Transhumanism, superintelligence and a global change

[Pantalimon]

I've been aware of this subject through science fiction writing and
films for a while but not in such a coherant way and not realising
that it was so far advanced.

I could spend a very long time explaining superintelligent AI's and
the sigularity that is predicted at the outside 30 years but as soon
as the first third of the next decade! However I'm not going to but I
will add some links to some very intresting and maybe more lucidly
written.

Reason to bother paying any intrest in this subject and following the
links include :-
1. Finding out you might never die
2. Ever
3. End of war, money, evils of the world... maybe
4. You could immerse yourself in a self created world of your choice
that would be just as real/more real than any lucid dream maybe even the equal of a high astral.
5. There are things you can do to help
6. It could happen as soon as the first 3rd of the next decade... 2012
anyone?

I listened to this programe yesterday
http://www.bbc.co.uk/radio4/science/lawsoflife.shtml (you can listen
via a replay link at the bottom). The progamme was very intresting
about moores law "the exponential growth of computing power" etc.

After listening I did some research into the issues raised and found
this article http://www.nickbostrom.com/ethics/ai.html by scientist
Nick Bostrom of Oxford uni. http://www.nickbostrom.com

http://www.acceleratingfuture.com/articles/whoaresingularityactivists.htm

The Singularity Institute http://singinst.org/ has been formed to
bring about the Seed AI.

http://www.extropy.org/index.htm

http://www.bbc.co.uk/radio4/science/smallworlds.shtml 3 programes to
listen to about nano technology.

The whole thing definately needs to be thought about closely
especially with a spiritual/astral eye. It fits well with my pantheistic
leanings as I don't need an individual soul to be attached to my flesh
body but I can see where its going to conflict with other systems of
belief and spirituality.... especially about the bring frozen dead
people back thing.

[JAW]

Ive read a lot about this subject as well and find it extremely interesting. Im not too for or against it, since in 30 years time Ill be 50+ anyway and will have lived a good life by then, so if anything strange happens either way, no huge loss.

The big thing I have with this not dying thing, is that it might interupt a natural process. If we are spirits having a human experience, then can we get trapped in human existance? If this stuff sounds a lot like astral realms, etc, why not just wait until youre dead to experience all of that, albeit at a higher more meaningful level, rather than essentially a glorified game? Who knows, perhaps once we reach this level of intelligence, it will be obvious that other states of existance exist. As they say a lot about this subject, however, its useless even thinking about what it would be like because we cant comprehend it, we arent smart enough, etc.

Cheers.

[Telos]

Just a thought,

It's possible that what we call astral experiences and these virtual reality experiences are the same thing.

Transhumanists speak of information as being immortal. We say that our souls are immortal. Virtual reality is the place where information directly interacts with other information dynamically. The astral reality is where our souls interact with one another directly.

Not dying is a natural process from both perspectives. There may not be any need to worry, Jaw.

[adaptandbeadept]

Transhumanism is probably my favorite hobby thought.  My favorite movie is Ghost in the Shell.  For those not familiar with the film: Its an animated movie about a female officer for the anti-terrorist government agency called section 9. The only organic parts of her body that remain are some pieces of her brain housed in a completely artificial body. Although she requires routine maintenance, the performance of her mechanical shell in combat situations is far superior to any normal human body. Throughout the course of the movie she doubts the existence of her rumored organic brain and human soul or "ghost." Its a fantastic movie that arouses skillfuly presented questions of humanity in a technologically dominated world. It has also been made into a tv series (GITS: Stand Alone Complex) that currently airs on Adult Swim, and the sequel (GITS: Innocence) which recently played in select theaters begs me to find it.
    So, having made my shameless plug I'll give you my thoughts on virtual realities in comparison to the astral and technological life extension. First, astral v. virtual. I've never had an astral experience although they interest me greatly, and hopefully I'll get mastering AP for Christmas! I'm a big time gamer in the sense that when I do have time to play a video game I try to play it cover to cover, backwards and forwards, inside and out, 100% completion. Depending on the game that can be a full time experience. I didn't play final fantasy X-2 simply because I saw my self spending an entire summer living two lives. One of those being a kid sleeping and eating, the other as a team of three girls saving a virtual world. But virtual experiences can do more than suck months from your life. Fatal Frame II is a game I spent my first semester in college playing surrounded by friends with the lights off on our big screen. You wander an inescapable village of ghosts in the woods of Japan with your twin sister who is very sensitive to spirits and gets possessed easily, leaving you to track her down and discover the secret, morbid rituals surrounding twins in the villages history. Your only weapon against the violent spirits is an antique camera with magic qualities. This leaves you with your eye stuck in a camera with literally tortured souls flying at your lens with the inention of stabbing at you from hell's heart. With a vibrating controller that simulates your character's senses and heartbeat, it caused more than one blood curdling scream. I love that game so much.
    So what am I getting at? Virtual experiences are valuable because they give us the ability to simulate experiences and emotions quickly and easily for human growth. By the same token, the have the ability to lure us away from our actual life experiences and to be abused the same as any drug. When we do in fact experience a technological singularity, there WILL BE people who get strung out on virtual experiences. When they return to the real world they will find themselves feeble and atrophied in the face of life and will crave their next virtual experience. But just as we have helpful and harmful drugs, there will be a useful and detrimental way to use virtual reality. The line will not always be clear.
    As for how similar it will be to the astral plane. I believe, even though I have never been there, that a distinct difference will always exist between the too and both will have something unique to add to a person's life. For one, a completely imagination influenced virtual world requires the assistance of a super computer that would have the ability to destroy or hyperevolve the human race. I see no reason why it would want to destroy us. We're much cheaper to make and require much less mainenance than an army of robots. bonking hollywood.....  Anyways, would this self conscious super computer with facilities housing different parts of its brain housed in every corner of the world and all linked together through the internet and through our own wirelessly nanotechnologically connected brains be able ::deep breath:: ::ahem:: to experience spirituality? Would such a powerful consciousness have a manifestation in the astral? It never falls asleep. Does it dream through our minds?  When our bodies fall asleep at night, is that part of the world the part of its brain that is in the astral? Thats what I lean toward. It would only be in the astral through our minds, making it permanently in and out of the astral planes. So it would probably be able to channel magic, spirituality, souls, higher astral intelligences and their messages into the virtual experiences. I think it would be more than a recreation of the astral, it would be a tap into it. Like astral light, in the beer sense not the photon sense.
    SOOOOO, as far as extending our lives with nanotechnology. I think that the implementation of nanomachines into our bodies to act as super macrophages is a great idea. I would love to have the cure for every disease even thinkable injected into my body simply by putting on the right shirt and having the nanomachines walk from it into my blood stream. The thing is when some of us evolve into hyper-humans we will need to establish hyper-ethics. Some of us will not and will become hyper-villains, but thats OK because there will be hyper-heroes to counterbalance. I personally would live a full life and then commit a very tasteful form of suicide because after having found astral pulse and a firm belief in the everlasting soul, death isnt as scary as it once was and life isn't as all important. Those so selfish as to live four thousand years at the expense of other humans and their livelihood may bump into me some day, and because I will know who they are and see the truth of their soul through the all seeing eye of a super computer that sees the secrets of everyone, I may just shuffle them off the mortal coil. In any case, a policy of who gives birth and population control, genetics everything would have to be in place. On that same note, our current laws and forms of government are just not gonna cut it when we experience a technological shift. But thats fine with me. I say, let those in power be metaphorically eviscerated by the inescapable force of change.
   I could go on about whatever new things pop into my head but I fell like maybe I've already gone on too many tangents. Thoughts? Thoughts?

[Ybom]

Wow, quick reply system. Ahem back on topic.

A&bA: I may be one of the lucky ones who have read over this topic on the web so much, that I could easily reorganize it in my mind pretty decently as I was reading through your post. I think you have a lot of the same ideas about it that I do!

Now, as for what I think, I believe scientists will very quickly find out that this virtual world will be impossible to create. I believe that there is a more than likely chance that our brain is just a simple circuit with a good buffer of memory that links to the hard drive/soul in another dimention. The good news is, I think the same scientists will then go and find out how to optimize our physical realm quite a bit.

The reason why I state this concept is that I believe there's going to be a major shift on the idea of birth control and mortality pretty soon. This is essentially the big problem I see with this whole virtual world concept. The smaller problem is current computers, meaning non-binary soul stuck in a binary world. I think that logic is making me queasy.

[RenaissanceMan]

This sort of advanced technology does what most other technology has done for humans -given us more control, more choices and the possibility of more finding more answers.
This time the choices are more profound - stay human or become something else, whether to die or not.
If it is wrong to live indefinitely physically, how do we find this out?  I say, let's try it and see.
Let's test the validity of astral/spiritual belief systems.
Personally, I would like to avoid dying until I knew for a fact that death wasn't the end.  I would like to hang around physically for a few thousand years anyway, to do everything that I want to do, as often as I want do do it.

[Ybom]

I agree with that logic. I know that the Afterlife is supposed to be better, but my life right now is very fragile comparatively. I will never have a chance to live this current life ever again after I die. I would like to cherish it; hold it dear to me, until I can transition from 'death' back to 'life' without much or any impedence.

There's no question in my mind that death isn't the end. I look forward to a celebration when I step over if I choose that path (assuming I have a choice then). However, my wish is to explore this realm as I currently am.

Is that so wrong?

[beavis]

Ybom

more than likely chance that our brain is just a simple circuit with a good buffer of memory that links to the hard drive/soul in another dimention

I think brain can do intelligent things without a spirit... if it didnt die from lack of energy. If brain is only a simple circuit, then artificial intelligence will instead try to do what spirits do now. Either way intelligence exists and can be copied in a computer.

[You]

Don't plan for it until it's here or you'll lose yourself and the privilege.

Do not dream too far ahead unless you plan to personally advance it.

[beavis]

I've been trying to build that software ("personally advance it") for a long time, but it looks like to exceed humans on a single computer, it will need a lot more cpu mhz (probably will have it in 25 years). Theres enough speed spread over the internet to do it today, and thats where it will live eventually.

The plan so far is to use many of the common kinds of artificial intelligence (and at least 1 new kind that I make) in the same program, many copies of each, and train a few of them to put the others together in intelligent ways.

Regardless of what kind of AI they are, they'll all have the same communication interface, which is mostly a list of decimal numbers (like 455.3337, 2.2, -7...) for input and an other list for output. What the numbers mean is unknown until the program starts and can be changed by a person or an AI after that. For text input, I could assign 1 number to each word: 5 for "I", 9 for "is", 3 for "alive"... (thousands of words) and (5, 9, 3) would mean "i is alive".

I'm almost finished with the part that lets the AIs write code (randomly at first, but without fatal errors) in a small programming language I made specificly for an AI to use. I can write parts of the AIs in that language, and they can try to rewrite those parts, or make new parts. Its going to have at least a few kinds of neural networks and some genetic/evolution types of AI, and I've been thinking about adding somebody else's AI that can see basic context info from sentences.

[Adun]

And what's your plan to make the AI self-aware?

[Blackstream]

It's about time a programmer actually steps in here.  AI is no where near sentience people.  They thought that by the year 2001 we'd have HAL-esque AIs but we were nowhere close then either.  The problem is not that we don't have enough mhz or ram.  If that's all it was, then we'd have sentient AIs already (there are AI scientists with much more resources than you Beavis that have probably already tried what you are trying).  The real problem with creating a sentient AI is that we are all missing a fundamental piece (or pieces) of the picture.  

I took an AI class and we looked into various techniques for learning and adapating and storing memory and the such.  It became quickly apparent that while you can make a computer specialize very well in one thing (chess for example or maybe vacuuming off the floor), making a computer that could end up like a human is something we have no idea how to do.  The closest we've gotten so far right now is Neural Nets, but while they have done some impressive things, they still don't get AI to where it needs to be.

Bottom line is, we have no idea how to make a computer learn and evolve and become sentient, partially because we don't know how we ourselves learn and evolve and are sentient.    And that's the grim truth of things.  

That singularity website is all a bunch of hooha that base their research off of some kind of "event" in the future that will magically give us the ability to create better AIs.  Right.  The first article link doesn't mean anything either.  The article with a list of so-called laws (which didn't mean anything and didn't have anything to do with AI) also doesn't promote the cause of AI.

Randomly generating code might work if you let the thing run for a few million years, but you probably should give it a bit more direction.

End of story, if you guys wanna make an AI, you need to figure out what we are all missing.

[RenaissanceMan]

I think that idea that Kurzweil mentions on his website, about reverse engineering the human brain, should give us significant insights into AI.

[Blackstream]

I don't think that reverse engineering the brain is possible, but that's a bit out of my realm.  It's one thing to say, yeah I see stuff going from here to here and I see activity here, but actually reverse engineering it and getting the "code" behind it all would either require some crazy science we don't have, or the public doesn't know about.

[karnautrahl]

"there are 1billion synapses in an area the size of a grain of sand in our brains"*. The sequences in which they work together in combination may be part of the overall picture in how we think on a biological level. If we can duplicate a conscious and preconscious way of absorbing information holistically and coherently and duplicate multiple internal information integration processes including emotional information, we might be able to consider creating an AI with some kind of sentience. But that looks really hard because we don't really know how this is done. We can point to structures firing when we learn/read or do anything at this time but more or less that's all.

*The Human Mind, by Robert Winston is the source of this fact I use LOL.

[Ybom]

Why try to make a human? I think that goal is a bit too far fetched. Why not try to make a worm, or an ant first?

[Blackstream]

Personally, I think emotions are really just a combination of thoughts and imagings of sensory feelings (such as the feeling of anger would bring up violent thoughts along with higher blood ratem, ect) that conincide together with things that affect us in some way (such as our sense of hunger or our sense of needing to belong or pain).

In other words, give the computer desires and goals and needs, give the computer thoughts, and give the computer a complex brain, and theoretically it should develop emotions.

Whether or not that'd happen I can't say.  Without the same needs as humans, it probably wouldn't develop the same emotions, but there probably would be some similarities.

By the way, I think I might be wrong about the thing about processing power, but there is no real way to test it yet.  When we get a few more leaps in storage power, and we are able to make a neural net the size of the human brain, we can see how that works and if there's anything we are missing.

[beavis]

And what's your plan to make the AI self-aware?

Its made of lots of little software tools (like add, minimum, show text). There are low-level things about those tools that nobody needs to know just to use them. At the high level, its "self" is a tangled equation that can view or change any part of itsself (or other equations). It doesnt have to interpret anything about what itsself is good at or is meant for, but it can be aware of every part of itsself as a literal group of code. In that small way, its already self-aware.

The problem is not that we don't have enough mhz or ram.

Thats one of the problems. My programming language is about 100 times slower than a normal. Ability to modify itsself at any time but never crash costs a lot.

That singularity website is all a bunch of hooha that base their research off of some kind of "event" in the future that will magically give us the ability to create better AIs.

No. They said the event occurs AFTER smarter-than-human AI is created. Humans can make better AI than existing AI, so a smarter-than-human AI can do at least that.

Randomly generating code might work if you let the thing run for a few million years, but you probably should give it a bit more direction.

Mine will be random at first, but code can always have the ability to make other code, so it should non-randomly determine the direction of its evolution when it gets smarter.

[Leannain]

can it increase my IQ to the likes of the alien races know as greys and alikes?

[CaCoDeMoN]

Beavis, could you post more info on your project?(like in which programming language you are creating it)

Thats one of the problems. My programming language is about 100 times slower than a normal. Ability to modify itsself at any time but never crash costs a lot.

Maybe dynamic recompilation would be a solution to this problem?

[beavis]

Compiling during execution would be faster if it usually made large changes. It costs a lot just to start the compiler, and I havent been able to find a way to do it without using files, which are the slowest of all. It often changes just 1 small thing at a time and executes it right after that. For that, my language is faster than most. I've seen the java binary format (my lang is made with java), and dont want to make my own compiler for it.

[CaCoDeMoN]

Java? I think that it would be best done with c/c++, you could optimize it to run only on Intel x86 architecture. The program in that language designed by your own would be stored only as a table in RAM, and compilation would look like this:
You could write your own compiler that would translate the code directly into assembler(it's not as hard as it may seem, you can create examples of such translation using c++ compiler). The "compiler" code would be always running, and it would not recompile entire code, only part being modified. Of course compiler's output would be not text assembler code, but x86 code that could be dumped directly into memory and called like a normal function. Of course it would not directly call other functions(I think it's too hard to do), only indirectly by modifying some area of memory. I think that it would be extremely fast then(nearly as fast as assembler code). The code of AI would be saved only on exiting program. I would love to see the source code of your program!

[beavis]

Each type of computer, even those as close as pentium3 and pentium4, have a different assembly language. Eventually the assembly language and computer will become obsolete, and my program with it, unless its updated. As a java program, its code will work in any computer for decades.

Java is not nearly as slow as it used to be. Just before some part of the java program starts, it repeatedly compiles itsself to a lower language, changes code to make it faster, and eventually gets to assembly. Computers beat people 99% of the time, at writing assembly code that does the same thing as code in a higher language.

Just now I measured a simple java program running at 70% the speed of assembly. Most of it isnt that fast, but it will be years later (even with the same speed computer). I know the fastest way to do it in assembly and count by that. Its an empty loop, which should have 3 instructions per cycle. That java test program counted to 500 million once a second. Does that sound slow? Java can optimize to VERY fast code, but limited by the complexity of the code. It gets smarter every year.

[CaCoDeMoN]

Each type of computer, even those as close as pentium3 and pentium4, have a different assembly language. Eventually the assembly language and computer will become obsolete, and my program with it, unless its updated. As a java program, its code will work in any computer for decades.

You are wrong, they are not as different as you think. Some commands are added, but code from older processors will have no problems with working on faster ones(like "Ancient Art of War" that was written for 286 or slower computers can run on Athlon 2000 without any problems). Assembly code will not be big part of a program, soconverting it to other architecture will take not more than one hour of work.

[beavis]

Ok I dont know about their specific hardwares, but I was right they have some commands added. I dont know if they're completely backwards-compatible, but eventually (on a newer computer) I'm sure any program that depends on their assembly wont run directly on the hardware. Many old programs only run today because it runs inside a software copy of the old hardware (emulate), and is a lot slower.

Assembly code will not be big part of a program, soconverting it to other architecture will take not more than one hour of work.

Computers never use anything but (the 1s and 0s of) assembly code. All the easily-readable high level code like java, C++, Lisp, etc... is automatically converted to assembly some time before its run. To convert a program from one kind of assembly to an other, depending on how different they are, could take an hour or months, depending on how much of the code they can get the computer to write for them.