Slashdot Mirror
NPR Looks to Technological Singularity
Rick Kleffel writes to tell us that NPR is featuring a piece with both Vernor Vinge and Cory Doctorow looking at the possibility of the "technological singularity" in the near future. Wikipedia defines a technological singularity as a "hypothetical "event horizon" in the predictability of human technological development. Past this event horizon, following the creation of strong artificial intelligence or the amplification of human intelligence, existing models of the future cease to give reliable or accurate answers. Futurists predict that after the Singularity, posthumans and/or strong AI will replace humans as the dominating force in science and technology, rendering human-specific social models obsolete."
...welcome our new post-human overlords. (Somebody had to say it.)
So, they first say that you can't predict what'll happen after that singularity because The World Will Be So Different Than Now, and then proceed to give predictions of what'll happen after that singularity?
Brilliant, real brilliant.
Since when have futurists have gotten anything right? If we believe them we would all be enjoying our flying cars, that can interact with us using voice control. We would talk to each other using video phones(first designed in 1969? AT&T).
The event singularity doesn't have to happen because the futurists are always wrong.
i thought once I was found, but it was only a dream.
First Post-Human!
I'll be your candy shop of infinite deliciousity if you'll be my discotheque of endless rump-shaking.
Well, I doubt it. I agree with most of the idea of the 6:17 cast and even agree that educational and social changes like widespread literacy may be considered a singularity, but I seriously doubt the timeframe of one generation/30 years they mention. Literacy was adapted over hundreds of years, network communities have been developing for at least 30 years and are still primitive and very far from a "collective mind". For me Wikipedia is "augmented intelligence", but before that I had the Encyclopedia Britannica on my iBook and before that an encyclopedia on my desk, so this to is evolved. And since the Wikipedia is created by so many, it may be considered a primitive product of the "meta intelligence" described.
Btw, the piece from NPR focuses (very trendy) on collaboration and advanced information management, they do not lay great hope on a major breakthrough in AI.
memomo: free web based language trainer DE-EN-ES-FR-IT
It will be a Technological Singularity ON WHEELS!
Willy on Wheels!
An author using /. to publicise their latest novel. Yawn.
From what I've seen we are as near to creating decent AI as we are to producing fusion power stations.
One of the toughest nuts to crack is what are going to want to do, that is what should our goals be.
If you look at most of the goals we have right now, they're pretty mundane and shortlived. Curing disease, stop killing eachother, end to hunger, creating objects that we find beautiful and pleasing, creating more living beings like ourselves.
Once we reach a singularity we'll have the technology to do away with all these problem oriented goals and I for the life of me can't really think of any obvious goals past that point. While I agree with the premise that we don't have any reliable way of predicting what our goals will become past the singularity, does anyone have any guesses?
You know, I used to have this technological post-human bent. Buried in C++ programming projects, I admired the order of all that I was creating. It was fun. I'd get a new set of behaviors programmed in the usual conditional branching - if/else, class polymorphism, you name it - and seeing it work was exhilarating. The idea that humanity could reinvent its world piece by piece - much like in the argument where if you replaced each neuron in your brain one by one with an artificial equivalent, at what point would you cease to be human, if at all? I still have Raymond Kurzweil's The Age of Spiritual Machines on one of my bookshelves.
;)
The thing is, we are still way surpassed at this by billions of years of evolution. We run on energy from fossil fuels and build from materials we've mined and shipped. On the other hand, we find bacteria living in the most surprising places, we find superior sonar in dolphins and bats to anything we make, and all of it runs on, ultimately, fresh plant matter. We get excited over a myomer that lifted some heavy weight, and I tell you, an elephant can do the same thing given enough food. The sheer variety and efficiency of the ecosystem virtually guarantees that most any way you can think to survive has been done somewhere, somehow, by some living creature. We're worrying about when oil will peak, if we can live another century, and outside our doors the world can go on for eons to come provided we don't break it with our silly toys.
And in a geek-intense environment like this one, I think I can say that it's difficult to beat the end product of a long-term evolutionary algorithm, which itself is an arguably good model of what the world around us acts like, and you all will understand.
I don't deny the coolness of my Apple notebook and I've got a decent number of shelves full of programming books, but I think biomimicry is where it's at. We can go a lot further learning from our world of proteins and DNA and RNA and using - or just having fun with! - what's already there.
We can also get out more and enjoy our analog, fuzzy-logic, neural-net-driven, molecularly-computed fleshy selves.
I keep wanting to find Vinge and slap him around a bit until he shuts up about "The Singularity". The thing is, there have been several "singularities" in human history: the Agricultural Singularity, the Industrial Singularity, the Computer Singularity, and so on and so forth. Or, to use the term that most historians use - rather than "Singularity", "Revolution." Yes, technology will change the context of human interaction. Yes, nifty and non-nifty things will happen. But, dammit, it's not as if technology has never fundamentally altered society before. Get over it, already.
We have a time-war with the Daleks.
Christ. Just wait until the "defend traditional marriage" crowd gets word of this.
I don't care if it's 90,000 hectares. That lake was not my doing.
Artificial primarily means that it comes from artifice (ingenuity) or art. It doesn't (directly) mean it's fake, it just means it's a consciously created work of humankind rather than nature. I think that in modern times with so many knock-offs of natural goods, such as artificial sweetener, the secondary definition has gained the upper hand.
Check out wictionary (It's the hive-mind wikipedia, it must be right!)
When you read enough literature from the 16th and 17th centuries you get more familiar with the original, literal meanings of words such as this one. A favorite subject was to compare art to nature, and they'd freely use the word "artificial" to mean that which comes from human arts. This is not to say that the secondary definition is wrong: for example, when in Book 3 of The Faerie Queene a troll creates an artificial woman to replace the girl who left him out of snow, "virgin" wax and some gold wire (and of course wackiness ensues) it is repeatedly underscored that this "False Florimell" is a cheap immitation.
Anyway, you can chose any definition you like. I sort of prefer artificial intelligence to synthetic intelligence or whatever, just because how you regard the word artificial says a lot about you and what you think of human creativity. And I don't like euphamism treadmills, which is effectively what we're talking about here.
The hard takeoff concept of a seed AI has as a prerequisit the creation of a computer program that can understand and write source code. I'd probably try to make something like that to make my job as a programmer easier, but there's no way I'd let anyone know I had.. otherwise they wouldn't need me. Which makes you wonder, maybe someone already has one.
How we know is more important than what we know.
from BoingBoing http://boingboing.net/ http://craphound.com.nyud.net:8080/man-machine-mer ger.mp3
Only bad things happen when people steal hard-science ideas to describe soft-science phenomena-- the ridiculous (and unaccountably persistent) idea of "social evolution" is one example, and as far as I can see, this "technological singularity" notion is another. History is a phenomenally complex system; even in hindsight, it's virtually impossible to find real patterns, and grafting the language of astrophysics onto a theory of social progress lends an undeserved air of gravitas and mathematical precision to what's essentially just fun speculation.
Sure, things change, sometimes quite suddenly and unexpectedly. But really, the relationship between the development of literacy (NPR's example of a past singularity) and the subsequent course of history is nothing like the relationship between a real singularity and... anything. It's just a bad metaphor, and I think I'd have a lot more respect for "future studies" if they dropped it and came up with a new way of describing whatever phenomenon it is they're predicting
I think Bruce Sterling gave a talk on this subject, it can be found a bit down on this page: Long Now Seminars.
My personal whimsical theo.. hypoth... idea is that alien civilizations turn into (towards us) apathetic singularities, and that's why we will never hear Chenjesu's crystaline humming calling us. Maybe the universe will end in some sort of rather dull uniform black technological singularity goo.
The Chair Corp. comic(*00-12)
The C-Prize is the path to superhuman AI.
And as for the "threat" of superhuman AI:
Even assuming AI were to develop the equivalent of genetic self-interest, (something that would take a long time even if humans turned them lose to reproduce without us selecting them appropriately) I'd much rather be in competition with a species that had the potential of being symbiotic due to having a different ecological nich. If it gets to the point that the solar output (forget the sun falling on Earth here -- that's too insignificant to consider important to a silicon based life form) is the limited resource, I suspect that the nich humans will fill will be orders of magnitude larger than they now fill on earth.
The best hope humans have of the transhumanist wishful thinking is to develop superhuman AIs that find utilizing the gas giants to their advantage given the limited supply of silicon. Humans, as the highest form of organic intelligence, would be the natural species to transit to higher intelligence.
Maybe the super AI's could get around this by using a straight carbon semiconductor form of intelligence or something but there is more going on in our brains than we understand. For example, I suspect there is a lot more quantum logic going on within our brains than currently thought by cognitive scientists and neurologists. It only makes sense evolution would have exploited every angle of the physics of the universe to create intelligence. My point in bringing in the possibility of quantum logic is that there are really many things we don't know about natural systems of high complexity and I suspect the same will apply even to super AI's. The fact that we might have the laws down cold at the quantum level doesn't mean we know how things operate in the higher complexity systems.
Human brains are very valuable repositories of ancient wisdom about the universe and the most optimal thing for the super AIs to do -- at least for a while -- would be to transhumanize our brains for us.
Moreover, if it is ok to pass laws to prevent the creation of intelligences greater than our own, why isn't it ok to pass laws dumbing down the smartest among us?
The self-determination argument applied to humanity as a whole -- striving to maintain control of its own destiny by preventing the creation of higher non-human intelligences -- applies also to people who want to maintain control of their own destiny against those smarter than themselves.
Personally I'm much more frightened of unenlightened self-interest than I am enlightened self-interest.
I really wish it were possible to make some of the "smart" people who are really good at grabbing control of resources intelligent enough to understand that they are using those resources in very stupid, self-destructive ways.
Indeed, it is this abysmal stupidity among the shrewdest among us that is my main motivation for promoting super AI.
Seastead this.
Will AIs produce AIs, and if so, will they be better, or equally flawed?
The current thinking is that we will make seed AI, i.e., general intelligence for manipulating software, and that it will improve itself, in an incremental fashion, all the way up to and beyond the level of human intelligence. Of course, this will be done with the help and guidance of programmers but the fear is that by giving it free reign to manipulate itself we will no longer be able to understand what it creates. Not only will this mean that we won't learn anything, but we'll also be unable to control it. As such, most people who seriously consider working on this stuff advocate a goal based higher level of functioning with "friendliness" to humans as being the primary goal and improve yourself as a secondary subgoal. That way, even if the beast gets out of control, the worst it will do is solve world hunger.
How we know is more important than what we know.
Now, let me see... when was the last Singularity? Was it Y2K? Or was it perhaps the Jupiter Effect (when all the planets lined up and the gravitational effect tipped the earth out of its axis?) Or am I confusing both of them with the beginning of the Aquarian Age? Or maybe I'm thinking of the Harmonic Convergence of August 17, 1987?
I'm way too young to remember the Millerites and the Great Disappointment of October 22, 1844, when Jesus failed to reappear, but I've been blessed to live through a veritable multiplicity of singularities.
Oooh, singularity! I like that word. So much kewler than, say, "Armageddon." It sounds so technical, so scientific, so free from ranting religiosity....
"How to Do Nothing," kids activities, back in print!
> From what I've seen we are as near to creating decent AI as we are to producing fusion power stations.
About 10 years away then...
"existing models of the future cease to give reliable or accurate answers"
The premise of this definition is that models of the future give reliable or accurate answers at present. What are the models they talk about? Special futurist models? Do these really give reliable or accurate answers today? Or do they mean all models of human behaviour, i.e. most models of the social sciences? Supply & demand will no longer determine price?
If the models are found not to be good predictors of behaviour, they will be modified or replaced. You know... sort of like how it works right now?
If patterns in human behaviour start changing rapidly because of rapidly evolving superhuman intelligence, then sure, our ability to model that behaviour will go out the window. But then, we wont be doing the modeling, superhuman intelligences will. I don't see why the emergence of superhuman intelligence would have to lead to a singularity.
I believe the models will cope. Not "existing models", but tomorrow's models.
The problem is also mostly with the expectations people have of computers. Everyone wants computers to return deterministic and easily tracable results. For example if I want a value from a database I want to issue a query and have the value returned. I don't want a system that would return it faster but only with 80% of correctness, I don't want any "fuzziness" only exact numbers. In other words people would rather have computers do what computers are doing - calculating stuff fast and exactly, they don't want computers to really act like humans. I think subconsciously we will just never allow computers to reach a human level of soffistication and thus they will probably never surpass us.
On the other hand, what would rather happen is that we will slowly integrate machines into ourselves - litteraly. As soon as the baby is born we will tag it with an RFID, we will implant sensors for infrared vision, ultrasound, we will inject nanoparticles to boost the immune system. In other words I see a cyborg future were we become one with the machines. If anything or anyone will destroy us it will only by ourselves, at the same time if anything helps us prosper, it will also be ourselves. The future is (mostly - short of a big meteorite hitting us) in our hands...
>>> the worst it will do is solve world hunger
"Thank you for using AI-net. The best solution to "world hunger" appears to be large-scale thermonuclear war. I have taken the liberty of releasing sufficient war-heads to destroy all humans who can get hungry. As a side effect and in accordance with my prime directive (being a friend to humans) all human suffering will be ended.
Have a prosperous existence."
Did anyone foresee that in the 90s the largest empire humans ever built would evaporate like a soap bubble? (Except Poul Anderson in the 1953 story "The Last Deliverer"). Talk about existing models of how things work falling apart.
Imagine an intelligent and curious human from rural Nepal, or Papua New Guinea. Could you explain your job to them?
Could you do your job without the embryonic augmentations we have now, such as Google?
We're partway up that vertical curve now.
From a 15th century monk's perspective, today's curve is vertical. Of course to us it's clearly not. Thus the flaw of the hand-wringing over "the singularity" is illustrated--it suffers from the classic error of attempting to evaluate the future in the context of today. Of course when we get to the future, we'll be in the future too--so it doesn't matter what we think now.
Ever hear of the generation gap? The youth of today are different from us--they've been raised from birth in a world of ubiquitous networked computing and ambient findability. (see? I can throw around stupid buzzwords too.) Talk of "The Singularity" is not much different from complaining that your kids spend all their time texting. It's making explicit the fact that you can't imagine keeping up as you age. Well duh. We won't be running the show in 2050--our kids and their kids will.
Build a man a fire, he's warm for one night. Set him on fire, and he's warm for the rest of his life.
As such, most people who seriously consider working on this stuff advocate a goal based higher level of functioning with "friendliness" to humans as being the primary goal and improve yourself as a secondary subgoal. That way, even if the beast gets out of control, the worst it will do is solve world hunger.
Isaac Asimov discusses that concept in one of his short stories; The Evitable Conflict. In that short story, there were huge computers that could assimilate vast amounts of information in order to determine the best course. Because of their reliability, the machines had been put in charge of things like food production and distribution. In the end, the machines began manipulating events to ensure that anyone who disagreed with the machines control was removed from a position of influence. They did this because obviously what was best for mankind was to be guided by the machines, who didn't start wars or squandor resources like they did. In order to maintain what was best for humanity, they had to act against individual humans and, in short, ensure that humanity was never ever the master of its own destiny.
It's fiction, yes, but even such simple goals as the one you suggested need to be interpreted. How should one weigh up the needs of the many against the needs of the few?
Just because you're paranoid doesn't mean there isn't an invisible demon about to eat your face
- Moronic politicians get caught up in the hype, form a gov't agency called National Art-intel Singularity Administration to make it happen, and the country's resources in AI are drained away into ineffectiveness and software that keeps crashing.
Nah, the gov't wouldn't do something that dumb.
Pavlov wouldn't be so famous if he'd used a can opener instead of a bell.