►
A
All
right
cool
we're
streaming,
so
this
is
me,
I'm
actually
streaming
one
camera
to
the
other
computer
so
because
yeah
this
is
actually
better
than
the
usb
camera.
I
have
believe
it
or
not.
Quality-Wise
all
right
so
deal
here
is
that
we
have
a
problem
as
usual,
so
I
need
to
figure
out
how
to
make
this
a
second
display.
A
That
is
virtual
duplicate
or
extend
to
connect
your
space
change.
The
resolution
windows
there's
some
way
to
do
this.
I
can't
remember
what
it
is,
so
we're
probably
gonna
have
to
google
it,
but
windows
has
a
way
to
add
like
an
extra
vga,
adapter
or
something
older
displays,
select,
detect
to
connect
them
now
to
a
wireless
display,
no
graphics
settings.
A
A
A
Okay,
you
know,
I
don't
think
we
really
need
to
deal
with
this
right
now.
I
will
make
the
usb
some
other
day,
because
this
is
just
getting
a
little
ridiculous.
All
right.
This
should
be
our
video,
oh,
but
yeah.
You
can't
see.
Can
you
see
me?
Oh
yeah,
you
can
okay
well
great.
Okay,
then
we
don't
have
to
do
that
anyways.
A
So
try
to
take
the
part
of
the
green
color
out
of
the
overlay
here
and
see
if
I
can
be
transparent,
so
always
trying
to
be
a
transparent
zone.
That's
why
we're
doing
these
videos.
So
in
the
spirit
of
transparency,
let's
get
to
it,
I
want
to
document
the
whole
process.
So,
okay,
where'd
we
live
off.
Okay,
I
think
I
kind
of
got.
I
got
down
a
rat
hole.
I
got
that
a
rat
hole.
I
cut
off
video.
A
I
made
another
video
got
down
another
rad
hall:
okay,
we're
back
we're
just
gonna,
get
down
to
business.
Okay,
so
first
thing
is
the
introduction,
this
maps
to
the
titling
of
the
past
video
and
this
video.
So
we're
gonna
start
with
we're.
Gonna
do
a
set
of
volumes
for
this
series
and
it's
a
big
project
we
got
planned
here.
So
we're
gonna
see
a
set
of
volumes
for
this
series
volume.
A
Zero
will
sort
of
set
the
stage
for
us
give
us
our
kind
of
operating
parameters
is
where
we're
going
to
do
our
planning
essentially
and
then
volume
one
will
be
alice's
alice's
adventures
in
wonderland,
where
we're
we're
actually
going
to
rewrite
wonderland
and
we're
going
to
then
change
it
into
something.
That's
you
know
applicable
to
our
case,
which
is
you
know,
architecting
this
software
architect
of
alice
right
and
how
we're
going
to
build
her
and
all
the
stuff
that's
going
to
go
into
it.
A
So,
okay
and
then
obviously
the
tutorial
series
is
to
teach
how
to
do
the
same
thing.
So
these
are
engineering
logs
on
how
we're
going
about
the
process
of
building
the
tutorial
series.
So
we've
got
a
few
people
involved
already,
who
you
know
we're
gonna
loop
into
different
parts
of
this,
and
then
you
know
us
in
the
community
who
are
active
and
and
watch
the
videos
on
youtube,
and
you
know
we
show
up
in
the
weekly
sinks.
You
know
we're.
A
We've
been
doing
a
lot
of
machine
learning
stuff,
that's
been
building
up
to
this
right,
and
so
basically,
I
think
I
think
we've
got.
We've
got
the
thing:
that's
gonna!
Let
us
tie
together
the
operations
and
the
cli
and
the
ml,
and
the
data
flows
I
think
it
came
together
is,
is
the
long
short
of
it
now
we're
gonna
go
into
the
long
long
of
it
so
yeah
it
all
came
together.
A
I
figured
out
shared
config
figured
out
unification
with
python
type
system
figured
out
what
was
the
other
one:
sugar,
oh
system
system,
local
resource
management
and
yeah.
A
There's
just
there's
a
lot
of
good
things
that
flow
from
figuring
those
out
and
oh
also
manifest
overall
architecture
involving
the
distributed,
the
a
possible
optimal
distributed,
compute
situation,
it's
gonna
be
based
on
a
project
called
kcp,
which
is
going
to
decouple
a
kubernetes
api
server
from
the
the
cluster
interface,
and
so
basically,
that's
going
to
allow
us
to
swap
out
and
create
a
bunch
of
custom
crds
under
there
and
then
we're
going
to
do.
A
The
first
thing
of
that
today
is
that
we're
actually
going
to
then
write
some
proxies
in
there
so
that
we
can
invoke
these
kubernetes
apis
instead
of
we'll
we'll
basically
run
a
little
service
within
kcp,
which
is
basically
my
understanding
is
they're,
all
basically
crds,
and
so
some
kind
of
background
job
that
we'll
spin
up
on
start.
A
That's
gonna
take
messages
from
that
are
we
can
interpret
as
being
from
like
the
what
web
three
space
and
we're
gonna
go
and,
let's
see,
can
you
see
me
better
over
here?
Okay,
maybe
I
took
too
much
green
out
of
the
background,
so
we're
gonna
basically
grab
that
did
spec.
That's
what
we're
gonna
be
doing
today,
we're
gonna,
try
to
understand
that
we're
gonna,
try
and
understand
how
that
relates
to
the
web
relay
notes
and
we're
going
to
try
to
understand
where
personal
data
stores
are
at.
A
I
don't
think
personal
we're
not
actually
going
to
touch
that
we're
not
going
to
touch
personal
data
source
for
now.
That
probably
will
be
when
we
get
to
identity,
aware
contexts.
These
are
contexts
running
within
a
data
flow.
A
So
that's
when
we'll
want
to
look
at
that's
going
to
be
when
we
want
to
look
more
personal
data
stores,
I
believe
so,
but
in
the
meantime
and
and
tonight's
activity,
the
the
main
event
is
is
is
what
the
hell's
going
on
with
these
tid
specs
and
the
a
few
of
you
who
have
heard
me
lately.
I've
basically
been
champion
this
thing
that
I
know
very
little
about
here.
So
I
read
a
little
bit
of
the
spec
and
it
seems
perfect,
based
on
what
I
skimmed
but
granted.
I
spent
about
two
minutes.
A
Reading
this
thing
most
of
it
was
looking
at
a
flowchart,
so
I'm
really
hoping
that
that
was
correct
assumptions
and
what
I
think
is
the
assumptions
I
gleaned
from
from
skimming
it
or
basically
that
is
we're.
Gonna
build
a
giant
linked
list.
So,
and
you
know
some
of
this
metadata
is
or
basically
you
some
of
this
stuff
goes
on
chain
and
that's
like
I'm
putting
something
in
the
blockchain
now
you
know
we
don't
really
want
to
put
a
lot
of
stuff
in
the
blockchain.
A
So
this
kind
of
follows
with
the
rest
of
what
we've
been
talking
about
with
alice
if
you're
following
the
discussions
thread,
which
is
that
you
know
we're
going
to
sort
of
encode
software
like
it's
dna
and
you
know,
different
little
organisms
have
different
jobs
right
and
their
job
is
encoded
in
their
dna.
A
A
Basically,
what
we
can
do
is
we
can
encode
a
manifest
and
a
manifest
is
basically
so
you
got
to
go
check
the
thread
that
one
that
one
requires
full
explanation.
There's
full
write
up
with
the
demo
in
an
example.
So
this
can
be.
This
is
gonna,
be
a
critical
first
part
here.
Essentially,
it's
we're
gonna
write.
These
json
schemas
paired
with
adrs
and
they're
gonna,
be
used
to
validate
they're,
going
to
be
used
to
validate
the
effectively
the
schema.
A
The
adr
defines
the
contract
that
accepting
this
as
a
so
think
of
it,
like
think
of
everything
like
we're,
building
a
big
distributed
system
right
and
we're
basically
going
to
have
this
sort
of
like
sin,
synoc.
A
Awk
flow,
I
gotta
go,
get
my
t-shirt,
and
so
you
know
basically,
let's
just
not
assume
that
something
gets
done.
You
know
if
it
just
because
the
job
got
accepted
is
what
I'm
trying
to
say.
So
you
know
basically
here's
the
job
and
then
you
know,
sometime
later,
I'm
finished
with
the
job
and
then
are
you
finished
with
the
job,
something
like
that,
but
that'll
all
be
facilitated
by
adding
more
links
to
the
chain.
So
so,
basically
what
we've
got
is
we've
got
these.
A
You
know
these
directed
acyclic
graphs,
which
form
our
data
flows
right
and
when
one
of
the
changes
that
we're
going
to
make
is
to
instantiate
the
the
operations
via
passing
them
in
as
inputs
to
an
operation,
and
that
way
we
are
able
to
support
this
shared
config
stuff,
because
we
can
pass
by
reference
to
the
instantiated
objects.
In
a
particular
sequence.
A
We
can
support
the
sort
of
secret
unlocking
from
various
sources,
because
we
can
chain
together
arbitrarily
complex
sequences
of
operations,
to
do
that
and
yeah.
It
does.
Oh.
That
was
another
big
thing
yeah,
but
that
wasn't
an
issue
particularly
so
when
we
do
this,
basically
we're
gonna
have
this
idea
of
cached
cache
data
flows
right,
and
so
this
is
like.
I'm
gonna
save
a
data
flow,
I'm
gonna
load
a
dataflow.
We've
been
talking
about
this
forever
right.
A
We
just
got
the
model
stuff
working,
great
great
job,
everybody
who
was
involved
in
that
but
yeah.
So
this
is
the
more
generic
form
right
and
we
were
playing
with
this.
We
were
like
what
do
we
do?
What
do
we
do?
What
do
we
do?
We
couldn't
figure
it
out.
A
I
mean
we've
been
playing
with
this
for
years
like,
and
so
I
I
believe
the
trick
is
in
sort
of
you
know,
and
we've
talked
about
unification
of
the
of
the
command
line,
interfaces
and
the
operations
and
the
classes
themselves
and
the
data
flows,
and
then
the
python
type
system
right,
like
all
this
stuff,
has
to
come
together
or
else
it's
just.
We've
got
this
kind
of
clunky
mess
right
now
right.
We
know
there's
something
going
on,
but
it's
a
little
bit
of
a
mess
right.
Apis
aren't
consistent.
A
A
Basically,
what
I'm
saying
here
is
that
I've
been
doing
some
cleanup,
conceptually
conceptual
cleanup
and
and
now
it's
time
to
go
plan
out
how
that
how
that
hits
the
code
and
what
new
features
are
involved
there
and
what
that,
how
the
future
road
map
impacts
that
cleanup
so
we're
going
to
basically
refactor
as
we
go
and
we're
going
to
map
our
refactoring
activities
to
our
feature
development
activities
to
our
bug,
fix
activities
and
we're
going
to
take
those
we're
going
to
map
them
into
these
chapters,
which
is
this
chapters
of
this
tutorial
series
that
we're
going
to
write
and
yeah
we're
going
to
have
a
we're
going
to
have
a
hell
of
a
great
time.
A
I
hope
so.
Okay,
so
so.
So,
basically-
and
I
think
I
I
I
I
rolled
on
a
train
of
thought
there,
but
I
was
talking
about
the
cache
data
flows
and
it
all
being
the
same
thing.
A
So
what
we
need
to
do,
I
believe,
is
we're
going
to
test
this
out
and
we're
going
to
test
all
this
stuff
out
is
we
need
to
have
a
concept
of
like
a
top
level
flow
and,
and
that's
not
necessarily
to
say
it's,
the
top-level
flow
and
we're
also
going
to
rely
heavily
on
the
on
the
overlay
concept
as
well
for
a
lot
of
this
stuff.
A
So
so
we,
if
you
think
about
the
overlays
right,
the
overlays,
are
effectively
and
I
think
there's
what
the
one
of
the
one
of
the
examples
covers.
The
overlays,
I
believe
it's,
the
the
the
one
where
we
take
the
automated
classification
data
and
the
should
I
data,
and
then
we
run
them
together
in
the
same
data
flow
via
the
http
service.
Okay.
So
what
does
this
represent?
A
This
represents
the
upstream
tracking
model
right.
So
so
what
is
this
overlay?
Well,
the
overlay
is
the
diff
right.
The
overlay
is
that
the
upstream
is
basically
the
ups
upstream
is
basically,
I
guess
I
should
just
make
myself
big,
because
I'm
just
talking,
maybe
I
should
talk
to
alice.
Okay,.
A
All
right,
so,
where
are
we-
and
this
is
why
we're
also
doing
this
videos,
because
I'm
gonna
we're
gonna
we're
gonna,
we're
gonna,
have
alice
we're
gonna
go
through
youtube's
transcription
or
we're
gonna
have
alice.
Do
this
transcription
we'll
turn
this
all
into
notes
and
we're?
Basically,
I
just
want
to
have
data
on
everything
everything
everything
all
the
time
and
that's
that's
part
of
what's
going
to
happen
here.
Okay,
so
we're
talking
about
importing
and
exporting
data
and
that
needing
to
have
the
top
level
contact.
A
So,
basically,
when
we
run
when
we
run
and
when
we
run
okay,
yeah.
A
All
right
so
you'll
notice
that
this
maps
onto
my
areas
of
focus,
because
you
know
probably
gonna
author-
a
fair
amount
of
this.
This
tutorial
series,
I
would
hope,
other
people,
you
know
want
to
author
it
with
me
here,
but
you
know
this
is
this
is
basically
this
is.
This
is
what
I'm,
seeing
as
the
places
that
are
really
need
to
be
nailed
down
to
build
alice,
and
so
these
are
the
things
that
I
am
therefore
involved
in
and
so
yeah.
That's
why
I
wrote
them
here.
A
If
you,
if
you
see
other
areas
that
are
critical
in
the
building
of
you
should,
you
know,
propose,
add
an
edit
to
this
document,
and
this
will
be
something
that
will
be
available
as
a
pull
request
to
comment
on
and
you
can
say,
hey,
you
know,
we
really
need
to
look
at
this
area
right,
and
so
this
is,
you
know
where,
where
we
can,
where
we
blink
blank
subfield
of
computers
is
where
we
need
to
look
to
understand.
What's
what's
the
state
of
art
happening
in
this
area?
A
How
does
it
apply
to
you
know
what
what
our
ai's
understanding
of
software
engineering
is.
So
please,
you
know
be
on
the
lookout
for
that:
okay,
god,
there's
so
much
stuff,
and
part
of
this
is
about
how
to
organize
it.
So
I'm
just
trying
to
get
everything
out
there,
okay,
so
okay,
so
the
importing
we.
This
is
just
that
frame
of
thought
is
obviously
way
too
far
gone,
we'll
we'll
cover
it
all.
I
think
what
are
we
gonna
get
to?
A
Let's
just
do
tonight's
thing
and
then
we'll
kind
of
go
piece
by
piece.
I'm
gonna
write
some
enough
stuff
up
more
and
then
I'll,
give
more
more
coherent
and
organized
thought
versions
of
what
I've
been
blabbing
on
about,
but
you
can
read
the
full
basic.
A
I'm
writing
down
all
of
my
thoughts
related
to
this
and
it's
all
here
and
I
am
asking
everybody
to
please
write
down
their
thoughts
as
well,
so
we
can
all
go
build
this
thing,
because
this
is
gonna
be
ridiculous,
so
yeah,
it's
gonna,
be
there's
gonna,
be
a
lot
of
work
involved.
It's
like
a
years
long
effort
across
you
know
n
amount
of
people
who
want
to
be
involved
so
just
putting
it
all
out
there.
A
So
we
can
all
just
like
figure
out
who
needs
to
do
what
right,
I'm
sure,
we're
all
thinking
about
this
stuff,
so
I
figured
it
might
be
nice
to
have
some
place
to
to
think
think
together.
That's
really
what
this
is
about
actually-
and
that's
all
in
here
too,
is-
is
to
to
come
together
to
think
together,
okay,
okay,
so
all
right!
Well,
let's
just
go
look
at
the
web
3
stuff.
I
hope
this
is
running.
A
A
Oh
yeah,
and
so
why
do
we
want
to
do
this?
Well,
basically,
things
are
going
this
way.
I
I
don't
think
anybody
would
argue
that
right,
we
know,
there's
gonna
be
blockchain
web3,
whatever
it's
it's
sort
of
a
given.
We
know
that
there
is
some
sort
of
next
step
in
what's
a
commodity
as
far
as
compute
is
concerned.
Right
so-
and
I
think
you
know
part
of
what's
gonna
happen
here-
is
that
this
commodity
is.
Is
this
this
ai?
A
Data
ai
operation
hybrid-
I
think
I
have
notes
on
it
somewhere,
but
and
these
are
basically
like
clumps
of
models
and
these
clumps
of
models
just
like
that
dna
that
you,
that
sort
of
will
create
for
this,
these
little
alices
that
live
on
chain.
These
are
like
entities
in
a
way
because
it
is
it
functions
like
dna.
So
it's
like
it's
like
software
dna
software
is
dna
because
basically
we
won't
need
to
write
any
code.
A
I
don't
think
because
I
think
I
think
alice
will
just
be
able
to
write
all
the
code,
because
I
mean
look
at
like
copilot
and
stuff
right.
So
I
think
that
the
main
thing
here
that
we're
trying
to
do
is
we're
trying
to
map
intent
to
code,
and
once
you
can,
map
intent
to
code,
she
can
go.
A
Learn
fast,
is
the
guess,
and
so
that's
what
we're
trying
to
do
here
so
and
and
then
obviously
part
of
this
is
going
on
these
overlays
right,
so
we're
gonna
overlay
all
these
strategic
plans.
A
On
top
of
these
things
and
right,
you
know
part
of
what's
going
on
with
machine
learning,
is
that
you
know
sometimes
we're
not
so
sure
how
things
get
done
right
and
sometimes
we're
not
so
sure,
if
we
remember
to
add
our
extra
input,
validation
on
the
output
of
our
model
to
check
for
these,
you
know
corner
cases
if
we're
looking
at
you
know
potentially
malicious
input,
data
right,
it
would
be
really
nice
effectively
to
have
static
analysis
across
all.
A
All
of
you
know
every
language
to
to
map
intent
to
code
in
that
language,
and
basically
you
know
alice-
is
going
to
need
to
do
that,
and
so
once
you
have
that
the
question
really
becomes
well
okay.
So
where
do
you
go
from
there
right?
Well,
how
do
you
decide
what
to
do
next?
Well,
how
do
any
of
us
decide
what
to
do
next
and
we're
going
to
follow
this
sort
of
mental
model
of
human
thought
to
figure
out,
you
know
how
should
our?
How
should
our,
how
should
our?
A
A
It's
it's
like
this
manifest
and
the
manifest
allows
for
you
know
arbitrarily
deep
levels
of
nested
manifests
which
is
sort
of
like
your
plug-in
instantiation,
and
all
of
these
you
know
subsequent
and
is
very
similar
to
like
the
kubernetes
api
server
crd
concept,
right,
where
you're,
defining
these
different
api
types
and
stuff
right
or
like
an
open
api,
v3,
spec
or
something
right.
A
So,
basically,
just
you
know
give
me
all
the
metadata
give
me
the
function
prototypes
right
and
if
you've
got
the
function
prototypes
about
everything
you
can
sort
of
just
and
you
understand,
intent,
mapped
to
the
data
involved.
You
can
sort
of
just
start,
you
know
building.
Also
you
can
build
all
sorts
of
stuff
right,
that's
that's
what
we're
gonna
do,
and
so
you
can
then
map
that
to
abstract
concepts.
A
By
doing
essentially
like
this,
I
believe
it
a
sort
of
two
by
two
matrix
conversion
like
in
in
basically
like
an
image
you
can
think
of
it
like
doing
doing
doing
a
gan
if
you're
doing
like
an
image
scan,
and
you
want
to
take
one
picture
and
make
it
in
the
style
of
another
picture
right
like
you're
painting
as
another
painter,
right
or
or
like.
A
In
this
case,
I
believe
what
it's
going
to
be
like
is
like
you're,
seeing
through
somebody
else's
eyes
right
effectively,
because
you
can
apply
all
these
sort
of
like
views
and
these
views
or
the
views
on
the
data
right
and,
and
so
so.
A
The
views
on
the
data
are
effectively
these
strategic
plans,
which
are
effectively
just
like
output
operations,
which
we're
going
to
overlay
on
top
of
data
flows,
and
this
this
is
coincidentally
and
then
probably
not
so,
coincidentally,
but
exactly
what
we
were
doing
for
this
data
flows
is
class.
Pull
requests
that
that
some
have
noticed
is
stale,
so
we're
going
to
get
back
to
that
right
away,
and
it
also
ties
into
this
concept
of
caching,
where
we're
going
to
god.
Damn
it.
I'm
on
another
rabbit
hole.
A
This
is
why
also
alice
in
wonderland,
okay,
so,
basically
we're
going
to
build
all
these
models.
The
models
are
going
to
help
us
understand
the
data
in
a
language
that
we
understand
right
and
so
therefore,
a
fundamental
piece
of
this
whole
data,
ai
sort
of
compute-
in
that
it's
that's
kind
of
wrapped
up
in
the
manifest
thing
with
the
spec.
A
Is
this
ability
to
sort
of
transform
one
view
of
something
into
another
view
of
something
right
like
anything,
you
can
sort
of
take
a
and
get
b.
As
long
as
you
give
me
some
context
around,
you
know
what
my
parameters
are.
A
You
know
you
know,
take
take
this
picture
of
a
fish
and
make
it
look
like
man
go
painted
it
right
but,
like
you
know,
take
the
concept
of
democracy
and
make
it
look
like
bank
van
gogh
painted
it
right,
like
that's
the
type
of
stuff
we're
talking
about
here
and
doing
that
like
in
sort
of
like
this
multi-dimensional
space
by
training
all
these
models
against
each
other.
A
And
that
way
you
can
say
you
know
like
mix
the
concept
of
democracy
and
picasso
or
whatever
right
to
create
like
code.
It
sounds
ridiculous,
but
it's
it's.
I
think
it's
possible,
I
think
it's
100
as
possible,
and
so
because
it's
basically
just
a
it's
everything
is
basically
like
a
encoder
there's.
A
lot
of
it
is
basically
like
an
encoder
decoder.
A
Translate
like
a
language
translation
model,
as
far
as
I
can
see,
but
we're
gonna
find
out,
and
so
essentially
the
point
of
this
is
that,
since
you
can
see
anything
through
anybody's
eyes,
you
can
get
together
as
this
organizationally
formed
groups
of
people
where
all
of
the
agents
sort
of
vote
for
these
strategic
plans
to
be
in
place
across
all
of
the
data
flows.
And
this
is
where
how
you
can
create
this
kind
of
like
system
for
iot
devices
and
all
of
this
stuff
to
where
you
can
go
out
in
the
world.
A
And
you
can
see
it
and
they
have
like
this
attestation
information
and
you
can
verify
whether
they're
running
these
strategic
plans
or
whether
they've
accepted
input
from
a
strategic
decision
maker
or
strategic
planner,
which
is
something
that
you
trust
right
or
that
that
you
are
okay
with
right
like
so,
you
could
be
at
the
grocery
store
and
you
could
wave
your
phone
over
a
set
of
items
right
and
you
know
it
could
do
the
image,
detection
and
the
barcode
reading
just
by
going
over,
and
it
could
look
up
the
company
and
information,
and
it
could
then
map
that
to
what
you
care
about
right.
A
So
maybe
I'm
like
you
know
I
care
about.
If
this
all
I
care
about
is,
is
this
food
keto
right
and
I
can.
It
will
like
look
up
calorie
counts
on
the
fly
right
like
that
is
my
organization
that
I'm
I'm
I'm
I'm
I'm
a
part
of
the
keto
organization
and
and
so
and
then
you
could
also
be
like
you
know.
A
Is
it
you,
you
see
what
I'm
saying
like
this?
Is
I
picked
the
keto
one,
because
it's
got
multiple
steps.
You've
got
to
look
up.
The
calories
and
stuff
you've
got
to
look
at
what
food
and
then
what
calories
right
and
you.
So
basically,
you
want
assurance
that.
A
Okay,
all
right
we're
just
going
to
look
at
the
did
stuff
and
I'm
going
to
write
the
plans
because
there's
just
too
much
it's
just
too
much
to
talk
about
it's
just
it's
too
much,
all
right,
okay.
So
how
is
this
structured?
So,
basically,
this
is
the
mission.
Why,
if
you
can
do
this,
I
can
see
is
somebody
else.
I
can
see
as
a
whole
organization,
and
I
can
understand
what
the
whole
organization
thinks
is:
freedom
and
privacy
preserving
and
then
I
can
basically,
like
you
know
only.
A
You
know
like
I
can
only
shop
at
at
stores,
who
I
know
are
supporting
these
strategic
plans
or
something
you
know
you
can
get
arbitrarily
complex
and
that's.
Why
that's
why?
The
first
thing
that
we're
doing
is
we're
going
to
look
at
this
web
3
stuff,
because
we
want
to
build
from
the
solid
foundation
here
and
if
we
can
just
plot
all
this
data.
A
I
don't
know
what
on
chain
means,
but
on
chain
or
something
like
that,
like
I
don't
know,
maybe
maybe
that's
the
way
we're
supposed
to
do
it,
but
we're
not
going
to
do
that.
I
mean
we're
not
going
to
put
all
the
data
on
there.
That's
what
the
manifesto
for
in
the
little
tiny
you
know,
versions
of
alice
in
these
tiny
dna
strands
so
that
we
can
put
the
smallest
amount
of
information
on
chain
as
possible.
A
Let's
go
look
at
the
spec
okay,
this
thing
is
full
of
full
of
stuff.
I'm
gonna
dump
it
to
markdown
documents,
okay,
so
the
id
oh!
This
is
in
here
like
a
million
times.
A
I've
been
writing
this
for,
like
a
while.
Now
I've
been
writing
okay,
so
this
is
the
most
relevant
comment.
Can
you
see?
Yes,
you
can
great
sorry
about
the
4k,
but
I
I
gotta
I
gotta
have
the
4k
to
figure
this
out.
Okay,
yeah
we're
gonna
have
to
see
a
lot
of
things
at
the
same
time.
Okay,
so
basically
I
think
dwn
was
distributed.
A
Web
node,
the
id
personal
data
store
and,
like
I
said,
we're
not
going
to
touch
the
personal
data
store
thing
right
now,
just
because
you
know
I,
I
don't
think
it's
worth.
I
don't
think
it's
worth
getting
into
that.
I
at
this
point
I
think
from
what
I
understood.
A
It's
not
I
don't
know
if
they
have
like
enough
examples
that
we
could
just
jump
into
it
yet
and
I
think
it's
more
relevant
to
the
authentication
and
authorization
problems
that
we'll
run
into
down
the
line
we're
just
going
to
ignore
those
for
now
you
know,
as
as
as
we
as
as
one
does
and
the
way
that
we
ignore
those
is
not
by
writing
insecure
code,
it's
by
writing
secure
code.
A
Just
writing
not
that
much
of
it,
because
it
takes
a
long
time
to
write
so
right,
one
secure,
complete
implementation
and
then
don't
write
a
million
little
examples
and
and
and
then
yeah
you
get
where
we're
at.
So
that's
why,
although
my
sql
source
has
the
tls
on
and
everything
so,
okay,
so
basically
yeah,
we
just
don't
we're
not
gonna,
do
any
examples
that
are
gonna,
be
you
know,
auth
related
and
then
we're
not
gonna
need
to
to
deal
with
that
so
much,
at
least
in
a
complicated
sense.
A
What
does
this
say
so
I
like
to
this
is
just
I'm
just
gonna
give
full
thought
process
for
the
sake
of
the
recording,
so
I
like
to
hit
the
middle
mouse
click
and
and
open
every
link.
I
I
find-
and
this
is
why
there's
so
many
links
on
here-
basically
I'll
skim,
something
and
then
I'll
go
through
and
I'll
go
boom.
I
need
that
boom.
I
need
that
and
boom.
A
I
need
that
and
then
I'll
go
through
here
then
I'll
click
I'll
give
one
each
one
a
quick
control,
tab
kind
of
see
which
ones
I
do
need
and
don't
need
close
the
ones
I
don't
need,
and
then
you
know
I
go
back
to
the
duplicate
right.
Okay,
here
we
are
now
we're
good.
Now
we
know
what
we
need
and
sort
of
like
a
binary
search
for
information
as
soon
as
you
open
all
the
links
right.
A
Okay,
so
this
is
the
spec
and
here's
the
the
quick
summary
of
the
spec,
and
you
know
yet
another
reason
listed.
Why
that's
to
pick
the
name
alice's,
because
it's
already
in
a
million
people's
stocks,
so
so
yeah.
Basically,
I
saw
this
picture
and
I
was
like
this
looks
great.
A
This
looks
like
the
problems
that,
like
gun,
is
trying
to
solve
and
and
things
like
that,
and
we
I
I
may
have
tried
to
rope
a
few
people
into
a
project
called
chill
chat
a
while
ago,
you'll
notice
things
like
like
very,
very
similar,
so
basically
asymmetric,
keys
and,
like
you
know,
sign
this.
I
sign
this.
You
sign
that
you
know
sort
of
like
here's.
My
data
pass
it
along.
A
I
have
to
sign
it
for
everybody,
that's
kind
of
where
it
gets
annoying
and
then
you
have
to
like
you
know,
grant
claims
to
other
people
to
do
things
with
your
data.
What
do
you
trust
them
to
do
once
again
same
thing
as
loops
back
into
the
operations
discussion
and
executing
all
of
this?
So
all
right,
I
should
probably
read
this
so
they
just
talk.
Okay,.
A
And
what
is
this?
Where
are
we?
Is
this
decentralized
web
node?
Okay?
So
this
is
interesting.
Maybe
we
should
read
this
first
because
I
think
this
seems
much
more
helpful,
no
offense
to
anybody
who
wrote
the
did
spec,
but
I
I
didn't
really
understand
what
was
going
on
here
and
this
seems
to
to
give
up
give
more
of
a
concrete
example.
A
So
decentralized
web
note
is
a
draft
specification
under
the
development
of
the
decentralized
identified
identity
foundation,
incorporates
requirements
and
learners
from
related
work
and
many
active
industry
players
into
a
shared
specification
that
meets
the
collective
needs
of
the
community.
The
specification
will
be
updated
to
incorporate
feedback
from
idf
members
they're,
not
id.
I
d,
I
f
members
and
wider
community
with
a
reference
implementation
being
developed
within
dif
that
okay,
perfect
reference
implementation.
I
love
to
hear
that
at
that
point
you
can
basically
you're.
A
Basically,
oh
this
guy,
this
guy's
great,
this
guy
knows
what
he's
doing
I've
been
following
him
on.
A
That's
how
found
out
about
this
stuff
as
long
as
well
as
I
saw
kelsey
start
taking
it
more
seriously,
and
I
saw
angie
jones
start
taking
it
more
seriously
and
then
I
started.
Then
I
saw
this
guy
daniel
and
I
saw
what
is
her
name
lincoln.
Brooklyn
said
linka
and
she's
very
smart.
I
could
tell
she
seemed
very
smart,
so
I
was
like
these
are
some
people.
I
should
go,
listen
to
and
sure
enough.
They
were
right
and
so
they're
leading
some
really
cool
stuff
here
and
okay.
A
A
You
can
see
why
I'm
thinking
that
the
logical
step
here
is
the
build
building
ai.
That
does
the
same
thing:
implementations,
javascript,
fantastic.
A
A
A
Like
that's
the
that's
why
I'm
also
interested
in
the
web
note
is
because
I
found
this
thing
that
like
implies
that
you
can
sort
of
claim
a
name
namespace
in
a
distributed
fashion
by
putting
it
under
a
dot,
well-known
directory,
which
is
apparently
some
spec,
that's
been
around
from
like
the
90s
or
something
I
don't
know
it's
been
around
for
a
while,
but
I
was
surprised
I
was
surprised.
Nobody
has.
I'm
surprised.
I
haven't
heard
more
people
using
this.
A
This
seems
like
a
really
great
way
to
anchor
root
of
trust
on
a
domain
for
an
alternate
protocol.
I'm
like
very
pleased
okay,
so.
A
Validated.Validate,
let's
see,
are
we
getting
any
of
this?
Are
we
not
oh?
No.
I
hope
that
we're
getting
this.
A
A
Oh
god,
this
stuff?
I
am
really
bad
at
this
stuff.
Oh
yes,
there's
me,
I
am
live.
Oh
look!
There's
one
person
watching,
thank
god.
I
hope
that
works
out.
I
know
somebody
will
watch.
A
A
This
is
ridiculous:
okay,.
A
A
A
A
What's
a
d
id
question
mark
and
then
we
can
figure
out
the
the
goal,
and
then
this
is
part
of
why
we're
doing
this
series
of
videos,
because
when
we
say
hello
moth,
when
we
say
what's
a
d
id,
it
would
be
nice
if
alice
could
just
chime
in
on
the
side
and
say:
oh
by
the
way
you
know.
Oh,
this
is
what
we're
we're
figuring
out,
but
this
is
what
a
cid
is
right.
That
would
be
great.
A
A
A
Basically,
I
think
I
think
I
think
I
think
I
think
it's
a
good
way
to
have
a
job.
I
think
so
I
think
so
in
the
event
that
you
lose
you
win.
A
A
A
I'm
guessing
that
the
cid
is
the
so
this
is
probably
where
in
ipfs
the
did
lives.
So
this
is
probably
like
I
was
saying.
Is
it
yeah?
I
think
this
is
prop
data,
I'm
not
even
reading
this.
Okay
read
it.
If
I
could
just
read
one
line
at
a
time,
I
might
figure
out
what
things
mean.
Okay,
let
ipfs
equals
optional.
A
A
Or
is
this
not?
Where
is
this
no,
this
interface
where's
the
where's
the
entry
point
for
this.
That's
what
we
should
be
asking.
This
was
a
bad
idea
of
me.
I
should
have
asked
that
this.
This
would
be
the
correct
thing
to
do
would
say:
where
is
the
entry
point
for
this
who's?
This
I've
seen
him
around
yeah.
A
A
Why
would
you
take
us
to
a
reduction
one
time
watch
this
look
at
this.
Look
at
this
look.
What
happens,
look
what
happens
just
just
just
imagine
imagine
imagine
you
really
really
really
really
want
to
know
more
about
this
file
format.
This
is
you're
desperate
to
know
you
type
in
e.
L
f,
really
excited
to
see
a
nice
design
created
of
the
whole
layout.
A
A
A
These
wonderful
people,
these
people-
are
great
technical
specification
development.
Oh
I
love
this
already.
A
God
this
is
okay,
this
is
beautiful
is
what
this
is
so
basically,
okay,
I
like
this.
I
like
this
a
lot.
A
A
A
I
think
claire's
project
is
going
to
take
this
even
further
woohoo
woohoo,
claire
and
michael
yay.
That's
going
to
be
great.
I
think
we're
going
to
be
able
to
merge
that
with
this.
Even
if
we
merge
that
with
this,
we
should
merge
that
with
this,
we
should
merge
that,
with
this.
A
Yeah,
yes,
this
is
basically
this
is
effectively
show
me
the
notes,
so
there's
a
demo
in
there.
That's
basically
like
hey
rewrite
these
notes.
Basically,
the
the
goal
is
take,
take
stream
of
audio
with
unrelated
topics
and
create
cohesive
thoughts
out
of
them.
It's
the
task,
so
obviously
we're
going
to
need
to
build
an
ai
to
figure
this
one
out.
A
A
And
I
mean
you
can
basically
take
that
and
scale
that
up
to
every
single
thing
that
you
see
on
the
internet
right
make
sense
of
it
all
so
or
in
this
did
web
okay.
So
basically
what
they
have
is
they
have
something
kind
of
like
our
console
test
situation,
where
they've
got
their
test
cases
and
their
docs
and
they're
linked,
and
so,
when
you
do
spec
up
no
watch
true.
So
there's
no
watch.