Quote of the day by ARC Prize co-founder François Chollet: "OpenAI basically set back progress to AGI by five to 10 years"

ARC Prize co-founder François Chollet recently declared that OpenAI's work has jeopardized the path to artificial general intelligence (AGI) by a period of five to ten years. Chollet, who co-founded the ARC Prize, explained that language models, which are based on neural networks, are a dead end when it comes to achieving AGI. While he believes there is room for improving LLMs in terms of their power, autonomy, and usefulness to businesses, he argues that this is not the same as achieving true AGI.

Chollet's concern is that pouring funding into LLMs at the expense of other avenues or architectures is ignoring the route to AGI, potentially setting progress back by up to a decade. Intelligent systems remain a hot topic, with recent disclosures from OpenAI and Anthropic sparking fears of potential threats to humanity. However, Chollet believes that the purpose of the ARC-AGI benchmark is to determine genuine progress toward AGI, and thus far, no LLMs have come close to achieving this milestone.