36c3 preroll music
Herald: OK, good. Gauthier Roussilhe
investigates the impact of the digital
industry on the environment and how we can
actually reduce this industry's footprint,
ecological footprint. One example is his
own home page. It's visually appealing,
but it only loads in 450 kilobytes.
Gauthier stage is yours.
Gauthier Roussilhe: So. So thank you for
inviting me here. Uh, so my talk will look
at digital industry, but in a broader
scope. We gonna look and analyze what is
digital industry nowadays, looking at what
is possible to do within transition. Which
should be our goal altogether. So to just
give me a sense of who I am and from which
position I am speaking of. Uh, first of
all, I'm a designer. I don't know if there
is much designers here, but it's quite a
nice practice, I recommend. Which means I
make digital services. I don't have a
technical expertise on programing or
coding, but I do understand a little bit.
But most of my work as a designer and also
as a PhD candidate, I've been looking at
transitions, policies, energy policies,
environmental policies, legal policies,
when it comes to, and the Anthropocene and
the paradigm change, the paradigm shift
that we to operate regarding that. So
within the scope of transition and climate
crisis, environmental crisis, I've been
looking especially at the digital
industry, its footprints. The way it will
evolve and if it's going far away from
transition goals or if it is going the
same direction. I use sometimes the term
low tech, which I don't really like. And I
will explain later why. But basically
looking at: What, what does, uh,
digital... sustainable digital industry
looks like? Which is quite a long way. And
also I have been doing a side research on
economics as it is very interesting to
look at that when doing this kind of
stuff. At the same time I was also the
director of online documentary called
Ethics for Design, looking at the
responsibility of designers. When you put
goods and services massively in people's
everyday life, what is your
responsibility? So today my position as a
speaker will be mostly looking at at least
making a transition politics arguments
linked to a social argument. But I don't
only... I will not focus on technical
arguments per se. I will focus on
techniques and technologies through the
scope of transition. So please remember
that I am not that much of a technician.
So first we need to set up the framework.
I will not give you a lecture on the state
of the planet right now. I don't think you
need me to go through that. And I think
there is a far better people to talk about
that. Anyway, I prefer to talk about
transition. So what is it? What are the
targets? There is a first target. We all
know it. That 2° target. Paris agreement,
which means we need to stabilize carbon
concentration in the atmosphere to 480 ppm
(particles per million). To stabilize
that, to stay under 2° average on Earth,
we need energy transition. We need to
apply an energy transition. And sometimes
we reduce it to just shifting our energy
mix to less carbon intensive energy mix.
When actually the first step is firstly to
reduce energy consumption, then you can
make it less carbon intensive. But if you
don't know, if you don't learn how to
reduce energy consumption, there not that
much to do. So I'm gonna go through that
first. And first, because I think we've
been talking about carbon quite a lot. Uh,
just before. And also this morning with
Chris Adams. But I realize that not that
much people understand why we picked
carbon between the halls of greenhouse
gases that are on earth. Why do we set up
a target on carbon? Uh, carbon has two
specificities: time lag, residence time.
First thing to know about carbon is on
average, a particle of carbon that just
being emitted from your car will on
average take 20 years to reach its maximum
heating effect on the atmosphere. So when
we are doing transition now or engaging
for transient policies, we are doing it to
have a result and effect 20 years ahead.
So when we are doing stuff now, we are
doing it for 2040 on average. So which
means also that the transition. I mean,
the emission for the next 20 years are
not... we have pretty good estimates.
Secondly, carbon has the highest residence
time from all the greenhouse gases, one of
the highest at least. It will take 20, uh,
10000 years for all the carbon that we
emitted so far to go through the
atmosphere. We add carbon. It doesn't go
through the atmosphere happily. And going
into the outer space. We add carbon. And
the carbon that we emit today, right now,
it will stay at least for 1000 years. And
some of it will go through the carbon
cycle and will stabilize. But we are
adding more carbon in the atmosphere and
carbon has the specific nature of staying
a very long time in the atmosphere. That's
why our targets are in looking at carbon,
because it has the maximum heating effect
regarding its residence time in the
atmosphere. So if we look at France,
because you have to remember that in this
talk I'm speaking from the perspective of
a French designer and most of the work
I've been using as a result, I've been
using or I've made is from the French
perspective and the French research. So on
average on France, we emit 12 tons of
carbon equivalent per person per year. In
Germany I don't know. I think you are
roughly around the same number. What's
interesting here is that digital industry
in this total is almost 1 ton and 200
kilograms of carbon equivalence. That's
where we are looking right now. With this
talk, we are looking only a small portion
in green, in goods and services, but it's
dynamically linked to all the other
sectors. So when I'm operating transition
in this sector, I'm also looking: How does
it link to everything around them? And we
have to know that if we want to reach
Paris agreement, we have to stabilize our
carbon emission per person per year to 2t.
So in France, we have a lot, we have to
reduce by 10t of carbon emissions per
person per year. So roughly, roughly the
same thing goes with Germany. So this is
another calculation with less, just a
smaller amount because they don't
integrate digital industry as much as a
precedent one. But basically the road here
is 11 to 2. And what's interesting: It's a
French study looking at what is my
responsibility as an individual to go to
the target, and they estimated that with a
realistic individual behavior change.
Eating less meat, not taking planes, using
less car, biking, cycling more. I can only
do one quarter of the effort needed. So
when companies are focusing on individual
behavior change it's bullshit. Because
individual behavior change always goes
with systemic change, which is three
quarters of the road we have to take. So
you cannot engage in individual behavior
change. We thought asking or fighting for
systemic change. So here most of the road
that we have to make is through
decarbonizing or making less carbon
intensive industry, agriculture,
transport, public services, energy
production, something I cannot do through
my own individual behavior, but my own
individual behavior is needed. If I want
to act on the political scale, on the
systemic scale. So we live in a paradox,
then that's every individual change is
necessary, but insufficient. But we need
it. So looking back at digital industry,
we need to frame digital industry in 0.6t
of carbon left for goods and services.
That's where we need to put new
imaginaries, uh, new ways of practicing
digital in this target. And we have to
share it with other goods and services,
clothing and so on. So going quickly
through the impacts of ICTs right now. In
2019, 3% of the worldwide energy is
consumed by ICTs, its main primary energy,
fuel, oil, gas, nuclear power,
hydroelectric power. Everything that we
need to power, uh, transport, boats, cars,
your phone, data centers. 3% of worldwide
energy is consumed by digital industry.
Accordingly to that, now, we are almost 4%
of greenhouse gases emitted worldwide are
coming from digital industry. But they are
only numbers. So we need to see how they
will evolve. Right now, the growth rate of
these two numbers are quite shocking. So
basically, the energy consumption of this
industry is doubling every 8 years, I
think. Yeah, 9% of growth rate. That's the
only industry worldwide, I think, with
that kind of growth rate. And it goes with
greenhouse gases emission 8%. So which
give us perspective that digital industry
in 2025 will be 5% of all of the worldwide
energy consumption and 7.5% of greenhouse
emissions. The fact is greenhouse
emissions are growing faster than energy
consumption is because the increase of
energy demand coming from digital industry
cannot be absorbed by renewables. You need
coal power plants, or carbonated energy to
go with this fast growth rate of digital
industry. So if I look closer at energy,
you can see that right now on the global
average, 45% to 50% of all energy is used
to manufacture. And the rest of energy is
for making things work, when you are using
them. But it's a global average. And when
we are looking at that... So in our
calculation methodology, we have 3 places
we are looking at: Consumer equipment,
networks, data centers. So we are looking
at the energy consumption of these three
places, both to manufacture them and to
use them. But it's a global average. So if
I look at a specific consumer equipment
like a smartphone, it will look like that.
So mainly when you are buying a
smartphone, 90% of all energy has already
been used. And if you charge it every day
for three years, it will be the 10% left.
So when you are using a phone, changing a
phone, you actually trashing 90% of
energy. And then there is water, minerals
and so on. But I don't have time to enter
in this topic. So today the main impact of
digital industry is manufacturing and
producing electricity to make all the
infrastructure work. So when we think of
web design or designing digital services.
It's little compared to these impacts. But
we need to think of services that enable
to reduce this impact. That's how we link
it together. And this infrastructure has
been used mainly for videos. So Chris
Adams, uh, showed this graph today. I was
part of the study that look at the impact
of online video on Internet. So mainly 20%
of all data moving in the world is not
video. 80% is video. So as Chris shows
this morning, most of it is Netflix, uh,
pornography, tubes and others. Uh, the
fact is Netflix with 150 million users
worldwide is basically representing 15% of
the global Internet traffic. And they've
been doing that for... And they launched
the streaming services ten years ago, I
guess. So that's quite a big growth rate.
But this graphic, I want to challenge it
because I was part of this study. So I
also know its limits. This doesn't show
pollution or energy consumption. It just
shows you data moving. And 1 gigabyte of
Netflix data, video data, has way less
energy consumption than once you get bytes
of banking data, especially if it's
running on old data centers that haven't
been be updated for 30 years. Netflix has
incredible infrastructure. So moving 1 GB
for Netflix is less energy. So it doesn't
represent energy consumption. It's just
data moving. So how do we deal with that
now? We have a good... We know the
transient framework in which we are
operating. We know the impacts. How do we
do differently knowing that? First we need
to challenge the discourse that have been
put up when we speak of Internet and
digital infrastructure. When digital
infrastructure arrived in civil society,
there was two discourses. First that it
was dematerialized. Secondly, that it will
create a global village. Well, I think now
we can know, we have enough data to say
that both of these discourses were myth or
lies. If anything, digital industry is
hyper-materialized. It requires an
astonishing amount of minerals, resources,
water, energy, infrastructures that never
been seen before for such a small. I mean,
such a young infrastructure. So when you
are dematerialized, you don't account for
resources. You don't account for energy,
because there is no impact, at least in
public discourse. When you think of a
global village, it's quite an aggressive
thing when you say global village. It
means basically your erase culture,
geography and history of places in which
you are implementing the infrastructure.
So I think we need to change this both,
these discourses. If we want to look at
what digital industry can be in a system,
a paradigm of sustainability. And we have
to understand that because of these two
things that have been said about digital
industry by default, at least in my
perspective, most of the usage of the uses
were created with the current digital
industry. By default energy intensive or
high energy based, by default. I can show
you quite easily with Netflix. So Netflix,
one of the biggest data movers on the
Internet. But actually, when you think of
it being able to broadcast to 150 million
users worldwide, high quality videos, is
already based on the fact that they don't
pay that much energy. And secondly, to be
able to incite people to watch more, to
create an interaction of auto play so
people can see more and more videos, can
watch more and more videos. It's because
you don't account for energy. It's not a
cost. It's not really something that
matters. But also to look at Netflix
precisely: Netflix also created one of the
very efficient network broadcast its
videos. So here to be quite precise, when
you are watching a Netflix episode, it
will never go through the Internet because
in each data centers of Internet service
providers, there is this little red box,
this little servers from Netflix that are
actually caching all the catalog every
morning. So when you are clicking on play
on Netflix, it's just streaming you a
video from your ISP data center. So it
will never go through the rest of the
Internet. So they actually don't, they
optimize a lot the streaming services. But
the fact is, even if it's a very energy
efficient, they are growing so much that a
gain of energy are being completely
overwhelmed by the growth rate that they
are fostering through their practice. And
yet Netflix account for if it was going
through the Internet for real, it will
account for 37% of all the peak internet
traffic. So if we look at the way we think
of designing websites, applications, so
on. Uh, normally we start with money.
Someone is giving you money and goals,
targets that, in design we say KPI. So key
performance indicators. And they will tell
you, I want that much audience. I want
that much engagements. I want that much
people buying my stuff. Do a service. A
web service. Uh, website application so
you can reach my targets. So from my
perspective, you are giving money for
people to move data because also data is
getting back to the people paying for it.
But in this framework, when do we think of
energy? When do we think of resources?
Because so far we've been very good at
creating efficient equipment and in
design, in the design practice, energy
never really mattered. In computer
sciences it's different. We created
fantastic efficient, energy efficient
equipment. But the fact is, the more
efficient our equipment became, the more
we consume of it. So there is a constant
rebound effect that it is not giving us
any possibility to transition in a less
intensive infrastructure. So the fact is
we never account for energy. We never
account for resources from the design
side. And I've been trawling quite a few
designers with that, asking them, can you
make me a website for 2 Watts per hour?
Nobody knows how to do that. No designers,
no designers can answer this question. And
they might be very senior. I asked senior
designers or students, junior designers,
nobody could answer this question. Because
energy never mattered. So from the way I
see it, I'm designing from energy, so I
start with energy budget because my goal
here is to reduce carbon emissions. To
reduce carbon emissions I need to reduce
the amount of energy I'm consuming. To
reduce the amount of energy I'm consuming,
I need to reduce the amount of data I'm
moving. And from that, I can decide how
much money I'm spending to design a
specific service. I don't start from
carbon. It is very inefficient and it's
unfair because for most countries, I mean,
if I give like a carbon threshold... In
France, we could do amazing website
spending a lot of energy because we don't
have a very carbon intensive energy mix.
Australia, USA will end up with very
crappy websites because they will have
maybe three kilobytes to move. So it's
better to start from energy than from
carbon, because energy is fair, to some
extent. And it's more efficient to reduce
because it is more important to reduce
energy than to reduce carbon emissions.
Because if you reduce energy consumption,
normally you would reduce carbon
emissions. So I did that with my website.
When I start to realize the also blind
spot in Zen practice, I go for my own
transformation. So my website is consuming
one kilowatt per hour for 1000 visits.
It's quite an average website. I mean, no,
it's not average anymore. It's 450
kilobytes on average. And I will add a new
thing in the next month, I will limit my
traffic to 5000 visitors a month, because
if you want to constrain energy budgets,
if you want to apply a real energy budget,
you need to constrain traffic. So you have
to decide how much visitors you want to
come every month. And once there is there
be 5000 visitors then. So the other ones
will wait for the next month and it's
fine. It's not that important to get
information all the time. So if you are
not up to do all this calculation that I
can explain with with you later. Uh, I
design a Firefox extension that shows you
the amount of data moving. And so the
energy consumption and the carbon emission
linked to this data moving on your
computer from your browser and shows you
what's moving them. What are the different
ways, where is the data going. And it
gives you some equivalences on charged
smartphones and kilometer in a car. So I
did that for, uh, for a lobby in France
called the shift project that also
produced most of... a lot of the studies
looking for impacts of digital industry.
It's called Carbonalyser. So as I was
saying to do an energy budget, you need 3
things: You need to describe or to reclaim
energy infrastructure. Where does your
energy come from? It's very important. You
need to reclaim the digital
infrastructure, which is: What is hosting,
what is the network, on which consumer
equipment? So you have to define in which
territory you're operating. So at that
point, the global village is dead because
you cannot be global anymore, you need to
precisely know where is your energy coming
from, which are the impact of the data
center you are using in a specific place
and you need to decide about traffic. So
now while I'm working sometimes with
clients, we decide how much traffic they
want to go for. And we put a hard cap on
that. So a good example of that, I think
you all know it, is the Low Tech Magazine.
First, they Reclaim Energy. They built a
solar panel, I mean, they installed a
solar panel on a balcony in Barcelona that
is powering a website. You can see the
battery in the yellow part on the website.
So first they reclaim energy. Then they
reclaim digital infrastructure. So they
are doing a self hosting with a Raspberry
Pi. And they are not looking at traffic so
far, but it will come. That's kind of the
example I want to show you, because the
territory here is paramount to the design.
You don't design stuff without knowing why
it actually be. What is the materiality of
what you are doing? So and you can see
here in the footers, they are also giving
the weather for the next coming days in
Barcelona, what's our base and where the
energy is coming from. And because I was
speaking of low tech, I just want to do
like a quick heads up on when we speak of
high tech or low tech. I'm coming from a
social science background, especially
anthropology of techniques, philosophy,
political sciences and so on. So I've
always been shocked by the word tech. What
does it mean? Is it technology,
techniques? Nobody defines it. So it was
interesting, to say that when we think of
low tech in our perspective of transition,
we are speaking of low technology and high
techniques. When when you are speaking of
high tech from the Silicon Valley
perspective, you are speaking of high
technology, low techniques, which means
you are relying on blackboxing technology,
making it the less open to people, which
means you will reduce drastically the
skills and the knowledge that people can
get from the technology you are deploying.
So low techniques, low skills. On the
other way, when you think of low tech, we
are also relying on technology. You cannot
do digital without thinking of high
technology Infrastructure. But you are
relying less on that. And you are relying
on techniques. How do you spread
knowledge? How do you share skills? How do
you learn to maintain stuff is important.
And I think you've been doing that for
quite a while here. Not telling you
something new. But what's changing here is
a perspective of drastically changing
living conditions on earth. And also the
material condition of production are
drastically changing. The way we've been
extracting minerals, producing energy,
using water for mining exploitation will
change forever and nothing will be the
same anymore on that level. And this is
more of my anthropology side speaking
here, but I've seen much more interesting
stuff of empowerment, of what is
technology, what is digital infrastructure
in other places in the world, especially
el paquete semanal in Cuba, which is
basically people coming to office in Cuba
with a hard drive. They get one terabyte
of TV shows, film, whatever, tutorials,
books, and they go back and they pay a
little fee for that. It's basically a
content distribution network, except that
you don't need network, you just need your
feet. Daknet in north India is quite
interesting, too. They don't have access
to cellular or mobile networks as we can
have here. So they deal with the problem
quite interestingly. So sometimes there is
this guy on a motorcycle with a little
antenna and sever in the back of the
motorcycle and is basically going into
every village in a specific place. He
broadcasts a signal. He creates a hotspot.
Everybody is sending the stuff and he goes
from village to village and go back to the
city. He plugs on the main network and
everything is sent. It's also do that with
bus that are picking up kids going to the
school. Also in Brooklyn: Every network
tells a story. Great initiative, people
getting to design their own networks and
to understand the materiality of networks
or the Association of French Internet
Provider in France. Fantastic initiative,
too. So I will kind of conclude on that.
This is my own framework to think of
digital industry now in the context of
transition. So we gonna start from the
inner circle: materialisation. That is the
issue with digital industry. That's the
only infrastructure that has been to my
knowledge that have been publicized and
thus a discourse of being dematerialized.
You cannot do that with roads, with roads
network or any other infrastructure,
that's unique to digital infrastructure.
So the first step is always to materialize
it. That's why we need the plugin. So you
can see that there is impacts, but this
impacts you need to frame it in two
different ways. You need to frame the
impacts on the territory in which your
energy and your infrastructure is hosted.
And also the impacts at the scale of the
earth. Watching porn is emitting carbon.
So you have a global impact with very
intimate use of the Internet. Then you
need to defend your territory. Very
interesting, because since we've be living
in the myth of the global village, we
never thought on how the territory can
actually influence the way we are
designing web services or websites. So we
need to start from the territory, as low
tech magazine did, accepting the
constraint of Barcelona and playing with
it. And then you need to understand that
we also are working on a planetary scale,
what I call terrestrialisation which is
like kind of a mouthful. But that's what
it is. We need to understand the effect of
digital industry on a global scale, on a
planetary scale. And the fact that the
living condition on earth are quickly
changing, are impacting territories, which
will also impact the way we think of
services. So starting from the territory
is a good place to start because that's
where there is a materiality of digital
industry. The ones that have been hidden
so far, or we don't want you to look at.
And I wanted to finish with this little
picture, because right now in France, we
are striking because of the reform of the
pension system. And there was something
quite interesting in the way the strike
evolved in the last days, because in
between... there is many people striking
in France right now, lawyers,
firefighters. So the Paris Opera Ballet,
because they want to change the pension
system. So we have moments now in Paris
where the ballerinas are performing for
the strikers. And it creates something
very interesting because it goes beyond
act of resistance. It creates beauty and
opportunities in the way we think of
changing the system. It goes beyond
resistance. It creates imaginaries. And
that's the most important thing that we
need to do right now for the digital
industry. Sustain and create imaginaries.
Thank you.
Herald: Gauthier, Merci beaucoup. We have
five minutes time for a couple of
questions. Please line up at the
microphones. And is there a question
already from the Internet? No question
from the Internet. Please to the
microphones. Number three, please.
Q: Okay. I'm still formulating it, but
I'll try. I've been looking a lot about
how the new push in the digital industries
is framed around the fourth industrial
revolution, which is pushing us more
towards Internet of Things, always on, the
artificial intelligence ideas the industry
is coming up with. And I'm wondering if
there is a way to push us in the opposite
direction, to go away from personal
devices and more towards library modes of
technology? So like trying to create
places like the hack labs, the hack spaces
where we go to use things instead of
people constantly having their devices on,
feeding the data surveillance capitalism
and so going against the grain of pushing
against this expansionism. And if you have
looked at that in that way?
Gauthier: Well, I can give you like a
prime experience from the French landscape
because I think I'm only legitimate to
talk about that. One thing that will be
quite dramatic for the way digital
industry is going to evolve is in my own
perspective, the deployment of 5G because
for 5G then you get autonomous cars, IoT,
4K videos streaming in a tube. It is not
going in a good direction. The
massification of [unintelligible] is not a
solution. And I was recently talking I
mean, giving a talk in the biggest, French
telecom company called Orange. And there
is actually like, an inner revolt inside
the company because engineers don't want
to deploy 5G because it's useless. We
don't need that. And that's right now,
that's kind of the shift that we are
observing in France. We think there is a
momentum of people. I mean, also, some
laws are getting passed in the parliament
regarding that. But companies in France, I
understood that they cannot do... they
will be accountable for environmental
impact of digital industries. Uh, several
cities are contacting me to influence or
to give them advice on the digital
strategy, going far away from the 4th
Industrial Revolution, the Rifkin thing.
So I think right now it is about
resistance and trying to stop the coming
flow of whatever techno solution is
incoming from the Silicon Valley to
actually stop at a specific moment. The
next big infrastructure, which would be
5G. Fighting against 5G in my regard is
what creates great space to rethink what
we want from the digital industry and what
digital use we want to foster.
scattered applause
Herald: Okay. We got time for one more
question. Microphone number two, please.
Q: Hello there. I found your model very,
very interesting of terrestrialization,
territorialization, materialization. I'm
looking for like worked examples of what
design decisions you would make
differently as a result for that. And I
didn't quite get that from the talk. Where
would I look to find a really concrete
example of this?
Russilhe: Yeah. So there is free projects
going on now. The first one. Well, I got a
European fund actually, to do a specific
project that I'm very keen on because I
don't come from a big city. I come from a
rural place in France. And I always kept
this perspective. What thinking from the
territory, thinking from the rural aspect
of life. Well, what digital use are also
less excessive. So I receive funding to
make low energy template to make cities
websites. And so I want to spread this
open source template. So all the little
village cities or little cities of France
can get the best of what we can do
regarding low energy web design and spread
it through the territory of France. That's
what we are doing right now. It will be
documented in, I mean, the first version
will pushed in March. Secondly, we are
also doing another website for the low
tech lab in brittany. What we're doing
here is documenting how to think
differently of maps, digital maps
especially. Because Google Map is not
something I want to foster, especially in
terms of energy impacts, because even if
it's very efficient, there's so much
growth regarding its use that we need to
think differently. And when you think of
digital maps, there is four. I look at it
from a design perspective. So I see for
uses. Localisation: Where I am or where is
the point I'm looking for. Orientation.
How those are related? Modelization of the
map. Or, uh, what is the fourth one?
Giving a route. When you are using Google
map, the photos that you are given at the
same time. But because it was thought on a
high energy perspective, but you don't
need to display the map if you don't know
where you want to go. So it's not
necessary to show the map if you haven't
decided where you're going next. So we are
just, most of the use I've been developing
on the digital industry so far, we are
trying to rethink it very differently with
the lowest energy possible. And it means
that we need to break down some of the
things that have been made. It will be
documented in February. So I have things
to show, but not yet.
Herald: Encore en fois, merci beaucop.
Russilhe: Thank you.
Herald: Gauthier
applause
postroll music
Subtitles created by many many volunteers and
the c3subtitles.de team. Join us, and help us!