Moonshot AI founder builds business in the mould of ByteDance, OpenAI - South China Morning Post

Yang is considered “an entrepreneur with potential”, said Cao Xi, founding partner at Hong Kong investment company Monolith Management and a former partner at Sequoia Capital China. Monolith is among the group of venture capital firms backing Moonshot AI, which declined to provide details of the latest funding round.

Moonshot AI’s Yang Zhilin, 31, has been honing his skills in the development of artificial intelligence large language models. Photo: Weibo

In an interview with mainland media in February, Yang shared his vision for Moonshot AI, which was to develop into a business that “combines the technology idealism of OpenAI and business philosophy of ByteDance”.

While Moonshot AI did not make him available for an interview, Yang’s stellar academic track record and industry experience have been closely scrutinised by investors.

A computer scientist who earned his doctorate at the prestigious Carnegie Mellon University in Pennsylvania after graduating from Tsinghua University, Yang worked at Facebook owner Meta Platforms and Google Brain during his doctoral study.

During his years in the United States, Yang served as co-author of papers with some of the world’s most influential information technology experts, including Meta chief AI scientist Yann LeCun, Google research scientist Quoc V Le and Canadian computer scientist Yoshua Bengio. Through those years, Yang became recognised as one of the world’s foremost young authorities on LLMs.

In 2020, Yang worked with Huawei Technologies on an early version of the firm’s Pangu AI model. The following year, he led a team to work on the development of the Wudao LLM at the Beijing Academy of Artificial Intelligence. These collaborations on LLMs gave his team early exposure to all the potential issues in building AI models, Yang told mainland media.

A white Yamaha digital piano is displayed at the entrance of Moonshot AI’s headquarters in Beijing. Photo: Ben Jiang

Soon after ChatGPT’s launch, Yang made a trip to the US in December 2022 to learn about this GenAI application and to recruit talent for his own AI venture. He told mainland media that he felt it was time to “ride the wave” in GenAI.

Along with Tsinghua University schoolmates Zhou Xinyu and Wu Yuxin, Yang co-founded Moonshot AI in March last year. He managed to raise US$60 million as initial funding and assembled a team of around 40 AI specialists within the first three months of the company.

Moonshot AI experienced a surge in interest after it launched its ChatGPT-like product last October. Kimi chatbot – built on the firm’s self-developed Kimi LLM – quickly gained popularity among mainland users, thanks in large part to its capabilities in handling extremely long text queries of up to 2 million Chinese characters.

Kimi’s user base has steadily grown, as Moonshot AI’s “lossless long context is better than OpenAI in many applications”, Yang told mainland media.

The name of his company, according to Yang, came from his love for classic rock music during his years at Tsinghua University, where he had formed a band. During the Post’s visit last month to Moonshot AI headquarters in Beijing, the 1973 album Dark Side of the Moon, by English rock band Pink Floyd, was seen on top of a white Yamaha digital piano at the office’s entrance.

READ SOURCE