Newcomb's Problem

We’re going to do a thought experiment called Newcomb’s Problem. Here’s how it works:

A generous predictor wants to reward people who trust them. They set up a game:

The predictor has already looked at you and decided what kind of person you are:

The predictor is never wrong.

What do you choose?

Stand up if you would take only Box B.

Stay seated if you would take both boxes.

Most people take only Box B and get the million. But some people argue: “The money is already in there or it isn’t — I can’t change that now, so I should grab both boxes for an extra $1,000.”

What do you think? Can you outsmart a perfect predictor? Or does trying to outsmart them mean they already predicted that?

Contributor(s)