►
From YouTube: 2019 05 30 memory team weekly
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
Oh
yep
recording
is
started.
Okay,
so
first
of
all
welcome.
Creek
Creek
is
our
new
manager
of
the
team
and
he
will
be
working
with
Camille
and
and
everybody
off,
your
recruitment
comes
okay,
so
I
repeat,
welcome,
Creek
and
quick.
So
just
let
you
know
on
the
screen:
Kenya
Kenya
is
a
key
manager
for
for
our
team.
Our
partner
and
Mac
Mac
is
a
queue
director
and
Camille.
A
A
C
B
A
Okay,
so
next
item
is
it's
just
a
reminder
that
we
agree
back
from
the
each
mas
from
the
first
day
to
the
service
day,
basically
June
1st
to
2/7,
we
are
in
the
planning
mode.
Let's
do
the
ASIC
planning
in
in
select
channel,
but
the
board
is
there.
The
links
to
the
board
is
at
the
channels,
topic
or
channel
summary
there.
So
two
leagues
there
use
the
planning
board
to
do.
A
The
planning
and
I
will
will
discuss
real-time
there
in
a
channel
like
a
since
it's
like
breaking
down
the
issues
to
smaller
pieces
and
what
we
want
to
do
for
next
to
release
those
kinds
of
planning,
&
Gumi.
So
yes,
just
a
kind
of
reminder
and
the
next
coming
releases,
twelve
point
one
if
I'm
correct
that
planning
is,
for
the
twelve
point,
one
release,
we
are
working
on
the
twelve
point
out
right
now,
yet
thank
you
that
is
status,
update,
afore,
12
point.
Oh,
that
was
c'mere.
That
must
be
you
right.
A
D
I'm
still
I
started
like
putting
the
notes
there
is,
but
we
did
a
lot
of
I
guess
like
in
one
of
the
future
meetings.
We
basically
could
work
for
the
part
when
we
have
more
people
joining
it's
kind
of
like
tweets,
so
Mario
what
things
are
happening
now,
if
you
have
like
any
questions
about
these
things,
I
feel
free
to
ask
right
now.
I.
C
D
D
Everything
is
kind
of
like
behaving
as
expected,
so
like
from
my
perspective,
I
didn't
expect
like
any
any
edits
like
changes
from
the
from
the
memorial
consumption
on
the
Puma
like
it's
it's
kind
of
like
it
seems
that,
like
for
this
more
installation
as
it
seems
to
be
the
Sun
and
that's
zero,
pretty
much
I
can
write
it.
This
thread
from
my
to
this
investigation
on
like
trying
that's
out
like
it
seems
that
we
get
just
another
strike
with
12.0.
A
A
A
Where
else,
so
so
something
else
or
was
in
the
discussion
that
Dahlia
was,
was
working
with
her
team
to
see
to
implement
a
little
change
to
groom
the
backlog
more
frequently
she's
thinking
about
the
weekly
grooming
and
to
keep
us
moving
faster
I
think
that's
a
good
idea,
but
we
don't
have
to
be
it.
You
know
weekly,
maybe
I
mean
either
at
a
pace
we're
comfortable
for
that
works
best
for
our
team.
So
it's
just
the
food
for
thought.
I
heard
Dalia
proposed
to
that
too
for
her
teams.
A
So
we
observed
that
it's
a
provider
it
will
help
us
moving
faster
because
it
more
frequently
and
more
frequently
prioritization
will
keep
us
moving
faster.
Definitely
so
food
for
thought.
I
haven't
made
a
item
here,
but
it's
up
to
the
team.
We
can
discuss
what
semester
for
us,
but
in
general
my
dearest
is
probably
a
good
thing
to
do.
A
D
It's
it's
actually
what
I,
what
I
mentioned
a
few
minutes
ago
so
like
who
might
serve
for
this
more
installation?
It
doesn't
really
make
a
lot
of
difference
from
the
memorial
consumption
like
what
I
explained
right
now,
when
we
unable
to
get
for
the
github.com
food
like
we
have
16
workers,
for
example.
This
is
wearing
a
shorts
like
see
the
difference
cool,
alright
so
like
who
might
more
like
the
it's
a
tricky
because,
like
from
the
performance
point
of
it
also
doesn't
seem
to
like
show
you
like
that
this
is
way
faster,
like
it's.
D
It's
better
memory
consumption,
but
this
kind
of
model
of
operation
like
each
task
more
features
basically
to
play
with
yeah
it's
technically.
It's
not
really
like
it
that
we
gonna
achieve
like
the
big
benefits.
We've
like
that
for
myself.
The
Puma
just
opens
us
to
big
more
efficient
in
the
future,
but,
like
today,
it's
fun
to
not
make
the
get
up
to
be
like
go
so
30%,
less
much
less
memories.
It's
not
gonna
happen
with
first
point:
zero,
yeah
cool.
A
D
Ahead
so
so,
actually
like,
like
we've
done,
if
we've
the
sightings,
that
we
have
to
pay
for
the
small
deployments,
it
seems
to
make
no
difference
at
all,
but
we
could
find
you
there,
like
the
small
deployments
to
make
it
like
editor
of
my
efficient
so
like
we
could
rightly
have.
We
would
use
something
like
between,
like
300
or
400,
less
memorial,
on
this
more
deployment.
D
So
if
we
find
that
like
we
can
skip
like
one
worker
for
the
Puma,
this
is
when
we
cannot
select
the
perb
like
my
more
improvement,
because
the
running
footprint
could
be
like
something
big
to
to
400s
memorial
for
this
morning's
durations
month.
I
still
like
our
recommendation,
is
eight
geeks
right
now,
so
it
doesn't
remake
it
like
way
better
than
that
right.
D
A
D
Thing
I
think
is
about
tuning
because,
like
well
from
what
I
saw,
I
saw
the
performance
degradation
on
Puma
of
like
10%,
because
of
running
16,
concurrent
stress
on
the
single
CPU,
so
unicorn
wars
in
the
model
that,
like
Iran
concurrently
marks
to
request
with
the
default
settings
from
the
Puma.
This
pool
of
threads
is
much
bigger,
so
actually,
like
I
was
seeing
by
the
contention
of
the
of
the
execution
so
really
like
right
now.
The
thing
that
we
have
to
work
on
is
figuring
out
the
defaults.
D
Before
that,
like
we
don't
see
the
performance
degradation,
because
we
see
some
small
performance
degradation
right
now
and
put
like
using
of
the
home
of
the
smile
installation
can
actually
create
a
much
bigger
contention.
Issue
then
like
having
unique
or
so
I
would
say
that,
like
we
have
to
tune
down,
I
need
to
like
the
Puma
settings
to
like
to
make
it
like.
Yes,
maybe
not
not
expose
like
16
threads,
but
rather
my
baby
exposed
like
4
threads.
D
So,
like
numeric
instruction
seems
to
be
like
pretty
much
equal
to
the
default
settings
that
we
have
weaved
a
unicorn,
but
the
belonging
of
the
memory
can
like
if
the
Puma
can
be
much
bigger
than
what
is
happening
with
the
unicorn,
because
in
it
we've
in
the
unique
or
model
right.
We
have
one
request
running
on
the
unicorn
which
can
write.
We
actually
give
a
pair
instead
of
something
like
100
or
200
max.
D
So
we
have
to
tie
to
an
eight
or
differently
the
the
single
worker
process
to
accommodate
to
like
to
this
multi
portraits
running,
so
it
can
erase
our
that,
like
we
might
have
a
little
higher
memory,
consumption,
the
single
worker,
but
because
now
maybe
run
just
a
one
worker
which,
in
the
end,
should
give
us
a
much
better
like
like
performance
and
a
little
better
memory
consumption
than
unicode.
Because
then
we
don't
really
need
to
like
to
unique
foreskin
correctly.
Maybe
just
one
Puma
with
like
fortress
or
a
trades.
D
This
is
actually
what
I'm
trying
right
now
to
figure
out.
Our
prayer
environment
actually
is
configured
like
with
single
CPU,
which
seems
like
the
idea
of
example,
of
like
joining
the
small
screens
narration
to
so
I'm
gonna.
Basically,
right
now
try
to
get
our
load
balancing
sorry
about
monastic
the
RT
very
test
hood
to
figure
out
what
is
the
speed
limit
between
my
Mario?
Such
mom
already
meets
a
number
of
workers
and
like
try
to
ship
something,
and
then
we
could
tune
it
further
for
the
22.0
cool.
A
D
Yes
and
I
think
that,
like
what
we're
gonna
do-
and
this
is
what
I
gonna
start
doing-
maybe
like
the
creative
issue,
I
would
like
us
to
like
to
make
an
hour
hour.
Each
hundred
come
canary
to
run
on
the
Puma
for
our
purpose,
and
we
could
gather
metrics
and
meter
from
there
to
to
better
understand
how
it
works
and
like
on
the
be
scale
deployments,
not
really
small
scale
department.
D
A
E
Yeah,
that's
a
hub
in
probably
guess
I,
don't
think
we
can
literally
set
up
environments
that
fast
right
now.
I
know
that
Jerry
is
working
on
tariffs
on
scripts,
to
set
up
newer,
smaller
environments
that
we
can
set
up
at
scale
yeah.
So
what
how
many
combinations
are
we
looking
at
Camille
and
what's
the
timeline,
do
you
know
I
would.
D
I
would
just
figure
out
to
like
just
a
small
thing.
First,
like
let's,
let's,
let's
make
note
of
our
engineer,
Mike
our
solutions.
I
just
thought
with
this
morphing
and
let's
see
exactly
how
it
behaves
on
deegeu.com
and
like
I,
feel
like
this
2x10
should
give
us
a
compression
what
we,
what
kind
of
isleta
we
are
looking
at
I.
Don't
really
think
that
right,
we
need
anything
more
at
this
point,
the
environments
that
we
have.
It
seems
very
good
enough
for
me
at
this
moment.
A
Mack
yeah
Mack
your
comment:
the
entire
release
of
planning
is
also
moving
to
twice
a
month.
That's
good!
If
so,
I
think
we
can.
We
can
do
either
weekly
or
bi-weekly
I
think
either
way,
but
the
teams
teams
the
team
makes
the
call
was
the
best
for
us,
but
in
general
that's
a
good
practice.
We
can
move
faster
with
the
frequent
planning
goomy.
A
Thank
you.
Mike
I
was
not
aware
of
that
and
then
Tanya
is
out
from
June
3rd
four
kings
and
I'm
out
from
8
to
10
users.
So
that's
it
just
a
heads
up,
yes
Mac!
Should
we
do
a
wrong
table
update
on
the
new
hires
timeline?
Yes
yeah,
my
fault
I
should
have
a
hiring
update
here
and
also
your
hiring.
Of
course
you
got
there.
She
was
in
your
test
automation,
engineer
starting
on
June
17th.
Let
me
catch
engineer
a
bit
here.