►
Description
The Decentralized Internet Infrastructure Research Group (DINRG) session at IETF 117 will be held 20:00-21:30 UTC on Tuesday, 25 July 2023. The session will include presentations by Cory Doctorow, "Let The Platforms Burn: Bringing Back the Good Fire of the Old Internet" and "Ecosystem Evolution and Digital Infrastructure Policy Challenges: Insights & Reflections from an Economics Perspective" by Volker Stocker and William Lehr.
A
A
B
A
A
C
A
C
B
A
B
B
B
B
F
D
A
B
I'm,
sorry,
that
I
cannot
be
with
you,
this
ITF
I'm,
currently
on
vacation
and
then
Denmark
okay.
So
before
we
start
the
usual
thing,
if
you
have
been
to
some
meetings,
you
know
this,
so
we
are
using
mid
Echo
for
both
the
online
and
the
on-site
participation.
So
if
you're
participating
in
person,
please
also
use
the
meteco
tool
and
to
sign
into
the
room
and
to
to
manage
the
queue
and
yeah
for
our
remote
participants.
B
Please
make
sure
your
audio
videos
off
unless
you
are
sharing
or
presenting,
and
we
normally
recommend
using
headsets
quick
note.
So
we
are
here
in
the
internet,
research
task
force
and
we
have.
We
are
largely
following
the
ITF
IPR
disclosure
rules.
So
please
familiarize
yourself
with
it.
In
short,
the
idea
is
that
you
will
let
us
know
if
you
contribute
or
if
you're
aware
of
any
ipr-related
contribution
and
do
that
in
a
short
time
frame.
B
B
So
this
is
the
list
of
objectives.
Let's
not
read
this.
We
have
discussed
this
a
few
times,
but
it's
it's
quite
important
that
so
this
we
really
want
to
spend,
say
quite
some
effort
on
say,
discussing
good
courses,
but
not
only
from
a
technical
side,
but
also
from
an
economic
side
and
political
side
as
well,
and
so
we'll
do
a
set
of
activities
that
we
think
are
useful.
B
It's
not
quite
clear
whether
this
is
the
best
format
for
all
irtf
research
groups,
and
so
Energy
may
be
also
one
of
the
groups
where
say
other
forms
of
documentation
and
say
research.
Outputs
may
be
more
useful,
so
we
are
currently
thinking
about,
say
some
kind
of
policy.
How
we
want
to
you
know,
work
with,
say
the
ietf
process.
B
B
Quick
information,
so
we
we
had
this
very
interesting
dinner,
GE
workshop
on
centralization
in
2021,
and
so
we
we
reported
on
this
earlier
and
so
now,
a
few
weeks
ago.
We
also
managed
to
to
write
up
the
report
in
a
more
polished
form
and
submit
this
as
an
editorial
to
ACM
CCR,
so
for
the
July
issue,
so
it
should
be
published
anytime
now.
B
All
right,
so,
please
join
the
main
list.
If
you
haven't
done
so
and
for
this
meeting.
As
for
all
the
other
meetings,
we
really
need
a
note
taker
to
capture
the
the
gist
of
the
discussions,
so
that
works
by
volunteering
efforts,
and
so
we
asked
for
one
notch
on
the
mail
list,
but
so
far
we're
not
able
to
convince
anybody.
B
F
B
All
right,
so
this
is
our
agenda.
We
have
three
talks
today
and
quite
interesting
from
like
different
angles
on
that
problem,
so
we
start
with
cori.ctoro
and
then
we
have
full
cast
record
and
William
Lear
and
then
Christian
Julian
was
a
talk
on
some
ideas
for
for
research
in
this
in
this
field,
and
that's
it
from
us
so
I
would
say
we
just
start
with
the
presentation
and
give
the
floor
to
Corey
and
I'll.
Stop
sharing
Corey.
A
G
A
One
and
after
a
couple
of
Decades
of
lots
and
lots
of
platform
growth.
We
are
now
in
act
three
and
that
gun
on
the
mantle
piece
in
act.
One
is
going
off
with
quite
a
deafening
fuselage.
We
are
in
this
moment
of
a
very
visible,
highly
contagious
platform,
Decay
from
flop,
from
Facebook
and
Instagram
to
Twitter
and
twitch
to
Reddit
and
Discord
and
Beyond,
and
there's
a
signature
pathology
to
platform
Decay
a
disease
with
a
three-stage
procession.
A
In
the
first
stage
of
platform
Decay,
you
have
firms
that
have
a
lot
of
access
to
the
capital
markets
and
they
use
that
to
generate
and
offer
surpluses
to
end
users
and
even
as
they
do
that
and
make
a
hospitable
place
to
lure
end
users
in
they
leverage
technical,
economic,
economic
and
behavioral
factors
to
lock
those
users
in.
So
they
can't
readily
leave.
A
And
then
you
reach
the
end
stage
stage
three,
where
the
platform
now
having
locked
in
business,
customers
and
end
users
be
begins
to
withdraw
all
available
surpluses
from
both
and
reallocating
them
to
its
own
shareholders,
and
the
idea
here
is
to
leave
behind
just
enough
residual
platform
value
that
both
groups
of
users
remain
locked
in
and
don't
just
rush
for.
The
exits
and
I
call
this
process
somewhat
infamously
in
shitification.
A
It's
the
process
by
which
platforms
start
off
as
robust
sites
of
value
for
both
users
and
business
customers,
but
then,
in
the
end,
turn
into
a
pile
of
with
no
value
for
either
and
and
as
a
use
case
here.
I
want
to
talk
about
how
Facebook
proceeded,
because
it's
pretty
canonical
so
when
Facebook
started
it
had
a
problem.
Everybody
wanted
to
use
social
media
already
had
an
account
on
MySpace.
A
The
only
people
were
allowed
to
use
Facebook
were
K-12
and
University
students
with
edu
addresses
and
it
needed
to
lure
those
people
off
of
Myspace
and
onto
their
platform,
so
Facebook
gave
them
two
pitches
the
that
to
convince
them
to
leave.
The
first
was,
if
you
come
to
Facebook,
we
are
only
going
to
show
you
the
things
that
you
ask
us
to
show
you
you
show
up
here.
You
articulate
your
social
graph.
A
Tell
us
who
matters
enough
to
you
that
you
want
to
hear
when
they
have
something
to
say
and
then
every
time
they
say
something
on
the
platform,
we'll
just
put
it
in
your
feed
and
the
second
thing
they
they
promised
was.
We
will
never
ever
spy
on
you
right,
not
like
that
service
run
by
the
evil,
Cricket
Australian
billionaire
Myspace.
We
are
the
Privacy
forward
alternative
where
you
can
come
and
have
a
non-commodified
social
relation
with
your
friends
now
with
social
media
lock-in's,
actually
quite
easy.
A
It
comes
built
in
with
social
media,
social
media
Dynamics,
because
once
the
people
you
like
are
on
a
social
media
platform,
they
hold
you
hostage
and
you
hold
them
hostage.
You
have
this
network
effects
driven
growth
where
people
join
because
they
value
the
people
who
are
already
there,
and
then
they
make
the
service
more
value
because
other
people
value
them,
but
the
the
coral
area
of
this
network
effects
driven
growth
is,
is
that
it's
very
hard
to
leave
that
there's
this
High
Collective
action
problem
associated
with
going.
A
You
know
those
of
you
who
are
there
in
person
today
in
San
Francisco,
when,
like
eight
of
you,
try
to
decide
where
to
go
for
drinks
tonight,
you're
going
to
have
to
argue
about
it
for
45
minutes.
If
that
were
you
know,
180
of
you,
in
a
social
graph
on
Facebook
all
trying
to
agree
whether
it
is
time
to
leave
Facebook
and
if
so,
where
you
should
go
you're.
A
Basically,
hopelessly
mired
in
this
Collective
action
problem,
so
Facebook's
customers
lock
themselves
into
the
the
network
and
they
stay
there
because
of
the
high
switching
costs
of
going
it's
not
that
they
like
Facebook,
but
they
like
the
people
who
are
on
Facebook.
They
they
value
their
communities,
they
value
their
family,
they
value
their
friends,
they
value
their
customers.
You
know,
axiomatically,
that's
that's!
A
What
they're
there
for
so
at
this
critical
juncture,
Facebook
starts
to
draw
down
the
surpluses
that
it
allocated
to
those
users
in
the
form
of
what
we
could
call
a
clean
feed
and
a
privacy
respecting
environment
and
started
to
reallocate
those
to
business
users.
So
it
turned
to
the
advertisers
and
it
said
Hey,
do
you
remember
when
we
promised
our
users
that
we
were
never
ever
going
to
spy
on
them,
not
like
that
evil
Australian
billionaire
it
turned
out.
A
We
were
lying
and
we're
going
to
spy
on
them,
From,
Dusk,
Till,
Dawn,
and
we're
going
to
build
fine-grained
targeting
tools,
we're
going
to
task
a
lot
of
engine,
an
engineering
resource
to
make
sure
that
there
is
an
ad
fraud
or
ad
abuse.
When
you
pay
for
an
ad
someone's
going
to
see
it
and
we're
not
going
to
charge
you
very
much
for
it
we're
going
to
offer
very
competitive
rates
and
to
Publishers
they
said
Hey.
A
Do
you
remember
when
we
told
those
users,
we
were
only
going
to
show
them
the
things
that
they
asked
to
see
from
the
people
that
they
chose
to
follow.
That
was
also
a
lie.
If
you
post
small
excerpts
with
links
to
your
own
website
that
you
have
some
means
of
monetizing,
we
will
just
like
non-consensually
cram
them
into
the
eyeballs
of
those
users
who
never
ask
to
see
them
and
will
generate
a
traffic
funnel
for
you,
and
some
of
those
users
will
even
subscribe
to
you.
A
You
won't
even
have
to
rely
on
our
recommendation
system
because
they're
they're
going
to
choose
to
follow
you
now.
Digital
tools
have
got
this
incredible
flexibility
that
is
lacking
in
the
traditional
analog
markets.
Digital
tools
allow
for
rapid
repricing,
rapid
changes
to
business
logic
and
so
on,
and
so
it
was
possible
for
Facebook's
management
to
subtly
alter
the
conditions
to
make
these
firms
more
reliant
on
Facebook
and
also
to
extract
value
from
them
one
little
driven
job
at
a
time.
A
A
All
the
value
has
been
extracted,
but
your
readers
are
there,
and
so
you
have
to
be
there
and
then
you
get
to
the
final
stage,
where
they're
really
drawing
down
all
the
value
to
the
bearer
minimum
and
they're
saying
things
to
they're,
saying
things
to
Publishers
like
well.
Even
the
people
who
subscribe
to
you
aren't
going
to
see
the
post,
even
if
it's
the
whole
Tech
text
of
the
post.
A
Even
if
you
haven't
linked
back
to
your
site,
you're
gonna
have
to
pay
to
boost
your
post
so
that
the
people
who
explicitly
asked
to
see
that
to
see
it
will
actually
see
it.
Likewise,
advertisers
face
a
an
incredible
burden
of
AD
fraud.
They
face
extremely
high
prices
and
then,
to
add
insult
to
injury,
you
have
the
Jedi
blue
collusion,
an
illegal
collusive
Arrangement
between
Google
and
Facebook.
A
A
That's
where
end
users
and
business
customers
vote
for
the
exits,
because
the
cost
of
using
the
platform
outweighed
the
benefits
and
the
reason
that
late
stage
and
end
stage
are
so
close
together
is
because
the
distinction
between
almost
but
not
quite
useless
and
useless
is
is
very
fine
and
the
equilibrium
that
sits
at
almost
but
not
quite
useless,
is
very
brittle
and
difficult
to
maintain
it
can
be.
A
fern
can
be
knocked
out
of
that
equilibrium
by
exogenous
shocks.
A
So
these
very
large,
durable
platforms
that
are
now
going
through
this
mass
extinction
event
now
teetering
on
the
brink
of
collapse.
They're
anomalous
in
tech
tech
has
always
had
the
potential
for
explosive
growth,
fueled
by
Network
effects.
We've
we've
had
firms
pop
up
overnight
and
grow
to
unimaginable
size,
but
firms
in
Tech
have
also
always
been
subject
to
implosive
contraction
because
of
low
switching
costs.
We
don't
know
how
to
make
a
computer.
A
That
runs
all
the
programs,
except
for
the
ones
that
are
disfavored
by
a
company
shareholders
and
that's
why
switching
costs
are
intrinsically
low
in
Tech,
the
only
computer
we
know
how
to
make
is
the
touring
complete
Universal,
Von
Neumann
machine
that
runs
every
program.
We
know
how
to
write,
and
that
means
that
it's
always
possible
to
turn
a
network
effect
on
its
head.
If
there's
a
dominant
Mainframe
company,
that's
making
a
lot
of
margin
on
its
peripherals,
someone
else
can
show
up
and
make
interoperable
peripherals
and
eat
their
lunch.
A
If
there's
a
dominant
printer
company,
that's
charging
High
margins
for
ink,
someone
can
can
show
up
and
make
an
interoperable
ink
cartridge
and
eat
their
lunch.
If
there's
a
dominant
operating
system
that
is
making
an
enormous
profit
on
its
office
suite
someone
can
show
up
and
make
an
interoperable
office
suite.
If
there's
a
dominant
social
media
platform
like
Myspace,
you
can
do
what
Facebook
did,
which
is
allow
users
to
depart
by
giving
them
a
bot
that
you
give
your
login
credentials
to
that.
Periodically
visits
Myspace
impersonates,
you
scrapes.
A
The
messages
that
are
waiting
for
you
on
MySpace
puts
them
in
your
Facebook
inbox.
Lets
you
reply
to
them
there
and
then
autopilots
them
back
out
to
your
Myspace
outbox.
The
the
latent
inescapable
interoperability
of
All
Digital
Services,
has
always
acted
as
a
check
on
growth.
It's
always
been
the
other
half
of
the
cycle.
Firms
grow
very
quickly
because
of
network
effects,
and
they
implode
very
quickly
because
of
low
switching
costs.
Remember
Jeff,
Bezos
early
on
in
Amazon's
history
told
the
book
publishers.
A
Your
margin
is
my
opportunity,
that
is
the
historic
rallying
Cry
of
new
technology
entrance.
A
new
company
can
disrupt
a
high,
highly
profitable
incumbent
by
selectively
attacking
just
the
profitable
parts
of
its
business.
The
more
value,
the
more
money
an
incumbent
makes
on
ads,
the
more
valuable
an
ad
blocker
is,
and
the
more
you
can
sell
it
for
the
network
effects,
driven
growth
and
low
switching
cost
driven
contraction
is
why
Tech
has
historically
been
so
Dynamic
with
big
firms
popping
up
and
then
going
away.
A
Debt
gives
way
to
compact
compact
becomes
a
division
of
HP,
and
then
it's
no
more
but
Tech
today
is
stagnant.
We
live
in
what
Tom
Eastman
calls
the
internet
of
five
giant
websites
each
filled
with
screenshots
of
text
from
the
other
four
and
it's
important
to
figure
out
how
we
got
here.
It
starts
with
the
drawdown
on
merger
policing.
A
Until
about
40
years
ago,
mergers
were
considered
highly
suspect,
especially
among
firms
in
the
same
sector,
both
horizontal
and
vertical
mergers,
but
about
40
years
ago,
with
the
rise
of
Reagan
Era
antitrust,
a
policy
that
has
been
Amplified
by
every
Administration
since
up
to
the
current
one,
which
has
reversed
this.
This
40-year
decline.
A
We
have
welcomed
mergers
as
in
as
axiomatically
efficient
as
producing
firms
that
are
going
to
do
better
for
their
users
and
we've
treated
Monopoly
as
evidence
that
a
company
is
doing
something
right
rather
than
doing
something
wrong,
and
that's
how
you
get
giants
like
Google
Google
made
the
greatest
search
engine
in
the
history
of
the
world.
World
I,
don't
know
if
you
remember
those
days,
but
it
was
like
sorcery
how
good
that
search
engine
was,
and
that
gave
them
a
lot
of
access
to
the
capital
markets.
A
And
then
that
was
the
end,
because,
while
Google
has
basically
not
made
a
successful
product
in-house,
since
you
know
from
its
RSS
readers
to
its
social
media
networks
to
its
video
platform,
basically
the
things
that
Google
makes
die.
It
has
nevertheless
been
able
to
acquire
a
video
platform.
A
mobile
platform,
an
ad
Tech
stack
server
management,
stack,
Maps
collaboration,
docs
navigation,
satellite,
all
of
this
stuff
by
buying
companies
by
buying
other
people's
ideas.
A
You
know,
Google
is
a
one
idea:
company
with
Deep
Pockets:
it's
not
Willy,
Wonka's
idea,
Factory,
it's
Uncle
Penny
bags
just
buying
up
everybody
who
has
a
good
idea.
So
as
firms
accumulate,
Market
power,
they're
able
to
convert
that
market
power
to
political
power,
when
an
industry
is
dominated
by
a
small
number
of
firms,
it
is
able
to
reliably
secure
regulation
in
its
favor
to
it.
A
First,
that's
because
when
you
have
a
small
number
of
firms,
they
don't
have
to
compete
as
much
as
they
they
used
to
think
of
Google
and
Facebook
dividing
up
the
ad
Market
with
Jedi
blue.
The
way
the
pope
divided
up
the
new
world
for
Spain
and
Portugal,
and
so
you
just
don't
have
what
what
Peter
Thiel
calls
wasteful
competition.
Peter
Thiel
says:
competition
is
for
losers,
so,
but
now
that
they've
got
all
this
dry
powder,
they
can
decide
how
to
mobilize
it
to
overcome
the
collective
action
problem
of
deciding
how
to
spend
that
money.
A
A
Meanwhile,
the
entertainment
companies
were
a
very
cozy
oligopoly
of
a
handful
of
firms.
Today,
it's
even
more
concentrated
five
Publishers
four
Studios
three
labels
back
then
it
was
there
were
two
or
three
more
in
each
of
those
sectors
and
they
were
able
to
all
get
together
and
decide
what
to
do
and
they
kicked
Technologies
butt
all
around
the
world.
A
It
would
be
much
harder
for
platforms
to
allocate
withdraw,
allocate,
withdraw,
dangle
and
take
away
the
surpluses
that
lure
end
users
and
then
lock
them
in
lure
and
business
customers
and
lock
them
in.
On
the
other
hand,
not
it's
not
just
that.
Tech
has
been
able
to
stop
regulation
happening
from
happening,
they've
been
able
to
get
regulation
and
Regulatory
interpretation,
that's
favorable
to
its
interests
and
that
cut
against
the
interests
of
New
Market
entrants,
who
would
make
interoperable
products
colloquially?
A
It's
what
Jay
Freeman
from
the
Cydia
project
calls
felony
contempt
of
business
model.
Ip
is
where
we
get
lock-in
and
the
monotonous
growth
of
Ip
has
allowed
the
firms
who
secured
a
potentially
a
potentially
temporary
advantage
in
the
first
decade
or
so
of
the
2000s
to
convert
that
into
what
seems
like
a
permanent
Advantage.
Ip
is
why
you
can't
reverse
engineer
the
Facebook
app
and
let
and
allow
users
to
defect
to
community-run
platforms
and
still
continue
to
exchange
messages
with
the
people
who
stay
behind
it's.
A
Why
ad
blocking
on
the
web
is
the
most
successful
consumer
boycotted
in
history,
but
ad
blocking
for
an
app
is
a
felony
punishable
by
a
five-year
prison
sentence
under
Section
1201
of
the
Digital
Millennium
Copyright
Act.
It's
why
we
don't
have
unified
inboxes
for
all
the
social
services,
multiple
Services
we
use
it's,
why
we
can't
create
Standalone
third-party
unsanctioned
recommendation
systems
that
draw
from
Tick
Tock,
Instagram
and
YouTube
and
allow
performers
to
opt
into
their
own
monetization
schemes.
A
In
other
words,
monopolization
allows
companies
to
forestall
policies
that
would
interfere
with
network
effects
driven
growth
and
it
allows
those
dominant
firms
to
secure
IP
policies
that
prohibit
the
marketing
and
sale
of
tools
that
enable
low
switching
cost.
Fueled
contraction
so
we're
about
to
get
California
fire
season,
you're
you're
in
the
Bay
Area.
Now,
if
you
were
to
come
back
and
and
not
too
long,
there's
a
pretty
good
chance
that
the
entire
city
would
be
wreathed
in
Smoke.
A
These
fires
come
every
year
and
for
Millennia
the
First
Nations
people
of
California
practice
controlled
fire
fire
burns,
Forest
Burns
and
the
burns
cleared
out
the
brush.
They
made
space
in
the
canopy
and
they
called
weak
trees,
but
a
couple
of
centuries
ago,
settlers
declared
war
on
fire
and
they
ended
that
cycle
of
controlled
Demolition
and
new
growth,
and
today
California
has
an
inconceivable
fire
debt.
Even
without
the
climate
emergency,
that's
around
us,
California
would
still
have
punishing
annual
fires
because
there's
just
too
much
dead
rotten
wood
in
our
forests.
A
The
way
out
of
the
annual
cycle
of
out-of-control
fires
is
to
restore
a
cycle
of
controlled
fires.
We
have
to
end
the
war
on
fire
and
allow
some
forests
to
burn
every
year,
but
instead
what
we
keep
doing
is
trying
to
make
the
Wildland
Urban
interface
safe
for
habitation
spending
more
and
more
money
and
resources
to
lure
people
into
those
combustible
zones.
We
should
be
evacuating
those
zones
not
making
them
more
hospitable.
A
Now,
platforms
should
be
turning
over.
There
is
nothing
wrong
with
platform
collapse.
We
don't
need
Facebook,
we
don't
need
a
better
Mark
Zuckerberg,
like
all
billionaires.
Mark
Zuckerberg
is
a
policy
failure,
like
all
dictators,
rolling
ruling
over
massive
platforms.
Mark
Zuckerberg
is
a
factory
for
producing
policy
failures
at
scale.
We
need
to
stop
focusing
on
making
Mark
Zuckerberg
better
at
running
Facebook.
The
problem
of
having
an
unelected
social
medias
are:
who
unilaterally
determines
the
social
lives
of
4
billion
people?
Isn't
merely
that
he's
bad
at
that
job.
A
The
problem
is
that
that
job
exists
at
all.
We
should
stop
trying
to
perfect
Zuck
or
replace
him
with
someone
who's
better
at
his
job,
and
instead
we
should
be
abolishing
Zuck,
it's
okay
for
Forest
to
burn,
and
it's
okay
for
platforms
to
collapse.
Our
Focus
shouldn't
be
preserving
the
crepascular
senescence
of
dying
platforms.
The
focus
should
be
on
easing
the
path
of
platform
users
as
they
depart
for
the
next
pasture.
A
Every
demand
for
platform
governance
is
a
demand
for
policing
platform
users
conduct,
whether
that's
sesta,
Foster
or
the
lawful,
but
awful
regimes
that
ask
platforms
to
police
an
inhospitable
speech
on
their
platforms
or
related
policies,
and
every
one
of
these
requirements
to
police
users
becomes
a
pretext
for
preventing
inter-operators,
because
you
can't
police
what
you
can't
control.
You
can't
control
what
you
can't
observe.
A
So
let
me
conclude
by
giving
some
examples
of
policies
that
let
us
treat
these
highly
combustible,
often
on
fire
platforms
as
damage
and
Route
around
them.
The
first
is
to
restore
the
end-to-end
principle
with
a
rule
that
requires
platforms
to
connect
willing
senders
with
willing
receivers.
If
I
subscribe
to
your
feed,
I
should
see
what
you
post
to
it.
If
I
search
for
your
goods,
giving
the
the
exact
model
number
or
SKU,
that
model
should
be
at
the
top
of
the
search
results.
A
If
I
hoist
an
email
out
of
a
spam,
folder
and
Market
is
not
spam.
That
sender's
mail
shouldn't
go
to
my
spam
folder
thereafter,
I
mean
it
was
bizarre.
Last
year
that
Congress
had
a
debate
about
whether
unsolicited
fundraising
emails
from
Hopeful
Congress
people
should
be
marked
as
spam,
and
they
ignore
the
much
more
Salient
debate
about
whether
the
solicited
emails
that
you
asked
to
see
from
your
elected
representatives
should
go
to
spam
even
after
you
mark
them
as
not
spam.
Now,
beyond
the
end-to-end
principle,
I
want
to
establish
the
right
to
exit.
A
You
may
have
heard
a
story
about
mastodon.lol.
This
was
a
fediverse
server
with
12
000
users
and
a
fight
broke
out
there
over
the
new
Harry
Potter
game
and
the
administrator
decided
to
shut
down
the
server.
Now
this
may
sound
terrible
to
you,
but
Mastodon
has
this
easy
mechanism
for
migrating
to
another
server.
You
have
sport
the
list
to
your
followers
with
one
click
and
then
with
another
click.
A
A
right
to
exit
rule
would
build
on
existing
data
protection
laws
requiring
both
large
and
small
platform
operators
to
provide
users
with
machine,
readable
files
that
can
be
used
to
re-establish
yourself
on
other
platforms
and,
unlike
other
intermediary
regulations,
Like
rules
about
lawful
but
awful
speech
or
copyright
filter
mandates.
This
rule
is
not
a
capital
mode
to
prevent
New
Market
entrants.
It
costs
nothing
to
run
a
server
that
hands
over
this
data.
A
When
the
fediverse
already
has
this
built
into
its
reference
code,
and
it
is
eminently
administratable
anytime,
a
Twitter
user
or
a
user
on
master
mastodon.lol
alleges
that
they
weren't
provided
with
their
data
when
they
got
kicked
off
the
server.
We
don't
need
a
long
fact-finding
mission
where
we
depose
the
engineers
who
run
the
service.
We
just
call
up
Facebook
or
Twitter,
or
we
call
up
macedon.lol
and
we
say
to
the
administrator
I
know
you
say
you
gave
this
user
their
data.
They
say
they
didn't
get
it.
A
Let's
just
settle
it
by
you
giving
them
another
copy
of
it.
So
we
don't
have
to
create
a
parallel
corporate
Civil,
Justice
System
to
adjudicate
content,
moderation,
calls
or
decisions
to
remove
users.
We
can
just
make
it
easier
for
end
users
and
business
customers
to
go
elsewhere,
and
this
makes
the
content
moderation
Stakes
lower
and
makes
content
moderation
easier
and
when
content
moderation,
failures
do
lead
to
mass
user
defections
Tech
firms
will
have
strong
incentives
to
improve
their
content.
A
Moderation,
departments,
regulation
disciplines,
firms
over
the
long
term
and
disastrous
quarterly
reports
sharpen
the
attention
of
even
the
most
Adderall
dependent
Tech
founder
in
certification
is
not
for
ordained.
It
is
the
predictable
inevitable
outcome
of
three
policy
choices:
the
choice
to
allow
firms
to
grow
through
anti-competitive
measures,
predatory
pricing
and
other
conduct
that
was
historically
prohibited
until
Reagan.
A
They
say
when
we
did
it
to
others
that
was
progress
and
when
you
try
to
do
it
to
us,
that
is
theft,
because
every
pirate
wants
to
be
an
admiral
Tech.
Barons
aren't
evil
wizards
who
hacked
our
dopamine
Loops
they're.
Just
the
ordinary
mediocrities
deploying
the
same
tactics
as
their
Gilded
Age
forebears,
but
faster
and
with
computers.
John
John
Rockefeller
couldn't
lay
a
railroad
or
put
in
a
new
oil
Pipeline
with
a
with
a
with
a
mouse
click.
A
But
Jeff
Bezos
can
change
the
business
logic
of
Amazon
by
just
twiddling
a
knob
and
in
any
shell
game.
The
quickness
of
the
hand
deceives
the
eye.
If
we
focus
on
the
material
factors
that
allow
firms
to
prolong
their
senescence,
the
deceptive
and
unfair
practices
that
luring
users
and
fuel
the
rapid
Network
effects,
driven
growth
and
the
wielding
of
legal
privilege
to
block
interoperability
and
artificially
increasing
switching
costs
allowing
for
lock-in.
Then
we
get
rid
of
the
self-serving
Mystique
of
the
tech
genius.
A
We
can
keep
users
safe
and
controlled
Burns
that
clean
out
the
rotten
Giants
that
have
crowded
out
the
sunlight
and
oxygen
that
would
allow
New
Growth.
So
to
close,
we
need
to
keep
up
with
our
antitrust
remedies
the
best
time
to
have
prevented
Monopoly
formation
was
40
years
ago
and
the
second
best
time
is
now.
We
should
be
cheering
on
the
FTC
as
they
challenge
rotten
mergers
like
Activision
Microsoft
and
as
they
seek
to
unwind
historic
mergers
to
undertaken
on
false
pretenses
like
Facebook
Instagram.
A
We
should
be
seeking
to
limit
a
platform
twiddling,
passing
and
enforcing
muscular
labor
and
consumer
protection
rules,
ending
worker
misclassification
that
allows
the
abuse
abuse
of
platform
gig
workers
and
enacting
long
overdue,
American,
Federal
Privacy
Law
with
private
rights
of
action.
And
finally,
we
need
to
reinvigorate
user
side.
Twiddling
by
reforming
IP
law,
enforcing
laws
like
Europe's
dma
and
making
interoperability
in
Ironclad
element
of
all
public
procurement.
A
Insidification
is
not
inevitable,
but
allowing
companies
to
gobble
each
other
up
in
an
orgy
of
consolidation
to
twiddle
their
knobs
with
every
hour
that
God
sends
and
and
prohibiting
users
from
twiddling
back
of
course,
results
in
a
world
where
we
are
all
being
twiddled
to
death.
But
the
alternative
where
we
seize
the
means
of
computation
could
give
us
the
web
that
we
want.
Thank
you.
D
H
Jeff
Houston.
Thank
you
very
much
for
that
talk.
It
was
quite
enlightening.
It
reminded
me
of
what's
Happening
Now
with
what
happened
in
the
1890s
in
the
Gilded
Age
in
this
particular
country
and,
interestingly,
once
you
bought
your
market
and
bought
all
your
competition,
you
bought
your
future
JP
Morgan.
Standard
Oil
is
the
real
issue
here,
not
the
fact
that
they're
buying
each
other
but
they're
actually
making
remarkable
grip
holes
on
the
future
of
all
of
us
and
not
relinquishing
it
and
I
Echo.
H
A
I
think
that
the
size
isn't
doesn't
necessarily
cause
distortions,
although
size
is
prone
to
a
lot
of
distortions.
You
know
you
think
about
the
coordination
problems
within
large
firms.
I
think
we
we've
all
probably
read
the
mythical
man
month,
and
we
understand
that
you
know
adding
more
people
to
a
project
makes
it
later
right,
and
so,
when
you
ask
yourself
like
how
is
it
that
Google
managed
to
make
the
best
search
engine
into
the
worst
search
engine?
A
I
think
that
you
know
you
get
some
of
that
out
of
scale
where
you
have
a
company
that
has
you
know
a
90
plus
percent
search
dominance.
Google's
not
gonna
grow
a
search
business
by
attracting
new
Searchers.
It
can
only
grow
by
extracting
surpluses
from
people
who
search
and
people
whose
things
are
searched
for
and
within
the
firm.
A
You
have
a
lot
of
different
product
managers
whose
kpis
and
whose
bonuses
which
are
quite
substantial,
depend
on
finding
ways
to
increase
the
profitability
of
their
corner
of
the
firm
and
so
they're
all
sitting
there
going
well
I
can't
grow
by
making
by
attracting
new
search
users.
I
can
only
grow
by
Will
away
at
the
value
that
the
two
sides
of
the
search
Market
get
and
so
I'm
going
to
pull
out
this
Jenga
block.
A
But
of
course
nobody
is
coordinating
a
Jenga
block
removal
strategy
right,
so
you
have
adjacent
Jenga
blocks
being
removed
and
important
things
fall
over
so
like
last
week,
every
major
airline
in
America's
phone
number
on
Google
was
replaced
with
a
phone
number
for
a
boiler
room
full
of
scam
artists.
Who
would
take
your
credit
card
number
and
steal
from
you
right
and
I?
Don't
know
exactly
how
that
happened,
but
presumably
it
was
something
like
someone
said.
A
Well,
we
have
a
cost
center
associated
with
validating
phone
number
changes
and
we
can
realize
some
efficiency
in
that
cost
center,
either
by
drawing
down
human
oversight
or
drawing
down
the
amount
of
oversight
that
each
human
does
or
by
having
self-serve
platforms,
something
that
made
it
cheaper
right
and
they
made
the
firm
more
profitable.
But
at
this
enormous
expense-
and
you
know
ultimately
like
the
end
stage
of
insidification-
is
the
pile
of
right.
A
So
the
end
stage
of
this
is
that
people
defect
right
and-
and
there
is
like
you,
you're,
seeing
this
with
Twitter
you're,
seeing
it
with
Facebook,
there's,
there's
a
real
hunger
to
go
elsewhere
and
to
try
other
things,
and
no
one
really
knows
what
that's
going
to
be.
But
you
know
the
scale
is
definitely
a
part
of
it
as
to
you
know
whether
all
of
the
things
being
equal,
if
scale
is
a
problem
anyway,
I
think
it
is
and
I
think
it's
a
problem
for
for
two
reasons.
A
The
first
is
even
if
you're
like
a
hardcore
libertarian,
and
you
think
the
only
reason
that
we
should
have
a
government
is
to
enforce
contracts
that
government's
contract
enforcement
function
relies
on
it
having
power
that
the
firms
whose
contracts
are
being
issued
and
signed
it
dominates
right.
If
the
referee
is
less
powerful
than
the
team,
then
the
team
can
cheat.
And
so,
if
you
want
a
small
government,
you
need
firms
that
are
smaller
than
that
government.
A
And
the
second
thing
is
that,
even
if
we
stipulate
that
these
companies
are
well
run
by
people
who
are
very
smart,
they
will
still
make
errors
because
nobody's
perfect,
and
when
you
have
that
much
scale
gathered
into
one
set
of
fallible
hands,
those
mistakes
were
downed.
Much
further
right,
Mark
Zuckerberg's
judgment
errors
are
much
more
consequential
than
mine,
I,
don't
think
they're
necessarily
worse
I
mean
you
can
talk
to
my
wife.
She
will
tell
you
about
my
judgment
errors.
They
are.
There
are
many
and
significant,
but
the
worst
thing
that
happens
is
I.
A
G
G
G
We've
spoken
a
lot
about
legal
remedies,
and
you
know
legislative
remedies
and
Regulatory
remedies
for
for
all
this,
which
is
frankly
like
above
the
pay
grade.
A
lot
of
the
people
here,
one
one
dynamicity
not
did
not
touch
on
which
I
think
is
relevant
to
to
like
the
sort
of
technical
ground
floor
level.
Of
a
lot
of
these.
These
phenomena
is
that
as
a
person
who
is
fairly
heavily
involved
in
improving
new
work,
one
thing
we
need
to
ask
is
here's
some
interoperability
standard.
G
Of
course
we're
dedicated
improper
interoperability,
and
you
know
we
very
much
share
that
value
with
with
what
you've
presented
here.
But
one
thing
to
ask
is
like
who's
going
to
implement
this
and
like
usually
it's
so
like,
if
Google's
not
going
to
do
it
and
Apple's
not
going
to
do
it
like?
Why
do
it?
G
And
that
turns-
and
you
know
it's
in
some
ways-
that's
a
reasonable
question
to
ask,
but
it
also
very
much
enforces
a
lot
of
the
Dynamics
that
you've
used.
That
I,
don't
know
if
you
have
any
thoughts
about
how
we
in
Technical
Community
can
can
help
combat
some
of
these
phenomena
without
like
getting
fired.
A
A
A
Think
of
this
through
the
lens
of
firm's
preferences,
where
companies
would
on
to
a
first
order,
would
like
to
have
walled
Gardens
right,
they'd
like
to
be
able
to
capture
value
or
at
best
to
offer
apis
that
represent
a
kind
of
commercial
advantage
to
them,
but
not
a
disadvantage,
and
they
want
to
be
able
to
draw
down
that
API
when
when
it
doesn't
make
sense
to
them
commercially
and
so
on.
But
what
they
really
just
prefer,
even
more
than
having
being
forced
to
offer
an
API
that
actually
allows
for
competitive.
A
A
You
know
make
the
intrusion
detection
system
better,
do
some
stuff
right
and
on
the
other
side
they
have
the
lawyers
and
they
can
say,
just
go
and
threaten
them
under
the
Digital
Landing
Copyright
Act
or
the
Computer
Fraud,
Abuse,
Act
or
tortious
interference.
They
always
go
to
the
building
full
of
lawyers
and
there's
a
couple
of
reasons
for
this
one
is
that
like
Engineers,
are
scarce.
A
So
they're
always
going
to
be
in
kind
of
a
world
of
hurt
when
they're
in
reverse
engineering
land,
so
they
go
to
the
lawyers
and
they
say
to
the
lawyers
go
and
make
an
example
of
these
people
right.
Let
the
world
see
that
if
you
build
powerventures,
dashboard
or
Og
app
for
Instagram
or
any
of
these
other
services
that
came
and
went
very
quickly
because
of
legal
threats
and
or
lawsuits,
let
them
see
what
happens.
Let
their
potential
investors
see
what
happens.
A
Let
hackers
who
are
thinking
about
doing
a
startup
find
out
what
happens
so
that
they
don't
even
try
right.
So
my
theory
is
that
if,
on
the
one
hand,
we
have
some
mandates
and
there's
different
ways
that
we're
going
to
get
them
in
Europe,
the
digital
markets,
Act
is
already
mandating
interoperability
in
order
to
allow
competitive
new
services
to
enter
the
market
and
users
to
leave
with
low
switching
costs.
A
A
That,
like
companies,
find
it
very
hard
to
color
within
the
lines
the
kind
of
micro
incentives
are
always
to
cheat
or
to
push
against
them,
and
so,
whatever
like
settlement,
they
last
came
to
where
they
pinky
swear,
they
weren't
going
to
do
the
bad
thing
anymore,
they're
they're,
breaking
it
right
now,
and
so
they
this
opens
the
opportunity
for
new
settlements
and
those
new
settlements
can
come
with
interoperability
requirements.
And
then
we
can
also
like
get
interoperability
requirements
through
things
like
government
procurement.
A
A
That
Google
did
when
it
showed
up
and
told
websites
in
1998
that
its
user
agent
was
a
browser
and
got
copies
of
all
the
websites.
Right,
like
that.
Reverse
engineering
that
that
interoperability,
when
you,
if
you
have
restore
that
power
right
to
to
interoperate
without
permission
to
do
adversarial
interoperability,
and
you
couple
that
with
a
mandate,
then
companies
for
preference
are
going
to
color
within
the
lines
for
the
Mandate,
because
they
don't
want
their
New
Market
entrance
to
go
from.
A
Like
oh,
the
you
know,
the
API
is
down
or
it's
slow
or
it's
ineffective,
we'll
just
switch
to
scraping
and
now.
You've
just
got
to
commit
like
an
unknown
number
of
Engineers
to
to
fighting
the
scrapers
right
and,
and
so
they'll
just
say:
okay,
let's
just
like
make
the
API
work,
because
while
we
would
prefer
not
to
have
a
competitive
API,
we
would
prefer
even
more
not
to
have
an
unquantifiable
risk,
and
so,
if
we
just
like
know
what
the
rules
of
the
road
are,
we
know
how
the
API
works.
We
can
observe
it.
A
We
can
know
how
our
competitors
are
using
it.
We
can
plan
for
it.
We
can
observe
Trends
in
it.
Then
that's
we're
going
to
focus
on
now
as
to
what
technologists
individual
Technologies
can
do.
You
know
I
I
like
to
quote
my
arch
nemesis
here:
Milton
Friedman,
the
architect
of
The
Chicago,
School
of
Economics.
You
know
in
in
the
the
years
when
he
was
in
the
wilderness,
and
people
were
saying
like
Milton.
A
You
have
all
these
ideas
for
destroying
everything,
people
love
about
the
post-war
consensus
where
people
have
like
health
care
and
education
and
pensions,
and
so
on.
How
are
you
ever
going
to
convince
people
to
give
up
these
things
they
loved?
He
said:
look
there
will
always
be
a
crisis
and
in
a
time
of
Crisis
ideas,
lying
around
can
move
from
the
periphery
to
the
center
right.
Now,
whenever
there's
a
tech
crisis,
we
just
do
the
same
thing.
We
did
last
time
but
harder
and
hope
we
get
a
different
outcome.
A
I
think
that
if
technologists
are
are
out
there
saying
hey,
you
know
like
the
root
here
is.
The
of
the
problem
is
not
that,
like
Facebook
made
a
mind,
control
Ray
to
sell
your
nephew,
fidget
Spinners
and
then
like
Robert
Mercer,
stole
it
and
made
your
uncle
a
racist,
Q
Anon,
but
rather
they
just
have
a
monopoly
over
what
you
see
and
that's
a
thing
we
shouldn't
have,
and
they
got
it
through
these
technical
means
and
here's
what
the
technical
countermeasures
should
be.
A
Like
you,
come
work
for
Google
for
five
years,
figure
out
how
the
business
Works
quit
and
start
a
business
that
competes
with
Google
to
come.
Work
for
Google
for
five
years
figure.
The
business
Works
start
a
business
that
competes
with
Google
and
Google
will
buy
it
to
come
and
work
for
Google
for
five
years
and
and
get
free,
kombucha
and
massages,
but
don't
think
of
starting
your
own
business
to
come
work
for
Google
and
you're
lucky.
You
still
have
a
job
because
we
just
fired
12
000
of
your
colleagues.
A
B
B
Thank
you
we'll
appreciate
it
so
for
those
who
are
not
aware,
so
we
have
a
chat
feature
here
and
there's
a
lively
discussion
already
in
the
chat
channel.
So
if
there
are
more
questions-
or
you
know
more
more
feedback
or
more
ideas,
please
use
that
or
also
use
our
main
list
for
more
substantial
discussions.
B
A
E
I
I
Perfect,
thank
you
very
much,
sir
hello.
Good
afternoon,
everybody,
it's
a
great
honor
and
pleasure
to
speak
to
you
today
thanks
very
much
for
the
gracious
invitation.
This
is
my
first
ITF
meeting
and
already
really
exciting
session.
So
so
thanks
very
much,
my
name
is
focus
stalker
and
alongside
my
esteemed
colleague,
Bill
layer,
we
are
here
to
talk
about
ecosystem
Evolution
and
digital
infrastructure
policy
changes.
I
Well
we're
not
really
here
in
San
Francisco,
obviously,
but
we
are
joining
you
from
Swiss
Alps,
where
it
is
rather
late
and
where
I
am
and
from
an
island
in
Maine,
not
a
main
island
where
Bill
is
so
since
both
of
us
will
speak.
We
hope
that
our
connectivity
will
be
decent
throughout
the
talk
so
so
bear
with
us
and
since
bill
and
I
our
economists.
I
That
being
said,
bill
and
I
both
have
broad
experience
and
interdisciplinary
research
often
collaborating
with
Technologies
from
Fields
like
engineering
or
computer
science.
So
this
session
really
is
a
wonderful
initiative
that
brings
together
experts
from
diverse
fields
and
with
different
perspectives.
So
my
compliments
again
for
facilitating
such
a
fruitful
exchange,
but
now
without
further
Ado,
let's
dive
into
a
presentation.
Now,
as
we
are
all
aware,
we
are
in
the
midst
of
a
digital
transformation,
while
the
pace
at
which
the
digital
or
virtual
and
non-digital
worlds
are
converging
is
arguably
increased.
I
The
internet
ecosystem
has
undergone
tremendous
Evolution
and
it
has
not
only
changed
significantly,
but
it
also
continues
to
do
so.
So,
let's
briefly
take
a
look
at
some
of
these
changes.
Note
that,
due
to
the
time
constraints
of
this
presentation,
of
course,
bill
and
I
won't
be
able
to
provide
a
comprehensive
or
detailed
overview
today,
but
we
will
concentrate
on
highlighting
some
major
issues
that
we
believe
warrant
tension
aiming
to
stimulate
and
inform
further
discussion.
So
in
doing
so,
we
lean
on
the
body
of
work
we've
conducted
on
related
topics
over
the
years.
I
To
start
with,
we
observe
an
Ever,
evolving
variety
of
applications
and
services
that
differ
greatly
in
their
demands
for
Network
performance
and
capacities,
cloud-based
computation
security
and
so
on.
Arguably
responding
to
the
need
for
more
extensive
capabilities
to
efficiently
support
ever
more
demanding
applications
and
The
Wider
range
of
requirement
profiles.
If
you
will
the
way
the
internet
is
interconnected,
where
servers
are
placed,
data
is
stored
and
processed,
and
how
and
from
where
content
application
services
are
distributed
and
delivered
have
experienced
a
seismic
shift.
I
In
fact,
interconnection
points,
network
computing
resources,
servers
and
thus
content
applications
and
Computing
capabilities
have
been
moved
closer
to
end
users
and
devices.
For
example,
a
large
share
of
today's
internet
traffic
is
delivered
via
highly
distributed
serving
infrastructures
that
make
use
of
ofnet
servers
deployed
within
the
networks
of
isps
now.
This
shift
to
the
edge
points
to
a
localization
of
network
traffic
and
of
network
computing
resources.
On
top
of
that,
we
observe
a
flattening
intensification
of
the
interconnection
fabric
that
has
expanded
routing
options
and
interconnection
diversity.
I
Now,
in
today's
ecosystem,
service
provision
is
based
on
a
complex
and
evolving
amalgam
of
interconnected
and
complementary
network
computing
and
Communications
network
resources,
so
basically
cables
routers,
which
is
service
and
so
on.
Many
of
those
are
owned
and
managed
or
controlled
by
different
entities
which
are
often
competing
and
have
different
goals.
Some
are
open
and
publicly
shared
While,
others
are
dedicated
or
managing
the
proprietary
fashion.
In
recent
years,
let's
say
the
focus
of
academic
research
and
policy
debates
on
the
internet
has
some
which
shifted
towards
some.
I
If
you
will
New
Kids
on
the
Block,
like
a
group
of
non-isp
actors
such
as
large
content,
application
providers,
sometimes
the
Fortress
hypergiants
and
including
large
digital
platform
providers.
Broadly
speaking,
many
of
these
entities
have
invested
in
their
own
Global
and
proprietorily
managed
networks,
comprehensive
servers,
routers
and
cables,
physical
or
virtual
that
they
utilize
for
Service
delivery.
I
When
we
shift
our
Focus,
though,
to
what
is
sometimes
called
5G,
plus
Next
Generation
networks
and
the
need
for
an
anticipated
role
of
local
or
Edge
clouds,
things
look
a
bit
different
right
and
other
actors
are
likely
involved
in
managing
and
orchestrating
the
relevant
resources.
So
it's
complicated
and
both
bill
and
I
have
worked
in
these
issues
for
several
years,
and
we
will
continue
to
do
so
now,
but
what
so?
I
What
we
can
say
that,
as
a
direct
result
of
the
Transformations
and
shifts
I've
just
highlighted
value
chain,
constellations
and
configurations
in
today's
ecosystem
that
we're
more
complex
diversion
Dynamic.
For
instance,
we
witnessed
a
complex
fabric
of
both
vertical
relationships,
such
as
between
platforms
and
complementers,
and
horizontal
relationships
like
the
interconnections
between
networks
via
query.
This
increased
complexity
is
manifested
in
a
richer
variety
of
collaborations
and
integration
strategies,
as
well
as
complementaries
and
synonyms.
These
changes
have
been
reflected
in
changing
ownership
in
governance
structures,
industry
structures
as
well
as
competitive
and
innovational
Dynamics.
I
I
Now,
with
this
in
mind,
I
want
to
talk.
I
want
to
walk
you
through
some
Reflections
from
an
economist
perspective
again
give
the
time
constraints
we're
facing
I
will
keep
this
brief
and
high
level
and
non-exhaustive
first,
when
we
talk
about
the
essential
digital
infrastructure
required
for
a
digital
economy
future.
This
is
about
more
than
just
connectivity
in
Broadband,
more
specifically,
broadband
and
ISP
access
is
not
all
that
is
needed
for
a
digital
future.
I
Indeed,
substantial,
complementary,
Investments
necessary,
both
upstream
and
downstream,
take,
for
instance,
cloud
computing,
which
is
becoming
increasingly
critical
right,
often
local
and
at
the
edge
for
the
effective
delivery
of
numerous
application
services
that
have
stringent
low
latency
requirements,
perhaps
tactile
internet
applications
think
of
dynamic
control
problems
also,
moreover,
when
we
think
about
essential
infrastructure
for
a
digital
economy
future,
this
will
be
better
about.
Let's
say
what
we
call
it
or
Bill
calls
it
mostly
smart,
ax
application,
where
the
X
in
smartx
can
be
replaced
by
the
specific
context
benefiting
from
ICT
augmentation.
I
Examples
are
smart
grids:
small
Healthcare,
smart
cities,
smart
supply
chain
smartages,
so
you
name
it
now
there.
There
are
many
others,
and
it's
not
just
about
building
a
more
capable
video
and
entertainment
distribution
platform,
or
why
would
we
invest
especially
on
public
money?
Finally,
there's
a
need
to
consider
codependencies
and
co-evolution
of
different
Investments,
as
well
as
components
owned
and
managed
by
different
entities.
This
links
to
the
point
I
made
before
about
more
complex
and
diverse
value,
chain,
constellations
and
considerations.
I
Second,
the
majority
of
the
requisite
investment
will
be
private
and
delivered
in
a
market
driven
fashion,
while
Public
Funding
will
likely
have
a
significant
role
to
play,
a
point
in
which
bill
will
elaborate
further
in
a
minute,
it
is
crucial
to
recognize
that
investments
will
come
from
a
variety
of
stakeholders
with
diverse
objectives.
Again,
this
group
includes
access,
isps,
content,
application
providers,
cdns
upstream
isps
and
also
end
users.
I
Third,
considering
the
variety
of
stakeholders,
along
with
the
complexity
and
heterogeneity
of
applications
and
usage
environments
and
the
subsequent
diversity
of
networking
and
Computing
requirements,
we
can
foresee
the
increasing
relevance
and
necessity
for
more
and
new
forms
of
research.
Sharing,
more
complex
and
regular
needs
for
resource
allocation
and
orchestration
in
real
time
will
emerge.
I
I
On
the
other
hand,
as
we
look
towards
the
future,
digital
platforms
and
Edge
clouds
are
likely
to
assume
an
increasingly
significant
role.
This
development
May,
for
example,
involve
a
greater
degree
of
end
user
perspection,
as
well
as
the
emergence
or
introduction
of
Novel
Edge
sharing
participants.
I
Finally,
given
the
diversity
of
stakeholders
and
value
chain
constellations
I
mentioned
and
considering
the
ongoing
Technical
and
Market
changes,
we
foresee
a
need
for
more
network
sharing
agreements.
It's
essential
to
acknowledge
that
this
includes
various
active
and
passive
Arrangements
sharing
arrangements
and
that
optimal
models
are
continually
evolving
bill
and
I
recently
published
a
paper
focusing
on
some
of
these
aspects
and
with
that
I
shall
pass
the
patent
to
my
colleague,
Bill
Lear
bill.
E
That's
clear
to
policy
folks
and
economists
that
have
looked
at
this
is
the
world
really
has
changed
at
the
Legacy
internet
is
not
the
internet
of
today
and
the
economic
environment
is
not
the
environment
of
yesterday
and
a
lot
of
the
goals
and
values
that
worked
in
the
past.
That
we
want
to
preserve
mean
something
different
today
in
the
Telecommunications
space.
There's
a
bunch
of
issues
that
always
show
up
and
always
have
showed
up
and
will
continue
to
show
up.
E
If
you
ask
yourself
what
neutrality
means
in
the
next,
the
next
Network
we're
building
some
version
of
5G.
It
means
something
like
Network
slicing,
which
means
how
do
you
share
infrastructure?
Answering
that
question
is
a
hell
of
a
lot
more
difficult
than
answering
the
question
about
net
neutrality.
We
wasted
10
years
backing
away
from
the
stupid
argument
that
all
packets
should
be
treated
the
same.
Anybody
that's
involved
in
network
research,
said
all
packets
should
not
be
treated
the
same.
E
That's
what
network
Management's
all
about
it's
designed
to
sort
of
figure
out
if
the
malware
packet
or
a
DDOS
packet
get
rid
of
it.
If
it's
good
traffic
treat
it
well,
when
you
get
to
something
like
5G,
it's
a
much
much
more
complicated
problem.
Universal
service
we're
now
going
to
spend
40
billion
dollars
in
the
United
States,
putting
fiber
to
every
Broadband
serviceable
location
in
the
United
States.
Why
we're
going
to
do
that?
I,
don't
know,
I,
think
that's
crazy,
but
it's
we're
going
to
spend
40
billion
dollars.
E
E
These
are
all
big
problems
and
then
finally,
the
first
talk,
Corey
Dr
Rose
talk
was
largely
in
a
lot
of
what
like
the
focus
of
the
group,
is:
how
do
we
restore
meaningful
competition
so
that
the
future
of
the
internet
is
in
the
hands
of
a
very
small
number
of
players?
All
of
these
are
complicated
questions.
E
There's
a
lot
of
sophisticated
analysis
going
on
by
a
bunch
of
different
people
in
the
technology
sector,
developing
Technologies,
trying
to
respond
to
this
and
by
the
economists
we're
in
the
midst
of
a
major
attempt
to
redo
competition
law
in
the
United
States
Jeff
Houston
commented
and
mentioned
made
reference
to
justice.
Brandeis
that
is
called
by
some
of
the
people
is
the
Neil
brandisia
movement.
E
That's
a
battle
we'll
continue
to
have
to
fight,
it's
all
Hands-On
deck
thing
and
we
are
going
to
be
facing
a
major
collision
between
privacy
and
lawful
access
over
and
and
encryption
and
what
should
happen
there
and
that's
an
international
battle.
That's
going
on
the
blockchain,
crypto,
Financial,
regulation,
stuff
and
fintech.
That's
a
whole
other
area.
You
know
it's
temporarily
self-destructed
over
the
you
know
the
saying
bankman,
Freud
nonsense
and
then,
of
course,
AI
regulation
was
a
new
thing.
So
these
are
all
issues
go
to
the
next
slide.
E
Folker
go
next
slide,
yeah,
so
so
three
things
one
is
we're
economists
that
want
to
engage
these
debates
or
engage
these
debates
with
the
regulatory
authorities
in
in
the
U.S
and
abroad
and
standards
bodies
and
other
technical
research.
We
want
to
have
collaboration
with
technical
people.
We
hope
that
when
we,
you
know
when
you
see
the
slides,
you'll
look
at
the
papers
and
you'll
get
in
touch
with
us
about
specific
papers
and
comments.
If
any
of
the
people
who
participate
in
this
have
an
interest.
Multidisciplinary
engagement
feedback
is
absolutely
necessary.
E
Corey
doctorow's,
point
of
view
and
things
he
expressed
in
the
beginning
is
a
point
of
view.
It
is
certainly
represented
by
a
number
of
people.
It
is
far
not
I
would
say
a
mainstream
point
of
view
or
a
dominant
point
of
view,
amongst
most
of
the
people
that
are
typically
involved
in
regulatory
policy
debates
and
so
to
have
these
kinds
of
arguments
and
have
them
be
able
to
take
place
in
a
coherent
way.
E
We
need
people
involved
in
this
fundamentally
we're
going
into
a
future
where
asymmetric
information
and
the
measurements
that
happen
in
that
world,
so
that
you
can
do
informed,
evidence-based
policy
making
are
going
to
be
critical
and
if
the
measurements
and
the
data
and
the
evidence
people
bring
to
the
table
actually
influences
decisions
as
opposed
to
just
being.
You
know,
nonsense,
blather,
like
so
much
of
the
crap,
that's
in
the
Indus
in
the
media,
the
popular
media.
E
Today,
that's
driven
so
many
people
away
from
Reading
newspapers
because
they're
like
what's
the
point,
it's
just
horse
race.
If
those
things
are
going
to
have
mean
any
thing,
that
does
is
going
to
be
strategic
and
you
have
to
think
about
the
Strategic
implications
of
Who's
involved
and
we
need
to
think
about
designing
an
ecosystem
which
means
that
that
the
provenance
of
the
metrics,
the
measurements
and
and
careful
thinking
about
how
these
things
are
used
and
analyze.
E
It's
going
to
be
important
and
that's
going
to
take
a
lot
of
help
from
academics,
technologists
that
really
understand
how
the
systems
work
and
social
scientists
and
economists
that
understand
how
these
things
are.
Thinking
about
how
these
things
actually
impact
the
real
world
and
then
finally,
the
ability
to
frame
X,
anti-rules
that'll
determine
how
the
networks
operate.
That
may
have
always
been
a
suspect
goal.
It's
completely
impossible
today.
The
state
space
that
we're
going
into
in
a
global
internet
where
the
everything's
digital
is
just
way
too
large.
E
We're
gonna
have
to
figure
out
build
mechanisms
that
are
capable
of
detecting
when
harms
are
happening
and
developing
more
reactive,
market-based
ways
to
respond
to
that.
In
looking
at
the
platforms
we
we
could
sit
there
and
pick
out
individual
platforms
and
say
that's
a
bad
actor
blow
up.
Facebook
blow
up
Amazon
blow
up
Microsoft
blow
up
Apple
blow
up
Cisco
blow
up
while
away.
Where
do
we
start?
What
are
the
companies
what's
the
ecosystem?
E
E
You
shouldn't
care
about
a
lot,
there's
a
lot
of
there's
a
lot
of
thinking
that
needs
to
happen
there
and
the
right
way
to
do
that
is
I
think
to
have
a
more
informed
dialogue
from
Engineers
people
in
the
communities
building
the
stuff
talking
to
the
policy
makers,
because
the
the
pot
we
we
can't
do
what
what
what
economists
have
said
for
a
long
time.
Regulatory
policy
is
the
captive
of
Economist
wisdom
from
20
years
ago.
B
J
So
good
morning,
everyone
tobs
feedback
max
Park
Institute
for
informatics.
Speaking
for
myself
and
not
my
affiliation,
you
mentioned
that
you're
expecting
significant
investment
from
the
private
sector
and
infrastructure
I
recently
claimed
that
infrastructure
operation
in
a
careful
and
maintainable
manner
is
incompatible
with
the
concept
of
profit.
E
The
telecoms
network
was
built
as
as
a
public
as
a
national
utility
in
most
countries
and
in
the
U.S
we
built
it
with
a
t
which
is
a
private
Monopoly,
and
you
know
we
argue
about
whether
that
was
a
good
model
or
not.
But
it's
how
we
did
it
almost
everybody
kind
of
agreed.
We
need
to
move
against
away
from
that
and
had
we
not
moved
away
from
that,
we
never
would
have
ended
up
with
the
internet.
E
The
way
we
have
today
and
the
original
investment
in
the
internet
was
a
small
amount
of
investment
at
the
edges
that
took
advantage
of
the
public
utility
aspects
of
the
affiliate.
Now
we're
in
this
place,
where
we
may
say
that
significant
portions
of
the
investment
ought
to
be
Public
Utilities
and
that's
a
debate
and
there's
some
people,
smart
people
that
actually
believe
that
and
I
wouldn't
dismiss
that
point
of
view.
E
But
I
would
say
that
if
you
thought
about
trying
to
actually
move
us
in
that
direction,
you're
going
to
need
a
really
good
argument
for
how
you
get
there
and
you're
going
to
face
some
very,
very
resistance
from
basically
everybody
that
that
that's
investing
in
this
space
And.
The
fact
of
the
matter
is,
if
you
actually
look
at
the
empirical
numbers
most
of
the
money
that
they're
talking
about.
E
If
you
look
at
the
the
investment
by
private
actors,
it
absolutely
dwarfs
the
public
investment
that's
available
here
so
either
we
say:
there's
a
whole
lot
of
money.
That's
got
to
come
from
the
public
sector,
general
tax
revenues
that
are
put
towards
this
public
utility
and
then
it's
regulated
or
we
have
to
figure
out
how
to
manage
private
actors,
and
if
we're
managing
private
actors,
we
have
to
figure
out.
You
know
how
we
allow
them
to
see
an
incentive
in
doing
what
they're
doing
so.
E
I
C
E
But
who
would
we
stand
up
the
itu
or
a
completely
new
formulated
icam
or
the
World
Bank
or
or
Amazon,
or
any
of
these
other
guys?
None
of
those
people
are
great
candidates.
The
best
hope,
I
think
is
in
standards
and
intelligent
architectures
that
allow
options,
and
one
of
the
problems
about
getting
standards
is
if
someone
came
out
with
a
good
technology,
is
how
do
you
get
people
to
adopt
it
and
deploy
it?
That's
a
big
economic
Challenge
and
a
tough
one.
E
So
definitely
entry
ought
to
be
feasible,
but
you're
you're
not
going
to
get
entry.
You
know
there's
also
security
reasons.
I
can't
just
sit
there
and
say
you
know:
anybody's
got
to
make
everybody's
got
to
make
any
source
code
has
to
be
public
and
it
all
has
to
be
totally
shared.
That's
just
not
going
to
work,
there's
a
lot
of
room
for
open
source,
but
not
everything's,
going
to
be
open
source.
B
Okay,
great
yeah,
thanks
again,
for
example,
for
for
sharing
your
book.
There
are
really
many
interesting
things
to
unwrap
and
respond
to
in
your
presentation,
I
I
hope
we
can
do
that
on
the
mail
list,
and
we
will
be
looking
forward
to
discussing
these
things
more
with
you
in
the
future.
Thanks
again
for
coming.
K
C
K
Once
please,
so,
a
month
ago
it
was
I
was
at
the
decentralized
web
camp
Camp
number
two
hours,
North
from
San
Francisco,
Tim
burnsley
was
there,
but
also
along
a
lot
of
young
people,
the
the
web
movement
so
to
speak,
and
they
kind
of
liked
the
atmosphere.
Things
are
crumbling
or
burning
and
we
have
Alternatives,
but
at
the
same
time
you
realize
they
played
the
same
game
as
the
others.
They
all
run
around
and
say.
Look
at
me.
K
Look
at
me,
look
at
me
and
we
are
the
better
ones
come
to
us
and
if
people
are
really
doing
that,
how
do
the
other
people
find
where
we
are
next
slide?
Please?
So
that
is
what
is
at
the
core.
We
have
a
lot
of
options
now
and
if
you
do
the
move
to
go
to
another
platform,
you
have
to
leave
traces.
You
have
to
inform
orders
how
you
can
be
reached,
how
you
keep
the
family
together,
how
important
things
are
kept.
K
So
even
if
things
are
burning,
there
are
some
things
you
have
to
preserve
and
the
networking
next
slide.
Please.
We
know
that
this
is
the
Rendezvous
problem.
Is
the
initial
problem
of
even
before
you
establish
communication?
You
know
how
to
know
where
to
communicate,
how
to
communicate.
It
is
that's
classic
example
of
two
people
meeting
in
a
large
Park
if
both
decide
to
not
move
around
and
find
each
other,
they
will
never
find
each
other
if
both
move
around
the
chances
are
very
slim,
maybe
it's
better
that
one
party
stays
the
other
searches.
K
But
how
do
you
coordinate
that
you
haven't
communication,
so
there's
really
an
intrinsic
hard
problem
there
in
networking
we
have
that
the
same
we
use
standards
and
everybody
doesn't
have
to
think
about
how
to
do
things.
We
have
to
address
that
in
that
now
burning
environment,
that
we
have
next
slide
please.
K
So
here
is
a
suggestion
to
fill
that
small
Gap.
That
I
think
is
there,
and
we
have
to
address
that.
We
shouldn't
think
about
having
a
kind
of
a
common
element
where
everybody
can
look
up
where
everybody
else
is.
We
would
just
build
a
new
platform.
We
also
shouldn't
overload
existing
platforms
like
DNS,
to
have
a
record
where
everybody
can
be
located.
These
are
single
points
or
failures
and
in
many
cases
we
know
DNS
is
not
working
when
it
should
and
when
it's
most
urgently
needed.
K
K
So
that
is
the
way
I
would
phrase
it
think
about
a
generic
right
that
each
human
entity
is
allowed
to
send
200
bytes
a
month.
That
is
the
novelty
that
you
are
allowed
to
ingest
and
to
use
any
means
to
convey
that
information.
I,
don't
talk
about
protocols
or
infrastructures
to
make
that
happen,
but
just
an
entitlement,
and
once
you
have
it,
you
will
find
your
ways.
Orders
will
come
and
provide
ways
to
do
that
and
why
I
brought
here
my
antenna
here.
This
is
a
shortwave
antenna,
typically
shortwave.
K
With
with
a
laundry
clip,
I
move
it
to
20
meters
bandwidth
with
that
antenna
in
such
a
transceiver
five
Watts
you
reach
Japan
and
back.
So
that
is
what
you
can
do
end
to
end
without
intermediaries.
So
this
is
not
science
fiction.
We
have
Technical
Solutions
to
let
people
speak
at
long
distances
and
there
are
ways
we
should
imagine
to
push
that
forward.
The
we
they're
also
kind
of
the
so-called
pocket
switch
networking
where
people
run
around
with
USB
sticks.
All
this
is
okay.
K
K
So
how
would
they
did
I
came
up
with
that
number?
It
might
be
not
an
exact
number.
You
could
use
that
exactly
for
spreading
your
new
social
media
handle
where
you
are,
but
we
can
also
have
cryptographic
usage
of
that.
If
you
want
to
say,
oh
I
had
a
compromise
on
one
of
the
key
use,
the
other
one.
You
can
also
have
a
redirection
how
we
call
it
in
computer
science,
saying
I'm
out
of
town.
Please
contact
this
person,
but
also
for
emergency
situation.
K
We
need
kind
of
something
that
doesn't
depend
on
turning
off
this
Skynet
or
whatever,
and
the
other
reason
how
white
to
go
low
is
not
because
we,
it
would
be
nice
to
have
more
but
200
bytes
a
month.
You
can't
do
much
harm.
This
is
something
that
is
still
manageable.
Next
slide,
please
I'm
almost
at
the
end.
So
what
is
the
overall
volume
volume?
It's
quite
small.
K
If
you
collect
everything,
if
you
do
the
Justice
the
multiplication
everything
fits
on
one
of
these
small
USB
sticks,
then
it's
also
something
that
is
not
really
novel.
Others
do
that
already
for
trust
reasons,
because
you
can't
have
a
trusted
Central
entity,
all
server
certificates
are
collected
and
brought,
and
globally
rebroadcasted
and
kept
in
multiple
repositories
that
everybody
can
check
that
everybody
is
doing
the
same
thing
and
you
need
broadcast
to
do
that,
and
this
is
the
the
industry
paying
for
that.
K
And
finally,
so
if
about
the
feasibility,
you
could
also,
if
that
is
a
common
right
to
do.
Why
not
also
ask
the
platforms
to
provide
you
this,
so
Apple
provides
emergency
channels.
Why
not?
The
200
buys
a
month.
It's
really
not
a
big
load
on
top
of
that,
what
they
want
to
offer.
You
next
slide
and
that's
the
final
one
I
think
so
that
is
at
the
core.
K
So
burning
everything
down
is
fine,
but
there
are
things
which
have
to
stay,
so
we
can
keep
contacts
to
solve
the
Rendezvous
problem
and
make
it
really
more
easier
to
do
that.
Switching,
don't
think
about
creating
the
new
platform,
but
expressing
the
needs
and
the
right
to
use
such
a
channel,
and
if
you
use
a
lighthouse
thing,
don't
build
the
giantic
lighthouse
that
collects
all
that
information
just
enable
allow
each
individual
persons
if
they
want
to
do
so
to
operate
their
own
channel.
K
B
L
Hey
Eric
Oswald,
George,
Mason
University.
That's
very
interesting
I
like
the
idea,
but
I
had
a
question.
So
it
sort
of
feels
a
little
bit
like
the
forest
tree
falling
in
the
woods
and
no
one's
there
to
hear
it.
So
do
you
need
a
platform
to
capture
the
broadcast
in
case
the
recipient?
You
want
to
hear
it
wasn't
listening
when
you
sent
it.
K
Yes,
but
I
mean
the
platform
you
could
try
to
have
several
platforms
offering
that
you
would
not
build
one
to
do
it.
I
agree
the
store
and
forward
or
what,
as
mentioned
the
asynchronous
communication
offered
by
Facebook
and
the
others,
is
one
of
the
value
that
helps
in
the
Rendezvous.
We
would
have
to
do
that,
but
who
knows?
Maybe
these
devices
they
just
listen
a
month
long
and
they
filter
on
what
you
are
interested
in
so
also
technically.
We
could
even
go
around
the
store
and
forward
necessity,
but
we
would
have
to
explore
that.
B
K
That
is
typically
observation.
Trust
is
created
out
of
band
and
then
you
use
it
in
the
system.
In
the
past
you
for
a
service
certificate.
This
is
really
90s.
You
had
to
physically
show
up
at
the
certificate
issuer
and
say
and
prove
that
you
are
representative
of
that
University
to
get
a
certificate.
For
that
now
we
have
a
let's
encrypt,
where
it's
just
proving
that
you
have
the
DNS
name.
So
it's
out
of
bands
in
principle
and
the
decentralized
people
do
that
they
have
signing
parties.
K
You
know
you
with
whom
you
talk
and
from
that
on
later
on,
you
can
reconnect
later.
If
it's
out
of
the
blue
and
with
talking
to
somebody
you
never
met,
then
you
have
a
problem,
but
that
is
not
addressing
this
one.
It's
we
have
already
some
trust
anchors
and
you
want
to
write
continue
using
that
you
simply
lost
the
person
out
of
sight
and
you
want
to
reconnect
later.
K
D
Thanks
I
put
myself
in
the
queue
sorry
that
security
is
based
on
identity.
You
mentioned
earlier
like
a
teen
assets
platform
or
somehow
centralized,
but
there
is
an
idea
is
merely
just
the
naming
system,
allocating
names
right
and
without
identifier.
How
would
you
attach
your
security
onto
it?
You
mentioned
the
institutions
can
issue
your
certificate
with
my
name
and
therefore
you
do
need
a
namespace
correct
if
nothing
else,
and
what
would
be
that
the
namespace.
K
Well,
the
namespace
I'm
using
right
now
is
from
Securus
cardboard.
It's
a
ed25519
key
randomly
generated
and
I
have
other
peoples
when
I
have
a
signing
party.
No
that
this
is
me.
I
can
transfer
that
information.
I
can
have
trusted
friends
which
collect
maybe
that
information
we
can
do
that
Grassroots
up
if
we
know
each
other,
if
it's
a
different
story,
if
we
talk
about
banking
relationship,
things
like
that
there
there
must
be
probably
some
trust
industry
thing
we're
vouching,
for
this
is
really
the
ed25519
from
this
bank
or
not.
F
J
F
K
The
end
I
was
anticipating
a
little
bit
that
yes,
I
did.
I
wasn't
participating
a
little
bit
that
comment
and
it
made
me
smile
because
I
have
found
out
that
the
radio
amateurs
that
have
internalized
the
Restriction
the
government
put
on
them
since
hundred
years,
and
now
they
are
the
cops
on
the
airwaves
and
they
will
try
to
make
sure-
and
you
seem
now
to
be
taking
the
role
of
the
government
and
making
sure
that
we
do
not
overstep
I.
K
Think
the
the
answer
is
that
if
you
think
about
the
device
here,
you
could
do
a
lot
of
problems,
and
we
have
had
that
email
that
you
shouldn't
use
your
personal
hotspot
here.
So
we
have.
We
like
give
people
already
that
technology
to
disrupt
things.
If,
if
you
make
a
cheap
device
for
doing
that,
minimal
global
Outcast
people
will
typically
not
open
it,
not
hack
it,
and
even
if
it's
worldwide,
a
hundred
thousand
people
abuse
that
the
rest
of
the
vast
majority
would
simply
use
the
tool
as
it
was
built.
B
K
You
do
a
lot
of
of
data
keeping
cleaning
up
keeping
up
to
date.
Your
contact
lists.
This
is
some
real
convenience
that
you
get
by
the
centralized
platforms
like
keep
that
for
you,
and
so
you
see
a
lot
of
manual
stuff
that
you
have
to
do.
I
I
stopped
doing
signing
parties
now,
except
when
I
came
back
to
to
these
decentralized
things,
but
it's
much
more
difficult
to
easily
grow.
Now,
I
just
exchanged
phone
numbers,
we
have
I
mean
these
contacts
still
exist.
K
It's
a
way
to
bootstrap
trust,
so
I
wouldn't
single
it
out.
It
has
become
much
easier
than
in
the
past
in
security
Albert.
We
have
a
you
scan
the
QR
code
on
the
on
your
friend's
phone
and
that's
all
you
need
you
never
have
to
type
in
something,
so
we
can
make
that
frictionless
or
easier
than
it
was
in
the
past,
with
the
typical
old
90s
style,
designing
parties,
okay,.