you know how much the industry of the
artificial intelligence about 15 thousand 700
billions of dollars with bar b
barbarity asked him to top million
down so it does not surprise me that what
more think you can generate
international tension around
artificial intelligence and can we
reach a third world war what
that I’m surprised a little more
is that I think the button
this time he will press a machine not
it’s going to tighten human and surprise me
because we are still really at
principle of artificial intelligence
there’s a lot of myth around because
come as a tsunami in the past
five years and then it scares us and
we started to project the latter
five years forward and we say that
is what awaits us the return of the
corner
Well, let’s see a little bit
maybe we do not wait so much so
comes and comes with power to stay
it’s a bit of timing, not time
times that are handled what maybe not
be so realistic says that three types of
artificial intelligence is the
specialized to the generalized the
specialized intelligence in which
we are right now the one that scares us
end do not stop being machines
focused on very particular topics that
make us evolve example
we want to win at chess garry
kasparov we tell di blu we want to play
to such as a machine or image
we want to do very cute series netflix
so the problem is that it takes
mythified about it the official boss and
if we scratch a little we see something very
important and that is based not on
five or ten years of research
this rate in 50 years of research
maybe I was a little bit and the trap
we are getting into the car that they have
mounted others for a long time
very serious, very educated people who have
led to where we are that is very
good but careful
the difference is that before they
they had almost no data to play with
without us we no longer live
only in three living physical dimensions
in four dimensions the fourth emission
the last of them the digital when
we leave fingerprint to the average day
tour
of data that’s a lot more still the
super cheap technology right now the
computers that we are buying today
eight years ago it was unthinkable
imagined 18 28 years ago the only thing that
has happened is that now profitable for
Business
For the last 56 years, it has been very
profitable and why it is appearing
massive form
I love everything from one point of view
of economist
if it makes economic sense yes I think
that you can balance what you can
keep moving
that yes another trapdoor that has is
that really is not intelligence to have
we are humans
you give me a lot of data you give me a lot
machine and I actually what I’m going to
do not be cognitive intelligence
relationship to the similar to the human I’m not going
to do brute force
there he took out as always we do not
brute force that works in any
case but it’s not smart if it were
smart the foundation of bill and
Melinda Gates did not have to spend 80
millions of dollars not for education
not for food not to increase the
fingerprint of women in countries
Developing
it’s that important to us 80
million because otherwise why not go
in the photos and they are not in the footprint
digital the machine can not have them
account and if you do not take into account that
is going to be racism sexism them
inclusion problems are at
agenda in this kind of
intelligence and that is one of the great
problems that we have that is not
smart to the next problem is that
the generalized intelligence that is like that
cognitive rational would not have 50
years
therefore it will take time to arrive and
we have to deal with the problems that
I told you before and
sterilizations of grace because in
reality once the machines are
more or less as smart as
us but at the same time have connection
to the servers and data are going to pull
brute force and they will overcome us
immediately
exactly like we have a
brute force and where would we get to
the superintelligent already and here is where
comes the biggest controversy because the
superintendence by definition not the
let’s understand is that a problem then
for me it’s not a problem it ‘s part of
the evolution
evolution has led us to manage
difficult situations therefore the
artificial intelligence
you have to have it or love it you have to
manage it then I keep wondering
I say good then we are doing it
well to manage what now that yes
that I start worrying a little bit more
I would say no because what is
propose are to force rules in favor of
men type the three laws of simof
in the code but when your forces’
things in the code because you have errors
humans and you have human errors
forced
this is hacker attacks and to me this
I’m also very scared of the
artificial intelligence does not scare me
I’m afraid of humans because we can
make mistakes and why can a
goal is going to see what happens if I remove
security to the typical hack
we could already and this is my question no
is murcia -30 of a one management
this one to super intelligence a little bit
more robust because I said that
It will take time to arrive but when it came to
move directly to your intelligence
we’re not ready where it comes out
half these rare concerns
auxerre roosters and a trader I was in
london calmly doing 35 mis
56 screen act like the movies do not
bunch of graphics bunch of
software and algorithms that told me more
or less what I had to do
and make decisions based on that and
I was doing very well and is quadrupling
the benefit less than a year has gained
a paste and that’s where I realized
and wait because this all if you
you set is work very much the fourth
dimension everything very digital and therefore
there is nuanced and the fact that
come the pasta was not so cool
that brought me a lot of risk
because the company would be willing
replaced by a cheap robot and that was
the key
so I did what nobody wanted to do
in 2009
I left the industry to make a
PhD in computer science
the ones he talks about the most
that in his tweet and there I was 29
years a scholarship since I only have to
thank my wife for the support they have given me
given always and a very clear objective of
break out from breaking in instead of being
collapsed did not wait did not want to wait
since it supposed me a machine and since
then if I’m going to break that machine
that was how I would want you to give me
break and I’m not going to leave you
programmer who does it for me I do not know
if you understand me there but that means
that I do not trust others and what else
I’m worth the environment
take a little control is not going to be
that is worse than not take the problem
the first problem because I found
were the doctrines of intelligence
artificial that there was not like all the
life there are two extremes were those of my
cronies of the job they said hears no
create the machine we will never be able to
replace
this is a chick and in fact like me
try to systematize I’m going to boycott the
project to deny me the companies the
artificial intelligence seemed to me
ballasting is absurd and therefore not
it’s an economic balance and I to me
style of economics and equilibria
in economics
I think they are the ones who are really
Moving the world whether we like it or not we
like in the highest percentage of
cases so I did not believe it and I insist
salt and what luck what luck because 2017
It was the year that has begun
Fashion for throwing workers
expensive financial instruments and replace them with
cheap machines blacrock one of the
most powerful agents in finance already what
dozens ago and I hope that the
since I had in the next two or
three years decimated
in fact in the last job I had
It is a trading platform in which
I proposed efficiency improvements of x
20 x 20 the platform took the
award for the best trading platform
of the industry in 2016
and this in a Spanish commercial bank
we won jp a merrill lynch students
this is possible there are people with criteria and
out can make some improvements
impressive and that can affect us
a lot to everyone
then there was the other end obviously
the one of my buddies of the university
those who was beginning to know what
did the algorithms they can do
absolutely everything
you give me the data and I start to see
If the machine and that replaces
who is needed above all the works
of the fourth dimension
honestly that despite everything seems
that is a little more meaningful
I have much more than what the other
end but I deny deny
professional in that rotund way
it seems absurd to deny ourselves
themselves in that way seems absurd to me and
that is happening was what was
doing
blackrock so it can not be no we
we can allow companies not to
can afford the luxury of digitizing
both men
because we our criterion
neocortex we do not leave it in our
fingerprint – not completely this
you have to explore it a little more like that
that looks at the midpoint
the midpoint was to increase the humans
with machines
the futurists that is very good what
they say a lot of imagination is more
science fiction that science and is very
expensive is not today if in the future our
brain will be communicating by wifi with
the servers and we have labor data
if everything is fine
yes but but it will come I want something
more short term what really we
can help and I also say one thing if
you increase with the machines what
you are doing is increase your
dependence on the machine therefore
the survival and the
probable of your presence of the machine
does not give the name and I want it backwards
among its relatively simple the
The only thing I have to do is increase
the intelligence of the machine with the
of man and so the machine depends
of man and man survives very
easy to say now you have to pass it to
science and not science fiction and there
is my thesis and my challenge in fact
get there what is in one of the
chapters because before and without news
of algorithmic trading of sciences
computational
but, well, the thesis started like that, I’m going to
talk about the two worlds not the ones
guys who do not believe in machines and
guys who do not believe in humans or
machines that humans here have
to imagine a kind of black box
a mess of mathematical algorithms that
given the data makes all the sense
but they are not interpreting everything
it occurred to me it was good because we’re going
to put a robot in the middle and even a
kind of translator’s api between
human intelligence and the robot is so
complex so here is the
key is the human
the professional who determines that he is going to
learn that robot from the machine
the machine to start moving with
a certain parameterization starts at
move starts to act and the robot
start learning what he has said
the human and that will have told you human
his idiosyncrasy all the criteria that
does he discuss with his boss with his
companions
before this is here apart creativity
of the user that makes that there may be
many different propositions for the
same very human task
in the end what we have is the robot
is learning from the machine and the
what he does is observe the
brain can compare us with the
his for example what he admired was
good if he if the machine starts to
earn a lot of money that I do if it starts to
act as randomly the
profitability as random I do and
if you start losing money that I do and I
I found robots that learned a
little psychopaths was not you follow you
follow you follow you still there to
final accepted and the truth that he won
a lot of money but that did not have
meaning to me is not using my
systematic user experience to
basis of calibrating the robot many times
the machine and let it behave
freely in the market was originating
different brains from different robberies and
in the end the only thing I had to do
was to look for one that was more like
me this is my avatar
then this is all very well does not
I left my avatar managing the machine
calibrated by this system just missing
that basically worked the
day defend the thesis that I have to
explain that there are two types of data not
are the sample data that are
where is the machine learning
et cetera etcetera and then there are the
new data
the new data the real world then
when you are going to defend the thesis what
you want is to have good results
in both worlds obviously and that’s how it came out
quiet and little brave because my
results in the world of training
of the machine were worse this is the
machine was better than man plus the
machine and being a bit weird
calm calm because it’s really the
world where the machine is cheated
to the lonely
this is the world of brute force
where you have to compare yourself with what you are with
humans are not a laboratory is the
real world
and that’s how I did it in that world
real
my approach was not only novel
but it was much better than
previous
so he had managed to prove that
he
man could inject knowledge the
machine and to which she tells this
knowledge we got machines
augmented with names and all
simply because I started a new
starting point I was not worth those who
there was
I started my own and yesterday to vouchers and
the machine plus the man is better than
the machine means that you earn more
then there is what for me is how much
my avatar so I had got
find an income that not what I
worried was that it made me leave the
industry in the medium term
now my avatar is worth money and the company
is willing to pay for me for me
avatar
I would have to see how long I am
working for it you will have the power
of negotiation of each one I do not know
but that avatar is worth money that I’m
insurance
and the most interesting thing is that there was
more consequences
unintentionally it had increased or
accelerated the world of those of the
humans augmented by machines in this
human cases augmented by machines
augmented by humans before when I said
it was good to open the skull expensive one
etc. etc. etc. it was expensive but
I had focused on just the
three physical dimensions but we’re
speaking the fourth dimension that there is of
the fourth dimension the court mentions
radiant simple increase us and half
going to the bank saying there to devise
I want the abbot to invest I want the
my neighbor’s fifth avatar and for
save my cousin’s avatar that is very
good rano and easy and if we have many
more digital dimensions
we can have a lot more people creating
avatars designing them proposing the
and in the end we have a kind of
avatar industry is that I see myself
working on for example my case
making avatars of cybersecurity and
algorithmic trading because there is not and
perceiving the income of sollamas here
something important is that if you see
managing time to put inject the
knowledge of the professional in the
Professional machine means that
those more than six million unemployed
which it is expected in 2020 affected
by artificial intelligence can
have an opportunity because they are from a
same generation and that’s what more
My concern is the way it works
all this we need to talk about the
management
the management of the ti also worries
full of the solution
bill and melinda instead of forcing the
rules what you have to do is accommodate
the in the knowledge of the machine
so you have a natural learning
then what we can not forget
are avatar type examples or the air is
that is one that we are creating now in
Our company also to try
demonstrate that he the man has value
for the machine and then when the
machines start thinking what happens
What do they do with us?
that we have increased for the moment
received some hacker and takes away those
signals would also have to those
lines of code would also have to
kick us out of the fingerprint
that we have left I would be much more
difficult so it would be much more
robust in more examples of
cybersecurity of other things but
I think the idea has become clearer or
less until here at the end simply
avatar is just an example of
of lateral thinking of creativity
for me creativity is the most
nice of the human is what makes us be
humans
creativity is what has helped us
to manage changes before and
adapt as a species
so simply if we want to leave
well standing from this world of the
artificial intelligence about
intelligence that we do not understand
we just have to remain
creative
we have to remain human
Thank you
[Applause]
me
”
Be First to Comment