OpenAI's focus on LLMs set back AGI progress by 5-10 years, says ARC Prize co-founder

François Chollet, co-founder of the ARC Prize, stated that OpenAI has set back progress toward artificial general intelligence (AGI) by five to ten years by focusing on large language models (LLMs). Chollet argues that LLMs, based on neural networks, are a dead end for achieving human-like intelligence, despite their increasing power and usefulness for businesses. He believes this heavy investment has diverted resources from other, more promising architectures. The ARC-AGI benchmark, created by Chollet, is designed to measure true AGI progress, and no LLM has come close to passing it. Meanwhile, some tech executives have used AI threat warnings for marketing, especially ahead of potential IPOs.