site banner

Friday Fun Thread for October 14, 2022

Be advised; this thread is not for serious in depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.

9
Jump in the discussion.

No email address required.

I've been waiting for the Fun Thread to post this, since it is manifestly not CW. I had a conversation about physics, namely about the expansion of space and the feasibility of intergalactic travel, in the CW thread a few days back. I came back to it, and some things I wrote have been bugging me enough that I want to issue a clarification.

Here's the relevant thread.

https://www.themotte.org/post/120/culture-war-roundup-for-the-week/15941?context=8#context

Basically I said that the expansion of the universe renders galaxies that are beyond a certain distance unreachable absent some form of FTL travel, since the universe's expansion isn't set at any constant rate of speed - any two points in space recede at a rate proportional to their distance from each other, and that distance is increasing, therefore the rate of recession of the two points is increasing. This means there will be some galaxies far enough away from us that they recede faster than the speed of light.

I got this question in response: "Is it not the case that, once we start moving towards those distant objects (in say a colony ship), the expansion behind us compensates for a growing portion of that total expansion? It's my understanding that there IS an inflection point as you describe, but we haven't reached it yet."

And this was my reply: "The case for an inflection point is pretty strong. It’s my understanding that for objects that have already crossed the boundary of the event horizon, no reduction of the distance between us and that object will occur. Think about it this way: There are objects far enough away from you that they are moving away at a rate that exceeds the speed of light, meaning without FTL travel they will be receding from you faster than you can travel to them. The space between you and any object beyond that horizon will only increase and the further they go, the faster they recede. If you try to reach it in a relativistic colony ship, all that happens is that you’ll be stranded from your original galaxy group and will never reach the new one as your galaxy of origin passes out of your event horizon. Sure, you are closer to the object and further away from your point of origin than you would've counterfactually been, but that does not equate to closing the distance."

The issue here is that there are additional real-life complexities absent in the model I was outlining which I neglected to address when I wrote this (note: do not write when you're tired unless you want to omit things you should've mentioned). Now, I don't have too much of a problem with what I wrote here - I stand behind my point that ceteris paribus, anything in a superluminally receding region of space in which expansion is driving the objects away from you faster than light would simply be completely unreachable, and you wouldn't be able to magically "close the distance" and catch up. The area where things are receding from us slower than light is called our Hubble volume, and is a sphere approx 14 billion light years in radius (everything outside of it is moving away faster than light). However, I want to clear something up: We can sometimes receive light from galaxies outside our Hubble sphere at the time the light was emitted, meaning light-speed information in a region of space which is receding from us faster than light can in fact travel to us. So how can this be possible?

The reason is because the Hubble constant (the unit that describes how fast the universe is expanding at different distances from a particular point in space) is decreasing, causing our Hubble volume to expand. This means that photons emitted by galaxies in a superluminal region can eventually enter inside of our Hubble volume and be able to reach us, and this is not because light is magically "catching up" - it is receding, but its recession doesn't outpace the growth of our Hubble volume. Similarly, the recession speed between light we emit and an object farther than the Hubble distance is initially positive, but can become negative as the Hubble distance increases. Here's a Veritasium video with a visual representation of how this can happen.

Of course, there's a limit to this too - there's a point beyond which light emitted from objects are receding from us so fast that they will never fall inside of our Hubble volume, and this boundary is delineated by the "cosmological event horizon" which is currently about 16 billion light years in radius.

There is another bigger issue that I want to correct here, and that's my statement that we might not make it outside our Local Group (and that this has something to do with expansion). I knew previously that our reachable universe was much larger than that, but of course practically speaking that's very optimistic and assumes we leave now and at the speed of light. What I was thinking was that 1: if relativistic speeds are hard to accomplish, our closest galaxies might be expanding away from us faster than we can travel, and 2: even if relativistic speeds are possible expansion might set a limit on how much we can progress before our closest galaxies are eventually isolated from us. Again, I wrote this bit while not thinking too deeply about it, and have since reasoned myself out of this position.

With regards to the first point, one of our closest galaxy clusters (the M81 Group) is currently 11.4 million light years away from us, and the Hubble constant (according to some estimates) is 68 km/s/Mpc. 11.4 million light years is equivalent to 3.5 megaparsecs, and that means the M81 Group is expanding away from us at 238 km/s. That is a very small fraction of the speed of light, and probably isn't impossible for us to exceed given that the Parker Solar Probe has already been able to reach speeds of 163 km/s. The rate of recession only becomes prohibitive for objects much further from us.

Note, this doesn't mean that I think travel to another galaxy cluster is actually that feasible, it just means expansion wouldn't pose too large an obstacle for us. The reason why it would be difficult in a practical sense is not just because of the difficulty of finding a reasonable propulsion method, it's also because of the time involved to travel the entire distance. Even in a colony ship travelling at 99% of the speed of light, the trip to the M81 Group would take an unrealistically long time, even accounting for relativistic effects from the perspective of the traveller. As viewed from the spaceship, an 11.4 million light-year trip would be 1,624,412 years long, which is far longer than the entirety of human history (here's a neat website that helps you calculate these things, for the lazy). This is assuming that we travel constantly at 99% the speed of light the entire way, it's not taking into account acceleration and deceleration to the destination, so this is a minimum estimate. There's simply no way to design for missions of that length, nor will there be for a very long time, if indeed ever.

Accelerating to 99.99999999...% of the speed of light would create enough dilation to get us there in an acceptably short time, but there's something else stopping us from doing that (even assuming we manage to find a method of propulsion which will allow us to go that fast, which is a big assumption). And that's space dust. At 99% of the speed of light, hitting a 4 milligram grain of dust in space gets you 2,188,941 megajoules of kinetic energy (here's the calculator I used, I use them because I want to mess with the variables without having to do the calculation again and again). A ton of TNT contains 4,184 megajoules of energy, so that 4 milligram grain of dust at 0.99c is going to be equivalent to 523 tons of TNT exploding. Even hitting a dust grain of 0.1 mg at that speed is going to yield you 54,724 megajoules of kinetic energy, equivalent to 13 tons of TNT.

Get closer and closer to the speed of light so that the travel duration becomes more reasonable, and eventually these grains of dust are going to start looking more and more like Hiroshima.

So it's not that I think that travelling to another galaxy cluster is feasible, it's rather that at this point, expansion is a red herring. If we can't travel faster than 238 km/s in the first place, the travel time would be far too long for us to even think about starting a mission even assuming that the M81 Group isn't receding from us. Even non-lethal relativistic speeds won't take us there in any reasonable time. Travelling to other galaxy clusters is probably FTL or nothing (we're talking Alcubierre drives and wormholes here and not actual travel faster than light, because of the constraints relativity poses, and there are still many problems with those methods which means there is plenty of reason to suspect FTL is not possible).

As to the second point about the time limit expansion imposes, it turns out the timeframe we have before our closest neighbours have receded into superluminally receding regions of space which we will never be able to reach (without FTL, that is) is hundreds of billions of years, so this timeframe probably doesn't pose too much of an obstacle. Whether exiting our Local Group is actually feasible or not in the first place is almost certainly the main factor. And the reasons why it might not be feasible are huge.

I love the cosmological analysis but would be curious to hear your thoughts about the practicalities of closing those distances from the following perspective:

  • The journey begins 100-1000 years from now

  • Civilization consists of strongly superintelligent artificial general intelligences that have harvested all of the mass energy of our planet / solar system / galaxy / local group to construct astronomical data centers to wring almost as much computer out of our available mass-energy as permitted by the laws of physics (dyson swarms, or a hyper-advanced computer suspended across the event horizon of a supermassive black hole, or something else of that nature)

  • The expedition to a distant galaxy cluster consists of a Von Neumann probe, whose design objectives are (1) get there, (2) repurpose the mass-energy of that region of space to build another giant computer, and (3) (optional) establish communication with the parent civilization to facilitate the transmission of digital intelligences or the exchange of information.

If you assume that scenario, it seems like many of your objections dissolve. The difficulty of accelerating a probe to greater than 238km/s seems comparatively trivial (the probe can be as small as needed to meet the objectives above, and the probe design, the mechanism of accelerating and then decelerating, and redundancies or solutions to the problem of space dust can all be assisted by galaxy-brain technology), and there's no population of primates that needs to reproduce and sustain a culture for over a million years.

I do find it perplexing how many accounts of humanity's far future seem to assume that we'll be tackling problems that span millennia with the technology and life-forms that exist as of (give or take) 2022. In my head, I call it the Primates In Space Fallacy.

On a cursory inspection I'd say your argument has some merit. It's plausible that artificial intelligence might be able to send probe after probe out which accelerates then coasts in the intergalactic void for millions of years, and if even one of them actually reaches the destination intact it could make copies of itself elsewhere in the manner you describe.

Success rates would probably still be low, though. Space dust is not the only limit to the speed you can travel (though it is a significant one), there's also other hazards to contend with in interstellar and intergalactic space when you're travelling at highly relativistic speeds. And there's also the fact that finding a method of propulsion that will create strong enough relativistic effects to shorten the travel time significantly is hard - as noted earlier, even if the probe is able to achieve 99% of the speed of light (a very tall order) it will still take it over a million years to get to the M81 Group. And the longer the travel time, the more problems can occur - even a minuscule chance of equipment damage or malfunction becomes an unreasonably big issue when the timeframe is millions of years. However, with the human element removed, tackling the issue through sheer volume alone is much more feasible and much less objectionable when your repeated suicide missions involve Von Neumann probes instead of generation ships carrying flesh and blood humans.

With regards to establishing communication with the new outpost, however, there are problems which I can't help but view as being insurmountable. Once the probe reaches the M81 Group it would take 11.4 million light years for the original AI to receive a response, and that time lag means that the utility of communication with the other AI would be negligible at best since all the information it'd receive would be coming from an outpost which could be long dead. The two AIs would effectively be cut off from each other, and expansion over a very long period would drive the two clusters further and further away from each other which would mean that they'd eventually end up being completely isolated and unable to communicate.

Perhaps I should've made it more clear that I was merely talking about the limits of what any manned mission could reach - while artificial intelligence making it outside of our Local Group is definitely something interesting it's also not us as in humanity making it there. The kinds of things that could make the journey are not biological in nature. I guess the argument could be made that a new humanity could be constructed in situ by the AI itself after it's been seeded in the new galaxy cluster and technically represents an "expansion" of human life (whether this is a likely outcome at all given the abundant hazards of AI is another question entirely), but the original humans in our Local Group are still basically stuck there and this is probably not what people imagine when they think of humans venturing out into the universe.

Success rates would probably still be low, though. Space dust is not the only limit to the speed you can travel (though it is a significant one), there's also other hazards to contend with in interstellar and intergalactic space when you're travelling at highly relativistic speeds.

238km/s is not highly relativistic. Also it would be silly to travel at 99.9% the speed of light when you could travel at 90% for a tiny fraction of the energy and risk and get there less than 10% later. The only reason to do it would be so that less time passes for your travelers -- but if it's a self-repairing box of electronics and robotics, engineered by a galaxy-brain superintelligence, it can probably while away the millions of years without issue. There are no primates on board who are aging, nor even who are consuming energy to maintain.

You also don't have to make the trip all in one shot. Your civilization is probably launching probes in all directions to absorb the galaxy and then neighboring galaxies and then outward from there -- basically an expanding sphere of civilization led by Von Neumann probes. And you can send a lot of probes -- depending on how small they can be and how efficient their propulsion is, even a very high loss rate can just be overcome with quantity. Another advantage of not having precious primates on board!

Once the probe reaches the M81 Group it would take 11.4 million light years for the original AI to receive a response, and that time lag means that the utility of communication with the other AI would be negligible at best since all the information it'd receive would be coming from an outpost which could be long dead.

Why would the outpost be dead? Galaxy-brained superintelligences don't seem like likely to be mercurial creatures that might just die off one day from a plague or civil war or something. Once they're established, I assume they're gonna be here till the end of time.

Perhaps I should've made it more clear that I was merely talking about the limits of what any manned mission could reach - while artificial intelligence making it outside of our Local Group is definitely something interesting it's also not us as in humanity making it there.

Right, that's why I framed my challenge as a hypothetical -- what would it look like if you made some different assumptions. Nothing seems to stand in the way of galaxy-spanning superintelligences, so if anything it would seem to demand more artifice not to have that future -- some reason why we don't have massive superintelligences that doesn't also involve the destruction of human civilization. I think humanity is going to let go of our sentimental attachment to meat-based life basically as soon as we have a digital alternative -- but even if we don't, as you say, presumably your Von Neumann probes could build "human manufactories" on the other side of their voyage, even from digitally reconstituted genomes from our local group, in which case I don't see why they'd be any less "ours" than whatever distant descendants clambered off of a successful million year generation ship after it arrived on the other side of the cosmic ocean.

Anyway... if I'm right about the trajectory of our species, how much of our light cone do you think we could in principle colonize? That's the interesting question IMO. We've already lost a bunch of the cosmos over the cosmological horizon, and we'll lose more over the next 100-1000 years while we prepare our ballochory, and we need to draw the bounds somewhat tighter than the cosmological horizon given that we won't be traveling at precisely c, but it seems to me that our descendant civilization still stands to inherit a staggering cosmic endowment even with those assumptions.

238km/s is not highly relativistic. Also it would be silly to travel at 99.9% the speed of light when you could travel at 90% for a tiny fraction of the energy and risk and get there less than 10% later. The only reason to do it would be so that less time passes for your travelers -- but if it's a self-repairing box of electronics and robotics, engineered by a galaxy-brain superintelligence, it can probably while away the millions of years without issue. There are no primates on board who are aging, nor even who are consuming energy to maintain.

I'm aware 238 km/s isn't highly relativistic, travelling slightly above that speed just means the probe will spend a painfully large amount of time cruising. And even non-relativistic travel poses issues. For example your probe is going to encounter the harsh radiation environment in space, even if it's not travelling at relativistic speeds (if it is it's much worse). Shielding could be possible if one was willing to add to probe mass, but if it fails to block all of the radiation it's going to be exposed to that for the entirety of the journey. This is fine when your mission duration is short. It's less fine when your mission duration is millions of years and your probe contains lots of delicate electronics that need to work properly.

Self-repair is hypothetically possible, but that requires usable energy and matter, and interstellar and intergalactic space is famously devoid of both of these things. And the longer your mission is, the greater the chance of a critical failure at some point. Even if that chance is small, if you're going to take millions upon millions of years to get there most of your probes might not reach. Travelling slow comes with its own costs and impracticalities.

And yeah, I know every single one of these problems can be solved by invoking [hypothetical future technology], and I'm sure the future will unceremoniously spit in the face of any prediction I make, but I'm not too convinced by any explanation that relies too heavily on handwavium.

And you can send a lot of probes -- depending on how small they can be and how efficient their propulsion is, even a very high loss rate can just be overcome with quantity. Another advantage of not having precious primates on board!

Yes, I agree, even with a high loss rate you could spread your probes as long as there's a nonzero probability of survival. As I said, the idea you posited is not out of the question. Of course, then the Fermi paradox rears its head, since not only are we seeing no sign of alien life from our own galaxy, but also from other galaxies and other galaxy clusters which should hypothetically be able to reach us.

Why would the outpost be dead? Galaxy-brained superintelligences don't seem like likely to be mercurial creatures that might just die off one day from a plague or civil war or something. Once they're established, I assume they're gonna be here till the end of time.

I'm not saying it would be dead, I'm more saying that communicating would be full of latency problems - any information you'd get from it wouldn't be timely at all and would be mostly of little value since it'd be impossible to act on. The point was that for all you know the outpost could be dead and you'd only know 11.4 million years later.

I think humanity is going to let go of our sentimental attachment to meat-based life basically as soon as we have a digital alternative

Personally, I wouldn't do it. Even assuming that you can replicate the phenomenon of consciousness in a non-biological substrate (something that could be the case but that's basically impossible to prove), there's the issue of continuity when you're uploading your brain. Sure, there's another version of myself now, but this is not me and I will not experience the change. I will live and die as meat-based life, and there will not be any "transfer" of consciousness. There will not be any change in my own state except now I possess the knowledge that there is an immortal version of me running around out there.

So this is not because I have any attachment to meat-based life - the benefits of a digital substrate would be very tantalising if I could genuinely transfer myself into it. Rather, I think my experience of being me is so intrinsically linked with my physical body that they basically can't be disconnected from each other. The incentive to digitise my brain kind of starts looking very weak then.

but even if we don't, as you say, presumably your Von Neumann probes could build "human manufactories" on the other side of their voyage, even from digitally reconstituted genomes from our local group, in which case I don't see why they'd be any less "ours" than whatever distant descendants clambered off of a successful million year generation ship after it arrived on the other side of the cosmic ocean.

The issue for me is that you don't actually get to colonise anywhere, nobody leaves, you just make another galaxy cluster full of humans. Maybe this is just an irreconcilable values difference, but I think this solution completely voids the point of the exercise. I don't intrinsically care about creating as many humans as possible and distributing them throughout the galaxy. I care infinitely more about where these humans come from.

Anyway... if I'm right about the trajectory of our species, how much of our light cone do you think we could in principle colonize? That's the interesting question IMO.

Let's assume we can go at, say, 50% light speed (149896.229 km/s). The expansion of the universe is 68 km/s/Mpc, and the value of Mpc that gives us a recession speed of 0.5c (the relevant formula here is 68 x Mpc = 149896.229) is 2204.36 megaparsecs, which translates to roughly 7 billion light years. Everything outside that distance is receding from us faster than that.

It's basically Hubble's law. You can take any speed of travel, divide it by the expansion rate, and find the distance beyond which everything is receding from you faster than your travel speed. There's probably additional complexity created by the aforementioned fact that the Hubble "constant" is not actually constant and is decreasing, but I can't be arsed to factor that in right now.

Personally, I wouldn't do it. Even assuming that you can replicate the phenomenon of consciousness in a non-biological substrate (something that could be the case but that's basically impossible to prove), there's the issue of continuity when you're uploading your brain. Sure, there's another version of myself now, but this is not me and I will not experience the change. I will live and die as meat-based life, and there will not be any "transfer" of consciousness. There will not be any change in my own state except now I possess the knowledge that there is an immortal version of me running around out there.

My typical rejoinder is basically, how do you know this isn't happening every time you go to sleep -- that "you" die and a new "you" comes into existence when your body and brain awake the next morning, a brand new consciousness tricked by your brain's memories into inferring false continuity of self? Or even that this isn't happening thirty times per second as part of your brain's natural operation?

My other rejoinder is, how about a Moravec Procedure, where you remain conscious and lucid the entire time during your gradual upload, in each moment satisfied (and confirming via repeated formal consent) that your consciousness has not changed?

The issue for me is that you don't actually get to colonise anywhere, nobody leaves, you just make another galaxy cluster full of humans. Maybe this is just an irreconcilable values difference, but I think this solution completely voids the point of the exercise. I don't intrinsically care about creating as many humans as possible and distributing them throughout the galaxy. I care infinitely more about where these humans come from.

Indeed, perhaps an irreconcilable values difference, but I think there's something ineffably beautiful about a quest to awaken the dead matter of the universe into sentience.

Fortunately, it's very cheap! The cost of sending von neumann probes beyond the radius of a given civilization sphere is likely to be low relative to the wealth of mass-energy contained within the sphere. That is true even when our sphere is just the earth, and becomes only more true as the sphere expands. Even if 99% of the sentiences or "mind-share" of the civilization agrees with you, the remaining 1% could probably fund the expedition as a hobby.

Let's assume we can go at, say, 50% light speed (149896.229 km/s). The expansion of the universe is 68 km/s/Mpc, and the value of Mpc that gives us a recession speed of 0.5c (the relevant formula here is 68 x Mpc = 149896.229) is 2204.36 megaparsecs, which translates to roughly 7 billion light years. Everything outside that distance is receding from us faster than that.

A 7 billion light year radius is pretty damn good, I'd say! We're well outside our local group and have begun eating superclusters like potato chips at that point!