In doing so, Clem Delangue, co-founder and CEO of Hugging Face, raised big concerns. What’s more, he is alarmed by the state of play in the big language model (LLM) sector. Delangue has 15 years of experience in AI. Llull thinks the LLM bubble is already on the verge of bursting, possibly as soon as next year. He points out that LLMs, despite receiving the hype and focus of the world (and a lot of money), aren’t the magic bullet for every AI problem.
In an interview with TechCrunch, Delangue recently outlined his vision for the future of AI. His big takeaway was that smaller, more specialized models are going to be more widely adopted in the public, academic, and private sectors. He thinks these models are cost effective and efficient. They allow industries to run on their current bandwidth without the need for massive investments.
“I think we’re in an LLM bubble, and I think the LLM bubble might be bursting next year,” he stated. Delangue’s comments are a great illustration of the broader issue at work here—how quickly the landscape of AI technology continues to evolve.
He also made an important second point—that the AI sector is already diversified and evolving beyond the confines of LLMs. “But ‘LLM’ is just a subset of AI when it comes to applying AI to biology, chemistry, image, audio, and video,” he explained. This increased diversification is a clear indication of the need for tailored solutions. These solutions have to address unique industry challenges rather than just relying on the one, sole model.
Delangue acknowledged his worries that the narrowly LLM-first approach will create a short-term mindset in companies that only look to capitalize on the hype cycle. “I think a lot of people right now are rushing — or maybe even panicking — and taking a really short-term approach to things,” he remarked. He’s concerned that in this rush we risk losing sight of how smaller models can better provide big solutions.
Even the CEO of Hugging Face, a firm thriving in the LLM boom, admitted that the impending collapse of the LLM bubble would be painful for his company. He viewed that as an opening to encourage a more grounded understanding of AI’s potential. “I think all the attention, all the focus, all the money, is concentrated into this idea that you can build one model through a bunch of compute and that is going to solve all problems for all companies and all people,” he noted.
Though net negative on LLMs, Delangue is still bullish about the future of AI. “I think we’re at the beginning of it, and we’ll see much more in the next few years,” he stated. But he envisions a reality where a multiplicity of models, each one fine-tuned to solve specific problems in specific sectors, are able to exist.