►
Description
Professor Prithwiraj Choudhury from Harvard Business School interviews GitLab CEO, Sid Sijbrandij on remote work
A
So
I
guess,
where
I
wanted
to
start
said,
is
just
to
get
a
sense
of
the
whole
background
of
how
you
use
you
thought
about
this
company
and
especially
as
you
were
thinking
about
this
company,
how
you
thought
about
all
remote,
just
just
going
back
in
time.
If
you
can,
if
you
can
walk
me
through
like
the
whole
process
and
the
thinking
as
it
evolved
for.
B
B
Then
we
hired
the
mitri
Dmitry
was
in
the
Ukraine,
so
that
wasn't
practical
either
and
then
I
hired
a
couple
of
people
in
the
Netherlands.
It
was
practical
for
them
to
come.
I
had
an
extra
desk
at
my
home
and
we
worked
from
there
and
then
I
think
kind
of
day
three
or
day
four.
They
didn't
show
up
and
thought
about
it
like.
Where
are
they?
They
were
all
lined.
So
we
just
continued
working.
B
It's
like
okay,
well,
I'm
gonna
tell
them
to
show
up
tomorrow.
That's
like
well,
it's
it's
two
hours
like
it's
one,
half
hours
for
them
extra
time
spent.
Are
we
gonna
get
one
half
hours
of
extra
efficiency?
You
know
I'm,
probably
not,
and
we
in
the
first
couple
of
days,
who
already
covered
a
lot
of
like
the
things
that
the
own
boarding
things
that
are
actually
a
bit
easier
to
do
in
person.
Just
asking
a
ton
of
questions.
B
Was
my
home
yep,
then
we
we
did
Y
Combinator.
We
lived
for
in
the
same
house
with
almost
all
of
the
company
that
was
effective
and
intense
and
I.
Think
there's
something
to
be
said.
If
you're
one
team
of
people
and
you
can
be
in
the
same
room
together,
you
have
to
move
quickly.
There's
I'm
definitely
supportive
of
doing
that
in
person.
B
After
that,
everyone
went
back
home
because
they
were
away
from
there
from
the
people
they
care
most
about
and
Y
Combinator
said.
Look
this
whole
remote
thing
are
coaches,
that
one
Combinator
said
it
kind
of
works
for
engineering,
doesn't
work
for
the
other
things
so
consider
getting
an
office
and
I
thought
that
was
good
advice.
You
see
typical
flaw
of
technical,
like
product
focus
CEOs
is
that
they
think
they
can
reinvent
management.
Thank
you.
You
tend
to
see,
there's
a
lot
where
we're
not
gonna
have
any
middle
management
here.
B
I
strongly
believe
in
that
I
believe
in
middle
management
and
sales
and
I
thought.
Look,
let's
not
fall
into
that
same
strap
trap
by
being
adamant
about
having
to
be
remote.
So
we
got
an
office.
It
housed.
15
people
before
I
think
we
had
two
that
was
a
bit
much.
They
had
to
cook
a
big
fat
fan
in
the
in
the
door,
opening
to
keep
it
cool
so.
A
B
B
I
want
to
simply
get
up
reading
things
here
and
then
we
thought
it
wouldn't
work
for
STRs
this
year,
so
their
sales
development
representatives
reach
out
to
enterprises
and
what
they're
just
out
of
just
out
of
school,
they
kind
of
it
and
it's
a
tough
job.
Again
you
hear
a
lot
of
nose,
so
it
kind
of
they
needed
camaraderie
love
of
being
together.
Then
we
found
a
couple
of
great
SEO
are
saying
Utah
and
we're
not
gonna
make
them
move.
B
So
even
that
ended
up
and
then
at
a
certain
point
we
just
decided
okay,
this
works.
Almost
nobody
in
the
company
actually
wants
to
come
to
an
office,
because
it's
a
waste
of
time-
and
the
worst
thing
you
can
do
is
have
a
hybrid
company
where
some
people
are
at
the
office
and
they
use
those
communication.
Styles
and
some
people
are
remote
and
they
use
different
communication
styles,
because
that
that
leads
to
a
lot
of
inefficiencies
and
we
decided
to
become
remote.
Our
investors
were
super
worried
about
it.
B
They
said,
look,
there's
no,
no
very
few
remote
companies
at
scale.
At
the
time
there
was
WordPress
and
and
inflation
was
already
coming
up
and
we're
like
okay.
Well
we're
going
to
be
pragmatic
as
soon
as
like
we
have
this
big
breakdown
in
the
leadership
team
or
something
we'll
all
reconsider
it.
That
didn't
happen
at
a
certain
point.
We
decided
that
yeah
we're
pretty
sure
their
skills
a
lot
better.
We're
actually
able
to
grow
faster
and
and
have
less
disconnect
between
the
leadership
team
than
in
other
companies
and.
B
They've
turned
from
skeptics
into
advocates.
They
seen
our
ability
to
hire
our
placing
their
portfolio
companies.
They
seen
our
ability
to
retain
people
and
our
cost
structure
being
much
better,
so
they're
now
actually
starting
to
focus
on
companies
that
are
all
remote
because
they
think
it's
a
competitive
advantage.
Mm-Hmm.
A
So
I'm
going
to
come
to
the
future
and
the
and
and
and
sort
of
scaling
up,
but
before
that,
so
you
know,
and
through
my
conversations
one
of
the
things
that
strikes
me
as
really
interesting
is
given
that
you're
all
remote
there's
no
headquarters
right,
but
you
know
the
30
40
years
of
management.
Thinking
has
been
about
how
the
headquarters
really
is
important
in
terms
of
performing
many
functions.
B
There's
two
parts
of
it:
there's
hardly
bring
multiple
functions
together
in
the
company
and
the
functions
that
a
headquarter
has
in
bringing
people
together,
for
example,
at
Pixar
they
before
about
how
to
design
their
headquarter
in
the
same
thing
with
a
new
Apple
headquarter
for
the
latter,
I
think
it's
really
important
to
be
intentional
about
informal
communication
and
even
more
so
issue,
not
co-located.
So
there's
a
lot
of
different
things.
We
do
and
I'd
love
to
elaborate
on
them
to
facilitate
remote
communication.
B
As
for
bringing
together
different
functions,
get
lab
is
a
functionally
organized
company,
which
means
that
we're
very
capital
efficient,
there's
great
career
progress.
People
have
great
managers.
The
problem
tends
to
be
the
interaction
between
the
functions,
different
functions,
doing
different
things,
not
coordinating
effectively
things.
B
We
do
to
mitigate
that,
for
example,
having
a
group
conversation
every
day
where
a
function
presents
what
they're
working
on
and
everyone
can
chime
in
having
okay
ours,
to
make
sure
that
the
goals
for
the
quarter
for
all
the
different
functions
are
aligned
and
there's
more
tea
and
a
couple
of
other
things.
But
that's
maybe
I'll
stop
here
and
not.
Allow
me
to
ask
further
questions.
Yeah.
A
So
let
me
ask
a
specific
question:
so
one
of
the
functions
in
my
you
know
research
and
consulting
work
with
large
companies
like
Microsoft
and
Infosys
and
and
Google.
You
know
one
of
the
critical
functions,
the
headquarter,
places,
resource
allocation,
internal
capital,
budgeting
right
and
the
way
it
works,
as
you
would
know,
really
well,
is
the
frontline
people
they
spot
new
ideas.
They
socialize
these
ideas
with
middle
managers.
Then
you
prepare
some
sort
of
like
a
forecast
of
profitability.
A
Then
there's
some
internal
budgeting
meeting
at
the
headquarters,
where
all
the
functions,
finance
and
marketing
and
and
in
technology
would
come
together
and
the
CEO
or
the
CEO
would
say
this
is
a
go.
This
is
a
Nobu
and
this
is
a
go
and
this
is
a
no-go
and
then
you
would
track
that
every
quarterly
based
on
QB,
RS
or
whatever.
So
this
whole
internal
resource
allocation
process
I'm
just
curious
how
it
happens
in
your
context,.
B
B
and
everyone
in
the
company,
whether
it's
sales,
whether
it's
external
people
external
to
the
company,
whether
whether
it's
engineers
can
argue
with
the
product
manager,
to
change
to
change
their
mind,
but
it's
their
responsibility
and
we
call
this
principle
directly
responsible
individual.
They
are
the
DRI.
They
make
the
decision.
You
can
try
to
change
their
mind.
So
it's
a
continuous
thing
like
there's
always
the
next
month.
We
can
even
make
adaptations
at
a
bigger
scale.
B
B
How
we
allocate
the
money
now
is:
is
a
zone
they're
all
on
a
slope
to
get
to
to
that
eventual
model,
so
every
every
every
department
has
their
allocation
and
it's
largely
up
to
them
to
decide
how
to
spend
that
money
within
within
their
company.
Now,
there's
always
examples
of
cross-functional
things.
For
example,
we
have
a
project
called
Mel
tano,
which
is
kind
of
a
separate
product,
separate
completely
separate
from
Gila.
A
B
B
We
have
our
eventual,
like
suppose.
Our
eventual
model
is
that
sales
will
have
twenty
percent
of
budget
right
now,
we're
growing,
fast,
I.
Suppose
it's
now
forty
five
percent
of
budget,
that
money
goes
to
sales,
I,
think
that
so
it's
very
clear
how
much
budget
each
function
has
I
think
the
only
hard
thing
to
so
far
has
been.
How
does
engineering
allocate
their
head
count
against
all
the
different
parts
of
the
project?
Engineering
can
say:
okay,
we're
spending
this
much
in
Q&A
we're
spending
this
much
in
documentation.
B
We're
spending
way
too
much
in
debt
that
they
can
decide,
because
they
know
they're,
gearing
ratios
how
much
you
spend
in
the
different
kind
of
parts
of
the
product.
We
have
like
a
planning
component
and
a
monitoring
component
and
eight
other
components
to
the
product.
Do
we
put
an
extra
team
on
planning,
or
do
we
put
an
extra
team
on
monitoring
that's
up
to
product,
because
product
knows
like
more
to
market
ones
where
they
want
to
go
where
they
want
to
drive
you're
opening
where
they
will
need
most
capacity?
B
So
that's
kind
of
been
the
only
thing
where
it's
there's
a
a
very
extensive
collaboration
product
we'll
make
a
proposal.
Engineering
we'll
make
sure
it's
all
practical
that
we
don't
end
off
with
half
people
or
infeasible
teams.
Even
I
have
a
look
at
it
because
I'm
a
product
driven
CEO,
but
it's
it's
mainly
product
who
sets
those
priorities
within
input
from
other
people,
but
they
in
the
end
they
decide.
We.
A
B
B
A
B
There's
two
components
to
it:
one
is
the
eventual
steady-state.
How
much
are
we
going
to
spend
on
every
part
of
the
company
finance
develop
that
by
looking
at
comparables
and
kind
of
okay
we're
more
product
heavy
than
others?
This
is
about
how
much
Enterprise
we
do
so
they
developed
those
numbers
and
it's
a
model
like
people
want
to
know.
B
If
you
like,
for
example,
go
public
investors
want
to
know
that
eventual
model
like
how
you're
going
to
turn
a
profit,
and
then
we
made
a
formula
to
get
on
a
glide
path
and
that
formulas,
basically
it's
been
inspired
by
the
rule
of
40.
Are
you
from
familiar
with
that
by
any
chance
who
no
I'm
not
so
the
rule
of
40
is
the
best
practice
that,
if
you
add
up
your
EBIT
are
margin
and
your
growth
rate,
you
should
be
about
above
40%.
C
B
B
B
Trick
is
to
not
run
out
of
cash.
You
can.
With
this
model,
you
can
predict
your
cash
flow,
so
the
only
way
to
run
out
of
cash
is
if
your
revenue
is
not
growing
at
the
rate
you
anticipate
it
and
the
action
you
take
if
that
happens,
is
like
you
slow
down
high
to
make
sure
you
fall
back
within
the
model,
so.
A
Maybe
what
I'll
do
is
sit
like
tell
me
as
a
CEO,
you
said,
you're
a
product
focused
CEO.
Tell
me
what
a
typical
week
in
your
life
would
look
at,
like
so
talking
to
the
investors
talking
to
product.
So
tell
me
like
what
would
be
the
set
of
activities
that
you
as
a
CEO
perform
in
a
typical
movie
yeah.
B
Shadows
I'm
gonna
hand
it
over
to
you
after
after
I
had
a
first
stab
at
this
typically
there's
different
things
like
there's
calls
for
the
entire
company,
for
example.
This
week,
I
did
a
group
conversation.
We
had
a
values
conversation
where
the
whole
company
was
invited.
It's
external
meetings
mainly
to
promote
like
a
remote
and
support
leadership.
Meetings
like
this
is
hiring
interviews.
There's
talking
with
advisers,
there's
my
one
on
once
with
my
direct
reports.
There's
skip
level
meetings
with
the
reports
of
my
reports.
B
B
C
B
A
B
A
A
Yeah
yeah
yeah
I,
like
that,
it's
so
and
and
beyond
that.
You
also
mentioned
that
you
periodically
you
co-locate
you
meet
somewhere
so
last
year,
I
believe
it
was
in
New
Orleans
and
then
you
doing
it
again
in
Prague
next
year.
Right
so
so
are
those
meetings
like
pretty
similar
to
the
meetings
you
have
on
zoom'
or
are
they
different
in
any
way?
They're.
B
We
not
only
are
they
not,
we
don't
do
the
regular
work.
We
don't
allow.
For
example,
when
we
all
come
together
for
a
week,
we
don't
do
like
death
by
PowerPoint.
We
don't
have
big
presentations.
We
have
an
opening
in
a
closing
ceremony
where
there's
where
it's
like
a
talk
at
a
conference,
but
throughout
the
week
there's
no
projectors,
there's
no
presentations.
A
You
know
managerial
question
that
people
can
debate
right
and
the
thing
that
I
was
thinking
about
was:
are
you
scale
this
model
to
the
next
level,
so
hypothetically
you're
at
about
1,000
something
now,
as
you
see
this
model
being
scaled
up
to
say,
five
thousand
or
even
ten
thousand?
Where
do
you?
Where
do
you
feel
given
your
experience
in
the
past
like
few
years,
this
could
break.
So
what
are
the
key
concerns
that
keep
you
up
at
night,
so
to
say,
yeah.
B
I'm
not
sure
that
sir,
a
good
question
to
big
debate,
because
it's
actually
this
model
skills
a
lot
better
than
co-located
models.
So
if
you're
co-located
all
in
the
same
room
that
works
that
works
well,
you're
on
the
same
floor.
Okay,
if
you're
on
the
same
building
man
you're
on
the
same
city,
same
state,
same
continent,
it
gets
harder
and
harder
and
harder
so
at
scale.
B
Ole
remote
works
better
because
the
practices
that
you
need
at
scale
like
extensive,
asynchronous
communication
like
documenting
and
writing
down
and
recording
things
those
practices
they
help
you
scale.
So
it's
not
this
old
remote
becomes
a
bigger,
different,
a
bigger
benefit,
as
you
scale,
not
a
smaller
one.
So
this
this
idea,
it
breaks
that
a
certain
point
is
very
foreign
to
me
and
if
you
think
that
many
people
would
debate
that
I'd
be
happy
to
work
with
you
on
that
question,
or
we
can
also
discuss
alternative
questions
so.
A
I
think
yeah
I
would
love
to
you
know,
brainstorm
alternative
questions,
but
let
me
tell
you
what
I've
heard
right
so
and
in
my
you
know,
four
or
five
calls
so
far.
So
if
I
list
the
sort
of
like
frictions
and
I'm
not
saying
none
of
these
can
be
solved,
but
there
are
frictions
like
in
any
models.
1:1
friction
that
came
that
people
talked
about
was
time
zones
and
the
challenges
of
communicating
across
time
zones
right
a
related
thing,
which
I
think
is
a
very
interesting
thing
to
debate.
B
B
B
First
is
communicating
on
chat
and
those
two
things
you
lose.
If
you
don't
have
synchronous
communication,
a
synchronous
communication
across
time
zones
is
it's
very,
very
hard
I,
don't
know
how
to
solve
time
zones
I,
don't
think
anybody
does.
We
do
know
how
to
do
as
a
lot
of
asynchronous
communication,
but
there's
we
don't
want
to
go
to
like
always
doing
asynchronous
communication,
because
we
know
it's
not
the
most
efficient
thing
for
every
situation
and
how
we
deal
with
it.
B
Regarding
new
hires,
that's
a
big
onboarding
is
so
important.
We
do
a
better
job
than
any
other
company
in
the
world,
but
it's
it
can
always
be
better.
For
example,
we
started
office
hours
for
people
with
kind
of
IT
problems,
I
think
we're
not
doing
a
good
job
of
making
people
with
familiar
with
our
values.
There
are
communication
methods
and
with
how
to
use
gitlab.
B
So
actually
we
had
that
conversation
earlier
today
and
where
I
put
in
a
proposal
to
RTP
for
officer
whether
I
could
have
a
an
early
start
with
making
a
learning
management
system
that
test
for
those
things,
so
that
we're
absolutely
sure
that
people
who
join
the
company
are
familiar
with
those
things.
Mm-Hmm.
A
B
Last
thing
is
about
consensus
and
how
we
make
decisions.
Is
it
helpful
if
I
elaborate
a
bit
on
our
model
on
how
to
make
decisions
yeah?
So
it's
documented
on
our
leadership.
Page
yeah
and
they'll
put
it
in
the
chat.
The
idea
is,
we
want
to
combine
the
best
of
decision
of
consensus
organizations
and
hierarchical
organizations
and.
B
B
How
we
keep
the
best
of
hierarchical
organizations
is
by
having
a
DRI
there's
one
person
who
will
decide
now.
There's
one
tricky
thing
here:
if
you
make
the
DRI
kind
of
convince
or
explain
the
decision
too
much
to
all
the
people
who
provided
input,
you
will
get
projects
that
fly
under
the
radar
and
that's
what
I've
seen
in
a
lot
of
companies
so
very
explicitly.
The
DRI
does
not
have
to
explain
why
there
taking
the
decision
and
they
absolutely
don't
have
to
convince
other
people.
A
B
I
think
we
should
do
a
better
job
of
surfacing.
It
I'm
pretty
sure
that
this
making
decisions
part
of
our
handbook
that
no
one
who
on
boards
it
actually
reads
this
part
of
our
handbook
or
no
one
like
less
than
5%.
So
we
should
make
it
part
of
like
the
curriculum,
and
we
should
test
whether
they
can
answer
a
basic
question
about
like
hey,
why?
Why
is
it?
Why
is
it
a
problem?
B
If
you
have
to
make
sure
everyone
feels
heard
and
gaben
Chania,
maybe
have
a
look
at
the
language,
because
I
think
we
have
want
to
have
some
strong
language
in
there.
That
says,
look
as
a
DRI.
It's
not
even
your
task
to
make
sure
people
feel
hurt,
because
that
will
lead
to
decisions
flying
under
the
radar,
because
the
more
people
will
know
about
a
decision,
the
more
work
it
is
yeah.
A
Yeah
make
sense
so
two
more
things
and
by
the
way
said,
I'm
a
huge
fan
of
this
work
from
anywhere
model,
and
that's
why
I
was
very
excited
to
work
on
this.
But
you
know
we
we
have
to
document
and
then
discuss
in
the
class,
because
people
will
be
thinking
about
these
right
things
that
you
know
we
heard
and
I
would
like
to
get
your
thoughts
on
relate
to
like,
if
I'm,
a
really
really
young
person
and
I'm
very
motivated,
it's
easy
to
get
into
a
mode
of
working
constantly
and
suffering
burnout
very
soon.
Yes,.
B
A
Relatedly
there
was
an
you
know,
I
think
one
of
the
people
I
spoke
to
yesterday
said
this
is
a
great
model
if
you
know
how
to
manage
your
social
life
proactively.
But
if
you
are
someone
who's,
an
introvert,
the
isolation
can
be
terrible,
right
and
I,
wonder
I'm
sure
you've
thought
about
it
and
then
I
just
I'm
curious
how
you
thought
about
dealing
this
that's
yeah.
B
B
One
of
the
things
we
require,
one
of
our
selection
criteria
is
that
people
should
be
a
manager
of
one
they
should
be
able
to
manage
their
own
time.
We
also
try
to
make
it
a
frequent
topic
of
mental
health.
A
frequent
topic
of
conversation
like
we
have
sections
in
the
handbook
about
it.
We're
working
on
old,
remote
section.
B
We
have
contracted
with
a
provider
called
modern
health,
that's
accessible
at
no
cost
to
anybody
in
the
company,
I'm,
not
something
it's
I'm,
not
saying
it's
something
we
solved,
but
we're
aware
of
it
we're
doing
better
than
most
companies.
It's
also
we're
trying
to
lead
by
example.
I
took
three
full
weeks
off
this
summer,
so
just
in
in
a
single
quarter
like
we
do
take
time
off
work
as
for
introverts,
we'll
have
a
hard
time.
B
B
A
That's
fascinating
so
and
I
think
that
that's
the
huge
tech,
not
a
tectonic
shift
here,
which
I
totally
agree
and
that's
why
I
was
trying
to
understand
how
you
do
the
budget
allocation.
But
beyond
these
five
concerns
just
to
wrap
up
the
concerns.
Conversation
said
so
we
talked
about
time
zones.
We
talked
about
new
hires,
we
talked
about
autonomy
and
and
how
you
you
re
familiarize
yourself
in
the
in
this
new
culture
of
doing
things
and
then
burnout
in
isolation
which
of
these
five
has
been
the
trickiest
issue
in
your
mind,
to
solve.
B
The
trickiest
one
has
been
some
thing
we
didn't
discuss
and
it's
working
handbook
first
and
it's
making
sure
that
every
change
that
happens
at
at
get
lab
is
affected
in
the
handbook
and
gate.
Maybe
you
can
you
can
talk
about
what
your
experience
has
been
in
your
previous
comfortable
company
if
you're
comfortable
with
doing
that
publicly.
C
Yeah
so
I
smaller
company-
and
we
made
the
mistake
of
trying
to
do
the
hybrid
approach.
Two
thirds
of
our
team
was
distributed
across
the
United
States
on
third
was
an
office
in
an
attempt
to
get
everybody
on
the
same
page.
I
push
down
from
the
top.
We
were
gonna
document
everything
in
the
company,
we're
gonna
do
similar
to
what
does,
but
just
using
github
pages
at
the
time
and
every
change.
C
Every
policy
was
gonna,
be
reflected
and
everybody
had
to
keep
doing
that,
but
because
it
wasn't
from
the
beginning
of
the
company,
the
sales
and
marketing
says
organization
never
bought
in
and
if
everybody
doesn't
buy
and
it
doesn't
work,
and
so
it's
nearly
impossible
to
to
manage
information
unless
it's
written
down
and
without
that
discipline
you
can't
work
for
a
month
effectively,
and
so
that's
a
big
challenge.
I
ran
into
and.
A
C
A
A
B
That
has
always
been
the
trickiest
issue:
it
takes
so
much
discipline
to
put
something
in
the
handbook,
instead
of
just
bringing
people
on
slack,
sending
an
email
doing
all
these
other
ways
of
asking
people
to
do
something.
All
these
other
ways
are,
namely
much
faster
because
you
don't
have
to
find
a
relevant
part
of
the
handbook
restructure
it
to
accommodate
your
your
change.
Make
the
change
then
communicate
this
the
change
and
also
because
it's
in
context
you'd
have
to
consider
the
impact
on
every
every
other
process
which
you
don't
have
to
do.
B
If
you
just
tell
want
to
do
something
different,
so
it's
slower
and
people
don't
like
slow
this
kind
of
go
slow
to
go
fast.
It's
like
I,
don't
know
writing
code
with
our
tests.
Yes,
it's
it's
faster
to
get
that
code
written,
but
in
the
end
it
prevents
you
from
making
changes
upon
changes.
It's
kind
of
that
is
where,
where
you
end
like,
you
will
not
be
able
to
do
to
accomplish
anything
beyond
that
because
kind
of
the
defender,
the
foundation
of
what
you're
building
is
starting
to
crumble.
B
It
is
just
more
painful
for
the
pre
people
who
want
to
have
a
change
take
effect,
but
of
course,
although
it's
quicker
for
them
in
that
moment
that
emails
gonna
get
lost
in
two
years.
No
one
will
know
that
that
is
the
process
and
they'll
have
to
rediscover
it.
In
the
end,
you
end
up
going
slower,
I,
wonder.
A
B
A
A
Thinking
about
so,
you
know,
I,
you
know
I,
so
I
go
to
my
PCP
in
in
the
US
and
that
gentleman
now
speaks
into
like
a
prescription
system
and
documents
and
I.
Guess
it's
just
easier
for
that
gentleman
to
like
speak
to
a
system
then
type
something
up
and
I'm
wondering
whether
contemporaneous
to
a
call
you
could
you
could
capture
the
voice,
transcribe
it
using
some
some
automation
and
then
just
automatically
update
the
handbook
without
any
human
intervention.
B
I
think
the
problem
is
the
structure.
So
the
example
you
talked
about
doesn't
need
any
structure
because
it's
your
prescription,
so
it
doesn't
affect
anything
else.
If
you
for
you
needing
to
find
like
where
in
the
hierarchy
or
the
structure
does
it
belong
and
then
insert
it
correctly
and
make
sure
that,
like
you,
rearrange
the
sides
of
it
a
bit
to
make
it
sure
it
fits
in
that
takes
extra
time,
and
instead
of
you
doing
that
you
can
do
to
say
well.
B
The
structure
is
in
people's
heads
and
I'm
just
going
to
push
in
this
new
message
and
it's
up
to
them
to
figure
out
how
it
works.
That's
way
more
inefficient,
because
now,
if
I
wasn't
people
I
have
to
do
it
and
they
won't
do
as
good
of
a
job
and
they'll
forget
about
it.
But
for
you,
the
sender
of
the
message
that
was
a
lot
less
work.
So
are
you
gonna
build
that
house?
Are
you
gonna
build
that
written
down?
Are
you
gonna
build
that
in
people's
minds
and
most
companies
answer
that
question?
B
It's
different,
the
house,
the
house
looks
different.
It
means
it's
weak,
it
crumbles
it
continually
deteriorates
and
needs
reinforcement
and
sending
the
same
message
again
and
again
and
again
and
again,
there's
also
problems
at
our
model
like,
for
example,
there's
not
a
natural
aging
of
messages.
Maybe
we
put
stuff
in
five
years
ago
that
need
to
be
taken
out,
we're
not
taking
out
a
lot
of
messages.
So
maybe
that's
gonna
be
something
that
will
hurt
us.
B
A
No
I
agree
and
I
think
this
is.
This
is
super
interesting
because
you
know
unless
and
and
you
know,
given
everything
you've
done
and
I
guess
you
can.
You
can
continually
improve
this,
but
this
is
critical
for
asynchronous
communication
as
well
right
because
unless
in
a
synchronous,
there's
no
reading
of
minds,
you
have
to
read
it
read
the
documentation.
Yeah
I.
B
A
B
It's
it's
both
making
it
easier
to
do
like
the
technical
editing,
but
also
convincing
people
that
this
is.
This
is
the
way
and
it
becoming
something
that
people
do
a
bit
of
a
bit
of
thing.
We
recently
added
handbook.
First,
you,
like
our
leadership
criteria.
You
cannot
be
a
good
manager
at
good
lab
if
you
do
not
reinforce
handbook.
First,.
A
Because
you
know
what
I'm
thinking
allowed
said,
but
if
this
is
such,
if
this
is
the
issue
that
bothers
you
as
a
CEO,
then
it
it
deserves
more
attention
from
me
and
Emma,
and
maybe
we
can
dive
deeper
and
ask
a
follow-up
call
with
someone,
even
you
trying
to
understand
the
handbook
first
process
and
then
trying
to
understand
where
it
breaks
and
trying
to
understand
and
making
it.
You
know
that
whole
thing,
the
discussion,
a
more
central
part
of
the
case,
study,
cool.
B
I
think
I
think
jape
is
helpful
because
he's
seen
it
break
down
so
I
think
Darren
is
helpful.
We
have
a
handbook,
a
usage
page
and
maybe
the
product
manager
for
the
static
site.
Editor
can
also
join
that
conversation.
Anything
with
those
four
people
you'll
be
well
served
and
I'd
be
happy
to
do.
A
follow-up
course
is
that
if
you
don't
get
everything
out
of
that,
call
that
you
want
absolutely.
A
So
maybe
we
can
reach
out
to
Gabe
and
schedule
that
meeting
yeah,
okay
boss,
set
it
up
awesome.
Just
one
last
thing
before
you
know:
I,
let
you
go
so
you
know,
as
we
were
doing
the
research
about
the
company
and
I'm
sure
the
students
will
bring
this
up.
We
found
the
the
news
article
about
hiring
from
China
right.
So
can
you
tell
me
your
thoughts
on
that
and
you
know
so.
This
is
clearly
one
more
friction
about
working
from
anywhere,
because
now,
because
of
regulatory
or
other
reasons,
you
cannot
hire
people
from
that
country.
B
So,
first
of
all,
there's
a
bit
of
distortion.
We
hire
people
in
China,
we
have
customers
in
China.
We
have
contributors
to
get
lab
in
China.
That
is
not
the
conversation.
The
the
conversation
is
for
the
people
at
gate,
lab
that
have
access
to
the
production
data
of
get
Lancome,
for
example,
some
of
our
site,
reliability,
engineers
are,
we
gonna
hire
them
in
China
or
Russia,
and.
B
The
second
thing
is
it's
almost
weird,
because
if
a
normal
company
wouldn't
want
to
hire
for
certain
role
in
a
certain
country,
teachers
wouldn't
open
up
the
vacancy,
and
where
does
all
vacancies
are
like
everywhere
in
the
world
unless
specified?
Otherwise,
so
we
made
it
a
bit
more
difficult
on
ourselves.
We
could
have
also
just
listed
the
250
countries.
We
hired
is
for
this
function
and
we
haven't
done
that.
Yet
maybe
we
should
maybe
not
maybe
most
people
understanding
complex,
that
forget,
mark
omit.
B
It's
mainly
used
by
US
companies
and
Russia
and
China
are
cyber
adversaries
of
the
US.
If
we
would
open
up
a
gate,
lap
saw
service
in
China,
maybe
wouldn't
hire
anybody
in
the
US.
For
that,
like
it's,
it's
not
about
one
country
being
better
or
people
being
better
than
the
other.
It's
about
hiring
near
data
like.
Where
is
the
data
stored?
Who
uses
it
and
making
sure
that
the
people
are
near
there,
because
there
will
always
be
tension
in
the
world.
C
A
B
One
thing
that
I
don't
think
it's
a
good
fit
for
the
old
remote
subject,
but
on
the
subject
of
China,
we
spent
a
ton
of
time
thinking
about
how
to
offer
get
lamp
as
a
service
in
China
and
whether
that
should
be
a
fully
owned
enterprise,
a
joint
venture
or
a
company
where
a
minor
shareholder
in
or
whether
we
shouldn't
do
it
at
all.
But
I
don't
think
it's
a
good
fit
for
all
remote.
And
that's
my
be
the
gist
of
your
article
right.
A
So
I
guess
you
know
the
process
for
us
would
be
I.
Think
we
have
a
couple
of
interviews
left
and
then
a
mine
I
will
work
on
the
draft
and
hopefully
send
you
something
early
January
and
then
you
can
read
it
give
us
feedback
on
that,
it's
very
collaborative
and
then
at
that
point
we
might
request
for
a
couple
of
more
calls
and
then
hopefully
at
least
spring.
We
publish
this
and
then
I'm
going
to
teach
it
in
my
second
year.
Elective
at
HBS
and
I
would
love
to
invite
you
to
the
classroom.
A
B
You
ok
with
publishing
this
corner,
unfiltered
YouTube
channel,
oh
yeah,
absolutely
no
way!
Awesome!
Chen,
yay,
we'll
get
in
touch
with
you
about
that.
That's
good
awesome!
Well!
Thank
you!
So
much
for
the
questions.
You
did
a
great
job
of
finding
things
that
we
are
struggling
with
at
the
company
and
it
was
fun
to
discuss
them
with
you.
Yeah.