?

Log in

No account? Create an account

Previous Entry | Next Entry

Take-off: Hard or Smooth?

Poll #691408 Singularity: Hard or Smooth Take-off?

Do you think we're in for a hard or a smooth take-off?

Hard
14(25.0%)
Smooth
18(32.1%)
No idea.
13(23.2%)
I remain pretty skeptical about this whole Singularity business.
11(19.6%)

What do you hope will happen?

Hard take-off starting RIGHT NOW!
7(12.5%)
Smooth take-off that gives everybody enough time to get their heads around the transistion before it happens.
8(14.3%)
Any Singularity will come as a hard take-off to some and a smooth take off to others.
28(50.0%)
Whatever happens happens.
13(23.2%)

Should John Q. Public, age 30, fund his IRA?

No point, the Singularity will render his efforts moot long before he reaches "retirement age."
14(25.0%)
Yes. You never know.
42(75.0%)

Comments

( 12 comments — Leave a comment )
ankh_f_n_khonsu
Mar. 16th, 2006 04:00 am (UTC)
I don't think a "smooth takeoff", given the rampant ignorance and endemic unsustainability, appears possible at all.

I think the people who think that a smooth takeoff can occur, given the present state of affairs, don't really know what has been happening around them.

As for the future finances, I'd say live each day as your last, but that doesn't give carte blanche approval for fiscal irresponsibility.

Namaste.
mungojelly
Mar. 17th, 2006 12:17 am (UTC)
Smooth vs Hard isn't about whether the Singularity is going to be a difficult transition for the human species-- I think just about everyone can agree that it's going to be "hard" in that sense. The question is how quickly we're going to transition.

What makes the Singularity so really fantastically singular, is the possibility of recursively improving intelligence. That is, once you have machine intelligence, one of the tasks it can be applied to is improving machine intelligence. The most serious hard-take-off singularitarians imagine a scenario like this: One day, an intelligent program is written which is capable of intelligently reprogramming itself. The start button is pushed, and it hits the ground running. An hour later, it's figured out how to alter its design to make itself twice as smart as it started. Half an hour after that, it does it again. Bang, the world ends.

I think that those scenarios are just as ridiculous as thinking that we won't have human-like AI for another thousand years. My opinion is that while the Singularity is going very very quickly, it's still fundamentally a social process, which is going to flow from the decisions made by human socities over time and not from any particular computer program.

I believe that augmented human intelligences are likely to be superior to independent AIs for the forseeable future. Not because it's impossible to make an AI which is more intelligent than a human-- no, I'm talking about augmented intelligences where the AI part is millions of times smarter in every way than a naked human. But human beings are essential centers of narrative & economic gravity in this world. Machine intelligence will therefore tend to cluster around humans to a much greater extent than it forms into its own separate structures.

(continued)
kmo
Mar. 18th, 2006 03:35 pm (UTC)
hard = fast ; hard = difficult
Smooth vs Hard isn't about whether the Singularity is going to be a difficult transition for the human species-- I think just about everyone can agree that it's going to be "hard" in that sense. The question is how quickly we're going to transition.

When I put the poll together, I was indeed thinking in terms of how quickly the transition will happen, but I didn’t provide an exact definition. I thought that most of the people responding to the poll would have encountered the hard/smooth take-off concept before and would know that it related to the swiftness of the transition rather than the psychological/cultural level of challenge the Singularity would provide, but I think it makes sense to equate hard and smooth with hardship and ease, and I’m interested in what people have to say in either case.

What makes the Singularity so really fantastically singular, is the possibility of recursively improving intelligence. That is, once you have machine intelligence, one of the tasks it can be applied to is improving machine intelligence. The most serious hard-take-off singularitarians imagine a scenario like this: One day, an intelligent program is written which is capable of intelligently reprogramming itself. The start button is pushed, and it hits the ground running. An hour later, it's figured out how to alter its design to make itself twice as smart as it started. Half an hour after that, it does it again. Bang, the world ends.

For the hard take-off to result from a recursively self-improving seed AI that AI needs to have an exponentially increasing computational substrate to expand into. For that to happen, I think that AI would need to control a very mature form of molecular nanotechnology (the kind that involves self-replicating nano-assemblers doing their thing outside of the laboratory or factory). An AI developed by a corporation may have the means to take control of chip manufacturing plants, but that will not provide the necessary substrate quickly enough for a hard-take off.

On the other hand, a bio-tech singularity seems more likely to produce a hard take-off a la Blood Music. I hope that we have a bio-tech singularity before we have an AI-based nano-tech singularity because the bio-tech version seems more dependent on enhancing human intelligence and doing it in a way that changes us profoundly, but not as profoundly or as quickly as a strictly AI. I think the bio-tech route will more likely produce a hard take-off because GM intelligent cells can increase their cognitive substrate in the exponential manner needed for hard take-off by mitosis.

I believe that augmented human intelligences are likely to be superior to independent AIs for the forseeable future.

I agree, and I see that as a good thing.

...human beings are essential centers of narrative & economic gravity in this world. Machine intelligence will therefore tend to cluster around humans to a much greater extent than it forms into its own separate structures.

I would say, "Human beings serve as essential centers of narrative, and persons provide essential centers of economic gravity." Of course, legally speaking, humans share the classification as persons, able to own property and control money and other resources, with corporations. Once S(uperhuman)AI comes on the scene, SAI-controlled corporations will wipe the deck with human-controlled companies, and quite soon the bulk of the globe's financial resources and economic activity will occur outside the realm of human control or comprehension.


mungojelly
Apr. 2nd, 2006 12:13 am (UTC)
Re: hard = fast ; hard = difficult
I certainly hope that the corporation as it exists today doesn't persist very far into the future. They're a strange beast.

There are lots of possible (& actual) locations of economic or narrative gravity. We could all start forming Pickle Clubs, where we go to have pickle tastings, and we could buy Pickle Club Computers and use them to calculate complicated pickle algorithms. The future would then belong to pickle clubs. I really think there's no rules on this thing.

By the time there's anything resembling autonomous intelligence, augmented intelligence will already be so qualitatively distant from unaided human minds. Hopefully by then we will have clearer ways of thinking about & dealing with the possibilities presented by autonomous intelligences.

I also expect that when we are able to make autonomous intelligences, we will also be able to define their goal structures. They won't necessarily have independent wills, IOW. They could be agents, following arbitrary agendas. Like corporations, but the memos from headquarters will be taken very seriously. Now that's scary; that's different.

<3
mungojelly
Mar. 17th, 2006 12:18 am (UTC)
It makes no sense to imagine a Singularity occuring as if it were outside of the people experiencing it. The humans who experience Singularity will not be human anymore, not exactly. They will have mental capacities that are difficult to imagine. To them-- to us, then-- the rapid transformations occuring in the world will be much more comprehensible. IMHO, therefore, the hardest of hard takeoffs won't seem so impossibly fast from inside. A million different fundamental changes on the same day to the way the world works might not seem like any big deal.

Some people have described this as "subjective time" being slower. I think that vastly understates the strangeness of what is going to happen. These creatures that experience the Singularity-- these Powers-- will not think in ways that are like human thought except for their speed. They will think in ways that are as different from human thought as human thought is from the thoughts of a flatworm.

Consider that Powers will have the ability to easily maintain a database with extensive information about every human being on earth. They will have instant access to this information, in contextually aware integrated ways, just as you have access to your memories of people that you know well. In other words, there will be a fantastic intimacy between every human being on earth. Everyone will know everyone, literally inside & out.

Powers will remember everything they ever see-- which will be much more than any unaided human has ever seen. They will not just store this information away, but will have access to it in ways which are similar to normal memory, constantly engaged with the world.

They will also have a tremendously enhanced ability to plan out their future. We'll start to see the beginnings of this very soon. Imagine how useful it would be, for instance, just to be able to tell your computer: "Hey computer, next time I'm at the supermarket, remind me to get sponges." We'll start to build long-term conceptual intentions into our electronic selves-- intentions about the world which are much more accurate & intense than unaided human intention.

These sorts of changes add up to something very strange. It's not "like a human but more/faster." It's something else. We shouldn't imagine any kind of human entering into a Singularity. Entering a Singularity transforms you into something else. That makes a lot of the hard/soft takeoff discussions seem hollow; they're discussions about how the normal human time-line sort of experience is going to relate to the progress of AI, and that fails to encounter the true strangeness of the matter.

<3
ankh_f_n_khonsu
Mar. 17th, 2006 02:45 am (UTC)
All good points, but there seems to be a very similar component among actualized humans. We can, after all, command tremendous accuracy among invariant representations (memories). Systems like NLP, Lilly's Metaprogramming of the Human Biocomputer, etc provide individuals massive increases in recollection.

You could ascribe many of the abilities of these "Powers" to actualized humans. "Subjective time" moves slower for them. They think in paradigms that have little in common with paradigms endorsed by your standard domesticated primate.

I'd rather see a push towards integralism than a drive towards frivolous reliance.

I think all this gets away from the original point of the poll.

Namaste.
mungojelly
Apr. 1st, 2006 07:11 pm (UTC)
What you're saying is that it's possible to think in ways that are quite different from the ways that most people think. I agree. I also think that this distinction is very tiny compared to the difference caused by entering the Singularity.

You can improve your memory a little by taking gingko, or practicing remembering things, or actualizing yourself, or what-have-you. Let's be generous and say that you could double your ability to remember things. Let's be really generous, let's say that could you multiply a thousandfold your ability to remember things. It's still so unbelievably fantastically tiny compared to the Singularity.

Some savants are able to remember long series of numbers. Most of them, in my estimation of the situation, do tricks like using their spatial memory to remember an imagined walk-- they enlist a huge part of the brain via some corridor. Their abilities are very impressive, when compared to normal humans. Much better than twice as good. Thousands of times as good. But compare it to the Singularity: After the Singularity, depending on how deep into the physical structure of our universe we manage to embed computation, there's every possibility that we'll be able to remember a GOOGLEPLEX of numbers. More mundanely, we can each already remember billions of numbers. Billions.

It's not a little bit of spice here, a bit more parsley there. The soup pot of the human experience is about to encounter a dumptruck full of computation. So I hope you're in the mood for information stew.

<3
ankh_f_n_khonsu
Apr. 2nd, 2006 03:33 am (UTC)
I agree on all counts, but I'd still rather see a push towards integral ideologies than a movement towards reliance on technologies. I think we can certainly use those technologies sustainably, but I'd rather have integral paradigms before rather than after.

Namaste.
mungojelly
Apr. 15th, 2006 01:07 pm (UTC)
Absolutely. I think we all need to expand & integrate our consciousnesses, in order to be able to deal with this situation. We're not at all where we need to be, and we don't have long to get there.

What seems dangerous to me is that the technology is just stuff, and so it bends itself so easily to our will. The paradigms that we've established are being written into code. We're talking here in a system created with a particular energy, and that energy is being reinforced and deepened. It matters a lot how we do things right now.

What techniques have you been using to get yourself together? You mentioned John C Lilly, I've read him years ago. I was thinking about him recently actually, when I was realizing how my unconscious mind really does take orders-- it's very directly programmable, even if it can be flakey. The main tool I've been involved with these days is Taoist Tai Chi. Tai Chi is weird ass shit-- it's a very different system of thought that I've been grappling with. Hopefully it will help to bring these kinds of understanding into the creation of these new spaces. They definitely need a little something.

<3
ankh_f_n_khonsu
Apr. 15th, 2006 01:12 pm (UTC)
I studied Tai Chi in my younger years, but I was moreso interested in Hung Gar, Baqua and Chi Kung.

I've found Lilly hit or miss. I presently prefer Center of the Cyclone best. It won me back after Programming the Human Biocomputer.

I'm also fond of Hyatt's "Undoing Yourself" models, Wilson's "model agnosticism" and "guerilla ontology", and Hyatt/Mumford's tantriks.

The central component I've found most dynamic thus far has been Kundalini.

Namaste.
mungojelly
Apr. 16th, 2006 01:33 am (UTC)
I'd like to get into some more advanced chigung, chi exercises, but I feel (for me personally) that it doesn't work to learn those things from books, and this is the teaching that I've been lucky enough to come in contact with. My school of Tai Chi isn't combat oriented, and it was invented with an intention of reawakening some of the original spirits put into the form by the old Taoist masters, before the art became militarized. So there's a lot of chi that gets built in our form.

Kundalini energy is powerful but dangerous. I've had some transformative experiences myself. But I've definitely gotten a sense for what can go wrong.

We need to do a lot of collective energy work as a society. We need to refocus ourselves on subtle energies, bring ourselves out of this materialistic realm. A lot of people are rejecting the Singularity outright because they see it as associated with "technology," which is associated with this very anti-body anti-energy anti-feeling materialistic atmosphere. If that holds up it could lead the future into some bleak places. We need energy work that ties this small work (of transforming our mere little species) in with the bigger picture.

There are definitely steps being taken in that direction. But still a lot of the things we're bringing in, to try to overcome our spiritual blindness, are just words and concepts. It'll take a lot more concerted committment to really bring in transformative energies, to build real sanghas & really form a collective understanding of the subtlties of human experience.
ankh_f_n_khonsu
Apr. 16th, 2006 03:57 am (UTC)
Kundalini can certainly go awry.

Namaste.
( 12 comments — Leave a comment )

Latest Month

August 2017
S M T W T F S
  12345
6789101112
13141516171819
20212223242526
2728293031  

Tags

Powered by LiveJournal.com
Designed by Ideacodes