►
From YouTube: CasperLabs Community Call
Description
Engineering update, Bonding Auction design and Enterprise Adoption of blockchain
A
All
right,
good
morning,
good
afternoon,
good
evening,
everyone,
wherever
you're
coming
to
us
from
thank
you
for
listening
in
dialing
in
joining
tuning
in
after
the
fact
I'm
going
to
give
us
a
give.
All
of
you
an
update
on
the
project
where
we're
at
this
is
meta.
Parla
car,
I'm
CTO,
caspere
labs,
and
we
do
this
call
every
Tuesday
morning
our
community
call.
If
you
want
more
details,
you
can
join
us
on
telegram.
It
is
an
open
call.
A
We
do
welcome
people
to
join
the
zoom
I'm
joined
today
by
Alex
lemon
off
one
of
our
economists
and
then
are
a
couple
of
our
community
moderators
OG
and
huddle
shark
Joe
Davies.
So
before
we
dive
in
I'll
just
give
a
quick
agenda
I'm
going
to
provide
an
development
engineering
update
and
then
we're
going
to
talk
about
what
the
community
has
requested.
They
wanted
me
to
talk
about
enterprise.
Adoption.
Is
that
right,
I
think
that's
right,
I,
believe
that's!
The
question
was
around:
how
are
we
going
to
get
Enterprise
to
adopt
blockchain
technology?
A
What
my
thoughts
are
around
that?
How
I
think
that's
gonna
happen
and
then
I
thought
we
were
going
to
have
an
economic
update,
but
it
might
be
a
short
call
today.
So
if
you
have
questions,
you
can
definitely
drop
them
in
our
telegram
Channel
or
ask
them
on
YouTube
feel
free
to
ask
me
anything
I'm
more
than
happy
to
answer
questions.
A
A
What
we
have
been
doing
now
is
we
drop
the
bits
into
a
test
framework
and
we
run
a
long-running
test
about
seven
days,
just
in
our
sre
site
or
site
reliability
team-
and
this
is
like
the
pre-production
readiness
testing
that
we
do
to
validate,
that
no
new
regressions
were
introduced
or
the
code
base
and
everything
is
solid
and
it
basically
comprises
of
you,
know,
running
up
standing
up
a
network
and
AWS.
That's
just
some.
A
You
know
kind
of
geographically
distributed
and
throwing
a
bunch
of
a
bunch
of
load
at
it,
and
we
observe
the
network
behavior
and
try
to
destabilize
it
and
make
sure
that
it
remains
up
and
robust
and
stable.
So
that's
that's
what
we
do
the
we
did
restart
the
test
net
last
weekend,
the
test
net.
Finally
de
stabilized,
that
code
was
almost
two
months
old.
Actually
we
had
quite
a
few
bug
fixes
that
we
had
introduced.
A
You
know
B.
We
had
fixed
actually
that
we're
not
introduced
in
the
test
night,
because
we
just
didn't
want
to
bounce
it.
So
what
we're
gonna
do
is
we
bounced
it
and
we
have
all
those
bug
fixes
in
the
test
net
right
now,
when
0.20
is
ready
to
be
released
with
our
beta
test
net
milestone,
that's
going
to
happen,
probably
in
about
two
weeks
we're
going
to
be
opening
up
the
validator
set
to
50.
A
A
A
lot
of
these
test
networks
have
even
seen.
Sla
is
around.
You
will
not
attack
the
network
during
the
test
net
timeframe
and
that's
actually
not
what
we're
going
to
do.
We're
going
to
ask
our
validators
to
actively
attack
the
network.
So
the
hope
is
that
in
the
second
phase
and
the
third
phase
they
will
be
focusing
on
how
to
build.
A
You
know
malicious
nodes
right.
Our
intention
is
to
offer
a
full
specification
of
the
node
and
we're
going
to
invite
validators
to
try
to
attack
the
network,
because
we
want
this
thing
to
be
stable
and
robust
and
permissionless,
and
we
can't
do
that
unless
we
know
it's
held
up
against
some
attacks,
so
that
is
my
hope
and
prayer
all
right
moving
along.
So
we
bounce
the
network.
The
last
test
that
was
up
for
40
days
and
everything
is
looking
good.
A
A
Given
the
current
block
production
rate,
because
the
blocks
being
produced
in
phase
zero
of
eath
2.0
with
the
beacon
chain
will
still
be
using
proof
of
work,
I,
don't
believe
the
hash
rate
is
going
to
go
down,
because
you
can
still
do
a
deep
chain
reorganization
and
can
take
control
over
the
next.
You
know
of
contractions
and
the
next
block
being
produced.
If
you
reduce
the
hash
rate,
so
I
think
the
hash
rates
gonna
stay
the
same.
So
if
you
do
that
math.
A
That
means
the
finalization
time
on
eath
2.0
is
gonna,
be
approximately
31
minutes,
so
I
thought
not
I
thought
8
minutes
was
too
long.
I
was
really
grumpy
about
that.
But
when
I
hear
31
minutes
for
eath
2
I
know
I'm
like
well.
Maybe
it's
not
so
bad,
but
I
think.
Definitely
when
we
do
maenette,
it's
gonna
be
lower
than
8
minutes.
A
Eight
minutes
is
simply
too
long,
so
I'm
hoping
that
we
can
get
to
a
place
where
we
can
have
round
lane
so
about
two
minutes
and
it's
going
to
be
adjustable
so
the
important
thing
to
notice.
If
the
protocol
has
processed
all
the
messages
it
expects
to
receive
and
around
it
will
speed
up
if
it
takes
longer
to
receive
all
the
messages
in
the
round,
it
will
slow
down.
So
the
good
news
about
highway-
and
this
is
being
implemented
on
the
rough
side
right
now
is
it
will
auto,
adjust
right.
A
The
nodes
will
actually
respond
and
with
finality,
depending
on
how
what
the
network
conditions
are
like.
So
in
terms
of
message
propagations,
so
if
you
get
enough
votes
in
a
given
round,
the
protocol
will
continue
to
speed
up
until
it
takes
longer
to
process,
receive
and
process
those
messages,
in
which
case
it
will
slow
down
so
good
good
progress
being
made
there.
A
Yes,
so
the
fixes
there's,
if
there's
maybe
one
more
bug,
fix
that's
going
in
the
dot,
oh
release
for
test
net,
but
otherwise
19
have
been
fixed
and
they
are
all
18
F
and
fixed
and
they're.
All
in
this
patch
that
are
currently
is
currently
running,
except
for
the
one
fix
on
highway.
We
are
implementing
the
consensus
component,
so
we're
implementing
the
basically
the
we've
implemented,
the
fourth
choice,
rule
we've
implemented
rounds,
we're
working
on
synchronization
and
the
arrow
supervisor,
and
now
we're
working
on
I
think
believe.
A
It
I
believe
it's
leader
selection,
so
the
random
verifiable
bits
that
allow
for
the
leader
to
be
selected
we're
also
implementing
a
test
harness
for
consensus.
So
there's
going
to
be
a
good,
solid
unit
test
and
we're
implementing
the
chain
spec.
So
the
chain
SPECT,
is
basically
the
manifesto
mole
and
they
count
CSV.
We
will
implement
something
similar,
something
just
as
decentralize
and
easy
to
use
such
that
anyone
can
just
download
the
chain
spec
and
join
the
network.
That's
the
intention.
A
Contract
runtime,
we
are
feature
complete
with
contract
headers.
You
can
go
onto
YouTube
and
check
out
the
contract.
Header
demo
that
was
given
by
ma
che.
He
did
a
zoom
call
where
he
was
doing
our
weekly
workshops.
If
you're
wanting
to
get
some,
you
know
hands-on
training
on
how
to
use
blockchain
and
how
to
build
smart
contracts
on
Casper
labs.
Those
weekly
workshops
are
open
and
available
for
anyone
to
join
joke.
Will
you
drop
the
calendar
location?
So
people
can
see
what
the
timings
are
into
telegram.
So
folks
know
it
would
be
very
helpful.
A
You
could
also
go
to
our
YouTube
and
in
the
playlist.
If
you
look
at
any
of
those
videos,
the
calendars
also
presented
there
in
the
description.
So
if
you
go
to
the
weekly
workshops,
playlist
you'll
also
see
how
you
can
join
the
weekly
workshops
and
get
hands-on
support
on
how
to
smart
and
do
you
know,
learn
how
to
build
smart
contracts.
A
lot
of
video
content
already
up
there,
but
if
you
want,
if
you
have
a
specific
question,
don't
hesitate
to
join
and
Macchi
can
walk
you
through
or
if
you're
trying
to
build
something
specific.
A
A
And
this
is
on
a
variety
of
network
sizes,
so
we're
gonna
try
throughput
on
five
ten
fifteen
and
twenty
node
Network
we're
gonna,
try
to
suss
out
what
the
consensus
overhead
is
on
these
networks,
so
really
trying
to
understand
what
is
the
consensus
overhead
for
highway
on
the
ecosystem?
Really
exciting
news
here?
So
with
the
dot
twenty
release?
For
those
of
you
that
don't
know,
we
are
adding
support
for
aetherium
keys,
I'm
happy
to
report
that
somewhere
close
to
the
dot
20
release
around
the
beta
test
at
timeframe.
A
We're
also
going
to
be
delivering
a
tool
chain
that
will
support
solidity
contracts.
So
you
will
be
able
to
transpile
your
solidity
into
web
assembly
and
we're
going
to
build
a
full
two
tool
chain.
That
makes
it
very
very
easy.
So
basically,
solidity
developers
will
need
to
will
be
able
to
transpile
their
salinity
contracts
into
web
assembly
and
I
believe
they'll
be
able
to
run
them
through
a
test.
A
Harness
I
need
to
validate
this
with
Macchi,
but
if
they're
going
to
do
what
I
expect
they're
going
to
do
in
terms
of
what
they're
building
they're
gonna
build,
something
that
will
just
take
it
straight
through
to
the
test
harness.
Our
intention
is
to
leverage
our
smart
contract
DSL
and
transpile.
The
AST
that's
generated
from
sohliang,
which
is
a
solidity
to
solidity
to
well
Wazza.
A
More
webassembly
compiler
will
take
the
AST
that
soling
generates
entrance
and
basically
transform
that
into
our
rust
DSL
and
are
basically
a
truss
contract
DSL
and
then
compile
it,
and
this
should
enable
I
believe
if
you
wanted
to
write
a
test
harness
for
your
contracts
to
know
that
they
ran
correctly.
You
can
do
that
using
the
rust,
basically,
the
rust
tooling.
This
is
super
exciting.
This
will
enable
solidity
developers
to
you
know
kick
the
tires
on
wasum
and
understand
what
the
solidity
to
was.
Experience
looks
like
we're.
A
Also
improving
the
Kasper
lab
signer
for
those
of
you
not
familiar.
We
do
have
this
nifty
little
Casper,
signer
I,
wonder
if
it's
going
to
work,
I
think
it
only
there.
It
is
so.
This
is
the
Casper
signer.
It
allows
you
to
basically
add
your
keys
to
a
vault
locally
that
is
encrypted
and
saved
locally
on
your
machine,
we're
going
to
basically
enable
the
Casper
signer
to
look
at
files.
Instead
of
the
actual
keys,
we
discovered
a
big
UX
problem,
so
that's
the
that's.
A
A
On
the
economics
front,
we
are
building
our
validator
bonding
auctions,
we're
doing
the
technical
specifications
for
that,
and
we
are
writing
the
formal
specs
for
the
gas
and
metering
model,
including
storage
and
bandwidth.
For
those
of
you
that
don't
know
for
storage,
we
have
talked
about
just
holding
token
and
or
potentially
burning
fees
for
token
for
storage
right,
because
storage
fle
doesn't
go
to
a
single
net,
a
single
node.
It
goes
to
all
the
nodes
in
the
network
when
you
store
stuff
in
the
global
state.
A
Okay,
let's
go
I'm
gonna
dive
directly
into
what
the
community
asks.
The
community
has
to
question
about.
What
are
we
going
to
do
to
get
enterprise
adoption
of
blockchain?
And
you
know
this
is
a
really
interesting
question.
I
think
a
lot
of
people
have
different
opinions
about
this.
Some
people
believe
that
they
should
build
full-blown
enterprise
type
applications
that
use
the
blockchain
or
specific
blockchain
systems
and
I
actually
think
that
enterprise
is
not
going
to
adopt
blockchain
directly.
A
If
you're
talking
about
big
fortune,
1000
companies
I
think
the
way
they
will
adopt
blockchain
is
going
to
be
more
along
the
lines
of
middleware
right.
So
for
those
of
you
that
know
a
little
bit
about
me
for
those
of
you,
don't
I
worked
for
a
Valera,
a
Valera
does
sales
tax
calculation
and
a
Valeris
strategy
to
go
to
market
has
been
quite
brilliant.
A
They
actually
built
in
they
built
a
software-as-a-service
core
architecture,
that
is
a
high
scale
tax
calculation
engine,
and
so
you
basically
make
an
API
call
and
you
pass
that
API
some
information
and
you
get
basically
a
tax
rate
back.
So
if
I
want
to
know
the
tax
rate
for
a
widget
in
the
state
of
Arkansas
I
pass
in
that
information,
what
you
know
what
the
information
that
it's
a
widget
and
I
pass
in,
where
and
Arkansas
I'm
transacting
and
I
na
the
system
already
knows
where
I
do.
A
Business
and
I
passed
those
those
data
points
in,
and
it
gets
me
back,
the
tax
rate,
the
tax
rate.
If
it's
a
widget,
it's
gonna
be
what
we
call
tangible
personal
property
and
the
tax
rate
for
basic
sales
tax
rate
in
Arkansas.
Let's
just
say
it's
7%,
and
so
that
widget
is
gonna,
be
taxed
at
a
rate
of
7%.
Well,
everybody
knows
that
enterprise
uses
software
called
enterprise,
resource
planning
systems
and
si
P
Great
Plains
Dynamics
NAV
dynamics.
You
know
ax.
These
are
just
some
of
the
names
of
big.
A
What
we
call
enterprise,
ERP
system,
enterprise,
resource
planning
systems
or
we
call
accounting.
Software
QuickBooks
is
another
example
of
a
county
software
and
for
those
of
you
that
don't
know,
there's
probably
over
a
thousand
different
accounting
packages
out
there
doing
an
ERP
implementation
in
your
business
is
not
a
small
thing.
It
is
a
really
really
big
deal
right.
So
folks
that
want
to
build.
You
know:
blockchain
backs
supply
chain
systems,
I
kind
of
chuckled
to
myself,
because
you're
actually
expecting
a
company
to
upgrade
their
ERP.
A
The
upgrade
cycle
for
ERP
systems
is
about
10
years
enterprise's,
don't
upgrade.
Their
system
is
very
often
in
fact,
they've
rarely
upgrade
their
systems,
because
it's
a
massive
massive
disruption
doing
that
kind
of
an
upgrade
literally
takes
12
to
18
months
and
call
it
good,
bad
or
ugly
or
indifferent.
A
They
aren't
willing
to
take
that
kind
of
disruption
on
very
often
because
it
even
bleeds
out
to
their
end
user
experience
right,
because
they've
got
integrations
into
their
end
user
experience
into
their
workflows
in
terms
of
how
invoices
are
sent
and
received
and
paid-
and
you
know
automatic
payment
by
ACH
and
whatnot.
So
it's
it's
a
non-trivial
thing
and
it
affects
cash
flow
and
it
affects
revenue.
So
it's
specifically
talking
about
ERP,
but
it
permeates
their
entire
business.
A
So
whatever
Lera
did
is
they
did
these
little
tiny
integrations
into
each
of
these
accounting
packages
and
then
you
basically
download
and
but
they
call
an
enhancement
or
a
widget
or
an
add-on
to
your
existing
accounting
packages.
That
gives
you
the
a
Valera
sales,
tax,
module
and
voila.
Your
existing
package
is
all
ready
to
use
a
bolero
sales
tax,
and
this
is
how
I
believe
that
enterprise
will
ultimately
adopt
blockchain.
A
The
Casper
allows
public
blockchain,
a
hybrid
private
blockchain
that
sends
hashes
to
the
public,
blockchain
or
potentially
even
a
full
private
implementation.
If
they
really
want
to
do
that,
but
we're
going
to
help
them
implement
this
using
our
technology,
because
what
we
built
is
very,
very
easy
for
them
to
use
right.
They
can
upgrade
their
contracts,
they
can
test
their
contracts.
They
confess
it
basically
test
their
on
chain
code
as
if
it
was
off
chain
code
and
provide
these
proofs.
A
These
guarantees
to
their
customers
that
whatever
they're
looking
at,
is
backed
by
blockchain
technology
as
a
trust
layer.
So
that's
what
we're
going
to
go.
Build
and
I
have
seen
this
business
model
work.
It's
worked
phenomenally
well
for
a
valera
alva
Lair
is
now
a
multi
billion
dollar
publicly-traded
company,
and
I
believe
this
is
the.
This
is
the
least
path
of
resistance
for
enterprise
to
adopt
blockchain.
So
that's
my
take
on
it.
If
you
disagree
with
me,
love
to
hear
from
you
on
telegram,
this
is
again
a
hypothesis.
A
This
is
the
future
that
I
see
happening
with
blockchain,
but
maybe
it's
more
radical
than
that.
I
could
be
wrong,
but
when
it
comes
to
enterprise,
they're
definitely
slow
to
adopt
and
slow
to
move
right.
They
don't
do
anything
quickly.
They
definitely
wait
and
they're
not
going
to
want
to
disrupt
their
business.
So
that's
my
my
pasta
pause
it
on
that
Alex.
Do
you
have
questions
I
mean
Alex?
Do
you
have
a
presentation
actually
not
questions?
We
have
a
presentation.
Yes,.
B
B
B
A
B
All
right
so,
basically
as
medicine,
we
are
finalizing
well,
we
really
finalized,
essentially
the
design
for
our
auctions
and
we're
in
the
process
of
working
out
a
technical
spec
to
make
sure
that
we
can,
you
know,
implement
this
effectively
and
in
the
ways
that
we
can
upgrade
later.
So
what
I'm
going
to
present
today
is
you
know,
sort
of
a
chain
of
reasoning
that
led
me
to
you
know
essentially
writing
down
the
designs
that
we
have
now
and
I
will
explain.
Why
is
it
choices
that
were
made
were
made
all
right,
so
we
will
start.
B
You
know
somewhat.
You
know
far
away
from
the
blockchain
and
first
discussion
is
a
very
basic.
You
know:
models
of
auctions
that
exist
in
economics,
so
there
are
basically
four
classical
formats
that
have
been
historically
studied.
You
know
all
of
these
also
exist
in
you
know
in
real
life,
so
the
first
ones
on
one
of
these
is
English
were
sent
in
bid
auction,
which
is
basically
you
know
what
you
think
of
as
an
auction.
B
Probably
there
is
a
Dutch
auction,
which
is
a
descendant
bit
version
which
you
can
think
of
us
having
a
big
cloak
like
Fink
on
the
wall,
where's
our
moves,
and
once
it
goes
low
enough,
somebody
raises
their
hands
and
their
hand,
and
you
know,
receives
the
goods
and
that
there's
a
usual
first
price
and
the
second
price
sealed
bid
auctions.
Alright.
So
what
we'll
be
discussing?
There's
a
basic
results
that
concern
these
four
formats,
the
applicability
of
the
theory
to
our
particular
case.
B
You
know,
will
ability
to
empirical
studies
on
you
know
all
these
matters
and
the
design
choices
that
were
made
as
a
result
of
you
know,
studying
this,
so
the
basic
models
for
all
four
of
these
auctions-
full
you
know,
have
the
following
assumptions:
right
is
it
bidders?
Are
risk
neutral,
so,
basically,
agents
do
not
differentiate
between
expected
them
going
T
to
payoffs
right
Z.
They
don't
care
about.
You
know
the
potential
variants
or
anything
like
that.
B
The
values
for
the
objects
are
drawn
from
independent
identically
are
part
of
me:
are
iid,
so
they're
independent
and
identically
distributed,
and
this
is
commonly
known
among
all
participants,
and
this
is
interpretable
as
bidders
be
draw
for
some
large
population
of
north
characteristics
right,
so
bidders
are
also
symmetric,
so
you
know
again
like
it's.
You
know
identically.
This
is
distributed,
the
values
plus,
you
know
same
strategy
sets
and
finally,
the
four
first
options:
the
payment
is
a
function
of
bids
alone
right,
so,
of
course
you
know.
The
most
obvious
example
is
the
first
price
auction.
B
You
know
we,
you
know
what
you
pay
is
woody
bit
right
in
the
second
price
variant.
Would
you
pay
is
what
the
second
guy
did
now,
what
these
basic
models
do?
Is
they
proceed
to
derive?
You
know
some
Bayesian
Nash
or
you
know,
dominant
strategy
equilibria
over
the
specified
games
and
there's
interesting
part
here.
Is
it
turns
out
that,
for
a
risking
for
a
principal
or
seller,
the
choice
between
the
four
formats
is
meaningless
as
they're
all
the
same,
and
so
this
is
relatively
easy
to
see
in
some
of
these
cases
right.
B
So,
first
of
all,
you'll
observe
that
there's
a
Dutch
auction
is
a
you
know,
for
revenue.
Purpose
is
the
same
as
the
first
price
sealed
bid
auction,
because
winner
spacer
bids
and
same
information
is
available
to
the
winner,
in
both
cases
at
the
moment,
when
Z
right,
because
either
you're
the
first
to
raise
your
hand
or
you
know,
you're
someone
who's,
a
bid
is
actually
the
Venant
bid,
but
either
way
you
don't
know
anything
about
anybody
else.
Hey.
A
A
B
Towards
the
classical
formats
as
assumptions
and
now
I'm
on
this
one
right
is
equivalence
between
all
these
four
right.
So,
as
I
said,
there's
the
Daksha
auction,
so
the
descendant
bid
auction
is
actually
basically
the
same
as
the
first
price.
Sealed
bid
auction
enters
ours
at
the
English
auction
or
is
ascending
bid
auction
the
same
as
the
second
price,
sealed
bid,
auction
and
well
more
or
less,
there's
also
a
fairly
easy
to
see.
Because,
basically
you
know
it's
everybody
bids
until
the
last
guy,
a
just.
B
B
You
know
from
the
sellers
perspective
as
far
as
revenue
goes
because
of
something
that's
known
as
the
revenue
equivalent
sphere,
and
this
is
the
general
result
from
mechanism
design
that
basically
says
that,
if
you
know
two
mechanisms
gives
the
same,
you
know
allocation
rules
that
they
must
have.
You
know
the
sync
transfers,
so
in
this
case
the
transfers
are,
is
a
you
know:
payments
for
goods.
You
know,
sold
its
auction
so
well.
What
does
this
mean?
You
know.
B
B
How
do
we
choose
them
right
so?
Well,
it
turns
out
that
there
are
some
differences
that
do
remain
even
in
the
you
know
in
this
way,
basic
second,
which
is
that,
for
example,
like
the
first
price
auction,
even
in
theory,
has
higher
variance
over
every
new
compared
to
the
second
price
auction,
although
of
course
the
extractor
is
identical
and
it
also
turns
out.
B
The
auctions
are
no
longer
equivalent
when
you
introduce
a
what's
called
the
common
value
component
right
and
we
will
discuss
a
slightly
later
but
moves
it
so
far,
we've
only
been
treating
what
is
known
as
private
values
right,
so
we're,
basically
how
much
ever
I
value
an
object
has
no
bearing
on
how
much
anybody
else.
Well,
he
said
you
know
the
object.
So,
additionally,
is
there
empirical
studies
of
searches
that
there
may
be
significant
differences
that
are
not
captured
with
the
theory
and
there
has
been
an
explosion
of
these
papers
since
the
late
90s.
B
But
it's
hard
to
really
come
away
with.
You
know
the
classification
scheme
from
disease,
because
a
lot
of
the
results
are
quite
contradictory
and,
of
course
you
know
it's.
It
requires
very
careful,
reading
and
parse
and
to
really
get
much
out
of
it
right,
so
I
will
basically
disregard
that
is
empirical
studies
mostly
for
now
and
focus
on
something
else,
and
so
now
so,
as
I
said,
believe
the
Pirkle
studies
alone.
But
look
it's
a
really
full
question
right.
B
So,
first
of
all,
so
how
closely
does
our
set
and
actually
matches
a
simple
theory
right
so
and
there
you
know,
we
knows
that
the
answer
is
very
poorly
right,
because
first
of
all
was
any
conceivable
designs
that
you
would
come
up
with.
We
have
situation
where
payoffs
basically
scale
with
your
bit
right
because
it
entitles
you
to
you
know
more
and
more
roses.
You
know
fixed
selects
income
screen.
B
We
have
these
very
strong
competition
effects
because
again
they're
trying
to
essentially
split,
you
know
finite
size
pie
and,
as
you
know,
this
common
value
components
that
I
mentioned
previously
and
I
think
Ruby
dress
was
next
slide
now.
The
other
thing
to
notice
here
is
that
our
see
shaker
who
does
not
match
you,
know
the
well-known
models
of
search
ad
options.
These
things
are
very
well
studied.
B
There
is
a
lot
of
theory
around
it,
a
lot
of
empirical
research,
but
those
options
are
actually
quite
different
because
in
our
case
there
is
no
random
action
right,
everybody
just
centers
into
the
same
auctions.
You
know
who
everybody
else.
Isn't
that
auction
more
or
less
right
and
again
becomes
a
scaling
of
you
know,
pay
off
Swiss
bids
which
doesn't
exist,
and
you
know
search
at
auction
because
there
it's
yeah.
Well,
you
know
it's
skilled
in
the
sense
that
the
higher
up
on
the
page,
you
get
the
better.
B
You
know
the
better
of
your
position,
but
you
know
it's
a
that
happens
with
discrete
jumps
and
once
you're
on
top
you're
on
top
increasing
your
bid,
that's
doing
it
so
we
essentially
have
a
you
know
a
case
which
is
not
covered
by
most
a
few
rewards
that
they
know
right.
So
what
do
we
do
about
that?
Well,
we
develop
a
modular
interim
designs
if
you
could
possibly
upgrade
later
right
and
there's
a
question
is
how
do
we
choose
this
design?
B
Well,
you
look
at
certain
results
regarding
common
value
auctions,
all
right,
so,
let's
dive
into
that
set.
So
in
this
set
and
bidders
receive
some
private
signals
right,
which
you
can
just
lead
by
so
observation
so
far,
you
know
certain
random
variables
that
were
inaccessible
to
other
bidders,
and
you
can
just
think
of
these
as
being
their
values
from
the
previous
setting.
Right,
as
these
signals
are,
you
know,
usually
correlated
with
some
hidden
common
value.
You
know
you
can
think
of
that.
B
This
means
true
values
that
you
know
everybody's
trying
to
find
out,
and
this
reflect
the
full
in
real
world
settings.
The
first
one
is,
of
course,
the
options
of
financial
instruments,
or
at
least
you
know,
liquid
financial
instruments,
okras
of
antiques
that
have
resale
value
right.
So
paintings
sculptures,
you
know
manuscripts
and
is
that
nature
and
well,
you
know
soon
an
office
is
going
to
be
used
to
sell.
You
know
cedar
extremes
in
their
proof
of
stake.
B
Worksheets
so,
and
turns
out
that
introducing
this
common
value
component
breaks
the
equivalence
between
the
auctions
and,
in
particular,
some
of
these
options
become
better
than
others
at
dealing
with.
What's
called
the
winners
curse,
so
winners
curse,
essentially,
is
a
regret
from
villain,
because
one
over
bit-
and
this
you
know-
can
happen
because
our
at
least
the
second
happen
visitor
of
naive
bidders,
because
the
guy
who
Vince
is
a
guy
who
receives
the
highest
private
signal
so
his
way
his
privates,
basically
an
outlier,
so
he
can
easily
know.
B
Basically,
he
receives
a
good
and
then
up
on
the
situation.
Where
turns
out
that
you
know
he
breed
Moors,
and
you
know
the
actual
true
value
of
his
object.
So
now,
of
course,
these
models
don't
deal
with
naive
bidders
and
you
know
strategically
sophisticated
agents
and
so
strategically.
This
is
basically
reflected
in
the
what
are
called
shaded
bits
right
so
bids
that
are
reduced
relative
to
what
one
would
expect
the
Versa
bidders
to
have
better
information,
and
it
turns
out
that
open
bidding
actually
helps
here.
B
And
out
so
well,
this
is
actually
going
to
inform
one
half.
Let's
say
like
a
one:
half
of
our
design
rates,
cumin
valley,
business
right
so
in
litter,
so
we're
launching
something
that
is
actually
function
is
soon
enough.
We
were
basically
making
choices
on
the
two
dimensions
to
reflect
it
in.
You
know
these
classical
auction
formats,
so,
first
of
all,
on
the
issue
of
first
price
versus
the
second
price,
we
choose
in
fact,
first
price.
B
So
we
know
is
that
you
know
at
least
anecdotally,
really
from
the
experience
of
Yahoo
back
in
the
early
2000s,
that
first
price
auctions
are
lower
volatility
and,
of
course,
as
we
discussed
previously.
This
is
also
the
case
in
theory,
in
fact,
but
this
would
suggest
that
you
know
one
who
prefers
the
second
price
version,
but
unfortunately,
in
our
case
second
price,
you
know
a
second
price
auction
is
revered
right
because
we
have
bidders
who
actually
in
some
sense,
prefer
to
pay
more
right,
because
spray
here
means
to
look
up.
B
You
know
more
tokens
and
therefore
come
eligible
for
more
Siegert.
Moreover,
it
creates
a
major
aux
problem
with
delegation
right
because
well,
if
it's
a
second
price
right,
what
happens
to
the
component
of
your
bid?
That
is,
you
know,
above
as
the
second
guy
right.
Where
do
those
tokens
actually
go
like
do
the
delegator
player
tokens
get
priority?
Do
your
tokens
get
priority
two
for
going
into
the
bond
or
not
it's
really.
It's
I.
B
If
I
just
discussed,
writing
basic
common
value
component
acts
as
a
tiebreaker
among
all
these
different
formats
that
otherwise
look
equivalent
right
so
and
the
basically
results
with
common
value
sure
suggest
that
you
know
choosing.
This
will
yield
the
greater
stake
and,
by
11
private
information
to
be
revealed
to
reducing
the
risk
for
participants.
It
also
avoids
technical
complications,
actually
keep
ins
a
bit
sealed
on
the
blockchain
until
it's
time
to
throw
rods,
auction
and
so
and
well.
What
do
we
intend
to
do
after
this
interim
design?
Launch
yes
well.
B
First
of
all,
we
will
prepare
for
future
prove
butts
by
designing
something
modular
right.
It's
gonna
be
like
several
contracts
sitting
on
the
chain
which
we
will
be
able
to
upgrade
separately.
For
example,
to
you
know,
swap
out
another
allocation
mechanism.
We
will
create
a
system
for
monitor
in
the
state
of
these
auctions
before
data
collection
and
to
realize
the
theoretical
benefits
of
open
bidding
right
here.
B
Actually
what
people
to
be
able
to
easily
look
at
the
bids
and
incorporate
that
information
to
zero
decision-making,
and
we
would
probably
absolutely
shoot
the
well
vertical
model
of
motion
design
here-
is
that
this
fit
for
our
particular
case
right.
So
what
the
perfect
there's
a
competition
for
a
fixed
number
of
kind
of
variable
size,
shares
of
future
income
streams.
B
You
know
the
common
value
aspect
and
the
additionally
use
and
use
this
to
do
comparative
statics
on
you
know,
comparative
statics
is
a
study
or
movement
of
outcome
variables
as
you
move
input
parameters
right
so
that
that's
something
that
economists
do
quite
a
lot
right,
so
you'll
do
compare
them
and
lumber
a
comparative
statics.
You
know
with
respect
to
stake
in
the
mouse.
You
know
with
numbers
of
participants,
quality
of
public
and
private
information
and
so
on,
but
that's
basically
it
for
today.
A
That
sounds
great,
but
that's
very
informative
on
how
actions
work
I'd
like
to
see
have
you
you
guys,
have
run
some
models
on
this
right
to
see
how
we
think
it'd
be
interesting
to
have
a
few
scenarios,
maybe
in
one
of
a
following,
because
I
think
I'm
sure
validators
are
gonna.
Have
questions.
That's
like
okay.
So
how
do
we
see
this
thing
playing
out
right?
A
Censoring
the
bonding
auction
or
something
to
that
effect,
so
it
would
be
really
cool
I.
Think
people
would
be
interested
to
say
to
see
like
okay.
So
what
happens
if
there's
200
slots
and
there's
so
many
people
in
the
queue?
What
what
does
the
game
look
like
right?
Because
there
is
gonna,
be
a
little
bit
of
a
game
here?
That
would
be
really
interesting
for
them.
B
Yes,
I
mean
if
we
do
that
to
do
it,
I
mean
this
is
sort
of
a
part
of
the
theoretical
work
on
this
particular
case,
where
you
will
be
able
to
say:
okay
guys
these
are
actually
the
optimal
strategies.
This
is
what
this
looks
like
you
know
in
reality,
but
even
before
that,
of
course
we
can
sort
of
do.
You
know
get
a
tutorial
explanation
of
what
happens
when
you
do
different
things
in
that
auction
right
I
mean
so
the
people
actually
understand
how
it
operates.
A
Yeah
yeah
I
think
that
would
be
great
because
we've
gotten
a
lot
of
questions
from
prospective
validators
on
well.
How
does
this
work
and
what
is
the
it?
What
is
the
you
know,
implications
for
you
know
state
delegation
right.
What
are
the
implications
with
exchanges
like
we
know
that
exchanges
are
whales
on
these
kinds
of
systems
right?
A
How
much
of
an
implication
is
around
exchanges-
and
you
know
them
hogging
up
or
even
potentially
sibling
out
a
bunch
of
validator
slots,
so
the
whole
exchange
problem
with
proof
of
stake
networks.
It
isn't
something
that
we
can
necessarily
solve
in
protocol,
because
people
ultimately
have
to
take
custody
of
their
private
keys
like
that's.
That's
the
endgame
right
as
custody
of
private
keys
is
what's
really
Matt.
What
really
matters
so,
hopefully
we
can
make
that
as
simple
as
possible
right
in
terms
of
wall
@ux.
A
Is
it's
really
up
to
the
wallets
and
their
user
experience
to
give
people
enough
comfort
of
managing
their
own
keys
and
knowing
that
I'm
not
gonna,
lose
my
keys
right
and
lose
all
my
money.
As
a
result,
it's
very
different
keeping
a
credit
card
in
your
wallet
versus
keeping.
You
know.
Crypto
on
your
computer.
People
are
notoriously
bad
with
passwords
and
lazy.
A
B
A
Yep,
that's
right,
cool,
any
questions
coming
in
from
folks.
It's
just
the
six
of
us,
no
questions,
terrific,
alright,
so
yeah.
Let's
look
forward
to
that
release,
dot
twenty
and
join
our
weekly
workshops
and
engage
with
us
and
we'd
love
to
have
you
participate
in
the
beta
test
net
building
something
kicking
the
tires
a
little
bit
given
us
a
spinner
on
the
block,
and
let
us
know
how
we're
doing.
Thank
you
so
much.
Everyone
for
dialing
and
Cheers
have
a
great
day
Cheers
right,
bye,.