►
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Alright,
so
that's
what
how
have
used
scenery
by
production
or
testing
of
our
just
a
curiosity,
okay,
city's
number
I,
said
the
rest,
lyrics
kind
of
evaluate
this
in
kind
of
and
with
how
to
grow
it,
which
is
the
blue
and
focus
on
for
you.
Basically,
I
vs
been
called
signal
on
windows.
Multi-Tech
I've
been
there
for
about
four
years
during,
say
three
and
a
half
of
them
working
of
Cassandra
in
some
shape
or
form
so
I'm
trying
to
share
is
basically
the
lesson
we
learnt
any
kind
of
like
why
we
chose
in
ye.
A
We've
got
this
technology
to
solve.
Some
are
really
red
rose,
so
what
signal
actually
does
we're
in
need
at
tech
space?
We
help
me
easy
for
marketers
to
configure
septic
and
interact
their
audience
and
hopefully
the
same
way.
We
kind
of
pull
out
some
of
the
necessary
of
the
development
cycle
to
actually
get
to
these
market
technologies
implemented.
It
was
sent
six
winner
from
web-based
stuff
channel
kind
of
across
channels
of
mobile
and
other
places,
so.
A
So
we
kind
of
have
three
named
Cassandra
new
spaces
and
I've
really
spent
focus
on
one
of
those
topics
that
need
some
more
interesting
out
of
all
them
lower
need
realistic,
I
almost
went
over
day,
so
the
first
piece
is
this
kind
of
maximum
service,
and
is
it
kind
of
a
wall
opening
a
nap?
Actually,
it's.
Basically
it's
a
way
to
like
associated
identifiers
between
different
companies
or
between
different
services.
We
use
it
to
provide
with
volume,
divided
customer
view,
so
when
our
clients
interrupt
their
customers,
they
can
see
a
bigger
picture.
A
A
lot
of
severe
tourist
resident,
so
much
man
is
clearly
like
panel,
looks
like
how
many
people
do
I
have
definite
sub
slicing
dicing
data
activation
is
more
about.
I
want
to
take
action.
That's
now
in
at
the
dr.
I
would
like
you
know,
send
a
promotion
to
be
whatever
it
the
second,
these
cases
actually
Bible
sanli
strict
standards
I
want
to
store
some
sort
of
times
today
to
you
about
metrics
or
about
values
and
things.
So
we've
played
around
a
couple
different
back
in
season
sandra
tyrosine
sign
after
both
open
source
projects.
A
We
turn
around.
We
had
some
success
of
them,
also
some
games
with
a
little
blood
flow.
Those
basically
is
something
we
have
better
knowledge
and
we
are
actually
using
the
production
and
then
the
last
piece
of
the
New
York
case
is
kind
of
an
Akata
blog
of
offensive
coming
to
move
on
tracklist
happy
december.
Eight
universe
or
the
data
model
is
kind
of
an
awesome
of
that
too.
So,
really
focusing
our
first
few
spaces.
A
All
these
colors
are
a
little
balance
play,
but
basically
we
have
a
lot
of
sources
where
our
clients
can
send
data
from
source.
Here
our
systems
website
is
mobile,
s,
etc.
Those
always
channeled
into
our
system
and
really
worries
container
is
filled.
The
blue
box,
it's
all
around
our
matching
service
and
the
functionality
and
ethical
thing
is
he's
too
kind
of
funnel
data.
Out
of
those
time
of
interaction,
are
so
two
simple
little
background.
I'm
excited.
Are
we
working
with
the
Center
for
about
readers,
so
certain
in
September
car
tires
summer
of
two
thousand
eleven?
A
Basically,
we
evaluated
standard
before
we
chose
that
we
compared
to
couch
Armando
and
react
it's
pending
idea.
What
are
no
simple
solutions
were
out
there.
We've
got
a
kind
of
possible
data
model,
but
we
need
me
different
one.
They
represent
even
teenagers
and
a
few
other
concerns
about
the
consistency
of
it
so
ultimately
decide
after
Sandra,
and
we
finally
got
the
Russian
one
hour
release.
A
So
we
deploy
it
to
to
AWS
regions
with
a
mems,
wasn't
impressed
at
the
fun
bubbly
felt
it
because
we
actually
had
a
replicated
across
the
regions
when
our
very
first
theme
was
really
simple.
It's
just
one
day
to
table.
This
is
back
when
there
wasn't
cgl
was
all
birthdays,
so,
instead
of
one
column,
family
there's
another
office
is
not
your
first
time.
That
was
basically
it.
So
what
did
it
actually
look
like?
What
we
were
trying
to
model
is
effectively
a
sparse
matrix,
and
it
worked
really
well
account
as
a
single
sword
here.
A
So
you
had
a
real
key,
which
was
some
sort
of
internal
identifier
and
that
amount
to
you
know
one
or
more
columns
of
some
data
to
spark.
So
there
may
be
there
something
someone
owes
not
for
others,
and
this
work
really
well
for
our
first
trial
is.
This.
Is
what
which
production
so
from
there?
We
have
learned
some
stuff.
We
had
a
prior
to
meeting
spaces
for
how
we
want
here
half
of
the
data.
A
We
also
started
really
money
to
get
standard,
because
you
will
in
time
to
deal
upgrades
and
think
about
how
we
want
to
really
model
in
prayer
data,
so
part
of
that
was
we
were
looking
at
other
tools,
help
us
manage
our
scanner
insoles
one
of
those
is
cream
by
catholics.
They
basically
help
you
organize
your
security
rules
and
do
backups
in
some
other
stuff.
Since
we
run
all
our
services
me
w
a
sequence
during
space.
This
is
a
great
way
to
kind
of
manage
the
security
rules
for
us.
A
That's
basically,
no
one
muscle
isn't
one
of
the
other
things
we
learned
it
attracts
to
you
is
we
wanted
to
not
have
just
a
single
identifiers
curb
of
common
terminal
wanted
questions
you
really
want
to
like
kind
of
a
mac
or
a
set
of
multiple
things
there
and
there's
usually
a
commemorated
fibers
office
and
I'm
secondary
ones
or
other
pieces
of
metadata.
We
want
to
store,
or
actually
our
Tyson
I
want
us
to
sort
form
as
part
of
this.
A
So
we
connected
is
swinging
at
this
there's
anything
called
composite
types
which
probably
what
we
should
really
use.
We
didn't
have
the
time
to
fully
proved
out.
The
Berkeley
launches
are
kind
of
this
hack
solution
network
for
a
while,
because
the
14
at
least
we
could
need
to
see
space
and
kinda,
keep
evolving
artist
and
a
limited
as
we
added
our
first
pdx
tables
so
like
they
mentioned
earlier
to
a
standard,
isn't
relational,
you
can't
do
it
going
across
tables
and
there
is
a
secondary
index.
That's
not
religious
use
it.
So
we
basically
do.
A
Is
we
built
our
own
deaths
tools
to
maps
with
how
we
want
to
create
a
main
table?
So
for
each
thing
that
was
a
potential
identifier
in
the
main
table,
we
would
enter
a
rose
the
index
table,
so
we
could
find
out
what
our
eternal
life
he
wasn't
it
from
very
able
to
discover,
because
it
is
this
once
again,
this
kind
of
evolving
you
space
of
how
you
want
to
access
and
what
we
want
to
do.
A
This
was
just
the
next
phase
of
that
evolution,
so
the
next
thing
that
evokes
eyes
you
kind
of
get
one
is
hockey.
Stick
my
blushing
points
where
it
had
enough
to
be
certainly
more
about.
It
really
need
to
look
at
anything
is
up
and
wonder
should,
as
we
did,
that
we
could
test
what
we
were
changing
and
test
new
versions
compared
to
the
old
version.
So
we
built
a
little
driving
framework
by
a
piece
of
like
junior
plug-in,
or
we
guys
noticed
something
a
little
more
fancy
tools
between
my
greed
and
rights.
A
To
give
us
an
idea
of
what
you
look
like
with
a
comparable
amount
of
generating
traffic
compared
to
a
real
shock
that
we
can
generate
a
rough
road
vehicle
in
proportion
with
the
load
every
framework,
so
yeah
she's
compared
does
the
newest
one
dot
2x
line?
Is
that
equivalent
to
like
the
one
that
one
line
of
your
arm?
A
It's
part
of
this?
We
also
working
we
designer
schema.
We
made
some
adjustments
to
it
to
support
some
of
our
new
torrey
pines,
and
the
final
thing
is
here
at
the
upgrades
are
performance,
so
they
let
s
not
too
much
day
before
virtual
notes.
Every
time
you
go
to
the
rink,
you
would
have
to
kind
of
sleep,
allocate
its
open
space
on
your
own
and
kind
of
slide
around
what
data
you're
responsible
virtual
notes,
legend.
Not
to
do
that
previously.
It
was
very
easy
to
scale
arranged
by
a
double
in
size.
A
It
wasn't
easy
a
letter
about
EE
babies
because
you
split
each
token
ate
half
burton
those
kind
of
takes
care
a
lot
for
you.
So
there's
a
big
one
for
us,
because
you
have
a
scale
graphical
traffic
volume
not
based
on
just
don't
learn
of
our
servers
so
as
part
of
this
and
switch
to
the
12
line
bill
for
the
first
time,
starting
up,
basically
a
goopy
space,
and
so
when
you
print
the
keyspace,
you
can
set
up
how
you
want
to
replicate
it
where
you
want
to
so.
A
The
initial
try
to
do
it
with
is
one
of
two
copies
of
each
of
the
four
regions.
We're
going
to
be
in
these
copies
are
beautiful
evening
tool,
honey,
bread,
guys,
say:
I
want
a
copy
sweet
everywhere.
Before
my
same
is
known
or
I
want
to
at
least
in
my
region,
then
you
know
Ralph
cater
royalty
of
gossip,
so
it
gives
you
some
possibility,
but
we
said
we
knew
we
want
at
least
some
coffee
date
ever
work
at
one.
A
It's
for
redundancy
also
disabilities,
that
if
we
want
to
you
know
at
a
higher
level,
an
application
staff
take
a
region
offline
temporarily.
The
data
still
available
somewhere
else
for
me
to
ship
traffic
is
still
available
somewhere
else,
so
our
DNA
will
now
a
little
more
interesting
taco.
So
what
the
data
table
really
is
is
its
own.
Our
own
signal,
internal
identifier,
identifier,
is
basically
who,
this
day
to
sport,
a
map
of
the
IPS
or
the
identifiers
we
have
available
along
with
an
internal
one.
A
So
internal
was
just
for
us
lettuce
to
things
like
a
effective
replica,
create
time
modified,
try
cetera
it.
Also.
Let
us
store
other
information.
It
was
useful
to
our
system
via
recording
hints
or
one
of
making
this
popular
access,
basic
kind
of
seasons.
For
privacy
information
here,
if
someone's
opted
out
of
it,
we
could
sorting
and
displacement.
A
So
what
does
would
actually
look
like
folks
are
looking
at
Fred
summation,
so
the
other
thing
to
be
changed
to
is
our
hierarchy
went
from
just
our
own
internal
identifiers
to
our
own
internal
identifiers.
Clustered
with
the
identifier
forgiveness
is
for
this.
Let
us
still
have
data,
that's
basically
fabulous,
but
we
can.
Cluster
sewing
is
returned
by
c2l,
looks
more
like
a
relational
storm,
so
what
these
tables
would
actually
like?
A
We
did
the
work
this
week,
an
example,
so
we
might
have
a
silicic
same,
sparse,
matrix,
vector
field,
but
it's
now
in
the
sea
build
more
relational
format,
so
you
would
have
you
know
this.
One
single
idea
of
a
101
and
have
you
know,
has
many
bros
associated
with
this
as
your
knee,
our
clients
configuration
or
whatever
set
up,
and
then
you
have
my
knees,
reach
one
and
completely
data
breach
one,
and
if
there's
nothing
there,
that's
fine
too,
that's
an
old
as
well.
So
any
questions
so
far.
A
So
also
me
reading
index
table,
so
the
next
table
is
now
set
up
where
we
have
identifiers,
if
I'm
going
to
be
a
long
with
a
partition
and
then
there's
a
hiking
game
and
I
e
value.
Those
are.
This
is
Billy
the
same
rare
person,
XO
Lewis
informations,
for
what
the
name
of
their
empire
is
what
the
value
of
it
is
and
then
what
our
eternal
idea
is.
A
So
it's
the
primary
public
memory,
keys
identifier,
partition
and
in
the
hole
he's
also
flustered
by
the
mean
value
gave
you
so
how
we
actually
be
able
to
partition
is.
This
is
purely
something
to
start
the
application
level
we
just
compute,
it's
basically
a
hash
mod,
the
number
of
partitions
we
want
of
these
values.
Us
we've
actually
use
this
in
a
couple
other
places.
Now
to
you,
it's
a
very
effective
way
to
partition
of
data.
A
So
if
we
want
to
run
a
query,
some
clients
may
have
a
lot
more
needed
here
than
others,
the
very
large
ones.
The
queries
always
turn
out
because
you
can
wear
it
all
and
one
go
or
you
get
clever
and
credit
report
is
part
of
the
so
period
of
time.
This
allowed
us
to
worry
about
a
partition
of
time.
So
it's
kind
of
an
artificial
constraint
is
like
saying
women,
you
know,
and
but
this
is
a
little
easier
way
for
us
crude
by
ocean
applications.
A
So
another
so
we
move
to
virtual
knows:
there
is
theoretically
a
ways
to
go
from
a
ring
with
non-virtual
notes
to
performance,
I've,
never
driven
even
became
work.
It's
all
shuffled.
We
were
bulleted
sex
and
not
to
use
that
because
they've
been
working
it.
So
what
we
did
is
we
took
our
old
ring
left
where
it
was
and
made
a
new
ring
with
Albany
gonna
set
up
a
with
a
new
schema
in
virtual
nodes,
enable
like
side
by
side,
and
then
we
change
your
beta
access
service.
A
So
we
could
read
from
both
and
then
only
raised
to
anyone,
so
that
allows
start
migrating
over
organic
traffic
from
you'll
drink
to
the
new
rate.
And
then,
alongside
that,
we
had
migration.
Script
originally
ran
through
the
old
reinforce.
The
right
set
happen
to
the
new
ring,
so
overall
that
whole
migration
type
of
course
of
about
a
week,
or
so
he
didn't
push
how
fast
you
go
if
I
want
to
go
to
the
slow
scale
so
that
we
could
maintain
a
roughly
equivalent
performance
level.
A
We
we
did
have
some
extra
late
to
do
to
actually
to
rights.
Are
sexy
reason
would
have
to
happen,
but
overall
we're
going
to
keep
the
system
up
and
do
the
migration
kind
of
on
the
fly
out
how
to
make
what
we
do
with
it.
So
we
might
actually
going
to
create
similar
process
again
with
summer
changes
will
get
near
future
to
supply
those,
because
you
can
do
basic,
alter
table
peanuts
with
CEO
some
of
the
more
advanced
ones.
A
You
it's
almost
harder
to
try
to
alter
my
pastry
because
the
system
is
just
going
to
copy
and
rebuild
so.
This
is
a
pattern
that
we
made
port
so
really
2014
was
mostly
about
scaling
and
focusing
on
performance.
It's
also
my
personal
and
crevice
TV
as
well.
So
one
of
the
things
prior
to
the
cinema,
which
I
said
we
were
running
it
was
on
I,
never
say
what
we
were
going
under.
My
name
is
kind
of
crappy
MTX
letters
and
then
we
operate
at
least
72
X
largest.
A
They
were
finely
grated
this,
but
they
weren't
that
fast.
You
had
a
lot
of
spiking
it
early
and
see
it
was
kind
of
all
over
the
board.
What
we
found
out
through
both
seeing
some
reports
published
by
templates
and
then
our
own
bit
of
experimentation,
if
you
put
the
same
thing
on
our
me,
I
the
itune
instances,
which
is
basic,
SSD
bad
consistent,
said
it
made
a
huge
impact
in
the
car
lady
kind
of
it
really
leveled
it
up
to
the
point
starting
to
impact
I
see,
but
basically
the
blue
one
here
is
our
right.
A
A
We
basically
apply
the
SSDs.
There
were
the
first
workers
and
that
you
know
smooth
out
a
lot
of
work
are
kind
of
thinking
in
service,
so
this
is
for
weeks
and
smooth
that
a
lot
of
our
my
penis
in
the
latency
and
then
once
we
actually
did
the
upgrade
to
the
12
16,
which
was
the
latest
patch
at
the
time
and
had
the
other
fixes
for
human,
a
specific
patches,
something
to
do
with
the
lien.
Seeing
how
affected
it
really
kind
of
really
smooth
our
overall
deity
sword.
A
A
nice
slow,
steady
humph,
regardless
what
the
actual
volume
is
so
we're
probably
over
provision
previously
to
handle
the
volume,
but
we
didn't
go
because
of
the
basically
bad
performance
of
our
hard
work
and
potentially
older
versions
of
the
code
itself.
So
risa
very
similar
change,
same
thing.
We
have
lost
my
peace
and
army.
A
Latency
is
also
saying
the
same
thing:
blue
line
is
latency
3-2
count,
and
then
it
has
me
15
upgrades
SSDs
and
then
did
the
further
our
way
to
you,
the
12
16
line
kind
of
smooth
that
I've
got
a
nice
slow
City
month
so
that
when
we
r
we
stopped
trying
to
be
poor
performance
weeks
that
we
had
queued
up
in
case.
This
didn't
work
worth
now
how
to
either
done
some
of
those
horrible
to
Matilda.
A
So
this
year
we've
done
he
knew
things
that
are
SMS
you're,
different,
we've
added
a
new
region
without
completely
rebuilding
the
ring
we
at
cine
to
our
cluster
of
four
regions.
We
also
rolled
out
or
not
a
lot
of
use
case.
You
have
some
plans
emery
design
stuff,
which
is
worse
than
we
might
be
the
whole
migration
process
in
here
and
then
kind
of
tiny
master
to
the
freebie
suck.
You
guys
saw
here
in
this
room.
A
We
started
looking
at
spark
for
kind
of
batch
jobs,
reporting
and
essentially
getting
symmetrix
about
the
data
be
sorting
to
sooner
so.
Most
of
this
is
driven
by
client
demands.
They
wanna
know
how
many
things
we
have
store
for
them,
so
vs
over
the
ingredients,
so
we
can
only
have
four
regions
of
our
money,
tokyo,
europe,
east
and
west
of
us.
Those
are
all
young,
take
up
an
enormous
inner
ring
excited
at
city
to
the
mix
it
well.
So
city
is
a
little
different.
A
It's
going
to
be
used
as
a
read-only
copy,
where
we're
going
to
run
the
batch
jobs
and
make
reports
up
against
this,
so
it
doesn't
have
that
everybody
you're
becoming
period.
So
if
you
want
to
apply
caching
or
do
other
strategies,
it
will
be
different
in
sydney
versus
with
one
of
the
other
for
you,
so
the
other
provisions
would
have
geared
towards
more
real-time
random
access
patterns.
Persisting
is
very
much
more
like
a
table
scan
type
operations
that
computed
its
data.
A
So
one
of
the
other
big
topics
that
we
need
to
focus
on
the
series
that
is
like
latency,
so
we
have
a
lot
of
spiciness
from
the
application
bowl
because
in
our
application
we
may
be
doing
more
than
one
reader
writes
a
service
request
because
there
may
be
based
on
the
set
of
parameters
to
come
in
or
based
on
what
r
cubed
you
several
lookups
to
get.
What
we
ultimately
want.
So
this
is
roaming
to
the
blue
line.
Is
the
mean
green
systemic
fifth?
Percentile
purple
is
the
95th
and
red
is
a
99.
A
So
you
kind
of
see
that
on
average
we
have
a
pretty
nice
setting
wait
see,
but
we
definitely
have
a
lot
of
spikes
that
we're
trying
to
figure
out
how
we
can
iron
those
absolutely
the
best
possible
experience
to
like
our
customers.
So
that's
one
of
our
challenges.
A
This
is
where
baby
we
look
at
more
aggressive
fashion,
layered
already
use,
except
you
consider
possible
to
some
sort
of
the
application
layer
or
other
places
to
try
to
influence
that
agreement
kind
of
down
and
the
same
thing
for
the
read
side
to
right
side,
repairs
mess
up
the
cash
counseling,
but
for
the
reason,
if
that's,
what
makes
more
of
a
difference,
maybe
I
should
care
about
keeping
this
latency
as
well.
One
study
as
possible:
we
have
some
internal.
All
personal
educated,
like
10
milliseconds,
would
be
awesome.
A
So
that's
the
metrics
are
reporting.
We
have
been
doing
this
for
a
while.
So
a
customer
client
requests
we've
been
ordered,
you
want
to
add
Python
scripts,
so
they
were
students.
We
never
kid
about
pleasure.
Things
can
get
kind
of
gnarly
parts
mark
lived.
In
summary,
this
isn't
healed
early,
their
clients
eclipse.
So
what
we're
trying
to
do
is
overlooking
I
guess
is
we're
considering
you're,
adding
specific
reporting
tables
or
those
we're
into
at
the
same
time,
such
as
I
said
data
duplication
went
to
our
real
online
kind
of
tables
and
our
reporting
tables.
A
At
the
same
time,
these
we
center
most
about
accept
membership
and
kind
of
evening,
like
that
kind
of
counts,
alternately
with
me,
okay,
using
something
I've
sparks
multiple
out
numbers
and
the
kind
of
computers
numbers
in
crunch
out
periodically.
There
gets
an
idea
what's
going
on,
so
we
haven't
decided
one
of
those
NeoGraft
by
AP,
some
spikes
internal
discovering
court
to
make
one
or
the
other.
A
But
those
are
the
things
we've
been
considering,
and
some
special
sets
water
rate
we've
had
we've
grown
a
lot
and
arms
are
like
eight
notes,
we're
now
to
140
so
you're,
nothing,
not
really
the
lot
of
the
growing
pains
and
curves
and
change
for
like
a
small
effectively.
You
know
test
your
kind
of
odd
since
the
earth
you
know
like
business
itself.
Up
to
you
like
a
full-size
system,
is
actually
used
impression.
A
Now
we
have
estimated
fifteen
billion
rows
and
our
data
tables,
so
it's
grown
quite
a
bit
from
article
you
space,
we're
really
over
and
a
happy
gnome
hundreds
of
thousands,
maybe
a
million
eight
we're
lucky,
so
repetitive,
pasta.
We
also
have
very
low
connects
able
to
support
all
the
ways
with
104
data
and
in
overall.
A
This
is
about
30
plus
terabytes,
on
this
across
under
1402,
so
they're
real
reason
why
we
partridges
not
just
because
of
things
that
can
do
force
there's
also
the
failure
cases
think
stuff
happens
all
the
time,
so
he
knew
about
it.
So
we
surviving
son
of
things
that
would
either
possibly
other
databases
or
make
a
very
hard
field
operationally
to
guaranteed
uptime.
So
the
delk
amazon
swallow
notes,
maybe
the
massively
be
supplies
grievous,
is
disability,
expresses
tables-
and
just
you
know,
normal
networking,
bushes,
between
regions
or
between
between
those
in
st..
A
So
they
said,
there's
some
advice.
One
is
this:
you
know
don't
skip
on
your
hardware
and
assistance.
Work.
Really
really
well
with
Cassandra,
we
don't
use
a
grade,
you
don't
only
using
AES
actually
to
use
the
discount
chain.
You
want
to
stay
current
with
the
releases.
Maybe
I
ought
to
turn
there's
sometimes
bugs,
but,
like
you
know,
rough
look
today
and
then
one
of
the
things
have
been
us
for
a
while.
It
was
keeping
up
on
repairs.
We
started
to
fall
behind
because
their
data
set
was
so
large
and
then
fortunately
they
introduced
way.
A
This
slice
of
repairs
into
smaller
ranges.
So
we're
going
to
use
that
to
like
minimize
amounts,
how
many
economic
eat
reparative
actually
able
to
keep
up
with
their
current
volume
and
load
data,
and
then,
like
I,
said
before,
testing
is
really
really
important
it.
So
we
spent
a
fair
amount
of
time
30
how
to
build
a
testing
frameworks.
Dry
bones.
This
is
to
make
sure
that
has
you
make
changes,
they
don't
negative
impact
performance.
They
can
you
know
we
can
try.
You
know
to
your
things
like
a
V.
We
can
simulate
different
levels.
A
Low
different
amounts,
re
vs
right.
So
that's.
This
is
really
part
of
maintaining
a
stable
system
and
possibly
the
best.
Second
career
to
think
about
how
to
model
your
data,
it's
really
really
what
questions
you
want
to
ask
and
then
how
you
want
to
store
the
data
to
answer
those
questions.
So
we've
spent
a
lot
of
time
kind
of
white
workin
for
you
down
like
what
is
our
day
to
look
like
and
kind
of
going
through.
Iterations
of
it.
So
I
said
we
have.