The Big Freeze Tuesday, Jan 13 2009
brain and intelligence 7:22 pm
No, I’m not talking about Heat Death or another Ice Age. I’m talking about what could happen if mind uploading becomes universally or near-universally adopted and every mind is accelerated by a factor of several million or billion. Such an outcome seems inevitable if mind uploading is actually possible.
You may have heard that mind uploading means emulating human brains in computers, along with expansive landscapes to keep them entertained. Mainstream voices, or even WIRED-esque voices, may disparage such a possibility as a fantasy of techno-nerds, but if functionalism is true, as many contemporary philosophers and cognitive scientists believe, then mind uploading will eventually become the new rage, whether or not it looks cool from the perspective of a 2009 journalist. Easy to make fun of, hard to actually stop.
With a complete readout of the data contents and algorithms of the human mind, combined with molecular rod logics, or even molecular nanocomputing, it seems feasible that million, billion, or trillionfold speedups could become possible. We will experience things millions or billions of times faster, a phenomenon known as subjective speedup. The rate at which we experience the world is dictated by the operation speed of our neurons. It is not set eternally by some Platonic ideal, which many unthinkingly assume. The fact that a “moment” is defined as a few seconds and not a few nanoseconds or a few thousand years is a simple consequence of the operation speed of our brain.
All biological neurons operate at about 200 Hz (cycles per second). If you look at much phenomena in nature and at the microscale, this seems relatively slow, but there is a reason — it was obviously built for feasibility and reliability, not maximum performance. If there were a God, he would have done the right thing and built our minds to operate at the maximum possible speed with both reliability and high performance, but there isn’t, so our minds operate quite slowly. Guess it will be up to us to change that. One element will involve making neurons smaller, so that they can switch at terahertz speeds without frying outright.
Emulating human minds will be quite the technological challenge, but luckily, and I agree with Kurzweil on this one, sufficient computing power and scanning capabilities will be all we need. No Unified Theory of Mind (physics envy much?), philosophical enlightenment, or other such nebulosities will be necessary, though they could be helpful if available. Reductionist materialism will triumph with this one, much to the chagrin of many, but very quickly the old rivalries will be forgotten and the unlimited potential of the technology will become the prime focus.
The utter power that uploading will confer to our civilization is mind-boggling. From the new perspective of a world with such power, all the old human achievements will seem inconsequential. Plato, Socrates, Einstein, Mozart, all mediocre. If I can pick up a fistful of sand on the beach and turn it into a thinking mind in short order, intelligence will become as common as drops of rain in the world’s storms. All problems that can be solved by the application of intelligence will be. Truly an End of History, if you will.
It becomes hard to pretend to care excessively about present-day ethical quandries and thoughts about stem cells, life extension, nutritional supplements, gene therapy, and the like, when such tremendous potential rests in the near future. From a utilitarian point of view, if functionalism is true, then getting to uploading as quickly and smoothly as possible becomes an overriding concern. Conversely, uploading could become an ethical disaster without sufficient oversight, as sadists could keep and torture virtual nations in computers the size of a bowling ball. The easiest solution might be to simply eliminate the predisposition to such sadism in the human genetic code, and prevent all existing sadists from reproducing, but apparently the actions and beliefs of Nazi Germany have made any such talk about eugenics verboten.
If mind uploading is possible, the physical world may become boring due to the slow rate of its operation in comparison to the subjective acceleration. I say “may”, because such an outcome is not certain. Phenomena that occur on nanosecond or even femtosecond timescales, such as the reaction of pigments in the eye to incoming photons, might command some interest in the new humanity with its hyperfast thought rates. Uploading is not a “release” or an “escape” from the physical world, but a window to experiencing it the way it was meant to be experienced. If uploading is possible, then all civilizations would likely converge to it, as numerous pressures — not the least of which being the desire for personal longevity — will trump any conservative blowback. If aliens exist, they are likely either worms or uploads. Meat-based general intelligence is a short-lived phenomenon.
Despite all my praise of the possibility of uploading, I am doubtful that AI will be achieved by a direct copying of the human brain in software, as Ray Kurzweil believes. It seems more plausible that such an advanced technology will be developed by AIs whose data requirements are measured in the hundreds of gigabytes, not the hundreds of terabytes. Before we can mimic the human brain, we will be able to sketch out the general contours that underlie its ability to function at all. Just like how we can understand biomechanics and build prosthetics before we can copy a muscle.
Some transhumanists seem to ignore the “high future shock” possibilities of uploading and superintelligence, but doing so is impossible, as these “fantasies” pump through the very veins of the movement. Mild transhumanism, thoughts about reprogramming the body using biotechnology alone, are a “gateway drug” to thoughts about superintelligence and uploading, so the logical progression is quite unavoidable. What is burdensome, on occasion, is trying to hide such thoughts from journalists for the sake of avoiding coming off like a kook, but media coverage in recent years has proven that trying to hide such things is impossible. Thanks to growing mountains of popular media and imagination that is on our side, speaking of such topics openly and freely is only becoming easier and easier.
27 Responses to “The Big Freeze”
Leave a Reply
You must be logged in to post a comment.

January 13th, 2009 at 8:31 pm
“If there were a God, he would have done the right thing and built our minds to operate at the maximum possible speed with both reliability and high performance, but there isn’t, so our minds operate quite slowly.”
Why do you think God would have done that?
January 13th, 2009 at 9:52 pm
The mere connection between nazism and eugenics is NOT why eugenics is taboo. The nazi thing is only brought up at all when discussing eugenics because that kind of thinking is part of what made them nazis in the first place. Prevent all existing sadists from breeding? Are you f’ing serious?
The God remark, by the way, was very silly. It’s to the detriment of your writing and makes you come off as a resentful ex christian with an axe to grind when you sprinkle off topic little missives like that into your posts. It also made no sense.
January 13th, 2009 at 11:05 pm
@Steve: b/c God is supposed to be NICE and what you quoted is assumed to be a NICE thing
@Ben: Yes, he is f’ing serious.
Micheal, do please keep on remarking about God… and writing about uploading
January 13th, 2009 at 11:10 pm
@ Ryan. How should we stop sadists breeding? Maybe we could sterilize them, or lock them away. Perhaps we could do neither, but threaten them with both unless they obey us. All we need are the votes and you and I can live in paradise.
January 14th, 2009 at 5:23 am
Unfortunately, Michael, the notion of eugenics is once again gaining ground and is widely debated, although its ties to nazism do inform that debate still.
Your subsequent post about a “singleton” is apropos to this discussion.
I agree with Ryan, btw. Keep, talkin’, Michael. Just keep talkin’.
Be at peace.
January 14th, 2009 at 10:37 am
Michael, I really like this post, it summarizes nicely some conversations I’ve been having with my brother. We were pondering what civilizations would be like in a rapidly accelerating Kurzweilian world, and came to the same conclusion you did in this post. I would suggest that it can be taken even further. For a human based entity accelerated a billion times, conversing with another one mere meters away would be tiresome and limiting. Yet, since an uploaded entity would have unlimited life span, they could also have patience to wait for slow things to happen. Perhaps these uploaded denizens will create “models” of themselves for different timescales. The possibilities indeed boggle the mind!
January 14th, 2009 at 11:44 am
Unfortunately, this problem isn’t limited to deliberate sadism. A couple of weeks ago, there was a debate on OvercomingBias about a sketch of a possible post-uploading economy presented by Robin Hanson; see. e.g. this response to Robin Hanson’s original post on the topic.
Even without deliberate malice, certain economic attractors could easily become Existential Risks in the presence of uploading technology.
January 14th, 2009 at 2:42 pm
Ben, yes I am serious. Sadists that act out their fantasies are actually prevented from breeding in the US, as you can’t have children if you’re in jail and forbidden from having sex with visitors. So take that up with Congress and the President. According to your savior’s teachings, sadists should be killed anyway.
I would still be resentful towards Christianity, the largest religion in the world, even if I weren’t raised in a lax Christian household that went to church only occasionally. It’s just obscene that you people believe that I’ll be tortured in Hell for all eternity just for not believing in your make-believe Hebrew sky god that was invented by nomadic herders in the Bronze Age. It’s also obscene that Christianity dictates that all homosexuals, blasphemers, people that disobey their parents, etc., should be murdered, as the Old Testament said, and as Jesus confirmed the validity of in Mark 5:17. All those silly rules in Leviticus, here’s what Jesus had to say about them:
“Whosoever therefore shall break one of these least commandments, and shall teach men so, he shall be called the least in the kingdom of heaven: but whosoever shall do and teach them, the same shall be called great in the kingdom of heaven.”
How about this idea — Jesus was wrong, and some of his ideas were morally reprehensible, not to mention megalomaniacal, but they exert too much influence over the people of Earth today. That influence ought to be curtailed, or at the very least, openly mocked, as is my right to do so. If you don’t like it, leave.
January 14th, 2009 at 5:39 pm
Michael,
I’m sorry but I never indicated I was a christian. You merely assumed as much and went on the attack, ranting about Hell and Jesus and whatever else you’ve clearly spent some time ruminating over. For what it’s worth I’m agnostic, not that my personal beliefs are at all relevant. You’ve only confirmed my point. You have an axe to grind on this issue. Okay, fine. Grind away. But my constructive criticism still stands. Taking off topic anti-religious potshots in a totally unrelated post about mind uploading undermines your credibility and makes you look childish.
All the same, and against my better judgment, in defense of Christians the world over: they don’t all think you’re going to hell. Many of them don’t believe in hell, and certainly wouldn’t be making judgments about you or anyone else if they did. Many of them are also openly critical of numerous passages and ideas within the bible, engage with their religion in an active, thoughtful, scholarly way, and are capable on occasion of counting without using their fingers or drooling. …So on and so forth. Now, tell me if you can perceive the irony (not to mention the hypocrisy) of your closed mindedness in this context? Ignorant prejudice is ignorant prejudice, no matter where it’s directed. And you should know better.
January 14th, 2009 at 11:57 pm
Great post Michael; thank you. A couple of thoughts:
- Is it possible to question whether faster thought is equivalent to more intelligent thought? Its a bit orthogonal, but a hundred million bloggers doesn’t seem to raise the level of discourse over that of a thousand. Dumb people thinking faster just means more dumb ideas, not necessarily better ones. Smart people thinking faster and for longer may suggest better and broader ideas, but… I wonder if they are necessarily ‘better’ or just more of the same. I don’t think me thinking faster is going to improve the quality of my thought output, just the quantity. Perhaps my reasoning would be more well thought-out, or perhaps I’d simply travel down a greater number of the same blind alleys I’m bound to via my biases and simple lack of imagination.
- Subjective reality doesn’t exactly seem to be the fashion trend in terms of the greater populace. The ever increasing emergence of ‘reality’ tv etc speaks to a thirst for authenticity by a greater section of the populace than those say who are drawn to ’second life’ scenarios. So it may be that uploaded minds may be just as fascinated with meatspace as modern urbanites are with ‘back to basics’ ideas of living and interacting. Also, I may well just want to skip the boring bits; if I’m digital surely I can just set an interrupt for a few years of objective earth time if I get bored with the present. I -could- think my way through a million years of subjective time, but referring back to point #1, I may well feel simply like a hamster turning his wheel over and over with no transcendence in sight.
January 15th, 2009 at 2:50 pm
Instead of regarding the human mind as something that gets (”merely”) speeded up, we may get something of an explosion of mind plasticity.
Imagine a starting state of a new mind such as you in the future - to start out as a new citizen it starts out as an idealized human mind - basicly ycu, the reader, on the most lucid day of your life ever, with curiosity, happiness, raging ambition and optimism.
Then slowly envision a metamorphosis, as soon as you are ready (after days, months, years or several human livetimes) imagine slowly all features of the perfect human cognition trickling in, the mathematical clarity of an idiot-savant autist *with no frailty or drawback*, synaesthetia, perfection in every sense, talent, agility, memory, empathy, musical ability, physical as well as mental.
Then imagie the fully liberally applicable option to accelerate subjective thinking, effectively making events in your awareness move with the consistency of cold honey, but having full sensory acuity, ie the world is just a bright to see and sounds just as intelligible as in the normal speed.
Imagine however a full ability to suspend boredom measured with a full grasp of prioritizing, wisdom, prudence, dilligence and restraint. A frozen eternity might be spent watching droplets of rain hover motionless from a lover’s face in the full understanding what will be needed in the seconds beyond that to save him from a raging pack of hyena’s - with only a knitting needle as a weapon.
Imagine then the timely emergence, (after days, months, years or several human livetimes) all unique mental, neurological abilties of all animalo that ever lived on earth, the strongest muscle, the most resilient physiology, every keenest sense, resistance to heat, cold, hunger, having no need of sleep, happy to faste a month and then run a triathlon. All these abilities in one physique.
Then add the ability to create mental versions of the self, each fully eager or willing to be edited, commandedor dissoluted at a moments notice - dozens that can be set of tasks of art, philosophy, detective work, art, science, strategy, planning, romance, multitasking, or ewenmore contradictory states or obsessive deramgements, eac one thinking several times as fast as a normal humas, where the main mind thinks thousand times as fast, fully multitasked, parallel and indistractable.
Then imagine to be able to create inside your head all items ever created by humanity before 2009, in an imaginary space the size of initially a large office, populated by anything ranging from libraries, laboratories, serer farms, with equipment and software (windows!) tested and debugged for effectively decades. A version of WoW as it wuold evolve over decades.
Finally add exabit internet access.
The above, ALL of the above would and SHOULD ccur, logically BEFORE any singularity, since rough implications of ann these are all more or less transparant and predictable.
Hence, Mike, keeping an eye on the above, just how grotesque the implifations are of what we dare discuss here… would you mind spending a few sentences on how to meaningfully reply to critics of H+ when all this?
What do I say to someone like Dale, when he says we are all complete loons? I mean, it is so *hard* squaring off both paradigms!
January 17th, 2009 at 9:56 pm
@Khannea: That was beautiful, thank you… I maybe ‘get it’ a bit more now.
Makes me wonder how the minds I met in the prison system would behave with that level of enablement though. I wonder how -I- might have used that level of processing power when caught up in my crime-spree days. Abnormal or maladaptive psychology given free rein under those kinds of conditions… bit of a future shock moment.
January 17th, 2009 at 11:40 pm
Before Khannea’s vision took it to another level, the following yanked my mind into a state of extreme positive focused lucidity:
“The utter power that uploading will confer to our civilization is mind-boggling. From the new perspective of a world with such power, all the old human achievements will seem inconsequential. Plato, Socrates, Einstein, Mozart, all mediocre. If I can pick up a fistful of sand on the beach and turn it into a thinking mind in short order, intelligence will become as common as drops of rain in the world’s storms. All problems that can be solved by the application of intelligence will be. Truly an End of History, if you will.”
Why? Because it’s true.
Michael and Khannea, may you live to see it. Minds like yours deserve it.
January 17th, 2009 at 11:56 pm
…I mean, if this kind of stuff doesn’t blow your mind, what does? I love nothing better about transhumanism than these types of visionary future scenarios. The higher the Shock Level, the better. Someone should gather links to the best of them and post on Wikipedia. Eliezer Yudkowsky’s writings and Nick Bostroms ‘Letter from Utopia’ come to mind. Anything else?
(Post-Singularity scifi links, with uploading themes, the likes of Accelerando would be appreciated too…)
January 18th, 2009 at 4:45 am
Hi Michael,
I can see one thing that would attenuate the slow-down, which is a corresponding increase in complexity. For example, we humans have brains far more complex than that of a hummingbird, however most field tests have shown that perceptually humming birds perceive the physical universe slower than we do, in part because of their increased metabolism, etc.
The question is to what degree will SI’s use the resulting increased computation speed for complexity versus subjective experiential slowdown? It’s an interesting question I’ve pondered for a few years now. I believe it was Anders that once attempted to discover if there was limits to mind complexity, and concluded that the speed of light would prevent them from getting too big (i.e. jupiter brains) because of internal latency, and that past a certain size, this latency would result in a Jupiter brain resembling something much closer to a community of many minds, rather than a single one.
What are your thoughts on this?
January 19th, 2009 at 1:03 pm
Hey Paul,
My thoughts on this are that I agree with your general line of reasoning. I have no idea how SIs would apportion computation between speed and complexity. At a given point you hit that latency barrier, but it’s pretty huge. I only talk about acceleration because it’s easier to illustrate the power of uploading to the unfamiliar by talking about speed first and complexity afterward.
January 19th, 2009 at 8:40 pm
It is consciousness that causes brain activity, and not the other way around.
January 24th, 2009 at 9:28 pm
I’m not well read in transhumanist literature, so possibly those points have already been made a hundred times, but still I’ll ask.
About the unlimited potential: If you assume that AI probably will be first, maybe even necessary for developing that kind of technology, what advantages do you think might the “uploaded human” still hold?
If you have an AI that is worth its name, which would be the tasks an uploaded human mind could perform better?
From a point of view of effectivity, most of what we as humans are, lumps of flesh evolved to procreate in a prehistoric savnnah, would constitute computational flab for an AI.
Let’s assume the necessary technology you need for a human upload exists. It implies a detailed (at least functional) understanding of all relevant cognitive processes and their relation to the brain. Once you have that, nothing stops you from simulating every single human cognitve process in an AI, subsequently cutting away anything that evolution failed to throw out and improving it further. Which would make the AI better than the transhuman counterpart in every aspect.
Freely after Peter Watts: In a truly transhumanist future, there might be no more place for humans.
And eternal life? Well, as beautiful as those subjective descriptions of a freezing world in super-slomo may sound, they might be as misleading.
While I admit that you can label the “digital human” as fully human, the flesh and blood original will still die. He might live very very long by means of advanced technology. But at some point in life disaster is statistically bound to strike and will eradicate the physical existence of that human utterly. Since the individual experience of being that person is inseperably bound to the lump of human flesh that he is (unless a soul is assumed), those uploads will be nothing more than very sophisticated children.
I admit, those children will live forever. But they will perform worse than an equivalent AI. And their existence will not be of any profit to you personally, as far as the dream of an ongoing personal existence is concerned (apart from the narcistic one of having a superhuman mirror-image).
So far the skepticism.
January 25th, 2009 at 8:13 am
Rip, while such a result might be emotionally and spiritually satisfying, it seems doubtful.
Wolf, eventually a human upload would hold no advantage over AI. Having an “advantage” isn’t the point of life, though, it’s more about enjoying yourself. In the long term, the number of tasks that a human upload could perform better than a custom-built AI is zero, unless the upload has greater computational resources.
I should hope there’ll always be room for humans. Apparently there is room for plants and animals even though humans are better at most tasks. I’m slightly confused by the way you’re approaching this whole issue: are you saying that if AIs are better at performing certain tasks than humans, then humans don’t deserve to live, or something?
As for uploading, since I believe in causal functionalism, a high-resolution upload would be the same thing as me. I don’t believe that humans are inseparably bound to the flesh.
I’m not so super-obsessed with living forever, I just think it would be nice. I think that many transhumanists would agree.
Happy to bounce the ball back and forth on these questions.
January 25th, 2009 at 2:56 pm
You’re right on that one, holding an advantage is not the point of life. Yet we do most of the things we do because of a reason. What puzzles me are the reasons that are given, why someone would want to make an upload.
It probably won’t be for the benefit of society, since the problems a human could tangle would be better solved by an AI.
What remains are personal reasons.
What I’m trying to say in that domain, is only that the decision for an upload is more alike to the decision to have a kid, than to the decision to become immortal. For me it goes more along the lines of creating something interesting that will outlive oneself, than gaining any more tangible personal benefit.
I probably have a slight problem with the more extreme notion of functionalism, stating that a high resolution upload would be the same as me.
Sure, for a third person observer those two wouldn’t be discernible.
What I’m imagining as a thought experiment is the moment after the uploading procedure. Let’s say you wake up, the doctor congratulates you, everything was successful and there is a version of you, thinking and acting inside the machine. Apart from the fact that you now have a digital clone, nothing has changed in your perception of the world. Unless you assume something like a magical “split of personality” the moment your digital twin goes online or the transfer of a soul into the machine. And I don’t buy that (and you probably don’t either).
When relating stories about the frozen worlds and radically altered perceptual landscapes, one has to consider that those kinds of experiences won’t be ours. That it’s the way our digital children might see the world. But not us. At least not by the way of an upload.
February 18th, 2009 at 10:51 am
I’m a never-was christian with a big fat athiest axe to grind. I thought the “but there’s not” remark was awesome for one thing (I laughed my ass off, it fits so perfectly), but also the entire concept of the article is truly mind-expanding. You should read “the sky road” by ken mcleod, it’ll blow your mind. It touches on a whole lot of this sort of thing, including a wide range of themes of mind-uploading.
February 18th, 2009 at 11:57 pm
Interesting, but throwing in the God comment just made the article seem a bit unscientific.
February 19th, 2009 at 12:01 am
p.s:
Please post some scientific proof that conciousness arises from brain activity, because as far as I know it has never been proven.
In fact new studies on the ‘holo’ concept of the universe, indicates that the brain is just a ‘reciever’ rather than conciousness in itself.
February 19th, 2009 at 3:56 am
The God comment is philsophically quite relevant, you know.
Post some proof that consciousness *doesn’t* arise from brain activity. That hasn’t been proven.
Proof == consciousness happens when the brain develops
consciousness doesn’t happen when the brain dies,
Thus, brain = consciousness
February 19th, 2009 at 5:48 pm
Just for clarity’s sake: Continuity of personal consciousness is not achieved through the uploading of brain data. Right? In other words, my brain is uploaded into a computer; the computer brain becomes conscious, just as I would be–but I do not experience its consciousness, only my own. And when my individual brain stops functioning, so does my conscious awareness–for good. You can’t be resurrected and you can’t extend your life by uploading your brain. I can’t imagine how anyone could argue otherwise, since conscious experience itself (the so-called “qualia” of life) is not like objective information that can be quantified and stored as data, but is rather a subjective, emergent illusion arising from the complex interactions of brain functions. The only way to get around this is to admit to some kind of duality. Correct? Or am I missing something?
February 19th, 2009 at 6:12 pm
The whole Sadist thing seems a non-issue to me, just let them carry on their acts in a sandbox on simulations. It’s not a difficult ethical question, as the simulated bots don’t have to actually have consciousness themselves. Besides, in a completely VR world, you “respawn” if someone stabs you in the face, and you don’t feel any pain. What’s the problem?
The ethical questions I’m more concerned about are a bit more complicated. In the not-so-distant future, we’ll be able to design human beings from an assembly line, and with uploading a simple set of a bit. Who gets to decide these new being’s personalities? Why do they have the right to decide? With the current “system” we simply don’t have a choice, our mindspace design is mostly the one chance made us be born with.
We’ll also be able to alter the mindspace of human beings after they’re born. See Yudkowsky’s writings on “Drive Remapping”. Problemistic Sadists will simply have their impulses “mapped out” and replaced with something more productive (Until mind uploading is perfected and they won’t be a problem anymore, of course). Since we won’t have to eat, sleep, or have sex anymore, these would probably also be mapped to something more productive, like scientific research. But are these things we should just get rid of? Are things like the pleasures of eating something we should deny future generations the right to experience? Are we just going to throw away those fundamental parts of our history?
February 19th, 2009 at 11:16 pm
Casey, you are missing something, you should expect your consciousness to be uploaded to the computer as well. Why? Because the “complex interaction of brain functions” you describe would preserved directly across the transition.
To understand what I mean, imagine “uploading” your consciousness from your current brain to another biological brain by replacing your neurons one by one. Would you expect to lose your consciousness because your original brain had been compromised, thrown out, replaced? No. Your original brain is compromised often in just that way, as the individual molecules making up your neurons get replaced. Your consciousness stays.
What if each neuron were instead an electronic part that functioned the same as biological neurons but had silicon instead of protein components? Unless there is some special sauce for protein neurons only, you retain your consciousness. This is because of functionalism — every single quality of your mind, including consciousness, comes from the functions, the pattern of ongoing operations — NOT the specific substrate it is implemented on. You could replace every single neuron with “clown neurons” — little neurons with clowns nano-etched on to them — and as long as they did the same thing, you’d expect to retain your consciousness.
Everything is objective information that can be quantified and stores as data. Some qualities may be substrate dependent — for instance, silicon can’t have the material properties of carbon nanofoam — but it doesn’t seem like consciousness is one of them. Unless we have an extra-special reason to believe that consciousness is protein-dependent (and we don’t), it’s unlikely to be so.
Pen, people would make and torture bots with consciousness, so it would be an issue. I’m not sure why you just throw that possibility out the window because it seems inconvenient to consider.