►
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
B
The
current
focus
of
the
project
is
to
make
it
easier
for
data
scientists
to
build
predictive
models,
and
we
have
some
ideas
about
how
to
do
it,
but
we
also
want
to
learn
more
about
how
to
commercialize
it.
We
think
if
you
want
to
provide
a
nice
product
with
a
great
user
experience,
we
also
need
to
commercialize
it
to
be
able
to
have
funding
for
attracting
good
people.
So
we
looked
at
github
and
we
thought
they
know
what
they're
doing
apparently
they're
doing
work
quite
well
and
I
would
love
to
ask
a
few
questions.
B
You
said
about
the
open
core
model
and
how
to
make
it
work
for
sure.
So,
let's
jump
into
the
first
question,
which
is
how
do
you
decide?
Actually,
if
you
are
an
open
core
project
which
features
you
make
available
for
free
and
which
parts
do
you
leave
out
the
open
core,
the
open
source
offering
great.
A
Question
this
was
really
hard
for
us.
We
did
many
many
things.
First,
we
did
like
stages
and
then
we
did
buy
maturity
of
the
feature
set.
We
did
buy
scale
of
the
company,
all
of
them
didn't
work
and
we
settled
on
something
that
we
call
buyer
based
open
core
depending
on
who
the
buyer
is
and
what
features
they
care
about.
We
monetize
them
or
not.
A
If
the
buyer
is
an
individual
contributor,
if
the
individual
contributor
cares
a
lot
about
the
features,
they're
open
source
and
then
if
it's
a
manager,
a
director
or
an
executive
depending
on
that
they
get
into
one
of
our
pricing
plans.
So,
for
example,
exactly
they're
scared
about
company-wide
governance,
so
company-wide
governance,
dashboard,
where
you
can
see
of
every
project
like,
are
they
running
all
the
tests?
Are
they
running
all
the
security
tests?
Are
things
signed
off?
A
That
is
an
executive
and
execute
cares
more
most
about
that
feature
see
so
in
a
company
for
example,
and
then
the
entire
company,
every
user
in
the
company
has
to
pay
the
highest
price
for
gitlab,
because
they
they
are
the
people
who
can
decide
to
spend
more
and
by
aligning
it
the
pricing
and
the
features
with
the
buyer.
We
found
that
there's
the
least
friction
there's
still
friction
and
there
still
debate
about
who
cares
most
about
feature,
but
it's
been
better
than
any
other
model.
We
had
it.
A
For
sure,
when
we
started
get
lab,
it
was
completely
open-source
when
we
joined
YC,
one
percent
of
the
users
of
the
common
users
were
paying
and
now
about.
10
percent
of
the
organization's
are
paying.
We
have
over
a
hundred
thousand
organizations
using
get
lamp
and
about
ten
thousands
of
them
pay.
That's.
B
B
A
Know,
that's
not
been
our
experience
like
we
tried
for
a
while
to
kind
of
monetize
for
the
bigger
companies
we
we
found
no
really
significant
difference
in
the
features
they
wanted.
Even
the
small
companies
wanted
to
ship
secure
code.
Even
so
small
companies
needed
certain
approvals.
So
that's
not
been.
Our
experience.
I
think
it's
very
hard
to
run
a
hybrid
go-to-market
motion
so
having
a
sales
organization
that
serves
both
the
smaller
organizations
SMB
and
the
enterprise,
for
example,
enterprise
sales
reps
really
need
to
speak
with
the
customer.
A
One
way
to
do
that
is
to
not
display
your
price,
because
then
the
customers
get
in
touch
and
you
get
that
opportunity
SMB.
You
cannot
afford
to
have
a
rep
talk
with
them.
So,
like
you
have
this
struggle
like?
What
do
we
do
now?
Caleb
is
easy.
We
are
transparent
company.
So
always
this
our
pricing,
but
other
companies
really
have
to
choose
at
gitlab.
It's
now
SMB
is
self-serve
and
we
answer
questions
it.
A
Market
is
inside
sales,
so
we
do
reach
out
and
we
have
people
reaching
out,
but
they
do
not
visit
the
customer
and
then
the
top
of
the
market
is
enterprise
with
this
field,
Salesforce
supported
by
solution,
architects
and
that's
72%
of
our
revenue.
So
what
we
could
could
have
done
is
just
focus
exclusively
on
that
wouldn't
make
a
giant
difference
in
our
revenue,
but
we
think
it's
important
to
serve
every
part
of
the
market,
because
otherwise
your
future
competitors
will
come
or
first
serve
SMB
and
then
move
up.
Yeah.
B
Well,
actually,
we
are
thinking
of
competing
with
some
of
our
competitors,
because
we're
definitely
not
alone
in
the
machine
learning
space,
as
you
might
imagine,
yep
that
we
will
compete
with
some
of
them
by
moving
on
the
SMB
market,
because
we
think
there's
a
lot
left
desired
there.
So
that
will
be
our
market
approach
and
I.
Think
you.
B
A
B
B
A
B
I
think
it's
really
incredible,
so
we
want
to
be
based
in
Europe
and
we
also
want
to
sell
to
North
America,
for
example,
but
we
believe
that
there
are
extremely
good
people
all
over
the
world.
We
want
to
have
access
to
those
people,
but
we
are
not
certain.
How
should
you
deal
with
the
early
stage
of
the
company?
Would
you
stand
still
recommend?
You
have
a
very
spread
up
team
from
the
very
beginning,
or
do
you
think
it
doesn't
matter.
A
I
think
time
zones
are
heart,
this
heart.
If
you
have
little
overlap
in
your
day
and
it
slows
you
down
and
in
the
beginning
you
tend
to
have
more
adjustments
to
the
course
of
the
company.
So
in
the
beginning
we
were
very
European
at
three
months
during
by
sea,
we
were
like
in
the
same
building,
sleeping
and
working.
So
that's
what
we
did
so
I
do
think
you
feel
like
a
seven
person
team,
there's
nothing
better
than
being
in
the
same
room
together,
yeah.
B
A
B
B
Think
that's
so
interesting
because
I
was
discussing
with
someone
the
other
day
about
the
difficulty
of
having
remote
teams
and
then
I
said
to
them.
Well,
I
worked
in
a
200-person
organization.
They
weren't
remote
first,
but
at
that
scale
even
it
was
super
hard
to
keep
track
of
what
was
going
on
and
I
noticed
that
in
get
love
because
everything
is
so
accessible,
even
at
that
size.
I
think
you
already
see
advantages
of
thinking
as
a
remote
first
company,
because
everything
is
documented
for.
A
B
A
B
B
Like
we
are
in
the
field
that
is
highly
competitive,
like
data
science
requires
people
who
are
extremely
talented
at
what
they
do
to
be
able
to
to
provide
value
in
that
area,
because
the
people
in
the
area
are
very,
quite
smart,
so
yeah
we
see
it
as
a
necessity
to
get
access
to
that
people.
We
would
love
to
work
with
yeah.
A
B
Time
zone
alignment
is
a
important
consideration,
so
this
question
is
about
product
market
fit.
We
are
in
the
early
stages
and
we
think
we
have
something
interesting,
but
we
need
to
confirm
that
by
seeing
how
many
users
start
adopting
it,
did
you
have
a
moment
for
you
that
you
said
well,
if
there's
something
like
product
market
fit,
that's
what
we
just
experienced
that
get
that
it.
A
Wasn't
it
wasn't
as
clear
wait
get
lap
it's
just
that
I
did
a
show,
hacking,
news
and
300
people
signed
up
I
thought
that
was
validation,
but
then
nobody
wanted
to
pay
so
people
want
to
use,
but
not
pay,
which
is
not
so
helpful
if
you're
starting
a
business
and
then
I
just
got
a
ton
of
a
ton,
but
I
got
some
inbound
emails
from
really
interesting
companies
like
fortune.
One
companies
like
hey.
B
A
A
B
Yeah
we've
been,
we've
been
thinking
like
what
what
is
our
first
wave
like?
What's
the
first
thing
we
want
to
do
in
like
the
next
12
months,
and
for
us
it's
just.
We
want
to
make
sure
that
people
actually
start
using
it
and
that
it
starts
helping
data
scientist
to
build
their
predictive
models
and
we're
going
to
be
all
open-source
like
every
everything
is
going
to
be
open
and
free
in
the
beginning
and
the
investors
we've
spoken
to
they.
A
A
Then
that's
playing
because,
like
hey
you're
gonna
put
that
behind
a
penguin
and
for
us
it
was
like
those
fortune.
10
companies
were
kind
of
the
trigger
to
that.
They
started
asking
things
so
we're
like
hey.
This
is
okay.
L-Dub
Logan
was
already
in
the
product,
but
LDAP
Group,
synchronization,
hey.
That
makes
sense
as
a
paid
feature.
So
you
don't
have
to
think
about
what
features
you're
going
to
monetize
like
think
the
features
up
you
just
have
to
select
from
the
incoming
stream
of
requests.
A
B
That's
really
useful
practical
advice,
because
you
can
directly
do
this
and
see
if
it
works,
it's
very
clear
how
you
should
do
it
another
question
I
have
is,
but
when
we
are
in
this
process
of
starting
a
start-up,
we
get
a
lot
of
advice
and
we
ask
for
a
lot
of
advice,
but
we
also
get
a
lot
of
advice.
How
do
you
deal
with
advice
like
how
do
you
filter
out?
What
are
you
taking
to
account?
How
do
you
mix
it.
A
A
Second
of
all,
the
circumstance
is
in
with
date
which
they
did
it
are
they
similar
to
yours
like
b2c
and
b2b
are
very,
very
different
endeavors,
so
that
might
lead
to
very
different
advice,
and
it's
look
if
you're
hearing
the
same
thing
from
everyone,
it's
probably
right.
If
you're
hearing
different
things,
you've
probably
done
enough
digging,
and
you
can
just
make
up
your
own
minds.
Yeah.
B
That
makes
sense
yeah.
So
so
one
of
the
things
that
I've
been
using
from
the
things
you
mentioned,
the
other
things
I
can
start
using
one,
but
that,
like
I,
try
to
understand,
do
they
come
from
a
position
where
they
really
have
the
experience.
That's
all
ready
to
give
advice
on
this
particular
topic,
because
I
notice,
even
like
super
smart
people,
they
will
say
something
in
an
area
that
they
don't
that
I,
don't
really
understand
a
lot
about,
and
then
people
still
take
their
yeah.
A
B
Okay,
so
this
one
is
more
of
a
practical
question
from
a
technology
perspective,
some
might
be
more
fun.
It's
about
tech,
so
we're
thinking
about
our
stack
and
we
don't
want
to
over,
complicate
our
stack
so
that
we
have
like
a
C++
code
base
where
we
need
specialists
and
it
slows
it
down.
But
we
also
want
to
understand
when
to
like
opt
for
a
low-level
language.
If
it
makes
sense
yeah
do
you
generally
have
advice
about
startups
and
how
they
go
about
their
tech
stack
I
would.
A
Go
I
would
optimize
everything
for
speed
like
I,
think
it's
all
about
building
an
appealing
product
and
then
performance
problems
you
can
solve
later
on
we're
now
hitting
that
point.
We've
got
it's
not
good.
We're
starting
up
a
special
team,
just
a
performance
team
to
make
things
multi-threaded.
The
lab
is
now
pretty
different
services.
A
A
If
you
already
have
like
a
template
and
know
exactly
what
you're
gonna
make,
so
you
want
to
kind
of
do
the
its
the
prototyping
in
a
high-level
language
and
then
implemented
in
a
low-level
language
and
only
a
part
that
are
performance,
critical
and
the
thing
with
performance,
and
you
probably
know
this
but
measured
like
it's
it's
it's
amazing
how
wrong
your
intuition
tends
to
be
on
performance.
So
it's
all
about
measure,
improve
measure
again
and
do
it
that
way.
A
Yeah
I'm!
Not
if
we
cannot
think
said
premature,
premature
optimization,
is
the
mother
of
all
evil.
I
strongly
subscribe
to
that.
So
let's
go
with
a
high-level
language,
get
as
much
out
as
you
can
and
then,
where
it
starts.
Hurting
start
start
spinning
that
out
in
lower
level
language
stuff
will
be
my
advice,
because
speed
and
velocity
is
all
you
got.
That's
the
only
the
only
advantage
you
have
as
a
start-up.
So
don't
blunt
blunt
that
advantage.
Okay,.
B
A
It's
it's
such
a
broad
question:
I,
don't
know
I
just
I,
just
watched
in
the
Nvidia
presentation
from
I
think
I
think
it
was
even
last
year,
but
just
a
rapid
improvement
in
the
field
of
machine
learning
and
AI
is
so
amazing,
like
that,
can
be
an
explosion
of
different
models.
The
5x
year-over-year
gain
in
speed
and
complexity.
A
It
says
something
that,
like
open,
a
I
had
to
become
like
semi
for-profit,
because
otherwise
they'd
run
out
of
money
like
there's,
there's
billionaire
supporting
that
cause,
but
it's
it's
just.
We
just
learned
that
the
more
compute
we
throw
at
something
the
better
it's
gonna
be
I'm
talking
a
bit
more.
A
There's
amazing
companies
like
that,
like
focus
on
data,
wrangling,
I,
think
that's
the
most
most
of
the
smart
machine
learning
people
tend
to
and
most
of
the
machine
learning
people
said
tend
to
spend
80
percent
of
the
time
just
cleaning
up
their
data
as
far
as
I
know
not
very
experienced
here.
So
it's
it's.
The
data
wrangling
seems
to
be
underappreciated
in
the
industry.
Okay,.
B
A
A
We
just
lost
control,
so
the
way
to
stay
in
control
is
to
set
goals
that
you
hit
and
there's
I,
don't
think,
there's
a
way
that
they're
like
hot
is
this
investor
super
warm
fuzzy
feelings
for
the
open
source
and
therefore
it's
going
to
be
okay.
Good
investors
are
also
they're,
not
jittery
like.
If
you
miss
a
quarter,
they're
not
gonna,
be
jumping
up
and
down.
A
A
That's
a
great
question
and
in
this
case
this
venture
capitalists
that
they
were
on
our
board.
They
were
and
you
referenced
people
they
all
know
how
to
do
the
talk.
So
the
best
thing
to
do
and
it
takes
more
time,
but
it
is
worthwhile
you
look.
You
ask
them
for
a
complete
record
of
their
past
investments,
mostly
not
public
information,
so
they
have
to
supply
that
and
and
yeah
don't
do
that
with
Peter
Fenton
like
if
you
find
out
what
he
invested
in
but
like,
if
they're
a
bit
less
nervous
than.
A
Just
give
them,
but
the
important
thing
is
to
ask
for
a
complete
record
and
then
just
go
with
the
filled
companies
and
those
CEOs
will
tell
you
whether,
whether
they're
good
or
not,
that
you
built
your
track
record
with
the
filled
companies
and
that's
that's
the
not
so
nice
thing
about
being
a
VC.
You
end
up
spending
a
lot
of
time
at
companies
that
are
not
going
to
be
the
greatest
of
your
return.
A
B
That's
actually
quite
actionable
advice
and
I
somehow
feel
like
if
I'm
talking
to
a
VC
that
I
know
it's
kind
of
famous,
for
example,
in
the
Bay
Area
visiting
SF
right
now,
then
I
feel
like
like
me,
going
and
asking
them
do.
Can
you
give
me
a
full
list,
but
all
the
failures,
as
well
of
your
own
nurse
I'll,
start
calling
them
to
see
if
you're
any
good,
I
don't
know
it
feels
a
bit
out
of
place
because
I
want
it
well.
A
I
think
if
you,
if
you're
getting
close
to
a
term
sheet,
that's
a
very
appropriate
question
and
good
good
VCS
will
actually
say
this.
They
will
volunteer
this,
so
you
can
even
ask
so
good
question
hey
what
would
be
the
best
way
to
reference
you
and
any
good
VC
will
say.
Ask
me
for
my
failed
company
start
calling
the
CEOs.
A
That
is
the
default
answer
and
also
call
a
few
of
the
ones
that
worked
out
really
well,
where
they
had
a
big
impact,
and
just
one
just
just
for
fun
and
also
important
for
your
network
to
know
people
who
succeeded,
because
that's
data
yeah,
but
they
won't.
You
can't
ask
this
like
in
the
first
meeting,
but
if
you're
getting
close-
and
you
know
like
oh
yeah,.
A
Then
I
don't
have
any
open
source
seed
investors,
I'm
I,
don't
that's
almost
too
big
of
a
focus.
Good
seed
investors
are
brought
because
you
need
a
lot
of
volume,
so
they
don't
focus
on
a
specific
area.
I
want
to
give
three
names
in
its
public.
Setting
liquid
to
the
joe
Montana's
fund
has
been
our
most
helpful
seed.
Investing
so
they're,
really
really
good.
Sv
angel
from
Ron
Conway
is
generally
regarded
in
the
valley
to
be
the
best
seed
investor
and
1984
run
by
my
friend,
Rami
I.