Press "Enter" to skip to content

A new data economy with power to the people | Trent McConaghy | TEDxFrankfurt


[Applause]
hi everyone it’s really a pleasure and
honor so let’s get started I’m going to
talk today about a new data economy with
power to the people so quick another
show of hands are a show of hands how
many of you use Gmail come on don’t be
shy
great so looks like the vast majority
myself too I use it I’ve been using it
for more than 10 years and I was
thinking about this a while ago if I
lost my account if it got shut off I
would be devastated all of my
correspondents many of my contacts are
on there so I would lose track of many
friends of many work colleagues and so
on but you might think well you know
that’s not going to happen right it’s
gonna be around for a while but just a
few weeks ago a National Geographic
reporter named Rachel Bale she was
writing an article in Google Docs just
writing writing writing walking talking
about so what with wildlife crimes and
suddenly her account got frozen out
Boosh there was there was a notice from
Google that says you have violated our
Terms of Service and she got cut out of
her digital life on Google so she
started digging around and she
complained and turns out others had that
problem too and Google then fixed things
and they said oh we had a glitch we did
it pushed to the code and it
accidentally filtered out people who we
thought were violating the terms of
service but we fixed that so those of
you doing the wildlife crime writing
about that that’s okay but if you think
about it they’re still cutting people
out there cutting people out when they
write about things imagine you’re doing
an essay of what al Qaeda from gone and
all of that they are surveilling us and
of course that’s not just Google do you
remember Snowden the Snowden revelations
with prism and all that five years ago
those revelations did not stop the US
government program to monitor us in that
program Microsoft has been sending their
data to the US government for more than
ten years Google for five years Facebook
Yahoo all the rest right and it’s
happening all the time and the we can
keep going with example after example
so overall you can see that we are in a
data crisis it is really very dangerous
there is new headlines every day just
this past week for example net
neutrality got turned off in the USA and
we should be worried why well if you
think about what it is to be a citizen
of a democracy the very very core thing
you need is to protest is to write about
issues that you see about the government
it to organize and to stand up to call
people out but if you start typing an
essay that has any hint of terrorism or
otherwise or even just complaints about
the government wooosh your account could
be off and you are cut out you were
exiled from your digital accounts are
you angry I am I’m amazing it’s just I I
don’t even have words um but the amazing
thing is we can do something about it so
what can we do about it
well first of all this is take a breath
it’s breathe so and you know look at
some pretty jellyfish
if you think about it by the way a quick
aside if AI were an animal what would it
be and I’ve come to realize that it
probably would be jellyfish they kind of
look like space aliens kind of look like
you know AI is right and that takes me
actually into my next topic what we’re
trying to do is understand what’s going
on at the core what led us to this data
crisis where we are being watched we are
having our freedom our democracy
threatened so I have to start with a
story here in 1997 I had my very first
job paying jobs I grew up on a farm so I
had many many many hours of unpaid work
but in 1997 I had been hacking an AI for
years and years and I finally got paid
to do artificial intelligence AI and my
job I was given a dataset of 20,000 data
points and I had to build a classifier
an algorithm that could detect between
birds chirping people walking clouds
floating through the sky they make a
sound not very much and and tanks and so
on and so I I wrote some very very basic
algorithms at the beginning of that job
I had 45 percent error that meant half
the time it was failing and after four
months
toil with genetic algorithms and neural
networks and all this I got it down to
25% error I was proud of myself of
course it was still unusable because one
in four failures it’s pretty terrible
but it was a start you know it was my
first job and you might think that’s
that’s kind of interesting right four
months to improve error by 20% and that
was actually the standard back in the
90s that’s how you did AI but things
changed in the early 2000s people
realized that if you add more data than
you get lower error
10x more data a hundred x a thousand X
10,000 X so you don’t you actually have
mountains and mountains and mountains of
data so instead of having you know
genius PhDs to improve your model you
have CSVs
the CSV files are replacing the PhDs and
this actually read to a great reduction
in the error and that’s why we started
to see that the deployment of more and
more AI throughout the 2000s but it
didn’t stop there what happened was you
had the Google’s the Facebook’s of the
world realize that if they horde that
data to themselves if they keep to
themselves then they get an it data
network effect they have more data which
means better models which means higher
click-through rate which means more
money so they call themselves AI
companies but really their data
companies the AI part assists a
head-fake because we could think a AI is
cool but really it’s data and there they
are data silos that’s what’s going on
and what this has led to of course is
well once you have silos of data your
data then those companies can yank you
away from those silos and you get cut
off you get cut out of your digital life
once again this is threatening our
democracy and our freedom now that we
understand what’s going on the the fact
that AI has actually catalyzed these
data silos what can we do about it what
if we can actually democratize data what
if we actually give power back to the
people power over our data imagine that
you have a data Commons where all of the
data that is open and that should be
open and free to use is open and free to
use and with your personal data you
control it it’s not in some silo of some
billion-dollar organized
halfway around the world instead you are
controlling and you are giving
permission to those who you see fit and
revoking permission whenever you want
this is a new deal for data leading to a
new economy what might this look like
how can we do this well what we want to
do is address two things we want to
address the cyllage data part and the
economic incentive to share if we
address both of those then that is what
can lead us to the solution let’s talk
to with the siloed data so out there I
talked about you know this handful of
big corporations the Facebook’s and
googles of the world who have both tons
of data and tons of AEI expertise it
turns out that most other organizations
small and large don’t have that they
either have tons of data this is the big
enterprises of the that are out there
you know the fortune 500 guys they have
just lots and lots of data sitting there
latent they want to turn that data into
value using AI but it’s really really
hard to hire AI people now you have to
spend a lot of money it’s two million
dollar signing bonuses if you just
graduate with a grad degree in AI so
what’s happening is these AI people
they’re actually starting their own
companies they have all these tons and
tons of AI startups out there and the
thing is they want to play with
algorithms but they need data there is
starving for data so we have on one side
a lot of data data sitting around on the
other side a lot of AI expertise needing
data let’s connect them and that’s
basically what we’re doing with a
project called ocean this is a project
that myself and my colleagues are
working on and I will get more into it
about how it’s a blockchain and so on
but the overall project is ocean and the
goal here is to to fill the ocean with
data the ocean is for the public a
public utility network and by filling
this ocean what they means is we can
flood with the silos render them
irrelevant so once we have connected
these enterprises with these startups in
order for them to talk to each other we
need data marketplaces and not just one
day to marketplace because that would be
centralized we would have new silos
instead we need 10 100,000 marketplaces
and of course if each marketplace has
its own data sets then that would be
very low liquidity and we’d once again
have problems so overall what we really
need is a substrate a substrate that all
of these marketplaces tap into it’s a
lot like kayak or X
BDR Hipmunk where they’re all tapping
into the same database for travel data
here we want to do the same thing this
shared global database um
powering marketplaces for data itself so
I’ve talked about the silo data what we
can do to connect these latent needs now
let’s talk about economic incentives
which brings us to blockchain and
Bitcoin so nearly ten years ago now this
networks released called Bitcoin at
first just a few cypherpunks were
playing around with it but bit by bit by
bit it grew and grew and grew it’s now a
100 billion dollar ecosystem the amazing
thing about Bitcoin is that no one owns
or controls it it’s just out there you
own it a little bit I want to bid all of
us to if we’re just running a node what
is it for it’s a store of value like a
cache or eagled no more no less so that
was basically this whole new idea and it
basically cracked open this idea of
decentralization but we didn’t stop with
just Bitcoin the people who were playing
with Bitcoin including myself you know
many many years ago started asking
ourselves what more can we do besides
just digital cash for digital gold so
starting in about 2013 many people
really sort of working on this including
myself on this idea of block chains and
a blockchain basically is the thing that
powers Bitcoin but it can be used for so
many more things think of blocking like
Google sheets without the Google so it’s
this giant giant list of transactions
about who has done you know sent money
to whom or sent to ins to whom and the
amazing thing is with block chains like
I said no one owns or controls it but
you can treat these as utilities just
like the water works just like the sky
and the soil and and nature this is
actually a public utility that’s just
out there for everyone to use but they
have a very special super power and that
is you can get people to do stuff how
you reward them with tokens so in the
case of Bitcoin what that means is
Bitcoin defines security as compute
power and bitcoin is trying to maximize
its security how does it do that well
every ten minutes
it gives 12.5 bitcoins to
his gobs of compute power what you see
here is a giant ASIC farm of chips
running running running mining Bitcoin
there are dozens hundreds of these
throughout the world of various folks
and what are they doing they’re all
trying to win that digital gold and the
amazing thing is every single one of
them is adding to Bitcoin security
bitcoins incentive system has been set
up so that people build things like this
and mining pools and so on there’s a
whole ecosystem around this well we
don’t have to do this just with Bitcoin
we can design our own systems asking
what do we want people to do well with
ocean what we’re trying to do is
maximize the supply of high-quality data
how do we do that well help we reward
people for it with tokens ocean tokens
so we reward people for supplying data
and curating it it’s as simple as that
so you can run a node and you have data
and if you say this is good if you put
your money where your mouth is where the
curation then you can get some tokens so
overall I’ve talked about the two issues
the siloed data and the economic
incentives and what we can do about that
building the system we’re calling ocean
here’s some of the benefits that it
brings
first of all self-driving cars every
year in USA alone more than 1 million
people die from automotive deaths most
of this is due to driver error but with
self-driving cars it’s possible to have
the error the accident rate radically
lower than the human accident rate the
the algorithms are already there but the
only thing holding us back is the amount
of data that we need one trillion miles
driven and right now Toyota Research and
Volkswagen and all these folks they have
nowhere near that amount but if they
pool their data together they can start
to get there so we’re working with
Toyota to do this to build shared data
exchange using ocean of all the
automotive data in the world and Toyota
is collaborating with many other big
automakers to do this this will get a
self-driving car technology sooner for
fewer accidents and fewer lost lives
another one is data there’s data silos
all the hospitals
my daughter Tova she’s three she was
born here in Germany and when she was
born she was born in one hospital and a
few weeks ago I had to take her to
the doctor emergency she was coughing
and stuff had bronchitis anyway it was
midnight and I went to the hospital and
they wanted to know her whole medical
history well I was super tired and I was
trying to remember but I really had
trouble but what if I could just give
them permission to use Tova’s data from
the past where it’s just accessed from
data that I control this is and we can
do this not just for my daughter but
what if we share our health data with
permission to the scientists of the
world so that we can have better better
data sets bigger data sets for better
models for things like cancer early
cancer detection and finally of course
the biggest benefit all is flooding the
data silos with this shared ocean of
data to regain control over data towards
our own democracy again so to conclude
AI has catalyzed the rise of the data
silos and that has led us to a data
crisis democracy and freedom itself are
at our sake but we can do something
about this we can build and fill an
ocean of data we can create a new
economy of data giving power back to the
people thank
[Applause]
Please follow and like us: