sponsored links
TED2014

Ray Kurzweil: Get ready for hybrid thinking

March 20, 2014

Two hundred million years ago, our mammal ancestors developed a new brain feature: the neocortex. This stamp-sized piece of tissue (wrapped around a brain the size of a walnut) is the key to what humanity has become. Now, futurist Ray Kurzweil suggests, we should get ready for the next big leap in brain power, as we tap into the computing power in the cloud.

Ray Kurzweil - Inventor, futurist
Ray Kurzweil is an engineer who has radically advanced the fields of speech, text and audio technology. He's revered for his dizzying -- yet convincing -- writing on the advance of technology, the limits of biology and the future of the human species. Full bio

sponsored links
Double-click the English subtitles below to play the video.
Let me tell you a story.
00:12
It goes back 200 million years.
00:15
It's a story of the neocortex,
00:16
which means "new rind."
00:18
So in these early mammals,
00:20
because only mammals have a neocortex,
00:23
rodent-like creatures.
00:25
It was the size of a postage stamp and just as thin,
00:27
and was a thin covering around
00:30
their walnut-sized brain,
00:32
but it was capable of a new type of thinking.
00:34
Rather than the fixed behaviors
00:38
that non-mammalian animals have,
00:39
it could invent new behaviors.
00:41
So a mouse is escaping a predator,
00:44
its path is blocked,
00:46
it'll try to invent a new solution.
00:48
That may work, it may not,
00:50
but if it does, it will remember that
00:51
and have a new behavior,
00:53
and that can actually spread virally
00:54
through the rest of the community.
00:56
Another mouse watching this could say,
00:58
"Hey, that was pretty clever, going around that rock,"
01:00
and it could adopt a new behavior as well.
01:02
Non-mammalian animals
01:06
couldn't do any of those things.
01:08
They had fixed behaviors.
01:10
Now they could learn a new behavior
01:11
but not in the course of one lifetime.
01:12
In the course of maybe a thousand lifetimes,
01:15
it could evolve a new fixed behavior.
01:16
That was perfectly okay 200 million years ago.
01:20
The environment changed very slowly.
01:23
It could take 10,000 years for there to be
01:25
a significant environmental change,
01:27
and during that period of time
01:29
it would evolve a new behavior.
01:30
Now that went along fine,
01:33
but then something happened.
01:35
Sixty-five million years ago,
01:36
there was a sudden, violent
change to the environment.
01:39
We call it the Cretaceous extinction event.
01:41
That's when the dinosaurs went extinct,
01:45
that's when 75 percent of the
01:47
animal and plant species went extinct,
01:50
and that's when mammals
01:53
overtook their ecological niche,
01:55
and to anthropomorphize, biological evolution said,
01:57
"Hmm, this neocortex is pretty good stuff,"
02:01
and it began to grow it.
02:03
And mammals got bigger,
02:05
their brains got bigger at an even faster pace,
02:06
and the neocortex got bigger even faster than that
02:09
and developed these distinctive ridges and folds
02:13
basically to increase its surface area.
02:16
If you took the human neocortex
02:18
and stretched it out,
02:20
it's about the size of a table napkin,
02:22
and it's still a thin structure.
02:23
It's about the thickness of a table napkin.
02:25
But it has so many convolutions and ridges
02:27
it's now 80 percent of our brain,
02:29
and that's where we do our thinking,
02:32
and it's the great sublimator.
02:35
We still have that old brain
02:36
that provides our basic drives and motivations,
02:37
but I may have a drive for conquest,
02:40
and that'll be sublimated by the neocortex
02:43
into writing a poem or inventing an app
02:46
or giving a TED Talk,
02:49
and it's really the neocortex that's where
02:50
the action is.
02:54
Fifty years ago, I wrote a paper
02:56
describing how I thought the brain worked,
02:57
and I described it as a series of modules.
02:59
Each module could do things with a pattern.
03:02
It could learn a pattern. It could remember a pattern.
03:05
It could implement a pattern.
03:07
And these modules were organized in hierarchies,
03:09
and we created that hierarchy with our own thinking.
03:11
And there was actually very little to go on
03:14
50 years ago.
03:18
It led me to meet President Johnson.
03:19
I've been thinking about this for 50 years,
03:21
and a year and a half ago I came out with the book
03:24
"How To Create A Mind,"
03:26
which has the same thesis,
03:28
but now there's a plethora of evidence.
03:29
The amount of data we're getting about the brain
03:32
from neuroscience is doubling every year.
03:34
Spatial resolution of brainscanning of all types
03:36
is doubling every year.
03:39
We can now see inside a living brain
03:41
and see individual interneural connections
03:43
connecting in real time, firing in real time.
03:46
We can see your brain create your thoughts.
03:48
We can see your thoughts create your brain,
03:51
which is really key to how it works.
03:52
So let me describe briefly how it works.
03:54
I've actually counted these modules.
03:57
We have about 300 million of them,
03:59
and we create them in these hierarchies.
04:01
I'll give you a simple example.
04:03
I've got a bunch of modules
04:05
that can recognize the crossbar to a capital A,
04:08
and that's all they care about.
04:11
A beautiful song can play,
04:13
a pretty girl could walk by,
04:15
they don't care, but they see
a crossbar to a capital A,
04:16
they get very excited and they say "crossbar,"
04:19
and they put out a high probability
04:22
on their output axon.
04:24
That goes to the next level,
04:26
and these layers are organized in conceptual levels.
04:27
Each is more abstract than the next one,
04:30
so the next one might say "capital A."
04:32
That goes up to a higher
level that might say "Apple."
04:34
Information flows down also.
04:37
If the apple recognizer has seen A-P-P-L,
04:39
it'll think to itself, "Hmm, I
think an E is probably likely,"
04:42
and it'll send a signal down to all the E recognizers
04:46
saying, "Be on the lookout for an E,
04:48
I think one might be coming."
04:50
The E recognizers will lower their threshold
04:51
and they see some sloppy
thing, could be an E.
04:54
Ordinarily you wouldn't think so,
04:56
but we're expecting an E, it's good enough,
04:58
and yeah, I've seen an E, and then apple says,
05:00
"Yeah, I've seen an Apple."
05:01
Go up another five levels,
05:03
and you're now at a pretty high level
05:05
of this hierarchy,
05:06
and stretch down into the different senses,
05:08
and you may have a module
that sees a certain fabric,
05:10
hears a certain voice quality,
smells a certain perfume,
05:13
and will say, "My wife has entered the room."
05:16
Go up another 10 levels, and now you're at
05:18
a very high level.
05:20
You're probably in the frontal cortex,
05:21
and you'll have modules that say, "That was ironic.
05:23
That's funny. She's pretty."
05:27
You might think that those are more sophisticated,
05:29
but actually what's more complicated
05:31
is the hierarchy beneath them.
05:33
There was a 16-year-old girl, she had brain surgery,
05:36
and she was conscious because the surgeons
05:38
wanted to talk to her.
05:40
You can do that because there's no pain receptors
05:42
in the brain.
05:44
And whenever they stimulated particular,
05:45
very small points on her neocortex,
05:46
shown here in red, she would laugh.
05:49
So at first they thought they were triggering
05:52
some kind of laugh reflex,
05:53
but no, they quickly realized they had found
05:55
the points in her neocortex that detect humor,
05:57
and she just found everything hilarious
06:00
whenever they stimulated these points.
06:02
"You guys are so funny just standing around,"
06:05
was the typical comment,
06:07
and they weren't funny,
06:08
not while doing surgery.
06:11
So how are we doing today?
06:14
Well, computers are actually beginning to master
06:19
human language with techniques
06:22
that are similar to the neocortex.
06:24
I actually described the algorithm,
06:27
which is similar to something called
06:28
a hierarchical hidden Markov model,
06:30
something I've worked on since the '90s.
06:32
"Jeopardy" is a very broad natural language game,
06:36
and Watson got a higher score
06:39
than the best two players combined.
06:41
It got this query correct:
06:43
"A long, tiresome speech
06:45
delivered by a frothy pie topping,"
06:47
and it quickly responded,
"What is a meringue harangue?"
06:50
And Jennings and the other guy didn't get that.
06:52
It's a pretty sophisticated example of
06:55
computers actually understanding human language,
06:57
and it actually got its knowledge by reading
06:59
Wikipedia and several other encyclopedias.
07:00
Five to 10 years from now,
07:04
search engines will actually be based on
07:06
not just looking for combinations of words and links
07:09
but actually understanding,
07:11
reading for understanding the billions of pages
07:13
on the web and in books.
07:16
So you'll be walking along, and Google will pop up
07:18
and say, "You know, Mary, you expressed concern
07:21
to me a month ago that your glutathione supplement
07:24
wasn't getting past the blood-brain barrier.
07:27
Well, new research just came out 13 seconds ago
07:29
that shows a whole new approach to that
07:32
and a new way to take glutathione.
07:34
Let me summarize it for you."
07:36
Twenty years from now, we'll have nanobots,
07:38
because another exponential trend
07:42
is the shrinking of technology.
07:44
They'll go into our brain
07:45
through the capillaries
07:48
and basically connect our neocortex
07:49
to a synthetic neocortex in the cloud
07:52
providing an extension of our neocortex.
07:55
Now today, I mean,
07:58
you have a computer in your phone,
08:00
but if you need 10,000 computers for a few seconds
08:02
to do a complex search,
08:04
you can access that for a second or two in the cloud.
08:06
In the 2030s, if you need some extra neocortex,
08:09
you'll be able to connect to that in the cloud
08:12
directly from your brain.
08:15
So I'm walking along and I say,
08:16
"Oh, there's Chris Anderson.
08:18
He's coming my way.
08:19
I'd better think of something clever to say.
08:21
I've got three seconds.
08:23
My 300 million modules in my neocortex
08:25
isn't going to cut it.
08:28
I need a billion more."
08:29
I'll be able to access that in the cloud.
08:30
And our thinking, then, will be a hybrid
08:33
of biological and non-biological thinking,
08:36
but the non-biological portion
08:40
is subject to my law of accelerating returns.
08:42
It will grow exponentially.
08:44
And remember what happens
08:47
the last time we expanded our neocortex?
08:49
That was two million years ago
08:51
when we became humanoids
08:53
and developed these large foreheads.
08:54
Other primates have a slanted brow.
08:55
They don't have the frontal cortex.
08:58
But the frontal cortex is not
really qualitatively different.
09:00
It's a quantitative expansion of neocortex,
09:04
but that additional quantity of thinking
09:06
was the enabling factor for us to take
09:09
a qualitative leap and invent language
09:11
and art and science and technology
09:14
and TED conferences.
09:16
No other species has done that.
09:18
And so, over the next few decades,
09:20
we're going to do it again.
09:22
We're going to again expand our neocortex,
09:23
only this time we won't be limited
09:26
by a fixed architecture of enclosure.
09:27
It'll be expanded without limit.
09:32
That additional quantity will again
09:35
be the enabling factor for another qualitative leap
09:37
in culture and technology.
09:40
Thank you very much.
09:42
(Applause)
09:44

sponsored links

Ray Kurzweil - Inventor, futurist
Ray Kurzweil is an engineer who has radically advanced the fields of speech, text and audio technology. He's revered for his dizzying -- yet convincing -- writing on the advance of technology, the limits of biology and the future of the human species.

Why you should listen

Inventor, entrepreneur, visionary, Ray Kurzweil's accomplishments read as a startling series of firsts -- a litany of technological breakthroughs we've come to take for granted. Kurzweil invented the first optical character recognition (OCR) software for transforming the written word into data, the first print-to-speech software for the blind, the first text-to-speech synthesizer, and the first music synthesizer capable of recreating the grand piano and other orchestral instruments, and the first commercially marketed large-vocabulary speech recognition.

Yet his impact as a futurist and philosopher is no less significant. In his best-selling books, which include How to Create a Mind, The Age of Spiritual Machines, The Singularity Is Near: When Humans Transcend Biology, Kurzweil depicts in detail a portrait of the human condition over the next few decades, as accelerating technologies forever blur the line between human and machine.

In 2009, he unveiled Singularity University, an institution that aims to "assemble, educate and inspire leaders who strive to understand and facilitate the development of exponentially advancing technologies." He is a Director of Engineering at Google, where he heads up a team developing machine intelligence and natural language comprehension.

The original video is available on TED.com
sponsored links

If you need translations, you can install "Google Translate" extension into your Chrome Browser.
Furthermore, you can change playback rate by installing "Video Speed Controller" extension.

Data provided by TED.

This website is owned and operated by Tokyo English Network.
The developer's blog is here.