The Download: a biological de-aging contest and why LLMs don't reason (www.technologyreview.com)

A new and innovative competition has been launched where approximately 500 participants will aim to reverse their biological age over six months, using various health measures and competing for a spot on a leaderboard. This initiative is significant as it potentially redefines how we understand aging and health, moving beyond chronological age. The biological age focuses on organ health and physiological measures, making it a more relevant metric in assessing well-being and longevity. However, questions remain about how effectively biological age can be measured and the feasibility of consistently reversing it. In a separate yet related discussion, Thore Graepel, a former member of the AlphaGo team at Google DeepMind, shared insights on the limitations of current large language models (LLMs), emphasizing that they lack the reasoning capability that characterized AlphaGo’s groundbreaking performance. Graepel argues for a new approach to machine reasoning, inspired by the architecture of AlphaGo, which could lead to more advanced and capable AI. This emphasizes an ongoing debate within the AI/ML community about the need for improved frameworks that allow AI systems to think and reason more like humans, a cornerstone for future advancements in AI technology.