This Paradox Splits Smart People 50/50

I take the super computer. It's worth more than any of the boxes

So the choice is now between me taking the $1000 or $10001000 because the computer already made its choice?

You take both. The box are already filled. Nothing in your behaviour after you entered the room matters. As you learn of the problem the payoffs are already set. There’s literally nothing you or the computer can do.

Watch our latest video, out now on YouTube!

Take part in our Google form about Newcomb’s Paradox, did you change your mind after watching this video?

This is easy. The computer already made up it's mind before you did anything? Then your choice doesn't change any outcome. So pick both boxes and then maybe you got lucky and the computer predicted you would choose one

You take both boxes because at that point the computer has already set it up and your choice has no bearing on the contents.

The amount of people who simply can’t accept the premise of the question is insane. You are not going to outsmart the predictor by definition.

I’m with Casper 2 boxes is definitely the route to go. I don’t see how you predicting whether or not the computer predicted right, is casually linked to the million dollars. You are still asking the same fundamental question: If I gave you the choice between two boxes, one of which always contains $1,000 and one of which has essentially a random chance of having $1000000 in it, and you can either take both or only take the mystery box which choice should you make? Why would this change just because I ask someone or something to make a prediction about which one you might choose, if I know that the prediction is right virtually every time?

You take both boxes since their contents are predetermined.

Take the mystery box. If the supercomputer is nearly always correct, then if I take the mystery box, it has likely correctly predicted so. Which means the mystery box almost certainly contains £1m. This assumes the SC knows that I WILL find out about the problem, hence why it has been so successful in the past. Trying to get just $1000 more, in the hope the SC is wrong, makes no sense.

If the supercomputer can truly predict your behavior and it is right most of the time, then your choice should depend on a random external event which it can't predict. Toss a coin onto which you blindly wrote 1 and 2. Why blindly? Because it could have predicted that choice too.

Yeah, no shadow of a doubt in my mind. I'd pick one box.

The computer has near perfect prediction, so it knows that if you pick the mystery box it isn't by a coin flip. 1000 is sweet and it pays my bills for a few months, but a million is life changing. You take the mystery box. The computer has near perfect prediction. You win.

If the idea is that you want to “win” then there is no merit in taking the $1000 box. That’s not a win, it’s a given. If the supercomputer is nearly always correct then I will “nearly always” win $1million. The conundrum then is that since I only get one chance, “nearly always” means a very small chance of losing. I’d rather lose $1k than $1m.

the two boxers are objectively wrong and there is no paradox. At first I though obviously take them both your current decision doesn't impact what the computer already chose, which is 100% wrong, what you choose now 100% impacts what the super computer have chosen.

I don't like this because you just end up arguing price. Let's say the mystery box had $1000.01 Is the 1 cent worth the probability that it guessed you wrongly? Let's not even use dollars, use a currency you're unfamiliar with -- Open box has 1 million Uzbekistani Som. Mystery box has 1 million + 1 Som. At what point is the difference meaningful to the point it overcomes the uncertainty?

I live in a world where Murphy's Law prevails. I will always make the worst decision. Probably the day I walk into the experiment room, the lab is closed.

I made a video about a satisfying resolution to Newcomb’s paradox!

There's no paradox. Is the fight over whether or not you're smart? You seem to think you are, but you're definitely not.

If its always correct at predicting then whatever i decide is what it predicted, no?

My issue with two boxes has to do with the values given. If I get $1M then an extra $1K is insignificant, and $1K isn't life changing enough to risk the $1M.

This problem ignores people who take only the $1000 and those who take no box.

This should be a mandatory test for every 18 year old, on their brithday. If they are a one boxer, they get to drink, have sex, vote, buy a gun, apply for a job, play taxes, do whatever they want.

You take the mystery box. Leave the 1000 dollars. Computer has high accuracy. You have a very good chance at a fantastic payout.

That was a lot of probabilistic nonsense... Multiply the amounts in the boxes by 10.000 and see how many would walk out of the room turning down 10M. The only reason to be a two boxer is if you absolutely need the money right now for your immediate survival.

Both boxes for me.

The key point there it's weather we trust supercomputer has 99.9% chances of preselecting well. If so, taking only mystery box would mean we have 99.9% there will be million. Because if we took two, computer would have had to be not correct. How it is correct? Well its mystery!

The computer's been very good at predictions so far, so it's probably predicted me correctly, I take just the mystery box. If I were to take both, it would've predicted that, and I'd be down $999,000.

Given that it has guessed correctly thousands of times in a row, we can reasonably assume that will guess correctly now. The proper procedure used by military strategists when dealing with an adversary with an intelligence advantage or covert surveillance is to flip a coin. Reasoning backwards that at least a few of those thousands of prior people would have military backgrounds, and the supercomputer was still able to guess correctly despite a randomized counter strategy, we can infer a near omniscient superintelligence. Knowing that I cannot defeat this adversary in a battle of wits, my strategy becomes not to win, but to maximin my losing position. Only take the one box to maximize payout. The other so-called smart people are just monkeys quoting books without actually knowing how to think.

Consider a different problem. First you lock in your choice, and then the computer fills the mystery box in response. This problem is obvious: you pick $1k or $1m But what's the practical difference between perfect predictive power and simply going second?

Try it here :)

I'd flip a coin for it.

The problem is that people are random and plenty of times go against their own interests so the computer will never have a 100% succes rate of guessing right, don’t trust the computer.