Chinese tech companies are making significant strides in the development of Large Language Models (LLMs), aiming to close the gap with OpenAI. A recent post from an Alibaba researcher provides a rare insight into the daily life of building LLMs at the e-commerce giant, shedding light on the parallel experiences of researchers in China and the United States.
Key Takeaway
Chinese tech companies, including Alibaba, are intensively focused on developing Large Language Models (LLMs) to compete with global counterparts. The dedication and long hours put in by researchers reflect the urgency and speed with which these companies are rolling out LLMs, aiming to establish a strong presence in the AI space.
Similarities in Daily Routines
Binyuan Hui, a natural language processing researcher at Alibaba’s LLM team Qwen, shared his daily schedule, revealing striking similarities with the experiences of OpenAI researcher Jason Wei. Both Hui and Wei start their day with meetings, followed by coding, model training, and brainstorming with colleagues. Even after returning home, they continue to run experiments and work on enhancing their models late into the night.
Differences in Leisure Time
While their work routines are similar, Hui’s characterization of leisure time differs from Wei’s. Hui spends his leisure time reading research papers and browsing the web to stay updated on global events, in contrast to Wei’s reported habit of enjoying a glass of wine after work.
Intense Work Regime in China’s LLM Space
The demanding schedule of researchers like Hui reflects the intense competition in China’s LLM space, where top tech talent is joining companies to build competitive AI models. This drive to match or outpace Silicon Valley companies is evident in the dedication of researchers and the long hours they put in.
Urgency and Speed in Rolling Out LLMs
Chinese tech firms, including Alibaba, are rapidly rolling out LLMs, with Qwen open sourcing foundation models trained with both English and Chinese data. The team has also introduced commercial applications, integrating Qwen into Alibaba’s enterprise communication platform and online retailer Tmall.
Alibaba’s Niche and Multilingual Expansion
Amidst fierce competition, Alibaba is striving to carve out a niche in the LLM space. The company’s multilingual move includes the release of an LLM for several Southeast Asian languages, potentially leveraging its cloud computing business and ecommerce platform to introduce the model to a wider audience.
