Caution Against Wasteful Resource Allocation
In a stark warning, Robin Li, the CEO of Baidu , cautioned against the widespread rush to develop large language models in the country. Li, speaking at an industry forum in Shenzhen, emphasized the potential for a wasteful allocation of resources and urged companies to shift their focus towards practical applications of artificial intelligence.
The surge in interest in generative AI in China, triggered by the release of OpenAI’s ChatGPT last year, has prompted numerous established companies and startups to enter the field. However, Li expressed concern over the lack of viable business models for companies engaged in developing large language models, leading to growing apprehensions in the market about a possible industry shakeout.
“I’ve observed a phenomenon in China where many industries, companies, and even cities are investing in hardware, stocking chips, and building computing centers to train proprietary large models from scratch,” Li stated.
Foundational Role of Large Language Models Baidu
While acknowledging that a large language model serves as a foundational element similar to an operating system, Li stressed that developers must eventually rely on a limited number of these models to create various native applications. Continuously redeveloping foundational large models, according to Li, results in a significant waste of social resources.
Despite the proliferation of large models in China, the CEO highlighted the scarcity of AI applications developed based on these models. Citing a third-party report, he noted that as of October, there were 238 large language models, up from 79 in June.
Baidu, too, has contributed to this landscape with its large language model called Ernie, which was made publicly accessible in August. In a bid to stay at the forefront, Baidu recently unveiled Ernie 4.0, the latest iteration of its generative AI model, building on the foundation laid by the initial version introduced in March. However, Li’s cautionary remarks underscore the need for a strategic shift towards practical and impactful AI applications to avoid unnecessary resource expenditure in the race for AI dominance.
Read More (AI)
