The article explores the concept of a potential "half-life" for the success rates of AI agents, examining whether the effectiveness of these agents diminishes over time and what factors contribute to this phenomenon. It discusses implications for AI development and the sustainability of AI performance in various applications.

Toby Ord explores a mathematical model explaining the declining success rates of AI agents on longer tasks, suggesting that each agent can be characterized by its own "half-life." The findings from Kwa et al. (2025) indicate that as task duration increases, the probability of success decreases exponentially, with implications for understanding AI capabilities over time. The study highlights the importance of measuring performance across various tasks and the challenges of generalizing results beyond the specific task suite used in the research.