The Biggest Issue with AIGC Today: Lack of Originality

On September 7, Li Di, CEO of Xiaobing Company, spoke at the 2023 Bund Conference, presenting insights on "Potential Business Models for Generative Artificial Intelligence." In an interview with Cover News, he addressed the practical implementation of AIGC and shared his views on the industry's challenges and opportunities.

Li Di highlighted a significant hurdle in the sector: a lack of diversity, with many companies adopting similar approaches. He remarked, "Innovation isn't just about catching up; it's about identifying unique ways to leverage strengths." Continuous improvement of generative AI models is essential, and maintaining a calm perspective on technology can provide valuable insights that drive advancements, particularly in helping with writing tasks.

At the recent "Baidu Smart Cloud" conference, Baidu chairman Robin Li called for companies to collaborate on "the application of large models" rather than competing solely in their development. Li Di supported this view, emphasizing that redundancy in both large models and their applications stifles innovation. He pointed out that the industry's repetitiveness contradicts the innovative spirit.

With the high costs linked to large models, Li Di predicted that companies would need to quickly identify viable paths forward, as prolonged spending without results is not feasible. "The landscape resembles a game with 20,000 cards, and uncertainty remains regarding the outcomes," he explained, suggesting a shift in market dynamics.

Reflecting on recent trends, Li Di acknowledged a prevailing misconception that larger models guarantee better success, a notion yet to be validated. He used Xiaobing's launch in 2014 as a case study, illustrating that its evolution, even at a nascent stage of technology, was not a linear progression. "Technology’s evolution is marked by fluctuations, not simply a straight line," he noted.

For startups exploring large models, Li Di advised caution. He warned that "if a startup's innovation stays linear, it risks simply chasing progress. However, labeling itself as a large model startup could lead to misalignment if market conditions change."

Ultimately, the decision to pursue open-source or closed-source ecosystems hinges on business goals. Li Di remarked, "Open-source models have shown their vibrancy. However, in the realm of large models, this openness has led to market confusion, favoring those centered on application development."

Expressing optimism about the mainstream adoption of future AIGC applications, he argued that while large models can't address every problem, they are integral to a comprehensive system requiring ongoing support. Currently, successful AI models for consumers or businesses remain elusive, highlighting untapped market potential.

Looking ahead, Li Di predicted that a national-level blockbuster application in AIGC is likely to emerge, driven by the rapidly evolving technology landscape. He concluded, "Today’s environment offers companies a range of options, moving beyond a singular path to innovation."