Chinese teen AI prodigy Nathan Chen said to be joining US start-up Discovery Loop
Summary
Nathan Chen Guangyu is reported to be joining Discovery Loop, a U.S. AI start-up, after leaving Moonshot AI last week. The 17-year-old researcher helped develop ways to make Moonshot’s Kimi models more cost-efficient during his time there, which lasted less than a year. Discovery Loop was founded by four former senior Google researchers, including Jeff Dean. The company aims to use AI to automate scientific research. Neither Chen nor the start-up has confirmed the reported hiring.
Key Facts
- Discovery Loop started two months ago and has hired researchers from major U.S. technology companies, including former Google engineers.
- Radical Ventures and Khosla Ventures led the start-up’s first funding round. Alphabet, Google’s parent company, also invested.
- Business Insider reported last month that Discovery Loop was seeking a company value of about $50 billion in another funding round.
- At the Agentic AI Summit at the University of California, Berkeley, in early August, Jeff Dean described how AI could run thousands of scientific experiments at the same time. It could then use the results to decide which experiments to run next.
- Du Yulun, a senior researcher at Moonshot, said he had several long conversations with Chen about leaving. He believed Chen should have the chance to explore his curiosity at his age.
- Published KST
Moonshot AI's logo is shown in this illustration taken on Sept. 5. Reuters-Yonhap
Just days after announcing that he was leaving Moonshot AI, young Chinese AI expert Nathan Chen Guangyu appears to have found his next job.
On Tuesday, Arfur Grok, a widely followed anonymous social media account that tracks AI talent moves in the U.S., posted that Chen, a former Moonshot AI researcher, was joining Discovery Loop. The AI start-up was founded by four former senior Google researchers, including Jeff Dean, who previously served as Google’s chief scientist.
The start-up began operating just two months ago. It has quickly drawn attention from the industry because of its well-known founders and its ambitious goal of using AI to automate scientific research.
Discovery Loop has already hired several well-known researchers from major U.S. technology companies, including other former Google engineers. At just 17 years old, Chen would likely be one of its youngest team members.
Chen did not immediately respond to a request for comment on Tuesday. Discovery Loop said it does not comment on staff matters.
Chen is considered one of China’s most promising young AI researchers. When he announced last week that he was leaving China for the U.S., many people began guessing where he would work next.
Although Chen worked at Moonshot for less than a year, he helped develop important technical improvements that made its main Kimi models cheaper to run. These improvements helped the models compete with major U.S. AI systems, including OpenAI’s GPT and Anthropic’s Claude.
Moonshot has become one of China’s leading AI model developers. Its K3 model has attracted worldwide attention because of its strong abilities and the fact that its model weights are openly available.
After Chen announced his departure, Moonshot senior researcher Du Yulun wrote on social media that he and Chen had “several long conversations about his decision” to leave. In the end, Du said it was “best for [Chen] to explore and follow his new curiosity as much as possible at his age”.
Discovery Loop has not released any products yet. However, Business Insider reported last month that the company was seeking a valuation of about $50 billion in another funding round.
Radical Ventures and Khosla Ventures led its first funding round. Alphabet, the parent company of Google, also invested.
Nathan Chen Guangyu / Image captured from X
In one of his first public appearances after leaving Google, Dean spoke at the Agentic AI Summit at the University of California, Berkeley, in early August. He said the start-up’s work would be similar to efforts to achieve recursive self-improvement (RSI). RSI means using AI to improve AI itself, and it is a stated goal of leading AI companies in both the U.S. and China.
Dean explained that AI could one day run thousands of scientific experiments on its own at the same time. It could study the results and choose the best experiments to run next. This could shorten scientific research that takes years to just hours or even minutes.
“By building models that are really good at understanding many, many different domains of science and engineering, you can get PhD-level expertise in a model across many different domains,” he said.
Read the article at SCMP.