Boris Smus

interaction engineering

Future Babble

The best predictors are foxes, but even they do poorly. Interesting weakness: foxes tend to be more swayed by scenario planning. Because they start vividly believing all of the possible scenarios, so even unlikely ones become more possible. Also interesting is that popular pundits tend to be hedgehogs because they are prone to more extreme predictions that the audience likes.

Lots of good interesting failed predictions, just like I asked!

Prediction is fundamentally hard because of chaos.

Good predictors: - Peter Schiff predicting housing crash and recession. - Ross Anderson predicting y2k mildness. - Alan Barnes, pioneering work around predictions for Canada. - James Fallows, essay

Bad predictors: - David Brooks predicting nonsense. - Paul Erlich predicting ridiculousness in End of Affluence and Population Bomb - F. E. Smith, The World in 2030 - Arthur Clark: profiles of the future 1962

Idea: correlation between difficulty of times and likelihood to turn to superstition and religion. I should keep that in mind in the future. Russian superstition and magical thinking is not caused by leading a happy life. I should be kinder to those afflicted.

Interesting conclusion: accurate predictions aren't necessary for good choices. The goal is to make choices that are good in any event. If needed, skew according to probability.

How to predict better:

  • Be fox like: aggregate different sources.
  • Meta cognition: consider how you arrive at your decisions. Know the biases. But to avoid falling into them is hard! Exercise: enumerate reasons why you might be wrong.
  • Be humble. Predictions are always uncertain. Predictions too far into the future even more so.