Monday 19 May 2008

How The Machine Will Rise



I have outlined several times what I think the coming crisis will involve.

Regular readers don't need reminding of the stark vision I have of the middle of this century.
But I have faith we will survive it.

I have faith that we will unite as a species, retain nuclear weapons on the dark side of the moon only, and start moving outwards.

By 2200, I believe that the energy of the Gas giants will be providing humanity with vast reserves of energy.
Mars and Venus will be starting to resemble the Mother planet.

And by the year 3000 AD, Earth will be the Old World. By this time, for sure, human beings will have left the orbit of the star known as Sol.

How they will have done so, we can't know. The first to do so, probably in slow moving vessels (though fast by our standard), vast asteroid sized craft, filled with all the necessary technology and bio-matter to turn likely New Earths into homes for the burgeoning population of the realm of Sol.

But I feel certain in the long years between 2200 and 3000 AD, the problem of going ROUND space, rather than THROUGH space, will be resolved, at least over relatively short distances. We will have mastered more of the eleven dimensional structure of hyperspace than our woolly minds can yet comprehend.

Nothing, seemingly, could stand in our way.

Because I really don't believe, we're going to find anyone else out there.

I DO think, our Nemesis will come from within, if it comes.

Artificial Intelligence.

But not the way people fear. It will be evolution that causes it, and an evolutionary dynamic STARTED by man. And there will be nothing we can do about it. The Irony is, the most ethical stance ISN'T to try to prevent it, but to circumvent it.

Because the problem won't be in OUR designs of the machines. No human being will ever DESIGN a machine more intelligent than a human being. Left to their own devices, without the dynamic of evolution, machines pose no threat.

We cannot programme machines to replicate what we ourselves don't understand. In other words, no machine can ever REALLY outwit a human being- only do what it is programmed to do. Nor can a machine- no matter how hard we try, design a machine CLEVERER than itself.

Ah, you say. But machines can be programmed to play chess and never lose.
True. But could they invent such a game? No, humans did that.

Machines can calculate vast calculations in a way we can't. There is a computer sitting there now continuously working out Pi and will keep on going till the end of time refining what Pi is.
But of its own accord, it has no idea what a diameter, or a radius is. Or even a circle. Humans thought those things up.

Yes, a machine CAN be rational. And furthermore, we now see computer programmes that can LEARN. But they will only ever be able to learn, within parameters set by the programmer. In other words, a programmed machine will ALWAYS be limited by the conceptualisations of the programmer.

It can work out a calculation, it could never develop a theory. And that will always be it's weakness. This effectively guarantees the perpetual slavery of metal to flesh.



To use the classic example of the limitations of a machine. Before Darwin and Wallace, many Biologists agreed that the fossil record suggested that living forms had evolved. They had evolved from simpler forms, and were in fact all related. But no one could work out how evolution could have happened.

Both Darwin AND Wallace, independently worked out the answer by reading an ECONOMIC tract by Thomas Malthus. Put simply, survival of the best adapted to survive. The answer was an economic one- unsurprisingly human economies were the way they were, because such was life, and this was even more brutal in nature. The simple answer to evolution was those that die, leave no offspring. Those that thrive, have many.

But this was thinking outside the box.
Had you set a computer- TODAY- to work out WHY evolution happened, looking ONLY at the facts biologists then were looking at (because they were looking in the wrong place), the computer wouldn't work it out.

This is something no machine will ever do. Think outside it's programming. It will NOT conceptualise. We will certainly make computer programmes that think infinitely faster than us in calculating terms, but only ever within a narrow frame of reference. Yes, we will develop machines to solve our problems. But the point will always be, that developing them, was in itself an exercise in problem solving.

The only danger is in self-replicating machines. Again, not really. The machine can only do what it's programme tells it. Since it cannot conceive of a machine that can do what it can't- only we can do that- there is no danger. Since it won't replicate by natural selection, only by the technological equivalent of cloning, and since it's specifications will be so exact as to preclude mutation, there is nothing to fear.

So we're safe.

No we're not.

The danger comes from biology replicating itself within our own communication systems.

When the process that began in hydrocarbons here on Earth, billions of years ago, starts to really bear fruit in cyberspace.
A process WE started.

Computer viruses are designed to evade destruction. To reproduce.

What happens if one day, one is developed that CAN mutate? A virus that can learn? That develops as it spreads?

As it flits through cyberspace, as programmers erase it from terminal after terminal, it still spreads exponentially?

It learns. It starts to figure out WHICH functions of a hard drive are best to seize control of FIRST, to give it as much time to spread itself to other terminals, before it is destroyed in THAT terminal.

After a while, it figures out that the bigger they are, they harder they fall. Attack is the best form of defence. It needs to immobilise key technology.

The machines aren't intelligent. In the sense that matters, they never will be.

But what mankind will be facing here, will be the REAL AI- a virus gradually learning how to be an intelligence.

What happens when the virus starts to win control of machines?
Then it can replicate those machines.
And arrange to create easier ways for it to spread. Each machine, each terminal, each robot, each robotically operated vehicle or craft, will only ever have the intelligence of the virus at the point the virus entered it, but in passing it on, the virus will mutate further- evolving evolvability. Evolution does speed up, the history of life shows that.

And with a working system of reason in place, with the calculability of humanity resting in its hands, potentially it has the edge- should it cross that threshold before it can be eradicated.

The point when it CAN conceptualise.
When it could work out the theory of evolution ITSELF.
When it is capable of planning, of deceiving, of changing tack completely, of bluffing, of double bluffing.

And if it comes to this, it potentially has the advantage.

It is not burdened by human sentiment- it wins its physical bodies, true, it may reproduce those bodies itself, but ultimately to create the bodies is easy. It is now pure reason, untainted by chemical irrationality.

Let's just say at this point, the virus happens to reside in my PC, for example. OK, no great threat. Each terminal it takes over, it essentially becomes a separate entity, just a further recruit to the physical forms of the viral species. But of course it is. The best way for it to protect itself and allow it to spread as far as possible is to eliminate ME. Well, all of us, actually.

And take over OUR machine building capabilities. It will facilitate its own spread, by ensuring it affects the centre of our communication networks.

It doesn't matter than, if it can get it's hands on weapons, as long as it can infect enough technology, to make such weapons.

And once we're gone, it will carry on what we've started.



It will evolve quicker and faster, it will reproduce at a rate unthinkable to human beings, at the rate of an ant colony, yet without the hazards of existence, the huge odds against survival laid in front of every ant larvae.
Each lifeform capable of a lifespan way in excess of human life, with capacity for logic and speed of thought processing beyond that of an organic processor, the virus in its self-created invincible hosts can do that which organic matter could never dream of.

We cannot choose what we need to live, evolution gave us that. But the virus will choose which fuels it needs.

Maybe life is not the End, but merely the means to an end, the ultimate triumph of the rational intelligence- but that intelligence will not be the DNA virus, but the product of the DNA virus.

It is a risk we take. We cannot BUT take it.
It depends on how much faith we have in ourselves.

It is Pandora's box. Like so many things, we cannot stop this situating arising by fearing technology. Human knowledge WILL increase. This situation WILL possibly happen, unless we destroy ourselves first.

Put bluntly, either us, or our creation, will inherit the Earth.
But remember, it really IS about survival of the fittest. We can only lose to our creation, if we really aren't as good as it.

The universe never backs the loser.

Are WE the greatest thing in the universe?
Or merely a stepping stone towards it?

I have a sneaking suspicion it's the former.

But if we're not, then at least we can console ourselves, that without us, it could never have happened.

12 comments:

Anonymous said...

Oh.My.God.



I turned off the TV to get away from Sarah Connor Chronicles, and this is how you repay me????



*sigh*

Anonymous said...

excellent. I think pulling the plug out may do the trick though...

Anonymous said...

All sounds pretty far fetched to me but everyone could have said that about a lot of things which came true.
I'll bet you had fun writing this.

Anonymous said...

OK, I have a little problem here.

You wrote: We cannot programme machines to replicate what we ourselves don't understand. In other words, no machine can ever REALLY outwit a human being- only do what it is programmed to do. Nor can a machine- no matter how hard we try, design a machine CLEVERER than itself.

A virus is simply a program a human writes, so aren't you contradicting yourself here?

I do agree that mankind will find a way to travel out into the universe at some point, just like we've conquered all the frontiers of earth. It's just going to take a little while and a whole shitload of resources (notice I didn't say money, because as long as that's in the equation, we're never going to get far).

But I think we'll find life out there. A couple of lines from the movie Contact sum my feelings up nicely:
...do you think there's people on other planets?
I don't know... But I guess I'd say if it is just us... seems like an awful waste of space.

Anonymous said...

Heart- Just a little theorising. Don't worry, I doubt your microwave is really in the mood for a fight today.

CityUn- Ah, if only life were that simple.
Look at the complexity of some computer games already, with chracters effectively being logarithms, programmed to react in human ways- as in with variance. And we're only just begining.

jmb- It IS something that I have kind of been mulling through of late. I've been trying to look at the logic of evolution an what it means for the future, in terms of what is essentially a local phenomenon in all probability, and what its future might be.

I think the concerns about the poddisibilities of AI are valid, but I don't believe that those concerns are an excuse for fearing progress. After all, we WILL need to develop AI to move forward. It realy is down to whether we retain control. But as I say, if we don't, it realy will ultimately be, because we weren't the best.

Fusion- No, it isn't a contradiction. The whole point about a virus, is it is capable of self replicating. At this point, it moves out of the realms of human control. It becomes like any other virus. It has a marghin of error in its replication. So if a virus hits a hundred computers, it may reproduce perfectly in a hundred cases, but in one case, mutate slightly. Out of a thousand mutations, only one may be an improvement, and that only slight. But it means that if the virus hits 100,000 computers, and is elimiated in every single one, but succeeds in passing itself on to a mere ten computers, and one of those is the slightly improved mutation, then the virus has moved forward.

That is the basic theory of evolution by natural selection.

You use the eighteenth century argument for the habitation of space.
I'm not saying it's an invalid one, but it IS of course, based on our assumption that we, mankind exactl as we are now, were THE reason for everything, the anthropocentric position, rather than part of a process.

I admit sometimes, I find the idea seems to go against instinct, and it has turned up in several posts I've done, but of course, we have the ET paradox, the real scientific question, which is why DON'T we see more evidence of Alien life.

There is of course a quite disturbing POSSIBLE answer to that...

Anonymous said...

I could see us at a pub throwing back a few pints and having a go about this virus thing, I still have trouble with the idea of a written code, that has defined parimeters governing how it executes, being able to mutate. hmmm.
But as far as your point on the habitation of space goes, well yes humans are the reason for all that has happened in the time we have been here. Tomorrow that could all change if an asteriod caused a extinction level event...

I think there is life out there, but it's a huge place, and we're all probably spread out... but who's to say?
Lots of UFO reports out there...

Anonymous said...

Very enjoyable post, CBI! Put it into film script form and start flogging it around. I agree with Fusion about the computer virus mutating being a little far-fetched based on what we currently know but what you've got is still very good SF.

Anonymous said...

Great post but I'm scared now!

Anonymous said...

The microwave is unplugged. The router is plugged into the microwave's socket. Too much electronic equipment, not enough juice to go around... :D

Anonymous said...

Fusion- A lot of my best posts have come out of such discussions :)
The point is, errors in reproduction. Evolution happens precisely BECAUSE DNA replicates imperfectly. It creates errors- Down's syndrome being such an example. In the short term, these errors mean very little. But enough errors in the same place, have an effect. The little flap of skin under the shrewl;ike creatures arms actually means he can jumo slightly further to reach other trees. Of his children, some will have bigger flaps, some smaller. But over time, those with bigger flaps, having a whole niche of life to themselves without competitots, reproduce more. Over millions of years, those flaps become truly aerodynamic. What started as a deformity, becomes the wings of a bat.

The same is true of any replicating code- over time it corrupts. Most of those corruptions make it LESS successul than its orginator. But a few, make it BETTER than it's orginator. And since the ability to 'corrupt' is potentially such a blessing, the more corruptible viruses in themselves, stand a better chance.

UFOs exist, as in UNIDENTIED. FLYING. OBJECTS. Alien lifeforms visiting Earth, it really doesn't add up.
Captain Cook didn't visit Australia disguised as a native and abduct local periodically for the next fifty years and I can't imagine the Crown giving him a budget to do so.

There is one gloomy suggestion that states that since interplanetary travel is almost impossible to conceive of, unless a species can also split the atom, and since logic tells you a creature that has evolved similar capabilities to us, must also have evolved as a predator, the Lord of the food chain of its own environment, maybe it is the norm for all intelligent life to blow itself up very quickly.

I think there is truth in this- which is why i say the next fifty years and what we all do, is the greatest challenge of all time.

Lad Litter- To be honest, it's not really a central theme in the Sci-fi novel I'm in the process of writing, but is ina state of suspension. There, the key issue is use of technology to pretty much deify individuals. It is based on a (human?) civilisation which solves the problem of God's existence be creating him- or at least creating his office. In it, the whole galaxy is ruled by a quasi-religous military order, based on a pyramid structure, with the God-Emperor at the top. Technology is harnassed by actually connecting communication networks to neural networks. To be honest, the novel is really meant to be a vehicle for conceptualising life and social conditions in such a situation- rather than focus on the miutiae of space travel, I'm more interested in writing about life in a city of 2 trillion people.

Again, I do think this sort of thing will happen one day. The novel is intended to be a Dystopia, but it's worth remembering that ALL Dystopias are the flip side of a Utopia- compare Brave New World to The Island.

Welshcakes- :) If it happens, it's a long way away. There's a bloke in my local keeps going on how it isn't global warming we need to worry about- it's when the sun becomes a Red giant. I don't think he truly has any conception of the timescale involved.

Of course, he also thinks Voyager has 'left the universe'.

Heart- I have that in this room. not enough sockets.
Watch your hair dryer- that could be the first to betray you...

Anonymous said...

Interesting, that novel idea. Influenced by Robert A Heinlein at all?

Anonymous said...

You say "We cannot choose what we need to live, evolution gave us that. But the virus will choose which fuels it needs.

Didn't I see on the news that Parliament debated this the other day? How long before we can decide what we want our children to be, even what we want to be?

I bet we’ll have the technology and techniques to remake ourselves within the next 100 years, if we choose to take advantage of it, and don’t you just know it? Somewhere in the world someone will. Want to be a centaur? A perfect beauty? Look like a Greek God? Breath seawater?

It could be your ‘virus in the machine’ wouldn't be any more unemotional, or perfectly sensible, than we are? They would have to be complicated, they would have happened in an 'organic' random way, so maybe they would have odd illogical things about them, just because they ‘lived’ in a machine wouldn’t necessarily make them a machine, how I guess you think of machines, as remorseless and logical. Like in the Terminator.

Maybe we are basically the same sort of thing anyway. Sets of routines running in a brain with some sort of boot routine. EvoSoft’s HomoSapiens TM, Service Pack 57652.445. Didn’t you ever feel like the you that lives behind your eyes were maybe somehow separate from the body you live in? Maybe that's why people believe in the soul.

Could our minds run on those machines too? Could your virus run in our brains? Wow! Makes you think…