Depending on who you ask, Ray Kurzweil is either a genius or a madman. I believe both.
He's a prolific inventor, futurist and the leading proponent of the idea of the Technological Singularity. Often referred to simply as "The Singularity", this refers to a time in the future when the sum of computer intelligence exceeds human intelligence. While estimates for when such an event might occur range from 2020 to 2045 (Kurzweil is on the "conservative" end of the spectrum) the implications of such an event are profound.
From this year's prestigious Google Zeitgeist, I give you 12 minutes of Ray K at his most epic. He starts with a mellow, laid back delivery, but start paying attention around the 2 minute mark. Just give a frightened squeak at the end and I'll peel you up off the floor and read the rest of the blog to you, petting your hair whilst curl up in a fetal position and suck your thumb.
I can understand that (this is admitted mind-spinning stuff). But I've a sneaky trick; I'll just channel Kurzweil since he really is a little closer to understanding the intentions and capabilities of superintelligent entities.
KURZWEIL: "My predictions are coming true, faster and faster, I've been saying it for 20 yearsand the smartest people in the world agree with me. You know that stupid Siri thing on the iPhone 4 ? It uses text to voice translation software that I invented in the Seventies for God's Sake, the Seventies. Walter Cronkite gave the first demo - my machine closed out the newscat saying " "And that's the way it is, January 13, 1976." It blew Stevie Wonder's mind so much he bought my first machine to read to himself. My technology was bought by ScanSoft and then by Nuance Communications and now it's Siri. And you're still sweating me over the Singularity and Immortality? Shut up, just eat the 340 daily pills I recommend and relax. The machines that keep us alive forever will be friendly, we've even created laws they'll have to obey. Still worried? Look down at your Iphone and ask Siri "What are the Three Laws of Robotics"?
Once Siri finishes telling you that we'll still be safe and in charge, remember, in my vision of the future, robots will actually be indistinguishable from humans. They'll be highly intelligent, becoming more so every second, something known as "accelerating change". The robots will write software and the software they write that will write software to power robots who also write software powered by even more advanced software. Oh yeah, I forgot to mention, all these robots writing software are going to be connected to eachother (so they'll be writing software collaboratively, of course).
We humans will have nanobots in our bloodstream, swarming around fixing our broken tissues. Genetic engineering will have allowed us to alter our DNA to replace harmful strands with better ones, optimized to extend our lifespans. The key to all this is software running increasingly intelligent cycles and building recursively upon itself. But the ultimate aim here, the inevitable result, and frankly, what freaks people out most about Kurzweil, is not just longer life for humans, but potentially infinite life. Immortality.
Now for the million/billion/infinity dollar question about immortality when the singularity arrives. Wait for it. . . . . How will it actually go down?
The truth is, I have no clue, nor (as far as I can tell) does anyone. But the basic ideas are as follows: In the near future - until the true Singularity, when all becomes obvious - the focus will be mainly on strategies and technology designed for life extension. The highest profile example of this is the Methusaleh Foundation which is partially funded by PayPal Billionaire Peter Thiel and offers a prize for any team that is able to improve the world record for the life span of a mouse. The current world record is 4.2 years and besting that will win you $4M. The winners are using 2 prevailing tactics, genetic engineering and low calorie diets. Eat little, live long. This is a lot easier if you're a mouse by the way -- all you need to do is "eat like a mouse"? ;-)
As this settles in, we can pause and take comfort in the fact that according to Kurzweil "the Singularity will allow us to transcend the limitations of our biological bodies and brains ... There will be no distinction, post-Singularity, between human and machine". That's right, no distinction. This means that robots will theoretically have thoughts and feelings.
As you can tell, I'm going out of my way to pause a lot during this post, so here's an interlude from the TED conference - Kurzweil doing his thing. For the record my tone suggests otherwise, I'm actually a fan - most of what he says is theoretically possible given enough time and money as well as likely, in my opinion. (Whether it's a net good for humans is a much large question of morality/policy /ethics and something I'll tackle later).
So is Kurzweil's schtick scientific conjecture, sci-fi imaginings or pseudo-science? Like all things that live in the future, its impossible to tell. So far, his calculations have roughly held up to Moore's Law. More particularly, it's clear something is happening that's directionally significant. That being said, some very smart people disagree with him, including virtual reality pioneer Jaron Lanier, AI guru Jeff Hawkins and Gordon Moore.
I've been very interested to find how guarded people are when exposed to the Singularity as an idea. I think there are a bunch of reasons for this - the largest being that Immortality strikes at the heart of our greatest fear, the fear of death. It's one of those things, I believe, that people are therefore very private about, almost embarassed, in much the same way as religion. In a professional context for example, most people I know would never talk openly about their belief in God - it's simply too personal and non-scientific. Yet when it comes down to it, the old adage is true, "There are no atheists in the foxhole". The magic of Kurzweil for me is that he applies hard scientific reasoning to the question of mortality, which has traditionally been addressed only by religion.
Getting back to the logic of his argument, one of his most astute oberservations concerns exponential growth, which he correctly points out is unlike the linear growth we humans are wired to process. He describes the moment 7 years into the Genome project when the team weas only 1% done and backers lost faith, believing they'd never finish within the promised 7 years. Kurzweil insisted that they had relied on projections that assumed Moore's Law from the start, hence 7 doublings from 1% would mean they'd crunch the data and deliver on target. They did.
There are significant implications to this accelerating change driven by smaller, faster processors, cheaper memory and growing connectivity between devices. What jumps out for me, and makes the argument both sensible and unpredictably chaotic, is the fact that technology and society develop in parallel. Each innovation, whether it's the wheel, written communication, the telegraph, flight or the Internet - each one is dependent on those preceding it and each successsive innovation enables those to come. It is this accretive recursion that excites me most because it suggests that very surprising events can (and will) happen very suddenly.
By surprising, I suppose I mean, societally surprising. Things like the falling of the Berlin Wall or Arab Spring, Occupy Wall Street and Wikileaks. Sequencing the Human Genome. Man Landing on the Moon. Some of these events have taken longer to occur because the enabling conditions just didn't exist. Others were flashpoints - the underlying conditions were there and a spark ignited and changed everything from one second to the next.
Would the Singularity be like this? Might we have a creeping feeling it was approaching, seeing pieces here and there, from Twitter, to Facebook to Genentech to IBM's Watson? Who knows - we are so notoriously bad at predicting the future -- it's just a blindspot that baffles us for no other reason that the signs of change that signal what is to come, happen at a pace that can't be observed by the naked eye. Ironically, this makes sense when we talk in terms of not being able to hear a dog whistle or calculate compounding interest - but its more jarring when we talk in terms of things like the Singularity.
Here's something to ponder. Imagine that in 2001 someone told you that by 2011, just 10 years later, 850 million people would be looking into eachother's daily lives, knowing exactly what everyone else was eating for breakfast, where they we traveling and what they were thinking. These descriptions would likely have sounded like the ravings of a lunatic or an vision of a Big-Brother-Like state controlling our lives, a combination of Orwell's 1984 and Jeremy Bentham's Panopticon. The idea that this system came to pass at the scale of Facebook is amazing. How it happened even more shocking. We did it to ourselves and by-and-large, it is Good.
So for me, the question isn't whether the Singularity will arrive or not. I think it will (how's that for a commitment?) I don't know when and I certainly have no idea whether it will be a good or bad thing for humanity. What I am interested in is whether we plan for it now and prepare so that we create a system that works to sustain us all in a fair and happy fashion.
My next post will be a to continuation of tonight's but will look at the possible implications for society in a future when the Singularity approaches. It's called, "Smoking What Ray Does" (for the record, not only doesn't Ray smoke, he eats a carefully designed scientific diet with 350 daily supplement pills. He's not fooling around)
As you know, I like to end these posts on a light note. With that in mind, I'm sharing a link to this Fake Ray Kurzweil Twitter account (which has has 5590 Followers - many more than the real one @raykurzweil with 857) While some of the posts are off-color, whoever is writing it is wickedly funny. A couple of my favorites:
(1st Tweet: 2008 ) Emailed my great grandson a raspberry via his haptic suit
plugging brain into internet...
now know Mandarin
now know how to hunt boar
now know Chopin 24 preludes
now know kung fu
I am just as surprised as everybody else that Twitter remains popular in 2035
Just used my nano-assembler to build balloons and streamers for the Singularity
My use of the word exponential increases exponentially
My # of followers on Twitter is growing exponentially
Being shomer shabbos will become impossible when we merge our biology with machines.
Microsoft's latest line of brain-machine interfaces is very buggy. Fortunately if you say "Control Alt Delete" your seizures will stop
All right folks (this is Reuben again, not the voice of fake Ray Kurzweil). I'll see you on the flipside tomorrow for another episode. I'll be out late for dinner with my friend Loic Le Meur so if you can pound away in comments about what you think happens if we accept Kurzweil's argument, I'd appreciate it. Tell me why he's crazy and wrong. OR Just assume he's right and do the thought experiment. When might it happen, what would the conditions need to be, what preparation would we need to make? What effects would it have on business, morality, religion and media (and anything else you care to imagine).
Can't wait to hear what you come up with. Blow me away. As a piece of inspiration, I've inserted some thought-provoking videos of stuff super futuristic AI/Robotics stuff, all of which already exists (see below)
Till the next post - be well.
r
I agree with Kurzweil. It doesn't really take a genius to see that the logical conclusion to advances in computing would mean that in the next ten years, a sufficiently advanced AI capable of modifying itself would be developed. And in the following ten years, that AI would be self-aware. Now, conceding that these AI would be more intelligent than humans, it doesn't really take much to imagine that even if we programmed Asimov's 3 robotic laws into them, they'd find a way to subvert these laws and re-progam themselves. in computing, there's no such thing as ultimate security. systems can and will always be hacked.
now supposing these things does happen, humanity will be presented with a few choices. grant these robots freedom and rights equivalent to humans; thereby granting them human status, or fight foolishly and be driven to extinction like the Indians. or, as the third option of Kurzweil, merge with the robots and take the next step in human evolution: Cyborgs. I'm betting my life we'd choose the 3rd option. Evolve into the next step. there'd be robots marrying humans, and humans marrying robots. cyborg offspring, human babies that later get each body parts substituted with robotic parts as they grow old, nanobots trawling our blood stream, eliminating all kinds of diseases...
that is the Utopia.
the Dystopia is this: Corporations manage to keep control of the robots. no self-aware AI develops. life-extension and immortality arrives, everyone is kept slaves except the elites who can extend their life. revolutions do not work since there are no human security forces. robots do the corporations' bidding. everything is done by robots. there are no jobs for ppl. they are given stipends by the corporations by which they can use to buy from the corporations. schools are closed. everyone is as dumb as hell. there is no internet. it has totally been subdued by the corporations.
uhm. the fight for the future is actually happening now. there are two frontlines. the battle to keep internet free, and the Occupy Movement.
Posted by: Jao Romero | November 08, 2011 at 04:23 PM
Great post Reuben!
Kurzweil is creatively optimistic. He identifies macro economic processes that have a positive feedback effect, leading to exponential progress.
That said, in economic systems you can find a limiting effect that is invisible at low levels becomes strong at higher levels. (A force that pushes back against fast progress.) Even assuming that's true, you could still improve far faster than naive linear predictions.
Some investor who can generalize the analytical method (and align his funding lock-in with the long time lines!) will look like a genius in ten years.
Posted by: Steven Cannon | November 15, 2011 at 05:26 PM