In a recent statement, Robin Li, the chairman and CEO of Baidu, highlighted the superior Chinese language performance of ERNIE Bot over OpenAI's GPT-4. He envisioned a future where traditional programming jobs become obsolete, with English and Chinese emerging as the primary languages for coding.
A new frontier in AI development
According to MyDrivers and IT Home reports Li has articulated Baidu's artificial intelligence (AI) strategy as being built upon four foundational pillars: chips, architecture, models, and applications. Despite facing restrictions from US export controls on the chip front, Baidu sees vast opportunities for innovation in the remaining three sectors.
Li confidently stated that ERNIE Bot's efficiency does not hinge on the most advanced chips, thanks to Baidu's alternative solutions and stocked inventories. He reassured that Baidu remains committed to continuously enhancing its large language models (LLMs) and exploring new avenues beyond US-based suppliers.
He further asserted that Baidu's ERNIE Bot outperforms GPT -4 when it comes to understanding and generating Chinese content. However, he acknowledged GPT-4's superiority in dealing with English material, attributing this to its extensive training in English language data.
The evolving landscape of AI
Li observed that the progression of AI technology has not been as rapid as some might have expected. He reminded me that the quest for AI has been ongoing for 70 years, with promises of its achievement cropping up every decade yet falling short of expectations.
However, Li remains optimistic about AI's transformative potential in the near future, particularly its potential to revolutionise the workforce. He anticipates a world where programming becomes so intuitive that it requires only basic language skills, foreseeing a reduction in the need for traditional programming roles. In this new era, Li imagines English and Chinese as the dominant languages for coding, simplifying the technological landscape.
Towards a more efficient use of resources
Li also touched upon the proliferation of companies investing in developing LLMs, questioning the necessity of such widespread duplication. He argued that this trend leads to an inefficient use of societal resources and labour, suggesting a more streamlined approach to AI development might be beneficial.
In conclusion, as Robin Li shares, Baidu's vision for AI's future paints a picture of a world where language barriers in technology are minimised and innovation flourishes despite external constraints. As Baidu continues to push the boundaries of AI with ERNIE Bot, the implications for global technology and society at large could be profound.
