This
is a great update on our understanding of just how our brains might
work. I have underlined parts and made a couple of notes. I am
myself very much a student of Einstein's thinking and have been well
served. What held him back on his agenda was an oversight of the
handful of eighteenth century Mathematicians. My efforts ended that.
Now we have to get past the mental block put up by scholarship.
This
is a must read. It will help you to understand what our future looks
like. Fifteen years is not a long time for the Baby Boom generation
and their children. Guide yourself accordingly.
I
have actually guided myself through the past forty years
understanding this acceleration in knowledge. Then best advice I can
give anyone today is to become conscious of its implications.
How Ray Kurzweil
Will Help Google Make the Ultimate AI Brain
BY STEVEN LEVY
04.25.13
Google has always been
an artificial intelligence company, so it really shouldn’t have
been a surprise that Ray Kurzweil, one of the leading scientists
in the field, joined the search giant late last year.
Nonetheless, the hiring raised some eyebrows, since Kurzweil is
perhaps the most prominent proselytizer of “hard AI,” which
argues that it is possible to create consciousness in an artificial
being. Add to this Google’s revelation that it is using techniques
of deep learning to produce an artificial brain, and a subsequent
hiring of the godfather of computer neural nets Geoffrey Hinton,
and it would seem that Google is becoming the most daring developer
of AI, a fact that some may consider thrilling and others deeply
unsettling. Or both.
On Tuesday, Kurzweil
moderated a live Google hangout tied to a release of the
upcoming Will Smith film, After Earth, presumably tying the
film’s futuristic concept to actual futurists. The discussion
touched on the necessity of space travel and the imminent resolution
of the world’s energy problems with solar power. After the hangout,
Kurzweil got on the phone with me to explore a few issues in more
detail.
WIRED: In the Google
hangout you just finished, Will Smith said he had a copy of your book
by his bedside because he’s been involved in a number of science
fiction movies. How do you view science fiction?
RAY KURZWEIL: Science
fiction is the great opportunity to speculate on what could happen.
It does give me, as a futurist, scenarios. It’s not incumbent upon
science fiction creators to be realistic about time frames and so on.
In this movie, for example, the characters come back to Earth a
thousand years later and biological evolution has moved so far that
the animals are quite different. That’s not realistic. Also,
there’s very often a dystopian bent to science fiction because we
can perceive the dangers of science more than the benefits, and maybe
that makes more dramatic storytelling. A lot of movies about
artificial intelligence envision that AI’s will be very intelligent
but missing some key emotional qualities of humans and therefore turn
out to be very dangerous.
What’s the key to
predicting the future?
I realized 30 years
ago that the key to being successful is timing. I get a lot of new
technology proposals, and I’d say 95% of those teams will build
exactly what they claim if given the resources, but 95% of those
projects will fail because the timing is wrong I did anticipate, for
instance, that search engines would start emerging. Fifteen
years ago Larry Page and Sergey Brin were in exactly the right place
at the right time with the right idea
You anticipated search
engines?
Yes. I wrote about
that actually as early as The Age of Intelligent Machines, in the
1980s. [The book was published in 1990.]
But did you predict
that you would be working for a company that started as a search
engine?
That’s exactly the
kind of thing you can’t predict. It would be very hard to predict
that these couple of kids at Stanford would take over the world of
search. But what I did discover is that if you examine the key
measures of price performance and capacity of information technology,
they form amazingly predictable smooth exponential curves. The price
performance of computation has been rising in a very smooth
exponential since the 1890 census. This has gone on through thick and
thin, through war and peace, and nothing has affected it. I projected
it out to 2050. In 2013, we’re exactly where we should be on that
curve.
[ and
that is why I immediately recognized the importance of google when it
came out and immediately developed my unique nom de plume 'arclein']
What are you working
on at Google?
My mission at Google
is to develop natural language understanding with a team and in
collaboration with other researchers at Google. Search has moved
beyond just finding keywords, but it still doesn’t read all these
billions of web pages and book pages for semantic content. If you
write a blog post, you’ve got something to say, you’re not just
creating words and synonyms. We’d like the computers to
actually pick up on that semantic meaning. If that happens, and I
believe that it’s feasible, people could ask more
complex questions.
Are you participating
in Jeff Dean’s program there to build an artificial “Google
Brain?”
Well, Jeff Dean is one
of my collaborators. He’s a fellow research leader. We are going be
using his systems and his techniques of deep learning. The reason I’m
at Google is resources like that. Also the knowledge graph and very
advanced syntactic parsing and a lot of advanced technologies that I
really need for a project that really seeks to understand natural
language. I can succeed at this much more readily at Google because
of these technologies.
If your system
really understood complex natural language, would you argue that it’s
conscious?
Well, I do. I’ve
had a consistent date of 2029 for that vision. And that
doesn’t just mean logical intelligence. It means emotional
intelligence, being funny, getting the joke, being sexy, being
loving, understanding human emotion. That’s actually the most
complex thing we do. That is what separates computers and humans
today. I believe that gap will close by 2029.
Will we get there
simply by more computation and better software, or are there
currently unsolved barriers that we have to hurdle?
There are both
hardware and software requirements. I believe we actually
are very close to having the requisite software techniques.
Partly this is being assisted by understanding how the human brain
works, and we’re making exponential gains there. We can now
see inside a living brain and see individual inter-neural connections
being formed and firing in real time. We can see your brain
create your thoughts and thoughts create your brain. A lot
of this research reveals how the mechanism of the neocortex works,
which is where we do our thinking. This provides biologically
inspired methods that we can emulate in our computers. We’re
already doing that. The deep learning technique that I mentioned uses
multilayered neural nets that are inspired by how the brain works.
Using these biologically inspired models, plus all of the research
that’s been done over the decades in artificial intelligence,
combined with exponentially expanding hardware, we will achieve human
levels within two decades.
Do we really
understand at all why someone’s brain can result in such an unique
expression of a human? Take the transcendent intelligence of
Einstein, the creativity of Steve Jobs, or the focus of Larry Page.
What made those people so special? Do you have insights into that?
I examine that very
question, in fact, with regard to Einstein specifically in my recent
book, How to Create a Mind.
Tell me.
There are two things.
First of all, we create our brain with our thoughts.
We have a limited capacity in the neocortex, estimated to be about
300 million pattern recognizers, which are organized in a hierarchy.
We create that hierarchy with our own thinking. I would not explain
Einstein’s brilliance based on him having 350 million or 400
million. We have approximately the same capacity. But he organized
his brain to think deeply about this one subject. He was interested
in the violin, but he was no Jascha Heifetz. And Jascha Heifetz had
an interest in physics, but he was no Einstein. We have a
capacity to do world-class work in one field. That’s
part of the limited capacity of the brain, and Einstein really
devoted it to this one field.
But lot of physicists
are devoted to their one field, and only one became Einstein.
I didn’t finish. The
other aspect is courage to follow your own thought experiments
and not fall off the horse because the conclusions are so different
from your previous assumptions or the common belief of
society. People are so unable to accept thinking different than their
peers that they immediately drop their thought pattern when it leads
to absurd conclusions. So there’s a certain courage to go with your
convictions. Clearly Steve Jobs had that. He had a vision and carried
it out. It’s that courage of your convictions.
What’s the
biological basis for that kind of courage? If you had an infinite
ability to analyze a brain, could you say, “Oh, here’s where the
courage is?”
It is the
neocortex, and people who fill up too much of
their neocortex with concern about the approval of their peers are
probably not going be the next Einstein or Steve Jobs. [
wow – I knew that there is a reason I have never given a rat's ass
regarding anyone's intellectual opinion - arclein ]
Is this something one
can control?
That’s a good
question. I’ve been thinking about that and also why do some people
readily accept the exponential growth of information technology and
its implications, and other people are very resistant to it. I make
the argument that hard-wired in our brain are linear
expectations, because that worked very well 1000 years ago, tracking
an animal in the wild. Some people, though, can readily accept the
exponential perspective when you show them the evidence, and other
people don’t. I’m trying to answer the question, what accounts
for that? It really isn’t accomplishment level, intelligence,
education level, socio-economic status. It cuts across all of those
things. Some people’s neocortexes are organized so that they can
accept the implications that they see in front of them without
worrying too much about the opinion of others. Can we learn that? I
would imagine yes, but I don’t have data to prove that.
Since we’ve been
talking about Steve Jobs, let me bring up one of his famous quotes,
from his speech at Stanford. He said, “Death is very likely
the single best invention of life. It’s life’s change agent.”
You are very famously trying to extend your life indefinitely, so you
reject that, right?
Yes, This is what I
call a deathist statement, part of a millennium-old rationalization
of death as a good thing. It once seemed to make sense, because up
until very recently you could not make a plausibly sound argument
where life could be indefinitely extended. So religion, which emerged
in prescientific times, did the next best thing, which is to say,
‘Oh, that tragic thing? That’s really a good thing.” We
rationalized that because we did have to accept it. But in my mind
death is a tragedy. Our initial reaction to hearing that someone has
died is a profound loss of knowledge and skill and talents and
relationships. It’s not the case that there are only a fixed number
of positions, and if old people don’t die off, there’s no room
for young people to come up with new ideas, because we’re
constantly expanding knowledge. Larry Page and Sergey Brin didn’t
displace anybody– they created a whole new field. We see that
constantly. Knowledge is growing exponentially. It’s doubling
approximately every year.
And you think that
dramatically extended life is possible.
I think we’re
only 15 years away from a tipping point in longevity.[
if that – arclein ]
No comments:
Post a Comment