Nick Bostrom uses this kind of reasoning to argue that there is a Great Filter lying ahead of us, and that we live inside a computer simulation.

The first is why we don't see any evidence for aliens. However, if the reason we don't see any aliens is because technological civilizations are extremely improbable, then there is no Great Filter lying in wait for us. Assuming civilization continues, then eventually we will make super realistic simulations, some of which will be ancestor simulations. The population of simulated people will far outnumber those of non-simulated kind. So therefor, odds greatly favor us living in a simulation.

So either we face extinction soon, or we're simulated. However, I don't really buy these arguments. They seem to be too simple, ignoring potential complications to building ancestor simulations or colonizing the galaxy in a few million years.

This thought experiment is mentioned in Stephen R.L. Clark’s book God, Religion and Reality, a book I actually brought up in a recent thread.

Whoever/whenever you are it’s always most likely that yours is the final generation. The rationale being that if the final generation is the largest one, and if a random person is always most likely to be among the largest generation/group, then it follows that you (a random person) are most likely among the final generation.

Imagine the following two population predictions:
In the first, humanity prospers. We venture into space and our population keeps growing steadily. All the humans born until now will be a fraction of all the humans that are to come in the existence of humanity. — Mind Dough

This shows how unlikely it is that we’ll ever expand out into the galaxy, since it would mean we’re all part of a tiny fraction of all humans, rather than the other huge group. Instead it stands to reason that we’re at the top of graph 2’s curve.

Nick Bostrom uses this kind of reasoning to argue that there is a Great Filter lying ahead of us, and that we live inside a computer simulation. — Marchesk

Funny that you mention this. I'm writing an article that also references the simulation hypothesis, that's actually the reason I came to ask this question. It is exactly this question, only a substantial part of the first graph being simulated.

Whoever/whenever you are it’s always most likely that yours is the final generation. The rationale being that if the final generation is the largest one, and if a random person is always most likely to be among the largest generation/group, then it follows that you (a random person) are most likely among the final generation. — AJJ

From a random person's perspective, this will always be true (if indeed population keeps rising). However, I am wondering if this (and therefor also the simulation hypothesis) is a valid argument to make: Personally I think this is a special case of survivorship bias, in which we simply cannot see (or know) the people that havn't been born yet. A 4 dimensional survivorship bias if you will.

What are your thoughts on that?

I wouldn’t say that bias is in play here. The argument simply acknowledges that it is far more likely we’re at the top of Graph 2’s curve than at the bottom of Graph 1’s. Either situation could be true - indeed because we can’t see or know who is to come - but all that can be said without knowing is that the former situation is more likely.

The first is why we don't see any evidence for aliens. — Marchesk

We also don't have any independent lineage of life existing on earth that we aren't descended from. We also have failed to recreate it even with all the king's horses and men going at it.
It seems likely life is something extremely rare and I don't see why that isn't recognized or taken into account, like ever.

Just experimenting with this chain of thought:


Using that same logic, would it not also make sense to say that we are "far more likely" currently the largest conscious civilization that has ever and ever will exist in existence?

I mean, if over the time of the universe there would be a bigger civilization in existence, it would be more likely to be born as one of them.

Yes, that seems right to me.

Scary thought...But even so, we can still hope to be the one in a billion lottery ticket.

After all, this would also raise the chance significantly on being either a Boltzman brain or the rest of humanity being philosophical zombies (I mean, what is the chance of being me in 7 billion?!). A reality I would rather avoid (as if it matters:P).

Yeah, I guess we can.

I’m not sure it raises the chances of being a Boltzmann brain, since for that to be likely the universe would need to have had a much longer past that it’s commonly (to my knowledge) said to have had.

And it seems to me it’s not actually possible for you (anyone) to have been anyone else, since obviously you’d not be you then. I don’t see why you being you makes it likely that everyone else is a zombie, rather than everyone else just being the particular conscious person they are.

My feeling says this is not an argument for having a larger chance that graph two is true, but I am wondering about the argumentation on that. I suspect something in the line of survivorship bias, but I am still curious whether this is an existing thought experiment (I think I saw this somewhere before) and what the arguments/criticisms are. — Mind Dough

It's not a named bias as far as I am aware. It amounts to magical thinking though. It's very similar in that regard to the "fine tuning" arguments. Math (in this case statistics) is used to mask the fact that the argument has no substance. You'll notice that this "doomsday argument" has only a single piece of information as input: that right now, humans exist. It puports to derive from that information a piece of entirely new information - the likelihood of humanities imminent demise.

But how does it get from one bit of information to the other? It does not account for current trends in global politics, the economy or the environment. It simply transforms one piece of information into another. That is, it creates information ex-nihilo, which is impossible.

The interesting question is: How does an application of mathematics that, on it's face, seems valid, lead to an absurd result?

This shows how unlikely it is that we’ll ever expand out into the galaxy, since it would mean we’re all part of a tiny fraction of all humans, rather than the other huge group. Instead it stands to reason that we’re at the top of graph 2’s curve. — AJJ

But according to that logic, it stands to reason that everyone who has ever lived was at the top of graph 2s curve, but they weren't. How is that possible?

This shows how unlikely it is that we’ll ever expand out into the galaxy, since it would mean we’re all part of a tiny fraction of all humans, rather than the other huge group. Instead it stands to reason that we’re at the top of graph 2’s curve.
— AJJ

But according to that logic, it stands to reason that everyone who has ever lived was at the top of graph 2s curve, but they weren't. How is that possible? — Echarmion

If we know that someone is not at the top of Graph 2’s curve then obviously they’re not at the top of Graph 2’s curve. But we don’t know where we are, and in that ignorance the probability that we’re at the top of Graph 2’s curve comes into play.

If we know that someone is not at the top of Graph 2’s curve then obviously they’re not at the top of Graph 2’s curve. But we don’t know where we are, and in that ignorance the probability that we’re at the top of Grapxh 2’s curve comes into play. — AJJ

But we're turning that ignorance into a probability without further information. That is impossible. If we're ignorant about what graph we are on and where we are on that graph, we can't magically turn said ignorance into new information using math

But we're turning that ignorance into a probability without further information. That is impossible. If we're ignorant about what graph we are on and where we are on that graph, we can't magically turn said ignorance into new information using math — Echarmion

No we aren’t, and yes we can.

We don’t know where we are. Mathematical reasoning tells us we’re most likely at the top Graph 2’s curve. We might not be, but what we can say is it’s more likely.

No we aren’t, and yes we can. — AJJ

So, which is it? Do we have further information that allows us to make a determination or are we creating information ex nihilo using math?

Just saying "it's mathematical reasoning" is the same as saying "it's magic".

I suppose a simpler example would be to say that I'm given a number at random from a range of 1-n. What is the probability that the number I was given is n?

Assume the number I was given is 10. If n is 10 then there would have been a 1/10 chance that I would be given a 10, whereas if n is 100 then there would have been a 1/100 chance that I would be given a 10. Given that a 1/10 chance is greater than a 1/100 chance, it's more probable that n is 10 than n is 100 (and the same is true for any n > 10). This seems to be the sort of reasoning used in your example. It doesn't really work out that way when we play such a game though:


If for each game the range is 1-100 then there's a 5.2% chance that the number I'm given at random is the max. If for each game the range is 1-1000 then there's a 0.7% chance that the number I'm given at random is the max.

All this shows is that the shorter the lifetime of the human race, the more likely that humanity is closer to extinction. But that's a truism.

You’ve blundered here and now you’re trying too hard to disagree with me.

The information we use is the assumption that the final generation will be the largest one. The mathematical reasoning we use is that it’s most likely for a random person to be part of the largest group, assuming we don’t already know where they are.

Isn't the probability of you, as the particular individual you are, living now, actually 1?

The notion that there's anything random about you, as the particular individual you are, living at any random time seems flawed, no? Graphs like that would only make sense if you "already existed" somehow, without living, and then the exact year that you wind up living is chosen in a manner akin to selecting lottery numbers. But that's not how it works. So we can't pretend that it's how it works if we want to say anything not purely fantastical about it.

The information we use is the assumption that the final generation will be the largest one. The mathematical reasoning we use is that it’s most likely for a random person to be part of the largest group, assuming we don’t already know where they are. — AJJ

This is misleading. Assume 5 generations with populations like this:

1: 100
2: 200
3: 300
4: 400
5: 500

It's more likely that I'm in generation 5 than in generation 4, and more likely that I'm in generation 5 than generation 3, and so on – but it's more likely that I'm not in generation 5 than in generation 5. The actually probabilities are:

1: 6.7%
2: 13.3%
3: 20%
4: 26.7%
5: 33.3%

If you were blindfolded, so to speak, and told you were part of one of those generations, which would you predict you were a part of?

Your answer wouldn’t be “I’m most likely part the first four”, but rather “I’m most likely part of the last four”. Of those four you’re most likely part of the last three. Of those the last two, and of those the last one.

Is that not right?

You’ve blundered here and now you’re trying too hard to disagree with me. — AJJ

Or perhaps this isn't about you and I'm just interested in the topic.

The information we use is the assumption that the final generation will be the largest one. — AJJ

But as you state, this isn't information, but an assumption. But even if we grant the assumption as essentially correct, it does still not contain any information about humanities demise, so the question of where that information comes from remains.

The mathematical reasoning we use is that it’s most likely for a random person to be part of the largest group, assuming we don’t already know where they are. — AJJ

As an analogy, take the hotel room example. You're in a hotel with 100 rooms, but you don't know which room you're in. If someone asks you whether you are in the first ten rooms, your answer should be no. But if someone asks you whether you are in room 2 or in room 50 your answer should not be fifty just because you are more likely not to be in the first 10 rooms. Because for that specific question, the probability of either is 1/100. This probability doesn't change if you arbitrarily divide the hotel rooms into groups.

So in order to make a meaningful statement, the groups need to be given in advance. Otherwise, just knowing you're "in the largest group" tells you nothing because groups don't actually exist as physical objects.

If you were blindfolded, so to speak, and told you were part of one of those generations, which would you predict you were a part of?

Your answer wouldn’t be “I’m most likely part the first four”, but rather “I’m most likely part of the last four”. Of those four you’re most likely part of the last three. Of those the last two, and of those the last one.

Is that not right? — AJJ

Say you roll a dice. If it's 2-6 then roll again. Repeat until you roll a 1.

According to your reasoning, your first roll is more likely to be 2-6 than 1 so you're more likely to roll a second time; your second roll is more likely to be 2-6 than 1 so you're more likely to roll a third time; and so on. You're then saying that you're more likely to roll 20 times in a row than to roll once. But of course that's obviously wrong. The chance of rolling just once is 0.167, whereas the chance of rolling 20 times is 0.026.

You're more likely to roll < 20 times than to roll 20 times. You're more likely to be part of generation < 5 than to be part of generation 5.

I'm wondering what kind of sampling mechanism could actually exist to generate individuals from previous time points. All probability is doing here is summarising the areas under curves with reference to the total area of each curve. The concept is doing no work, since there's no probability model - the sampling mechanism is completely artificial.

But as you state, this isn't information, but an assumption. But even if we grant the assumption as essentially correct, it does still not contain any information about humanities demise, so the question of where that information comes from remains. — Echarmion

I would say a reasonable assumption is information. It’s certainly something we can reason from.

No it doesn’t contain information about humanity’s demise. It’s the mathematical reasoning that shows we’re most likely close to that demise.

As an analogy, take the hotel room example. You're in a hotel with 100 rooms, but you don't know which room you're in. If someone asks you whether you are in the first ten rooms, your answer should be no. But if someone asks you whether you are in room 2 or in room 50 your answer should not be fifty just because you are more likely not to be in the first 10 rooms. Because for that specific question, the probability of either is 1/100. This probability doesn't change if you arbitrarily divide the hotel rooms into groups. — Echarmion

The probability does change if you divide the rooms into groups. If I say you’re either in rooms 1 or 2 or in any of the rest, then it’s most likely you’re in the second “any of the rest” group.

Your dice example isn’t analogous to the group situation described. We’re not rolling a dice to find out which group we’re in, but using the amount of people in each group to work out the likelihoods.

A dice analogy would go like this: Imagine you’re a spot on a dice. Which side are you most likely on? You’re more likely on the 2-6 sides than the 1 side. Of those you’re more likely on the 3-6 sides, and so on.

Your dice example isn’t analogous to the group situation described. We’re not rolling a dice to find out which group we’re in, but using the amount of people in each group to work out the likelihoods. — AJJ

I did that here:

1: 100 people (6.7%)
2: 200 people (13.3%)
3: 300 people (20%)
4: 400 people (26.7%)
5: 500 people (33.3%)

A dice analogy would go like this: Imagine you’re a spot on a dice. Which side are you most likely on? You’re more likely on the 2-6 sides than the 1 side. Of those you’re more likely on the 3-6 sides, and so on. — AJJ

OK, and then of those you're more likely on the 4-6 side, and then of those more likely on the 5-6 side, and then of those more likely on the 6 side? So you're more likely to be on the 6 side than on one of the 1-5 sides? No. There are 21 dots on a dice, and only 6 are on the 6 side. There's a 29% chance that a dot taken at random is on the 6 side and a 71% chance that a dot taken at random is on one of the 1-5 sides.

I did that here:

1: 100 people (6.7%)
2: 200 people (13.3%)
3: 300 people (20%)
4: 400 people (26.7%)
5: 500 people (33.3%) — Michael

Doesn’t this prove my point? I think the mistake here is making groups 1-4 one group. You can’t be in multiple generations; since you can only be in one it’s most likely you’re in 5.

Doesn’t this prove my point? I think the mistake here is making groups 1-4 one group. You can’t be in multiple generations; since you can only be in one it’s most likely you’re in 5. — AJJ

I'm not saying that I'm in multiple generations. I'm saying that I'm more likely to be in generation 1 or generation 2 or generation 3 or generation 4 than to be in generation 5.

I have a 33.3% chance of being in generation 5, and so therefore a 66.7% chance of not being in generation 5.

Let's phrase this another way. There's 1 red ball, 2 orange balls, 3 yellow balls, 4 green balls, and 5 blue balls. I give you a ball at random. It's more likely to be not-blue (2/3) than blue (1/3).

What you’re actually saying is you have more chance of being within generations 1-4. What you actually have to do is pick one; the one most likely for you to be in.