►
From YouTube: The Baseline Show: w/ RISC Zero
Description
RISC Zero is a RISC-V virtual trusted/trustless computer based on zero-knowledge cryptography. The RISC Zero zkVM is an open-source, zero-knowledge virtual machine that lets you build trustless, verifiable software in your favorite languages.
Baseline Protocol is an emerging standard for state synchronization across different systems of record. This standard can be applied to traditional corporate systems of record and utilize any kind of database or state machine, including different blockchains or distributed ledger technologies (DLTs).
Join us every Wednesday at noon EST on our live show, 'The Baseline Show'. Each week, we host industry leaders who share their insights and expert opinions on data governance, security, digital supply chain, multi-party workflows and collaboration.
Tune in via YouTube or LinkedIn Live.
#BaselineProtocol #DataManagement #Blockchain #DataSharing #MultiPartyWorkflows #DataGovernance #InformationSecurity #BusinessAutomation #BusinessSynchronization #DataIntegrity #IndustryTrends #LiveShow #LiveStream #LiveDiscussion #ExpertOpinion #YouTube #LinkedInLive #B2BBlockchain #DigitalSupplyChain #Innovation #SupplyChain
A
All
right,
I
see
us
live
hello,
everybody
Welcome
to
our
Baseline
show
it
is
Wednesday
April
5th
and
we
have
risk
zero
joining
us
today.
I
first
started
learning
about
risk
zero.
Actually,
a
few
months
ago,
when
our
r
d
unit
was
talking
about
a
very
cool
risk,
five
approach
that
the
risk
zero
team
was
taking
to
ZK
virtual
machines
and
then
at
each
Denver.
A
They
were
at
ZK
day
and
we
got
to
start
learning
more
about
the
team
and
then
have
some
follow-up
calls
to
learn
more
since
and
send
resources
about
the
Baseline
protocol
and
see
how
we
can
connect
with
the
work
that
they're
doing
so.
We
brought
them
here
today
and
we
can
learn
more
about
what
they're
working
on
and
how
it
can
be
used
in
our
Baseline
protocol
implementations
in
the
future.
So
with
that
Brian
and
Chung,
would
you
like
to
start
with
some
introductions
on
yourselves.
B
A
C
A
lot
of
the
core
team
actually
has
a
lot
of
DNA
in
infrastructure,
software
sort
of
Bridging,
the
Gap
between
research
and
practice,
sort
of
advanced
algorithm
implementation,
and
even
as
far
back
as
around
2015
Jeremy,
one
of
the
co-founders
was
real
interested
in
ZK.
He
was,
you
know,
doing
his
own
sort
of
personal
time,
research
and
talking
about
it
Non-Stop
and
as
the
field
got
closer
to
an
area
where
practical
implementations
could
have
industrial
value,
we
could
actually
build
a
ZK
VM
that
could
be
used
by
normal
people.
C
That
was
where
the
conversation
got
more
serious
and
I.
Think
that's
where
Brian
and
Frank
kind
of
got
got
looped
in
and
I
was
sort
of
present
working
on
other
things
at
the
time,
but
that
was
the
Genesis
of
risk.
Zeros,
really
Jeremy's
deep
interest
in
ZK,
going
way
way
back
and
so
Brian.
How
many
people
do
we
have
now
it's
hard
to
keep
track
about.
C
C
The
DNA
of
the
company's
really
an
infrastructure
software
which
you
see
in
our
approach
to
ZK,
really
like
a
lot
of
experience
in
building
building
software
from
xero
and
starting
companies
and
building
a
product.
You
know
clean
sheet
designs
and
a
lot
real,
deep
familiarity
with
a
lot
of
the
issues
that
go
into.
What
does
it
take
to
really
build
applications?
A
C
Okay,
so
so
what
is
as
EK
VM?
What
is
the
risk
of
zero
V
ZK
VM,
and
what
does
it
mean
to
use
risk?
Five
I
think
those
are
sort
of
three
intertwined
questions
about
what
we're
doing
that.
A
lot
of
people
have
so
I
think
everybody
watching
this
probably
is
surface
level
familiar
with
what
it
means
to
do,
ZK
that
you
can
do
an
operation
consume.
C
Some
inputs
run
some
algorithm
and
have
an
output,
that's
verifiable,
without
ever
seeing
the
inputs
really
useful
for
for
doing
rollups
really
useful
for
certain
kinds
of
security
applications
and
what
we
did
when
we
started
looking
at.
How
do
you
implement
a
ZK
VM?
That's
really
developer
friendly.
C
That
runs
risk
five
machine
code
that
you
can
already
produce
using
compilers
for
existing
languages
like
C
or
rust,
but
we
we
built
it
in
a
VM
that
actually
emits
proofs,
and
so
we
said
your
workflow,
you
can
write
what
you
want
to
write
in
Rust.
You
can
take
existing
code
in
rust
or
where
C
or
C,
plus
plus
compile
that
down.
You
can
run
it
inside
our
ZK
VM
and
you
don't
have
to
change
anything.
C
You
do
about
how
you
write
that
code,
and
so
that
stands
a
little
bit,
in
contrast
to
some
of
the
other
earlier
approaches
where
people
have
to
build
custom
circuits
and
there's
a
lot
of
work
by
the
application
developer.
That
requires
basically
Computer
Engineering.
You
know
processor
designer
circuit
design,
along
with
application
engineering.
We
say
you
know,
leave
the
circuit
design
to
us.
We
will
provide
a
compiler
Target
that
uses
the
tool
chains
that
you've
already
got,
and
so
that's
that's
kind
of
the
thesis
behind
adopting
risk.
Five.
It's
risk
the
risk.
C
C
And
then
one
one
thing,
that's
interesting
about
when
we
talk
about
VMS
that
I've
I
was
confused
about
initially
coming
from
some
familiarity
with
ethereum
and
the
evm,
and
what
does
it
mean
to
build
a
ZK
VM
and
how
is
that?
Does
it
replace
the
evm
or
is
it?
Is
it
compatible
with
it
or
what
does
that
mean?
One
of
the
key
things
that
makes
it
complementary
is
that
you
have
two
different
needs
in
ZK.
C
You
have
a
need
to
operate
on
sensitive
data
and
then
erase
that
data
and
share
only
the
result
and
in
something
like
a
a
shared
blockchain
like
the
evm.
You
have
a
single
virtual
machine
where
the
state
is
visible
to
everybody.
That's
actually
a
very
important
part
of
that
that
design-
and
so
one
thing
that's
really
different
about
the
zkvm
from
the
evm
aside
from
the
kind
of
code
that
it
consumes.
C
Risk
5
machine
code
versus
evm
byte
code
is
that
the
ZK
VM
actually
runs
once
per
there's
one
instance
per
execution,
one
one
instance
per
transaction.
We
run
it
and
then
we
throw
away
all
this.
The
confidential
State
and
leave
only
the
the
output
that
you
want
to
share.
So
that's
kind
of
an
important
architectural
difference
between
what
we've
done,
with
the
ZK
VM
and
and
what's
in
evm,
and
you
can
integrate
those
two
things
right.
D
B
Is
there
yes,
the
responsibility
of
whatever
the
hose
chain
is
to
pass
State
into
the
VM
like
right.
Now
there
isn't
a
particular.
You
know
it's
not
it's,
not
a
VM,
that's
driving
an
actual
blockchain.
You
certainly
could
build
something
like
that
using
it,
but
it's
really
just
a
ZK
virtual
machine
that
takes
code
yeah.
D
E
C
Both
it
fits
into
some
things
that
are
on
our
roadmap,
which
Brian
might
talk
about,
but
also
would
be
probably
the
level
at
which
you'd
integrate
into
something
like
a
baseline
system,
and
that
there
is
a
split
between
the
guest
code,
which
is
stateless
and
the
host
code,
which
could
be
anything,
could
be
part
of
a
blockchain.
It
could
be
part
of
a
game
that
you're
running
on.
B
B
B
B
Depending
I
don't
know
if
you
have
slides
in
here
to
talk
about
you
know
so,
because
this
ZK
VM
can
run
code
in
any
language,
it
can
actually
run
the
evm
on
top
of
it,
and
we've
actually
started
to
see
some
really
interesting
use
cases
with
people
doing
this.
So
there's
the
sort
of
ZK
proof
of
exploit
project
that
that
won.
C
Perfect
example
of
an
application-
I
don't
have
slides
for
that,
but
that's
a
perfect
example
of
an
application
where
it's
really
important
not
to
share
the
the
internal
state
of
ZK
VM
before
you're.
Ready,
like
the
whole
point,
is
to
to
be
able
to
prove
that
you
have
a
runnable
program
that
has
different
properties
without
disclosing
it.
B
Yeah
and
people
have
also
used
ref
running
on
top
of
the
ZK
VM
to
produce
like
proofs
proof
of
assets
and
so
forth.
So
there's
a
lot
of
stuff
you
can
do
by
leveraging.
The
general
purpose
framework
to
you
know,
use
existing
code
bases
to
produce
proofs
of
all
kinds
of
interesting
facts.
Depending
on
what
you're
doing
we
actually
have
some
people
using
using
Baseline.
C
Yeah
and
then,
as
far
as
our
current
implementation,
so
it's
open
source
which
is
important
and
we
can
walk
through
just
like
the
first
couple
steps
of
the
repo,
but
we
have
focused
on
for
the
guest
code,
the
code
that
actually
runs
the
inside
of
the
ZK
VM
on
supporting
rust
first
and
that's
sort
of
a
priority
Choice
based
on
the
idea
that
rust
has
a
couple
interesting
properties,
one
it
lets.
You
write
very
compact,
efficient
programs
that
are
highly
reliable.
C
It's
a
Memory,
safe
language,
so
you
can
do
the
kinds
of
things
that
you
might
might
be
just
sort
of
on
the
edge
of
what
you
might
do
in
a
smart
contract
with
solidity.
Those
kinds
of
patterns
work,
but
it's
also
a
robust
language
that
gives
you
a
lot
of
Headroom
to
grow
and
build
more
complex
applications,
and
so
we
sort
of
picked
rust
as
our
first
stop
for
language
support,
although
internally
we've
prototyped
at
least
CC
plus
and
I-
think
go,
but
really
anything
runs
at
risk.
C
Five
conceptually
should
work
we're
supporting
rust
first,
so
that's
kind
of
the
near
term
is
we
want
to
make
rust
a
great
experience,
but
there's
a
lot.
There's
a
road
map
towards
support
of
pretty
much
any
mainstream
language
and
doing
that
not
by
re-implementing
the
language,
but
by
porting
the
existing
compiler.
B
C
C
This
is
a
one
slide.
Example
of
rust,
guess:
code,
it's
real
straightforward!
This
is
just
proving
that
you
have
two
numbers
that
multiply
together
produce
some
product
without
disclosing
the
the
factors-
and
this
is
really
we'll
look
at
this
in
the
the
well
we'll
look
at
this
in
GitHub,
so
you
can
see
all
the
scaffolding
around
it,
but
there's
really
not
very
much.
It's
pretty
concise.
C
C
C
B
Yeah,
while
you're,
while
you're
pulling
up
the
sort
of
code
I
can
so
we're
going
to
look
at
after
Jingles
it
up
is
an
example
of
using
the
risk,
zero
VM
to
basically
parse
a
Json
file
that
contains
anything.
You
can
then
basically
produce
proofs
that
Json
file
or
two
Json
files,
or
any
number
of
like
csvs
any
kind
of
input
you
might
expect
from
the
kind
of
like
Enterprise
context.
You
can
basically
extract
information
from
that
transform.
B
It
run
Logic
on
it
and
produce
a
proof
that
you
know
whatever
kind
of
interaction.
You're
trying
to
coordinate
was
done
according
to
the
specifications
that
that
companies
expect
and
they
should
be
able
to
reuse
their
existing
kind
of
logic.
Generally
speaking,
without
needing
to
build
new
circuits,
and
so
here's
the
context
of
Baseline.
C
And
so
make
this
a
little
bigger
one
thing:
that's
really
interesting
about
this
code.
Is
that
there's
no
implementation
of
Json
right
like
this,
is
just
using
the
existing
crate,
so
this
shows
both
It's
relatively
simple.
Overall,
it's
also
simple
to
reuse
existing
code
and
get
acceptable
efficiency
out
of
it.
B
C
B
C
C
C
C
C
So
one
question
for
a
topic
I'd
like
to
open
up
is
talking
about
how
having
something
like
ckvm
having
that
capability
fits
into
Baseline,
architectures
and
and
where
both
where
technically
and
then
we're
in
the
design
process
for
BPI.
That
would
come
up
and
in
my
reading,
what
of
of
the
Baseline
documentation?
I
saw
three
spots
where
it
seemed
like
it
made
sense
to
integrate.
C
D
Yeah,
so
you
you
could
you
could
I
agree?
Those
are
those
are
those
sort
of
potential
places
where
you
can
use
it.
If
you
have
trust
issues
with
an
operator
right,
I
mean
that's
a
that's
a
that's
like
that's
a
key
question.
That's
the
that's!
The
key
thing
right
and
in
in
in
this
case,
are
you
you
know
it's
like.
Is
everything
that
you're
taking
in
right?
D
So
if
you
look
at
the
first
one,
the
abstraction
layer
right,
if
you,
if
you,
if
you
think
of
a
of
a
oven,
you
know
a
traditional
API
Gateway
right,
can
you
prove
that
all
require
all
requests
incoming
that
pass?
Let's
say
the
you
know
an
excess
control
layer
are
passed
into
the
middleware
logic
right,
so
is
there
so
in
otherwise.
In
other
words,
it
there's
no
censoring
right.
D
So
you
have
a
you,
have
a
you,
have
a
proof
of
of
of
censorship,
resistance,
minus
any
access,
control,
rules
right,
and
you
can
even
then
prove
that
the
access
control
rules
were
were
appropriately
applied.
Right,
I
mean
you.
Can
you
can
basically
every
single
code,
execution
could
be,
could
be
verified
and
you
could
you
could
pass
that
along
and
and
basically
verify
the
entire
stack.
D
D
The
effort
versus
the
output
right
I
mean
it's
like,
or
you
know
that
that
you
need
to
do
you
know
if
we
are
talking
about
so
I
could
see
that
you
know
it's
like
if
you
think
of
an
application
of
risk
zero,
where
you
are
settling
in
the
FED
overnight
window,
the
outstanding
you
know,
somewhere
between
five
and
eight
trillion
dollars
worth
of
of
of
settlements,
and
that
that
is
done
correctly
right
now.
Everybody
trusts
the
FED
that
the
the
the
the
that
their
mainframes
operate
correctly.
D
Right
and
basically
everybody
just
sucks,
yeah
I
expect
this,
and
this
is
out
so
everybody
validates
independently
that
they're
not
being
cheated
and
there's
like
you
know,
or
at
least
they
pretend
to.
But
there
is
no,
but
there
is
no
collusion
resistant
way
to
prove
to
the
outside
world
that,
even
though
you
don't
see
anything,
you
know,
there's
no
collusion
going
on
right.
So
similarly,
you
can
then
think
of
this.
D
As
a
you
know,
if
you're,
if
you're,
if
you
put
that
as
if
you're
settling
invoices
right
orders
versus
invoice
map,
you
know
it
was
matching
proper
application
of
taxes
all
of
this
stuff
in
in
in
a
in
a
it.
Not
only
that
that
the
two
sides
right,
which
is
the
key
thing
about
Baseline,
do
the
do
the
is
the
synchronization
of
of
two
systems
of
Records
right,
and
the
idea
is
that
the
proof
that
is
then
verified
on
chain
is
is
is
is
is
proves
that
right,
but
you're
you're
there
are.
D
C
C
Enter
the
purchaser
and
say:
oh
you
charge
me
this
amount
of
tax,
you
say:
there's
a
government
tax,
but
it
was
it's
different
from
last
invoice
and
you
could
automatically
validate
that
all
the
way
through.
So
you
know
that
I
as
a
purchaser
would
know
that
I'm
paying
the
correct
amount
you're
withholding
the
right
amount
of
taxes.
D
Yes,
yes,
I
mean
that's
yeah,
it's
it's
it's
it's
a
proof
of
correct
execution
of
of
the
tax
line
and
and
everything
right.
So
it's
it's
it's
based
on
based
on
on
you
know,
the
current
state
right
do
are,
are
all
the
discounts
or
the
penalties
applied
properly
right?
Are
there
any
any
outstanding
credits
from
a
previous
invoice
right
that
you
know
you
typically
haggle
about
right?
But
if
you
don't,
if
you,
if
you,
if
you
do
this
consistently,
then
there
is
a
you
know.
D
C
It's
interesting
to
think
about
the
in
the
long
term.
You
could
end
up
in
a
situation
where,
as
a
company,
your
books,
you'd
actually
know
you'd,
have
integrity
guarantees
on
the
transactions
coming
in
as
well
as
internally.
So
you'd
have
a
lot
of
confidence
that
that
you
weren't
going
to
have
to
reach
back
and
restate.
D
B
So
if,
for
some
reason
you
know,
however
you're
settling
this
or
the
wherever
you
need
to
actually
record
the
the
proof
and
verify
it,
you
know,
oftentimes
snarks
are
a
lot
friendlier
than
Starks,
especially
in
the
ethereum
kind
of
context,
so
yeah,
so
the
circuit's
very
recruiting
friendly.
At
this
point.
D
B
B
Make
kind
of
AD
State,
you
can
actually
add
State
bonus
using
this
mechanism
as
well,
if
you
think
about
more,
like
Amazon,
oh
again,
what
they
call
it.
Their
stateful
function,
workflow
graph
stuff,
which
is
really
interesting,
I
think
using
ZK
to
kind
of
run
up
to
a
certain
point,
pause
your
sort
of
business
logic
and
then
react
as
as
new
information
comes
in.
D
C
D
D
One
is
the
size
of
the
proof
carrying
data
as
well,
as
you
know
with,
and
as
you
carry
forward
right,
that
the
size
of
the
proof
carrying
data,
as
you
start
to
add,
recursion
levels
increases
right
because
you
need
to
right,
you
need
to
you,
need
to
increase
you're,
adding,
basically,
all
your
all
your
all
your
conditions,
right
of
the
previous
Pros.
Otherwise
you
can
prove
it
in
the
same
circuit
right
so
improve.
B
It's
yeah,
although
it
actually
reaches
as
a
stable
Point
and
pretty
much
maxes
out
at
about
like
two
I
think
or
like
sort
of
400
kilobytes.
You
could,
in
this
proof
system
at
least
and
at
least
in
ZK
Summit.
You
know
this
year,
which
is
why
I'm
like
Lisbon
right
now,
there's
a
lot
of
research
coming
out.
You
know
about
exact
how
to
minimize
the
amount
of
proof
carrying
data
and
the
amount
of
proving
time
in
between
all
these
different
steps.
D
B
D
Not
that's
not
allowed,
so
you,
so
you
you,
you
need
to
right,
so
so
you're
you're
you're.
You
need
to
have
more
Cycles
because
you
have
to
have
more
Cycles
in
your
in
your.
The
group
has
to
be
larger,
but
it
needs
to
map
onto
the
same.
The
the
problem
is
that,
if
you're,
if
you're,
if
you're
the
problem,
is
the
group
size
right?
It's
it's
like
it's
like.
If
you're.
If.
D
D
B
D
B
Yeah,
so
it's
probably
I
mean
we
use
a
slightly
different
field
than
everyone
else.
Actually
we
use
something
called
bait,
we
call
it
baby
bear
and
it
actually
fits
into
32
bits,
and
then
we
use
the
4X
expansion
field
to
achieve
the
100
bits
of
security
that
we
typically
Target.
However,
these
things
are
like
fairly
parametric,
so
I
think
that's
actually
an
interesting
point
for
lots
of
Enterprise
use
cases
is
that
that
the
desired
and
needed
level
of
security
is
could
be
quite
different,
depending
on
the
use
case.
B
B
So,
but
currently
you
know
the
the
general
mode
our
proof
system
runs
in.
We
actually
have
100
bits
of
security,
and
typically
we
also
support
Shot,
2
or
Poseidon,
but
generally
you
can
sort
of
pick
which
hash
function
you
want
to
use
as
well.
Although
you
do
have
to
implement
those
in
a
circuit
yeah
a
lot
of
the
sort
of
problems
of
of
getting
a
ton
of
roofs
for
general
purpose,
Computing
down
into
one
proof,
we're
actually
mostly
solved
at
this
point
for
general
purpose.
Computing,
it's
it's.
B
I'm
still
not
sure
so
you
know
I'm
not
I'm.
Definitely
not
so
Jeremy's,
not
here,
right
now
he's
in
the
air
but
he's
RNG
scientist
and
the
inventor
of
the
proof
system.
Obviously
you
know
our
work.
We
stand
on
the
shoulder
of
giants,
many
many
but
yeah.
It
was
very
fascinating
to
see
all
of
the
the
different
approaches
to
to
proving
system
optimization
that
people
are
taking
nowadays
at
CK,
CK
Summit.
Nine
really
blew
my
mind
this
year
and
yeah,
but.
C
Although
there's
some
some
discussion
about,
you
know
how
do
we
make?
How
do
we
do
extensibility
in
a
way
that
maintains
compatibility
like?
Is
there
a
way
that
we
can
make
it?
You
can
bolt
on
an
accelerator
for
a
particular
operation
that
you
need,
if
you
have
some
very,
if
you're
like
not
just
a
normal
application
developer,
but
you
have
some
like
exotic
algorithm.
There's
one
step
that
really
gets
a
lot
more
efficient
is
a
way
to
bolt
that
on
and
then
have
that
not
break
compatibility
with
all
other
instances.
Yeah.
B
That's
basically
also
a
solid
at
this
point
at
this
point,
but
but
anyway
we
don't
need
to
get
super
far
into
the
the
nerdy
depths,
the
zkavm,
unless
you
really
want
to.
But
yes,
we
do
have
like
ways
to
build
acceleration
circuits
for
for
like
ecdsa,
so
we
actually
just
built
the
big
into
accelerator,
so
you
can
do
very
large,
integer
math
in
a
way
that
doesn't
sort
of
incur
the
overhead
of
decomposing
it
into
risk.
532-Bit
integer
operations.
B
Screen
your
screen,
your
screen's
still
on
you
might
want
to
stop
sharing
yeah.
Sorry
yeah,
thank
you,
so
composable
ZK
I
think
is
critical
Trend
and
we
actually
have
some
an
open
source
user.
Actually
did
this
a
while
ago,
they've
verified
Cairo,
Stark
net
proof
inside
this
live
by
just
importing
the
prover,
and
so
you
can
actually
do
that
in
a
lot
of
instances.
You
can
you
know
import
other
proof
systems
and
run
them
inside
different
proof
systems
and
then
produce
a
different
proof.
B
So,
for
instance,
as
I
mentioned
earlier,
you
know
right
now,
Jeremy's
implementing
a
start
to
start
conversion
system.
Basically,
so
we
will
take,
you
can
recurse
all
these
general
purpose
proofs
down
into
one
single
proof.
That's
the
you
know,
400
kilobytes,
dark
proof
or
whatever,
and
then
you
can
actually
run
that
through
surcom
and
produce,
you
know,
draw
16
proof
or
you
can
use
the
flank
or
F
flunked,
I'm,
not
sure
what
people
are
calling
it
yeah.
D
B
And
so
I
think
how
we're
planning
on
making
our
offering
approachable
for
supply
chain
system
users,
wow
I
mean
the
main.
The
the
main
way
we're
planning
on
doing
that
is
by
letting
people
use
the
code
that
they
already
have
for
sort
of
settling
transactions
or,
comparing
you
know,
insurance
claims
any
of
this
stuff
that
you
should
generally
be
able
to
use,
if
not
exactly
the
same
code.
You
know
a
very
similar
algorithm
in
in
something
else.
So
users
shouldn't
generally
need
to
think
about
this.
B
C
D
B
D
C
B
A
B
So,
oh
so
there's
the
other
direction
so
well
right.
So
there
are
two
ways
that
you
can
use:
sort
of
risk:
five
in
an
ethereum
context
and
one
of
them
so
being
able
to
take
programs
that
are
run
in
Risk,
zero
and
prove
them
on
chain
and
ethereum.
That's
actually
the
bit!
You
could
do
that
right
now.
It
would
just
cost
a
lot
of
gas,
because
the
Stark
proofs
are
so
large
from
the
recursion.
So
that's
why
we're
building
this
start
to
snark
converter?
B
So
once
that's
done,
you
will
be
able
to
basically
take
one
of
these
general
purpose
programs.
You
can
pass
in
the
state
from
the
evm.
This
is
why
I
can
find
the
link
to
the
sort
of
proof
of
assets
work
that
people
have
done
where
you
basically
pass
in
your
program
in
this
context,
is
like
our
evm
and
you
pass
in
a
block
State,
and
you
know
your
key
and
prove
that
you
actually
have
up
to
you-
know
10
million
dollars
and
assets
about
revealing
where
they
are.
B
So
that's
one
way
to
do
that
and
we
will
also
be
building
this
Bonsai
system.
That
will
let
people
submit
proofs
requests
for
proofs
into
the
system
and
then
we'll
roll
them
all
up
and
sort
of
pay
the
the
costs
across
ethereum,
but
that's
useful
across
sort
of
Enterprise
applications
as
well,
not
just
public
blockchain.
So
sadly,
there
isn't
an
example
code,
there's
example:
code
that
proves
ethereum
State.
You
can
look
up
ZK
proof
of
exploit
or
this
I
can
find
this
proof
of
assets
thing
really
quickly.
B
B
D
B
D
Reducing
the
the
you
can
reduce
the
size
of
the
of
the
chain
by
by
having
by
having
the
by
doing
exactly
that,
you're
you're
running
you
know
all
the
transactions
you're
you're,
the
you
don't
need
to
keep
the
transactions
anymore,
you
just
you
just
proof,
State
and
State
Transitions,
and
then
your
your
posting
that
just
the
the
the
chain
becomes
just
just
the
just
the
recursive
proof.
So
it's
like
it.
It.
D
Time
what
Mina
has
been
doing,
but
it's
like
maintaining
state,
so
the
the
point
is
that
they,
you
still
have
you're
still
you're,
still
stateful
right,
so
you're
you're
you're
in
that
was
the
original
question
is
who
maintained
state
right,
so
the
validators
still
maintained
State.
It's
just
that
all
the
proofs
are
it's
all
compressed
down
into
into
into
proofs
and
no
one
needs
to
maintain
the
state
except
the
validators
right.
The
elevators
maintain
exactly
right,
because
if
you're
going
stateless,
then
you
really
have.
D
You
still
need
consensus,
but
the
problem
is
that
the
users
don't
know
what
the
consensus
is
of
the
state
that
they
need
to
base
their
yeah
their
their
proof
on
right.
They
generate
the
proof
right,
they're,
they're,
generating
the
state
transition
and
they're
just
sending
in
the
proof
right
and
it's
in.
And
but
the
point
is
this
like
it's
like.
If,
if
you're
impacting
the
same,
the
same
account.
B
A
B
A
John
walpert
one
of
the
OG
Baseline
leaders
who's
a
product
exec,
has
the
comment
that
there's
very
little
on
the
site,
referring
to
blockchain
on
risk
zero
site.
The
current
discussions
seem
to
indicate
in
ethereum
Focus
potentially,
but
he
is
interested
in
learning
about
the
choice
of
not
putting
blockchain
front
and
center
on
all
of
the
collateral
as
well
as
interested
in
learning.
If
this
can
be
used
without
a
blockchain.
B
Yes,
so
so
part
of
the
current
sad
state
of
the
website
has
just
been
it's
taken
a
while
for
us
to
get
our
sort
of
branding
in
a
state
that
that
we
really
are
going
for.
That's,
hopefully
going
to
be
resolved
pretty
soon,
so
you
know
we're
big
Believers
in
blockchain
and
blockchain
Technology,
but
you
know
our
background.
I
would
say
we're
fairly
new
to
the
space.
At
least
the
co-founders
are
so
you
know
we
do.
Think
of
this.
B
Technology
is
much
broader
than
than
blockchain
itself,
but
we
are
very
much
focused
on
on
the
blockchain
side
of
the
world.
However,
you
know
we
also
this
Bonsai
Network
that
we're
building
you
know.
Bonsai
is
intended
to
be
more
of
a
you
know,
sort
of
public,
good
and
network,
rather
than
a
company
efforts
that
we
see
a
receiver
as
building
that,
and
so
Bonsai
will
really
be
where
we
focus
us
most
of
our
sort
of
blockchain
story
yeah.
B
But
we
have
been
talking
to
various
other
teams
about
things
outside
of
the
blockchain
context
too.
I
don't
know.
If
you
want
to
talk
about
that,
but
yeah
it's
certainly.
It's
certainly
come
up
in
conversations
that
even
the
limited
amount
of
blockchain
that
we
mentioned
on
the
receiver
site
does
occasionally
cause
concern
in
certain
quarters,
so
yeah.
C
Yeah
I
think
there
are.
There
are
a
lot
of
applications
outside
of
blockchain
or
like
Baseline,
which
sort
of
sits,
certainly
outside
of
defy
and
like
Baseline,
sits
in
between
Enterprise
and
traditional
blockchain,
and
we
don't
want
to
neglect
that
side
of
the
company
or
that
side
of
the
the
potential
applications
for
ZK
and
ZK
VM.
Specifically
because
of
how
we've
designed
it
could
fit
relatively
well
into
existing
software,
where
it
can
be
a
security
ad.
D
The
the
key
here
is
really
that
you
want
to
be
able
to
have
these
these
synchronizers
either
between
two
or
more
parties
and
that
they
they
inherit.
They
basically
take
advantage
of
the
security
properties
of
a
of
a
of
a
public
chain.
That's
like
that's
the
that's
the
primary
thing,
so
anybody
can
validate
anything
at
any
time
without
you
know
it.
D
If
you're
in
the
US
or
whatever
other
public
record
filings
companies
publicly
traded
companies
must
do
it
also
would
help
you
know
to
to
to
create
more
trust
in
in
other
entities
that
do
not
have
public
record
requirements
filings,
but
they
could
do
that
now
in
a
way
that
doesn't
reveal
anything
but
still
can
say,
hey
we're
we're
not
like
cheating
anywhere
right,
which
is
also
again
kind
of
important.
If
you're,
if
you're
like
tax
Authority
again,
they
can
or-
or
you
know,
Regulators,
they
can
now
say-
oh
yeah
you're.
D
B
C
B
Like
proving
me
approving
that
various
insurance
claims
or
whatever
potentially
are
following
whatever
rules,
they're
supposed
to
things
like
that
very
interesting,
yeah
I
will
say,
I,
don't
know
some
of
the
other
questions
or
can
I
have
this
yeah
when
I
sit
around
and
think
about
the
utility
of
this
technology,
I
I
started.
You
know
occasionally
really
getting
into
this
idea
of.
You
know:
solving
supply
chain,
stuff
and
mass,
and
really
enabling
people
to
to
communicate
orders
more
effectively
for
various.
B
C
D
E
This
is
a
way
to
use
your
your
existing
systems,
but
just
prove
consistency
just
prove
that
everybody
is
working
off
the
same
data
set
and
it
didn't
involve
cloning
and
moving
it
to
some
Central
repository
like
a
blockchain.
You
know,
which
is
also
a
massive
threat
surface
for
any
hacker
that
wants
to.
You,
know,
come
in
there
and
pick
some
locks,
and
you
know
really
where
it
went
from
there.
E
I
can't
share
it
to
you,
but
I
need
to
prove
to
you
that
I
know
the
secret
and
you
need
to
be
satisfied
that
I
know
the
secret
and
getting
those
elements
there
is
requiring
a
lot
of
gymnastics.
At
least
conversationally
yeah
definitely
yeah
the
the
the
method
that
I've
been
using
recently
just
trying
to
demonstrate
a
zero
knowledge
exchange
is
say.
You
know,
Brian
I
want
to
prove
to
you
that
I
know
your
phone
number,
but
I
can't
tell
you
because
somebody
might
be
listening.
E
Somebody
might
be
you
know,
looking
in
there,
so
I
can't
say
the
10
digits.
So
what
am
I
going
to
do?
I'm
going
to
pick
up
my
phone
and
I'm
going
to
call
you
and
if
your
phone
rings
I
have
proven
to
you
that
I
know
your
phone
number
and
I
had
given
you
no
additional
knowledge
other
than
I
have
proven
it
so
totally.
That
is
that's
a
good.
B
Example,
oh
yeah
great
example,
I
go
ahead.
I
was
I,
did
want
to
like
talk
about
that
like
an
emerging
thing
that
might
be
of
interest
to
the
audience,
which
is
the
viability
of
zkml.
We're
not
super
focused
on
that,
but
we've
actually
seen
like
some
weight.
Some
of
the
latest
research
people
being
able
to
prove
things
about
data
without
revealing
the
data
by
using
a
machine
learning
model,
inference
model
inside
ZK.
So
I'm
personally
optimistic
that
you
know
over
the
next
read
a
10-year
time
frame.
E
Right
I
mean
a
lot
of
people.
Sit
there
and
say
you
know:
I
I
I
need
to
be
convinced.
You
have
to
show
it
to
me
before
I
believe
it
and
I'm
like
well.
I
can't
show
it
to
you,
but
what
I
can
show?
You
is
a
mathematical
proof
saying
that
proving
and
I'm
not
just
a
testing
you're,
not
taking
my
word
for
it.
I
am
proven
that
I,
you
know
I'm
working
off
the
same
secret
that
you
are.
B
Right,
like
the
crazy,
the
crazy
example
I
came
up
with,
in
my
mind,
to
say
you
had
like
a
very
complex
part,
and
you
wanted
to
place
a
large
order
for
it
that
was
gonna
like
take
up
a
lot
of
some
factories.
You
know
very
expensive
CNC
machines
for
a
long
period
of
time,
and
they
don't
want
to
necessarily
commit
that
capacity
to
you
without
proof
that
you
actually
need
it.
So
you
could
actually
provide
at
some
point.
You
know
proof
of
similarity
of
a
part
without
necessarily
like.
D
Actually,
turning
it
around
is
is
is,
is
really
critical
by
saying.
I
need
this
at
this
point
in
time
or
do
you
is
it
actually
available
right?
Because,
because
that
is
that
is
that
is
that
is
much
more.
You
know
it's
like
it's
like.
Can
you
produce
this
deal
right
in
this
amount
at
this
quality
at
this
point
in
time?
You
know
by
this
point
in
time
and
deliver
it
right.
D
So
this
is
obviously
hard
in
the
physical
world,
but
in
the
digital
world,
it's
much
easier
and
it's
actually
very
very
relevant
for
for
actually
what
we're
doing
right
now
right,
because
what
we're
doing
is
we're
utilizing
what
probably
15
different
carriers
to
deliver
data
on
the
map
right.
It's
like
my
feed,
runs
through
like
a
t
and-
and
that
is
that
is,
but
you
know
Brian
you're
in
Portugal,
which
means
the
data
that
you're
seeing
right
now
is
being
it
it.
D
B
D
Right,
so
it's
that
that's
and
that's
more,
that's
more!
That's
more!
That's
more
doable
where
you're
then
committing
to
certain
things
to
be
available
at
a
certain
period
of
time
and
paying
penalties
if
they're.
If,
if
that
inventory
is
not
available
right
because
you
you
have
a
you
committed
to
it,
you
got
to
commit
that
prove
this
is
available.
And
then,
if
it's
not
available
right
you
can
you
can
prove
that
it
was
not
available
at
that
time
without
revealing
anything.
That
was
that's.
D
E
C
D
Connectivity
supply
chain
basically
is
like
data
on
Demand
right.
Is
it
there
when
it's
on
demand,
are
the
circuits
working?
If
a
circuit
is
not
working,
who's
at
fault
right,
because
then
you're,
then
you
have
you
know
you
meet
your
Atlas,
you
don't
or
their
credits,
then
on
the
on
the
invoice,
because
you're
you're
billing
bulk
right.
So
it's
it's
that
that
goes
through
the
entire
life
cycle,
from
coding
to
ordering
to
delivery
to
invoicing.
C
B
B
There's
not
a
lot
of
audit
capacity
for
a
circuit,
so
you
know
zero
can
by
using
that
and
using
pre-ordered
code
can
simplify
sort
of
Code
Compliance
mechanism,
like
complexities,
pretty
much
all
levels
and
then
additionally,
I
think
the
you
know.
The
part
that
I
find
really
interesting
is
the
potential
to
write
the
interesting
kinds
of
ZK
application
logic
and
that
counter
parties
prove
things
to
each
other
about
you
know
about
whatever
was
done
without
revealing
exactly
you
know,
all
of
the
data
behind
it.
D
We
have
an
interrupt
demonstration
coming
up
so
interrupt
between
two
different
Baseline
implementations.
Baseline
protocol
implementations
coming
up
around
the
the
there's,
a
working
group
around
that
we
have.
We
have
a
use
case,
which
is
synchronizing
banking
data,
which
is
a
huge
problem
globally
right.
So
apparently
our
approach
is
that
we're
you
know
one
side
generates
a
proof
and
then
sends
the
proof
over
to
the
other
side
without
the
public
input.
D
The
public
input
is
just
a
root
right,
and
can
you
generate
if
you
you
you're,
proving
synchronicity
by
being
able
to
generate
the
route
and
verify
the
proof
cool?
That's
that's
a
very
simplest.
It's
a
simple
thing,
however.
I
think
there
might
be
some
interesting
Synergy
there.
First
zero
wanted
to
to
like
participate
and
and
and
maybe
make
that
a
bit
more
extensible
on
on
at.
D
So
we
we
have
a
regular
meeting
Fridays
at
7am
Pacific
and
if
one
of
you
guys
wanted
to
like
sniff
in
and
see
whether
they
want
to
participate,
there's
also
an
RFP
that
is
out
that
you
know
organizations
can
participate.
We're
raising
funds
for
that,
so
you
know
you
can
fund
means
either
dollars
or
in-kind
contributions
that
work
that
that
work
too
so
happy
to
to
to
take
that
discussion
to
the
next
level
and
and
if
you're
interested,
we
would
be
happy
to
to
incorporate.