> "Beautifully shows how most predictions that extrapolate from past data are too simplistic"

IMO the actual mechanism is people ignore non-itineraries and assume that trends will continue upwards at the same exponential rate with no down force. Often at the start of a trend it hasn't triggered or hit against the hard limits, for example Moore's law hitting quantum mechanics.

It's so common I wish there was a pithy "logical fallacy" name for it, almost every extrapolation suffers from it, from superintelligence to housing prices and the stock market.