In recent years, U.S. productivity growth has been significantly lower than average growth from 1930 to 2000. Recent developments in artificial intelligence (AI) with large language models (such as ChatGPT, Claude and Llama) present a ray of optimism for future productivity growth. There are three primary considerations for evaluating productivity growth for any given set of technologies: (1) Improvements in productivity for individuals because of AI; (2) the extent to which AI is adopted by individuals across different firms, industries and cities, and (3) the speed of AI’s diffusion. Several recent studies have tried to estimate the improvement in the productivity level for workers using large language models in online experiments. In this blog post, we highlight the diffusion patterns of some recent technologies.