►
From YouTube: The Graph’s Community Talk #6
Description
The Graph’s Community Talk #6
This video was recorded: Tuesday, Dec 21 @ 8am PST, 2021.
0:00 Intro
1:25 DCentral Recap
5:00 The Graph Academy announcement
7:23 Governance
10:18 The Graph Birthday Recap
15:25 Review Milestones
19:00 Network stats
21:40 The Guild
33:48 Semiotic
50:55 Holiday Discussions
Follow The Graph on social media
Twitter: https://twitter.com/graphprotocol?s=20
Instagram: https://instagram.com/graphprotocol
LinkedIn: https://www.linkedin.com/company/thegraph/
GitHub: https://github.com/graphprotocol
Website: https://thegraph.com
A
Hello,
everyone
welcome
to
community
talk,
number
six,
it's
the
holiday
season,
and
that
means
it's
also
the
graft
birthday
which
happened.
Last
friday.
We
had
a
big
event,
I'm
sure
most
of
you
will
have
seen
everything
that
went
down.
We
have
noel
brian
and
george
with
us
today
who
will
give
us
the
key
highlights
of
the
birthday
event.
Last
friday,
we
also
have
made
two
major
announcements
on
the
core
dev
site.
A
Coming
up
today
on
the
curator
side,
graphgod
will
join
us
and
present
to
us
his
vision
for
the
curator
catalyst
website,
which
has
some
really
fun
and
exciting
features
which
I
think
can
be
a
great
tool
for
curators
and
he's
going
to
walk
us
through
that.
But
we're
going
to
start
off
with
the
community
updates
and
we're
starting
with
the
central
land
descent
from
miami.
There
is
sorry
it's
time
for
miami
that
where
we've
been-
and
I
think
it
was
end
of
november
right,
how
did
it
go.
B
Yeah,
hey
everybody.
I
can't
sing
like
oliver
but
happy
holidays,
and
hopefully
everyone
has
some
exciting
plans
coming
up
to
wrap
up
the
year
we
broke
through
december
and
started
off
december
by
wrapping
up
decentral
miami.
B
It
was
a
nft
conference
hosted
by
decentral,
or
I
think
it
was
actually
central
land
and
this
one
was
very
different
than
the
ones
that
we've
typically
gone
to
and
the
reason
being
is
because
for
the
first
time
we've
actually,
instead
of
just
had
a
booth
and
talked
about
the
graph,
we
showcased
how
the
graph
could
be
used
across
different
ecosystems.
B
So
you
know
within
the
nft
marketplace
and
within
d5,
so
what
we
did
is
set
up
a
booth
and
collaborated
with
audience:
a
decentralized
streaming,
music,
streaming
platform
known
origin,
an
nft
marketplace
on
the
ethereum
blockchain
and
upshot,
which
is
an
nft
market
data,
analytic
tool
where
you
could
look
at
all
the
data
across
different
nft
marketplaces
and
what
was
really
cool
about
it
is.
I
think
it
was
really
one
of
the
first
times
where
you
got
to
see
the
graph
in
action
without
actually
seeing
the
graph.
B
So
a
lot
of
the
folks
that
came
by
our
booth
were
able
to
put
on
these
wireless
headphones.
We
had,
I
want
to
say
about
50,
to
60
of
them
and
then
participate
in
like
a
silent
disco,
and
that
was
insanely
cool.
A
lot
of
folks
have
definitely
thought
it
was
one
of
the
coolest
things
that
they've
experienced.
It
was
very
immersive.
B
On
top
of
that,
we
also
had
known
origin
auctioning
off
a
few
pieces
of
art
that
we
displayed
across
the
booth
and
some
of
the
art
pieces
actually
were
pretty
much
visualizations
and
representations
of
data
flowing
through
subgraphs
across
the
graph
network.
So
it
was
really
awesome
to
kind
of
see
all
of
these
three
different
tools
and
decentralized
applications,
kind
of
in
a
way,
work
together,
but
also
really
amplify.
B
You
know
the
use
case
and
the
need
of
what
the
graph
could
do
and
I
want
to
say
you
know
not
to
toot
our
own
horn,
but
a
lot
of
folks
did
say
that
we
had
one
of
the
most
amazing
experiential
unforgettable
booths
at
the
conference,
and
maybe
this
is
a
little
bit
of
a
heads
up.
B
You
never
know
what
city
we
might
pop
into,
and
you
never
know
what
sort
of
tricks
up
our
sleeves
we
might
have
if
you
wanted
to
come
and
participate
in
some
of
these
experiences
that
we're
putting
together
so
shout
out
to
a
a
ton
of
graph
contributors,
especially
the
folks
at
edge
and
node
that
put
this
all
together,
but
also
to
our
community
that
came
out
and
supported
and
took
some
photos
and
really
dove
into
you
know
really
seeing
how
this
all
comes
together.
A
B
Yeah,
that's
a
great
question,
so
I
think
the
there
was
an
auction
that
happened
on
december
3rd
of
that
week.
I
believe
it
was
the
friday
of
the
art
pieces,
so
I
believe
they
were
auctioned
off,
but
maybe
we
can
actually
maybe
share
some
photos
that
folks
have
taken
while
we
were
there
and
we
can
maybe
dig
up
something
to
share
with
the
community.
A
C
The
graph
academy
is
all
about
collaboration,
so
if
you
would
like
to
contribute
to
our
joint
endeavor
of
building
a
universal
knowledge
base
for
the
graph
ecosystem,
you're
more
than
welcome
to
do
so.
With
this
new
initiative,
we
can
provide
grants
to
contributors
who
are
creating
materials
that
we
are
looking
for
right
now.
We
have
two
live
rfps,
one
is
an
index
or
onboarding
course,
and
the
second
one
is
a
subgraph
development
course.
So,
if
you
are
interested
in
creating
educational
materials
from
these,
two
areas
feel
free
to
apply.
C
If
this
has
piqued
your
interest-
and
you
want
to
apply
simply
visit
the
graph
dot
academy
grants,
that
is
the
graph.academy
grants
to
find
more
information
about
the
grants
program.
You
can
also
find
a
link
with
more
information
in
the
show
notes,
and
if
your
area
of
expertise
is
not
covered
in
one
of
our
rfps,
you
can
feel
free
to
reach
out
to
us,
and
let
us
know
what
you
would
like
to
work
on.
We
would
definitely
love
to
hear
from
you.
A
Awesome,
thank
you.
Stefan,
I
think
the
big
takeaway
from
this
message
really
is
a
continued
decentralization
of
community
grants
that
we've
typically
handled
with
the
foundation
and
we've
now
opened
it
up
with
a
graph
academy
on
something
that
is
critically
important,
which
is
you
know,
writing
education
and
tutorial
contents,
and
you
know
doing
so
in
a
way
that
you
know
pushes
also
sort
of
the
decision-making
part.
You
know
to
the
more
to
the
community
and
it's
a
great
step
forward
and
expect
to
see
more
of
that
happening
next
year.
A
Okay,
let's
switch
to
governance,
we
have
a
few
governance
updates.
There
is
a
forum
post
out
there
with
the
headline
community
snapshot
voting
strategy
and
I've
explained
it.
You
know
in
the
paragraph
itself,
in
the
end,
it's
something
that
we've
done
now
for
for
a
year,
we've
gained
a
ton
of
experience,
we've
also
matured
as
a
protocol.
We've
introduced
the
curator
role
in
middle
of
the
year
and
that
right
now
is
not
yet
reflected
properly
in
the
community
snapshot
strategy
voting.
So
it's
a
good
time
to
revise
it.
A
There's
an
open
discussion
with
specific
suggestions
feel
free
to
chime
in
and
provide
your
feedback
on
that
we
have
two
gips
that
have
been
approved
by
the
council,
both
of
them
curation
related
one
we
have
actually
implemented
already
last
week
and
that
is
for
saving
gas
cost
reduction
for
the
first
mint
on
the
bonding
curve.
That
is
a
significant
improvement
of
about
60
percent
opening.
A
The
bonding
curve
has
always
been
a
bigger
process,
so
the
gas
costs
associated
with
that
were
higher
and
that
is
actually
quite
exciting,
especially
for
subgraph
developers,
who
also
put
the
first
signal
on
the
bonding
curve
that
cost
has
been
reduced
now.
The
other
piece
that
we
will
probably
release
early
next
year,
possibly
in
january,
is
the
change
of
ownership
transfer
and
you
know,
expect
some
more
information
to
come
out
of
that.
It's
not
just
about
changing.
A
Lastly,
on
the
community
update
side,
I'm
told
that
we
have
for
the
rest
of
the
year
in
respect
to
the
holidays,
there's
going
to
be
no
more
curation
catalyst,
aragon
working
group
or
index
or
office
hours
for
the
rest
of
the
year.
They
all
will
resume
again.
Beginning
of
next
year,
so
that
is
with
response
to
the
to
the
holidays
that
are
coming
up
all
right.
Let's
switch
gears
to
our
birthday
event.
Last
friday
we
have
noelle
brian
and
george.
Welcome.
All
of
you
to
community
talk
noel
is
with
edgar
node
and
she's.
A
D
Yeah
thanks
oliver
hello,
everyone.
I'm
super
excited
to
be
here.
We
had
a
really
exciting
birthday.
Last
friday,
as
oliver
touched
on,
we
had
a
few
things
roll
out.
We
kicked
off
the
day
if
you
follow
us
on
twitter,
with
a
lot
of
really
incredible
charts
that
george
helped
create
and
the
design
team
at
edunode.
D
A
D
Thank
you.
We
yeah,
we
kicked
it
off
with
a
a
twitter
spaces
and
our
very
own
nick
hansen
emceed,
and
we
had
yaniv
there
all
the
founders
of
the
graph
brandon
giannis,
and
then
we
had
eva,
of
course,
the
director
of
foundation,
and
then
we
had
someone
from
almost
every
core
dev
team,
except
the
guild
wasn't
able
to
make
it,
but
we
had
taken
from
edge
and
node
mark
from
streaming
fast,
joseph
figment
alex
from
streaming
fast
and
sam
from
semiotics.
D
So
it
was
really
great
just
to
hear
everyone's
thought.
They
kicked
it
off.
With
the
first
question
asking
giannis,
you
know
about
the
early
days
and
some
things
that
were
important
and
he
touched
on.
You
know
the
graph
giving
the
developers
a
solution
to
a
problem.
They
had
how
the
graph
stuck
to
the
original
promise
of
progressively
decentralization.
D
What
was
really
interesting
is,
I
think
the
second
question
maybe
was
if
the
vision
had
changed
over
the
last
year-
and
you
know
yaniv
said
no,
you
know
we're
really
staying
strong
to
the
vision
that
they
had.
Originally
you
know
thought
of,
and
if
anything
you
know
it
is
more
complex
and
vivid,
but
they
really
are
staying
strong
to
everything
that
they
had
like
envisioned
for
it.
So
I
thought
that
was
really
great.
D
We
have
some
other
great
quotes
taken,
saying
she's
grateful
every
day
and
excited
to
continue
growing
the
graph
and
inviting
anyone
to
you
know
join
the
ecosystem,
and
I
love
eva's
quote
too
that
you
know
we
aren't
not
necessarily
the
talent
we
are
neces.
We
aren't
not
necessarily
the
core
depths
but
making
sure
the
core,
devs
and
grantees
have
the
resources
they
need
and
the
strategic
thinking
so
really
interesting.
D
We're
going
to
be
hopefully
putting
the
whole
recording
out
in
case
you
missed
it,
and
so,
if
you
didn't
get
a
chance
to
listen
in,
I
really
would
suggest
that
you
did.
It
was
a
really
great
hour
with
with
everyone,
and
then
it
really
just
turned
into
a
party
from
there.
We
had
an
incredible
community
led
events
around
the
world
to
help
celebrate.
D
It
was
supported
by
edunote
and
the
graf
foundation,
but
we
actually
had
one
that
started
in
lagos,
nigeria,
with
over
50
people,
so
so
great
to
see
everybody
there
celebrating
they
actually
had.
This
is
a
graph
curator
tea.
So
it
was
really
great
to
see
everybody
there.
D
Having
a
great
time,
then
in
bang
galore
india,
we
had
over
a
hundred
people
show
up
on
a
rooftop
and
celebrate
and
then
in
new
york
city
over
100
people
there,
where
we
shared
some
of
those
great
sharks
that
you
can
see
in
the
background
there
and
then
in
argentina.
Of
course,
they
all
came
out
over
200
people
celebrating
the
graph
and
and
then
sf
we
had
about
75
people
there.
You
can
see
a
little.
You
need
sighting
over
there
in
the
back,
but
it
was
super
fun
and
great
conversation
and
just
so
nice.
D
You
know
I
went
to
the
one
here
in
san
francisco,
but
it
was
just
really
great
to
just
talk
to
all
different
people.
I
met
people
that
have
had
grants
that
have
just
been
following
the
graph
that
our
delegators
in
the
graph-
and
it
was
just
a
really
nice
time
to
just
connect
and
celebrate
and
be
together
so
yeah.
We
had
so
much
fun
and
just
so
so
thrilled
to
be
a
part
of
the
ecosystem.
A
Awesome,
you
know
the
one
thing
that
I
like
to
see.
You
know.
First
of
all,
it
was
community
driven
which
is
awesome
and
that
it
happened
really
across
the
globe
on
in
five
different
locations
in
india,
nigeria,
in
the
us
and
in
argentina.
Really
many
continents
represented
there
and
to
see
that
sort
of
engagement,
and
you
know
the
attention
that
we
can
get
on
a
short
notice
fairly
right.
D
A
A
E
Everyone,
thank
you
guys
for
joining.
My
name
is
brian.
I
support
marketing
at
edunode.
It's
been
a
tremendous
year
and
we're
really
glad
to
have
had
you
guys
along
for
the
journey.
I
wanted
to
take
a
moment
and
review
some
of
the
milestones
of
the
year.
It's
really
amazing,
just
looking
in
retrospective.
E
One
thing
to
note
is
the
chain
expansions
that
have
happened
throughout
the
year,
so
I
guess,
starting
in
february
the
graph
expanded
to
its
first
layer
to
blockchain,
announcing
indexing
querying
on
polygon
with
the
hosted
service
and
another
layer.
One
was
added
in
february,
as
phantom
support
was
announced
for
the
hosted
service
and
then
just
two
months
later,
cello
was
integrated
on
the
hosted
service
as
well,
adding
to
the
group
of
layer
ones
and
another.
Two
months
after
cello
in
june,
both
arbitron
and
optimistic
ethereum
support
were
announced
for
the
hosted
service.
E
Then
october
came
along
and
the
first
non-evm
blockchain
was
supported
with
near
protocol.
Super
exciting
and
more
integrations
are
being
worked
on
every
day,
so
stay
tuned.
I
also
wanted
to
take
a
moment
and
just
reflect
how
the
ecosystem
has
grown
with
the
core
dev
expansions.
E
So
for
the
first
half
of
the
year,
edunode
was
working
as
a
core
developer
for
in
the
graph
ecosystem,
but
it's
amazing
how
much
can
change
in
just
a
few
months
in
june,
streaming
fast
joined
as
a
core
dev
for
protocol
r
d
and
to
improve
indexing
performance.
E
Just
one
month
later,
figment
was
awarded
a
grant
as
well
to
join
as
a
court
of
bringing
experience
with
proof
of
stake,
networks
and
node
infrastructure.
Just
as
we
thought
things
calmed
down.
In
december,
semiatic
ai
was
announced
to
support
r
d
implementing
artificial
intelligence
to
explore
smarter
and
more
efficient,
query
price
negotiations,
and
then,
just
one
week
later,
the
guild,
which
was
a
waveone
grantee
for
sub
graph,
tooling,
also
joined
us
as
a
core
developer
to
develop
new
subgraph
features
and
improve
graph
nodes
querying
capabilities.
E
So
we're
looking
at
five
core
devs
in
just
the
span
of
six
months
now,
and
I
think
it's
just
amazing
to
see
the
community
grow
and
just
see
where
the
graph
can
go
next.
Now
that
there's
so
much
so
many
folks
working
to
build
it
up,
so
just
really
really
great
to
have
you
guys
along
for
the
ride.
It's
been
a
tremendous
year
and
we're
excited
for
the
next
year.
A
Thank
you
for
that
recap.
It's
amazing
to
see
the
growth
many
of
us
might
have
joined
the
graph
about
a
year
ago,
when
launch
maintenance
occurred.
Like
myself.
In
fact-
and
you
know
thinking
back,
I
can
only
envision
how
big
of
a
step
that
must
have
been
for
the
original
team
and
the
early
supporters
that
were
all
part
of
that,
and
now
here
we
are
a
year
later
and
to
see
how
much
growth
we've
seen,
and
there
are
more
things
that
obviously
have
happened.
A
You
know
curation
on
you
know
over
the
summer,
for
example,
is
one
of
the
things
that
that
I
think
of
it's
it's
great
to
see
how
the
community
has
grown
and
not
just
the
community,
but
also
the
network
itself,
and
that
leads
us
to
george,
who
can
talk
a
little
bit
about
just
some
of
the
network,
stats
that
we
have
now
and
how
that
has
grown
over
the
last
12
months.
F
Hi
everyone-
my
name
is
george,
I'm
the
finance
and
ops
manager
at
the
foundation
and
yeah.
I'm
excited
to
be
sharing
some
network
stats
with
you
today.
So
a
year
after
the
graph
network
launched
we're
home
to
150
plus
indexers
over
2
300
curators,
and
over
8
000
delegators,
there's
been
over.
528
million
lifetime
queries.
To
date
we
have
over
15
000
developers
using
the
network
to
build
and
to
query
subgraphs.
F
The
last
three
months
we've
averaged
28
month
over
month,
growth
on
developers,
so
we're
actually
starting
to
see
some
parabolic
growth.
On
that
front.
Query
fees
are
also
growing
on
the
network.
We've
eclipsed
over
270,
000,
grt
and
fees
on
the
network
to
date,
and,
finally,
I
thought
it'd
be
cool
to
share
a
success
story
from
one
of
the
network's
biggest
users.
F
I'm
sure
a
lot
of
people
on
the
talk
here
have
heard
of
the
constitution.
Dial
and
a
lot
of
people
don't
know
this,
but
the
community
funding
platform
that
they
use
juicebox
is
actually
powered
by
the
graph.
F
You
know
throughout
the
week
of
the
fundraise,
the
dap
was
making
over
2000
queries
per
hour
and
at
its
peak
it
was
making
over
250
000
queries
per
hour
and
the
reason
I
point
this
out
as
a
success
story,
even
though
the
dow
didn't
end
up
acquiring
the
constitution
is
because
it
showed
the
power
of
daos
and
the
power
of
human
coordination.
F
We
saw
thousands
of
people
mobilize
and
raise
historic
numbers
in
less
than
a
week
and
they
came
together
for
a
common
purpose,
and
this
was
all
made
possible
because
even
during
really
heavy
traffic
hours,
juice
juicebox
was
able
to
leverage
the
graph
network
and
operate
at
over
99
uptime,
so
yeah.
In
summary,
we've
had
an
incredible
year,
but
we're
really
just
getting
started
and
we're
just
so
excited
to
see
what
will
continue
to
take
place
over
the
next
year.
A
A
F
G
So
yeah
the
guild
is
a
world-class
team
of
software
engineers.
Really,
members
of
the
graphql
foundation
we've
been
building
roughquilt
tools,
libraries
sdks,
I
believe,
for
the
last
five
six
years
since
the
since
facebook
launched
the
giraffe
ql.
So
if
you're
the
dev,
who
has
used
graphql
in
the
past,
most
likely
you've
used
some
of
these.
These
tools
that
that's
that's
been
built
by
the
team.
The
apollo
client,
for
instance,
is
a
prime
example.
G
The
guild
has
greatly
converted
to
this
project
in
in
the
past,
so
it's
suffice
to
say
that
they
are
probably
the
best
graphql
focus
team
in
the
world
and
we're
pretty
excited,
obviously
to
have
them
joining.
The
graph
did
a
great
complement
to
the
current
pool
of
core
devs
and
I
believe
a
lot
of
people
will
be
happy
to
know.
There's
a
new
team
focused
on
bringing
these
new
features
to
sub
rough
consumers,
developers
and
yeah.
G
Most
of
these
features
are
highly
requested
things
that
we
actually
need
to
get
on
onto
the
network
for
sure
things
like
current
position,
but
yeah
we
have
yuri
and
dotan
in
the
call
so
I'll
I'll.
Let
them
speak
to
them,
add
more
color
explaining
how
we're
already
working
with
other
core
devs
and
just
talk
a
little
bit
about
our
future
plans.
H
Yeah
thanks
for
having
us
hi
everyone,
I'm
muy
from
the
guild
and
bhutan
is
also
here
with
me
and
we're
really
excited
to
be
here.
We've
been
excited
for
a
while.
Now
that
everything
is
out
that
we're
joining
it's
yeah,
it's
really
exciting.
To
talk
about
all
the
things
that
are
coming
like
pedro
said
we're.
Probably
today
the
largest
group
of
open
source
developers
in
graphql
yeah,
you
probably
are
using
our
tools.
If
you
use
graphql
both
on
the
backend
and
front
end,
we
really
believe
in
open
source.
H
We
believe
that,
from
the
start,
we
have
like
a
unique
model
that
we
execute
for
years
about
how
to
basically
build
this
real,
sustainable,
open
source,
and
I
personally
met
yaniv.
You
know,
I
think
it
was
a
while
ago,
a
couple
of
good
months
ago
and
my
mind
was
blown
by
the
you
know
everything
that
the
graph
is
doing
and
the
community
is
doing
and
the
more
I
researched
and
the
more
I
learned
it
just
sounded
so
exciting.
H
H
So
we
started
with
one
brand,
and
that
was
really
nice
and
it's
you
know
it
was
for
the
schema
prototyper
and
we
basically
built
a
really
cool
tool
that
I
think,
would
be
valuable
for
any
graphql
developer,
especially
graph
developers-
and
you
know,
working
with
with
the
foundation
and
working
with
on
the
graph
was
just
so
good
and
we
just
we
decided
why
not
to
do
more,
and
we
saw
there
were
so
many
opportunities
to
improve
the
graphql
api
and
the
code
and
things
that
were.
H
You
know,
there's
so
much
to
do
in
the
graph,
so
things
that
were
just
there's
always
more
to
do
so
yeah.
H
So
we're
really
excited
to
join
and
there's
a
lot
to
do
and
we
already
started
so
you
know
you
can
find
in
the
you
know
in
the
announcement
there's
a
lot
of
things
we
want
to
do
like
you
know,
composition,
analytics
and
aggregations
mutations,
bring
subscriptions
and
even
looking
at
real-time
queries
in
general,
like
live
queries,
maybe,
and
basically
like
basically
like
you
know,
looking
at
in
general
and
like
how
we
can
improve
the
experience
and
the
performance
and
the
code
in
general
and
we
already
started
autumn-
can
maybe
also
share
like
a
couple
of
words
about
what
we
already
have
done.
H
And
you
know
part
of
the
work
is
also
for
us
now,
we're
you
know
pretty.
You
know,
like
I
said,
the
largest
open
source
group
on
graphql
in
the
javascript
ecosystem,
but
now
we
started
working
in
rust,
which
was
also
really
really
exciting
and
we've
seen
also
a
lot
of
things
that
could
improve
the
rust
ecosystem.
H
So
one
of
the
things
we're
already
started
doing
was
basically
bringing
a
lot
of
the
you
know:
powerful
ideas
and
things
that
we
have
on
the
javascript
ecosystem
and
basically
enhancing
the
graphql
rust
ecosystem
in
general.
So
yeah,
it's
very,
very
exciting.
The
time
you
want
to
share,
maybe
just
a
few
words
about
the
things
that
you're
already
are
working
on
and
there's
already
some
open
pr's.
I
Yeah
sure
thank
you,
aurie
hi
everyone,
I'm
dotan,
I'm
a
member
of
the
guild.
I
just
want
to
share
quickly
what
we
started
working
on,
so
we're
working
around
making
the
graph
graphql
api
compliant
to
the
graphql
specification,
we're
working
on
adding
extra
validations
and
extra
checks
to
make
sure
that
every
query
is
being
executed
and
checked
before
accessing
the
actual
data.
I
We're
also
working
on
improving
the
the
way
that
the
sub
graph
is
being
queried,
we're
adding
more
filtering
capabilities
like
filter
by
a
specific
block
filter
by
child
entities
who
are
adding
ore
and
and
filtering
and
tons
of
other
ideas
that
could
really
improve
the
the
experience
of
querying
your
subgraph
we're
also
working
on
adding
a
graphql
versioning
mechanism.
I
This
is
very
important
because
we
want
to
make
sure
that
every
change
to
the
graphql
layer
that
we
introduce
is
being
checked
and
tested,
and
we
want
to
make
sure
that
every
change
that
we
do
is
being
communicated
in
a
version
that
matches
the
changes
that
was
done.
This
way
we
can
make
sure
that
the
data
that
is
being
returned
from
the
graph
is
always
deterministic
yeah
and
a
lot
a
lot
of
improvements
around
the
ecosystem
of
graphql
in
general,
so
yeah.
I
think
it's
going
to
be
super
exciting
and.
D
I
Very
happy
to
join
the
graphics,
cardiff.
A
Yes
and
welcome
you've
been
in
the
ecosystem
already
for
a
while,
as
you've
mentioned,
and
welcome
and
excited
to
have
you
part
of
the
ecosystem
as
a
core
deaf
now
I
want
to
go:
go
back,
maybe
to
your
origins
a
little
bit
that
for
those
of
you
here
on
the
call
that
might
not
know
you
so
well,
can
you
talk
a
little
bit
about
the
guild?
How
has
it
come
about
and
and
what
was
your
initial
mission
when
you
put
the
guild
together.
H
When
we
started,
I
used
to
work
by
the
way
before
that
the
meteor
that
apollo,
which
is
also
like
a
very
big
graphql
developer
and
throughout
the
years
I
so
for
years
me
and
the
tan,
were
working
together
and
we
work
with
a
lot
of
other
freelance
developers
and
we
try
to
basically
see
how
we
can
how
we
can
make
individual
developers
the
most
productive
for
themselves
and
open
source
was
like
the
the
main
part
of
it,
and
the
thing
is:
there's
all
kinds
of
ways
to
open
source.
H
You
know
I
one
way
is
like
working
in
a
big
company
like
facebook
and
then
working
on
react
or
something
like
that
which
has
plus
and
minuses
and
then
there's
another
way
of
working
in
a
startup
that,
like
you
know,
you
raise
money
from
fbc
and
then
you
try
to
sell
some
product
or
something
like
that,
and
and
I
what
we
thought
was
maybe
there's
another
way
like
there's,
plus
and
minuses
for
both
and
both
are
very
necessary
and
good
for
the
open
source
in
general
in
the
open
source
economy.
H
But
we
thought
maybe
there
was
another
way.
Maybe
there
was
a
way
it
was
a
bit
slower
like
you
know,
we
that
will
take
the
time
and
each
developer
could
work
slowly
and
think.
H
Well,
what's
the
right
solution
and
how
to
but
also
stay
there
for
the
long
way
like
not
like
there's
just
too
many
like
open
source,
libraries
that
are
big
open
source
for
marketing
reasons
or
whatever,
and
then
three
months
later
nobody's
touching
that,
and
it
just
happened
too
many
times,
and
we
just
decided
to
create
this
group
to
basically
create
every
library
we
create
will
both
people
could
rely
on
for
the
long
term,
and
it
doesn't
mean
it's
the
best
library
at
the
beginning.
H
We
don't
think
we're
like
geniuses
or
something
like
that,
but
we
just
think
that
if
you
iterate
over
something
for
years
and
years
and
listen
to
the
community
really
take
the
time
to
listen
and
to
learn
from
the
community,
you
will
end
some
with
something
that
is
very
powerful.
So
when
we
started,
nobody
realized
why
we
are
doing
it
today.
H
We
just
were
the
largest
in
the
ecosystem,
so
everyone
is
like
knowing
us,
but
it
wasn't
obvious.
It
was
like
yeah.
It
took
many
years
for
people
to
realize
that
there's
actually
something
special
here
and
also
for
us
to
actually
realize
this
is
something
special
and
over
the
years
we
took
libraries.
H
Also,
we
took
over
libraries
from
companies
that
deserted
those
libraries,
so
I
think
yeah,
I
think
that's
the
main
goal
of
the
guild
to
build
open
source
that
is
sustainable
and
for
the
long
term
and
the
floor
like
makes
developers
more
productive,
and
I
think
that's
why
it
was
so
natural
for
us
to
to
to
join
the
graph,
because
it
was
basically
the
it's
kind
of
like
the
same
vision,
just
like
bigger,
so
yeah
yeah.
There's.
A
H
H
You
know,
follow
us
and
and
basically
suggest
and
share
with
us
your
experiences
because
we're
now
in
the
process
we're
already
started,
and
we
already
have
a
lot
of
things
to
do,
but
we
now
can
be
in
a
position
where
we
can
hear
your
feedback
and
we
can
execute
on
that.
So
bring
your
feedback.
It's
very
valuable
for
us
and.
A
A
K
K
J
Yeah,
so
I
I
think
at
at
a
high
level,
I
think
many
of
you
have
have
heard
from
me
before
and
I've
given
I've
given
some
background
of
semiotics,
so
I'm
going
to
try
to
share
some
new
information
today
before
I
before
I
get
into
that.
J
One
of
my
one
of
the
things
I'm
proud
of
with
respect
to
the
graph
is
that
I've
I've
joined
every
community
call
that
has
been
held
since
I
believe
it
since
the
beginning,
which
was
in
the
spring,
and
I
I
personally
before
semiotic
got
involved
with
the
graph.
I
I
got
involved
back
in
december
at
the
the
launch
on
mainnet,
so
yeah,
so
I'm
sam
green
and
I'm
the
cto
of
semiotic
and
we
are.
J
We
are
one
of
the
new
core
dev
teams
and
we
we're
the
first
core
dev
team
focused
on
research
and
development
and
we're
located
in
silicon
valley.
Although
we
have
our
team,
we
have
team
members
spread
around
spread
around
the
world
right
now
and
we
focus
on
our
team,
focuses
on
ai
and
machine
learning
and
cryptographic
engineering
and
with
the
ai
and
ml
front.
J
We
focus
on
zero
knowledge,
proofs
and
so
nick
asked
about
my
role
as
as
a
cto,
so
as
as
a
t
as
a
cto,
I'm
responsible
for
guiding
the
long
term
technical
priorities
of
semiotic
and
I'm
currently
the
acting
lead
for
both
our
ai
efforts
and
our
cryptographic
engineering
efforts.
J
K
Sam,
we
talked
a
little
bit
before
this
call
and
we
decided
we
would
take
a
a
structured
approach
to
introducing
you
and
semiotic
to
everybody
on
community
talk
and
one
of
the
things
that
I
brought
up
was.
I
don't
know
if
everybody
will
understand
how
ai
will
lend
itself
to
the
graph
protocol
and
some
of
what's
going
on
there.
So
maybe
we
can
take
a
minute
here
and
just
talk
about
what
a.I
will
do
in
terms
of
benefiting
the
graph
protocol
and
maybe
an
example
or
two
of
how
that
might
look.
J
Sure
so
we
we
hack,
we
actually
have
several
ai
oriented
efforts
that
we've
been
working
on
since,
since
we
got
involved
with
the
graph,
which
was
during
the
wave
one
grant
during
way,
one
wave
one
grants
and
I'll
unpack
the
details
of
that.
So
one
of
our
focus
areas
in
ai
is
reinforcement,
learning
and
reinforcement.
Learning
is
a
family
of
algorithms
for
automated
decision
making
in
sequential
for
sequential
problems.
J
So
let
me
let
me
explain
what
that
means.
So,
if
let
me
tell
you
first
of
all
I'll
give
an
example
of
where
you
don't
need
sequential
decision
making,
and
that
would
be
if
you
were
trying
to
identify
if
there
is
a
cat
or
a
dog
in
an
image.
That's
that's
classification.
That's
a
one!
That's
a
one
step
process,
an
example
of
where
we
do
need
decision
making
would
be
an
indexer
who
is
trying
to
run
a
a
profitable
business
doing
indexing.
So
why
is
it
just
sequential
decision
making?
J
Well
as
an
indexer
you
you
want
to
be
around.
You
want
to
run
your
operations
for
a
long
time.
You
don't
want
to
just
you
know,
run
for
a
month
and
then
have
to
close
up
shop
or
that's
not
how
you're
thinking
as
an
indexer.
This
is
a
long
term
activity
and
so
what
you're
doing
as
an
indexer
as
you're
picking
you
need
to
pick
which
sub
graphs
to
to
sync
and
these
sub-graphs.
They
take
a
long
time.
J
Some
of
them
take
a
long
time
to
sync,
and
so
it's
an
important
decision
you
need
to
make
and
you
need
to
decide
when
you're
going
to
stop,
maybe
you're
going
to
stop
serving
subgraphs
in
the
future
you're
going
to
switch
your
resources
to
serve
other
subgraphs,
and
in
addition
to
that,
you
may
have
periods
of
activity,
for
example
with
the
constitution
dow.
Maybe
you
have
periods
of
activity
where
you
have
a
lot
of
demand,
so
maybe
you
need
to
scale
up
your
indexing
operations
or
maybe
you
need
to
scale
down
your
indexing
operations.
J
So
this
this
is
an
example
of
sequential
decision
making
and
some
of
the
tools
that
we're
building
will
help
indexers
with
their
sequential
decision
making
tasks
and
it's
our
goal
to
also
build
tools
to
help
other
other
users,
for
example,
delegators
and
curators,
with
their
with
their
decision
making
tasks
in
the
future,
and
in
particular
we
are.
Our
team
focuses
on
deep
reinforcement,
learning
and
that
just
means
you're,
combining
reinforcement,
learning
the
concepts
of
reinforcement,
learning
which
are
quite
old
with
neural
networks,
so
it
deep
implies
deep,
neural
networks.
J
So
what
have
we
done
in
the
past?
Concretely
in
with
these
techniques?
Well,
our
initial
work
was
we.
We
did
some
quick
experiments
where
we
basically
created
simulation
environments
of
the
gateway,
and
we
we
connected
simulated
consumers,
slash
aka,
dapps
to
the
gateway
and
we
connected
indexers
to
the
gateway.
And
then
we
used
some
rl
algorithms
to
help
the
the
dapps
select
the
prices.
J
So
right
now
in
the
system,
indexers
have
a
a
domain
specific
language
called
agora
that
that
was
created
by
the
graph
and
agora
lets.
Indexers
gore
lets
indexers
pick
prices,
so
you
could
say
okay
for
this
sub
graph.
These
this
type
of
query
is
going
to
cost
this
much.
This
type
of
query
will
cost
this
much
and
so
on,
and
why
is
it
important
for
indexers
to
have
this
ability?
J
Well,
when
you
get
a
when
you
get
a
subgraph
query,
these
queries
could
be
anything
from
it
could
ask
for
anything
from
a
single
cell
in
the
database
or
a
query
could
basically
request
the
entire
database
contents
and
these
two
or
it
could
be
anywhere
in
between
and
and
so
it's
important
for.
Indexers
to
accurately
charge
for
the
queries
in
a
way
that
reflects
the
underlying
cost.
Like
I
mentioned,
and
right
now,
this
is
a
very
manual
process
for
indexers.
J
They
have
to
by
hand
set
their
prices,
and
so
what
we're
working
on
is
a
version
two
of
agora.
With
that
will
automatically
look.
The
indexer
can
use
this
tool
to
automatically
look
at
their
most
common
queries
and
analyze,
it'll
be
able
to
analyze
the
underlying
cost
for
their
most
common
queries
and
then
capture
those
underlying
costs
and
the
the
templates
for
these
queries
in
the
in
their
agora
cost
cost
models
so,
but
but
what
does
that
have
to
do
with
ai?
J
Actually
it
that
what
I
just
said
or
what
I
was
just
explaining
with
this
these
automatic
costing
is
that
that's
not
actually
ai?
That's
that's
going
to
be
statistics
and
the
the
statistics
that
we
are
the
statistics
that
we're
capturing
we're.
Actually,
what
we're
doing
is
we're
productizing,
an
intermediate
step
for
our
ai
efforts.
J
Actually
we
this
this
agora
v2
is
a
is
basically
a
spin
out
product
that
we're
creating,
and
it's
just
a
byproduct
of
our
of
our
research
and
it's
something
that
it's
something
that
is
one
of
our
goals
is
to
productize
byproducts
of
our
research
whenever
possible.
J
Okay,
so
with
the
with
the
ai
effort,
so
now,
with
the
gore
v2,
indexers
will
be
able
to
accurately
charge
for
the
most
common
queries.
But
what
about
the
novel
queries
that
an
indexer
may
get,
and
this
could
happen
if
someone
who's
trying
to
do
data
science
with
the
graph
which
is
becoming
more
more
common?
J
If
someone's
doing
data
science
you're
going
to
have
people
doing
a
lot
of
experiments
with
novel
types
of
queries
or
where
you
may
need,
you
may
need
to
access
lots
of
data
for
your
data
science
efforts
and
what
we're
doing
is
during
the
agora
v2
process.
We
basically
had
to
dig
dig
into
the
guts
and
make
modifications
of
postgres,
and
why
did
we
need
to
do
those
modifications?
Well,
we
need
to
get
these.
We
need
to
get
very
accurate
data
from
from
postgres.
J
So
postgres
is
the
database
that
we
use
in
the
graph
to
map
subgraphs
to
actually
store
the
subgraph
data
when
it's
being
synced
by
an
indexer,
so
we're
we're
digging
down
into
the
guts
of
postgresql
modifying
postgres
so
that
we
can
get
the
actual
costs
of
a
query
from
the
postgres
database.
These
actual
cost.
The
reason
we
need
them
for
the
ai
efforts
is
because
we
we
need
these
are
what
are
called
labels
in
machine
learning.
So
we
need
labels
that
we
can
use
to
learn
what
the
what
the
mapping
is.
J
J
I
guess
I
want
to
mention
one
other
thing:
that's
come
out
of
this
work,
and
that
is
during
the
process
of
generating
all
this
data
that
I'm
that
I
mentioned
that
we
need
we
we
created
this
we've
actually
created
two
random
query:
generators,
the
first
random
query,
generator
that
we
we
made
was.
Basically,
it
was
we,
we
reused,
some
open
source
tool
from
ibm
and
we
we
noticed
that
it
put
a
significant
load
on
indexers,
and
so
we
we
basically
wrapped
that
and
released.
J
We
wrapped
it
so
that
it
worked
well
with
the
for
the
graph
and
we
released
it
as
an
open
source
tool
and
it
allowed
indexers
to
stress
test
their
infrastructure.
J
But
during
the
more
later
efforts
we
we
created
a
random
query
generate
from
scratch
and
this
this
query
generator
was
actually
pretty
hardcore
and
it
would
really
stress
the
the
indexing,
the
indexer
node
software
and
it's
actually
not
something
that
we
have
released
to
the
public
yet,
but
we
have,
we
are
using
it
internally,
it's
being
used
internally
for
development
of
the
graph
node,
because
it's
basically
it's
basically
being
used
as
a
fuzzer
to
make
sure
that
the
graph
node
is
stable
and
eventually
we
will
really
will
release
this
tool
too.
J
So,
basically,
at
a
high
level,
you
see
that
you
know
we
have
these
long-term
goals
of
making
these
ai
driven
ai
driven
products.
But
in
that,
in
that
process,
we're
seeing
and
we'll
continue
to
see
that
there
will
be
these
intermediate
tools
that
we'll
be
able
to
capture
and
package
and
make
and
make
available
to
the
community
so
that
we
can
just
have
better
better
experiences
I'll,
I
don't
wanna,
I
wanna
make
sure
we're
doing
okay
on
time,
so
I'm
gonna,
I'm
gonna
briefly
mention
some
reinforcement.
J
Learning
work
that
we're
doing
so
everything
I
just
talked
about
earlier
is
basically
classification
or
regression
where
we're
trying
to
get
we're
getting
queries
and
we're
going
to
predict
the
resource
cost
of
these
queries.
That's
not
sequential
decision
making.
That's
that's!
J
That's
actually
called
regression
with
three
on
with
respect
to
reinforcement,
learning
we've
been
working
closely
with
brandon
ramirez
and
we
basically
brandon
thinks
a
lot
about
okay,
how
it
what's
the
state
of
the
graph
and
is
everything
running
smoothly
and
are
there
changes
to
the
protocol
that
can
make
things
even
better,
so
we're
working
with
brandon
and
several
other
external
groups
to
to
improve
the
pro
and
protocol
through
various
fields
of
expertise,
so,
specifically
we're
working
with
al
begman.
J
He
is
a
game
theorist
and
we're
working
with
prism
group
and
basically,
basically
prism
group
is
a
they're,
a
bunch
of
phd
economists
and
we're
working
with
block
science
and
block
science
specializes
in
complex
systems,
which
is
a
field
that
combines
emergence,
non-linearity
and
feedback
loops
and
block
science
is
an
expert
in
addition
to
all
those
fields,
they're
an
expert
in
building
simulators.
They
have
a
a
tool
called
cadcad
and
cad
cad
basically
gives
us
an
environment
you
could
think
of.
J
If
you
can
think
of
an
environment
in
the
context
of
ai
is
like
it
is
sort
of
like
game
engine,
so
they're
building
this
a
very
accurate
game
engine
that
captures
the
dynamics
of
the
graph
protocol
and
this
reinforcement
learning
our
reinforcement,
learning
efforts
which
are
part
of
these
greater
efforts
and
in
terms
of
making
changes
to
incentive
mechanisms
where
we're
collaborating
with
all
these
teams.
So
all
of
us
are
like
looking
at
this
game
engine
and
we're
going
to
run
simulations
in
this
in
this
game.
J
Engine
which
models
the
graph
to
test
new
changes
to
the
protocol.
K
Sam,
thank
you
so
much
for
sharing
all
that.
It's
a
great
overview
of
yourself
and
the
team
at
semiotic.
I
want
to
let
everybody
know
that
grt
iq
podcast
is
featuring
sam
and
the
team
at
semiotic
this
week.
So
if
you
want
to
learn
more
about
some
of
the
contributions
that
sam
and
the
team
are
going
to
make
to
the
graph
along
with
an
in-depth
background
overview
of
semiotic,
you
can
listen
to
the
grt
iq
podcast
this
upcoming
week
and
with
that
sam.
Thank
you!
So
much
I'll
throw
it
back
to
oliver.
A
Very
nice
well
welcome
sam
and
semiotic
ai.
You
know
as
a
core
dev,
it's
a
mind-blowing
stuff
that
you've
covered
here
sam
and
I'm
excited
to
hear
more
on
the
grt
iq,
podcast
and
obviously
you
know
for
next
year
to
see
the
work
coming
out
of
semiotic
and
how
it
applies
to
the
graph
ecosystem
awesome.
A
A
My
father
used
to
be
a
chef,
and
food
was
big
theme
over
the
holidays
for
us,
and
it's
something
that
I
carried
forward
and
one
of
the
things
that
we've
always
done
over
the
holidays
was
eating
roast
beef
on
toast,
with
a
remoulade
sauce
and
the
remoulade
sauce
was
made
this
special
cake.
But
anyone
who
loves
roast
beef
try
that
out,
it's
just
so
superb
and
it's
a
nice
little
snack
and
it's
something
that
I've
been
doing
and
continuing
with
my
own
life
now
as
well.
That's
something
that
I
always
look
forward
to.
A
F
I
could
go
I'm
renting
out
a
a
restaurant
in
chicago
called
toppo,
gigio
italian
restaurant
for
me
and
a
few
friends
from
college
to
throw
a
holiday
christmas
party.
So
I'm
pretty
excited
about
that.
B
I
can
go,
this
is
going
to
sound
a
little
lame,
but
if
anyone
I'm
not
going
to
share
any
spoilers,
however,
I
saw
spider-man
3
last
week
and
I'm
really
looking
forward
to
seeing
it
again
with
my
dad.
I
haven't
seen
a
movie
with
him
in
a
while,
so
it's
I'm
really
looking
forward
to
spending
some
time
with
them
with
some
family.
A
A
Reminds
me
of
die
hard
home
alone
kind
came
out.
At
the
same
time,
a
little.
H
A
Earlier,
I
think
in
the
late
80s,
when
you
watched
that
back
in
the
days
you
had
this
like
awesome
image,
it
really
sort
of
kicked
off
the
the
special
effects
in
a
different
way
that
we
now
have
come
to
use
to.
But
if
you
go
back
and
watch
those
old
movies
they're,
not
quite
as
impressive
anymore,
as
you
used
to
think
about
it,
which
is
kind
of
interesting,
and
I
watched
that
lately
die
hard.
You
know,
diet.
3
was
actually
a
little
bit
of
a
letdown,
so
we
stopped
watching
the
other
two.
J
Maybe,
following
on
from
the
food
related
theme,
I'm
I'm
excited
I'm
I'm
traveling
to
san
antonio
texas
to
meet
up
with
my
family
and
my
sister
she's.
She
has
been
specializing
in
she's
perfected.
Well,
maybe
not
perfected,
but
she's
she's
really
skilled
at
making
chinese
dumplings
and
they're
amazing,
and
she
also
has
been
spending
about
10
years,
making
homemade
chocolates,
which
are
which
are
epic.
So
I'm
excited
about
both
of
those
things.
A
K
Hey
oliver
I'll
jump
in
here,
real
quick.
First
of
all,
I
want
to
wish
everybody
a
merry
christmas
and
a
happy
new
year,
and
I
guess
jim
cousins
and
myself.
We
have
the
same
holiday
tradition,
it's
december
21st
and
I
need
to
go
shopping
last.