This talk will present three works on the impact of Large Language Models, particularly on digital public goods, labor markets, and the potential to model economic behavior. First, we will look at how ChatGPT may be diminishing the flow of publicly shared, human-generated content on platforms like Stack Overflow, which is crucial for programming insights. Our analysis reveals a 16% decrease in weekly contributions, threatening the future of AI training data and open knowledge sharing. This trend underscores concerns about preserving digital public resources and competitive markets. Second, we will delve into preliminary results on LLMs’ influence on job demand, highlighting a 29% reduction in roles LLMs can replace. Moreover, complementary jobs face both increases and decreases in demand. These findings contest the notion that technological progress always boosts labor demand for augmented skills, suggesting a more nuanced reality.