Ajay Agrawal, Joshua Gans

Prediction Machines

1. IDEA 1

   AI makes prediction cheap

   Agrawal, Gans and Goldfarb, economists, argue that the simplest way to understand modern AI is as a technology that makes prediction cheap. When something becomes cheaper, people use more of it and find new uses.

2. IDEA 2

   Cheaper prediction changes everything

   The authors explain that as prediction becomes cheaper, it will be used for tasks that were not seen as prediction problems before, such as driving or translation. Falling cost leads to using it more widely.

3. IDEA 3

   Judgement becomes more valuable

   The authors argue that as machines handle prediction, human judgement, deciding what outcomes matter, becomes more valuable. The two work together. A machine can say what is likely, but people must still decide what is worth wanting.

All 8 ideas from Prediction Machines →