There was a conference out of Sylmar that
made headlines around the world. The brightest
minds of artificial intelligence converged
onto Sylmar and a reporter asked them a question"
"When will this fabled singularity take place? 
When will the machines take over?  When will
machines become smarter than us?"
Well the answer was quite interesting. Among
the top people assembled in one place the
answers were anything from 20 years in the
future to 1,000 years in the future—with
some AI experts saying never. Some people
put it at 2029. They even give you an exact
date.  2029, that's going to be the moment
of truth that one day a robot will wake up,
wake up in the laboratory, look around and
say, "I am aware."  "I'm just as smart as
you."  "In fact, I could be even smarter
if I put a few more chips in my brain." 
Other people say: "Not so fast, not so fast
because Moore's law is going to break down."
The reason why many people are so confident
about this prediction of the so called singularity
is because of Moore's law that computer power
doubles every 18 months and it's a curve that
has held sway for 50 years. If you go back
100 years back to the time of mechanical hand-crank
computers, put that into Moore's law and you
still get a nice fit, so believe it or not
Moore's law has been in operation for about
100 years, going back to hand-crank calculators
with computer power doubling every 18 months. 
Well can this go on forever? And the answer
is no because eventually physics takes over
and that is physics says that silicon is unstable
at the molecular level. Transistors get so
small, so powerful and they generate so much
heat that the silicon chip melts and electrons
leak out because of the Heisenberg Uncertainty
Principle.  You don't know where the electron
is anymore. Therefore, we physicists are
looking for replacements for silicon. The
post-silicon era will be about 10 to 15 years
in the future.  Silicon Valley could become
a rust belt.  Think about it, every Christmas
your PCs, your computers, your gadgets will
be just as powerful as they were the previous
year and then the question is are you going
to buy?  Are you going to buy any more computer
products for Christmastime knowing that they're
just as powerful as they were the previous
year?  Probably not. Which means that the
computer industry could begin to shake as
a consequence.
So we physicists are looking at optical computers,
quantum computers, DNA computers, protein
computers, all sorts of different kinds of
architecture down to the molecular, down to
the atomic, down to the microscopic realm,
but none of them are ready for primetime yet.
So my answer is I don't know. All I'm saying
is there is vast uncertainties in projecting
Moore's law into the future. However, I would
say by end of the century it is definitely
conceivable that if we work out the technical
problems we might be able to create machines
that are as smart as us. Right now our machines
are as smart as insects. Eventually they'll
be smart as mice. After that they'll be smart
as dogs and cats. Probably by the end of
the century, who knows, they'll be as smart
as monkeys. At that point they could become
potentially dangerous because monkeys can
formulate their own plans. They don't have
to listen to you. They can formulate their
own strategies, their own goals and I would
say therefore at that point let's put a chip
in their brain to shut them off if they get
murderous thoughts. Isaac Asimov advocated
something like that with his "Three Laws."
I say hey, put a chip in their brain to shut
them off if they start to get murderous.
