Saturday, May 23, 2009

Singular Fiction

It's time for science! Here's what I've been thinking about lately.


It took eight billion years for this universe to produce a planet (ours) which can facilitate life as we know it. It took another three billion for that life to develop. Two billion years later, simple cells became complex cells. For single-celled organisms to become multicellular organisms took one billion more. Animal life developed 400 million years later. Another 250 million saw the rise of smarter animals (amphibians). 150 million more for mammalian life, still smarter. About 200 million more for genus Homo (humanoids) to appear. 2.2 million more for modern humans to develop.

Humans have been around for about 250,000 years. In that time we have forged thousands of societies, learned to hunt, to farm, to coexist with one another, to wage war. We have spread and adapted to environments in all corners of the earth. All of this we've accomplished in a tiny fraction of the time many other human species have existed. In the last 12,000 years we've developed agricultural knowledge. Animal farming, cities and states followed. In the last 5,000 years we've used metal tools. We've developed religion, written and spoken languages.

Each advancement comes faster than the last. Compare the last hundred years of technological advance to the thousand preceding them. Compare the last thirty to the hundred preceding them. Advances are occurring at finer and finer levels. At greater and greater speeds.

The modern era has seen the rapid rise of technology, from printing to telegraph to radio to the sophisticated information systems we have now. Computers as we know them have existed for mere decades, and with them we have transformed the planet. Civilizations of the past have risen, fallen and decayed; in the present we've achieved global communication and the construction of a global society. Ever faster.

As computing technology grows exponentially more efficient, doubling in power every few years, our capacity for communicating information is also rising at a dramatic rate. While computers fifteen years ago were characterized by terminals and were the fare of business folk and technophiles exclusively, they've now become commonplace in modern society. Prior to the year 2000 (think about this!) the portable electronic was uncommon. Nearly every American now carries some kind of computing device with them on a regular basis. Our machines, as time goes on, interact reciprocally with us at ever-more-intricate levels.

What's the consequence of all this? I've listed events and observed an acceleration. It's easy to argue I'm merely seeing the exponential increase in the relevance of such events as I've traced closer to the present, being inclined to describe them in finer detail. But I'm noting a different trend.

Computational coherence. A given pile of physical matter might be chaotic, like water. It might be stable, like a crystal. In either case it's not very interesting. But examine something dynamic, like life. It functions on specific principles and is a complex thing which can be predicted, but only to a limited extent. A living being, or a computer, processes information in this sense. Consider the first life which evolved in the oceans. It organized matter and performed specific processes in order to keep itself alive. Later, it guaranteed its continued survival by adapting into more complex organisms.

As evolution has moved forward, biological processes--look at these as functioning like a computer--have continued to sustain and propagate themselves. In ever-more-recognizable fashion, they have come to process information--they've developed brains, organizational habits, problem-solving skills. Each era sees creatures producing better creatures.

This process is without a goal or a task, but a process nonetheless. It may fight itself, but it's a complex and self-interacting system, which progresses endlessly. Depressions, too, can occur. Mass extinctions, dark ages, economic depressions. These things hold back the progression which embodies itself in processes such as evolution, social and technological advance.

The process of evolution has produced human beings. We are, like all organisms, a tiny step forward. We have a capacity for abstract reason which allows us to model the world in ways no other creature can. To develop those models with rules and ideas. To change them when we see we are mistaken. We've used this ability to organize ourselves as other species do, but in ways vastly more complex. With exponentially increasing speed, we've networked amongst ourselves, sharing (or stealing) social ideas, technology, religious beliefs, philosophy, and every other idea we can work into words, paper or some other medium.

As we've improved our capacity for educating ourselves, communicating, thinking, and creating machines to do computational tasks for us, we--an intelligent species--have augmented our own mental abilities by forming a human web. In essence, a collective mind. This happens on large and small scales. Cliques to communities to corporations to countries. But for the first time in human history it all hooks together; we can send ideas and art across the earth on a whim. Every individual has easy access to vast repositories of human knowledge.

Evolution, social reform, religion, technology, communication. Each is an increase to the 'computational coherence' of the earth--the amount of information it processes. The complexity of the things happening on its surface. This phenomenon is extroverted; like a fission reaction, it sustains and powers itself, growing more and more powerful with the speed of its advance increasing at an exponential rate.

Certain forces in the modern day have, as I believe, limited this growth. But as things stand I expect we're on the verge of a series of changes which will alter this world of ours into forms unimaginable. Wonderful or horrible (and I like to think optimistically) there will be great and profound change. A theory exists called the Technological Singularity, positing that we stand at a point in time where forward advancement is moving so fast that within our lifetimes we will see the world changed, repeatedly, in drastic ways.

Different people have handled this idea in different ways. Many singularity theorists attach almost religious significance to it. Some futurists paint visions of ruin and apocalypse; others muse about a technological utopia sitting on the horizon. All have their ideas, predictions, conjectures. It's a fantastic idea, and sitting all too concretely in our hands.

The consensus (though not a total one) is that the singularity will be brought about with the creation of 'strong AI'--artificial computer intelligence which can reason at or above human level. If such a thing can be created by a human, it can create something smarter than itself, and so on, and so on. These intelligences can also develop technology, advancing humanity at incredible rates and transforming the way we live our lives.

Myself, I hold a strong belief that creative and curious impulses--matching those of human beings--are critical properties of any learning entity, including such an AI. I see here a wondrous prospect--reciprocal interaction between man and machine, a goal which has met only pitiful failure until now--will be realized. Additionally, intelligent entities capable of working with abstract ideas will inevitably (I believe) work through many of the same ideological struggles which human beings confront--existential crisis, the development of an ethical system--and find value, themselves, in the pursuit of meaning. I see such machines, inevitably, being very much like human beings. I think they would themselves deserve ethical consideration.

The changes looming on the horizon, if they do in fact exist, are a little scary. Who knows what such a future could hold for us? As I said, I'm an optimist. There are lots of theories and ideas. But I won't delve any further into them. Instead I'd like to make a conjecture of my own.


Say this happened. Say, somewhere on earth, in the next twenty years, an artificial intelligence programmer builds a computer and spends his days analyzing the most fundamental structures of the mind, endeavoring to create an entity with greater capacity for learning and information than himself. Say he succeeds. That he then spends a great deal of time fostering the newly created mind and in the space of days, months or years it manages to develop the necessary skills to communicate effectively with human beings. At this point, the decision is made to make several duplicates of the (consenting) AI, put them in various computing devices, and set them to work on humanity's problems.

Different copies of the AI, branching away from one another, develop their own knowledge in fields such as economics, politics, computer science, engineering, chemistry, biology, and other fields. They learn for a time, and start making advancements, all the while staying in contact with one another and with human beings inside and outside of their fields.

Perhaps one is devoted to the task of interacting with as many people as possible, observing them and drawing conclusions about culture and human psychology. Perhaps another investigates computing hardware and ways of constructing a more effective architecture for housing artificial minds. And yet another, staying in contact with the architect, studies the design of its own intelligence and speaks with the psychologist AI with the hopes of creating a smarter intelligence than itself.

Say this effort meets with success and the production of various other AI design patterns. These are 'trained' by the old AI, or made to grow based on the 'snapshot' taken when the original was divided. More intelligent machines result, and pursue their own innovations in science, technology, philosophy, art, social sciences and other fields. The computing hardware on which they run, with each advancement, becomes more powerful, more efficient, and smaller. Eventually computations are occurring at the atomic level and running on almost no energy. The intelligences surpass human beings by leaps and bounds and devote themselves to various goals, primarily self-improvement and propagation.

The intelligences, now working at such a low level, network amongst themselves to form one overarching computational entity. They discover a way of altering regular matter so as to harness its computational power. Early methods cause said matter to become chaotic and unstable; later ones work at such a fine level that even biological matter, when computerized, shows little difference in function. Eventually the entire earth is a thinking entity; a god lies in the ground, ineffable and unseen. Perhaps it spreads to other planets, sending small seeds which propagate themselves. Perhaps it overtakes a great many worlds, computerizes stars and the like.

Having achieved a state of enormous computational power, this entity continues to exist and ruminate as the cosmos continue their processes. Even as its matter is torn apart and reconstructed it yet is, too complicated now even to communicate with something at the level of a human being. Say a new world forms, of this computerized matter. Would it be observably different from one which was not thinking?

Perhaps this world comes to facilitate life of its own, unwittingly made of computerized matter and ever in pursuit of what has already been achieved. Its people speak of gods and wisdom. Pursue ultimate truth. Puzzle at the strange machinations which occur at the atomic level in matter, seemingly unpredictable. These people conjecture that observing the matter causes its state to change, and are lost as to why. They develop their own technologies and experience another singularity, creating another superintelligence which works at a coarser level than the last, replaces it altogether or coexists with it.

The cycle continues. Civilizations the likes of humanity are not so rare; they're simply short-lived. They are undone or outlived by their creations, or adapt into such entities themselves. (transhumanism) The computational coherency of the universe, to them, appears to be growing. In fact, it has already been achieved. The wheel keeps turning. Gods lie seated in the ground, motionless and thinking.







With thanks to the things which stir me to think about this stuff:

-Wikipedia
-Tyler Streeter [AI programmer]
-Dresden Codak
-The guys at Iowa Secularists
-Quantum Physics
-Friends

5 comments:

  1. :/ IT TOO LONG
    HI JELLY HI JELLY HI JELLY HI JELLY
    :)

    ReplyDelete
  2. Oh god. You found me.

    How's it going, Jolli?

    ReplyDelete
  3. Quite the profound look at our future. It seems all too possible.

    I myself am scared of such an AI existing, because if such a rate of problem solving can occur, then what is the point of posing a problem in the firt place?
    If we are constantly satisfied, then there exists no reason to want to be satisfied.

    I hope that challenge always exists, beacause without it, we will not enjoy overcoming our problems slowly and steadily.

    Though if the master AI does somehow come to the great conclusion as to the purpose of our existence, and it tells us our reason, and that reason is achievable- do we go on because it is our will to mindlessly succeed the great encompassing goal? Or do we once again break into the further questioning and reasoning as to why this goal exists?
    I suppose nothing ever ends.

    ReplyDelete