Mindscape 330 | Petter Törnberg on the Dynamics of (Mis)Information

In this episode of the Mindscape podcast, Sean Carroll interviews Petter Turnberg, a professor of computational social science with a background in physics, about the application of physics-like reasoning to social science, particularly in understanding social media dynamics. They discuss the concept of "physics envy" in social sciences, agent-based models, and the use of large language models (LLMs) to simulate human interactions on social media platforms. Turnberg shares insights from his research, including a modern take on Schelling's segregation model applied to online communities, and a study using LLMs to model social media behavior, revealing the emergence of echo chambers, attention inequality, and polarization even in the absence of algorithmic manipulation. The conversation explores the challenges of mitigating these negative outcomes and the broader implications of social media on political discourse and information ecosystems.