►
From YouTube: Governance and Risk | Ep. 167
Description
Agenda:
Introduction:
@GovAlpha-Core-Unit : Hosting, Introduction, Agenda.
@gov-comms-core-unit : Slides
Governance Round-Up:
@LongForWisdom Poll + Executive Status
@blimpa - MIPs Update
@Artem_Gordon Forum at a Glance
Selected Updates / Discussions:
SAS-001- Sidestream Auction Services Demo
@Oracles-Core-Unit - Oracles; Costs, solutions, next steps.
Other Discussions and General Q&A:
(Time permitting, after general Q&A) - Ilk Hole size, Keeper health, liquidation ecosystem discussion.
Full written version:
https://forum.makerdao.com/t/agenda-discussion-scientific-governance-and-risk-167-thursday-november-11-17-00-utc/11588
A
All
right,
hello,
everyone
and
welcome
this
is
the
167th
scientific
governance
and
risk
meeting
here
at
maker
dao.
My
name
is
peyton.
I
go
by
pros
11
online
and
I
am
one
of
the
governance
facilitators
here.
A
A
We
have
a
awesome
group
of
people
here,
both
directly
involved
with
maker
and
just
interested
in
in
general,
with
what
we're
up
to-
and
this
is
our
weekly
call,
where
we
discuss
kind
of
the
key
things
that
are
going
on
within
the
dao
hold
discussion
topics,
as
well
as
update
the
community
on
recent
governance
actions.
A
So
a
few
notes,
I
guess
we
do
encourage
questions.
We
encourage
engagement
of
all
types,
so,
if
you're
comfortable
hopping
on
the
mic,
if
there's
a
good
moment
to
hop
in
feel
free
to
do
so,
if
you're,
not
comfortable
or
otherwise
unable
feel
free
to
drop
something
in
the
chat,
and
I
will
be
more
than
happy
to
bring
it
up
inappropriate.
A
As
a
reminder,
this
is
being
recorded
and
saved
to
youtube
later.
So
do
try
not
to
talk
over
one
another
and
be
mindful
your
comments
can
be
viewed
later
excellent.
I
think
that
more
or
less
covers
things
so
we'll
get
to
the
agenda
here,
we're
starting
in
with
the
governance,
roundup,
and
that
will
be
done
with
me.
A
Then
we'll
move
into
some
community
discussion
and
if
we
still
have
time
we'll
kind
of
open
the
floor
to
more
broad
questions
that
we
can
engage
with
with
everyone
who's
here
awesome.
A
So
governance,
roundup
start
with
the
votes
of
the
day
generally
with
the
polls
was
a
pretty
busy
week
for
for
polls.
If
you're
a
maker
voter,
you
notice
a
lot
of
different
options
to
give
your
opinion
on
as
far
as
the
ones
that
ended
today,
we
had
five
pass
and
one
that
was
rejected.
I've
got
them
listed
on
the
screen
here,
but
I
will
run
through
them
quickly.
A
One
pass
was
adding
the
wbtcb
volt.
This
again
would
be
a
lower
collateralization
ratio
vault
for
wbtc
on
people
to
seek
more
leverage.
Should
they
choose
to
do
so.
We
also
have
the
g-uni
parameter
adjustments
as
well
as
ones
for
wrapped
stick:
beef
and
the
d3m
are
of
a
direct
die
module.
I
might
be
getting
these
backwards
there.
I'm
sure
someone
can
fix
that
later.
The
other
bowl
that
got
passed
was
the
increase
to
the
recognized
delegate
compensation.
A
So
again,
this
is
our
first
month
of
doing
the
recognized
delegate
compensation
trial,
so
that
increase
will
be
reflected
when
we
do
the
payments
later
this,
I
guess,
at
the
end
of
the
month,
sorry,
the
poll
that
was
rejected
was
our
ns
covenant
modification
that
was
proposed,
and
you
can
read
up
about
that
in
the
forum
two
green
light
poles
still
up
that
do
end
in
four
days,
md
here,
irradiance,
so
llc.
Hopefully
I
didn't
butcher
that
too
badly
and
olympus.
A
Now
I
won't
steal
pablo
thunder
who
will
come
talk
to
us
a
little
bit
about
the
mips
update,
but
there
are
eight
radical
ratification
polls
that
are
also
ongoing
for
vote
and
those
close
on
the
22nd.
A
So
give
you
one
brief
more
update
on
the
executive
proposals
before
I
pass
it
off.
The
last
executive
was
passed
and
executed.
This
included
parameter
changes
from
the
maker
open
market
committee,
as
well
as
setting
the
two
psms
10
parameter,
which
is
the
fee
in
to
zero
percent,
and
then
we
also
had
a
core
unit.
Budget
transfer
for
the
development
in
ux
coordinate
the
executive
proposal
that
will
go
for
tomorrow.
You
can
expect
that
to
include
returning
die
to
from
the
paws
proxy.
A
That
is
the
die
if
you're
watching
the
forums
that
my
core
unit,
the
gov
alpha
sent
back
due
to
over
budgeting
or
under
spending.
However,
you
want
to
slice
it,
and
there
will
also
be
some
diameter
changes
included
from
the
polls.
I
just
mentioned
on
g
uni,
rhapsody
and
the
d3m
awesome
got
through
that
without
stumbling
too
much.
So
I'm
gonna
pop
I'm
gonna,
pass
it
off
to
pablo
here
and
see
if
he
can
do
a
little
better
job.
B
C
Pablo
here
not
much
newsworthy
from
me
really,
so
this
is
going
to
be
a
shorter
than
usual
presentation.
C
So
the
ratification
polls
went
up
on
monday.
There
are
eight
of
them,
one
less
than
I
predicted
and
we'll
now
take
a
peep
at
how
they're.
Looking
as
you
can
see,
we
had
a
three
coordinate
proposal
sets
formulas
formally
submitted.
They
resubmitted
lico01
immune
file,
security
and
the
sidestream
auction
services.
C
Also,
we
had
four
coordinated
budgets,
the
resume
from
content
production,
a
budget
for
governance,
communications,
a
budget
for
real
world
finance
and
also
for
real
world
finance
and
mkr
compensation
plan.
C
And
finally,
we
had
a
meep
six
template
amendment
that
had
some
new
real
world
assets,
exclusive
questions.
So
last
time
I
mentioned
the
special
purpose
fund
as
one
of
the
formal
submissions
and
sadly,
that
proposal
was
missing
a
couple
of
details
that
we
thought
would
be
provided
before
the
ratification
polls
went
up.
That
wasn't
the
case
and
it
had
to
be
left
out
this
time
now
for
stuff
in
rfc
again
no
changes
here
since
last
week.
C
So
we
have
a
couple
of
maps.
We
have
a
partial
coordinate
proposal
set
for
pairwise
fixed
rate,
and
we
have
those
two
controversial
facilitator,
affording
and
boarding
proposals.
I
suggest
we
just
skip
ahead
and,
yes,
I
don't
know
how
many
of
you
were
excited
about
this
zero
in
all
likelihood,
but
yeah.
We
finally
have
a
dark
mode
on
the
mips
portal
and
it's
looking
pretty
rad
yep
and
that's
pretty
much
it
from
me
this
time.
Thank
you.
A
Thanks
pablo
appreciate
that
update
and
I
for
one
am
excited
about
the
dark
ones,
so
you
can
get
at
least
one
there.
Next,
we'll
have
a
form
at
a
glance
update
this
segment
where
walk
through.
What's
been
going
on
on
our
maker
forum,
I
believe
arden
will
be
taking
that
away.
D
Yep,
hey
everyone
welcome
to
the
maker
forum
at
a
glance,
a
weekly
review
of
what's
happened
on
the
forums
this
time
for
the
week
of
november
4th
to
november
11th.
So,
let's
begin
with
the
top
announcements.
First
off
gfx
labs
announces
their
delegate
platform.
The
gfx
labs
community
desires
to
be
active
participants
and
maker
governance
and
provide
communication
and
reasoning
behind
their
votes.
D
D
Next
up
maker
wormhole
definitely
takes
the
cape
of
their
logo.
I
mean,
like
I
saw
it.
It
looks
nice,
I
love
their
logo,
but
yeah,
so
they
introduced
die
wormholing,
which
is
the
ability
for
almost
instantaneous
teleportation
of
dye
between
supported
layer,
twos
and
ethereum.
The
design
is
fully
trustless
and
allows
system
participants
to
bridge
virtually
unlimited
liquidity
through
instantaneous
minting
of
dye
towards
the
destination
domains
and
l2
to
l1.
Fast
withdrawal
should
be
available
around
q1
of
2022,
and
the
complete
wormhole
feature
should
be
available.
D
Next
up
low
light
from
collateral
engineering
services,
core
unit
introduces
dss
blow,
which
is
a
new
helper
smart
contract
that
helps
make
it
easy
for
core
units
to
repay
excess
dye
to
the
maker
dial
surplus
buffer
with
dss
blow.
You
can
send
erc20
die
directly
to
the
dss
globe
contract
address
once
the
I
has
been
received
by
dss
blow,
anyone
can
call
in
the
new
blow
function
to
send
all
deposited
die
and
dss
blow
as
internal
die
to
the
surplus
buffer.
D
Moving
on
to
the
top
three
discussions,
historically,
we
have
always
tried
to
keep
up
the
surplus
to
non-stable
coin
die
ratio
around
between
one
and
two
percent.
However,
controlling
the
surplus
buffer
becomes
more
difficult
with
more
collaterals
onboarded
and
the
rapid
growth
of
die
supply
forming
conflicted
community
sentiment
in
response,
schupe
submitted
a
pre-map
discussion
proposing
that
we
introduce
an
instant
access
module
that
would
allow
any
user
to
adjust
the
surplus
buffer
based
on
parameters
decided
by
maker
governance.
He
provides
dimensions
and
parameters
in
the
post
and
seeks
feedback
and
community
discussion
on
his
idea.
D
Next,
up
with
successful
rollout
of
d3m
to
ave
paper,
imperium
wants
to
gouge
the
sentiment
of
the
community
to
make
it
a
priority
to
extend
d3m
to
compound.
He
lays
out
the
pros
and
cons
and
asks
whether
we
should
integrate
compound
into
d3m
as
a
priority.
If
yes,
votes
outnumber,
the
no
votes
on
chain
poll
will
be
presented
to
vote
on
the
prior
prioritization
of
d3
and
the
compound
and
finally,
ruli
introduces
centrifuge
collateral
and
presents
an
arranger
idea
with
mr
dao.
D
D
And
our
next
signal
request,
dimension
from
wbtc
has
recently
been
increasing
and
the
amount
of
dye
has
been
has
been
tripled
within
the
last
month.
So
a
large
market
downturn
on
wbtc
volts
running
with
higher
liquidation
ratio
is
less
likely
to
undergo
liquidations
less
likely,
resulting
in
loss
for
the
protocol
and
default
owner
and
in
response
should
be
signals
to
the
community
asking
whether
we
should
add
a
new
wbtc
vault
with
a
higher
liquidation
ratio.
D
E
Which
I
just
want
to
jump
in
here
really
quickly
if
any
of
the
facilitators
are
actively
hiring,
please
reach
out
to
artem
and
let
him
know
and
he'll
add
whatever
role
you're
hiring
for
in
the
help
wanted
section,
along
with
some
contact
info
for
who
they
should
contact.
If
they're
interested
in
liberal.
A
I
know
well
thanks
for
that
harding.
We
really
appreciate
it.
Yeah,
we'll
move
on
to
our
next
segment.
Today
we
are
kind
of
fortunate
in
that
we
get
a
presentation.
It
doesn't
happen
every
week,
so
without
further
ado,
I'll
pass
it
over
to
the
sidestream
auction
services,
who
will
be
proposing
or
showing
us
a
demo
right.
E
F
Cool,
so
can
I
see
my
screen
great,
so
yeah,
thanks
for
having
us
I'm
daniel,
I'm
the
facilitator
of
sidestream
auction
services,
and
here
we
are
an
ses
incubation
team
that
is
going
to
form
a
new
car
unit
and
we
joined
the
incubation
in
summer
and
now
since
monday,
we
are
up
for
voting
and
hopefully
will
be
a
permanent
core
unit.
As
of
december.
F
And
yeah
so
we're
an
engineering
core
unit
and
our
mission
is
to
improve
security,
transparency
and
accessibility
of
the
maker
protocol
by
providing
and
maintaining
auction
services
through
open
source
development,
so
yeah.
We
are
really
diving
into
all
these
things
around
liquidations
and
auctions
and
yeah.
F
So
we
actually
so
one
slide
about
us,
we're
from
sidestream
and
that's
a
german
software
development
company
that
myself
and
my
co-founder
niels
started
like
two
years
ago
and
we
have
grown
it
to
20
people
and
we
basically
do
a
lot
of
custom
software
development
services
for
like
bigger
customers,
a
lot
of
in
traditional
finance,
but
yeah.
F
We
are
really
happy
to
work
together
with
maker
dao
and
we
imagine
to
have
a
dedicated
business
unit
with
the
team,
that's
already
working
on
on
the
auction
services
since
summer,
and
so
that
we
are
very
focused
on
this
project,
but
can
also
leverage
the
software
agency
know-how
that
we
have
from
like
other
projects
that
we
do
and
the
team
that
is
working
on
the
auction
services.
So
it's
myself
as
a
facilitator.
F
Lucas
as
a
product
manager,
valentin
and
band,
has
first
act
developers
and
neos
as
a
devops
engineer,
and
so
some
of
us
work
half
time
as
you
can
see
on
the
fte
count
and
some
full
time
for
for
maker
and
the
last
slide
before
the
actual
demo
a
bit
more
detail.
What
we're
going
to
do
so.
F
Also,
a
lot
of
our
work
is
about
providing
intuitive
uis,
but
also
well-maintained
apis,
so
interfaces
for
non-technical
and
technical
users
to
to
basically
increase
the
accessibility
and
adaption
of
auctions
and
liquidations.
F
All
the
auction
liquidations
functionality
and
all
the
apis
and
uis
are
in
place
to
cope
with
that,
but
also
in
case
of
inefficient
market
conditions.
Through
our
software,
we
hope
to
basically
increase
inquiry,
increase
participation
to
protect
the
protocol,
but
also
vote
on
this,
and
here
with
that,
I
will
just
show
you
a
video
of
what
we
have
built
so
far.
C
Everyone,
my
name,
is
lucas
product
manager
for
sidestream
auction
services
and
I'm
very
happy
to
give
you
a
quick
demo
walkthrough
of
our
mvp.
We
created
the
first
ui
that
facilitates
participation
in
collateral
auctions
via
flash
landing
feature.
This
means
there
are
no
capital
requirements
for
participation,
so
user
doesn't
need
to
have
any
dye
balances
available.
The
only
requirement
is
actually
to
have
some
ease
in
order
to
pay
for
the
transaction
fees.
Let's
dive
right
away
into
the
ui.
C
What
you
can
see
here
is
the
main
auctions
table
that
contains
all
of
the
active
auctions
and
those
that
need
to
be
restarted
for
demo
purposes.
We
are
currently
on
kovan
testnet
indicated
here
in
the
header
as
well,
and
that
is
also
why
we
can
see
some
very
small
eth
auctions
currently
ongoing
here
now.
The
selection
of
one
of
these
auctions
provides
in-depth
information
on
this
particular
auction,
like,
for
example,
the
auction
time
till
it
ends.
C
Then
we
have
the
current
auction
price
per
collateral
in
die
and
the
equivalent
collateral
price
on
another
marketplace
for
our
mvp.
We
focused
on
uni
swap,
but
the
idea
is
to
connect
to
other
taxes
in
the
future
as
well,
so
this
should
give
the
user
already
a
quite
good
indication
on
the
potential
arbitrage
opportunity
to
due
to
price
differences
between
the
first
marketplace,
which
is
basically
the
auction
and
the
second
marketplace
in
this
case.
C
Uni
swap
this
becomes
even
more
evident
when
clicking
on
the
directly
swap
into
profit
button
again
here
you
can
see
the
different
prices
auction
and
on
uni
swap,
and
we
also
get
an
indication
on
the
potential
profit
that
can
be
made
by
participating
in
this
auction.
We
give
an
approximation
of
the
involved,
transaction
fees
and,
finally,
a
given
indicator
on
the
overall
transaction
outcome
now
before
the
actual
execution
of
the
flash
landing
to
all
steps
are
required.
C
This
is
a
one-time
authorization
of
the
choice,
contract
and
the
authorization
of
the
clipper
contract,
which
needs
to
be
happen
once
per
collateral
type.
Now,
since
these
have
been
done
for
the
wallet
in
use
in
the
past,
we
do
not
have
to
do
them
and
can
directly
execute
the
flash
loan.
But
before
that,
let's
have
another
look
into
the
wallet
to
re-emphasize
the
point
that
there
is
really
no
die
requirement,
so
no
capital
requirement
to
participate
in
this
auction.
C
Now,
if
we
hit
the
execution
button,
we
are,
as
always,
prompted
with
the
meta
mask
window
to
confirm
the
transaction
fees.
We
are
doing
that
and
now
we
get
a
visual
indication
here
with
this
yeah
toast
notification
in
the
top
about
the
current
status
of
the
transaction,
of
course,
depending
on
the
current
network
congestion.
This
can
take
some
time,
but
with
the
notifications
here
and
also
the
spinners
who
want
to
prevent
to
leave
the
user
in
the
dark
about
the
current
transaction
state.
C
Now
that
the
auction
was
successfully
mined
and
confirmed,
we
do
display
that
here
at
the
top,
stating
that
the
auction
is
finished
and
let's
have
a
quick
look
into
our
wallet
again.
What
we
can
see
here
now
is
that
we
were
able
to
exploit
the
arbitrage
opportunity
obtain
some
dye
by
participating
in
this
auction.
So,
to
put
this
feature
a
little
bit
into
perspective
in
case
of
an
inefficient
keeper
ecosystem,
meaning
the
bots
that
would
automatically
detect
and
engage
in
these
auctions.
C
So
clearly,
this
was
a
lot
of
information
and
context
that
is
required
for
a
user
to
actually
understand
all
the
steps
and
actions
on
this
ui
for
a
new
user.
This
might
be
complex
and
even
intimidating,
which
in
the
end
might
hinder
auction
participation,
and
that
is
why
we
also
implemented
an
explanatory
mode
changeable
with
the
switch
in
the
header
yeah.
It
provides
more
context
on
what
this
ui
is
good
for
and
especially
why
and
how
users
should
interact
with.
A
C
C
For
example,
when
the
user
tries
to
switch
from
within
metamask
to
a
network
which
is
currently
not
supported,
we
do
reflect
that
with
a
pop-up
that
enforces
the
user
to
change
to
one
of
the
two
networks
that
are
currently
supported
yeah.
So
this
is
basically
it.
I
hope
this
gave
a
good
first
insight
on
the
capabilities
of
our
mvp.
There
is
more
to
come
in
the
future
and
we
are
really
excited
and
looking
forward
to
build
for
now
feel
free
to
have
your
own
look
on
auctions.makerdial.network
code
is
also
open
sourced.
C
F
E
D
B
B
I
have
a
question
so
I
remember
when
we
originally
put
enabled
like
flashlight
at
auctions.
We
were
not
concerned,
but
we
were
like
aware
that,
like
like
mev
might
be
an
issue
right,
like
I
mean
having
like
flashbots
and
things
like
front
run,
the
flashlight
auction
transactions.
B
Yeah,
because
the
flash
loan
auction
transactions
are
quite
easy
to
front
run
right
because
they're
kind
of
self-contained
and
they
don't
require
any
capital
like
so
this
kind
of
ui.
It
looks
awesome
right,
it's
amazing,
but
if
every
time
anyone
uses
it
like
the
transaction
fails
because
they
get
from
one,
then
that's
not
ideal.
B
F
Yeah,
so
I
mean
the
the
main
purpose
of
the
ui
is
to
to
yeah,
have
basically
another
backstab
solution,
so
that
if
there's
no
participation
that
you
can
that
participate
through
the
ui
and
then
the
next
step
will
be.
We
are
basically
providing
writing
scripts
or
apis
to
participate
in
these
auctions.
But
currently
we
yeah,
we
don't
see
front
running
as
a
problem,
but
rather
I
mean
it's
good.
If
there's
someone
bidding
on
these
options.
A
Awesome,
do
we
have
any
other
questions
for
about
the
side
stream
or
the
core
unit
in
general
or
the
demo
itself?
I
guess.
F
I
have
one
question.
First
of
all,
I
think
this
is
really
cool
for
our
keeper
ecosystem,
but
I'm
curious,
you
said
you're
you're
only
supporting
cuny
swap
is
it
v2
or
v3
and
then
second
does
the
price
in
the
ui?
Does
it
include
slippage
as
well,
so.
F
E
To
support
also
automatic
routing
at
some.
C
E
And
then
yes,
it
includes.
E
E
Will
be
sold
using
the
uniswap
sdk?
F
A
Awesome
appreciate
you
coming
on
and
giving
this
demo,
as
mentioned,
that
vote
is
live,
so
do
take
a
look
and
give
your
opinion
if
you're
a
mkr
voter
and
yeah
thanks.
So
much
thanks
for
having
us.
A
All
right
appreciate
it
and
now
we're
gonna
move
on
to
the
oracle
score
unit.
I
believe
nick
has
put
together
a
little
something
for
our
consideration
today.
G
G
Okay,
you
need
to
enable
screen
sharing.
For
me.
G
There
we
go
okay,
so
so
we
we
talked
about
this
on
previous
calls
right
that
right
now
the
gas
costs
for
the
oracles
are
not
included
in
the
current
oracle
core
unit
budget
and
that
those
would
come
at
a
later
date.
G
I've
also
covered
in
the
past
on
here
that
the
maker
foundation
generously
offered
to
front
the
gas
cost
since
august
1st
in
order
to
give
the
core
unit
more
time
to
kind
of
figure
out
how
to
to
best.
Do
this,
and-
and
what
I
mean
by
that-
is
that
the
tooling
that
we
had
back
then
to
really
delve
into
you
know
what
oracle's
cost
us
from
from
an
infrastructure
point
of
view
were
were
not
very
granular,
so
we
basically
had
the
holistic
view
of
well.
G
We
know
how
much
eath
is
going
in
and
so
that
that's
how
much
we're
spending
in
total,
but
we
didn't
have
this
more
granular
kind
of
scalpel.
Like
view
of
oh
well,
the
eth
oracle
is
costing
us
this
much
right.
The
the
zrx
oracle
is
costing
us
this
much
right.
You
know
the
uniswap,
lp
oracle's
cost
cost
this
much,
and
so
we
we've
recently.
G
You
know
started
creating
tooling
and
and
dashboards
that
gives
us
insight
into
these
types
of
things,
and
so
it's
kind
of
time
to
finish
the
transition
of
the
oracle
core
unit.
You
know
and
the
oracle
infrastructure
out
of
the
foundation
and
and
to
have
it
be
kind
of
completely
decentralized.
G
So
the
the
way
that
we're
thinking
of
of
structuring
this
is
that
we
will
ask.
We
will
typically
be
asking
for
kind
of
a
rolling
three-month
budget,
so
we
basically
will
submit
a
mip,
a
subproposal,
and
we
ask
we
say
we
expect
to
spend
this
amount
in
the
next
three
months.
G
What's
different
from
this
first
proposal
is
that
it
will
also
include
the
repayment
to
the
maker
foundation
for
fronting
the
cost
since
the
beginning
of
august
and
right
and
since
that's
kind
of
a
rolling
kind
of
a
timer
we've
kind
of
assumed
that,
due
to
the
way
kind
of
governance
works,
is
that
we
we
think
okay
by
end
of
december,
this
mip
should
be
passed
or
some
version
of
this
mip
should
be
passed
and
so
we'll
we'll
say
in
the
mip
that
the
cutoff
for
that
is
this
end
of
december
right.
G
But
if
it
gets
passed
earlier
than
that
or
or
later
than
that,
right,
we'll
have
to
adjust
the
the
the
payment
amount
to
the
foundation.
But
what
I,
what
I
really
want
to
get
across
today
and
then
with
this
mip,
is
really
just
the
the
analytics
themselves
right,
where
they're
coming
from
and
how
we
deduce
the
numbers
in
the
budget
for
the
method
itself.
G
So
these
are
basically
dune
analytics
that
we've
hooked
up
to
the
to
the
smart
contracts.
So
we
are.
We
can
happily
provide
the
link
for
these,
so
you
can
verify
the
data
for
yourself,
but
it
basically
uses
the
transaction
data
directly
from
the
blockchain
in
order
to
compile
these
metrics.
G
So
this
is
for
the
cost.
Since
august
1st
to
november
8th,
which
was
a
few
days
ago,
keep
in
mind
that
this
is
just
the
medianizers.
This
is
not
the
rest,
like
the
entirety
of
the
cost.
So
before
you
get
all
happy,
oracles
are
way
more
expensive
than
this,
so
the
medianizer
right
is
just
this
continuous
oracle
kind
of
price
right.
This
is
not
the
value
that
the
maker
protocol
uses
right.
There's
there's
other
layers
to
that.
G
This
is
just
a
bunch
of
feeds,
reach
a
consensus
that
the
price
of
you
know
eth
at
time
x
is
blah,
so
we
spent
about
a
little
over
six
hundred
thousand
dollars
on
that
since,
since
august
and
we'll
kind
of
go
later
more
more
in
detail
into
what
that
extrapolates
to
over
the
course
of
a
year
here,
you
can
kind
of
see
the
breakdown
by
asset
pair,
I'm
not
sure
how
well
you
can
see
this.
G
So
let
me
try
to
zoom
in
a
little
bit,
but
you'll
notice
right
that
some
assets
have
these
very
large
spikes,
and
it's
because
you
know
their
volatility
profile.
You.
D
G
Just
is
is
very
high
right
or
especially
in
the
kind
of
current
markets
right
where
everything
just
has
a
random.
I
don't
know
70
pump
or
something
right.
You
can
easily
have
days
right
where
you
spend.
You
know
the
amount
that
you
would
normally
spend
in.
You
know
say
two
three
weeks
of
operating
one
of
these
oracles
in
a
single
day
and
and
that
kind
of
leads
to
a
lot
of
uncertainty
when
it
comes
to
trying
to
predict
the
cost
of
any
single
oracle.
G
G
Okay,
so
this
table
and
again
I
apologize.
I
know
the
text
is
quite
small.
This
table
kind
of
shows
what
we
spend
on
the
individual
medianizers
so
compared
to
that,
you
know
600
thousand
dollar
number
above
right.
This
is
what
we
spent
since
august.
G
First
on
the
individual,
medianizers
you'll
notice,
that
eth
is
the
highest,
and
that
is
because
there
are
certain
special
factors
about
the
eth
oracle
that
we
use
to
make
it
update
more
frequently,
because
eth
is
such
a
very
big
part
of
our
portfolio.
G
G
What
this
means
is
that,
every
time
the
market
price
of
eth
differs
from
the
price,
the
medianizer
is
reporting
by
over
half
a
percent
the
it
gets
triggered
to
update
the
price.
Compare
this
to
say
something
like
tether
right,
where
you
know
the
price
of
tether
does
not
move
very
much
right.
It's
a
it's
a
stable
coin
right
or
we,
we
think
it's
a
stable
coin
right,
and
so
the
the
costs
for
updating
the
tether
oracle
are
substantially
lower.
G
It
really
likes
jumping
around
here
right,
so
here
you
can.
G
Here
you
can
see
the
daily
gas
consumption
by
and-
and
this
is
just
the
medianizers,
so
this
is
basically
all
of
the
medianizers
together.
How
much
are
we
spending
on
gas
you'll
notice
that,
even
though
it's
quite
volatile,
there
is
clearly
an
uptrend
since
the
beginning
of
august,
and
this
will
kind
of
be
important
later
when
we're
trying
to
create
the
budget,
because
it's
not
you
know
what
we
think
the
budget
has
to
be
based
not
on
well.
G
We
spent
this
much
in
the
last
three
months,
so
this
is
how
much
we
think
we'll
need
in
the
next
three
months
right.
You
have
to
assume
that
this
trend
will
continue
or
or
maybe
even
accelerate
right,
and
so
the
budget
itself
comes
down
to
not
really
the
realized
costs,
but
basically
the
an
expression
of
a
defensive
like
or
more
of
a
defensive
expression
of
in
the
worst
case.
G
Okay,
so
now
going
down
to
the
oracle
security
module
and
the
spotter
and
the
uniswap
lp
cost
so
from
a
dashboard
kind
of
tracking
perspective.
Until
now,
we
haven't
been
able
to
segregate
these.
Yet
in
a
in
a
automated
type
of
way,
so
to
kind
of
explain,
kind
of
what
the
the
architecture
here
looks
like
is
right.
So
you
have
this
medianizer
here
that
we
talked
about
earlier.
That
is
this
real-time
price.
G
You
have
the
oracle
security
module
right,
which
just
applies
this
delay
to
the
price,
so
it's
poked
once
an
hour
and
it
it
takes
in
whatever
price
is
in
the
medianizer
and
whatever
price
was
queued
up
previously
kind
of
becomes
the
canonical
price.
So
so
that's
the
the
osm,
the
oracle
security
module,
and
then
you
have
one
more
module
on
the
maker
protocol
side.
That's
part
of
the
core
maker
protocol,
which
is
called
the
spotter,
and
the
spotter
is
essentially
how
do
you?
How
does
maker
read
from
the
osm
right?
G
So
the
spotter
needs
to
be
basically
checking
the
osm
to
map
that
value
internally,
because
what
you
don't
want
is
you
don't
want
the
maker
oracle
protocol
kind
of
core
contracts
to
be
reaching
out
to
this
external
oracle
security
module
contract
constantly?
So
it
would
be
a
very
expensive
kind
of
gas
operation,
and
so
this
is
kind
of
the
spotter.
G
Is
kind
of
just
a
you
know
import
at
once,
and
now
everything
in
the
core
contracts
can
use
it
as
an
imported
value,
rather
than
always
having
to
fetch
this
much
more
expensive
kind
of
external
value,
and
so
you'll
notice,
right
that
this
number
is
is
higher
than
the
medianizer
number.
But
it's
essentially
because
it's
this
homologation
of
this
oracle
security
module
of
this
spotter
module-
and
it
actually
also
includes
the
uniswap
v2
lp
oracles
and
the
new
g-uni
oracle.
So
it's
it's
kind
of
this
superset.
G
So
from
this
top-level
view,
yeah.
This
is
kind
of
how
much
we've
spent
since
august.
First,
we've
spent
about
almost
1.4
million
dollars,
but
we'll
kind
of
break
it
down
a
little
more
into
the
components
themselves.
In
a
little
bit
again,
you
see
the
gas
consumption
right.
Trending
up
this
corresponds
exactly
what
we
saw
with
the
medianizer
right
going
up
since
august
1st.
G
And
so
here
here's
where
we
start
to
break
it
down
a
little
bit,
so
these
are
kind
of
the
this
is
the
gas
profile
kind
of
distribution
of
a
of
using
this
omega
poker
kind
of
contract
that
we
use
so
effectively.
What
this
means
is,
every
time
you
poke
this
omega
poker
contract
and
it
updates
the
osms
it
updates
the
spot
or
it
updates
the
uniswap.
You
know
lp
oracles,
to
uni.
G
This
is
kind
of
the
distribution
of
the
gas
costs
within
that
call
right
and
so
you'll
see
that
you
know
we
spend
30
percent
on
the
spotter.
We
spend
24
on
the
oracle
security
module.
We
spend.
You
know
thirty
three
point:
three
percent
on
the
uniswap
v2
lp
oracles
and
seven
percent
on
g
uni
right
and
then
there's
kind
of
five
percent-
that's
just
like
other.
G
That's
just
kind
of
the
the
scaffolding
so
using
this
right
top
level
number
that
we
had
right
along
with
this
gas
profile
breakdown,
we
can
start
to
calculate
you
know
what
we're
spending
for
each
one
right.
So
this
is
just
the
those
proportions
all
right.
G
So
in
that
time
right
we
spent
this
much
on
osm's
this
much
on
lp
oracles,
this
much
on
spotters
and
this
much
on
g
uni,
you'll
notice
that
all
of
those
amounts
are
kind
of
less
than
the
medianizer,
and
that
makes
sense
because
the
medianizer
does
a
lot
more
kind
of
computational
work.
G
So,
let's,
let's
break
it
down
kind
of
further.
You
know
how
much
does
an
osm.
How
much
does
each
osm
cost
each
osm
costs
about
75
000?
G
Sorry,
during
over
the
course
of
a
year
each
you
know
swap
lp
oracle
kind
of
costs,
us
184
thousand
dollars
a
year,
each
spotter
kind
of
costs
us
60
000
a
year.
Okay,
great
now,
that's
not
the
total
cost
for
the
oracles,
but
now
we
have
all
the
pieces
that
we
need
to
to
come
up
with
those
costs
right,
so
a
standard
oracle
right
it
starts
with
a
medianizer.
G
It
then
has
an
oracle
security
module
and
it
then
has
a
spotter
right.
So
it's
the
accumulation
of
of
all
of
these,
so
for
the
eth
oracle
extrapolated
over
a
year,
we're
spending
total
331
thousand
dollars.
To
put
that
into
perspective.
Compare
that
to
something
like
yfi,
which
has
more
kind
of
lacks,
spread
parameters
and
is
kind
of
like
maker
and
that
it's
a
stable
coin,
that
it
doesn't
fluctuate
very
much
in
price.
G
The
wi-fi
oracle
is
259
000
a
year
right.
So
you
have.
You
can
see
this
kind
of
range
right.
These
are
the
exact
same
architecture,
and
the
only
difference
is
here
is
the
volatility
profile
of
the
asset
and
kind
of
these.
These
slight
little
tuning
things
like
the
the
spread
parameter.
G
So
there
is
some
room
to
make
oracles
more
or
less
sensitive
and
tune
them
if
we
want
to
tune
the
costs
up
or
down.
G
Looking
at
the
uniswap
lp
oracles
right,
you
have
the
oracle
itself
plus
you
have
this
spotter,
so
those
cost
about
245
000
annually,
and
then
you
have
g
uni
now
with
g
uni.
This
is
we're
not
a
hundred
percent
confident
in
this
number,
because
we
just
on
boarded
it
very
recently.
G
So
the
data
set
that
we
have
to
work
with
is
quite
limited,
but
it
shouldn't
be
off
by
more
than
say,
like
I
don't
know,
five
percent.
So
this
is
it's.
It's
still
very
expensive
right,
it's
significantly
more
expensive
than
the
other
type
of
oracles.
So
that's
something
that
we
need
to
consider
when
we
think
about
onboarding
more
of
the
gelato
based
tokens.
G
G
So
governance
already
had
a
signal
request
to
off-board
ave,
balancer
and
comp,
so
those
are
all
right,
there's
pulling
votes
that
are
active
for
that.
So
I
hope
governance
kind
of
considers
the
return
that
these
vaults
are
making
relative
to
what
they're
costing
us
just
in
oracle
expenses
right,
that's
not
even
compensating
for
the
risk
component
that
mkr
is
insuring.
It's
that's
just
the
oracle
expenditures,
but
there's
there's
a
couple.
Other
ones
right
that
are
that
are
unprofitable
as
well.
G
Here,
right
that
really
a
steak
death
ride
is,
is
limited
due
to
debt
ceiling
right
now,
so
that
one's
kind
of
a
an
outlier
here
but
but
the
rest,
we
should
have
very
good
reasons
for
why
we
continue
to
have
these
collateral
types
when
they're
they're
losing
us
money
by
the
second
okay.
So
what
does
this
cost
on
an
annual
basis,
and
this
is
kind
of
the
sticker
shock
number
right?
This
was
what
we
were
spending
on
medianizers.
G
This
is
what
we
spent
on
the
right,
the
osm,
spotter,
uniswap,
v2,
oracles,
right
and
now
extrapolating
since
august.
First,
extrapolating
that
over
a
year,
you
arrive
at
this
number-
that's
that's
7.2
million
dollars,
which
is
a
huge
chunk
of
change
right
and
what
you
have
to
keep
in
mind
is
right.
We
talked
about
this
earlier
right.
Gas
costs
are
trending
up
and
we
have
to
assume
that
the
kind
of
worst
case
scenario
that
they
will
keep
going
up
right,
and
so
the
budget
has
to
reflect
that.
G
So
what
do
we
need
for
this
particular
proposal?
Well,
extrapolating
that
you
know
7.2
million
number
over
3
months
gives
us
this.
1.8
million
dollar
number
and
then
because
we
need
to
guard
against
gas
cost
increases.
G
You
know
we
put
in
a
multiplicative
factor
of
two,
so
we're
going
to
be
asking
for
three:
almost
3.6
million
dollars
to
kind
of
safely
cover
the
oracle
gas
cost
budget
for
through
the
end
of
march,
2022.
G
Then
we
also
need
to
refund
the
maker
foundation
right
for
the
costs
that
they've
incurred
thus
far
and
the
cost
that
we
expect
they're
going
to
incur
before
before
this
proposal
is
kind
of
passed
and
remember,
we
use
the
end
of
december
as
a
kind
of
cutoff
date,
but
if
it's
earlier
than
that
right,
of
course,
the
foundation
would
get
less.
If
it's
more
than
that
right,
the
foundation
gets
more
right.
G
It's
really
just
moving
in
and
out
of
this,
this
bucket
up
here
right,
so
someone
is
paying
for
it.
It's
either
coming
out
of
this
bucket
or
the
foundation's
paying
for
it,
but
ultimately
right
it's
it's
the
same
pool
of
money
and
then
as
a
further
defensive
measure.
G
What
we
wanted
is
we
want
effectively
an
extra
three
months,
but
not
kind
of
modulated
by
this
gas
scaling
factor
just
the
vanilla
three
months
in
kind
of
an
emergency
kind
of
separate
multi-sig,
which
you
know
we
if,
if
like
we
literally
blow
through
our
expectations
that
gas
costs
are
going
to
double,
then
we
can
dip
into
into
this.
G
We
obviously
no
one
expects
to
dip
into
the
emergency
funds,
but
you
know
I'll
give
you
an
example
right
on
black
thursday
we
saw
gas
prices
that
went
up
to
three
thousand
weight
and
because
the
sky
was
falling
and
because
you
know
the
difference
between
you
know,
3000
way
was
either
you
got
liquidated
or
you
didn't
get
liquidated
right.
People
were
easily
willing
to
pay.
You
know
eight
thousand
dollars
a
transaction.
G
You
know
five
thousand
dollars
a
transaction
to
ensure
that
you
know
they
were
safe,
so
in
in
a
day
like
that
right,
depending
on
how
close
we
are
to
that,
you
know
end
of
the
three-month
cycle.
G
If
you
have
one
of
these
outlier
events
right
this,
that
would
be
one
of
those
situations
where
you
would
have
to
dip
into
the
into
the
emergency
fund,
which
basically
leads
to
the
kind
of
big
sticker
shock
number
here
of.
So
what
is
what
does
all
that
sum
into?
And
it
sums
to
you
know
8.4
million
dollars,
so
I
I
want
to
kind
of
open
it
up
to
questions.
I
I
know
we.
We
don't
have
a
lot
of
time
either.
G
One
note
that
I
do
want
to
bring
up
is
well.
I
I
think
a
common
reaction
to
this
might
be
why
why
the
hell
are
we
spending
this
much
on
oracle's?
When
you
know
we
could
use
some
external
provider
or
something-
and
the
answer
to
that
is
one.
Is
you
don't
want
to
outsource
your
security
to
someone
else,
but
two?
Is
it
wouldn't
actually
save
us
that
much
money?
G
So
the
actual
you
know
oracle
itself.
Piece
right
is:
is
this
piece
right
here,
the
other
pieces
that
maker
has
built
around
the
oracle,
to
ensure
that
you
know
there's
were
you
know,
defended
against.
You
know
an
oracle
attack
to
ensure
that
you
know
we're
not
going
to
just
automatically
liquidate.
You
know
a
couple
billion.
G
You
know
from
one
moment
to
the
next
without
giving
people
a
chance
to
pay
back
right.
That's
that's
this
number
over
here
right.
So
really,
only
a
third
of
the
costs
actually
are
coming
from
our
oracles.
The
other
two
thirds
of
the
costs
are
really
from
the
kind
of
maker
scaffolding
that
we've
built
around
around
the
oracles.
G
It's
based
on
the
current
run
rate,
so
so
right,
I
I
don't.
I
don't
think
the
sampling
from
the
last
you
know
since
august.
Is
that
relevant
like
if
you
zoom
out
from
the
chart
right,
it's
just
very,
very
volatile
so
trying
to
establish
some
kind
of
trend
using
that
type
of
limited
data
set,
I
think,
is,
is
quite
foolish,
so
I
wouldn't
focus
too
much
on
on
the
2x
number.
G
It's
very
much
like
a
spitball
kind
of
number,
because
I
don't
think
there's
a
the
data
set
is
really
big
enough
to
to
come
up
with
with
anything
more
kind
of
algorithmic
or
deterministic.
E
Yeah,
I'm
more
looking
at
just
like
the
average
gray
assumption
used
for
the
forecast,
essentially.
G
Yeah
well,
okay,
so
so
it
depends.
If
you
want
to
use
the
the
today
value
or
write
the
the
sampling
over
over
the
last
period
right,
I
I
don't
take
much
stock
and
today
the
gas
price
is
this
because,
a
week
ago
you
know,
I
saw
like
60
away.
B
G
Yeah,
so
basically,
the
other
thing
that
came
out
of
this
right
is
these.
Are
the
numbers
that
really
governance
should
consider
when,
when
like
thinking
of
onboarding,
a
new
collateral
type
is
well?
If
it's
costing
us,
you
know
let
right
east
cost
us
like
has
cost
us
historically
330
k
since
august
right
and
so
now,
let's
say
you,
the
gas
price
is
double
right,
and
so
now
you're
looking
at
you
know,
you
know
660k.
G
So
any
vault,
that's
going
to
make
less
than
660k
like
you
know,
don't
even
bother
right.
It's
you're,
just
coin
flipping
based
on
you
know,
is
the
gas
price
not
going
to
go
up
and
and
screw
us
right.
So
we
really
need
to
be
thinking
of
only
like
the
really
big
ticket
vaults
that
can
generate.
You
know,
you
know
at
least
two
million
of
stability
fees
at
the
very
least
right
and,
and
even
that
is
kind
of
you're
still
paying
people.
G
You
know
very
high
salaries
to
to
onboard
those
things,
so
you
should
really
be
thinking
higher.
G
Yeah,
so
so
I
didn't
put
that
implementation
in
the
mip,
because
I
wanted
to
give
the
core
unit
more
flexibility
in
trying
out
different
solutions
with
that
regard,
but
yeah.
So
so
what
what
I?
What
I
think
we
would
do
is
we
would
want
some
of
it
in
eth
and
and
some
of
it
and
die,
and
I
think
the
the
very
spitball
and
like
the
the
kind
of
rough
number
off
the
top
of
my
head
would
kind
of
be
like
a
50
50.
Where
you
don't
want.
G
You
know
you
want
some
exposure
to
eat,
but
you
don't
want
all
of
your
exposure
in
each
ideally
at
some
point
in
the
future
we
get
actual
gas
futures
that
we
can
use
to
hedge,
our
gas
costs
kind
of
like
airlines
use.
You
know,
oil
futures
to
hedge,
their
their
their
jet
fuel
costs,
but
but
but
yeah.
G
So
that
that's
that's
kind
of
where
I
wear
my
heads
at
with
this,
but
we
will
definitely
try
a
few
different
type
of
models
to
kind
of
see
what
works
better
right.
The
goal
with
this
is
not
even
necessarily
to
get
the
ultimate
lowest
cost
for
running
the
oracles,
it's
more
just
to
bound
those
costs
right
so
to
make
them
more
deterministic
and
basically,
the
more
we
can
constrain.
You
know
what
the
range
of
of
cost
is
going
to
be
right,
the
the
better
off
we
we
are.
E
G
G
G
Yeah,
I
I
let's
discuss
it
more
on
the
forum.
My
the
reason
I
don't
want
it
all
in
eth
is
essentially
because
of
these.
You
know
black
thursday
type
scenarios
where
everything
is
crashing
right,
so
prices
are
going
down
and
yet
gray
could
be
obscenely
obscenely
high
right
with
people
willing
to
pay.
You
know
five
thousand
dollar,
ten
thousand
dollar
transaction
fees
right,
and
so,
if,
if
we
are
completely
an
eth,
we
would
be
screwed
by
by
that
type
of
by
that
type
of
event.
E
Yeah,
so
in
thinking
about
this
problem,
like
long
term
before
knowing
what
you're
gonna
do
nick,
I
always
kind
of
imagined
that
the
maker
protocol
could
use
the
surplus
buffer
directly
in
sort
of
like
like
a
split
solution
where
you
know
some
of
it
is
held
in
the
circle's
buffer,
a
little
bit
of
burns
on
kr
and
then
a
little
bit
is
used
to
buy
this.
E
You
know
dollar
cost
averaging
over
time
and
basically
making
a
native
pool
within
the
protocol
of
east
that
can
maybe
automatically
like
pay
oracles
or
something
like
that's
kind
of
like
a
the
imagination
I
had
in
my
head
like
years
ago
like
a
year
ago.
I'm
curious
what
you
think
about
that
vision
and
number
yeah.
Here's
what
you
think
about
that.
G
I
I
think
it's
something
the
community
should
discuss
more.
My
initial
inclination
is
that
probably
shouldn't
do
that.
Just
because,
like
essentially
you
don't
wanna
like
like
the
purpose
of
the
core
units
is
to
delegate
responsibility
of
things,
so
governance
doesn't
necessarily
have
to
be
an
expert
on
all
of
this
different
stuff
and
so
taxing
kind
of
governance,
with
the
inventory
kind
of
management
for
specifically
for
the
oracles.
G
I
I
think
would
would
be
a
mistake
now.
If
governance
wants
to
in
general,
you
know
have
exposure
to
ether
and
diversify
some
of
the
surplus
to
eth.
I
mean
you
know
have
at
it
right,
but
I
wouldn't
do
it
specifically
because
of
like
oracle
kind
of
eth
inventory
management.
A
I
guess
I
had
a
question
in
terms
of
the
uniform
oracle's,
you
kind
of
mentioned
the
v2
ones
and
their
costs
there.
Are
they
any
cheaper
when
we
are
already
running
oracles
for
the
underlying
assets
like
I
don't
I
guess
I
don't
understand
how
that
works.
Oh.
G
So
this
was
purely
for
the
gas
cost
of
running
that
uniswap
oracle
smart
contract.
So
if
you
need
to,
in
addition
to
those
create
new
medianizers
right
for
the
component,
tokens
then
go
ahead
and
add
the
medianizer
costs
on
top
of
that
right
right
now
we
don't
have
that
for
any
of
the.
Let
me
let
me
check
one
more
second,
I'm
pretty
sure
that's
a
true
statement.
G
Yeah.
We
don't
currently
have
that
for
for
any
of
the
tokens
right
where
we
onboarded
a
uniswap
lp
token
pair,
and
we
didn't
have
the
medianizers
already.
You
know
quote-unquote
for
free,
because
we'd
already
onboarded
the
the
those
those
assets
in
maker.
E
I
was
wondering
really
talking
about
the
onboarding
new
collaterals.
It
will
be
great
if,
in
your
risk
assessment,
you
could
at
the
cost
that
well,
I
know
you
we
don't
know
but
like
what
will
be
the
cause
of
that
collateral
and
I'm
talking
yeah
because
of
the
gelato
lp
token.
I
was
surprised.
A
We
did
switch
back
to
david's
screen,
fyi.
G
Oh
oops
yeah,
I
was,
I
guess
if
you
want.
E
G
I
don't
know
I
was
you
guys
have
already
seen
it
but
yeah
it's
it's
377k
for
the
g
uni,
one
plus
minus
five
percent
uncertainty,
because
the
the
sample
size
is
quite
small.
B
A
And
would
that
be
like
similar
for
other
v3
implementations,
or
is
it
unique
to
gelato.
G
All
right
so,
for
example,
kurt
is
creating
the
curve
lp
oracle
right
now.
So
we
could,
you
know,
do
the
staked
eath
curve
pool
right
as
a
as
a
collateral
type
right
and
then
people
can
basically
get
staked,
eat,
yield
and
the
curve
lido
yield
and
and
such
right,
but
we
don't
necessarily
right,
know
beforehand,
right
one,
how
much
demand
there's
going
to
be
for
that
vault
and
two.
You
know
how
much
that
oracle
is
is
going
to
to
cost
until
we
kind
of
really
see
it
running.
G
So,
there's
there's
always
going
to
be
cases
like
that
right,
where
it's
not
as
clear-cut
for
governance
on
you
know
what
what
the
real
costs
are,
what
the
real
upside
is,
you
know
so
to
a
degree
you
know
we
have
to.
We
have
to
make
some
some
educated
guesses
at
some
point,
but
for
this
other
stuff
right,
the
the
standard,
oracles,
the
you
know,
just
a
normal
collateral
type.
G
We
know
very
precisely
right
in
you
know
what
you
know
to
to
a
degree
of
uncertainty
with
gas
prices,
but
other
than
that,
like
we.
We
we
have
experience
in
knowing
how
much
this
is
going
to
cost
right,
and
so
we
can
very
clearly
say:
okay,
look!
If
your
collateral
type
doesn't,
you
know,
generate
like
2
million
instability
fees
like
let's
not
bother.
G
It's
not
necessarily
that,
like
once
you've
built
it
you've
kind
of
built
it
right,
so
the
the
cost
to
just
deploy
another
one
is
is
rather
low
right,
they're,
just
kind
of
incremental
costs.
It's
the
building
it
out
the
first
time
right
and.
F
G
It
audited
and
feeling
comfortable
that
it's
secure
right
that
that's
the
really
expensive
part
right
and
I
I
I
think
that
there
are
things
that
the
core
unit
can
add
to
its
roadmap
to
limit
some
of
these
costs
that
we're
spending
on
oracle's
right
like
we
can,
for
example,
just
by
offboarding
some
of
the
unprofitable
vaults
right.
G
Not
only
you
know
like
we
that's
gas
that
we
wouldn't
be
spending
on
on
those
vaults
right,
and
so,
if,
if
governance
does
that,
for
example,
then
you
know
we
might
not
even
spend
this
estimated
budget
right,
we
would
be
spending
less
right.
I
said
you
know
when
I
was
talking
about
ethan
and
wi-fi
right,
that
there
are
tuning
parameters
we
can
make
on
the
oracles.
That
makes
them
cost
a
little
bit
less.
G
You
know
not
like
50
percent
less,
but
you
know
maybe,
like
you
know,
20
25
percent
less,
and
that
has
to
be
a
discussion
that
has
to
be
had
with
risk
right
where,
if
you
are
decreasing
the
sensitivity
of
the
oracle
right,
how
much
more
risk
does
that
put
on
the
maker
protocol
in
terms
of
holding
assets
longer
than
it
would
otherwise
like
right?
G
But
you
know
it
does
come
with
with,
with
the
the
decreased
kind
of
operation
cost
for
the
oracles
right.
Another
thing
you
know
is
well
you
know
what
happens
with
l2.
Well,
that's
still
kind
of
an
open
question.
G
We
suspect
that
l2
oracles
will
be
cheaper
than
layer,
one
oracles,
and
so
if
there
were
to
be
some
kind
of
maker,
I
you
know
either
the
full
protocol
or
some
kind
of
shell
of
a
protocol
on
layer.
Two
right,
then
maybe
l2
could
support
collateral
types
that
l1
can't
support
because
of
because
of
the
oracle
costs.
G
Thing
that's
going
to
fix
all
of
our
problems,
and
I
I
want
to
caution
against
that.
For
for
like
a
couple
reasons,
one
is
you
know
all
l2
does.
Is
segregate
computational
costs
from
really
kind
of
storage?
So
right
you
don't
have
to
do
computation
on
l1,
which
is
expensive
right.
You
do
the
computation
on
l2,
but
all
of
the
call
data
also
all
of
the
data-
that's
basically
passed
to
the
smart
contract
on
layer
2
to
do
those
transactions.
G
All
of
that
call.
Data
still
needs
to
be
recorded
on
layer,
1
and
oracles
are
very,
very
heavy
on
call
data
at
least
the
medianizers
are
osm
and
spotter
are
not
not
so
much,
but
the
medianizers
are
very
heavy
on
call
data
right
because,
and
the
reason
is
because
you
need
to
get
a
consensus
of
a
bunch
of
feeds
right.
G
So
if
you
have
26
feeds,
you
need
at
least
13
to
reach
a
quorum
of
you
know
the
price
of
eath
is
this
right,
and
so
for
each
of
those
feeds
you
need
the
price,
the
timestamp
and
the
signature
right
saying
that
I've
verified.
You
know,
through
my
identity,
that
the
at
you
know
time
this
price
this.
This
is
I'm
corroborating
this
and
you
need
that
for
each
of
the
13
feeds
right.
So
there's
a
lot
of
data
that
gets
passed
to
these
functions,
and
so
we'll
we'll
have
to
do
some.
G
E
Oh
shoot,
I'm
going
to
ask
a
question:
do
we
have
a
number
on
a
per
collateral
like
what
it
costs?
I'm
trying
to
understand,
get
a
handle
around
like
amount
of
f
that
we
need
to
operate
the
oracles
per
collateral,
and
it
seems
like
that
that
that
should
be
roughly
linear.
I
know
there's
probably
some
it's
like
a
year
and
a
half
ago,
everyone
was
hot
to
add
lots
of
collateral
types,
and
I
was
like
be
careful.
E
G
Yeah,
no,
I
I
agree.
I
I
also
think
gas
costs
used
to
be
a
lot
lower
right
and
so
things
that
made
sense
back
then,
from
an
economics
point
of
view,
now
make
less
sense
or
or
don't
make
any
sense
right.
G
G
I
don't
know
about
drama
or
whatever
right
with
with
thematic
token.
Is
that,
like
there
was
an
agreement
that
they
would
mint
a
lot
more
die
and
we
onboarded
them
based
off
of
that
agreement
and
then,
while
they
did
meant
some
die,
they
didn't
mint
enough.
According
to
the
agreement
that
we
had
so
right
now,
they're,
given
how
much
they've
borrowed
right
now,
it's
an
unprofitable
vault.
G
So
I
really
think
you
know
unless
they
can
be
convinced
to
uphold
the
original
agreement
and
mint
significantly
more
diet
than
they're
doing
right
now.
We
should
just
off-board
them
because
it's
unprofitable
for
us.
E
E
You
know
when
gray
was
on
gas
was
like
one
way
and
f
was
like
200.
We
had
a
completely
different
ecosystem
and
network
than
a
hundred
way
without
that
five
thousand
dollars
but
realized.
You
know.
Unless
something
radically
changes
again,
we
could
have
at
the
ten
thousand
dollars
and
be
at
200,
gray
and
heartbeat
right,
and
so
I
mean
I'm
kind
of
asking
just
a
general
question
which
was
do
we
know
how
much
f
it
takes
for
a
particular
collateral
type.
E
You
know
I'm
saying
just
one
of
them
so
that
we
can
we
can
target
and
try
and
get
to
levels.
You
know
I
I
was
looking
at
actually
the
matic
fault
and
it
looked
like
there
unless
I
was
mistaken,
it
looked
like
there
was
79
million
in
collateral
in
there
and
they
were
minting
about
nine
and
their
cap
was
ten,
and
so
I
think,
they're
kind
of
at
a
point
where
they
don't
feel
like
they
can
move
because
they
wanted
to
go
beyond
ten,
and
so
I
agree
with
you.
E
We
can
address
these
issues,
but
we
kind
of
also
need
to
think
about
a
future
that
has
even
a
higher
up
price
higher
gas
costs
and
work
this
against
what
vaults
were
actually
going
to
have,
so
that
we
can
maintain
profitability,
but
we
might
get
to
a
point
where
we
have
to
run
unprofitably
on
oracles
just
because
gas
prices
might
be
high
temporarily.
We
don't
want
to
just
close
vaults
because
they're
slightly
unprofitable
just
to
kill
that
kill
that
vault.
E
If
we
can
make
it
grow,
if
we
can't,
then
we
definitely
need
to
axe
them
and
there's
a
number
of
vaults.
We
kind
of
need
to
look
at
carefully
to
act.
So
I
just
wanted
to
know
what,
if
there's
a
number
on
f
for
like
one
collateral
type,
what
does
it
cost
us
an
f
per
collateral
type,
so
we
can
run
these
numbers
for
the
future.
That.
G
G
It's
not
like,
there's
a
it's,
a
hundred
f
to
run
this
oracle
and
if
the
price
of
eth
goes
up
or
down
it
doesn't
matter,
it's
it's
a
hundred
f
right
that
that's
not
how
gas
prices
work,
they're,
they're
highly
volatile,
so
you
typically,
the
trend
is
if
if
eath
goes
up,
right
and
and
the
causality
here
right
is
questionable,
but
there
is
a
relationship
between
eath
going
up
and
gas
prices
going
up,
or
maybe
it's
right,
a
gas
price.
G
G
Right,
like
black
thursday
right
where
eath
can
go
down
and
get
gas
prices,
go
inverse
right.
G
That
that
being
said
right,
your
other
argument
of
okay
yeah,
you
know
gas
prices
could
could
go
sky
high
right.
We
do
need
to
have
a
plan
in
place.
You
know
for
what
happens
if
you
know,
maybe
it's
okay
for
vaults
to
be
a
little
bit
unprofitable
for
some
period
of
time,
but
we
need
to
define
what
is
a
little
bit
unprofitable,
that
we
are
willing
to
to
eat
and
you
know,
and
what
is
a
small
period
of
time
right.
G
So
when
we
violate
those
thresholds
that
there
is
a
clear
course
of
action
and
we
don't
spend-
I
don't
know
two
three
months
of
you-
know
bureaucracy
trying
to
decide
what
to
do
right,
that
there's
already
a
plan
in
place
and
we
just
execute
on
it.
E
Yeah
I
agree
I
was
just
trying
to
I
mean
I
guess
I
had
to
learn
more
about
medianizer,
because
I'm
just
trying
to
get
a
handle
on
cost
per
collateral
type
right
within
medianizer.
If
we
do
n
what
is
the
cost
for
n
plus
one
I'm
just
trying
to
get
a
handle
on,
I
know
there's
a
total,
a
total
cost
for
the
whole
thing,
that's
probably
fixed
to
some
element,
but
then
there's
a
cost
per
collateral
type
and
I'm
trying
to
understand
what
that
is.
G
So
those
are
the
numbers
that
I
presented
in
the
mip
and
I'll
be
publishing
this.
So
you
can
you
can
take
a
look
at
those
numbers
more
closely,
but
they're
they're
they're,
you
know
they're
just
what
we've
seen
in
the
past
they're,
not
necessarily
indicative
of
what
we
see
in
the
future
right.
So
you
have
to
look
at
this
number
and
be
like
okay,
330k,
let's,
let's
say:
okay,
how?
How
does
that
scale?
G
Okay,
let's
assume
really
really
bad
scenario:
gas
prices
triple
okay,
it's
a
million
right
instead
of
just
fixating
on
the
330k
number,
and
you
know
that
that's
what
it
costs
us
to
onboard
a
collateral
type.
That's
what
it
costs
us
today.
It's
not
what
it
might
cost
us
a
month
from
now
three
months
from
now
a
year
from
now.
B
Yep,
the
reason
we
can't
do
this
make
amend
is
because
the
cost
is
denominated
in
units
of
gas
right
and
the
price
of
gas
changes.
There's
great
change
and
the
price
of
that
changes
based
on
eth.
So
it's
like
we
can't
like
we
could,
I
guess,
figure
out
the
total
gas
units
each
things
costs
per
month
right
and
that
would
be
a
fixed
amount
ish.
E
E
Okay
yeah,
so
I
was
living
in
this
fantasy
world.
I
realized
where
I
thought
level
twos
were
you
know,
gonna
solve
our
problems,
but
it
it
sounds
like
do.
We
even
have
a
plan
for
for,
like
addressing
you
know
like
the
stock
exchanges.
Have
thousands
of
stocks
like
russell
3000
and
new
york
stock
has
changed
like
it
doesn't
sound
like
we
even
have
a
plan
for
supporting
that
kind
of
throughput.
G
Yeah
it
it
kind
of
remains
to
be
seen
right,
sharding,
which
is
kind
of
this
fairytale
unicorn
thing
that
people
talk
about
you
know
is
still
years
off
and
that
kind
of
helps
with
with
the
data
component
right
l2s
are
supposed
to
solve
the
computation
component
shards
solve
the
data
problem.
G
You
know,
is
there
some
things
that
maybe
we
can
do
well,
it's
it's
all
just
trade-offs.
It's
all
compromises
right
so
like
and
it's
just
what
compromises
are
is
is
maker
governance
willing
to
accept
in
other
areas
in
order
to
reduce
things
like
oracle
costs.
So,
for
example,
you
know
we
looked
at
the
medianizer
cost,
only
being
like
a
third
of
of
the
total
oracle
cost
right
of
of
this.
You
know
oracle
security,
module,
spotter
kind
of
combination.
G
If
governance
is
willing
to
go
an
oracle
attack
on
certain
vault
types
right,
you
know
that
are
maybe
smaller
in
debt
ceiling
size,
you
could
just
hook
up
a
medianizer
to
a
vault.
Is
that
a
good
trade-off
to
make?
I'm
don't
even
remotely
want
to
comment
on
that
right?
Everyone
is
going
to
be
very
opinionated
on
that,
but
there's
no
there's.
No
magic
bullet
right,
it's
just
it's
it's
just
kind
of
where,
where
you
want
to
pile
the
risk
of.
A
Awesome-
I
am
noting
the
time
here.
You
know
we
had
a
few
other
questions
and
potential
discussion
topics
in
the
chat
we
might
have
to
postpone
those
to
a
later
meeting
or
perhaps
to
the
forum,
but
we'll
give
everyone
one
last
chance.
If
there
was
something
you're
trying
to
say
or
ask,
that
would
be
your
time
to
do
it.
A
Right
now
all
right.
Well,
thank
you,
everyone
for
coming
and
engaging
like
I
mentioned
earlier.
Let's,
please
keep
the
discussion
going
on
the
forum.
Now
there
are
a
few
topics
we
didn't
quite
get
to
today,
we'll
make
sure
to
include
them
in
a
prompt
governance
call
and
yeah
thanks
everyone
for
coming.