►
Description
This talk was given at IPFS Camp 2022 in Lisbon, Portugal.
A
B
Years
ago
is
looking
for
a
way
to
index
my
book
collection
and
now
we're
indexing
about
700
million,
unique
cids
and
it's
yeah
growing
at
a
ridiculous
rate
about
1
million
a
day
at
this.
At
this
moment,
we've.
A
Been
especially
in
the
past
year,
as
you
saw
in
the
previous
slide,
we've
like
doubled
the
index
size
almost
no
more
than
doubled,
and
we
also
have
been
building
a
lot
of
cool
new
features
to
make
it
a
lot
more
usable
and
more
integratable.
We
have
been
scaling
out
with
the
help
of
filecoin
to
be
able
to
reach
much
higher
level
traffic
than
we
could
before.
So
we
can
hit
1000
hits
per
second,
we.
A
We
have
launched
just
three
days
ago,
a
brand
new
queue
service
that
piggybacks
on
the
pin
API
of
ipfs,
where
you
can,
which
you
can
use
to
push
your
cids
directly
into
our
Crow
queue.
So.
B
A
Would
love
to
have
your
input
on
if
you
use
it,
because
it's
so
brand
new,
it's
still
beta
and
we
really
need
to
get
some
user
feedback
on
this.
We
have
also
launched
recently
a
brand
new
update
of
our
front
end.
So
it's
a
lot
more,
a
lot
better.
With
a
lot
more
features.
The
most
coolest
new
feature
is
planetarify,
which
you've
got,
which
is
fully
integrated
audio
player
with
playlists
that
you
can
just
listen
to.
While
you
keep
searching
cat
pictures
on
ipfs,
for
example,
and
well
I.
B
Also
created
the
not
yet
another
thumbnail
service,
the
idea
is
that
we
have
a
universal
thumbnailer
for
ipfs,
so
you
have
a
file,
a
CID
could
be
a
video,
could
be
an
image
and
we
can
generate
thumbnails
out
of
them.
The
thumbnails
are
immutable,
so
infinitely
cachable
and
we're
actually
doing
that
over
ipns,
so
we
can
cache
forever.
This
is
all
super
experimental.
It's
why
we
haven't
shouted
it
from
the
top
of
the
tower,
but
you
know
play
with
it.
It's
on
npm.
A
B
A
little
bit
about
you
know
where
we're
coming
from.
You
know:
there's
there's
a
lot
of
habits
among
startups
like
oh,
we
want
to
have
double
digit
growth
and
we
want
to
have
yeah.
We
want
to
really
get
rich
and
have
a
yard.
Well,
we
don't
want
a
yacht.
What
we
want
is
search
to
be
free,
so
our
search
engine
is
open
source.
Our
data
is
open,
of
course,
the
underlying
isn't
we're
completely
objective
and
neutral.
So
it's
not
like.
B
We
have
an
AI
that
knows
you
or
generally
an
AI
sitting
between
your
search,
query
and
our
index,
and
we
have
a
public
API
that
we
really
want
you
to
use
and
play
with
and
make
your
own
clients
or
integrate
in
your
existing
ipfs
solution
to
do
search
and
discovery
so
where
we
really
want
to
be.
Of
course,
you
know
we
want
to
decentralized
search.
We
want.
We
want
to
solve
what
I
would
call
the
heart
problem
search
is
incredibly
resource
intensive.
B
We
use
a
lot
of
memory,
CPU
Network,
I
o
just
all
of
those
things.
So
it's
it's
incredibly
hard
to
really
solve
this
problem,
and
you
know
we
want
to
play
with
ideas.
We
want
to
collaborate
with
researchers,
we
want
to
collaborate
with
other
Engineers
other
projects
to
see.
For
example,
we
can
Outsource
part
of
the
crawling
and
indexing
to
content
providers,
because
the
moment
we
can
decouple
our
resource
usage
from
our
growing
user
base.
We
basically
hit
gold
because
we
we're
at
a
point
where
no
existing
search
engine
can
compete
with
us.
B
It's
just
about
Dynamics,
so
yeah
we're
really
looking
for
collaborators
people
that
play
with
our
stack,
but
also
people
that
really
help
us
iterate
on
this
model
on
the
idea
that
search
is
centralized.
It
doesn't
have
to
be,
but
it's
also
not
to
refuel,
to
just
decent,
realize
it
and
throw
some
blockchain
magic
on.
A
A
B
Probably
effectively
have
more
data
on
ipfs
than
anyone
we
want
to
integrate
with
content
providers,
so
the
the
fleeks
and
the
pinatas
and
other
pin
Services.
We
want
to
integrate
with
them,
so
we
index
stuff
as
soon
as
it
goes
on
ipfs.
We
also
want
to
work
towards
offering
build
API
services
to
browsers,
so
we
they
can
integrate
directly
to
meta
search
engines
like
duct.go
and
start
page
and
brave
search
these
kind
of
things,
and
we
are
very
interested
in
indexing,
structured
data
like
the
previous
stock.
B
The
icsc
I
think
it's
very
interesting
to
have
this
idea
and
we
really
would
like
to
have
mutable
data,
and
particularly,
we
want
your
pull
request.
We
want
to
work
with
you.
We
are
proper
open
source
project.
Our
stack
is
a
bit
complicated,
but
we
will
guide
you
through,
because
we
we're
really
looking
to
see
what
you
can
do
with
our
stack.
We.