►
From YouTube: Decentralizing the Future of Compute: A Keynote by Tom Trowbridge at Disco Summit 2023
Description
Join Tom Trowbridge, as he takes us on a journey through the landscape of web3 technology at the Disco Summit 2023. In this keynote presentation, Tom sheds light on the importance of decentralizing serverless compute and the role that Fluence is playing in this transformative field. He highlights the potential risks of centralized systems and the importance of building a decentralized ecosystem. He also discusses the challenges of scalability, verifiability, and compute capability, and how Fluence is addressing these challenges. The presentation concludes with a look into the future of Fluence, setting the stage for the next evolution in Web 3 technology.
DISCO SUMMIT, for the founders, teams and investors building the decentralized computing stack.
Hosted Apr 26, 2023 in Austin, TX by keiretsu and Web3 Working Group.
Web3 Working Group is a nonprofit organization that educates about web3 infrastructure and how it is distinct from cryptocurrency. We educate regulators and the public at-large about how web3 will permanently restore user control of cyberspace, replacing centralized control with transparent and permissionless protocols through incentives aligned with users. We focus on Blockchain / Cryptocurrency.
🎆FOLLOW US🎆
Website: https://web3wg.org
Twitter: https://twitter.com/web3wg
Tiktok: https://www.tiktok.com/@web3wg
Instagram: https://www.instagram.com/web3wg
Twitter: https://twitter.com/amyofalexandria
Instagram: https://www.instagram.com/amyofalexandria
A
A
Co-Founder
of
fluence
labs
evgeny,
another
co-founder
is
in
the
back.
So
if
you've
got
very
hard
questions,
ask
evgeny
not
me,
and
we
have
Nadia
also
here,
helping
orchestrate
everything
so
great
to
be
here,
I'm
going
to
talk
about
fluence
what
is
fluence
and
why
is
it
important
and
Echo
some
of
the
things
that
we've
heard
from
other
other
speakers
as
well,
so
fluence
is
a
decentralized
serverless,
compute
platform,
and
so
what
does
that
actually
mean?
A
And
so
first
I
think
there
are
a
lot
of
words
being
thrown
around
and
people
even
in
this
room
will
have
different
definitions
potentially
so,
first,
let's
talk
about
what
decentralized
means
all
right,
and
so
it's
very
important
we're
on
the
same
page
on
this,
because
different
people
have
different
definitions.
So
for
us,
decentralized
has
three
at
least
key
attributes,
but
three
are
minimum,
which
is
is
verifiable,
and
so
that
makes
it
trustless
that's
a
critical
piece.
It
also
is
Unstoppable
and
which
also
is
another
way
of
saying
censorship
resistant.
A
If
you
are
not
censorship,
resistant
I'd,
say
you're,
not
decentralized,
and
that's
a
foundational
premise
and
you
have
to
be
fault.
We
also
think
fault
tolerant,
and
these
are
not
always.
These
are
also
not
always
the
same
things.
You
could
be
fault,
tolerant
and
still
be
subject
to
censorship
and
vice
versa.
So
these
are
all
attributes
you
need
to
have
and
we
are
focused
on
building
and
what's
a
compute
platform,
a
compute
platform
basically
replicates
the
cloud
stack
and
so
think
of
it
as
everything
that
you
need
to
run.
A
Applications
besides
storage,
which
we've
have
we've
heard
Sam
from
our
weave
and
we've
heard.
Obviously
you
know
filecoin
well,
but
you've
got
to
do.
You
need
a
lot
besides
storage
to
actually
run
applications,
and
that's
what
compute
is,
and
so
you
know
what's
what's
in
more
detail,
and
why
are
we
doing
this?
And
so
what's
our
objective?
The
objective
is
to
run
arbitrary
code
without
the
limitations,
without
limitations
with
a
similar
speed
and
utility
as
the
cloud
all
right.
A
That
is
a
significant
goal,
so
we
want
to
basically
build
the
ability
for
people
to
run
applications
at
the
same
speed
as
a
cloud
but
being
decentralized,
and
that
is
a
when
achieving
that
is
significant,
and
so
a
couple
more
detail
on
sort
of
what
we're
doing
what
flips
does
it
can
run
anywhere
on
the
edge
on
your
on
a
browser
locally.
There's
no
consensus
required
I'm
going
to
talk
about
this.
It's
highly
performant
it's
off
chain
which
is
critical
and
fault,
tolerant,
secure
and
importantly,
it's
compatible
with.
A
You
know,
multiple
kinds
of
inputs
or
outputs,
whether
they
be
l1s
or
different
storage
layers,
and
so
I
want
to
put
this
slide
up
quickly,
to
make
to
emphasize
one
earlier
point
about
off,
Chain
versus
on-chain
and
so
in
this
ecosystem.
Here
everyone
considers
by
default,
blockchain
and
yet
key
components
of
decentralization
have
very
little
to
do
with
blockchain.
We've
got
to
take
a
step
back
and
remember.
The
first
decentralized
application
of
scale
had
nothing
to
do
with
blockchain
and
long
period.
A
Satoshi
was
Napster
right,
so
decentralization
and
decentralized
applications
predated
blockchain
and
one
of
the
most
successful
scaled
decentralized
applications.
Today
is
ipfs,
which
or
sorry
filecoin
built
on
ipfs,
which
is
also
off
chain.
It
uses
chain
to
validate,
verify
and
compensate,
but
it's
actually
off
chain,
and
so
these
are
important,
so
decentralized
Computing
we
have
today,
it's
called
smart
contracts,
works
very
well
and
there's
massive
amounts
of
value
which
is
transacted
on
it,
but
it
has
significant
limitations
right.
A
It's
deterministic,
it's
constrained,
it's
expensive,
so
99.9
percent
of
the
web
runs
on
general
purpose:
compute,
that's
not
suitable
for
blockchain
smart
contracts,
given
their
limitations
and
that's
what
fluence
is
trying
to
solve
and
has
solved
is
building
arbitrary
compute.
That
is
also
decentralized
and
is
not
on
chain,
and
so
you
know
why
bother?
Why
are
we
doing
this
like?
What's
what
are
the
benefits
of
this?
Like
it's,
not
just
an
exercise
as
a
academic
experiment.
There
got
to
be
reasons,
and
so
there's
a
couple
key
reasons
here.
A
First,
it's
cheaper
and
when
you
have
decentralized
compute
marketplaces,
we've
heard
from
Akash
it
gets
a
lot
cheaper.
So
that's
point
one!
You
want
things
to
be
cheaper.
You
also
want
to
be
free
from
cloud
lock-in
when
those
Cloud
ecosystems
are
very
easy
and
cheap
to
start,
as
we
all
know,
but
then,
as
you
scale,
the
pricing
ramps
up.
So
you
want
to
be
out
of
the
cloud
ecosystems.
A
If
you
can,
you
want
to
be
censorship,
resistant,
de-platforming
resistant
and
you
want
to
have
Innovation
by
being
able
to
build
without
being
without
fear
of
your
API
or
the
API.
You
build
on
being
cut
off
or
terminated,
so
we
think
you
can
actually
enhance
the
Innovation
dramatically
with
a
decentralized
decentralized
compute.
So
talk
about
a
couple
of
these
one
at
a
time,
so
price
I
don't
have
to
spend
too
much
time
on
this.
A
One
I
think
people
kind
of
grasp
that
when
you
have
a
Marketplace
that
drives
price
down,
that's
pretty
clear
and
so
there'll
be
we
have
a
compute
Marketplace
influence,
there's
others.
You
know
that
exist
and
will
exist,
and
that's
great.
We
have
a
very
clear
guide
to
follow
with
both
our
weave
as
well
as
filecoin.
You
know.
The
storage
cost
on
filecoin
is
about
one
percent
of
Amazon
S3
right.
A
So
there's
a
massive
massive
savings,
and
so
when
you
combine
the
incentives
of
a
Marketplace
price,
automatically
almost
automatically
goes
down
next
Cloud
lock-in,
you
know,
Amazon
has
terrific
margins
they're
growing
significantly,
as
is
Google
as
azer,
that
money
doesn't
come
from
anywhere
right
from
nowhere.
That
comes
from
all
the
applications,
all
the
users
there
that
are
actually
paying,
whether
it's
directly
or
indirectly,
through
higher
prices
to
their
customers.
A
You
know
there's
no
with
no
switching
costs,
because
it's
all
open
that
drives
price
down,
because
you
cannot
lock
people
in
so
whenever
there's
no
barrier
to
switching
your
cost
and
your
price
has
to
fall
because
there's
you
can't,
you
can't
retain
the
People
based
on
a
higher
price,
and
that
is
a
critical
piece
of
what
we
are
building
in
wide
Cloud
lock-in
and
eliminating
Cloud
lock-in
is
so
important
is
because
of
pricing.
Next
is
z,
platforming
and
I.
Think
we've
seen
centralized
companies
get
de-platformed,
and
you
know
parlor
is
an
example
of
that.
I.
A
Don't
even
hear
is
weeping
for
parlor
being
D
platform,
but
it
shows
the
power
of
cloud
compute
in
their
58.
You
know
terms
of
service
agreement
that
can
basically
be
pulled
out
from
a
company
with
very
little
notice,
so
with
a
decentralized
compute
Network.
You
cannot
de-platform
somebody
it's
a
decentralized
network
as
long
as
there's
any
provider
any
host
willing
to
serve
that
compute
compute
exists
that
application
exists,
and
that
is
that's
that's
a
critical
component.
A
This
is
less
talked
about,
but
is
equally
important.
Our
whole
ecosystem
here
has
a
deep
platforming
risk.
The
Secret
of
web
3
is
most
of
web
3
applications
rely
actually
on
centralized
apis
and
rely
on
Amazon,
and
so
you
know,
infuria
runs
on
Amazon
right.
So
this
is
something
we
don't
like
to
talk
about,
but
we
all
need
to
solve
design
centralized
compute
because
we
are
at
risk.
A
The
longer
we
run
on
Amazon-
and
all
you
have
to
do-
is
pay
attention
to
a
guy
named
Gary
Gensler,
to
know
that
we
all
run
risks
given
regulatory
regulatory
regime,
and
so
getting
in
a
decentralized
ecosystem
is
critical.
The
other
point
is
the
centralized
back
ends
of
dapps
also
exists
where
they
can
run
it
in-house.
It
may
not
be
on
the
cloud,
but
they
may
run
the
server
themselves.
That
has
all
kinds
of
other
risks
and
certainly
isn't
decentralized.
A
So
the
last
kind
of
Point
here
on
the
centralized,
compute
kind
of
flaws
is
the
Innovation
it
prevents,
and
so
some
people
here
may
remember
you
used
to
be
able
to
build
applications
on
top
of
Twitter.
You
used
to
be
able
to
build
on
top
of
Facebook
that
led
to
lots
of
interesting
applications.
Those
are
all
shut
down
when
you
can't
build
on
a
centralized
company.
Well,
you
you
can't,
because
you
can
be
cut
off
at
any
point
with
decentralized
compute.
A
Those
applications
run
and
you
can't
be
turned
off
that
inspires
a
terrific
amount
of
innovation,
because
people
are
free
to
build.
No,
they
won't
be
cut
off
and
can't
be
cut
off.
That
is
an
innovation
accelerant,
which
we
can't
really
forecast
what
that
means.
But
we
know
it's
significant,
and
these
are
the
reasons
why
we
think
we
decentralize
serverless
compute
is
the
next
iteration
of
compute.
We
start
off
with
mainframes,
then
over
a
decade
or
two,
we've
moved
to
the
centralized
cloud
cloud
and
that's
been
enormously
lucrative
and
enhancing
move
across
multiple
ecosystems.
A
But
the
next
web
is
a
decentralized
serverless,
decentralized,
serverless,
compute
platform
and
that's
where
we
all
go
for
the
reasons
I've
mentioned.
But
why
aren't
we
there
now
and
why
hasn't
this
happened?
It's
because
it's
hard
and
why
is
it
hard
it's
hard
because
you
want
to
have
generalized
compute
capability,
with
verification
and
being
scalable
and
those
three
things
are
very
hard
to
do,
which
is
why
they've,
for
example,
you
have
verifiability
on
blockchain,
but
you
don't
have
scalability
and
you
don't
have
generalized
compute
capability.
You
can
also
have
Central.
A
You
also
can
have
generalized
compute
capability,
easily
decentralized,
but
with
no
verification,
that's
also
not
ideal.
So
putting
and
wrapping
these
three
things
together
is
what
you
need
and
that's
very
hard
and
so
we've
solved
the
verification
without
blockchain
through
a
variety
of
different
methods.
I'll
mention
I'll
talk
about
in
a
minute,
but
these
are.
This
is
why
it
hasn't
happened
so
far.
So
we,
our
solution,
has
two
pieces:
there's
a
developer
platform
and
there's
a
Computing
Marketplace,
and
you
need
both
of
these
things
to
work.
A
And
importantly,
it's
not
just
the
compute
Marketplace
or
just
the
developer
platform,
it's
a
comprehensive
solution,
and
so
first
what's
the
developer
platform
and
we've
basically
built
a
decentralized
version
of
the
AWS
and
the
cloud
stack,
and
so
that
allows
native
composition
of
decentralized
applications
completely
free
of
the
cloud.
And
so
it's
not
a
Computing
marketplace
where
you
rent
a
server
somewhere
else
and
run
your
application.
This
is
actually
the
the
code
that
allows
you
to
compose
applications
natively
in
a
decentralized
manner,
and
that
is
a
you
can
see
the
different
pieces
here.
A
But
these
all
these
have
been
have
been
written
and
are
are
complete,
and
the
next
piece
of
this
is
the
Computing,
Marketplace
and
I.
Think
everyone
here
knows
the
Computing
Marketplace,
you
understand
the
price
Discovery
elements
of
it.
This
is
I,
think
you
know
pretty
powerful.
We
think
this
is
for
us.
It's
not
about
laptops
connecting.
We
think
not
unlike
Akash,
that
data
centers
that
are
not
you
know.
Amazon
scale
will
participate
in
these
types
of
marketplaces.
We
also
know
there's
a
terrific
amount
of
compute
capacity
at
some
of
the
storage
providers.
A
That
is
unused,
that
our
real
potential
Partners
as
well,
so
we're
confident
in
the
ability
to
find
this
compute.
This
unused
compute
that's
not
kind
of
retail,
necessarily
compute,
and
we
think
that
is
the
the
real
solution
to
how
you
build
a
decentralized
Computing
Marketplace.
This
I
want
to
put
in
because
this
is
what
makes
it
all
work
you
need
to
have
verifiability
and
incentives,
and
so
first
there's
there's
three
pieces
to
this,
but
proof
of
processing.
The
aqua
proof
basically
proves
that
the
verification
of
execution
has
happened.
A
A
And
so
where
are
we
in
this?
And
what
have
we
done
so
far
so
test?
Nets
live
decentralized
execution.
Protocol
Aqua
is
written
functioning.
Some
people
here
have
used
and
played
with
it.
This
2023
is
a
big
year.
Fluence
has
been
around
since
day
since
2017
so
been
building
for
a
long
time
and
have
raised
15
million
over
that
period
of
time.
But
this
year
we're
going
to
launch
it
down.
A
Mainnet
is
scheduled
for
Q4
verifiability
and
then
the
compute
economy
will
be
live
this
year
and
in
2024,
probably
early
in
the
year,
we're
adding
a
couple
of
languages
and
also
the
ZK
proofs.
So
that
is
that's.
That's
the
core
of
what
fluence
is
thank
you
for
paying
attention.
I.
Think
the
main
point
I
want
to
mention
is
that,
like
the
web,
when
Tim
berners-lee
was
starting
it,
he
didn't
know
exactly
what
would
happen.
He
knew
that
if
he
opened
up
these
types
of
hyperlinks,
innovation
would
come
right.
A
He
didn't
couldn't
predict
what
happened
and
we
think
the
same
thing
about
fluence
and
about
compute
when
we
open
up
compute
allow
people
to
build
on
each
other
in
a
permissionless
and
trustless
way.
We
can't
predict
exactly
what
will
happen,
but
we
know
it
will
be
big
and
powerful
and
we
also
know
we
need
to
do
it
and
if
we,
because,
if
we
don't
we're,
subject
to
Amazon
and
Google
for
the
rest
of
Eternity,
which
is
not
what
we
want.
So
thank
you,
foreign.