Our universe might be fined-tuned for life because there are a huge number of universes each with different laws of physics and only under a tiny set of these laws can sentient life exist and we shouldn’t be surprised to live in one of these fine-tuned universes. Our universe might also be fine-tuned for the Fermi paradox, especially if advanced civilizations often create paperclip maximizers.
Perhaps if you look at the subset of all possible laws of physics under which sentient life can exist, in a tiny subset of these you will get a Fermi paradox because, say, some quirk in the laws of physics makes interstellar travel very hard or creates a trap that destroys all civilizations before they become spacefaring. Civilizations such as ours will constantly arise in these universes.
In contrast, imagine that in universes fine-tuned for life but not the Fermi paradox civilizations often create some kind of paperclip maximizer that spreads at the maximum possible speed making the development of further life impossible. As a result, these universes tend to contain very few observers such as us. Consequently, even though a far higher percentage of universes might be fined-tuned for just life than for both life and the Fermi paradox, most civilizations might exist in the latter.
Our universe might be fined-tuned for life because there are a huge number of universes each with different laws of physics and only under a tiny set of these laws can sentient life exist and we shouldn’t be surprised to live in one of these fine-tuned universes.
I was assuming Solonomoff Induction over the full space of computable universes, which is a more principled take on fine tuning and selection effects. We should expect to find ourselves in the universe described by the simplest theory (TOE) which explains our observations.
Our universe might also be fine-tuned for the Fermi paradox, especially if advanced civilizations often create paperclip maximizers.
Paperclip maximizers are a specific absurdity with probability near zero, and I find that discussing them sucks insight out of the discussion.
Perhaps if you look at the subset of all possible laws of physics under which sentient life can exist,
This full set is infinite, complex, and irrelevant—for all we know much of this space could have life radically different than our own. It is more productive to focus on the subset of the multiverse with physics like ours—compatible with our observations. In other words—we are hardly a random observer sampled from the full set of ‘sentient life’.
in a tiny subset of these you will get a Fermi paradox because, say, some quirk in the laws of physics makes interstellar travel very hard or creates a trap that destroys all civilizations before they become spacefaring.
Interstellar travel does look pretty hard, but not hard enough to prevent slow colonization. For this argument to apply to our section of the multiverse, it would need to involve new unknown physics. This is one possibility—but it seems low probability compared to other options as discussed in my post.
creates a trap that destroys all civilizations before they become spacefaring. Civilizations such as ours will constantly arise in these universes.
But then they are destroyed. I didn’t describe this in my post, but from an observational selection effect, it is crucial to consider the effect of deep simulations. The universes that produce lots of deep simulations with sentient observers will tend to swamp out all other possibilities, such as those where civilizations arise but do not produce deep simulations.
The models I discussed in my post all tend to produce enormous amounts of computation applied to deep simulation—which creates enormous numbers of observers such as ourselves and vastly outweighs universes where civilizations are destroyed.
In contrast, imagine that in universes fine-tuned for life but not the Fermi paradox civilizations often create some kind of paperclip maximizer that spreads at the maximum possible speed making the development of further life impossible.
This full set [of all possible laws of physics under which sentient life can exist] is infinite,
How can you be so sure of this?
Presumably there is some level of resolution at which changes to the fundamental constants no longer have appreciable effects. Maybe it’s the thousandth decimal place; maybe it’s the googolth decimal place; but it seems extremely unlikely to me that there isn’t such a level. Given this assumption, the set of possible laws is clearly infinite.
It doesn’t, because the reals are infinite in two ways: any given interval is infinite, but the number of intervals is also infinite.
Also this only applies to changing the constants but keeping the general structure the same; you can also create further laws by changing the structure of the laws itself. There are a lot of degrees of freedom.
Our universe might be fined-tuned for life because there are a huge number of universes each with different laws of physics and only under a tiny set of these laws can sentient life exist and we shouldn’t be surprised to live in one of these fine-tuned universes.
So what you’re saying here is that we can conclude we live in one of these fine-tuned universes simply by updating on the fact of our existence, correct?
Our universe might be fined-tuned for life because there are a huge number of universes each with different laws of physics and only under a tiny set of these laws can sentient life exist and we shouldn’t be surprised to live in one of these fine-tuned universes. Our universe might also be fine-tuned for the Fermi paradox, especially if advanced civilizations often create paperclip maximizers.
Perhaps if you look at the subset of all possible laws of physics under which sentient life can exist, in a tiny subset of these you will get a Fermi paradox because, say, some quirk in the laws of physics makes interstellar travel very hard or creates a trap that destroys all civilizations before they become spacefaring. Civilizations such as ours will constantly arise in these universes.
In contrast, imagine that in universes fine-tuned for life but not the Fermi paradox civilizations often create some kind of paperclip maximizer that spreads at the maximum possible speed making the development of further life impossible. As a result, these universes tend to contain very few observers such as us. Consequently, even though a far higher percentage of universes might be fined-tuned for just life than for both life and the Fermi paradox, most civilizations might exist in the latter.
I was assuming Solonomoff Induction over the full space of computable universes, which is a more principled take on fine tuning and selection effects. We should expect to find ourselves in the universe described by the simplest theory (TOE) which explains our observations.
Paperclip maximizers are a specific absurdity with probability near zero, and I find that discussing them sucks insight out of the discussion.
This full set is infinite, complex, and irrelevant—for all we know much of this space could have life radically different than our own. It is more productive to focus on the subset of the multiverse with physics like ours—compatible with our observations. In other words—we are hardly a random observer sampled from the full set of ‘sentient life’.
Interstellar travel does look pretty hard, but not hard enough to prevent slow colonization. For this argument to apply to our section of the multiverse, it would need to involve new unknown physics. This is one possibility—but it seems low probability compared to other options as discussed in my post.
But then they are destroyed. I didn’t describe this in my post, but from an observational selection effect, it is crucial to consider the effect of deep simulations. The universes that produce lots of deep simulations with sentient observers will tend to swamp out all other possibilities, such as those where civilizations arise but do not produce deep simulations.
The models I discussed in my post all tend to produce enormous amounts of computation applied to deep simulation—which creates enormous numbers of observers such as ourselves and vastly outweighs universes where civilizations are destroyed.
Right—we don’t live in that kind of universe.
Strongly disagree, and in general it’s dangerous to dismiss an argument by asserting that it’s stupid and that merely discussing it is bad.
How can you be so sure of this?
Presumably there is some level of resolution at which changes to the fundamental constants no longer have appreciable effects. Maybe it’s the thousandth decimal place; maybe it’s the googolth decimal place; but it seems extremely unlikely to me that there isn’t such a level. Given this assumption, the set of possible laws is clearly infinite.
Doesn’t this all imply that the set of meaningfully different laws is finite? Also, what if there is a smallest possible level of resolution?
You are aiming at meaningfully distinct. nbouscal is aiming at the functionally equivalent.
It doesn’t, because the reals are infinite in two ways: any given interval is infinite, but the number of intervals is also infinite.
Also this only applies to changing the constants but keeping the general structure the same; you can also create further laws by changing the structure of the laws itself. There are a lot of degrees of freedom.
So what you’re saying here is that we can conclude we live in one of these fine-tuned universes simply by updating on the fact of our existence, correct?
Our existence + the Fermi paradox, and it’s a high probability rather than a certainty.