Yearly Archives: 2017

Newcomb’s Problem and Relationship Rules

In my defense of egoism, I discussed Newcomb’s Problem, summarized in the Less Wrong wiki as:

a superintelligence called Omega shows you two boxes, A and B, and offers you the choice of taking only box A, or both boxes A and B. Omega has put $1,000 in box B. If Omega thinks you will take box A only, he has put $1,000,000 in it. Otherwise he has left it empty. Omega has played this game many times, and has never been wrong in his predictions about whether someone will take both boxes or not.

Classic decision theory states that you should “two-box,” i.e. take both boxes, because nothing you do at that point will change what is in the boxes. The way to “win” the game is not to take one box. The only way to win is, at the time Omega makes their decision, to convince Omega that you are going to take one box.

Omega, though, is superintelligent, as has never been wrong about a prediction. Trying to outsmart Omega is risky at best, doomed at worst. The most effective strategy is to commit, before Omega fills the boxes, to only taking one box. And the commitment must be real. You have to change something in your brain to cause you to one-box despite the fact it is ostensibly in your best interest to take both.

This dovetails nicely with my post about why we make rules. In that post, I discussed how relationship rules function as psychological barriers. When we make a commitment, we create an obligation in our minds that encourages us to stick to the commitment, even if it seems like a bad idea at the time. This is exactly the way to win Newcomb’s Problem.

So when considering relationship rules, the operative question is whether the situation that the rule addresses is Newcomblike. If it is, making the rule is a good idea. But if it isn’t, making the rule unnecessarily limits your future behavior for no good reason, which can lead to unhappiness, resentment, toxic power differentials, and shaming. Questions to ask are:

  • Do I want to be the type of person who would follow this rule?
  • Do I trust my decision-making enough to make the correct choice without this rule?
  • Even though I know now what the right choice is, do I think that I will make the wrong choice in the moment?
  • Do I have enough information to make this decision now? Could I learn more information that would change my decision?
  • Is there a compelling reason why this decision needs to be made right now?

The dangerous part about gaming Newcomb’s problem is that to work, commitments must be irrevocable. In practice, that’s not really an option, but making a strong commitment can still be disastrous if minds are changed later. Trust me. I’m a divorce lawyer. So it’s important to really examine whether the situation is Newcomblike. If it is, make the commitment; but if not, it’s time to look for another solution.