In a surprising revelation, Google software engineer François Chollet has criticized OpenAI for allegedly hindering the progress of Artificial General Intelligence (AGI) research by 5 to 10 years. He voiced his concerns during a conversation with podcaster Dwarkesh Patel, highlighting that OpenAI's focus on Large Language Models (LLMs) has diverted crucial resources and attention away from broader AGI research.
Chollet explained that a few years ago, the field of AI research was more open and collaborative, with state-of-the-art results being widely shared and published. However, he now observes a significant shift, attributing this change to OpenAI’s influence, which he claims has led to a "complete closing down of frontier research publishing." This shift has, in his view, stifled innovation and slowed down the overall progress towards achieving true AGI.
Moreover, Chollet argues that OpenAI's promotion of LLMs has created a burst of hype that has "sucked the oxygen out of the room." This has resulted in many researchers and resources being channeled into LLM-related projects, thereby neglecting other promising avenues in AGI research. His comments underscore a broader concern within the AI community about the allocation of efforts and the direction of current research trends.
Chollet and Mike Knoop, the co-founder of Zapier, have also introduced the $1 million ARC-AGI Prize to encourage more focus on frontier AGI research. This competition aims to measure AGI's ability to acquire new skills and solve novel problems, pushing the boundaries beyond what current LLMs can accomplish. Through this initiative, they hope to reignite interest in diverse research directions and accelerate the journey toward true AGI.
