"MOSS is still a very immature model," the Fudan University Team in China said. (Representational)
A team from China's Fudan University apologised on Tuesday after a ChatGPT-like chatbot platform they developed crashed hours after it launched to the public, due to a sudden surge of traffic.
The team's announcement on Monday of the platform they called MOSS instantly went viral on Chinese social media, generating tens of millions of hits on China's Twitter-like Weibo. State media described it as the first Chinese rival to OpenAI's hit ChatGPT platform.
But MOSS, which bears the same name as a superintelligent quantum computer in the Chinese sci-fi blockbuster "Wandering Earth 2", crashed soon after and by Tuesday the team said it would no longer be open to the public.
The launch of MOSS and the public response to it underlines the fervour for generative AI and ChatGPT in China and the challenges its domestic industry faces, as several of China's top universities and tech companies race to produce a Chinese version of the Microsoft-backed chatbot.
While the Fudan University team had on Monday initially described MOSS as a conversational language model like ChatGPT, on Tuesday they played down the comparison, saying they had much to improve.
"MOSS is still a very immature model, it is still has a long way to go before reaching ChatGPT. An academic research lab like us is unable to produce a model whose ability nears ChatGPT," a statement published on its website said.
"Our computing resources were not enough to support such large traffic and as an academic group we do not have sufficient engineering experience, creating a very bad experience and first impression on everyone, and we hereby express our heartfelt apologies to everyone."
ChatGPT, the fastest-growing consumer application in history, has also crashed several times due to heavy traffic.
While few users were able to share their experiences of the platform before the crash, a journalist from the Shanghai Observer shared a detailed account of an interaction with MOSS and said that the chatbot's English was better than its Chinese.
The team's leader, Qiu Xipeng, a professor at Fudan's School of Computer Science, told the Shanghai Observer on Monday that the main gap between MOSS and ChatGPT was that the number of parameters put into MOSS' language training was an order of magnitude smaller than ChatGPT.
Qiu did not immediately respond to a request for further comment.
(Except for the headline, this story has not been edited by NDTV staff and is published from a syndicated feed.)