I'm Daniela Rus, and I make robots.
I'm a robotics researcher at MIT's Computer Science
and Artificial Intelligence Laboratory, or CSAIL, and also
the director of CSAIL.
And today I will be moderating this panel
on the future of computing.
So in this panel, we hope to explore with you
some of our latest insights and the challenges
that our researchers are tackling
in the field of computing.
We also will discuss what the future of the digital landscape
might look like and how computer science research will
make a difference in how we live, we work,
and we play in the future.
I will tell you more about CSAIL in just a minute.
First, I would like to introduce our panelists.
Tim Berners-Lee probably needs no introduction,
but he invented the world wide web 25 years ago in 1989
when he was a researcher at CERN, which is the European
Particle Physics Laboratory.
And today Tim is a crusader for a world wide web that
is of even greater benefit to humanity
to easy access to information and through providing
more transparency on how governments operate.
Andrew Lo was named by Time Magazine
one of the most influential people
of the year for his innovative approach to financial markets.
And Lo is an economist by training.
He is a faculty member at MIT Sloan School of Management.
He directs the laboratory for financial engineering,
and he's also principle investigator at CSAIL.
And Russ Tedrake can make robots that
fly like birds, run like ostriches,
and these days, he's directing his innovative approaches
to creating robots that can assist in search and rescue.
Russ finds his inspiration in nature,
and today he will tell you about what
we might expect from the field of robotics and from robots
into the future.
So I will share with you just a few words
about CSAIL, the organization we all belong to,
then we will hear from the panelists.
And at the end of their brief speeches,
we will open the floor for general questions.
So first of about CSAIL.
We have over 100 senior researchers.
That means faculty and principle investigators.
And we have over 500 students.
That means postdocs, graduate students,
and undergraduate researchers.
And we are the largest interdisciplinary research
laboratory at MIT.
We were founded 50 years ago as Project Math,
and since then, our laboratory has
played a very important role in the technological revolution,
starting with the early days of time-sharing computers
to the birth of data encryption.
Our objective is to teach our students
to become world-class computer scientists, to be big thinkers,
and to become constructive innovators,
that we take the field of computing to the next level
and also take on some of the world's
most pressing challenges.
Our researchers include robotic researchers like myself
as well as experts in all fields of computing--
security, mobile, wireless, robotics,
artificial intelligence, databases, theory.
You name it, we have experts in the field.
And because of this diverse range of interests,
we are very excited and very well equipped
to take on interdisciplinary challenges.
And all together, we imagined a world
where computing can turn the world into a better place.
We imagine a world where physicians
can use the most advanced knowledge on genomics
to engineer better treatments that are individualized
We imagine a world where data can help us all
make better decisions by predicting
the consequences of our decisions
and where the most recent advances in encryption
research, for example, and differential privacy
and homomorphic encryption put people
in charge of their own data.
We also imagine a world where robots
will be as common as smartphones and cars
and where robots can do everything,
from helping with search and rescue activities
to folding your laundry, which is my least
favorite chore at home.
So we really believe that computing
has a lot to offer in the way our world will
be shaping the future.
And with this, I would like to stop and turn it
over to Tim Berners-Lee, who will talk to us
about the web we want.
So it's great to be here.
Thank you for coming.
If you have a--
put your hand up in the seat next you, if you have a seat?
So the awesome seats over here, if you need them.
So thank you all for coming.
Great to be here.
Thank you back at South by Southwest.
It's 25 years ago that I originally
wrote a memo proposing the world wide web project.
And so-- I want to just check all of the other [INAUDIBLE]..
So we've got three the world wide web project
I wrote a memo about.
That was 25 years ago.
I published a memo, which has got these scribblings
on the front about reorganizing information.
And for the first three years, it
was actually quite difficult to persuade everybody
that it would be a good idea.
It wouldn't be too complicated, and that they should give out
their own proprietary system to open it up,
the psychological thing of getting
open about putting some of the things on the web.
But there were people who got it.
And just like now when you look around, South by Southwest,
you see a lot of people with a twinkle in their eye
because they get the idea of how the world means everything
in the media.
25 years, I want to ask what's the web we want.
So step back very easy for us to use the web every moment.
It's only when you're under the Mubarak regime
in Egypt turns the web off or other things happen on the web
that you realize that actually, it's not just
like the air we breathe.
You actually have to step back and think
about what is behind it and think about whether it's always
going to be there and think about the properties
that it can.
But it got us this far now we're going to continue to happen.
There are many threats to rights from the web.
Recently people have started to talk
about human rights on the web.
Many human rights were established like the pre web
builders work, from what you call and you
talk about the Magna Carta in 1215,
and we talk about the US Bill of Rights.
If you're brought up in South Africa,
then you're probably very proud of your new constitution.
It's put together with a very good bill of rights,
but, hopefully, to the right, none of them
mention the internet.
And to some extent, we can just demand
those rights we have shouldn't exist on the internet.
But in fact, it turns out that when you go from there,
then it turns out that some of the things that you
wanted like the ability to connect to anyone--
Yeah, but that's not really.
That needs to be called out because somebody
could tell you, yes, we're giving you connectivity.
You couldn't get these people, connect to our partners
through our partner or companies.
[INAUDIBLE] and say, oh, I want to be
able to connect to any new upstart company.
I don't want to just watch movies,
I want to watch movies with anybody
who has started making a little startup, making a film company,
and so on.
There are a number of countries who
are putting internet regulation legislation, ideas
about the internet being a right,
make sure it's something obviously the country
There are various declarations people got out there,
the Declaration of Internet Freedom is one.
You can check those out.
You can sign them, and share with the people somebody
you might know, somebody that might
be hanging around the South by Southwest who've signed
that declaration of freedom.
So what do we mean when we talk about rights on the internet?
Well, of course, we're talking about those thieves or things
like expressions, free expressions.
That means I don't want to be censored.
I will need to be able to get to what I want.
And also, as a person I need to put myself out
there so that anybody else can see it in any country.
The web blew up as a non-national thing.
At this point, it's an international thing.
International sounds like nations getting together.
Web grew up irrelevant of nations.
It's ended up, the internet was connected up,
and the time and the server when you connect the web browser
to the web server, they don't care about national boundaries.
That has had some really sweet side effects in a way,
but its international community has grown up.
Censorship blocks really messes that up.
It doesn't happen in every country.
Happens in different areas.
We have to keep track of that in the future,
but very important to be aware where it's happening.
Just ability to access the web.
So about now, the number of people
who got internet, could access the web technically
is around 40%.
So that is great for the 40%.
And the web internet is becoming more and more powerful in lots
of ways, every moment.
And this component is more and more cool,
and it's more and more empowering for people
to do their own medicine, have their own [INAUDIBLE]
later, in another session you'll be excited about.
All these new things you've done that
make it more and more powerful and makes
the gap between the 40% that do have
and the 60% that don't have it even bigger.
Obviously, for people who don't, many things get in the way.
Sometimes it's poverty, sometimes it's illiteracy.
Some obviously, are grouped for mobile.
In Saharan Africa, in developing countries, in general,
it's mobile, the mobile web, which is taking over.
Critical things are the cost of the gadget that
can do web and the cost of the beta plans,
not necessarily broadband, but some sort of broadband,
something that will let me do email, be part of the browsing
the websites, see how they do information society
and recreate it myself.
I want our people just not just to go and see the extent of we
put together by a videographer, then get on, make it themselves
in their own language so you see more and more stuff
in foreign language.
So one of the properties of the web is that it's open.
That means lots the things.
Open means lots of things.
Teach people part how to build some standards.
Obviously, many of you may have worked
on building web standards following the latest
versions of HTML5, the APIs.
And so keeping those open, keeping it royalty free.
A really crucial point, the place
I worked for in Switzerland, CERN,
that physics lab are taking months.
We persuade them to actually provide
a document, because you know what,
CERN is not going to charge a lot for the web technology.
That was a really crucial point of uptake.
Since then, it was the point in the consortium's history
where we realized we had to get all consortium members who
are involved in what we do.
They also, when we've come to this fact,
we aren't going to try to get royalties for people
who have implemented them.
That's the main part of it.
Now, the corporate part of being a open support
is the neutrality of media.
Censorship is when it's block.
Net neutrality, it's not just not being blocked
or parliament neutrality, many challenges,
but you don't mess with it.
The ISPs, for example, the famous one in Holland
that slowed down Skype because they noticed
the Skype was starting to eat into their telephone revenue,
Turned that around pretty quickly.
In just legislation, you can't do that.
In America, the convention is that internet service providers
provide you with connecting to everyone you want,
and they don't slow down packets go
into their commercial enemies, but they're always
under such tremendous commercial pressure
to do so because they can make such a mint if they
can actually get a stranglehold on you as a consumer
and just connect to who you connect to.
So this is the Central American illustrating
the picture of the happy ion couple
underneath all the people or all the demons trying
to get up the workings of these net and use it, some of you
are on there.
So many times it's important, many think,
I think it's important for the commercial world,
for an open market, of course, it's
important for the marketers, it's important
that you can go to any physical site you like.
It's also important for innovation, the fact
that you can build a new website and very quickly one person
providing the program, two people promote this program.
They can send their package through the cloud
or the internet.
By the way, we're using the cloud diagram
for the internet on before the cloud became proudly into our--
So why you care?
Because you don't care what's inside it.
You can write a program, she can write program,
they can communicate, and a new innovative protocol is born.
New innovative system is running over the web,
a new social media is spreading across the web, for example.
That's really important for innovation,
so that's why see that happen in the next 25 years.
All this, of course, thank you all for coming here.
We know you had also, to go where you had streaming live,
and the obvious moment, this is last year,
there's been a lot of focus on privacy.
It's a shame that it took us some revelations
to get us all thinking about the internet.
Obviously, we have to think about to what extent
we give the agencies that's supposed to protect us,
the ability to sample the internet, spy on the internet,
how I think the really interesting question is
how we build systems that will hold them accountable.
Checks and balances traditional in the Madisonian government.
Checks and balance is equally built like they've never
been built before.
We need very strong agencies accountable to the public who
have the power to go in and investigate the investigators,
and we don't know how to build that yet.
When it doesn't happen, it's really important
that we can protest.
We've got our website can go as Wiki--
actually, went back, not the other.
So that was when Wikipedia went back
is when the SOPA legislation was going to go through.
Also, people went out into the streets
protesting SOPA in the US, protesting ACTA in Europe.
ACTA was an agreement which was going
to give too much power to copyright holders
and take away freedoms from individuals.
So this sort of stuff is important to be
able to do that.
There are lots the things you have
to be able to protest about, lots of things
when actually individual voices have
to be added up so that this thing is kept open and free.
Those of you who were here last year,
we were talking about Aaron Swartz's life
and how he was a huge part of software so far.
Let's keep doing, in his memory keep doing
the things what would Aaron do.
So those are some of the things which I think.
Now, back on my agenda for the web
we want, so there is a web we want node called Web at 25.
Well, web 25 that's going to be live.
On the top of March, which is another lecture of the web.
I want you all to think about what sort of world web
I want you to not just think about it.
I want you when necessary to take action.
I have faith that by the end of the year,
we know what legislation is properly enacted,
in which countries.
In all of your countries, what sort of pressure
is going to be put on the companies necessary.
What sort of spirit do we want behind the web
that we're going to have for the next 25 years.
Thank you very much for your time.
So next, I would like to turn it over
to Andrew Lo, who will talk to us about big data, big brother,
and the financial crisis.
Also, while we get the slides set up,
I want to start by thanking Daniela
Rus and my other colleagues at CSAIL
and the organizers of South by Southwest
for inviting me to participate in this great panel.
I have to admit that as a financial economist,
I feel a little bit out of water and intimidated
because if you look at the composition of the panel,
we've got two of the leading roboticists of our generation
and the inventor of the internet and me.
So who ordered that?
But when I told a friend of mine that I
was coming to South by Southwest to be on this panel,
he was very jealous and said, why are you going?
Isn't South by Southwest all about disruptive innovation?
And I have to say that absolutely,
and financial economists belong here
because at least the first pact we've got down.
Finance has been incredibly disruptive.
And so what I want to talk about is, what do we do about it.
How do we deal with the challenges
of a financial crisis, and particularly
big data and the privacy issues that Tim mentioned.
So I'd like to start by talking a bit about what
we know of the economy.
And it turns out that there's actually quite a lot
that we know if we actually take a look at the data
that we have.
I can break this up.
Over the course of the last several years,
the economy has been going up and down,
but over the last few months, it looks like things are actually
looking pretty good.
So for example, if you take a look at recent announcements
by the Bureau of Economic Analysis--
I'll actually bring up the slides in just a minute.
There we go.
So what do we know about the economy?
Well, as I mentioned, we've got a lot
of data coming in, so we know that inflation is about 1/10
of a percent; unemployment, 6.6%; GDP growth, 2.4%;
non-farm payroll, 113,000; housing spark, 180,000.
The fed balance sheet is $2.6 trillion.
All of these data tell us where we are in the economy,
that basically things are getting better,
but we've still got a ways to go.
However, there is a huge part of the economy
that we actually don't know much about,
for example, hedge funds and the shadow banking
system that were central in creating the financial crisis.
What do we know about that?
Well, let me show you the data that we've
collected about hedge funds and the shadow banking sector.
This is what we know about it--
And the reason we don't know anything about it
is because there is actually no legal requirements
for hedge funds or the shadow banks to report.
They're not even actually under any regulatory jurisdiction.
And so the old adage goes, you can't
manage what you don't measure, applies to this sector.
And clearly, we want to understand
what these data are telling us about that part of the system.
But at the same time, a lot of these financial institutions
have argued, well, we're developing
proprietary trading strategies and investments
for our shareholders.
And if you believe in private enterprise,
you don't want to force us to disclose.
After all, you don't disclose Coca-Cola's formula,
you want to give people an incentive
to develop new innovations.
So the question is, is there some kind
of a compromise between privacy and total transparency?
And it's pretty clear that there are certain things that
have to be kept private.
We don't want total transparency.
The answer is, yes, there's actually a compromise,
and it has to do with a new technology,
not so new to computer science literature,
but new in its applications to finance.
And the technology is known as secure multi-party computation.
Let me tell you what it is with a simple example.
I've always wondered what the socioeconomic statuses
of participants at South by Southwest.
So one of the things I'd like to know
is what the average salary is of people in this room.
So let me start by asking, anybody
you want to tell me what their average salary is so we
can start computing an average?
Not so much.
Well, we obviously acknowledge that salaries
are a quite confidential thing.
So let me tell you what my annual salary is
so we can start trying to compute
the average in this room.
But I'm not going to tell you what my salary is by itself.
What I'm going to do is to tell you
what my annual salary is plus a random number that I
choose and add to my salary.
So my salary plus the random number is $15.
So hopefully, I added a negative number to my salary.
So I'm happy to share that with you, $15.
And I'm going to give that number to somebody else
in the audience, say, Alice.
And I'm going to ask Alice to take that number, $15,
and add to that her salary plus her random number
that only she knows and she chooses.
And so Alice does that, and the sum is maybe $37,000.
And Alice takes that number and gives it
to somebody else who adds his number and his salary
and passes it along to somebody else
and to somebody else and somebody
else until the last person in the room, Bob, gets the number.
And Bob adds to it his salary and his random number.
And at that point, the number turns out to be $15 billion,
and Bob gives it back to me.
I take that number, $15 billion, and I subtract from that
my random number.
I give it to Alice.
She subtracts her random number.
She passes it and gives it to the next person who
subtracts her random number and so on and so forth until Bob
subtract his random number.
And at that point, when Bob subtracts his random number,
what Bob has is the sum of everybody's salary, which,
when you divide by the total number of people
in the room, which everybody can see, and is public information,
yields the average salary of everybody in the room.
Nowhere did anybody have to reveal his or her salary,
but at the end of the day, we actually have the average.
Now, of course, this algorithm is a pretty simple one
in the sense that if I got together with Alice,
I could actually cheat and figure out
what Bob's salary is, but there are
more sophisticated algorithms that
allow no cheating to go on and yet permit us to calculate
numbers like average salaries or average loan portfolios
or average concentration indexes of credit default swaps,
things that will tell us what the systemic exposures are
for the aggregate economy without jeopardizing privacy.
So let me conclude by pointing out
that with the right technology, transparency and privacy can
both be achieved.
You don't have to give it up as long as you
have the wherewithal and the right tools
to be able to deal with these kind of challenges.
Thank you, Andrew.
Next, I'll turn it over to Russ Tedrake, who
will tell us why the robots are coming here,
and how they're coming here.
So it's sort of a cheesy title, The Robot Are Coming.
I had to think a lot about whether I
was going to use the title, but I
feel like there's never been a time before where I would have
felt right using the title, and there may never
be a time again where I'll feel right
using such a cheesy title.
But let me tell you why.
Robots are trending in such a way
right now, just in a massive proportion
that this felt like the one time where I could get away
maybe using the title.
And maybe it's not even their coming,
but they're actually here.
And I want to make sure you guys know the extent to which they
And it happened with a lot of progress
in academia and industry.
And then it got quickly spurred on
by this relatively massive investment
first by the government.
So there's a National Robotics Initiative
that happened three years ago now where the robotics
community got together, formed a robotics caucus,
and the Congress made robotics a national agenda
and produced research money for academic research in robotics.
There's another big huge another $80 million or so
coming from DARPA that was injected into building
robotic hardware in the US.
So you're just seeing many, many more robots
moving around, walking around in the United States.
It's a worldwide phenomenon.
All over, the European Union has massive funding for robotics.
Japan and Korea have been leaders in robotics,
even at MIT where we've been spending
our precious few junior faculty slots filling in
with roboticists because we know that robotics is hot.
It's hot now.
Google just shocked the world in December
saying that they purchased eight of the big robotics companies
in the last six months.
So why did all this happen?
Why is it all so much energy?
Because in the last five or six years, robots got awesome.
You must have seen this before, right?
So this Wild Katherine Boston Dynamics.
This is UADs that are flying around individually doing
amazing things, but collectively doing amazing swarm
computations that scale.
That's at UPenn.
This is an airplane.
Airplanes don't normally land on a perch.
This airplane would developed at CSAIL.
It can stall its wings and like a bird and land
on a perch like a bird.
Maybe it will land on power lines
to triple charge its batteries then keep flying.
And in fact, it's so robust that now we
have Joe just throwing airplanes at wires all day long.
Manipulation capabilities of robots,
the hardware technology that we've got now,
this is from Germany.
This just showcases the robotic arm technology
that is mature, the robotic hand technology
able to do relatively advanced manipulation paths with flair.
The sensing capabilities of robotics
has been just phenomenal progress.
This is a connect sensor that one of John Leonard's
colleagues, Tom Whelan, is walking through his apartment
once with a Microsoft connect sensor, prime sensor.
And after just a single traverse of the apartment,
look at the kind of map he can pull out.
But robots can see now.
They can't necessarily understand what they're seeing,
but the geometric representation of the world problem
is on a wave.
Robots can see the environment.
And now robots are going out and taking all these technologies
and putting them together and solving hard problems out
in the real world, but one of the ones
that I've been involved in the last year
is this project by DARPA, a competition
by DARPA, where we had to field robots.
They didn't have to be humanoid robots,
but used the humanoid robot that had to go out and handle
a disaster response scenario.
So this is the scene where they put us
in front of a debris pile.
We had to clear debris and get through--
to imagine the door at the end of the debris.
But the point is, we didn't set this up and pick one video
and put it on YouTube.
This is a challenge that had to work on game day.
We had one shot with the real robot.
The robot we some human operators that
were connected to the robot with a very degraded network link.
So they were allowed to get high-level commands where
they couldn't steer and control the whole robot,
and I'll just show you a really quick summary
of the kind of things we do.
I set it up, so I didn't use too much of my time.
But the robot is pretty amazingly capable.
In fact, this was a good thing for the bipeds in the world.
We just had this big competition in December.
Everybody said that the people who
are bringing bipedal robots, all the robots
are just going to fall down an embarrassing themselves.
But in fact, the robots did very well.
The bipedals did very well.
This guy happens to be so wide that.
I don't know if you remember Hans and Franz on Saturday
The only way we can get through the door was going sideways.
It cost is valuable time, but that was it.
This is taking up a drill, turning it on,
and then you got to cut of shape out of a wall.
And these aren't walls.
These are things that had to work on competition day,
so picking up a fire hose and mating it with a spigot,
we can do this in the wild now.
And DARPA put on this amazing event
where they've actually rented a NASCAR stadium, the Homestead
Speedway, and we had all the 18 most advanced humanoid robots
in the world.
And each of these could walk down Pit Crew Lane
and see all these humanoids and all
the most stressed out programmers
in the world working on their robots the day before.
And they put it through all these competitions.
All right, so it's easy to say that and then
make you think that robotics is solved,
but there's lots of things we can't do yet.
So the robots today were moving not very gracefully,
We wanted them to move like ballerinas.
I would be happy if our Atlas robot, our big humanoid
had the manipulation capabilities
of a two-year-old or three-year-old here.
If I have a factory where I have a perfectly specified object,
maybe I'm OK.
But if I have unknown objects I haven't seen before,
it's still a very hard problem for us.
Object recognition, scene understanding,
high-level decision making, these
are all technical problems that we're
going to be tackling in the next few years.
But there's big problems with robotics
too, bigger than these one-off research problems, the thing
that's going to get the robots out of the YouTube world
and into the real world is solving problems about robots
being tipped gradually.
They fall down, they break.
It's very depressing when you work that hard on a robot
and they break.
But there's new fabrication technologies
out there that, for instance, Daniela
is working on that are making soft robots that
will be able to recover from-- this is one of Daniela's videos
of a earthworm robot that she hits
with a mallet at all times.
And if we can figure out how to not only build these things,
but control these things to be as dexterous as the videos I've
been showing you, that can totally
reshape the robotic landscape.
Robots are too expensive.
We spent millions of dollars on that humanoid robot you saw.
And if I want to send them out into the world,
we're going to have to make them much, much cheaper,
and soft robotics can help with that too.
They're using too much energy.
The humanoid robot I showed you actually,
the reason it had a tether was because it pulls 15 kilowatts
off the wall when it's standing still.
So if you've seen the big dog robot,
the big dog has a 15 horsepower motor, and it's about 1/8
of a horse.
So I guess if you do the math, it's about 100 times less
efficient than a horse.
We're going to have to solve that if we
want to get them out there.
We also have to make it easier for people
to work on robots out in the field
and give them legislation to support this community
and foster the environment.
But this is a good time.
This is an amazing time for you to get
involved if you're thinking about a robotics startup.
There's never been a better time where
there's cop sensors out there, off-the-shelf sensors,
actuators, even robots.
The robotics community has finally
embraced the open source movement,
both for hardware and software, the maker movement
has impacted robotics in a major way.
You can download plans for entire humanoids
now and get involved.
We just released the entire planning and control
infrastructure code base for that humanoid robot I showed.
If you're interested in using that, send me an email.
I will help you build this because I
think this is so important for robotics to get out there.
You can downline machine shops.
I think if you guys can dream up new consumer applications
and we can work together, then we can drive down costs,
and we can solve really hard problems.
And there's never been a more important time
for academia to team with industry and other academics
to bring in the next wave of robots.
I think we're really entering an exciting era for robotics.
Thank you, Russ.
Well, I hope you can see the enthusiasm
that we have for the future and for what computing can
do to improve our lives and give us better ways to work,
to play, and to live.
And so please join us in this enthusiasm,
and please join us in the Q&A. Let
me start the Q&A session by asking each one
of our panelists one question.
And I see that there are microphones in the aisles.
Please join us with your questions.
So Tim, you created this really exciting view
for the future of the world wide web, for the web we want.
What new technologies might be needed to get to your dream?
Well, so the current state we're at, web developer, HTML5,
what's different about HTML5 compared to the original web I
started 25 years ago is that every web page instead,
is completely programmable.
And the power that we're giving as we
get more APIs, like the system APIs to the web page
means that you can program your web page to do anything
the programmer computed with.
And with real-time communication, WebRTC,
you could get two web pages or two versions of the same web
page run by different users talking to each other
without necessarily going through the server.
So you can build a new distributed application
using peer-to-peer protocols across the internet based
on web app technology.
So that is happening right now.
It's going to happen in browsers like we use [INAUDIBLE]
And I imagine that will cause people
to build really exciting, new platforms on top of that.
So that's an example that we have from web
on top of the internet, web hoopla,
web app on top of the web and exciting things on top of that.
Thank you, Tim.
So in your view of the web, you have articulated the importance
of privacy, and that is actually a common theme in your dream
and in Andrew's dream.
And so let me ask you, Andrew, in terms
of this multi-party protocol you have outlined for us,
is it used today?
In what feels like we imagined using those ideas?
Well, I mean, parts of it are certainly used today
because it's based upon ideas that were pioneered by some
of the early cryptographers, folks at CSAIL,
like Ron Rubesta, Shapi Goblosser, Sylvia Tolley,
these are giants in the field, and they pioneered
a lot of applications.
When you use a credit card transaction online,
you're using part of these kind of encryption methods.
Secure multi-party computation, though where
you're allowing individuals to encrypt their data
and then share the encrypted data
to be able to compute with it, that's really new.
And part of the newness comes from the fact
that it's only been a few years ago that we
solved the problem with so-called fully
That's a scary-sounding word.
It simply means that you can encrypt your private data,
then give somebody that encrypted data,
and they can take that data and compute certain statistics
as if they had the unencrypted data, but at no time
can they access the underlying data.
And you can imagine the applications are huge.
For example, not just in financial application,
but health care.
Imagine if we could share all of our health care
data to figure out what truly causes cancer or diabetes,
but without having to disclose details
about our own private medical condition.
That's an example where I think it
will transform our society if we don't get the technology out
So that's really great, and it was
very exciting to see in your presentation
the collaboration between computer scientist and finance.
And I just wonder if you can elaborate just a little bit
more on this.
Over the past decade and even before it into the next decade,
how will computer scientists and economists
come together to solve even more important problems?
Well, you know more and more, I think
computer science is actually getting
involved in a number of economic transactions.
Actually, part of that is thanks to Tim Berners-Lee
on the internet.
Of course, e-commerce now is something
that we all take for granted.
So the idea between computer science, particularly mechanism
design, option theory, and how you construct the better
marketplaces for interactions is really coming to the fore.
So I think that over the course of the next four years,
we're going to see even more powerful algorithms that
will ultimately allow us to do a better
job of allocating scarce resources across lots
of different needs.
So thank you, Andrew.
So Russ, before you stand up, I want to ask you a question.
I'm very excited to about your presentation
because you and I share the dream of robots coming.
So there has been a lot of talk about self-driving cars
Where else are the robots coming?
Can you comment on this?
I actually think self-driving cars
is a perfect example of both where we need legislation
to help so it makes it easier for people
to do the experiments they need to do to make
self-driving cars a reality.
I also think that so Google has got the most visible
self-driving car product, but all
the major automated factories have self-driving car products.
But these cars are driving right now,
they're out there doing meaningful things,
but they're only driving in small scenarios
where they have perfect maps, and they haven't
solved the whole problem.
So it's a perfect example of how the robots are at a point
today where they can do meaningful work.
They can make an impact on people's lives,
but there's still really hard set
of problems waiting around the corner
to feel that these things are a reality.
Thank you, Russ.
So please join the conversation.
We have a first question.
Would you like to introduce yourself, and--
Programmer from Brazil.
Thanks for creating Inspirational Talks.
And little bit louder.
Alex, programmer in Brazil.
My question is around--
so 40% of us have access to the internet after all this time.
And we're talking about robots, and I've
read that most of the data going around internet,
it's not human generated any more.
So maybe you guys can comment a bit
about the internet of things and how if robots aren't internet
would impact or maybe even overlap the human axis to it.
I'll take the first crack, then I'll defer to Tim.
But there is actually a great panel yesterday about robotics.
And Daniela brought hats.
So here you go, so yet get CSAIL hat for asking the question.
So ask more questions.
You'll get hats.
It's a nice hat.
I had to work really hard to get mine.
So there was a great panel yesterday on cloud robotics.
So Google is talking about cloud robotics and Ken Goldberg
So there's a dream in the field that an individual robot can
only have so many experiences.
It takes time to have experiences, get data.
We can have many robots out there feeling all the time they
can share experience.
They can learn from each other.
They can map the world, and it's much, much faster.
They can leverage the compute power of the servers.
So cloud robotics is a thing.
It's happening, definitely.
And for humans too.
So a lot of the data out there, people
are worried about privacy because they
worried about the data thing about them.
Well, actually I'm excited about them being about me.
For me, the value of the data about me is to me.
So if some marxing person can makes some money
out of knowing something about my detailed medical history,
they will get a certain amount of limited value
out of knowing about my demographics.
But if me, that's data about me is
going to tell me where I'm going to get a heart attack,
when I'm going to die, where I'm going to die,
so not go there, then that's more exciting.
So my favorite app at the moment is the one that's runs on here
and checks my movements just by listening.
The accelerant on my nose gives me now back as far
as of October, I've got when I'd been running and walking
sleeping and so including where I've been,
a pretty good journal of my life.
Adding to that, to my house.
It can tell me quite a lot about when I was in the house,
about how I was moving and turning the lights on and off,
adjusting the temperature.
We're getting to the point where certainly, the house
gets smart enough to know when I'm there.
Because after all, the list knows when I'm getting close,
so now I can program this took when
I get within it, certain radius of the house,
a couple of miles, to make sure it's
going to be warm enough for me.
So bit by bit, there's lots of ways
in which I want this world of data
to know a lot about me because, dah, there
is so much stuff that it could then do with it
that I wouldn't have to do.
And I can move on to being more creative
and get the [INAUDIBLE].
So what I'm interested in personally,
I'm using parallel with myself as opposed
to worrying about who else is using it for their own affairs.
So if I could just jump in.
I'm also very excited about my data.
I'm also excited about how I can improve the way I live
with my data, so take my data.
But the other side I wanted to add to what you said,
Tim, is that we are now developing technologies
to give people control over who accesses the data and how.
And so why with these new computational
approaches to ensuring privacy, it
will be very easy and very comfortable
for people to give their data for whatever
causes and greater goods and personal objectives
that we have, while at the same time
to ensure that the data is not used where
people feel uncomfortable.
And at CSAIL, a the labs is we are starting a information
privacy project, which is going to be proof
within a center within the map.
[INAUDIBLE] White is heading, he's
a great one of accountable systems,
putting accountable systems into government
so the government can make sure that people use or kind
of the way they used data.
Thank you for your question.
Let's take the next question.
My name is today Amadeus Ponticata, a really fascinating
And I have a question about the future and some
of the limitations in hardware where robotics examples we saw.
And there was some recent examples of the swarming
and playing, a bit like a bird.
And it seems like there's a lot of software, which
is trying to recreate political systems
and teach you these neural networks and machine learnings
and real cool stuff like it was brain.
And so I think it seems like people
have latched onto this idea of natural and software.
What about hardware?
Where have you got to in terms of biological and hardware?
There's some great famous examples of mimicking hardware
to the point where your robot doesn't work at all, where you
have every kinematic, every joint, every link,
every motor is in the same place as the biologic counterpart.
And you have a beautiful robot that doesn't move [AUDIO OUT]..
Is that me?
So I think the inspiration-- so for instance,
if you can take the Wild Cat video from Boston Dynamics,
it doesn't have every link in the same place
that a cheetah robot would have every link,
but its center of mass moves surprisingly,
the ground reaction forces are surprisingly
similar to what a cheetah runs.
So it's at that level of abstraction
where we take the governing principles of biology,
but skew them towards the hardware
that we're good at building.
And so the cheetah actually has more power
available at the joints than a biological cheetah,
and we should be able to do better
than that animals in a lot of cases.
But we have to build the hardware differently.
So it's not a perfect kinematic, it's
by inspiration instead of biomimicry is the trend.
Thank you very much.
Questions for Tim.
To protect freedom and access to information,
would you recommend it would be better
to apply this policy and maybe technology
changes for our existing internet,
or are there fundamental changes that
could happen to the internet that would change the playing
field all together and create a way better internet?
Should we reinvent a new internet?
The internet is constantly--
Give it somebody else.
The internet is constantly waiting to be reinvented.
But there has been a huge problem
without every group trying to get people to move on to IBD 6
from IBD 4, partly because it was a step change.
While HTTP, if we couple being continually, however, yeah,
there'll be quite a step change in HTTP 2,
which is very much rethought much more for real time
by any protocol.
But HTTP has been changing just by the addition of new headers,
and it's been tweaking sort of way.
It's really wanting to preserve the existing structure.
To Britain, it's not going to be from a dramatic change
in the protocols where we're going
to get privacy or something.
Well, moving to lots of encryption, but encryption
is largely there for you to be using HTTPS, involving HTTP.
We should be using SSL.
I use BGP.
Hands up, anybody who uses BGP encryption on an email?
The person does.
And you don't.
Ask him about it afterwards.
So one of the things we need to do
is to use the stuff, which is already there like BGP.
Make it much more easy for normal people
to use in the case of BGP, particularly, BGP, P2P.
But in order to resist the--
for example, the government is doing the wrong thing.
We have to work on the government.
We have to put pressure on them.
We have to change the laws.
We have to establishing a non-national constitution
bill of rights of internet.
And there's no quick technical fix to the fact
that we have a bylaw.
Let's take the next question.
My name is Jeff, and I'm dovetailing off
of that a little bit.
Considering that it's taken 25 years
to get this current state of infrastructure,
where do you see things coming together so that we're not
relying on telephony.
There's a ability to get PORT K, things like that.
And we rely on telephone companies
to provide the service, with so many municipalities not being
able to lay down the pipe for fiber, et cetera.
So that's a very big question.
We leverage telephone companies is really important,
at every layer in the internet at the connectivity layer
and content layer that independent markets and also
independent from the device layer.
So I can buy a device in one place.
I can completely separate to choose
my internet connectivity, choose which
movies I want to choose on.
So keeping those separate and keeping each market
so it is a really competitive market is important.
The last mile, which you mentioned, will we
be stuck with--
where the town where I live, the electricity company
is just going into internet provision.
So because there wasn't enough people,
there wasn't enough competition in the phone
companies delivering internet, I've lost my faith.
Other companies can get internet.
Some large nameless internet companies and search companies,
for example, even can get into providing internet.
So I think it's really important that we don't assume
that we have to rely on the existing top telcos
for supplying our internet.
It's important that the challenge
by all companies of new business models
and new companies every State.
And you see that changing over time?
Getting away from that?
The question is how do you see that changing?
I think it may change in lots of different ways,
and how it changes is up to us.
It's not clear how it's going to be.
Thank you, Tim.
And next question.
Thank you so much.
Can you describe a bit about the future
as far as the type and nature of jobs
that will exist when the future of robots
become well and truly ubiquitous?
Could you speak on that a bit?
So I had to--
maybe Daniela is also robotics, so
let me let her kill it first.
So we're very excited about working
on the next generation of robots that will
that will help people in manufacturing,
and that will bring manufacturing back to the US.
And we imagine a world where workers
will be able to synthesize the robots for any task using
computational tools and easy interfaces, a kind of a robot
keep costs, if you will, where instead
of printing your poster, you will be printing your robot
to solve that particular task on demand
at the time when you need a task.
But the robot cannot be fully in control of a factory.
People will still be in charge.
The robots will be available to help people.
And in my dream, I would like to see manufacturing environment
where we change from design in the US
and fabricated in China to design in the US,
customize at home in your living room by yourself
and fabricate it locally.
I totally agree.
So in the first robot revolution,
happened right around the 1980s where
the PUMA robots came out and revolutionized
And if you look at the automotive industry
just before that happened and just after it happened,
we were producing basically after 5 or 10 years,
we were producing twice as many cars,
but we were also employing 15% more people
to make those cars work.
So every time we've added a huge technology leap,
it's been good for people.
It's been good for jobs.
The challenge is going to be educating people so that they
can work with the robots.
And the next level of technology leap, we
have to work right into the education system.
We have to teach people computer science and robotics when
And the Dean Kamen, the FIRST robotics team
is here at South by Southwest, and I just
think this is so important for the future
to make this transition happen.
Thank you for that question.
I have a follow-up question on that,
and then a question on internet policy.
So my question on the previous question was,
do you see that as we start introducing technology
like driverless cars--
Hold the mic.
As we start to introduce technology
like driverless cars, you can very much
see where the taxi industry might become quickly obsolete
once that happens.
How do you see markets where manual labor
or entry-level jobs will be affected by robotics
and how that might change?
I think you're right.
I think that's a tough question.
I think people's jobs will ultimately change.
But I think in the short term, it's important for us
to embrace the technology, make policies
so that people can explore this technology, and educate people
so they can make the transition.
But I think some jobs will inevitably change.
I think self-driving cars are going to take time.
They're not as close as you might feel watching YouTube.
But I hope that I'll be reading a book next time when
I'm in a car in 20 years instead of being stressed out
behind the wheel.
Can I just jump in to re emphasize
that the jobs will not change with the flip of a button.
They don't change instantly.
It will take some time, and so the question
is, how do we retrain to go along with this transition.
And hopefully, during this transition,
the jobs will become more exciting and more interesting
for all of us.
Just like what you experience with the agriculture revolution
and with industrial revolution.
All of that, I think would be worthwhile
thinking as a community, how can we institute training
for people who are just entering the job market on how they can
move on from a market that perhaps is being replaced
by newer technology and help them in their transition.
So that's a good question, and here's the dream.
What if every high school kid in this country
learned how to program in high school.
Wouldn't that be cool?
I'm a big proponent of that.
It's just as important as reading and arithmetic.
Not just because we need more programmers,
but because we need the people who are lawyers and politicians
to know how to program so that they can understand what's
possible with technology.
So in the 21st century, I believe
that programming is as important as reading, writing,
and calculating, doing math.
The other question was--
We have some Trojans in the audience.
It's very important for lawyers as well,
but to have a foundation of technology and as a manager
as well, I think it's really important for managers
also to understand technology and working with technology.
You should keep it.
So my follow-up question was--
Why don't you join the back of the queue.
We have time.
We'll get to you again.
Thank you very much for the [INAUDIBLE]..
So to Professor Lo touched on the subject a little bit
with the financial data, but my question
is, do you think we can actually make
a balance between censorship and privacy and sometimes could
they conflict, and who should do this?
Who should be the agency or media to make this balance?
So I think that's a very important question
that you raised, and I was actually going to bring it up
when Tim was talking a bit about the data that is
on the internet all about him.
I think that there's a lot of benefits
to having all of your data out there,
but your question really is about potential costs.
And let me give you one example, and then I'll
tell you what I think you can do about it.
If I were an insurance company, I
would love to get that data that Tim
has on his app with his accelerometer.
Because I'd like to know how fast he drives.
I'd like to know where he goes.
I'd like to know what his health condition is
because once I find out all that information,
I may decide to refuse to insure him.
Because he's not a good risk.
No way, buddy.
And this is the problem.
It's that when your data is out there,
it can be used in ways that are unanticipated.
And this is why I think Tim's point about new legislation
is so critical.
We don't have legislation to protect our privacy
and to prevent the misuse of data
because, frankly, when the founders of our country
were framing the Constitution, they
didn't really understand the notion of data and privacy.
So I think that with better legislation-- and legislation
is going to have to come from the people.
You're going to have to write your congressman and say,
this is an important issue.
We want you to be at the front of it and not in back.
I think that that's going to be a critical part of what's
going to happen over the next 5 to 10 years.
So you think someone should be able to sensor
data and software because it violates someone's privacy?
I think that censorship is very different from preserving
privacy, and I think this has to do with individual rights.
In other words, you have the right
not to share your data with me, but nobody
should have the right to limit our free speech.
Those are two very different things.
And I think that right now the law does not
reflect that distinction, and we need to have better legislation
to deal with those issues.
So thank you for that question.
We're officially out of time, so let
me suggest that those of you who haven't had their questions
answered, join us at the podium, and before that, let's thank
the speakers one more time.
And also thank you for those who were online.