Particle.news
Download on the App Store

Hugging Face CEO Warns of ‘LLM Bubble,’ Predicts Shift to Smaller Models

He says AI’s broader trajectory remains intact despite a likely correction in giant chatbots.

Overview

  • Clem Delangue told an Axios event that today’s exuberance centers on large language models and said the bubble might burst next year.
  • He contends many tasks are better served by cheaper, faster, specialized models rather than one model meant to handle everything.
  • As an example, he pointed to bank customer chatbots that can run on a company’s own infrastructure without relying on a general-purpose LLM.
  • Delangue said a downturn in LLM valuations could affect Hugging Face to a degree, but he argued the wider AI field remains diversified.
  • He emphasized financial prudence, noting Hugging Face still holds roughly half of the $400 million it has raised and is focused on long-term sustainability.