►
From YouTube: 2023-03-29 Code Review Weekly Sync
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
Yeah
right
in
the
agenda,
we
have
right
it's
an
announcement
of
the
formation
of
the
new
working
group.
That's
going
to
be
looking
into
AI
integration
as
it
says
there.
That's.
The
goal
is
to
aim
to
Define,
coordinate
and
ramp
up
integration
of
AI
capabilities
into
all
product
areas.
A
So
you
can
look
at
the
page
there,
the
handbook
page
of
the
working
group-
sorry
in
our
group,
why
we
have
some
members,
so
we
do
have
me
representing
for
code
review
and
eventually
the
membership
will
grow
as
necessary
in
the
future.
I
copy
pasted
this
from
the
source
code.
That's
why
I
mentioned
okay,
but
it's
relevant
I
guess
to
to
be
aware
that
they
are
present
there.
A
So
we
can,
if
you
have
any
product,
focused
things
or
engineering,
Focus
things
you
can
leverage
me
in
Sarah,
I
guess
to
get
them
in
the
idea,
then,
is
there's
the
MLA
mlops
team
and
the
AI
assisted
team
to
or
the
AIS
Institute
to
work
on
their
own
models.
Then
we're
going
to
be
looking
at
potentially
integrating
with
far
beyond
that
and
that's
what
we're
going
to
be
looking
to
enable.
A
But
then
the
features
themselves
they'll
be
built
out
by
the
product
teams
consuming
those
apis
that
will
be
built
in
the
infrastructure
that
we
built.
So
that's
kind
of,
like
my
my
announcement
questions.
C
D
C
A
So
there
isn't,
there
is
a
a
search
by
customers
for
this,
and
what
we're
looking
is
for
reasonable
features
that
enable
and
Empower
our
customers
that
do
want
to
play
with
this
tool
or
these
tools
to
allow
that
to
enable
them
to
do
the
to
do
so.
What
this
working
group
will
be
looking
to
do
is
to,
instead
of
having
each
group
trying
to
figure
out
ways
to
do
this
in
their
own
way.
A
A
To
ship
API
power
features
in
how
that
looks
like
it
looks
like
we'll:
have
a
Consolidated
API
to
get
access
to
this
internal
and
external
models,
and
then
eventually
it
will
allow.
You
know
us
to
enable
certain
features
if
the
users
are
interested
in
that.
So
that's
my
attempt
at
answering
that
question,
but
I'll
keep
it
posted.
B
Okay,
yeah
I
was
a
reminder,
a
friendly
reminder
that
code
review
has
be
only
or
one
of
I'm
pretty
sure.
It's
the
only
current.
B
The
product
at
all
and
suggested
reviewers,
and
so
I,
was
talking
to
Taylor
about
this
yesterday
and
trying
to
understand
like
where
usage
was
anyways,
so
we
should.
We
should
also
not.
You
know,
look
for
opportunities
that
we
could
highlight
those
more
right
like
those
are
things
that
are
also
good
from
a
user
perspective
like
we
know,
Finding
reviewers
is
hard,
but
also
from
a
business
and
and
broader
Market
perspective.
Currently
so.
B
A
But
we're
going
to
say
on
the
same
subject
on
the
next
point,
so
I'll
move
along.
If
you
have
any
questions,
we'll
bring
it
up
so
in
the
spirit
of
experimenting
with
potential,
cruel,
Integrations
and
useful
and
intuition
that
we
could
do,
we
have
been
discussing
to
organize
some
sort
of
like
in
for
very
informal
hackathon
for
Rapid
prototypes,
so
what's
been
happening,
I
created
a
draft
issue
that
I'm
kind
of
like
spitballing
and
brainstorming.
A
What
this
could
look
like
the
idea
would
be
to
set
one
day
aside
for
the
engineers
to
collaborate
in
building
things
into.
You
know
our
areas
of
the
product
that
Implement
some
of
the
IDS
that
are
falling
around.
The
reason
for
this
is
essentially
to
get
us
acquainted
with
the
topic,
but
also
to
explore
potential
problems
with
UI
or
things
that
we
need
to
be
aware
of,
or
even
get
us
on
board
with
exactly
what's
going
to
be
the
hurdles
that
will
face
once
we
start
building
these
things.
A
This
will
facilitate,
or
this
will
help
planning
future
actual
implementations
of
these
features.
This
will
not
be
rolling
out
to
production.
This
is
just
going
to
be
just
playing
around
with,
but
the
ideas
didn't
just
to
get
everybody
on
board
with
the
topic.
A
A
The
hackathon
is
exactly
experimentation
and
there's
a
bunch
of
there's
some
issues
linked
there
with
ideas
and
use
cases
that
fall
on
source
code
and
code
review
and
a
bunch
of
other
groups
that
they
all
look
great
on
paper.
But
do
the
actual
look
good
once
we
start
playing
with
them.
A
So
that's
what
we're
trying
to
see
the
idea
will
be
to
have
some
sort
of
functional
prototypes,
to
call
the
to
call
the
apis
to
get
the
response
to
see
if
they
are
even
good
enough,
because
just
because
we
use
AI
doesn't
mean
that
this
is
going
to
be
useful.
So
we
definitely
want
to
build
the
right
thing
for
our
customers
and
for
our
things
so
yeah
just
announcing
a
hackathon.
A
It's
aligned
with
the
way
they're
building
these
things
and
yeah
I
I
am
thinking
of
using
some
type
of
our
team
to
work
with
that
I
believe
most
of
these
things
will
Leverage
front
end,
mostly
efforts
of
like
consuming
apis,
but
probably
backend
also
can
help
in
terms
of
talking
to
these
apis
as
well,
making
that
smooth
but
yeah
just
to
let
you
know
that
that's
gonna
happen,
I'll
keep
you
posted
on
once.
A
B
Licensing,
whatever
challenges
I
know
like
the
current.
There
are
some
current
concerns
around
like
the
models
that
do
exist
out
in
the
wild
and
like
using
those
for
anything
or
trying
to
like
do
that,
I
guess,
like
yeah,
how
do
I
you
know
how
do
I
hackathon
if
I
have
nothing
to
like
go.
Ask
questions
to
sort
of
my
question.
That's.
A
A
We
definitely
want
to
make
sure
that
we
don't
again
we're
not
shipping
we're,
not
merging
any
code,
so
the
session
will
be
experimentation,
so
there's
a
chance
that
we
might
just
use
personal
API
Keys,
not
personally
guys,
but
our
own
API
key
is
just
to
experiment
and
then
we'll
have
to
you
know
when
you
check
out
the
code
you'll
set
in
your
variable,
with
your
key
to
call
with
your
own
account,
if
that's
possible,
so
we'll
or
we'll
be
giving
keys
to
experiments.
A
So
there's
still
a
couple
of
things
starting
out,
but
they're
still
so
yeah,
we'll
we'll
do
it
once
we
have
answers
about
how
to
Rapid
prototype
these
things.
So
yeah
I'll
keep
you
posted
there.
That
makes
sense.
B
D
A
So
Ben
we
don't,
we
don't
want
to
limit
too
much,
so
it's
going
to
be
just
one
day
across
all
time
zones,
but
it's
just
going
to
be
one
day,
so
we
don't
want
to
like
be
too
ambitious
here,
but
we
will
not
be
too
restrictive.
A
So
if
you
have
an
idea
that
is
slightly
related
with,
with
with
the
group
that
you're
in
and
and
potentially
could
be
interesting
to
experiment
who
knows,
maybe
we
can
learn
from
that
experience
and
then
turn
that
into
a
more
concrete
proposal,
so
I
wouldn't
say
no
right
off
the
bat,
because
it's
not
me
to
say
it's
again.
It's
going
to
be
an
informal
I
have
to
stress
this
out.
It's
gonna
be
an
informal
hackathon.
Do
not
expect
this
whole
bunch
of
prizes
coming
out
of
the
you
know,
it's
gonna
be
for
fun.
A
A
Well,
bragging
rights
are
prizes,
so
I
mean
okay,
so
essentially
yeah.
We
have
to
keep
it
open,
there's
no
much
restricting
right
now
on
the
things
that
could
be
built
into
the
context
of
the
hackathon.
We're
looking
to
play
around
with
this
and
learn
from
it,
as
a
group
we'll
also
be
pulling
in
people
from
source
code.
A
Global
search
has
also
showed
some
interest
in
playing
around
with
it
at
this
level
of
like
experimentation,
loose
experimentation,
loose
collaboration,
so
there's
a
thread
there
already
so
yeah,
it's
gonna
be
a
fun
experimentation.
That's,
but
that's
the
point.
A
Yeah,
that's
a
nice
experimentation.
The
GPT
thing.
A
Yeah
Thomas
I
know
that
it's
really
going
but
I
think
what
what
they're
talking
about
is
having
that
internal
API
built
by
gitlab
be
a
pass
through
to
that
API,
but
eventually
yeah
we'll
need
to
have
some
ways
to
interact
with
this,
but
from
the
front
end
stuff.
So
you
probably
need
some
back
in
once.
We
have
some
of
those
answers
and
like
have
a
streamlined
way
of
experimenting
with
this.
It
might
not
be
close.
A
Very
close
to
I
will
be
shipping
into
the
product
will
be
enough
to
because
you
know
like
how
to
call
the
API,
whether
it's
internal
or
whatever,
that's
not
the
biggest
question.
We
know
how
to
do
those
sort
of
things.
What
we
don't
really
know
is
the
unknown.
That
comes
after
that,
like
Which
models
are
better
for
certain
questions.
Will
we
be
a
hard-coding
prompts
into
our
code
to
send
to
the
external
apis
we'll
be
allowing
the
users
to
prompt
it
directly?
A
Like
a
lot
of
those
questions,
we
don't
really
know
because
we've
never
seen
them
in
the
action
with
our
product
with
our
code
things
again
once
we
start
talking
about
shipping
this
to
production,
working
with
production
data
a
whole
bunch
of
other
questions
rise
up,
so
we
don't
have
those
answers
yet
right
now,
just
local
experimentation,
demoing
it
with
some
videos
and
stuff.
That's
the
level
of
the
expectation
for
the
hackathon
and
then
we'll
be
dedicating
one
day
of
capacity
for
it
and
looks
will
be
in
this.
A
Milestone
looks
like
not
so
potentially
the
next
Milestone.
We
might
do
the
planning
recording
reducing
capacity
for
it
once
I
have
more
I'll
share
Amy.
Do
you
want
a
voice.
C
A
We
have
a
demo
for
that.
Oh
did
we
yes,
I'll,
find
you
a
link
for
it.
Okay,
so
Tim
is
Tim's
playing
around
with
this
one
of
the
things
he
did.
Is
he
used
a
a
large
language
model
to
to
train
sorry,
you
trained
a
large
language
model
to
understand
the
docs,
and
the
demo
is
basically
a
prompt
to
ask
questions
about
the
docs.
Sorry,
it's
the
hand,
it's
a
handbook,
not
the
docs.
Sorry,
so.
D
A
C
A
Else,
I'll
add
something
to
the
agenda
because
a
lot
of
the
ideas
so
Tim
has
been
talking
about
this
concept
of
like
tiered
levels
of
complexity.
Where
one
is
like
a
simple
column
response,
you
send
something
and
like
hey
analyze,
this
that's
one
level.
The
other
is
like
hey
get
trained
on
my
stuff
and
then
help
me
try
to
find
this
whatever
just
kind
of
like
what
the
docs
are
and
there's
likely
even
more
complex,
more
custom
built
models.
That's
the
level
three.
A
The
first
level
should
be
fairly
in
a
fairly
simple
to
implement
is
like
API
calls,
which
I
have
to
figure
out
the
legal
Parts,
a
UI
Parts,
because,
for
example,
if
you
think
about
explaining
this
code
or
summarizing
this
Mr
or
what
give
me
a
walkthrough
for
this
Mr,
a
lot
of
that
is
is
does
a
lot
of
options
right?
Are
we
gonna
just
throw
them
in
the
menu?
Are
we
going
to
allow
people
to.
A
So
those
are
questions
that
we
really
need
to
figure
out
for
a
product
which
goes
even
beyond
all
the
legal
ethical
questions.
It's
just
like
some
of
our
customers
do
want
to
use
it.
That's
kind
of
like
the
the
trigger
for
all
this
or
not
the
trigger,
but
at
least
the
the
reason
for
us
to
be
looking
into
all
this
so
yeah
once
we
figure
out
all
the
legal
stuff
and
all
that
we
can
start
shipping
this
stuff.
A
A
C
C
If
you
follow
the
link
to
that
merge,
request,
I
would
suggest
you
get
a
cup
of
whatever
has
caffeine
in
it.
It's
long
long
story
short
we're
on
it.
We
know
it's
going
to
be
big
and
it's
probably
all
going
to
come
down
to
the
end
of
1511
when
my
availability
is
going
to
drop,
so
this
could
get
fun
as
the
Milestone
goes
on.
C
A
C
B
Oh
yes,
I
did
put
a
note.
Mine
is
just
a
thing
thanks
for
continuing
to
stay
involved
in
the
effort.
I
know.
There's
there
are
a
lot
of
moving
pieces
here.
So
I
appreciate
your
yeah
depth
of
searching
and
finding
and
all
of
those
things
to
stay
on
top
of
it.
I'm
hopeful
that
they'll
they'll
get
it
done
and
the
feature
flag
will
give
us
some
breathing
room
there.
So
it.