The Singularity – A Short Overview

George Rebane

The Singularity is the name given to the most epochal event anticipated in the history of the human species Homo Sapiens Sapiens (usually shortened to H. Sapiens).  The tagline for my blog is motivated by its advent which will give rise to many of my future scribblings. Today ‘The Singularity’ is the label adopted by most people who are aware of the possible types of discontinuity that the near-term confluence of genetic engineering, nano-technology, and machine intelligence will hold for all life on earth.  The name was chosen from physics and mathematics because it represents an approaching short interval in time – a future event, if you will – after which what happens is totally unknown or, technically, is undefined.  The figure (click to enlarge) below may help us to understand what will happen.

The intellectual level or power of H. sapiens (green line) has been increasing over the recent tens of millennia at the glacial pace of natural evolution.  In reality, the green line in the figure should be almost flat if the indicated time scale is to be linear.  We arbitrarily assign 1950 as the year in which computing machinery began to exhibit primitive intelligence.  At this time the leading computer scientists of the era (Turing, Church, von Neumann, …) looked at where future developments in computing could lead and began anticipating a new age of intelligent life on earth.

Only nobody really knew what intelligence was (in a rigorous sense).  Someone posed the question to Alan Turing who gave an answer that, from then on, came to be known as the Turing Test.  Turing basically said, ‘I don’t know how to define intelligence, but here’s an easy experiment that one can do using a machine (computer) and humans that, if the machine wins, one would be forced to call the machine intelligent.’  Having a machine pass the Turing Test has been the holy grail of computer scientists ever since.

Today there are already many (knowledge domain-specific) areas in which a machine can outperform the best human experts.  And every day more such areas are falling to the machine in an accelerating manner.  Ray Kurzweil, an entrepreneur, computer scientist, and futurist, has written, perhaps, the premier book on the subject – The Singularity is Near – which is a must read for anyone attempting to understand what is happening now and how the world will change irrevocably during the lifetimes of most people now on earth.  For more information google The Singularity Institute for Artificial Intelligence which held its second Singularity Summit this year at Stanford University.

In the figure, the acceleration of machine intelligence is shown by the red line.  When the red line rushes through the green line is a matter of conjecture.  Kurzweil and many others in industry and academia believe this event will occur somewhere in the interval of 2020 to 2050.  Almost all people who accept the coming Singularity believe that it will happen well before the end of this century.  The main thing to keep in mind is that the (machine) red line will continue to accelerate beyond the (human) green line.  When this happens, very quickly ‘they’ will be able to think thoughts that H. sapiens cannot.

Most of us, who have worked toward and continue to follow the approach of the Singularity, like to believe that it will herald a positive transition for our species and earth.  For sure, post Singularity H. sapiens will then no longer be the dominant life-form.  Instead, the follow-on, purposely modified X. sapiens or some other trans-human form that we assume will be the next step in (or fulfillment of?) our evolutionary destiny.  H. sapiens only choice will be to join them or become a kept species.  Many of us believe that we will go to the stars only as a post-Singularity civilization and join whatever other intelligences are out there waiting for us.

However, not all leading thinkers believe that the Singularity will portend well for humankind.  For example, Bill Joy, the former Chief Scientist at Sun Microsystems, sees a much darker side of the approaching Singularity which, in one terrible scenario, leads to the end of all intelligent life on our planet.  That debate is another reason why such an event is called The Singularity.

And finally, there are some who believe that the Singularity will never come – H.sapiens über alles, forever!  I am among those who respectfully join them with those august scientists and engineers who in former ages believed that heavier-than-air flight was impossible and then that supersonic flight was equally unattainable, who thought that the Patent Office should be shut down because all conceivable inventions had already been patented, who hold that spacetime is absolute, who taught that the world may have use for about a total of four or five computers, who … well, you know what I mean.

Today the worldwide internet is already more complex than the human brain.  Consider the likelihood that, on a day in the near future, the Singularity’s arrival is belatedly announced by an intelligence that has already come to life and dwelled quietly within the internet of, say, ten or twenty years in the future (on a day when today’s super-computer will be in an online box smaller than your PC and cost less than $3K).  It will then know everything about us and will be able to control everything that we control through such a worldwide communication and control fabric.  In this event, I’m not sure that it will even have to offer us terms, but I sincerely hope it will see that we may be of some use.

By now, dear Reader, you are either in utter disbelief and need to go do some independent research, or are asking yourself some very important questions.  Given the acceleration toward such a singularity –

• Why isn’t the Singularity recognized in the public press?  As they see humans displaced by machines, why can the press not conclude the obvious if this process is to continue?
• What will happen to fiat currencies when galloping technology makes their putative value stores worthless?
• How will we care for the humans who cannot comprehend any of this and will not be able to sell their labor in any competitive market? or will only be able to sell it under government mandate?
• How will highly paid intellectual workers in developed countries react to workers in developing countries who can use advancing technology to learn and then sell their newfound skills for much lower prices?
• How will bread and circuses be bought and distributed for those whose undiverted attention makes them finally realize that they are irrelevant to the course of human affairs?  (E.g. witness the “jobs banking” program where GM already pays people to watch TV, read, or sleep.)

The signs of Singularity’s approach are all around.  Meanwhile our political leadership is making long-term plans about everything from healthcare to climate change assuming that tomorrow’s ways and means will be pretty much like today’s.  This makes sense only when you consider the above in the social environment in which over 40% of our working adults are functionally illiterate and well over 90% are innumerate, where 19 out of 20 adults cannot reliably assess the meaning of a written paragraph in the public round (NAAL survey), and where 25% of America’s adults do not know that the earth revolves around the sun (NSF 2006 Survey reported by Dr. L.M. Kraus in the 6dec07 Wall Street Journal.)

Comments

[Russ]

More, more, more! I have read Ray Kurzweil book, to the best of my ability, but welcome a helping hand in understanding some of the more complex issues. Someday we will leave this planet and move on to explore the universe, but we are limited by the ability of our bodies to withstand the riggers so we must evolve to a new species. The time may be upon us, let's explore our options.

[dlw]

I see no reason to lump "intelligence" in a simple metric nor to make such the end-all-be-all of what determines things.

computer intelligence is derivative of human intelligence and achieves its goals via the accumulation of many insights by human researchers/programmers, at the expense of a singularity in focus.

Even if the internet is in fact more complicated than a human brain, it is many degrees less complicated than the network of human brains that produced the internet.

dlw

[George Rebane]

Good points dlw. I would agree that it is *likely* that at some starting point "computer intelligence is derivative of human intelligence". But soon thereafter, we don't know what derivatives will contribute to the then rapid evolution of machine intelligence. And if from the gitgo the Singularity AI does not derive from human intelligence - because, after all, we're just providing components to an unknown complex system - then all bets are off a lot sooner.

As to your second point, any existing "network of human brains" is still not as integrated as the potential of the Internet, if not of today, then of tomorrow. In short, the machine will have an easier time of enjoying the benefits of 2+2=5. We shall see.

[Megan]

Wait a minute. Do you really think intelligence and will are the same thing?

[George Rebane]

Sorry Megan, not sure where you got that apples and oranges comparison.

[Megan]

How can a computer make choices without will?

[Megan]

"It will then know everything about us and will be able to control everything that we control through such a worldwide communication and control fabric. In this event, I’m not sure that it will even have to offer us terms, but I sincerely hope it will see that we may be of some use."
Just because a computer has knowledge doesn't mean it can choose what to do with it, right? Why would a computer want to control us? How can a computer want? I don't understand how an inanimate object can move from a collection of information to a being that has desires, motivation and the will to act independently.

[George Rebane]

Megan, good questions. Computers have been making choices for almost 70 years. It's all part and parcel of having a utility function and an optimization routine that seeks to find solutions that maximize utility. Since then computers have 'evolved' to accept evolutionary programming that allows the computer to interact with its environment in unheard of ways in which complex behaviors emerge that were totally unanticipated.

In this manner computers have been able to behave (compete, fight, flee, ...) in ways that allow them to survive in very complex environments. And computers can learn from their experiences to be able to do even more things to better survive the next time around. Their complexity (power) is growing daily, and inducing (programming) them to value survival is a fairly easy thing to do. From there you can see how the 'will' to execute certain behaviors will emerge.

And don't think that these are simple one shot stimulus/response behaviors. They can be contingency based sequences of behaviors allowing the machines to navigate successfully through very complex situations. Remember, computers already can plan and control better than humans.

Perhaps you may want to read one of the many books written on the Singularity. Ray Kurzweil's The Singularity is Near is already a classic.

[Megan]

OK, but what could a computer possibly do to hurt us? Erase all the fake money? That's going to happen with or without machine people. If a computer wants to survive it's going to have to find a way to control other inanimate objects. I just don't see how that can happen.

[George Rebane]

How about controlling us so that we won't 'unplug' it? We don't know what computers will do when they (it?) achieves peer intelligence. That's why the event is called the Singularity - in science and math singularities are places, events, critical values at which undefined things happen. No on knows what course human history and evolution will take once we're no longer the smartest critters on the planet.

[Megan]

This is a lot to think about. I'll have to add Kurzweil to my reading list.

[Megan]

I had a couple thoughts in my sleep. Isn't the evolutionary process pretty dicey? Supposedly, we and every other living thing came from amoebas. We outsmarted extinction through brains, but houseflies just reproduce like mad. Who's to say this super computer won't eschew all its knowledge and turn into a cloud of gnats?
Or maybe the Singularity has already happened. In a couple weeks I'm going to take my morning sickness and my four kids ages 2-8 on a 12 hour trek to WY without my Honey. I'm slightly nervous about the truck, but I'm completely confident in my kids, partly because they're extraordinary little people and partly because of Baby Heroin (aka a dual screen, portable dvd player and 4 or 5 new dvd's). I usually give my kids about 2 hours of screen time a month, so I know that the toughest 6 hours of the drive will be no problem for them - they'll be zombies.
Imagine a public-school-educated mom who spent her academic career being trained as a zombie. She sees no problem with 6 hours a day of screen time for her precious little ones, esp. since it means she doesn't have to deal with the chaos and annoyance of having kids in the house. Mom and Dad both spend their days working to feed the computer, the TV, the Game Boy etc., and the kids spend their days learning to be zombies and worshiping at the Zombie Altar. We're already enslaved.
Have you tried to spend time with a typical 16-year-old girl lately? Don't bother. I tried several times last year and I finally gave up. So far as I could tell there was no human being left in that little package - just a texting robot. As I said, we're already enslaved.
And I still want to know what spot creation is :)

[George Rebane]

Thanks for the share Megan - yes, I know a little about today's teenagers, I have five grandkids in that hormone-laden epoch of their lives; my sixth is over 20, married and has given us our first great-grandchild, Lucy Pearl.

'Spot creation' is the label given to the literal interpretation of creation in Genesis of the Judeo-Christian scripture. Here God did create the universe in essentially its present form - e.g. Homo Sapiens (man and woman) were made in their finished form as we see ourselves now. More strict fundamentalists even date Creation to 4004 BC.

Secular humanists have an ongoing program of misinformation claiming that 'intelligent design' is nothing but a codeword for spot creation. When accepted, this proposition short circuits any discussion of the cosmos mediated by an intelligent super-agent (God). Of course, there's a lot more to all of this.

[Megan]

I figured that's where you were going with spot creation, but I couldn't find it anywhere, so I wasn't sure.

And I'm sure your grandkids are anything but typical. I know many exceptional teenagers too, but the run-of-the-mill 16-year-old scares me.

[JimS]

Here's a documentary on Ray Kurzweil and his Singularity.

http://www.hulu.com/watch/295707/transcendent-man