This week’s podcast asks about the benefits and problems associated with both hard “mathematical” prediction and soft “storytelling” prediction. We discuss the limits of mathematical prediction in terms of theory, randomness, chaos, and non-computability. We discuss the limits and benefits of storytelling and scenario planning as predictive tools as well, and we also discuss the self-reference problem, which can apply to both types of prediction. Finally we discuss the fictional discipline of psychohistory as imagined by Asimov, and wonder whether truly working prediction machines could exist without lacking transparency.
Relevant Links
- Article with video: Limits of prediction: stochasticity, chaos, and computation
- Article about social data-based predictions: Smart People Prefer Curly Fries
- The Halting Problem on Wikipedia
- Book: Isaac Asimov’s Foundation