Extrapolation: the most used and worst way to predict the future

I often hear "look what Gen Y or Z communicate and buy, they'll revolution the consumer world!!" or stats saying "In 2050, following this trend, the number of cars in China will be... the number of people living 100+ years will be..." or headlines saying that "if book/email/hotel/taxi industry continue to decline at such rhythm, it will die in 10 years!" or "Tesla plans to revolution car business, but at this pace, it will only represent 2 percent or car business!"

They all bet on 1 thing VERY often WRONG: linearity. But things are NOT linear most of the time.

Is your behavior linear from your childhood? No! We change, change again, again, and so on.
Are you sentiments evolving linearly? No
Do you buy the same brands as you did 2 years ago? No!
Do you eat the same food? No!
Do you do the same thing after work? No!
Do you do the same thing while commuting? No!
Is your revenue increasing (or decreasing) linearly? No!
Are companies revenues evolving linearly? No!

Linearity is an exception. Very few things are linear. Most things change in a very random way. As a result, most extrapolations are wrong. The reassure us but they are wrong.