Clem Delangue, CEO of Hugging Face, which serves as a repository for machine learning models, recently commented on the current state of artificial intelligence. At an event organized by Axios, he stated that we are in a large language model (LLM) bubble that could burst as early as next year. According to him, however, this does not mean the end of artificial intelligence as a whole. Delangue emphasized that LLMs are only one part of the broader field of AI, which includes applications in biology, chemistry, and image, audio, and video processing. He believes that major growth in these areas is only just beginning.
Delangue described how all the attention, funding, and effort are currently being directed toward the idea that a single enormous model, powered by massive computing resources, can solve every problem for every company and individual. In his view, this is causing the market to overheat. Instead, he predicts a future full of diverse models tailored to specific tasks and specialized in particular fields.
Hugging Face and Its Role in AI
Hugging Face, the company led by Delangue, serves as a GitHub-like platform for sharing machine learning models. It hosts both large models from companies such as OpenAI and Meta—such as gpt-oss and Llama 3.2—as well as smaller, modified versions created by developers for specific needs. Delangue sees this approach as the key to success. According to him, companies and researchers will rely on a variety of specialized models that solve specific problems more efficiently than a single versatile giant.
This view is also supported by research firm Gartner, which predicted in April that companies would be three times more likely to use small, task-specific models than large, general-purpose LLMs. The reason is the need for greater accuracy in workflows and better adaptation to specific data. Delangue thus suggests that the era of one model for everything is ending and will be replaced by diversity that delivers tangible benefits in practice.
Examples of Investments Beyond LLMs
Delangue's optimism beyond the field of LLMs is also underscored by recent developments. For example, Jeff Bezos, former CEO of Amazon, became co-CEO of a new startup focused on machine learning applications in engineering and manufacturing. The project is launching with more than $6 billion in funding, equivalent to approximately CZK 138 billion. This shows that AI investment extends beyond language models and is focusing on practical applications in industry.
Delangue emphasizes that while LLMs attract all the attention, real progress will come in other areas. His remarks at the Axios event serve as a reminder that the term AI is much broader and that we are only beginning to discover where these technologies will lead.
Sources: techcrunch.com and arstechnica.com
