JAKARTA - Hugging Face co-founder and CEO Clem Delangue stated that the world is currently not in an “AI bubble,” but rather an “LLM bubble” — and that bubble may soon burst.
Speaking at an Axios event on Tuesday, November 18, Delangue called the tech bubble debate a “trillion-dollar question.” However, he believes the future of AI as a whole will not be threatened even if the LLM bubble does burst.
According to Delangue, large language models (LLMs) like ChatGPT, Gemini, and various other chatbots are receiving too much and disproportionate attention, so the hype is unlikely to last.
“I think we’re in an LLM bubble, and it might burst next year,” Delangue said. “But LLM is just a subset of AI — when AI is applied to biology, chemistry, images, audio, video — it’s just getting started, and we’ll see more innovation in the next few years.”
LLM Isn't the Solution to Everything
Delangue argues that LLM is often pushed as a universal solution, but that's a mistake. He predicts that smaller, more specialized models will become more widely used because they are more efficient.
"All the attention, focus, and money right now is concentrated on the idea that one big model can solve all problems. In reality, in the next few months or years, many more specialized models will emerge to address different needs," he explained.
He cited bank chatbots as an example. "Bank chatbots don't need to answer the meaning of life. You just need a smaller model that's cheaper, faster, and can be run by the company itself. That's the future of AI," he said.
Impact on Hugging Face
Delangue acknowledged that the bursting of the LLM bubble might impact his company, but believes the AI industry is already highly diversified. This means that if some sectors—like LLM—are overvalued and ultimately collapse, the impact on the entire AI industry will be minimal.
Furthermore, Hugging Face still holds half of the $400 million (Rp6.6 trillion) investment it raised. “By AI standards, that’s called profitability,” he said, noting that other AI companies are spending not hundreds of millions, but billions of dollars.
A More Efficient Strategy
Hugging Face is taking a more efficient approach to capital than many other players in the LLM space.
“Many people now seem to be rushing—even panicking—and taking a very short-term approach. I’ve been in AI for 15 years and have seen the cycles,” Delangue said.
He added that Hugging Face is striving to build a sustainable company with long-term impact, not simply riding the LLM hype.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
