OpenAI reports near breakthrough with "reasoning" AI, reveals progress framework

OpenAI has introduced a five-tier system to assess its progress towards developing artificial general intelligence (AGI), with the ultimate goal of creating AI that can perform tasks like a human without specialized training. The system ranges from basic chatbots to AI capable of managing entire organizations. OpenAI's current technology, like GPT-4, is placed at Level 1, while they are reportedly on the brink of achieving Level 2, known as "Reasoners," which involves human-level problem-solving abilities. The company's CEO, Sam Altman, has expressed confidence in achieving AGI within this decade. However, the classification system is still a work in progress, subject to feedback and potential refinements. OpenAI's framework is seen as a communication tool to attract investors, reflecting the company's ambitious objectives rather than a precise measure of technical progress. The broader AI research community lacks consensus on measuring progress towards AGI, highlighting the challenges in defining and achieving such a goal.