Hugging Face co-founder and CEO Clem Delangue says we are not in an AI bubble, but an LLM bubble, and it may be poised to pop. At an Axios event, the entrepreneur behind the popular AI platform agreed that bubble talk is a major topic, but he said he does not believe AI’s future is at risk if the bubble bursts. Instead, Delangue sees large language models, like those powering ChatGPT and Gemini, as receiving outsized attention that may not last.
He explained that he believes we are in an LLM bubble that might burst next year. However, he noted that LLMs are just a subset of AI. When it comes to applying AI to biology, chemistry, image, audio, and video, he believes we are at the beginning and will see much more development in the next few years.
One issue he raised is that LLMs are not the right solution for everything. He argued that smaller, more specialized models will see increased adoption in the future. Delangue stated that all the attention, focus, and money is concentrated on the idea that one large model can solve all problems for all companies and people. He believes the reality is that we will see a multiplicity of models that are more customized and specialized to solve different problems.
As an example, he suggested a banking customer chatbot. He pointed out that such a chatbot does not need to discuss the meaning of life. A smaller, more specialized model would be cheaper, faster, and could be run on a company’s own infrastructure. Delangue believes that this is the future of AI.
The Hugging Face founder admitted that an LLM bubble bursting could impact his company to some extent. However, he noted that the AI industry is already so large and diversified that even if one portion is overvalued, it will not have a massive impact on the AI field itself or his business.
He also shared that Hugging Face still has half of the 400 million dollars it has raised in the bank. This cautious approach to spending represents a different strategy from other AI companies, especially in the LLM space. Delangue said that in AI standards, having that capital is called profitability, while other companies are spending billions of dollars.
By comparison, Hugging Face is taking a more capital-efficient approach. Delangue added that he believes many people are rushing or panicking and taking a short-term approach. Having been in AI for 15 years, he has seen some of the cycles. He said the company is learning from that and trying to build a long-term, sustainable, and impactful company for the world.