Hugging Face CEO says we’re in an ‘LLM bubble,’ not an AI bubble

Hugging face Co-founder and CEO Clem Delangue says we are not in an AI bubble, but an “LLM bubble” – and it may be about to burst. At an Axios event on Tuesday, the entrepreneur behind the popular AI platform and community site agreed that bubble talk is today’s “trillion dollar question,” but said he doesn’t believe the future of AI will be in jeopardy if the bubble bursts.

Instead, as Delangue sees it, the large language models (LLMs) – such as those powering ChatGPT, Gemini, and other chatbots – are receiving outsized attention, and that attention may not last.

“I think we’re in an LLM bubble, and I think the LLM bubble will burst next year,” Delangue explains. “But ‘LLM’ is only a subset of AI when it comes to the application of AI to biology, chemistry, imaging, audio, [and] video. I think we are just getting started and we will see a lot more in the coming years,” he noted.

One problem, he argued, is that LLMs are not the right solution for everything, and that smaller, more specialized models will become increasingly common in the future.

“I think all the attention, all the focus, all the money is concentrated in this idea that you can build one model using a lot of computing power and it will solve all the problems for all the companies and all the people,” Delangue said. “I think the reality is that in the coming months, in the coming years, you’re going to see sort of a multitude of models that are more customized and specialized, that are going to solve different, different problems.”

As an example, he presented the use case of a chatbot for banking customers.

“You don’t need it to tell you about the meaning of life, right? You can use a smaller, more specialized model that will be cheaper, that will be faster, that you can maybe run as an enterprise on your infrastructure, and I think that’s the future of AI,” Delangue emphasized.

The Hugging Face founder admitted that the bursting of an LLM bubble could impact his business to some extent, but noted that the AI ​​industry is so big that it is already diversified. That means that even if one part of the industry is overvalued, such as LLMs, it won’t have a huge impact on the AI ​​field itself or its business.

Moreover, he noted, Hugging Face still has half of the $400 million it raised in the bank. This cautious approach to spending represents a different strategy than the some other AI companies are doing this nowadays, especially in the LLM space.

“In AI standards, that’s called profitability because the other guys aren’t spending hundreds of millions. It’s clearly billions of dollars,” he said.

In comparison, Hugging Face takes a more capital-efficient approach.

“I think a lot of people right now are rushing – or maybe even panicking – and taking a very short-term approach to things. I’ve been in AI for 15 years now, so I’ve seen some cycles,” Delangue added. “And so we learn from that and try to build a sustainable, long-term impactful company for the world.”