Taleb makes "long-term predictions on the basis of the prevalence of forecasting errors"

Nassim Taleb talks in the Economist about the world in 2036 - "what will break and what won't".

A system that is over-reliant on prediction (through leverage, like the banking system before the recent crisis), hence fragile to unforeseen “black swan” events, will eventually break into pieces. Although fragile bridges can take a long time to collapse, 25 years in the 21st century should be sufficient to make hidden risks salient: connectivity and operational leverage are making cultural and economic events cascade faster and deeper. Anything fragile today will be broken by then.

I picked up a copy of the second edition of Taleb's The Black Swan which appears to be massively revised and expanded from the first edition (including a new section "On Robustness and Fragility").