►
From YouTube: Podcast 003 - CoScreen
Description
A conversation with the founders of the collaborative tool CoScreen
A
Hi
everyone
and
welcome
to
this
new
episode
of
our
little
podcast
here
today
we
are
talking
to
a
very
exciting
company
from
the
us
called
co-screen.
I
I
think
they
have
a
a
fascinating
way
to
to
to
look
at
collaboration,
and
so
when
I
both
wanted
to
make
sure
that
you
are
familiar
with
them,
because
they're
interesting
from
a
product
perspective
and
also
from
a
technological
perspective
and
we'll
try
to
touch
on
both,
but
first
till
and
jason.
B
Yeah
nice
to
meet
you
good
to
be
here,
one
we'll
try
not
to
make
it
too
long.
Why
not?
So
until
originally
from
germany,
I'm
the
co-founder
and
ceo
of
coldscreen
we've
been
working
on
coolscreen
for
quite
some
time.
It
was
once
an
sap
kind
of
skunkworks
project.
Looked
a
little
different.
You
know
we
thought
that
it's
really
something
there.
I
was
able
to
take
it
out
of
sap
working
it
on
the
site
as
a
hobby,
more
or
less.
B
That's
when
we
ended
up
raising
our
seat
round
roughly
one
and
a
half
years
ago,
put
a
team
together,
we're
now
a
team
of
eight,
mostly
in
the
us.
As
you
mentioned,
I'm
in
menlo
park
in
the
bay
area,
and
we
have
one
team
member
in
germany
and
yeah
now
we
launched
the
beta
of
course
running
on
jazz
six
months
ago.
It's
exciting
hundreds
of
teams
of
around
the
world
use
cold
screen
to
pair
program.
B
Debug
have
interactive
sprint
demos,
and
the
key
thing
about
us
is
that
we
make
any
meeting
collaborative
and
the
way
this
works
is
it's
a
mac,
os
and
windows
app,
and
you
can
share
any
window
on
your
desktop
with
a
single
click
by
just
hovering
over
the
window
and
clicking
on
share
on
the
share
window
tab
that
appears
above
the
window
when
you
do
that,
it
appears
for
all
fellow
team
members
and
they
can
instantly
interact
with
his
windows.
Okay
and.
A
A
D
D
You
know
that's
sort
of
like
where
I
first
got
into
video
conferencing
and
I,
and
actually
where
I
first
came
in
contact
with
your
guys's
project,
the
sip
communicator,
which
eventually
turned
into
jitsi
and
from
there
I
went
to
work
for
a
tele
health
startup
and
I
got
all
sorts
of
exposure
to
you
know
I
wrote
my
own
sfu.
I
wrote
my
own
mcu.
I
wrote
I
worked
with
lie
code.
D
I
worked
with
all
sorts
of
different
solutions
and
I
think,
like
the
most
stable
solution
over
the
years,
was
you
know
jitsi
right.
It
seems
like
you
guys
kind
of
went
built
in
webrtc
or
webrtc
support
things
like
that,
and
so,
like
I've
kind
of
you
know,
as
I
worked
through
my
career,
you
know
like
jitsi's,
always
kind
of
been
there
in
the
background
a
little
bit.
D
You
know
it
wasn't
there
in
the
very
beginning,
but,
like
you
know
like,
as
far
as
when
webrtc
started
becoming
out
there,
you
know
I
started
you
know
being
really
aware
and
getting
involved
in
jitsi.
I
went
to.
I
worked
at
higher
view
in
its
early
stages,
which
was
all
at
the
time
because
there
was
no
webrtc,
it
was
flash
based,
which
was
terrible
and
later
I
actually
after
the
telehealth
startup.
D
It
went
back
to
higher
view
and
helped
them
move
a
lot
of
their
infrastructure
onto
jitsi,
and
there
was
you
know
like
a
very
poor
light
code
implementation
there
and
things
like
that,
and
I
ended
up
just
I
think.
While
I
was
while
I
was
actually
at
the
telehealth
startup,
I
was
working
on
co-screen
behind
the
scenes,
so
I
went
and
found
the
ad
that
till
and
max
had
made,
and
at
ibond
years
and
years
before
I
had
made
like
a
system
called
imedia
which
was
like
for
just
for
hotels
right.
D
D
There
just
was
no
way
to
do
that,
and
so
I
made
this
application
streaming
platform
and
it
was
a
big
product
that
we
actually
sold,
because
then
you
could
go
into
these
hotels
that
had
all
this
hardware
already
and
then,
like
stream
applications
to
it.
I
was
like
you
know.
I
always
wondered
like
why.
Why
isn't
there
such
a
thing
for
the
internet?
Why
don't
people
have
a
way
to
like
virtualize
and
stream,
individual
applications
on
the
internet,
and
so
when
I
saw
what
you
know,
co-screen
was
doing.
D
D
Of
course
I
built
it
against
jitsi
on
the
back
end
and
and
we
got
a
lot
of
attention
and
we
got
funding
and
and
then
that
and
then
I
got
sucked
in
to
work
full
time
and
then
you
guys
at
this
very
serendipitously,
released
jazz,
which
was
great
for
us
because
building
all
of
that
infrastructure
in
the
back
end.
I've
done
it
before
it's
pain
in
the
butt
and
we
didn't
want
to
do
it
so
and
I
think
that's
kind
of
where
we're
at
today.
So
that's
how
I
got
here.
A
And
why
don't?
Why
didn't
you
guys
show
us
now
and
everyone
what
is
crow
screen,
because
I
I
think
it
is
very
important:
you're
you're,
that's
not
your
typical
screen
sharing
tool
by
any
stretch,
so
I'm
really
anxious
for
for
our
viewers
to
see
it
and
we'll
try
to
describe
it
in
words
for
those
who
listen,
yeah.
C
D
B
B
It's
an
animation,
but
it's
exactly
how
the
app
is,
as
you
know,
so,
if
you
run
cool
screen
on
your
mac
as
a
windows
computer,
you
boot
up,
you
can
join
the
same
cool
screen,
it's
like
a
jitsi
session
or
like
a
zoom
meeting
and
you
send
a
link
around
and
everyone
can
just
join.
I
can
go
and
call
people
and
you
see
them
on
the
top
right
of
your
desktop.
So
there's
video
chat
built
in
and
audio
chat,
but
these
are
just
features
for
us
just
adjacent
things.
B
That
is
not
the
core
of
the
experience.
The
core
of
the
experience
is
what
you
see
in
the
middle,
which
is
here
in
this
case
mary
jason
and
robert
each
share
one
window
from
the
desktop,
so
you
can
see
mary
shares
command
prompt.
If
you
look
carefully
it's
on
windows,
json
shares
a
chrome
winner
from
mac
os
robot
shares.
I
don't
know
what
vs
code
from
mac
os
in
the
way
they
do
this
to
just
move
their
mouse
over
the
window
just
to
to
share
the
window.
Only
that
is
transmitted.
B
It's
not
the
full
desktop.
It's
not
like.
You
know
other
remote
desktop
tools
where
you
always
give
access
to
the
full
thing,
but
you
only
give
access
to
specific
windows
and
we
only
allow
remote
control
of
these
particular
windows.
And
if
you
look
carefully
also,
you
can
see
that
jason
is
now
typing
to
mary's
window.
Robert
is
typing
into
jason's
window
scrolling
and
mary's
typing
into
robots
windows.
So
it's
multi-editing
across
multiple
users.
That's
in
its
essence.
B
What
code
screen
enables
very
fast
screen
sharing
instant,
you
know
getting
the
window
across
to
other
folks
recreating
these
windows
as
if
they
were
locals,
they
can
interact
with
windows
instantly
as
soon
as
they
are
shared
with
you
and
everyone
can
share
at
the
same
time,
and
then
there
are
some
edge
things
like
you
can
copy
and
paste
between
windows.
I
could
grab
another
piece
from
barry's
window
and
paste
it
into
robots
window,
which
is
especially
great
if
you're
lying
between
you
know,
designers
and
engineers.
B
D
So
I
I
mean,
I
think
it's
the
right
way
to
share
for
particular
circumstances
right
so
like
if
you
really
break
this
down,
like
the
you
know,
there's
several
things
that
we're
doing
different
than
the
typical
meeting
situation
right.
So
the
situation
we
have
right
now
is
we're
all
videos
we're
big
videos.
Everyone's
screen
is
currently
taken
over
and
they're,
focusing
on
everyone's
picture
right.
We
know
from
a
lot
of
research,
especially
during
the
pandemic,
that
this
is
something
that
has
led
to
you
know,
burnout
and
stuff.
You
know
this
is
one.
D
This
is
the
zoom
hell
that
everyone's
talking
about
everyone's
stuck
they're
staring
at
people's
pictures,
all
the
time,
they're
they're,
getting
tired
from
doing
it
and
there's
like
psychological
reasons.
Why
that's
the
case
and
we
really
built
co-screen
around
people
who
are
working
together
right
so
co-screen?
Isn't
that
water
cooler
conversation?
It's
not
for
this,
like
direct
communication,
that
you're
doing
all
the
time
and
when
you
come
at
it
from
that
perspective
of
okay.
These
are,
these
are
people
that
are
maybe
sitting
in
this
room
all
day
and
they're
working
together.
D
D
It
would
be
like
if
you
were
in
the
same
room
with
someone
and
they're
sitting
there
over
your
shoulder
and
you
share
a
window
with
them
right
and
so
having
it
so
that
you
can
give
remote
control
implicitly
and
then
you
can
disable
it
explicitly
right.
That
seems
you
know.
That's
a
time,
saver
right!
That's
something
that
people
you
know
if
you're,
if
you're
having
to
come
in
every
time,
ask
okay!
I
want
to
share
this
window
with
you.
D
Okay
is
everyone
seeing
it
and
then
somebody's
asking
well,
can
I
have
access
to
it
and
you
know
like
oh
joe's
got
access
to
it,
but
not
mary
right.
All
of
these
things
cost
people
time
right.
So
so,
there's
there's
multiple
reasons
why,
when
you're
working
together
the
way
you'd
want
to
share
things
would
be
very
different
right,
which
has
which
led
us
to
this
very
different
paradigm.
Right,
which
is
people
can
share
whatever
they
want
whenever
they
want.
D
That
means
you
can
take
two
pieces
of
content,
put
them
next
to
each
other
and
compare
them
which
for
like,
if
you're
doing,
incident,
response
right
like
right
now
with
incident
response,
if
you're
using
remote
collaboration
tools,
you
share
your
a
window.
So
let's
say
it's
splunk
right.
Somebody
else
has
logs
open.
It's
like!
D
Oh
now,
joe
needs
to
share
his
logs
and
there's
no
way
to
compare
those
two
pieces
of
information
together
right,
whereas
if
you
guys,
you
know,
I've
been
in
war
rooms,
I've
been
in
situations
where
you
have
a
devops
team
and
you're
all
trying
to
maintain
some
sort
of
or
you're
trying
to
figure
out
okay.
What
happened
with
this
problem,
so
everyone
all
rushes
around
a
single
laptop
somebody
else
bringing
their
laptop
in
and
shoving
it
in
out
of
a
corner
of
the
eye
kind
of
saying:
hey,
look
at
this
log
here.
D
I
found
this
line
and
co-screen
kind
of
lets.
You
do
those
sorts
of
things
remotely,
because
now
you
can
just
without
even
asking
okay,
here's
the
log,
here's
the
splunk,
you
can
move
the
windows,
look
at
them
next
to
each
other.
You
can
move
your
mouse
over
the
sections
of
the
log
that
you're
talking
about,
and
you
see
you
get
that
context
of
what
they're
talking
about.
A
Right:
okay!
Well,
you
know
when,
when
when
people
talk
about
startups
or
or-
and
I
think
business
schools
like
to
take
that
approach
a
lot-
they
they
often
try
to
frame
a
new
product
in
in
in
this
manner
of
speaking
like
comparing
it
to
a
reference
offer
and
specifically
the
reference
software.
Is
you
know
this
problem
that
you're
solving
right
now
before
you
existed?
A
B
A
So
would
you
say
that
this
is
like
you
know
a
decent
reference,
for
you
know
this
is
this
is
this
is
this
is
what
co-screen
is
trying
to
make
better?
For
example,
that's
one
scenario
where
you're
juggling
cables,
maybe
you're,
not
even
in
the
same
room
right,
that's
a
big!
That's
a
big
deal.
That's.
B
A
B
You
can
share
from
three
people.
Windows
from
three
people
on
the
same
joint
workspace
does
not
exist
today.
You
have
some
tools
there,
where
you
can
share
multiple
desktops
from
different
people,
but
they
appear
in
different
black
boxes
and
they're,
not
kind
of
cool.
You
can
join
on
the
same
workspace
and
you
can't
interact
with
them.
So
there
is
no
tool
that
merges
the
experience
of
multiple
team
members
in
one
there's,
always
this
handover,
you
say:
okay,
no,
you
it's
your
turn.
I
unshare
you
share,
or
you
know.
A
D
D
And
that's
you
know
like
you
know,
we
have
a
solution
that
can
deal
with
like
what
people
might
use
a
kvm
switch
for
right
or
what
they
might
be
juggling
around
hdmi
cables
for,
or
you
know
like
for
for
me
like
when
I'm
using
you
know,
we
use
code
screen
every
day,
we're
using
pro
screen
to
develop
code
screen
and
whenever
I'm
interacting
with
another
engineer
right,
it's
a
lot
like
the
to
be
honest,
the
biggest
inflection
point
is
that
whole.
You
know.
D
You
know
some
place
in
the
code
that
I'm
working
on
and
explain
it
or
take
them
to
a
you
know
a
portal
where
we're
you
know
aws
or
something
and
show
them
like
where
I'm
registering
a
server,
and
so
it's
like
we're
doing
a
lot
of
those
moments.
That
would
be
this
in
office
work
collaboration,
but
using
co-screen
right.
So
it's
like
co-screen
is
kind
of
this.
It
is
a
collaboration
tool.
It's
how
we
build
it
right.
D
D
I
mean
I
mean
I
I
I
don't.
The
way
I
feel
is
like
you
know,
remote
communication
is
never
going
to
be
as
good
as
in
person
communication.
You
know
like
I
am
you
know,
I'm
having
this
conversation
with
you,
but
there's
there's
always
this
kind
of
layer
of
you
know
where
you
have
that
human
physical
connection
so
like
co-screen's,
never
going
to
replace
that
right.
D
But
in
terms
of
like
do
I
have
to
go
hunch
over
my
my
friend,
you
know
or
my
co-workers
keyboard
and
he
happens
to
have
you
know
I
use
this.
You
know
this
advantage
too
and
I
have
a
very
different
keyboard
layout
and
I
can't
figure
out
his
keyboard
layout
and
stuff,
and
I
can't
really
use
his
window
or
I
can't
work
on
anything.
That's
on
his
computer
right
like
there's.
It's
obviously
much
better
in
terms
of
that
right.
D
So
there's
you
know,
I
I
think
we
all
know
that
there's
trade-offs
to
remote
work,
you
know,
and
it's
you
know
I
could.
I
I'm
sure
I
can
make
a
huge
list
of
advantages
and
disadvantages,
like
everyone
can
now,
because
everyone's
been
working
remotely
for
some
time,
but
you
know,
I
think
that,
in
terms
of
like
actually
offering
that
you
know
ability
to
just
ad
hoc,
you
know
share
something
like
you
can
in
reality.
Co-Screen
does
that
better
than
in
person
right
so.
C
So
after
you
guys
have
gone
public
beta
because
from
what
you
say
it
like
it
sounds
to
me
that
it
could
definitely
help
developers,
because
we
mentioned
this
use
case
for
sharing
code
editors
war
rooms
logs.
Would
you
say,
or
are
you
willing
to
share
this
with
us,
that
development
or
like
devops
kind
of
scenarios,
are
like
the
main
use
case?
B
No,
no,
I
mean,
I
would
say,
ninety
percent
of
our
users
are
engineers.
That
is
true.
It's
just
our
the
wedged
into
orgs.
It
is
the
bridgehead
of
you
know
how
you
land.
First,
it's
individual
engineers,
often
that
find
ourselves.
Many
of
them
are
pair
programmers
that
you
know
write
code
always
together.
Many
of
them
only
write
code.
B
Where
you
have
a
senior
engineer
and
junior
engineer
senior
engineers
online,
all
the
time
kind
of
office
hours
student
engineers
can
just
drop
in
share
where
they
struggle
the
senior
engineer
shares
you
know
what
the
solution
is
and
they
align
very
quickly.
So
it's
all
these
use
cases
that
jason
mentioned
earlier.
They
do.
We
do
think
we
do
see
and
that's
the
exciting
part
lots
of
use
cases
that
we
haven't
thought
about.
We
always
knew
that
code
screen
is
useful
for
a
thousand
different
things.
B
Competition
is
stiff
and
they
think
you
know
this
could
be
a
game
changer
for
them
because
they
can
just
be
all
on
the
same.
On
the
same
page,
all
the
time
you
know
all
hands
on
deck,
one
can
share
the
latest
list
of
you
know
whatever
restaurants
to
go
after
another.
One
can
just
share
the
latest
outreach
and
there's
no
more
debating
of
hey.
Can
you
do
this
link?
No
go
page
back.
Can
you
go
to
slack?
I
just
sent
you
something
open
this
link,
no,
no,
the
other
one
and
so
on.
B
D
Is
actually
a
really
small
number
of
our
users?
I
just
wanted
to
put
that
out
there
like.
We
have
had
users
that
have
gotten
co-screen
and
approached
us
and
used
it
for
specifically
that
purpose.
You
know,
that's
mostly
like
what
we're
using
it
for
internally
right
is
incident
response,
mentoring,
onboarding
and
you
know,
like
I
think,
the
most
surprising
use
case
that
I've
seen
is
like
we
have.
D
We
had
someone
like
make
this
really
cool
video
for
accounting
tools
right,
and
it
makes
a
lot
of
sense
so
apparently,
like
some
group
of
accountants
found
us
they're
using
us
for
sharing
like
excel
and
like
going
through
and,
like
you
know,
doing
doing
a
lot
of
the
stuff
that
they
used
to
collaborate
on
in
office
right.
My
wife
is
a
accountant
right
and
so,
like
I've,
I've
seen
her.
D
You
know
at
goldman
sachs.
She
had
like
some
bizarre,
like
terminal
mainframe
things
she
had
to
log
into,
and
it's
like.
How
do
you
even
share
that
these
days
you
know
but
like
with
co-screen?
They
could
just
like
share
their
window
and
stuff.
So
it
was
a
you
know.
It's
definitely
not
a
huge
use
case.
Most
of
our
users
are
developers,
but
you
know
I
could
see
you
know
I
could
definitely
see
utility
in
other
areas.
It's
not
how
we're
marketing
it
they're,
not
our
target
customers,
but
you
know
right.
C
I
was
thinking
that
I
have.
I
have
used
other
tools
in
the
past
to,
for
instance,
you
know
help
a
relative
help.
My
brother
help
my
father
fix
a
problem
like
requiring
this
remote
control,
for
almost
anything
is
a
thing
that
you
need
to
have,
and
I
think
the
last
one
I've
used
was
like
teamviewer
or
something,
but
I
don't
know
if
it
does
now,
but
it
didn't.
You
know
it
doesn't,
have
these
video
things.
C
C
You
know,
of
course,
is
a
standalone
application,
but
you
you
I
mean
in
my
mind,
you
need
to
get
into
call
screen
from
somewhere
like
you're
already
chatting
in
slack
or
you
are
in
some
other
business
or
personal
communication
means
that
you
didn't
escalate
into
into
this
this
kind
of
session.
So,
do
you
see
yourselves
going
into
like
maybe
deeper
integration
with
with,
like
an
ide,
for
example,
built
writing
to
visual
studio
code
or
intellij?
D
We
recently
released
in
the
setup,
so
people
that
are
using,
like
the
teams
product
from
setup
that
they
already
have
it
for
their
business.
They
can
all
use
co-screen
yeah,
but
I
think
integrations
is
a
big
part
of
our
strategy.
You
know,
I
don't
think
we
want
to
be.
We
don't
want
to
be
slack.
You
know
we
don't
want
to
like
in
some
ways
you
know
like
if
you
look
at
our
user
interface,
it'll
be
very
natural
for
us
to
go
back
to
that
old
kind
of
skype
or
icq
or
whatever
style.
D
Appreciate
you
guys
saying
that
you
know
because,
like
I
think
you
know,
people
want
to
use
slack
and
they
want
to
use
the
tools
that
they
already
have
for
accomplishing.
That
and
like
there's,
been
very
few
tools
that
I've
seen
that
have
done
a
good
job
at
that
so
like
for
us.
It
would
just
be
an
a
distraction
or
engineering
effort
and
so
like.
We
are
really
focusing
like.
A
Well,
it's
going
to
be
now
and
if
that's
the
case
is
everyone
available,
is
it
going
to
be
later
and
that's
the
case?
How
are
we
going
to
have
no
scheduling
and
it
always
becomes
a
very
people
routinely
underestimate.
The
complexity
of
this
problem
is
that
is
that
something
that
you've
seen
as
well,
that
that
you
know
getting
people
to
meet,
isn't
actually
a
trivial
issue?
I
mean.
D
A
D
B
And
also,
and
also
to
adopt
them
to
adopt
a
new
workflow
right.
What
we
do
is
different,
and
I
think
that
is
that
is
a
challenge
that
we've
seen
right.
It's
there
are
folks,
like,
I
would
say,
process
nerds,
like
us,
productivity
nerds
that
enjoy
the
new
way
of
doing
things,
and
they
can
ideate
about
what
the
what
cool
things
they
could
do
on
code
screen.
B
But
there
are
also
many
folks
that
just
can't
see
that
just
yet
and
they
have
to
be
convinced,
then
you
have
to
guide
them
through
the
process,
so
we
actually
launched
in
the
onboarding
process
where
the
first
time
you
join
a
co-screen
there's
a
bot
that
joins
you
if
you're
alone.
So
it's
it's
it's
in
that
case
smart.
If
there's
no
other
person,
it's
then
it
joins
it
force
brought
upon
you
that
collaborates
with
you.
B
You
know
she
asks
you
to
share
a
window
interacts
with
you
play,
stick
lecture
with
you,
so
you
get
the
whole
experience,
even
if
you're
alone,
because
we
know
by
based
on
our
own
stats
that
the
five
majority
of
users,
either
connect
alone
or
connect,
are
at
least
alone
for
the
first
two
minutes,
because
they
never
join
at
the
same
moment
and
now
we're
using
that
time
to
teach
them
our
goal
screen
works
so
as
soon
as
they're.
You
know
the
second
team
member
joins.
They
have
learned
how
the
tool
works.
B
They
got
how
it's
different
and
you
know
we
can
see
our
activation
goes
up
and
that's
something
we
will
continue
to
invest
to
you
know
smoothly.
On-Ramp
integrations
are
one
piece
is
this
is
kind
of
distribution
and
awareness,
then,
once
there
in
a
session
teaching
them
how
to
use
that
getting
them
inspired
about
what
else
they
could
do
with
this
and
then,
of
course,
getting
them
back
to
different
means,
and
these
are
now
kind
of
the
day
two
problems
after
we
built
the
foundation
over
the
first
one
and
half
years
I
mean.
D
C
D
D
C
D
It
this
way,
sometimes
it's
not
really
how
we
designed
it
right,
it's
kind
of
more
like
what
teamviewer
is
doing
and
stuff
we've
had
a
lot
of
people.
Talk
about
that
and,
like
I
mean
it's
definitely
possible
you
just
you
know
I
have
a.
I
have
a
mac
mini.
I
leave
code
screen
running
on
it.
People
join
into
the
session,
I
have
the
desktop
always
sharing
and
it
works
that
way
right.
D
B
Maybe
just
add,
you
know
that
it
was
funny
going
back
to
use
cases,
there's
a
researcher
that
I
think
works
on
the
lab
with
lasers.
I
don't
know
what
it's
a
big
installation
of
different
of
like
a
big
range
of
different
devices,
very
complex,
very
expensive,
four
different
computers
along
that
test
test
bench,
and
then
he
installs
cool
screen
on
all
these
four
devices
and
on
the
center
computer.
He
just
gets
one
window
from
each
computer
and
you
can
see
that
he
sees
the
whole
setup
at
the
glance.
B
You
know
that
wasn't
something
we
built,
but
it's
true
that
you
know
this
is
now
possible
right
now
and
you
would
have
had
to
find
other
ways
to
kind
of
share
four
desktops
at
the
same
time
and
kind
of
squeeze
them
together
on
this
fifth
computer
to
see
what
that
is
going
on.
But
now
we
can
just
say:
okay,
this
window
from
this
computer,
this
window
from
this
computer.
You
arrange
them
and
you
can
see
instantly
okay.
What
are
the
signals
at
each
of
these
steps
of
his
of
his
lab
set
up?
B
D
I
think
so
just
going
to
like
some
of
the
struggles
with
retention
and
stuff,
I
think
one
of
the
the
big
things
is.
You
know
I
think
in
terms
of
like
what
what
we
have
to
accomplish.
You
know.
Latency
is
a
really
big
deal
for
us.
You
know,
because
our
tool
is,
you
know
very
much,
it's
not
just
the
quality
of
the
stream
that's
coming
through,
but
it's
also
like
how
you
know
how
much
latency
is
there
right,
because
the
illusion
gets
destroyed
very
quickly
if
there's
any
typing
latency?
D
If
there's
you
know
it's
like,
even
if
you
just
get
down
to
the
you
know,
we'll
have
people
that
have
just
the
poor
experience
and
you
go
look
at
like.
What's
the
rtt
for
their
connection
and
it'll,
you
know
200
milliseconds
or
something
right
and
you'll
just
be
like
okay,
you
guys
aren't
going
to
have
a
very
good
time
with
that,
because
you
know
your
base,
rtt
is,
is
quite
high
right
and
and
then
also
like,
I
mean
all
the
same
struggles.
D
I've
had
it
with
other
places,
with
just
general
video
conferencing,
packet
loss
and
people
that
are
too
far
from
their
wi-fi
and
all
these
other
things
you
know,
and
it's
like
you
know,
I
think,
we're
we're
trying
to
build
a
good
support
and
triage
process
for
dealing
with
that.
But,
like
you
know,
if
you
go
look
at
like
stadia
stadia's,
docs
right
there,
like,
basically,
you
have
to
have
10
megabits
up
and
down.
You
have
to
have
this
latency
and
they're
like
very
draconian
about
their
network
requirements.
D
You
know:
we've
built
coast
green,
to
have
much
lower
requirements
in
terms
of
like
what
what
network
requirements
are
there,
and
you
know,
I
think
you
know-
we've-
we've
done
a
pretty
good
job
at
it
in
terms
of
like
what
what
we
provide
but
like.
It's
definitely
going
to
be
a
long
term
struggle
for
our
product.
You
know
like
working
on.
How
do
we
deal
with
that?
How
do
we,
how
do
we
educate
customers
about
it?
D
We're
also
like
dealing
starting
to
get
like
tons
of
inbound
traffic
from
banks
and,
like
I
worked
at
a
company
called
bvs
performance
solutions
that
built
you
know.
I
built
a
video
conferencing
product
for
them.
All
of
our
clients
were
banks,
and
it
was
a
huge
struggle
because
they
just
have
these
absolutely
draconian
horrible
firewalls
that
do
all
sorts
of
janky
things
that
you
know
like
you
know.
We
had
to
basically
make
extremely
small
sets
of
endpoints
for
them
to
whitelist
and
connect
to
and
all
this
stuff
it
was
a
a
big
pain.
D
You
know,
and
it's
you
know,
I
I
think
you
know
I'm
trying
you
know.
I
think
that
was
our
strategy
right
was
to
try
to
like
keep
it
simple
stupid,
but
it's
probably
more
like
we
need
to
get
better
at
understanding
like
what
is
it
that
their
security
people
actually
want
in
terms
of
these
things
you
know,
but
that's,
I
think,
that's
like
in
terms
of
growing
pains,
that's
a
big
big
part
of
it
right
now,.
C
I
was
muted
oopsie.
No,
I
was
sorry
okay,
so
you
did
mention
that
you
guys
have
also
you
know
struggled
with
networking
stuff
and,
and
you
did
touch
a
bit
on
what
you've
built.
I
was
wondering
if
we
can
dive
a
little
bit
into
the
tech
side
of
of
things,
so
you
did
mention
that
you
guys
are
using
jazz
gt
for
the
video
conferencing
side
of
it.
But
then
you
know
is
the
so.
I'm
gonna
throw
you
like
a
a
random
bunch
of
questions
that
popped
in
my
head.
D
Majority
of
the
actual
streaming
stuff
is
done,
natively
and
it's
mostly
written
in
c
plus,
and
we,
you
know,
we
actually
have
a
libweb
rtc
and
a
wrapper,
and
we
take
the
electron
windows
and
we
grab
the
eight
wind
on
windows.
We
grab
a
handle
to
it
to
the
window
and
we
inject
an
opengl
surface
on
on
a
mac
and
we
inject
a
direct
3d
surface
on
windows
and
we
render
all
of
our
content
directly
into
the
window,
and
we
do
all
of
our
slicing
and
stream
composition.
There.
D
It's
a
terror.
It's
not
actually
like.
It's
not
really
an
insertable
streams,
api
that
you
know,
chrome.
That
was
a
big
thing
that
they
you
know.
I
hate
that
they
do
this
with
library
tc,
where
they,
you
know,
they
have
like
a
webrtc
api
like
concertable
streams,
and
then
they
go
and
basically
put
the
actual
implementation
of
the
api
into
chromium,
and
so
it's
not
even
there.
So
they
have
basically
like
a
very
rudimentary
insertable
streams.
C
C
Not
but
don't
get
me
wrong,
I
meant,
from
the
perspective
of
you
know,
we
provide
a
javascript
api
for
gta's
last
jazz,
so
I
was
thinking
yeah
at
least
this
window,
where
I
can
see
myself.
This
is
most
likely
yeah
electron,
but
the
rest
definitely
doesn't
feel
like
collecting
yeah.
It
meant
more,
like
the
the,
because
I've
seen
these
kinds
of
applications
that
they
use
electron
for
a
little
bit
of
the
of
the
stuff
and
then
and
then
they
venture
into
other
things.
I
think.
D
D
I
may
I
mean
maybe
we
would
have
made
a
very
different
decision
and
done
flutter
or
done
something
different.
Had
we
approached
it
today,
you
know
I
would
be
qt
or
something
like
we
like
the
the
truth.
Is
that
electrons
actually
been
somewhat
painful.
You
know,
there's
there's
a
lot
of
things
about
electron
that
I
really
don't
like
and
there's
a
lot
of
things
I
really
do
like
too,
but
like
they're,
you
know,
like
the
amount
of
engineering
resources
that
have
gone
into
debugging.
D
C
B
C
Very
often
so
yeah
there
you
go
when
you
mention
stadia
and
their
draconian
requirements.
The
thing
that
comes
to
mind,
when
I
think
of
stadia
is,
is
that
they
were
very
you
know,
they're
very
strict
on.
They
were
thinking
a
lot
about
the
codec
they
use
because
of
the
quality
that
they
get
back
in
a
video
game.
You
cannot
really
tolerate
the
blurriness
and,
and
we
have
battle
with
blurriness
ourselves
in
screen
sharing.
I
don't
know
if
you
have
any
stories
to
share
about
codex
that
you
guys
have
tried
discard
it.
D
I
mean
we,
we've
tried
using
h.264
multiple
times,
but
there's
incompatibilities
with
the
video
bridge
that
you
know.
I
think
we've
quickly
learned
that
trying
to
swim
upstream
there
was
a
was
a
bad
idea.
You
know
in
the
past
for
these
types
of
things
you
know,
like
I
mean
one
of
our,
you
know,
I'm
sure
everyone
here
is
familiar
with
screen
hero
right
screen
hero.
D
You
know
they
ended
up
using
like
libx
264
and
doing
a
lot
of
optimization
down
on
the
low
level
for
trying
to
reduce
latency
and
stuff.
In
my
opinion,
vp8
is
not
very
great
at
doing
this.
You
know,
I
think
they've
gotten
a
lot
better
in
terms
of
like
low
latency
and
such
there
was
a
period
of
time
where
you
know
it
was
very
hard
to
tweak
things.
There's
certainly
a
lot
less
knobs
and
levers
to
pull
in
vp8
than
there
are
in
h.264
yeah
this
the
blurriness
issue.
D
D
You
know
high
frame
rate
content,
low
latency
and
I
know
there's
like
I
think
it's
proton
or
something
right,
there's
other
companies
that
are
building
that
but,
like
you
know,
I'd
love
to
be
super
focused
on
on
the
codec.
You
know:
we've
done
a
bunch
of
things
to
try
to
tweak
vp8
as
well
as
we
can.
We
did
a
bunch
of
work
with
h.264,
but
we
haven't
released
it
because
of
you
know,
there's
a
lot
of
complexity
behind
it
in
terms
of
like
our
infrastructure
and
everything.
D
I
think
you
know
it's
like
so
far
like
we're,
trying
not
to
swim
upstream
or
trying
to
get
the
best
to
do
the
best
we
can
with
vp8.
And
you
know
we
have
a
host
of
tweaks
that
we've
made
in
libwebrtc
in
order
to
try
to
produce
better
lower
latency
and
to
you
know,
we
still
have
blurring
when
the
scene
changes
are
very
rapid
and
stuff,
but
we've
done
some
tweaks
to
try
to
improve
that.
D
I'm
not
sure
if
we're
doing
as
good
as
like,
for
instance,
screen
hero
was
doing
with
libex
264.
That's
definitely
something
I
want
to
do
at
some
point,
but
you
know
I
think
the
vp9
support
that
you
guys
are
having
in
the
bridge
and
stuff
like.
I
think
that
if
we're
gonna
be
working
in
the
future,
we'll
try
to
make
use
of
that
as
much
as
possible.
D
D
Yeah
and
that's
that's
actually
like
part
of
our
secret
sauce
right
is
this
whole.
I
mean
that's,
that's
one
of
the
things
you
can
do
with
co-screen
that
that
nobody
else
can
do
right,
which
is
that
we
actually
consider
the
native
resolution
of
the
windows
and
we
try
to
preserve
that
ratio,
and
we
have
an
algorithm
for
actually
determining
okay.
This
screen
is
just
way
too
small
to
actually
preserve
that
ratio,
but
you
can
always
scale
it.
D
You
can
scale
it
to
that
same
ratio
eventually-
and
you
know,
we've
kind
of
come
up
with
a
compromise
of
how
we
end
up,
showing
those
windows
originally
to
try
to
make
them
look
the
best
in
some
screens
like
if
somebody's
coming
from
a
huge
4k
screen
and
they're
coming
to
a
really
low
resolution
screen,
it's
just
going
to
show
up
tiny,
but
then
you
can
scale
it
to
where
it
looks
good
for
you
and
it's.
I
think
it's
a
pretty
good
compromise
that
we've
made
so
far.
D
Our
original
paradigm
was
like
all
the
windows
I'd
actually
like
track
and
stay
together,
and
when
the
remote
person
moved
it
would
move,
and
then
everyone
had
this
kind
of
shared
desktop.
We
moved
away
from
that
to
individual
windows
and
in
the
future,
we'll
probably
move
into
actually
like
composing
the
contents
of
the
windows
in
the
stream,
totally
independent
of
each
other
and
independent
of
the
remote
position,
and
I
think
that
gives
us
a
lot
of
leeway
in
terms
of
like
you
know.
D
We
could
actually
like
if
the
window's
in
the
background
and
it's
obscured,
we
could
scale
it
down
a
lot,
so
we're
not
sending
as
much
data
there.
So
we
help
preserve
some
of
their
bandwidth
and
things
like
that.
There's
a
lot
of
leeway
in
terms
of
how
we
send
windows,
based
on
the
actual
context
of
them
being
used
that
isn't
really
available
in
a
standard
screen
sharing
scenario.
In
order
to
try
to
make
it
look
better
to
the
end
user.
You
know
so
it's
kind
of
exciting.
D
Might
you
know,
would
give
us
a
lot
more
tools
to
be
able
to
to
really
customize
co-screen
in
a
way
I
get
that
it's
not
really
the
core
focus
of
of
jitsi
right.
It's
maybe
not
even
super
helpful
to
jitsi
and
the
tools
that
are
there
already.
Actually,
we've
made
good
use
of
it
because,
like
the
paradigm,
we're
using
right
now
is
whoever
the
current
focused
user
is
is
the
on
stage
user
and
stuff.
D
A
C
C
D
D
Mostly
when
we've
had
any
issues,
it's
been,
you
know
due
to
a
change
in
the
infrastructure
and
usually
us
just
being
unaware
of
some
of
the
consequences
of
those
changes,
and
so
you
know,
I
think,
we've
done
a
lot
of
work
to
try
to
like
you
know
we
have
like
an
environment,
so
we
have
our
own
jitsi
environment
right.
We
have
a
server,
that's
on
our
last
stable
version.
We
have
a
server,
that's
on
on
the
most
current
version.
D
We
test
against
the
most
current
version
and
and
then
try
to
to
detect
any
breaking
changes
before
they
end
up
going
into
production,
which
there
tends
to
be
a
pretty
long
runway
before
that
happens,
and
then
and
then,
when
something.
If
something
does
happen,
then
we
try
to
bisect
to
find
between
the
two
servers
to
find
like
where,
where
did
this
happen?
D
Go
and
think
about
it
and
come
up
with
a
solution,
and
so
far
that's
that's
worked
pretty
well
when,
when
something
does
go
south,
it's
it's
a
pretty
harrowing
thing,
especially
if
it's
in
production
we're
just
like
up
all
night
and
stuff.
It's
definitely
not
the
funnest
thing
in
the
world
to
do
so.
We're
really
trying
to
make
that
work
happen
before
it
gets
there,
because
the
the
more
we
can
do
that,
the
more
we
can
actually
like
space
out
the
engineering
effort,
that's
required
to
to
reconcile
those
those
sorts
of
problems.
A
A
B
It's
an
interesting
question.
I
think
for
sure
could
have
happened
technically.
You
know,
if
you
might,
you
might
know
the
teamviewer
and
any
desk
most
remote
desktop
tools.
The
most
popular
ones,
as
far
as
I
know,
are
based
in
germany,
built
in
by
german
teams
in
germany.
So
there's
nothing
against.
You
know
their
great
italians
over
there.
The
journey
that
I
think
is
is
that
that
was
ours
wouldn't
have
been
possible.
B
Like
that,
it's
more,
you
know,
I
mean,
I
said
everything
makes
sense,
but
the
way
we
we
came
here
that
wasn't,
I
didn't
come
here
to
to
build
a
startup.
I
came
here
because
we
found
an
exciting
position.
Our
wi-fi
decided
we
didn't
want
to
go
back
to
germany.
We
were
in
switzerland
at
the
time
and
rather
spend
some
time
in
the
us.
At
least
a
couple
of
years
came
here.
B
A
B
I
don't
think
so,
especially
in
the
west
coast.
The
spirit
is
different
right
folks,
folks,
it's
it's
not
always
this
thing
about
yeah
folks
are
excited
about
failure
here,
or
value
failure
and
silicon
valley.
I
don't
know
if
that's
the
thing,
it's
just
more
common,
that
you
do
crazy
things
and
that
you
do
do
leave
very
cushy,
cozy
jobs.
They
call
it.
The
gravy
train
here,
you
know,
worked
for
google
for
a
while
had
a
great
job
enjoying
my
my
life.
At
the
same
time,
it
was
obvious
for
me.
B
I
wouldn't
stay
for
much
longer,
because
I
want
to
build
something
on
my
own
with
a
tiny
team
just
like
jase-
and
I
have
done
it
over
the
past
two
years
now
and
that's
that's
something
that
is
very
uncommon
in
germany.
Unfortunately,
I
think
in
europe.
Overall,
it's
catching
up.
Berlin
is
a
great
hot
spot
for
startups,
but
it's
still
way
too
rare.
There
are
way
too
few
folks
that
build,
try,
try
things
and
that's
that's
truly
unique
to
to
the
area,
and
you
know,
hopefully,
it's
changing
globally.
A
B
Yeah
right
I
mean,
of
course
you
know
now
you
can
raise
from
anywhere
on
zoom
and
so
on,
but
the
truth
is
we've
raised
from.
We
see
that
it's
literally
600
feet
away
from
where
I
am
right
now
there
is
a
benefit
of
being
in
the
space
and
having
being
able
to
meet
up
and
have
some
connections,
join
connections
and
like,
if
you
think
about
it
after
we've
raised
and
talked
event
investors
many
times
over,
you
can
see
it's
a
very
small
ecosystem.
It's
not
truly
global.