the open-source community," – Wei Sun
Fiona Zhou, an intrigued user, echoes the widespread curiosity surrounding DeepSeek's capabilities.
"People call it 'the glory of made-in-China', and say it shocked Silicon Valley, so I downloaded it to see how good it is." – Fiona Zhou
As DeepSeek continues to gain traction, it raises significant questions about global AI leadership. Some experts suggest that China may be closing the gap with the United States in AI development. Liang Wenfeng remains optimistic about China's ability to cultivate top talent domestically.
"The Top 50 talents in this field might not be in China, but we can build people like that here." – Liang Wenfeng