Robin Li Yanhong, founder and CEO of Baidu, China’s largest search engine, said the country has too many language models and too few practical applications. Yanhong made the announcement during a recent panel discussion at the World Artificial Intelligence Conference (WAIC) in Shanghai, as reported by the South China Morning Post.
“In 2023, intense competition has emerged in China among more than 100 LLMs, which has led to a great waste of resources, especially computing power,” Li said. “I noticed that many people still mainly focus on fundamental models. But I would like to ask: what about real-world applications? Who has benefited from it?”
The World Intellectual Property Organization (WIPO) announced last Friday that China has outpaced the United States in artificial intelligence patents by a ratio of six to one over the past decade. However, WIPO data also showed that the country lags behind in terms of citations, with the Chinese Academy of Sciences being the only Chinese organization to make the list of the top 20 organizations with the most citations in the field of research.
Publicly available LLMs in China must be approved by Chinese regulatory authorities, to ensure that the Chinese Communist Party (CCP) can effectively control the Chinese people. More than 200 AI companies have applied for licenses as of March 2024, of which 117 have received approval from Beijing. Having so many LLMs means they are all fighting for a piece of the pie, and not everyone will win. Yan Junjie, CEO of AI startup MiniMax, said he “expects major consolidation in the industry in the future, with LLMs being primarily developed by just five companies.”
Many large companies have begun rushing to capture the market that OpenAI will leave when its API is no longer accessible in China, starting July 9. The largest tech companies — Tencent, Baidu and Alibaba — have begun offering discounts and incentives to entice customers to buy their products, something smaller companies may not be able to sustain.
While competition is good for any market, too much of an offering can also lead to decision fatigue among customers, who are simply overwhelmed by the number of services offered at similar prices. Large companies would likely win in this case, as they have more working capital that they can use to acquire or eliminate smaller competitors. As Bernard Leong, CEO of Singapore-based Dorje AI, put it, “There will likely be a bloodbath in the big language models and I think there will probably be very few players left.”
