If you prefer to listen rather than read, this blog is available as a podcast here. Or if you want to listen to just this post:
Knowing that the 50th anniversary of the first Moon landing was this month, I felt a strong compulsion to say something about it, but what exactly? There is definitely no shortage of commentary related to the occasion. Obviously the most interesting part of the anniversary is the fact that we haven’t been back since the end of the Apollo Program, nearly 47 years ago. Perhaps I should wait and say something when the 50th anniversary of the last man on the moon rolls around. (Does anyone think we’ll make it back before then?) But I’m definitely not the only one to have noticed this distressing fact, most of the commentary surrounding the anniversary mentions the fact that we haven’t been back. What can I say on this occasion that would be unique?
I do think there’s something interesting to be said about the connection between space travel and human salvation, but I’ve already covered that connection and however unique that observation is, I don’t want to just rehash what I’ve said previously. So what can I bring to the table that isn’t being served in dozens of different locations by hundreds of other commenters? Well, as I survey the commentary I think there’s a definite dearth of extrapolations. Sure, humans will probably make it back to the moon. (I assume that if no one else gets around to it China will, at least, if only for reasons of national prestige.) But if we extrapolate things out and look at the trends, when are we likely to be there permanently, and what about Mars? And, perhaps most important of all, if current trends continue when would humans actually leave the solar system? Obviously this exercise will produce only the crudest of numbers, but I expect that whatever comes out will be pretty depressing even if I end up off by a factor of 2 or more.
To start, though, for those who never read or can’t recall my post on the connection between space travel and salvation, and who don’t have the time or inclination to go back and read it, I should briefly explain my point, which is: If you want to ensure that humanity continues for as long as possible, and you don’t believe there’s any external force capable of helping with that (religion, aliens, vaguer forms of spirituality, etc.) then, ultimately, this is going to require getting off the planet in a sustainable and ongoing fashion. In that post, I further pointed out that most of the large scale goals we’re pursuing at any given moment have very little to do with this endeavor and in fact work against achieving it, if for no other reason than opportunity cost.
It’s upon considering this last point that branching off into an extrapolation of trends starts to look like an important next step. Yes, occasionally when a technology becomes available, things can change dramatically, and trends before this change become meaningless, a great example of this is the internet. But in the case of space travel we’ve had all the relevant technology for at least 50 years (and yes, I’m aware of the EMDrive) but yet so far there’s been no dramatic upward spike in space travel, particularly if we view the Apollo Program as an outlier (as I am inclined to do, see my post about S-Curves). Accordingly if our salvation depends on getting off the planet, and we have 50 or more years of data on the rate at which that’s actually happening, and every expectation that this rate is unlikely to change very much, then, it would definitely appear to be worthwhile to extrapolate out these rates and see where they get us.
None of this is to say that the rate of space exploration and colonization isn’t increasing in an exponential fashion. In fact, for all of my trend extrapolation, I’m going to assume that there’s some underlying law along the lines of Moore’s Law, where a given quantity doubles every X years. Meaning that we merely have to decide what a reasonable rate of doubling would be, using the last 50 years worth of data. I’m not actually saying that there is a parallel to Moore’s Law when talking about space, I’m more saying that there had better be, because the distances from one destination to the next are already exponential. Meaning that we’d better hope things are growing exponentially because otherwise space colonization is definitely doomed. Also something that doubles every two years is going to already assume significant ongoing technological advancements. Meaning that if you do think something like fusion or the EMDrive is going to come along and drastically change things, those advances are probably already built in to the model
For our first example, let’s start off by making the hugely optimistic assumption that the current trend is for the distance humans are capable of travelling to and returning from to double every 10 years. And if we then take 1970 and travelling to the Moon as our starting point, we wouldn’t make it to Mars until sometime in the 2040s, Jupiter would be about 2075, Neptune around 2100 and Alpha Centauri would not be reached until the year 2230. And If we, instead, made the more reasonable, but still fairly optimistic assumption that the distance only doubled every 15 years, then we’d get to Mars around 2075, Jupiter would slot in at 2120, Neptune would be 2180 and Alpha Centauri wouldn’t be until sometime around the year 2360…
That last one may not seem especially optimistic, but recall if we’ve decided that space travel is important for our long term salvation it’s not enough to get there once. Surely some nation can massively divert resources for a single moonshot, which is what the word came to mean, and possibly put people on Mars a half dozen times and bring them back, but in order for it to assist with our salvation we have to be able to do it on an ongoing, perpetual basis. And, of course, not only is all of the above optimistic, but based on a single data point: putting a man on the Moon. Not only have we not gotten any farther than that, we haven’t even been able to do it on the ongoing and perpetual basis I’m talking about. But before we leave this example, let’s conduct the exercise one more time, and assume that, as he has predicted, Elon Musk manages to put someone on Mars in 2024 (and by the way here we would appear to be in the realm of the insanely optimistic). This would finally give us a second data point and putting that into the crude model I’m using it would mean a doubling approximately every 7.5 years. Which gets us to Jupiter in 2045, Neptune in 2075 and Alpha Centauri in the year 2165. Not bad, but still a lot slower than most people imagined 50 years ago, and here we touch on one of the problems.
In many respects we’re living in a science fiction world more incredible than anything anyone imagined in 1969, and in other respects, particularly when one looks at space travel, someone reading Heinlein, Clarke or Asimov would be profoundly depressed by how little progress we’ve made. And yet the idea that any day now things will change and suddenly we will be living in that world is hard to shake. Certainly there could always be some dramatic new invention that would change whatever curve we’re currently on, but at the moment there’s good reason to think that, absent some massive space exploration/colonization inflection point in our future, the current rate of plodding along isn’t going to get us anywhere very fast. Now it may be that it doesn’t matter how long it takes, as long as we get there eventually, but a lot can happen between now and even 2024, to say nothing of 2040, 2075 or 2360. Recall that there’s good black swans and bad black swans, and while the former may be exactly the positive inflection point we were hoping for, there are a lot more things which could happen that would make this whole project much more difficult rather than less.
Moving on, what other trends are there that we can extrapolate? Above I talked about something being continuous, and we have had continuous human presence in low earth orbit (with occasional gaps) since 1973 when Skylab was launched and occupied. All of this has occurred at around 250 miles from the surface, but I’ll be generous and round up to 300. With this as our new starting point we can once again imagine this distance doubling every so many years, only this time it gives us the distance from Earth where humans will be able to sustain a continuous presence. If we once again start with, what I feel, is an incredibly optimistic doubling time of 10 years, we will have a continuous human presence on Mars in 2140, Jupiter (or one of its moons) in 2175 and a continuous presence at Alpha Centauri around the year 2300. If we instead assume a more realistic trend of doubling every 15 years, then Mars is 2225, Jupiter is 2280 and Alpha Centauri is not until the year 2500.
Now I understand that certain things might get easier, for example just getting out of the gravity well of the Earth is a major hurdle, and perhaps we should take that into account, but when you’re talking about a continuous presence, I would argue that getting out of the Earth’s gravity well, is perhaps the least of your worries. Also recall that to a certain extent productivity gains are built into the model of exponential growth we’re already using. Finally, these extrapolations are not meant to be especially precise, but rather to illustrate that even using some fairly generous assumptions space colonization is going to be a lot harder than I think most people realize. Particularly given how spectacularly unimpressive our manned efforts have been since the end of Apollo. But, perhaps that’s where I’m going astray, by so far only focusing on manned efforts.
Unmanned exploration really is the easiest way to explore space. And while unmanned probes do not directly accomplish that “salvation of humanity” I keep coming back to, they are at least a reasonable potential stepping stone along the path to that. With that in mind, what kind of Moore’s Law might we extract if we turned our focus to unmanned exploration? Here, at least, we have multiple data points, one for each celestial body, and if you graph it, it looks like a pretty nice exponential curve:
There are a couple of things to note about this data. First given that Uranus and Neptune were both first visited by Voyager 2, I’m not sure if Neptune should count as a separate milestone from Uranus (or perhaps it’s the other way around). Also you’ll notice that I didn’t include Pluto, if I did you’d see that nice exponential curve flatten out into something that looks a lot more like a plateau, since, at the time New Horizons visited it, Pluto wasn’t that much farther out than Neptune and we didn’t get to it until 2015.
Mapping this to our simple model of deciding on a doubling rate is messier with actual data, but after fiddling with it a little bit it looks like seven years fits fairly well. Taking that and anchoring it around Voyager 2, I came up with an arrival time for the first probe to Alpha Centauri of around 2110. Which is almost exactly NASA’s current estimate of a 2113 arrival for the probe they plan to launch in 2069. (You’ll have to take it on faith that I came up with my number before I found the number from NASA). These estimates might be pessimistic, given that Yuri Milner, the Russian billionaire, is proposing to launch a probe by 2036, which might arrive as early as 2056. But when you get into the details of that proposal there’s reason to question whether it should necessarily be placed in the same category with all of the other probes. The probe proposed by Milner’s team weighs only a few grams and would enter the Alpha Centauri system at 20% the speed of light. Which means the probing part is going to end up being some infinitesimal fraction of the entire trip.
Of all these trends, the trend in unmanned probes is the only one that seems a little bit promising, and even there, it’s going to take quite a while to get anywhere we haven’t already been.
Fifty years ago everything seemed so promising. What happened? What happened to the science fiction dreams I grew up on? Instead the best way to describe space exploration over the last 50 years, is vaguely depressing with occasional all to brief glimpses of triumph here and there. And perhaps even worse than that, there is no sign that the future is going to be any better. Instead most of our energy seems focused inward, and the Great Silence of the universe becomes less and less paradoxical.
The harvest is past, the summer is ended, and we are not saved.
Space: the final frontier. These are the writings of a slightly unhinged blogger. His five-year mission: to explore strange new topics. To seek out new controversies and new weirdness. To boldly go where no man should ever go period! If you’d like to help with this mission consider donating.
For all but one of these metrics you’ve applied a continuous model to an entirely discontinuous data set. I’m not sure extrapolating a continuous trend on that kind of data set even helps inform our intuitions here.
My key point is that I’m pessimistic about things, but I think the best even the optimists can hope for is some kind of Moore’s Law progression. But even using the most optimistic method for modeling and optimistic intervals on top of an optimistic model, it’s still looks like it’s going to take longer than people think. So the question is not is there a better model, but is there a more optimistic model that’s more than “and then everything worked out really well, better than we imagined.”
Okay, so let’s use a better analogy than Moore’s law to model the system we want to consider. We need something with a discontinuous data set that matches the challenges of space exploration. The obvious parallel, to my mind, would be the oceans. Ignoring examples of travel along ocean currents to populate tiny islands (Pacific islanders?) let’s just focus on European attempts at ocean exploration.
1. At first they stayed close to the shore. Yes, there were people who would strike out due West, but those ships never returned. Without a good way to determine longitudinal positioning they couldn’t navigate very well once they got away from shore. The most they could do is get away from shore a little bit, then quickly return to shore by traveling due East. Lots of sailors circumnavigated Africa, with some evidence that this was being done regularly for hundreds of years before 1492.
2. Once sufficient navigational tools were available, the biggest problem was in the distances involved. They knew the circumference of the Earth to a pretty good degree of accuracy, and since they assumed there was nothing between them and East Asia, there was no point going on an endless voyage they couldn’t possibly survive. It wasn’t until Columbus convinced Spain (mostly Isabella?) that the Earth wasn’t as big as it actually was that he got funding for a trip due West. He made it to land, even if it wasn’t the land he expected, and began ‘exploring’.
3. Once the Americans were discovered, within a very short period of time relative to the amount of time it had taken for Europeans to develop the technologies to get there, they were exploring and conquering everything of value. Sure, it took a lot longer to get to Seattle and Alaska, but at the time there was nothing there worth getting. They weren’t exploring for the pure human delight of knowing what’s out there. They were exploring and colonizing because there were resources to get.
Now let’s look at space parallels:
1. We’re clearly still in the ‘staying close to shore’ era. Why? We haven’t developed the technologies sufficient to get ourselves out and back again safely. It’s deceptive, because it feels like we should just be able to build the space ships that will take us out there, but there are a lot of complexities involved that you don’t discover until you start working on the problem directly. Like scurvy, for the sailors.
2. Once those tools are in place and we’re able to go out exploring, we can’t assume everyone will just sit around as though the new tools are unimportant. Yes Spain got a head start, but Portugal soon got into the game followed by the British, Dutch, French, and (once Bismark brought them together) the Germans.
This has also been happening with Space. Elon Musk proved he could put stuff into orbit on the cheap. Now the Europeans, Chinese, Indians, and everyone else are getting in on the action. I saw a startup the other day (based out of Australia?, can’t remember the details, sorry) is trying to launch very small rockets at a million dollars each. Once it can be done, people start pushing out farther and faster. There’s little reason to doubt we will see something similar with deep space, where the quantum step from no market to an open market changes the rate of growth dramatically.
3. Although Elon Musk really wants to get to Mars to diversify humanity’s presence in the solar system, the capital that will make that dream happen doesn’t share the same goal. That capital, so far as we can see right now, is interested in natural resources. Currently those can be found more cheaply here on Earth, but there are massive amounts of rare resources on the moon and especially in the asteroids.
Of course, nobody is going to invest in a project with massive unknown costs that so far has only been achieved by one government project of such massive proportions that it is its own saying. But if we succeed in the current plan to build an international deep-space station capable of switching orbits between the moon and the Earth, not only will the costs of going to the moon go down, so will the number of unknown variables. Thus, capital will be able to reach the moon – and beyond – and we should expect it to begin flowing in space similar to how currently does on Earth. In other words, I expect the next moon landing to be like the iPhone launch, where proof of concept is followed by a competitive race to fill in the gaps in potential. For cell phones, this was more than just adding a multi-touch screen to existing mobile technology. It meant the massive explosion in deployment of MEMS devices that make modern cell phones capable of such an impressive array of functions. Similarly, we’ve been developing huge amounts of technology over the past few decades (and doing large numbers of experiments at the ISS) that we currently employ in limited ways, but which could be more broadly applicable once applied to space.
I’ll borrow a theme from one of your recent posts here: we should expect technological development in space exploration to follow a punctuated path, rather than a gradual one. Sure it takes lots of small steps too, but to get there must mankind take several giant leaps along the way.
I’m wondering if distance is really the measure to look at here versus population density.
Let’s consider European colonization of the Americas. I’m finding from wikipedia in the 16th century (that would be the 1500’s) , they estimate 240,000 Europeans entered American ports. In the 1700’s slavery to the 13 colonies alone was 387,000 and total slavery to the Americas had to have been a few million at least. On top of that we have to add European emigrants to the Americas.
In terms of civilizational survival, increasing density is more important than miles. Going from 1,000 to 10,000 people living in the Americas was more valuable to the European civilizations than sending one man on longer and longer journeys.
I suspect then you have to consider the number of people sustained at any given time away from the earth. For example, getting to 10,000 people living at any given time between low earth orbit and the moon would essentially provide a huge amount of insurance against Earth bound catastrophe. Getting 100,000 people living between low earth orbit and between either Mars or Venus would essentially solve the problem of the sun expanding in the next few billion years. Populate the solar system with a few millions of people and you’re more or less safe from just about anything except maybe a nearby star going supernova (and even that would probably be ok given we’d have the tech to put settlements under deep shielding from radiation). So the measure should probably be ‘are we filling out’ nearby space rather than how fast are we going the next million miles.
But this is difficult to measure. Columbus was 1492 and while I wish I could find the numbers I suspect even 50 years later the Europeans in America were probably not that impressive. However the rest of the century exploded and the century after that was even larger. Basically when you’re dealing with any exponential function and your datapoints are all near the bottom, it’s impossible to see a curve or a flat line. They look almost identical.