questioneer

This is what I look like

This is the weblog of one Fredrik Bränström.

Topics covered include minds, science, future, philosophy, programming & graphic design.

Me elsewhere:

We're Headed Inwards

Homo Sapiens is getting to know itself a whole lot better lately, and the accumulation of cultural knowhow is growing into a more powerful phenomenon than this planet has ever known before. That much is obvious if you look at the world today. We would have big trouble surviving without technology. This change will shift to a higher gear once we start making real headway in reverse-engineering the brain — the hub of our interconnectedness.

The rise of massive multiplayer online roleplaying games and a growing market for accompanying non-physical merchandise point toward a coming period of new creative self-expressions and experimentation in virtual worlds. At the same time, we’re decoding the human genome, finding out exactly how gene expressions work in our real bodies.

Soon enough, regardless of whether we’re mostly real or emulated, we’ll start (re)designing our bodies. When this starts to happen in the real world, we might even be removing unattractive personality traits, reconfiguring our own minds. All while getting more connected. This will intensify our psychosocial complexities.

Increasing machine intelligence & connectivity will also result in more unimpressive job titles taken over by machines. They’ll compete with humans in the job market, and we’ll adapt and get smarter, do more high-level administration and lead the way, but eventually machines will become more generally & deeply intelligent and able to accomplish very complex tasks, growing in competences entirely unrestrained by messy and inefficient DNA. Hopefully by this time they will have learnt everything we (think we) have to teach about what’s right and wrong for them to continue along our trajectory of ethical thinking.

The algorithms that my brain uses to understand the world can be scaled out, sped up and improved upon. Evolution is very, very stupid. It is random variations sifted through by natural or artificial selection and so it has no foresight; it cant make engineering decisions of large imaginary scope or progress towards anything other than the survival of that which replicates. It layers kludge upon kludge. We can do much, much better. The difference in quality of our R&D work compared to blind evolution is immense.

Our smarts have enabled language, which has given us a digitally transmitted culture, which in effect has exploded our knowledge and allowed us to accumulate practical intelligence unfathomable by a human from only a few thousand years ago. Fire, the simplest weaponry & tools, the wheel, clothes, all this is technology, we’ve had it as long as we’ve been human beings, but tech amplifies itself: with every invention it frees up resources to deal with remaining problems. When it starts to not just augment, but copy & improve upon human intelligence, we will have put into motion a feedback loop more disruptive than anything else in the history of life, causing weirdness and levels of organization truly novel to this 14 billion year old universe.

In the coming decades we will of course very much appreciate advanced robotics & AI helping with more of the dreary administration of our existence; leaving us with more philosophizing to do, jumping in the pool, making exquisite food, overseeing & steering the course of progress, traveling around the world, falling in love, eating lots of chocolate, having all kinds of mysterious & beautiful experiences available to our minds.

But that won’t do. The human psyche is like a timebomb. We are in every moment driven forward in thought & action by hedonic states toward the goals inscribed in the digital code of our genes. Having grown smart enough to adapt and automate our environment, we will direct that inventiveness inward and strike the match that ignites the grand journey of exploration of physically possible psychologies different from the one we started with.

Sciences concerned with the brain will start teasing out an increasing number of safe shortcuts we can take to maximize the duration & scope of positive emotions and explore new forms of pleasure. This will gradually become more socially acceptable as we lose the current negative images associated with addiction to psychotropic drugs. Eventually all external preoccupations will be thrown to the wayside, recognized as superfluous, along with all other “meaningful” attributes for interacting with the world that we now possess or desire. We will prune and refine our minds in a kind of short-circuiting of our current behavioral reinforcement loops, and the sole moral imperative of our descendants will by the law of the Hedonistic Imperative be the relentless exploration of all possible positive mental states in mind design space, as well as the wiping out of involuntary suffering.

I suspect that in this scenario, there would just not be any point in having more than one mind around. Why have more than one perfect being in paradise? Except for the purpose of redundancy; backup copies to be on the safe side.

We’re making slow progress on the question of what constitutes human consciousness. The question of whether there will ultimately be one or many minds might become a non-issue as neuroscience progresses further and we discover new phenomena related to the phenomenon of consciousness, that are also allowed by the natural laws of this universe. Human consciousness is probably just one flavor in a vast candy-shop. The lines between individual and society will likely be increasingly blurred.

One of the problems I can see with this vision is that this unified hedonic entity has to be evolutionarily stable. Its survival will literally mean everything. Our human feelings are inherently inadequate to describe the importance of this. One could imagine a solution to this problem in that the entity could surround itself with a layer of reactive, superintelligent robotics to permanently serve to protect the inner core of hyper-sublime experience.

Considering it would probably imply that we condemn part of all the matter under our command to be less happy than the other part, would this be an ethically sound decision? It would be equivalent to slavery, wouldn’t it? How would the moral equation look like here? Is this unavoidable?

I suppose one way to escape the problem is to simply usurp as much of the universe as possible into our own being, so that there are no imaginable threats left to the continued exploration of pleasure, with the exception of the heat death of the universe.

I’m fairly convinced this is actually not as arrogant as it sounds. Especially as we’ll carefully study any complex phenomena we hit upon in our cosmic expansion to ensure we’re not encroaching on anything valuable. Yet regardless of what it might deem of value: against this vastly intelligent cosmic might of our descent, resistance would most certainly be futile.

Of course, we would have the responsibility to spread outwards as fast as possible anyhow; to abolish biological suffering taking place on possible (and probable) life-harboring exoplanets.

There are ways in which we could fuck all this up. More on that exciting prospect later.

Comments

  1. David Pearce says:

    I agree about the need for cosmic rescue missions if the Rare Earth hypothesis turns out to be false and there are other suffering sentients in our Hubble volume. But merging all consciousness into one is tricky. I can't see any technical objection a Jupiter-sized brain; but a Jupiter-sized unitary conciousness is another matter. How the human mind/brain solves the binding problem and creates the unity of perception is currently unknown. I suspect it's a manifestation of quantum coherence; but if so, presumably there is a maxiumum size above which a unitary experiential field is impossible because of thermally-induced decoherence.

  2. Fredrik Bränström says:

    I'm less certain about the unitary consciousness proposition than the rest. The problem of consciousness might be solved for us once we get human-level AI or full-scale human brain emulation, or we might need to solve it before we can construct a human-level AI or get successful brain em. I'm just thinking it's unnecessary duplication to have multiple consciousnesses in paradise, no matter how large one mind-unit can be made. They would just be making the same kinds of experiences at different points in space. But I really need to think and read more about this.

    Could you elaborate on the quantum coherence theory? Maybe you can give a link to a writeup on that?

  3. David Pearce says:

    See perhaps http://consc.net/papers/moving.html#4.4 - especially Seager, W. 1995. Consciousness, information, and panpsychism. Journal of Consciousness Studies 2:272-88.

    If unitary consciousness is indeed a form of macroscopic quantum coherence, then it becomes harder to see how human minds could be scanned, digitized and "uploaded" to a classical digital computer - since their properties depend on the fundamental quantum substrate of the world. This (hypothetical) dependence is one reason I'm still much more cautious than most transhumanists about the prospects of whole-brain emulation and indeed an imminent Singularity.

    But for a critique of macroscopic quantum coherence, perhaps see physicist Max Tegmark's "Importance of quantum coherence in brain processes," [2000] Physical Reviews E, 61, pp. 4194-4206. Tegmark calculates that quantum states in microtubules could survive for only 10[ superscript]-13 seconds. Note that while quantum coherence might explain the unity of perception, this only works if "mind-dust" or fields of microqualia are the stuff of the world [see Galen Strawson's "Consciousness and Its Place in Nature: Does Physicalism Entail Panpsychism?" (2006) Needless to say, most working scientists find this hard to swallow.

  4. Fredrik Bränström says:

    Thanks!

    It almost sounds like something Philip Pullman would pull out of his ass. :D I'm intrigued, had not heard of this specific idea before, only similar, sort of fuzzily articulated ones. I will take a look at those papers but am skeptical. If quantum coherence turns out be key though, wouldn't it just be a question of waiting for when we can build circuits that small? And similarly if we need to use specific chemical elements we could just use carbon etc instead of silicon... Adapt to what's physically possible. It wouldn't necessarily mean we can't be conscious in any other way than we are now: a human brain is very likely far from the only construct that can make use of the tricks of consciousness.

  5. Ian Burleson says:

    Have you both forgotten about quantum computing? It will be founded in quantum logic, which will accomodate biological consciousness' necessary marriage to quantum (de)coherence. And its quantum logic will be the essential feature of a greater-than-human intelligence, as it will allow it to think in terms of reality we humans simply cannot grok, by the very nature of being a product of evolution in a macro world dominated by Newtonian physics (or so it would appear to the untrained eye).

    So yes a human mind uploaded onto a classic, digital computer is inpossble. However, on a quantum computer, it seems to be more than possible, if not intuitively feasible. Perhaps I'm being too optimistic, but if the argument against uploading is the problem of replicating quantum behavior in a computer, then quantum computing is the cure.

  6. Fredrik Bränström says:

    @Ian: "So yes a human mind uploaded onto a classic, digital computer is inpossble."

    If the mind demands quantum effects to work. I'm not convinced we do. I think the organization that matters probably lies mostly in the slightly higher levels. It follows that I don't think quantum processing is fundamental to AGI although I do think it can very greatly extend its capabilities.

  7. Fredrik Bränström says:

    And we're probably assuming in some naive conceptual blur that once you put the word quantum in front of some other word you're basically talking about the same phenomena. But quantum logic and the possible mind-dust/nanoqualia-fields of Galen Strawson do sound like different concepts.

  8. Ian Burleson says:

    I simply meant in terms of simulating the biological substrate of gray matter, using a quantum computer will allow for a full simulation because it doesn't break down when simulating all the way down to electrons in the synaptic clefts. A digital computer, however, cannot simulate quantum phenomena accurately, or efficiently, and so to try to replicate the structure of the brain, atom by atom, shouldn't be possible or at least not feasible using a digitized computation system. I'm not suggesting quantum computing will fix all the problems of uploading, but it does, by its very nature, seem that it will help a great deal in solving the problems of quantum effects in the brain, which I think may (or indeed may not) be pertinent. More research will tell. But I think any accurate portrayal of a human mind will probably need a full simulation of said mind in a system that allows for <i>all</i> physical possibilities to be computed, since that is how it works in reality. If we don't go full simulation (to whatever extent may be necessary), I'm convinced we will never have a fully operating mind.

    I also haven't had a chance to read the links David has provided, so do forgive me for not addressing those arguments specifically. I'm simply speaking in generalities of the underlying concepts (possibly) at work.

  9. godthepantheist at gmail says:

    Fredrik, you might seriously want to take a look at this:

    http://rapidshare.com/files/314768041/Kolak__Daniel_-_I_am_You.zip

  10. Fredrik Bränström says:

    Thanks, pantheist. And thanks for supplying me with the entire book. Sounds like it might be an interesting read. ( So, it's I Am You — The Metaphysical Foundations for Global Ethics, by Croatian-American philosopher Daniel Kolak ).

  11. Dustin Juliano says:

    Fredrik, I have to agree with you that consciousness should not be necessarily limited to a cerebellum. To argue against quantum effects governing consciousness, this may be true for biological organisms, but that probability does not specify causality for all consciousness.

    There needs to be a philosophy written that segregates levels of abstraction from each other. Reductionism is great for following the line of determinist causality between lower and higher order systems. It is not, however, great for understanding what those systems are. As a result of there being no unifying philosophy that brings together the separation of abstractions with the causal relations that create them, we are stuck with narrow notions of reality and logic, like reductionism.

    I feel that the physical stuff doesn't matter as much as its arrangement, both on the tiny and the large scale. The contrast between existence and non-existence is what allows for any intelligent pattern to be known or emerge. What we know of as matter and energy are just substrates. If alternative theories prove true in the decades and centuries ahead, and we find ourselves on a brane, strait out of m-theory, or some other experimental hypothesis in physics proved true, what we know as quantum physics in those realms will be forever changed and made only local to our reality. To assume consciousness would not, could not, exist in such alternative physicality is the most heinous of assumptions.

    We are seething in bias, as human beings, and the cure is skepticism. We must question the universe, but most especially, we must question ourselves.

  12. Joni Hanski says:

    @Bränström: "I'm just thinking it's unnecessary duplication to have multiple consciousnesses in paradise, no matter how large one mind-unit can be made. They would just be making the same kinds of experiences at different points in space. But I really need to think and read more about this."

    Two person feeling good is better than one person feeling good, so based on utilitarism, plurality of minds in paradise would be preferable to a single happy mind.

    If this wasn't the case, and two people being in identical, happy state of mind wouldn't be better than one person in such a state, then we could reverse this. Let's assume an infinite universe where all possible mind states exist in plurality*. Now it doesn't matter if I torture you or if I give you a bliss, since both states of mind already exist somewhere in this infinite space. Since in any case the result would be redundant copy of the mind state that already exists somewhere, any choice we make becomes meaningless, or, for the very least, it's really, really hard to argue why we should care about anything at all.

    Thus, two people in paradise has to be better than just one.

    *World could very well be like this, but even if it wasn't, this thought experiment still works, simply because infinity of the world shouldn't affect our decisions like that. I read this thought experiment from Nick Bostrom's paper http://www.nickbostrom.com/papers/experience.pdf

  13. Fredrik Bränström says:

    Joni, that's a powerful counterargument... it seems. I will certainly continue letting my mind wander into this. And read Bostrom! Thank you for the food for thought.

    Dustin, ... woah. :)

  14. Alex Vance says:

    Fascinating post, and something that I've thought about frequently. Ever since I encountered Isaac Assimov's idea of a culminating universal consciousness (http://en.wikipedia.org/wiki/Gaia_%28Foundation_universe%29#Galaxia) I've been sure of its inevitability. The specific pros of Galaxia are that it is composed of hierarchies of individuals, all of which are fluidly united within the same mind.

    One point about evolution, which is not mutually exclusive with what you wrote: though Darwinian biological evolution is cruel, blind, and inefficient, natural selection as a design tool is incredibly powerful, and will in fact probably factor hugely in engineering the upcoming intelligence boom. If you haven't already, I highly recommend you read about the computer scientist who created a hyperefficient computer chip by inventing a method to weed out the worst and mate the best over thousands of generations (http://www.liveleak.com/view?i=885_1244368713).

  15. Jonatas says:

    Evolution has empirical testing, this may be lacking to our engineers for ethical reasons... but we can learn from it, at least in the initial stages of changing, before we can simulate how completely different DNA would manifest itself in a person... and then eventually we can just abandon the framework of DNA and redesign things another way.

    "I suspect that in this scenario, there would just not be any point in having more than one mind around. Why have more than one perfect being in paradise? Except for the purpose of redundancy; backup copies to be on the safe side."

    While having thought the same a while ago, I now am in doubt about this, or maybe I even oppose it, on the grounds that for utilitarianism quantity may be important, not just quality or average. My reasoning goes something like this: in an empty universe, having a conscious being experience positive feelings is better than not having. Does this fact change by there already being another? I'm not sure, but I'm tending to think that more could be better, even though terribly less efficient.

    Bostrom's argument about an infinite universe making any difference is very bad in my opinion. It makes no difference in quantitative terms or in a local level, or in respect to class action, even if we consider mean utility as the relevant aspect of utilitarianism.

    Anyway, if we're talking about the future, maybe people with intelligence that surpasses what is possible today could be able to figure out this sort of problem better than we can.

    "Considering it would probably imply that we condemn part of all the matter under our command to be less happy than the other part, would this be an ethically sound decision?"

    I think isolated "universes" of lower utility constitute a major ethical problem. However in my opinion intelligent robots need not be conscious... it is optional, not necessary. Consciousness seems to need a whole complex software running exclusively for it.

    That said, beautiful text, very well informed and good. I should pass this for friends to read.