►
From YouTube: New Revenue Stream for FIL Miners | FIL Paris 2023
Description
Speaker: Dmitry Kurinskiy
#DecentralizedStorage #filecoi̇n #filecoinfoundation #web3 #blockchain #DWeb #cryptocurrency #crypto #decentralization
A
A
A
A
You
want
to
be
a
full,
tolerant,
secure,
asynchronous.
That
means
that
transactions
or
operations
or
function
calls
shouldn't
be
in
a
sequence.
We
should
be
able
to
run
them
in
parallel
if
you
want
to,
and
we
want
it
to
be
compatible
with
a
multiple
data
inputs
and
outputs
like
ipfos,
falcon
or
we've
different
chains,
other
protocols,
proprietary
data
sources
and
so
on
so
forth,
because
mainly
we're
about
being
the
clouds
not
about
doing
everything
for
all
the
world.
A
So.
Speaking
about
the
consensus,
we
can
say
that
the
rod
two
very
different
kinds
of
distressed
compute.
On
one
hand
we
have
blockchain
compute,
we
have
a
smart
contracts
on
chain
and
a
fundamentally
is
sufficient
for
deterministic
processing
for
a
redundant
constrained,
expensive
expensive.
Why
expensive?
Because
we
repeat
the
same
competition
many
times
and
the
same
data
mean
times
so
there
are
some
limitations
which
also
open
possibilities,
but
they
are
constraints
for
the
general
purpose
compute,
which
is
99
of
all
the
well
wearable
compute.
A
A
How
could
it
be
cheap,
so
fluence
runs
on
a
digitalized,
compute
Marketplace
and
anyone
any
anyone
who
has
a
spare
CPU
time
can
contribute,
can
be
the
capacity
open
protocol,
eliminates
switching
costs
between
providers
and
with
no
switching
costs.
There
is
competition
which
drives
down
the
compute
costs
yeah,
and
we
already
saw
how
it
works
with
a
filecoin.
A
And
for
a
prediction:
clouds:
it's
not
the
keys.
If
you
providers
of
cloud
services
control
the
majority
of
the
market
and
their
revenue
is
growing
pretty
fast,
because
once
you
enter
it,
one
of
the
computer
providers,
one
of
the
cloud
providers
it's
very
hard
to
live,
it's
very
hard
to
switch,
and
that's
why
destroys?
Compute
networks
are
good
because
providers
can
be
changed
at
will
not
switching
costs,
no
login,
because
that's
the
protocol.
A
Another
point
is
that
we
have
de-platforming
risks.
Clouds
can
shut
off
an
application
for
any
reason,
and
not
only
clouds.
Apis
can
also
burn
the
users
of
these
apis
or
they
can
change
or
shut
down
the
API
itself
and
the
majority
of
infrastructure
we
have
three
decentralized
infrastructure
is
either
the
centralized
applicants
of
the
apps
Aura
is
centralized
developer
apis
that
the
apps
use
with
decentralized
compute.
We
have
no
API
to
shut
off,
not
the
platforming
risk,
and
another
point
is
that
centralized
compute
prevents
collaboration
and
Innovation.
A
You
know
this
world
Garden
model
when
all
the
data
and
all
the
compute
resources
is
surrounded
by
the
walls
of
some
corporation
like
Twitter,
Facebook
and
so
on,
which
prevents
developers
from
adding
to
existing
applications
and
platforms.
In
decentralized
networks.
We
have
compatibility
of
protocols
which
is
nice,
so
all
of
these
constitutes
a
new
generation,
a
new
step
in
the
evolution
of
a
compute
we
had
mainframes,
we
had
centralized
Cloud
platforms
and
now
we're
opening
the
era
of
decentralized
serverless.
A
Compute
platforms-
that's
happening
right
now,
but
that
means
that
it
hasn't
happened
before
and
that's
because
making
the
real,
open,
decentralized
compute
hard,
especially
it's
super
hard
extremely
hard
to
do
it
without
blockchain,
because
we
want
to
be
verifiable
and
scalable
at
the
same
time
and
I
want
to
run
a
compute
on
generalized
compute
capacity.
That's
correct,
but
now,
thanks
to
the
concept
like
web
of
trust,
different
consensuses
and
GP
proofs,
we
can
build
the
model
that
will
help
with
that.
A
So
we
can
look
at
fluency,
centralized
serverless,
compute
platform
from
two
different
angles:
it's
developer
platform
and
for
the
developers
it
offers
serverless,
compute
functions
and
workflows
and
for
the
compute
providers
for
data
centers,
it's
a
Computing
marked
place,
a
decentralized
network
for
compute
providers
for
the
developers.
The
simplest
analogy,
not
super
precise
but
simple,
is
that
we
have
something
like
AWS,
Lambda
Cloud
functions
and
we
have
a
web
assembly
runtime
for
that
called
Marine.
A
We
have
an
analogy
for
AWS
step
functions,
distributed
algorithms
and
that's
aqua
and
influence,
and
the
protocol
itself
ends
on
this
step
and
all
the
other
behaviors
are
actually
distributed.
Algorithms
like
lower
balancing
routing
scaling,
orchestration
deploy
and
all
of
this
can
be
done
in
the
scripting
fashion
at
the
aqua
libraries.
A
So
that's
about
the
new
Revenue
stream
for
a
few
miners.
So
influence
is
another
way
to
bring
compute
to
the
data
we
had.
The
previous
talk
about.
Fvm
fluence
solves
similar
problem,
but
with
a
different
approach.
So
miners
earn
from
and
use
CPU
compute
capacity,
not
GPU.
We
are
using
utilizing
CPU
and
compute
unlocks
values
in
store
data.
As
a
film
miners
already
have
data,
they
can
benefit
from
it
more
after
selling.
Cpu
is
free
and
compute
providers
can
monetize
CPUs
yeah.
A
We
expect
that
compute
demand
will
grow
because
more
data
more
needs
to
compute
more
ways
to
compute
more
different
jobs.
So
monetization
will
increase
the
revenue
and
to
have
this
revenue
from
a
useful
work
made
on
the
network.
We
need
to
have
verifiability.
We
need
to
have
a
variety
of
proofs
to
distribute
the
incentives
and
we
have
three
kinds
of
proofs.
The
main
one
is
the
proof
of
processing
that
proves
are
quite
the
distributed
workflows,
so
it's
verified
by
every
beer
on
the
way
of
execution
of
the
distributed
algorithm.
A
And
after
that
we
have
probably
clarification
on
chain.
You
can
ask
which
chain
I,
hope
and
the
goal
is
to
use
fvm
we're
going
to
run
on
fem
verification.
Should
be
done
on
filecoin,
another
part
is
proof
of
execution,
that's
the
proof
of
actual
compute,
and
this
topic
is
a
bit
more
difficult,
because
if
you
want
to
have
effects,
then
it
means
that
not
everything
can
be
verified.
A
A
A
We
raised
15
millions
and
we
have
our
testnet
live
just
a
few
words
about
the
world
map.
As
we're
on
the
test
net,
currently
it
means
that
we
aim
going
live
and
we
expect
to
have
a
our
mainnet
running
on
fvm
in
Q4,
and
this
release
should
be
mainly
focused
on
the
most
fundamental
proofs
and
on
the
proof
of
capacity
for
the
computer
providers
and
next
year
we'll
be
working
on
more
proofs,
more
languages
and
better
developer
experience.
A
Yeah.
We
raised
15
million
from
top
investors,
including
PL,
which
is
super
nice.
We
have
some
Founders,
Tonto,
Bridge,
evgenie,
panorof
and
and
me
yeah,
and
we
have
a
mission.
Fundamentally,
it's
kinda
slightly
insane
project
because
we
really
want
to
make
this
Choice
compute,
really
decentralized,
like
no
compromises
and
that's
kind
of
very
Visionary,
and
we
really
want
to
empower
the
next
wave
of
Internet
Innovation.
Remove
this
blocks.