Sunday, November 29, 2009

What Alien Colonization Would Look Like

In the previous blog entry, I explored what galactic colonization must inevitably look like to any civilization conducting it.

Note that any civilization that fails to expand its technological and industrial base to the point where galactic colonization is easy and cheap will eventually become extinct due to an asteroid collision, supervolcano eruption, glaciation or global war. This argument was first made years ago and it is well-established.

Note also that such homebody civilizations have average lifespans of at most a few million years. Given the many hundreds of millions of years in our galaxy's past such blatantly suicidal fucking idiots are of no consequence to us. I would even say, if they want to die, why should we not help them die?

But that isn't the subject of this post. Here we explore what galactic colonization looks like to any hypothetical backwards inbred hillbilly civilization unfortunate enough to be trampled underneath its expansion.

As already established in the previous article, AI makes starflight cheap and distance irrelevant. It takes just as much resources to accelerate a starship to near light speed then hibernate the AI for a 10 year flight as it does to accelerate a starship to near light speed then hibernate the AI for a 100 year flight. Or a 1000 year flight, or a 10,000 year flight. Or even a 50,000 year flight. There is no friction in space.

The only difference is the increased risk of a micrometeorite vaporizing the starship. And that can be easily surmounted by building an AI with an appropriate martyr complex and fanatical dedication to its mission. Why does that matter? Because it means that the wave of colonization will NOT be lackadaisical with sporadic branching to the nearest possible systems. Rather, the wave of colonization will be perfectly spherical.

The homeworld will be launching starships one after another, with the nearest star systems first in line, and it will never stop. Because it won't matter that the nearest 1000 systems are successfully colonized, the only thing that will matter is that there's another system out there that you can send an AI colonization ship to. And it won't matter that there is no immediate benefit. At some point down the road, perhaps only in 10,000 years, the homeworld will benefit in some way. If only due to exchange of technological information. And if it doesn't benefit, it's not like the starship was expensive anyways.

Let me make the point again. Many people (even supposedly educated people who study technology) have the deeply mistaken notion that the homeworld will send a short burt of colonization ships only to the nearest star systems. Perhaps even only to the nearest "inhabitable" planets. Then, after suitable millenia when each of those systems are developed enough, they will each send out their own wave of starships. That's not going to happen. This is ludicrous Star Trek fantasy.

What's going to happen is this. The homeworld will send a single giant wave of starships to every possible star system within line-of-sight. First to Alpha Centauri then, as soon as the first ship is accelerated to near-c, then to Barnard's star, then to Wolf 359, then Lalande 21185, and so on until the last star system in the galaxy within line of sight is targeted. Once the starships are all launched then the Homeworld will simply begin targeting the globular clusters and Andromeda. This wave of colonization will occur over centuries or at most millenia.

Speed

Now some uneducated people scoff at the idea of starships going near the speed of light. They claim that humanity has no idea how to accelerate starships to that speed and that our civilization will never develop such technology even given a million years. From fire to controlled nuclear fission in 10,000 years? Certainly. But from slow starflight to fast starflight in a million years? No never. Never ever! Apparently, technological development is something that only ever happens in the past.

As it happens, we already have a good idea of how to accelerate a starship to near-lightspeed. Isaac Kuo of Bad Astronomy and Atomic Rocket fame came up with a Starship Design Concept which can easily be accelerated to such velocities using current industrial technology. The best part is that it's cheap because it's reusable. Once the launch system is built, it can accelerate 1000 starships (one after the other) as easily as it can accelerate a single one.

For those who are interested, the idea involves an array of free electron X-ray laser beam emitters in solar orbit. The beams are focused using a giant fresnel lens. They can easily target the back of a starship's solid lead plate to a distance of several light-years. This enables the starship to be driven to anything from 0.9c to 0.99c. It also enables the starship to be decelerated using the exact same mechanism (x-ray laser beams near the homeworld) within a few light-years. Not that deceleration is the difficult part.

Furthermore, since every gram of matter carried by the starship at 0.9 c has an impact of a small 9 kiloton thermonuclear explosion, the starship can spray a fine mist of gas ahead of itself to annihilate all but large meteoroids in its way. The resulting spray of elementary particles can be swept aside with a magnetic field. To carry more gas for longer journeys, it's sufficient to make the starship longer. So long as you don't increase its cross-section, you don't make it a fatter target at all.

So yes, we actually do have a good idea how to go near the speed of light. And a civilization that's dismantling its home system to build a Dyson sphere will have centuries to perfect the design before implementing it.

What You Would See

So what would you see if some alien civilization were out there, colonizing our galaxy? Well for a millenium or so you would get radio waves. Then you would see the star associated with the radio waves becoming markedly darker. Within decades or centuries you would see the nearby stars go dark. Then you would see a hemispherical wave of darkness engulf every star at an astonishing rate. Within a bare thousand years from the first star going dark, you would see every other star from that origin point to your own sun going dark. And you would see this no matter where in the galaxy you are because the Dyson-sphere building starships would be only barely slower than the light you use to see them.

This is why everyone who yammers on about invisible or "hidden" aliens is a useless twit not worth listening to. Because you can't hide the stars going dark. If aliens were out there, it would be one fucking impressive sight. And as for the SETI notion that there is an alien civilization out there in the miniscule window between "doesn't have radio technology" and "is turning off the stars you see in the sky" ... Or perhaps SETI is interested in talking to blatantly suicidal fucking idiots? I don't know. I don't care either.

Do you see those stars in the sky at night? That's all the evidence an educated person needs that aliens really aren't out there. Unless you think the Great Void between galaxies is caused by aliens. Unfortunately, it isn't because dark stars emitting in the infrared aren't even remotely the same thing as no stars at all. Dyson spheres are quite distinctive and our astronomy hasn't found any. So are partial Dyson spheres for that matter, because the stars they partially cover would vary in brightness as regularly as pulsars.

Next up, what a galactic civilization might choose to do.

What Galactic Colonization Really Looks Like

I was wondering where the enduring interest in my post Aliens Don't Exist was coming from and discovered it is a small hit among Russian bloggers thanks to Alexander Semenov's translation, expansion and commentary. One thing I'm getting frustrated with is the people who persist in thinking galactic colonization, and galactic civilization for that matter, is some kind of Star Trek fantasy.

Do these people not realize that galactic colonization takes place over 100,000 years? Do they not realize what one hundred millenia, means?! 100,000 years ago, homo sapiens didn't exist! 10,000 years ago, homo sapiens weren't human. They did not have fluid language, they did not have consciousness. They had nothing that we would recognize as distinctly human. They were animals. It wasn't until relatively recently that these animals learned to control fire and hit rocks against each other to get a sharp edge. And it took much longer for them to develop anything we would recognize as consciousness.

Even a mere 1000 years ago, barely an eyeblink by astronomical standards, humans had not yet mastered steel. The universal speed limit was governed by the muscle power of the horse. Water mills were primitive and dams unheard of. Concrete had been abandoned after the Roman empire fell. Not that the Romans had ever used their poor concrete to its full potential. Think about this, an eye-blink ago, there was barely any steel and no concrete. And now suddenly we have multi-million inhabitant metropolises full of high rises and skyscrapers.

What's going to happen in the next eye-blink? What's going to happen in the next 10? 10,000 years is chump change in the galactic colonization game. Anything that happens in the first 10,000 years will be dwarfed by what's accomplished in the last 90,000 when galactic-scale plans start being made. But with absolute certainty, assuming that civilization on Earth doesn't collapse and take us out of the game, three very important things are going to happen in the next 1000 years.

Artificial Intelligence

Does anyone seriously doubt that artificial intelligence will not only be developed but will come to absolutely dominate civilization in the next thousand years? If you do then please stop reading and end your life as expeditiously as possible because you are a waste of perfectly good oxygen. There are no words to describe the stupidity of the notion that the technological status quo will continue unchanged for the next thousand years.

What does artificial intelligence mean? It means that intelligence is plentiful and cheap, that it is no longer a bottleneck in the economy. It means hopefully that rationality will become a fact of life and not the exception it currently is. It means that custom design (design requiring attention which can only be produced by an intelligence) will be the default. Every consumer good will be tailored to your specific needs. Every political, economic and intellectual opportunity will be intelligently evaluated, judged and explored.

Artificial intelligence will lead to a transformation of social relations to something most people can't begin to imagine. But for space travel, the consequences are very simple to imagine. Artificial intelligence means that you can send a 1 tonne solid cube of metal with a single specially chosen AI (one with no social needs and a low chance of becoming psychotic) on a flight that lasts thousands of years. As opposed to sending a multi-million tonne fat hollow shell of a target with 10 people on a flight that can only last a few years. Suddenly, colonization becomes cheap and distance no longer matters.

Molecular Nanotechnology

Same deal as AI. If you don't believe molecular nanotechnology is going to be developed and dominate in the next thousand years in the natural course of events then there are marmots that deserve the oxygen you're using up more than you do.

Molecular nanotechnology will utterly transform society. Suddenly, automation and general construction becomes dirt cheap. Homo sapiens are going to cease to exist. Quite likely, human beings will cease to exist too. And that's not a bad thing so long as technological civilization continues on.

You can whine about it all you want. You can agonize about whether humans will transcend flesh to become disembodied intelligences or whether they'll be crushed into extinction. Nothing you say, none of your whining, will prevent it. Molecular nanotechnology (unlike synthetic biology) offers too many advantages and too few hazards for anyone to stop its development. It's going to happen.

In comparison, starship technology offers only high advantages at the cost of enormous hazards. A starship such as the nuclear-powered Orion enables its crew to steer a large asteroid towards the Earth and exterminate the human species. Easily. This is the reason why nuclear starships are prohibited by the Nuclear Test Ban Treaty. Because the Americans were afraid the Soviets would threaten to, or cause, asteroids to fall on the USA.

Note that cheap starship technology is the only stable solution to the problems caused by expensive starship technology. Though MAD, or Nuclear Parity as the Russians call it since they were convinced that a nuclear war is surviveable with enough preparation (and they weren't exactly wrong either), seems to have worked so far. Nuclear weapons actually make the world safer because they prevent their owner from being invaded (thus making them feel secure) while making them terrified of invading anyone that might even remotely possess nuclear weapons (thus making them peaceful). The best part about nuclear weapons is those who seek them are so blinded by 'I cannot be subjugated' they don't see 'I can be destroyed'. [link]

Getting back to the subject, what does molecular nanotechnology mean for space-flight? Well if AI makes it cheap then nanotech makes it dirt cheap. Any dirt-poor moron will be able to afford to build and launch his very own starship. Are you starting to get the picture? Galactic colonization isn't going to be like on Star Trek or 2001 or any of that cheap unimaginative trash. It's going to be done by some guy with his trusty AI and his hand-me-down obsolete SUN Nanosystems Mark III nanoforge.

Unless some giant industrial consortium beats him to it. Which will likely happen too.

So .... given the above FACTS,
  • Artificial Intelligence will be developed in the next thousand years
  • Artificial Intelligence will come to dominate and transform society
  • Molecular Nanotechnology will be developed in the next thousand years
  • Molecular Nanotechnology will come to dominate and transform society
  • Together they will make enormous wealth attainable and fantastic projects affordable to everybody

then there is one inescapable conclusion. This is FACT #3 of the future:

Dyson Sphere

Yes people, we will construct a Dyson Sphere. Because there's another thing that those idiots with a Star Trek future in mind don't comprehend. That thing is arithmetic. Exponential arithmetic.

With fantastic wealth attainable to everyone, there is no longer any need to reproduce. And generally speaking, as people discover that it is extremely expensive to offer their children the kind of lifestyle they want for themselves, the birth rate crashes. But what happens when fantastic wealth enables an effectively immortal lifespan?

What happens when Artificial Intelligence eliminates fatal "accidents" caused by lack of attention? What happens when having children becomes cheap again? It is currently expensive only because education and high quality parenting take up a huge amount of human attention. What happens when you can raise children for 100 years and still have 99% of your life left over for other projects?

For that matter, what happens when the AIs demand more and more computing power? What happens when ever more sophisticated Pan-Dimensional Universe of Warcraft multiplayer games get online? What happens when someone wants to bring their favourite fictional character to life?

On the one hand, the environmentalists keep harping about the fact that we're already consuming more resources than one planet can provide. On the other hand, their proposed "solution" to this is to reduce the entirety of humanity to poverty and (when that causes an increase in destruction of resources) to exterminate the poor people through disease and starvation.

Well, poverty isn't in our future. It isn't considered a real solution by anyone who deserves to live. The only solution that is in our future is expansion beyond the confines of the Earth by building artificial habitats. And we aren't going to stop at one artificial habitat either.

We will construct giant rotating cylinders in solar orbit. We will construct more and more of them until all the resources of the asteroid belt are exhausted. Then we will start dismantling the Moon to construct more. Then we will dismantle Mars. And when there are no more moons and planets left, then we will dismantle Earth. This is inevitable. It's going to happen.

It's a simple consequence of exponential growth. Even a 1% economic growth rate compounded over 1000 years (or 0.1% compounded over 10,000 years) implies a total growth in excess of 20,000. If Earth is insufficient NOW then we will need the equivalent of 20,000 planets in the year 3000. Fortunately that will not be much of a problem since planets are the most inefficient habitats possible. Any space habitats we construct will make much more efficient use of material resources.

Economic expansion will always, always hit physical limits. It cannot be otherwise since a 0.1% per annum expansion over 100,000 years yields growth in excess of 10^43. The entire Milky Way galaxy masses only 10^42 kilograms, to put that into perspective. Currently, there is more than one metric tonne of steel and concrete being produced for every man, woman and child of industrialized society each and every year. In 100,000 years, if galactic colonization is successful, there won't be even a single kilogram. And that's assuming we can keep population growth rates at an ultra-low 0.1%.

So you still believe in the Star Trek future? You still believe that "humans" will go out to the stars in great big million-tonne starships? You still believe that they'll "settle" so-called "habitable" planets? Then you're a good example of how intellectually decrepit and illogical our species is.

It isn't "planets", habitable or otherwise, that civilization will settle. It's star systems. By building dyson spheres around the star using whatever available junk there is at hand. Junk like asteroids, moons and entire planets. And if some of those planets happen to have biospheres then it won't matter. The biosphere will either be bulldozed under, or possibly, transplanted. But it sure as fuck isn't going to stop civilization from using the planet underneath in the most efficient and intelligent manner possible. And "using" a planet means dismantling it. Because a highly technological civilization neither needs nor wants planets.

This isn't your Star Trek future.

So this is my response to the following arguments:

  • There's no need for galactic expansion. -- Speak for yourself you treehugging dirtmonging druid.
  • Nobody would want to bulldoze Earth. -- I would. And there's a trillion credits in it for anyone who votes for dismantlement.
  • Civilization would stop everyone from trampling "preserves" through totalitarianism. -- Die you fanatical anti-human scum.

Next up is Part 3 of Aliens Don't Exist - What An Alien Galactic Civilizaiton Looks Like. Or: Why SETI is retarded.

Monday, November 23, 2009

C++ Programmers All Ought To Die, Die Motherfuckers, Die.

So I'm porting this 3D engine called Horde3D to Smalltalk. Why? Don't ask, it's too involved. Why Horde3D? Cause it's short. Naturally enough, it's written in C++ as every 3D engine seems to be.

Now, the quality of the code involved is pretty poor by my standards. Then again, I've got seemingly ridiculous standards. Seemingly because nobody seems to be able to meet them consistently. But this is something else.

There's this whole crap about templates. You see, instead of creating a Node and filling it with the parameters you want, you instead create a String with all those parameters. Then you create a NodeTemplate by cutting up the String. Then you create a Node by parsing the NodeTemplate. Niiiice.

Supposedly this whole rigmarole is so that you can make lots of copies of a standard Node. Unfortunately, it doesn't make any fucking sense. The right way to make copies is to ... copy. You create a Node and then you copy it lots of times. It's not difficulty.

But the rationale is more involved than that because IFF creating a node is ridiculously more expensive than creating a template then it can be a small win to create 1 template + N nodes rather than N+1 nodes. You know, cause that extra 1 makes a whole lot of difference.

Meanwhile, it doesn't seem to matter that there are two classes (or one struct and one class) instead of a single class. Or that there's code to check whether you've got an appropriate struct when you're trying to convert it to a node. As opposed to, you know, never doing any conversion in the first place.

But you know what? That doesn't even nearly take the cake. Cause now I've just discovered that matrix multiplication, yes simple matrix multiplication, is reversed! It's all fucking backwards. The operators and functions take two input matrices, m1 and m2, and instead of doing the seemingly inevitable of multiplying m1 * m2, it does m2 * m1.

Seriously, WTF?!

I'm not even gonna go into the complete lack of any coding standards in this piece of dreck. The morons can't even standardize on "return a NEW object based on performing this operation on this object" versus "become the new object".

And the worst part of it is that this engine is supposed to be the cleanest.

There is only one possible conclusion from this. C++ programmers must die. ALL of them. Every single fucking one.