►
From YouTube: NERSC Nobel Lecture Series: George Smoot, June 3, 2014
Description
George Smoot, Lawrence Berkeley National Lab
2006 Nobel Prize in Physics
"Mapping the Universe"
A
Okay,
I
wanted
I
wanted
to
thank
you
all
for
for
coming.
This
is
our
third
lecture
in
the
nurse
keynote
lecture,
seminar,
series
and
and
we've
actually
gotten
great
attendance
on
all
of
them,
and
so
we
appreciate,
I
recognize
the
number
of
people.
I
know
that
a
number
of
you
come
to
each
of
these,
and
this
has
been
a
fantastic
opportunity
to
learn
more
about
the
ways
in
which
computation
is
making
an
impact
on
on
leading
edge.
A
Science
they're
also
about
a
probably
somewhere
around
100
or
200
people
who
are
watching
online
and
then
we're
also
making
the
lectures
available
so
that
people
can
watch
them
later,
and
so
so
I'll.
Ask
professor
smoot
to
tell
me
if
the
the
lights
are
too
bright,
but
they're.
A
For
the
it's
for
the
recording,
see.
A
So
professor
smoot
is
an
astrophysicist
cosmologist
and
nobel
laureate,
who
has
been
recognized
worldwide
for
his
groundbreaking
research
in
cosmic
microwave
background
radiation
and
the
origins
of
the
universe.
Professor
smoot
was
one
of
the
first
astrophysicists
to
devise
ways
to
conduct
experiments
that
produce
data
and
information
about
the
early
universe,
using
tools
such
as
the
cosmic
background
explorer
a
satellite
launched
in
1989.
A
He
and
his
colleagues
made
the
first
measurements
of
temperature
fluctuations
in
microwave
background
radiation
and
set
the
stage
for
later
measurements.
These
findings
also
enable
them
to
discover
the
seeds
of
present-day
galaxies
and
galaxy
clusters
and
map
the
early
universe
in
unprecedented
ways.
A
In
2006,
professor
smoot
and
fellow
astro
physicist
john
c
mather
were
awarded
the
nobel
prize
in
physics
for
their
work,
with
the
cosmic
background
explorer,
which
validated
the
big
bang
theory,
professor
smoot
has
been
the
recipient
of
the
einstein
medal,
the
gruber
prize,
the
orster
dead
medal
and
the
ernest
o
lawrence
award
he's
been
involved
in
observational,
astrophysicist
and
physics
and
cosmology
research
at
berkeley
lab
since
the
early
1970s
and
is
a
professor
of
physics
at
uc.
A
B
Thank
you.
I
was
invited
here
by
several
people,
some
of
whom
you
see
in
the
room,
but
one
of
them
was
horse
simon,
who
liked
to
remind
me
how
old
I
was,
and
he
said
you
had
some
great
pictures
of
old
computers
and
so
forth.
Can
you
please
show
those
and
somewhere,
and
so
he's
forced
me
to
face
the
fact
that
I'm
a
little
old
and
then
50
years
ago,
two
things
happened
that
have
had
a
profound
impact
on
my
career.
B
So
one
of
the
obvious,
namely
the
cosmic
microwave
background,
was
discovered
50
years
ago.
It's
actually
discovered
50
years
ago,
but
published
one
year
later
and
the
the
second
one
was
that
I
learned
how
to
use
computers
for
scientific
research,
and
so
those
were.
Those
were
two
really
major
events
that
profoundly
influenced
my
my
career
I'll
talk
about
them
indirectly,
along
this
way,
but
like
always
and
anywhere
you're,
doing
research
or
anywhere
you're,
making
technical
progress.
There
are
always
precursors
so
I'll
remind
you.
B
There
was
an
eniac
computer,
which
was
the
first,
the
first
electronic
general
purpose
computer
and
that
was
in
1946
and
it
had
17
000
vacuum
tubes,
which
means
it
was
working
about
half
time,
because
the
vacuum
dudes
don't
last
very
long,
but
that's
the.
So
if
you,
if
you
really
link
back
the
guys
that
had
the
foresight
to
build
this
computer,
which
cost
about
the
same
as
the
supercomputer
system
today
in
terms
of
what's
going
on
they're
the
ones
that
started
this
path
and
they
even
started
a
course
once
it
was
released
and
made
now.
B
They
even
started
a
course
to
train
people
how
to
build
computers
and
how
to
design
and
build
computers,
and
so
there's
a
lot
of
seeds
that
come
from
that.
The
other
thing
that
happened
near
the
same
time
and
both
of
these
are
sort
of
outgrowths
of
activities
that
happened
during
world
war
ii,
as
are
several
other
things
and
was
the
prediction
of
the
cossack
background
radiation.
B
So
in
1948,
george
gamau,
who
was
a
refugee
from
the
soviet
union
in
earlier
days,
had
come
to
the
united
states
and
worked
in
the
war
effort
and
worked
in
nuclear
physics
with
edward
teller.
In
fact,
ian
teller
did
a
lot
of
work
together.
He
got
teller
out
of
the
out
of
the
soviet
bloc.
B
They
were
able
to
estimate
the
temperature
of
this
cosmetic
background
should
be
around
5
kelvin.
They
published
it.
It
was
mostly
forgotten
for
a
while,
but
in
1965
vincent
wilson
published
their
results.
So
here's
the
pictures
and
they
got
the
nobel
prize
in
78
and
so
forth.
I'll
come
back
in
a
second
so
immediately.
This
confirmed
because
there
turned
out
to
be
gas
clouds
that
have
molecules
in
them
and
the
cosmic
microwave
background
were
exciting.
Some
of
those.
So
very
quickly
people
had
old
measurements,
they
could
reinterpret.
B
So
within
a
year
the
cosmic
microwave
background
I
forgot,
I
had
this
fancy
tool.
You
know
the
cosmic
ray
microwave
background
had
had
observations
from
the
you
know:
ch
molecules,
cn
cyanogenic
molecule
and
from
the
berkeley
group
from
the
princeton
group
princess
wilson
made
two
measurements.
They
originally
made
the
measurement
here,
because
that's
what
bell
labs
was
paying
to
do
was
to
see
if
they
could
send
radio
signals
to
the
sky
and
back
and
they
need
to
know
the
background.
B
They
wanted
a
measure
of
21
centimeters,
so
they
had
build
a
receiver
for
that
in
order
to
measure
the
hydrogen
and
so
already
within
the
first
year.
You
were
starting
to
trace
out
this
curve,
and
this
is
the
prediction
of
a
three
kelvin
black
body
right,
so
that
was
pretty
good.
Didn't
take
too
long
65
to
the
announcement
in
65
to
1978.
B
You
know,
11
years
later,
after
the
nobel
prize,
they
declared
this
a
national
historic
site,
so
historic
site
is
takes
longer
than
a
nobel
prize.
So
just
bear
that
in
mind.
But
if
you
hadn't
destroyed
my
lab
and
the
equipment
on
the
roof,
it
would
be
a
national
historic
site
and
two
more
years
night
2017.
B
That's
the
way
it
goes
right
so,
back
to
this,
it's
a
big
antenna,
so
bell
labs
had
built
this
antenna
in
order
to
look
at
the
telstar
satellite
and
then
it
hired
pencils
and
wilson.
Wilson
was
actually
straight
out
of
graduate
school.
The
first
paper
we
ever
wrote
was
the
discovery
of
the
cosmic
microwave
background.
Good
start,
okay,
and
so
this
is
their
equipment,
and
I
went
to
visit
their
equipment,
which
is
in
the
deutsches
museum
in
munich,
and
I
took
this
picture
and
this
picture
particular
picture.
B
I
pulled
out
because
this
is
the
computing
power.
Okay,
you
guys
may
not
recognize
this
as
computing
power.
This
is
a
thing
we
call
a
strip
chart
recorder,
okay,
this
is
before
this
is
the
old
days
of
analog
computers
and
doing
things
by
hand
I'll
give
you
an
idea
what
that
looked
like.
Okay,
so
remember,
here's
the
antenna,
you
see
it
can
rotate
here's
the
scan
where,
if
you're
at
90
degrees,
you're
looking
through
the
least
atmosphere
and
as
you
swing
down
towards
the
horizon,
you
see
this,
which
should
be
as
a
cosecant
curve.
B
You
see
the
temperature
from
that.
You
can
estimate
the
temperature
of
the
atmosphere
and
then
you
do
the
trick.
This
is
the
best
slide
I
happen
to
have
at
the
time
of
the
well
here's
the
equipment
in
the
museum
but
of
the
actual
discovery
chart,
but
I
found
this
now.
This
is
how
we
used
to
take
data
in
the
old
days
the
really
old
days.
B
This
is
the
strip
cart
chart,
there's
a
calibration
here
on
cassipia,
a
a
nice
known
radio
source
and
then
the
calibration
where
they
switch
between
the
sky
and
the
cold
load
and
then
do
the
zenith
scans
you
saw
later
on.
So
this
is
the
discovery
and
you
see
what
we
used
to
do
is
use
these
chart
recorders
and
just
count
the
number
of
little
squares
and
estimate
what
the
level
was
and
write
it
down
and
do
additions
and
subtractions
right.
That's
how
we
did
it
right.
Okay,
so
that
was
the
first
impact
you
can
see.
B
B
B
Now
I
was
working
at
mit
and
the
laboratory
for
nuclear
science
bought
one
of
these
one
of
the
early
ones
of
these,
and
I
was
very
proud
because
I
took
the
course-
and
I
got
my
own
key,
so
I
could
analyze
data
using
this
ibm
360.
and
I
you
know,
could
come
in
at
night.
You
know
when
nobody
else
was
here.
Was
there
and
use
my
key
and
run
the
computer
and
put
the
data
in
and
do
it
now
just
in
case,
you
guys
want
to
know
what
computer
nerds
look
like
in
those
days.
B
This
is
well
if
you're
at
ibm,
you
have
to
wear
kind
of
a
suit,
but
they
don't
wear
the
blue.
You
know
they
don't
wear
the
you
know
the
kind
of
stuff,
but
you'll
notice,
they're
registers,
they're
flashing
lights.
That's
why
they're
still
in
science
fiction
film
flashing
lights
on
the
computers,
because
this
was
the
one
that
really
started
it
all
and
here's
the
picture
she's,
not
a
nerd
she's,
a
model
that
was
hired
by
ibm
to
advertise
it.
Ours
is
blue.
B
Most
of
the
panels
are
blue,
but
some
of
them
were
red,
but
so
the
thing
that
I
remember
about
this
is
not
only
was
it
great
as
a
young
kid.
I
got
to
run
this
super
computer
right
because
that
was
the
superhero
in
those
days.
But
the
other
thing
I
remember
very
much
is
the
punch
cards,
so
you
guys,
may
not.
B
B
This
is
computer
played
music
right,
so
it's
very
different
and-
and
there
were
people
who
eventually
learned
how
to
do
things
when
we
got
to
faster
computers,
how
to
actually
put
your
radio
on
the
computer
and
play
through
the
play
play
the
sounds
of
the
circuits
changing
through
the
radio.
So
that's
the
good
old
days
right,
so
the
you
know
the
the
computer
center.
B
You
know
there
are
lots
of
little
little
chads
and
things
like
that
that
we
we
had
in
those
days,
but
here
was
the
breakthrough
that
made
a
lot
of
a
science
possible
for
us
digital
equipment.
Corporation
was,
while
I
was
at
the
same
area
building
in
their
in
their
garage
or
their
parents
garage
they
what
turned
out
to
be
the
pdp-11.
B
So
what
was
great
about
that?
Instead
of
being
in
this
big
temperature,
controlled
room
with
humidity
everything
it
was
in
the
garage
and
had
to
work
in
a
garage
that
meant
we
can
build
stuff
that
could
go
out
in
the
field
and
take
data
right.
That's
the
and
the
other
thing
that
it
had
was
it
had
a
real-time
operating
system
with
a
foreground
task.
So
you
could
run
two
jobs
at
once.
B
B
We
had
a
big
step
forward
with
these
computers
because
we
started
having
terminals
and
we
had
that
the
p
what
people
did
was
they
adapt
teletype
terminals.
So
this
is
the
paper
tape
reader,
and
this
is
the
this-
is
the
sort
of
sending
messages
like
equivalent
of
of
telegraphs
and
so
forth,
and
then
we
had
a
big
step
forward
here.
So
this
is
a
picture
of
one
that
that
we
had.
It
was
a
big,
a
big
advance.
Look
at
this
terminal
compared
to
I
mean
that's,
really
modern
right!
B
Well,
you
don't
quite
feel
that
way
and
then
the
thing
the
next
big
step
forward
that
really
made
the
difference
for
us
was
that
came
out
with
a
lsi
1123.
lsi
stands
for
large
scale,
integrated
circuits
no
longer
was
it
whoops.
I
gone
too
many
here,
no
whoops,
sorry,
no
longer
was
it
transistors
right
and
my
first
introduction
to
ray
weiss,
who
some
of
you
may
know
of
as
the
person
who
invented
gravity
wave
laser
interferometer
systems
right.
B
C
B
B
Okay,
that
meant
we
didn't
no
longer
have
to
have
the
foreground
program
in
less
than
32k
of
memory.
It's
you
know
it's
it.
It
was
a
little
tricky
to
get
everything
to
work
in
that
kind
of
kind
of
days.
Nowadays
we
we
don't
think
that
our
cell
phones
should
should
have
less
than
four
megs
of
memory,
or
it's
going
to
be
impossible
right
and
then
the
lsi
11
20,
the
73,
came
on
a
little
while
longer
and
it
ended
up
to
four
megs
okay.
So
why
was
this
useful?
B
Well,
it
allowed
us
to
run
experiments
real
time.
So
here's
an
actual
picture,
one
of
the
ones
horse
wanted
me
to
show
you.
This
is
an
lsi
11,
I
think
23.
It
might
have
been
a
73
by
then
these
are
disk
drives.
I'm
sure
these
were
several
megabytes,
because
otherwise
we
had
these
big
floppy
disks
and
there's
a
big
floppy
disk
reader
in
here
and
there's
the
old
hello
type
and
here's
a
newfangled
crt
right,
and
but
this
was
when
we
were
building
the
experience
we're
doing.
B
This
is
the
prototype
for
kobe,
where
we're
trying
to
measure
very
tiny
differences.
So
we
look
at
two
different
parts
of
the
sky
and
measure
the
difference:
interchange
the
horns.
So
it's
on
a
rotating
table.
So
we
could.
We
could
do
that.
So
this
was
a
prototype
for
what
was
going
to
go
on
the
kobe
satellite.
So
this
we
were
building
in
the
1970s.
B
At
the
same
time,
we
were
starting
to
make
measurements
of
the
spectrum
of
the
cosmic
microwave
background,
which
we
thought
was
the
relic
radiation
from
the
big
bang.
But
we
had
to
prove
it
because
you
know
it:
it
never
had
curled
over.
Yet
we've
just
seen
the
the
shape
and
we
want
to
measure
it
more
precisely
so
we
I've
formed
an
italian
american
collaboration
and
we
actually
ran
seven
instruments
and
a
lot
of
other
activities
here
in
this
cold
place,
where
we
made
it
colder
with
nitrogen,
liquid
helium.
B
The
computer
was
in
this
hut
right
it
got
treated
special,
it
got
to
be
inside
the
rest
of
us
had
to
sit
out
in
the
snow
and
the
same
thing.
We
went
to
the
south
pole
same
group.
Now
you
can
tell
this.
It's
really
italian
americans,
because
we
had,
we
decided
we'd,
take
this
picture,
doing
that
and
I
have
a
close-up.
So
you
can
tell
that
I
was
there
see
well
with
the
lights
on.
It
doesn't
come
out
so
bright,
but
it's
not
the
ceremonial
south
pole.
B
This
is
the
real
one,
or
at
least
the
real
one
that
day,
because
the
antarctica
chief
is
moving
a
little
bit,
but
the
geological
survey
comes
and
put
a
puts
a
little
stamp
in
every
year
and
they
move
the
sign
so
that
you
know
where
it
is
right
because
they,
actually
you
know
the
gla.
It's
a
big
glacier,
as
you
can
see,
it's
almost
two
miles
thick
and
it's
moving
moving
slowly
and
but
that
wasn't
the
south
pole
wasn't
remote
enough
and
cold
enough.
B
For
us,
we
went
about
a
kilometer
or
two
away
and
dig
a
big,
a
big
pit,
and
we
had
this
screen
to
keep
us
from
the
wind
so
we're
out
here
on
a
balmy
day,
taking
it,
we
have
our
liquid
nitrogen
liquid
helium
around
to
make
it
going
on
at
the
time
we
were
doing
these
measurements
up
in
the
mountains
and
at
the
south
pole
we
went
to
south
pole
twice,
you
know
their
other
deck
was
making
some
more
advances,
so
dex
first
advance
was
to
go
from
a
12
bit
to
a
16
bit.
B
This
was
the
major
thing.
32
bits
you
guys
cannot
imagine
how
much
that
opened
up
right,
because
you
know
it's
just
it's
just
kind
of
so
we
we
got
the
vaxes
and
we
took
our
legacy
software
from
our
pdps
and
we
migrated
them
into
the
maxis,
and
then
we
could
do
all
kinds
of
crazy
things,
and
eventually
we
created
dec
alpha
clusters
and
by
that
time
I
realized
how
important
computers
were
to
us.
So
we
actually
were
rolling
our
budget
where
we
were
buying
some
new
computers
every
year.
B
So
we
were,
our
cluster
was
always
upgrading
right.
So
nurse
has
a
different
problem.
It
has
to
get
a
whole
new
cluster
every
now
and
then
right.
So
at
some
point,
you'll
figure
out
how
to
do
rolling
clusters,
but
we'll
see
well,
hopefully,
but
anyway,
we
eventually
built
a
cluster
farm
like
this,
in
which
we
did
the
data
analysis
for
kobe
and
our,
and
it
was
very
challenging
for
us
to
do
that,
even
though
it
seemed
like
we
had
a
simple
problem.
B
We
had
a
6144
by
6144
matrix
and
many
multiple
right
hand
sides
to
the
matrix
equation.
We
had
to
invert
that
and
find
the
solutions
and
that
actually
ended
up
taking
us
about
three
months
of
running
in
order
to
do
it
on
our
on
our
alpha
cluster,
which
the
time
was
like
30
alphas
right,
it's,
they
were
really
fast
and
they
were
really
cool.
At
lbl
we
were
using
the
walnut
computer
as
we
used
to
know
it
affectionately,
the
cdc
7600.
B
I
don't
know
if
any
of
you
guys
are
old
enough
to
remember
that.
But
this
is
the
beginning
of
making
computers
in
a
more
compact
configuration
in
order
to
not
be
limited
by
the
speed
of
of
signals
right
just
set
by
the
speed
of
light
and
the
other
thing
that
was
the
great
improvement.
Was
we
had
the
people's
printer
and
card
reader
right?
We
knew
it
here
as
a
user,
but
we
had
some
visitors
from
the
soviet
block.
C
B
And
there
was
two
great
advances
here:
you
could
go
in
and
read
your
own
programs
in
the
computer
and
you
get
your
own
printout,
so
you
could
run
your
job
anytime
right.
So
thank
you,
computer
center.
For
doing
that.
That
means
we
could
be
running
stuff
all
the
time,
but
the
really
big
thing
was
they
started
to
allow
us
to
store
images
of
our
programs
in
the
computer.
They
gave
us
a
little
space
in
the
computer.
Where
no
longer
do
I
have
to
read
five
file
doors
worth
of
cards
in
every
time.
I
want
to.
C
B
Really
big
long
program
with
a
lot
of
data,
I
could
just
run
an
update
deck,
so
it
would
be
a
little
thin
deck.
That
said,
please
replace
these
cards
in
the
program
with
these
these
cards
and
then
run
the
program,
and
that's
that
just
made
your
life
a
lot
easier
right
and
you
guys
didn't
come
along
until
we
actually
had
terminals
where
you
could
do
this
all
remotely
and
there
weren't
cards
there
weren't
anything
like
that,
but
that
was
the
steps
along
the
way.
B
D
B
Neglect
saying
simulations
are
not
to
be
very
important.
We
were
working
on
the
instruments
for
kobe,
so
fire
a
stands
for
far
infrared
absolute
spectrophotometer
and
what
it
was
designed
to
do
is
the
same
trick
that
we
were
doing
before.
There's
a
big
antenna
that
looks
at
the
sky,
there's
a
small
antenna
that
looks
at
an
internal
load
whose
temperature
is
controlled
to
whichever
temperature
you
want,
because
we
ran
the
dewar
about
2.7
kelvin.
B
We
could
heat
it
up
and
down
so
we
could
know
what
was
going
on
and
then
we
had
a
thing
to
replace
the
universe.
This
long
thing,
which
is
protected
that
could
rotate
into
place,
and
so,
instead
of
looking
at
the
universe,
we'd
be
looking
at
this
calibrator,
which
we
very
precisely
made,
and
so
as
soon
as
we
got
up
there
and
started
taking
data,
we
knew
the
answer
because
here's,
the
predicted
black
body
curve
and
here's
our
error
bars
multiplied
by
400
right
now,
that's
pretty
good!
B
So
that's
the
first
part
of
what's
going
on
here's
the
prediction
of
a
2.72
kelvin
black
body
at
least
dashed
line.
Your
red
is
the
fire
ass
data.
You're
our
ground
base,
you
know
from
wide
mountain
and
from
the
south
pole
your
species
and
wilson's
original
point
there's
a
some
of
the
other
ground-based
stuff.
That's
in
there
and
there's
a
couple
balloons
and
so
forth,
but
you
can
see
and
here's
the
difference
and
you
see,
byras
errors
are
incredibly
small.
There
was
a
university
of
british
columbia
crumb
university,
a
british
columbia
rocket.
B
That
was
very
much
like
a
small
version
of
fire
ass
and
it
was
launched
about
a
year
after
less
than
a
year,
afterwards
called
cobra,
and
it
made
this
sort
of
blue
curve
that
sort
of
verified
that
we
were
in
the
right
ballpark.
So
I
have
one
more
that
has
more
blown
up.
There
was
a
tremendous
amount
of
work
to
make
these
measurements.
You
know
going
all
around
the
world
and
doing
that,
look
at
how
that's
nothing
compared
to
the.
B
We
got
from
doing
it
in
space
and
doing
it
precisely
and
analyzing
the
data
fairly
in
a
fairly
complicated
way.
It's
a
it's,
a
fourier
transform
spectrometer,
but
there
are
many
other
issues
in
there.
So
the
data
analysis
actually
took
a
fair
amount
of
time,
but
its
computing
constraints
were
not
as
severe
as
measuring
the
the
temperature
variations.
So
here's
a
picture
of
kobe
when
it
was
at
vandenberg
air
force
base
and
I
went
down
to
inspect
to
make
sure
the
shields
there's
a
long
story.
We
were
supposed
to
go
on
the
shuttle.
B
So
we
took
these
pictures,
so
this
is
the
door
in
which
fire
ass
and
the
derby
are
that's
covered
over
and
here's
two
of
the
three
dmr
instruments
so
they're
back-to-back
horns,
that's
the
one
you
saw
the
prototype
of
this
is
at
a
frequency,
that's
twice
as
high,
so
the
wavelength
is
twice
as
short
and
there's
one
that's
three
times
a
sign
on
the
other
side
that
that
are
shown
there.
So
this
is
us
the
final
time
that
we
saw
the
satellite
up
and
close,
and
I
figured
at
the
launch
that
was
the
last
time.
B
I
would
see
it,
but
that
night
I
did
see
it
again
because
it
was
in
orbit
and
you
was
in
the
terminated
orbit.
So
it
was
in
the
sun
and
I
was
in
the
dark
and
you
could
see
it.
So
this
is
what
it
looked
like,
except
no
sun
was
on
the
inside.
It's
just
signing
on
the
solar
cells.
There's
three
soil
cells.
B
It's
rotating
like
a
chicken
on
rotisserie
to
keep
the
temperature
uniform,
but
it
scans
nice
circles
in
the
sky
and
by
doing
that,
and
by
having
it
precess
with
the
earth
going
around
the
sun,
we
were
able
to
map
the
whole
sky
and
we
made
these
maps.
The
first
map
which
is
off
here
is
the
universal
diffuse
glow.
It's
just
green,
so
I
didn't
show
you
that
data,
but
I
have
some
real
data
that
shows
that
and
then,
if
you
go
down
a
factor
of
a
thousand,
you
see
what
we
call
the
dipole.
B
So
if
you
look
90
degrees
that
way,
you're
looking
down
the
spiral
arm,
so
you
see
a
bump
if
you
look
90
degrees,
the
other
way
you're
also
going
down
the
spiral
arm.
So
you
see
these
two
bumps,
which
are
the
spiral
armor
and
the
galactic
center.
Is
there
that's
why
this
doesn't
have
a
nice
smooth
thing?
The
galactic
center
is
contributing.
B
If
you
then
subtract
away
the
dipole
and
blow
the
scale
up.
Another
factor
of
100
is
the
galaxy
saturates
along
here
the
galactic
center,
the
spiral
arm
that
we're
in
the
opposite
anti
anti-center
direction.
But
then
you
see
these
blue
regions
together,
red
and
yellow
regions
together,
red
and
yellow
regions
together,
blue
regions
together
blue
regions
together.
Those
are
the
things
that
we
claimed
are
the
are
the
signals
from
the
beginning
of
the
universe,
and
here
is
with
the
all
the
known,
galactic
contamination
stuff.
B
You
know
stamped
out
and
you
can
see
the
signal
that's
left
behind,
but
if
you're
more
adventuresome
you
can
make
a
plot
in
colors,
so
here's
the
dipole
and
if
it's
stronger
in
the
red,
it
shows
up
red
if
it's
right
in
the
right
cmb
it
shows
up
in
in
white.
So
you
see
hot
to
cold.
It's
supposed
to
be
white
to
dark.
You
take
away
the
dipole
below
the
scale
up
now.
You
can
still
see
the
spiral
arms,
but
you
can
begin
to
see
the
signal.
B
That's
coming
from
the
universe
and
if
you
take
a
model
of
the
galaxy
out,
you
end
up
with
a
map
that
looks
like
that
right
so
that
but
modeling
the
galaxy
is
the
tricky
part.
So
we
did
with
cuts
and
we
did
with
with
subtractions
and
got
pretty
similar
answers,
but
we
were
being
careful,
okay
and
so
not
so
long
after
that
we
did
two
more
experiments.
E
C
B
There's
an
actual
map
there.
The
size
of
the
moon
is
about
this
big.
So
this
covers
the
reasonable
part
of
the
sky,
but
you're,
seeing
very
clear
structures,
and
when
we
first
when
I
first
saw
this
map-
and
we
were
having
our
team
meeting,
I
said
the
universe
is
flat
and
radix
stomper
who's.
One
of
some
of
you
may
know
he
says:
no,
no,
don't
don't
jump
to
conclusions.
B
You
get
these
bumps
and
wiggles
and
so
forth,
and
those
bumps
and
wiggles
we'll
I'll
show.
You
will
tell
us,
what's
the
universe
made
out
of
right
and
how
much
ordinary
matter,
how
much
dark
matter?
How
much
dark
energy
so
forth
that
those
bumps
and
wiggles
allow
you
to
do
that.
Well
about
this
time
that
we
were
doing
this,
the
the
maps
out
of
the
microwave
anastasia
probe
later
renamed,
the
wilkinson
micro
vendor
scientific
probe,
because
one
of
our
collaborators
and
and
and
also
competitors
wilkinson,
died
during
the
during
the
mission.
So
they
renamed
them.
B
B
But
now
you
have
a
big
antenna,
so
you
have
much
smaller
angular
resolution
and
so
wmap,
instead
of
being
in
the
terminated
orbit
near
the
earth,
gets
down
send
out
to
the
earth's
sun,
lagrange
0.2
and
that's
the
place
where
the
combined
gravity
of
the
sun
plus
the
earth
causes
it
to
orbit
at
the
same
time
as
the
earth.
So
you
always
keep
that
same
jump,
the
all
coal
orbit,
it's
almost
stable,
but
not
quite
stable,
and
that
keeps
the
sun,
the
moon
and
the
earth
in
the
backward
direction.
B
B
B
So
the
issue
was:
how
do
you
compute
this
orbit
to
get
it
there?
Okay,
non-trivial
I'll,
show
you
this
other
one,
but
but
think
about
that
orbit.
This
is
just
pretty
pictures,
so
I
can
talk
about
that
computers
couldn't
do
it
in
those
days
there
was
a
guy
who
could
kind
of
visualize
it
and
get
the
right
range,
and
then
our
computers
in
the
us
could
make
that
calculation,
the
russians
and
the
european
space
agency
had
to
build
more
powerful
rockets,
so
they
could
go
direct
eventually.
We
eventually
the
us
has
done
that
too.
B
B
I
don't
my
rocket's
not
powerful
enough
to
take
this
thing
out
to
l2,
but
if
I
get
a
gravity
assist
boost
from
you
know
from
the
moon,
I
can
make
it
right
and-
and
you
all
have
heard
about
using
gravity-
assist
from
jupiter
because
that's
a
easier,
simpler
calculation
to
do,
but
here,
where
you
have
to
do
this
orbit
and
sync
yourself
to
the
moon
and
avoid
a
lot
of
stuff,
that's
a
little
tricky!
Okay!
B
So
that's
how
you
get
out
to
this
region
where
you
want
to
have
the
sun
out
here
the
earth
and
the
moon
behind
you
and
you're
and
you're
out
here
rotating
around.
You
know
like
like
the
chicken
on
the
barbecue,
but
now
only
your
backside
is
getting
warmed
right,
so,
okay,
so
up
in
the
corner.
I
have
one
last
little
thing
here.
So
here's
the
the
sun,
the
earth
now
here's
the
shadow
of
the
earth.
B
Now,
if
you're
going
to
a
lot
of
trouble,
keeping
your
stuff
stabilized,
you
don't
want
to
go
into
eclipse.
That
is,
you
don't
want
to
go
in
the
shadow
because
there's
a
big
temperature
change,
so
the
orbit
is
the
orbit
around
the
unstable
place.
So
you
need
a
little
station
keeping.
So
if
you
look
the
orbit
as
it
goes
around
has
to
orbit
around
and
avoid
that
now
here,
you're
you're
starting
to
do
computing
power,
you
have
to
make
these
maps
and
then
you
have
to
combine
them
and
subtract
away.
B
The
galaxy
make
a
different
model
for
the
galaxy
and
separate
out
and
make
your
map
of
what
the
temperature
anaesthetics
look
like.
Now,
that's
not
so
hard
right.
It's
the
sky,
except
these
variations
are
about
apart
100
000
of
the
total
signal
coming
from
the
sky
or
the
of
the
main
signal
coming
from
the
sky,
and
it's
actually
less
than
a
part
and
a
hundred
thousand
for
all
signal,
because
at
the
lowest
frequency
the
galaxy
is
fairly
bright.
B
But
when
we
get
the
plank,
we're
doing
10
million
by
10
million
and
we're
trying
not
to
approximate
we're
we're
solving
for
iteration
and
so
forth,
we're
trying
not
to
approximate
so
here
is
the
cobie
data
and
then
we're
going
to
neatly
transform
it
into
wmap
data.
So
if
you
paid
attention,
you'd
know
that
they're
almost
the
same,
but
just
in
case
you
weren't
paying
attention
here,
you
can
do
the
comparison
right,
dark,
dark,
red
red.
The
big
features
are
there
both
of
them.
B
What
kind
of
rotation
angle
is.
This
is
just
separation
angles
and
you
see
the
cobie
data
was
down
in
this
region.
This
is
the
first
peak
and
the
second
peak
and
now
we're
beginning
to
fill
in
the
third
fourth
and
the
fifth
peak
right.
So
now
we're
in
the
early
2000s
and
every
every
step
along
the
way
we've
challenged
computers.
First,
it
was
to
actually
just
take
the
data.
B
Now
it's
the
process,
the
data,
but
the
level
of
data
we
have
so
much,
and
sometimes
I
see
people
doing
experiments
now
where
they
say
we
don't
want
to
actually
make
that
much
of
a
trigger.
Let's
just
record
all
the
data,
and
I'm
going
oh
well,
you
can
now,
but
that's
not
the
design
experiment
before
we
could
not
record
all
the
data
we
had
to
make
decisions
we
had
anyway
and
hydrogen
physicists.
They
have
to
make
triggers
and
it's
always
an
issue,
but
there's
a
lot
of
experiments.
B
Now,
where
people
just
record
all
the
data
and
they've
got,
you
know
some
number
of
mini
terabytes
or
petabytes
of
data
and
they've
got
to
sift
through
it
to
find
that
little
bit
of
stuff.
That's
good
here,
we're
still
really
dragging
the
computer,
but
from
that
we
have
made
what
we
call
our
current
model
of
cosmology.
So
imagine
there's
a
sphere
around
us.
B
This
is
a
slice
of
it
and
we
follow
that
sphere
backwards
in
time,
and
we
we
have
this
model
that
that
you'll
presumably
hear
some
about
in
the
next
lecture
from
saul
that
we
have
this
period
of
accelerated
expansion,
which
we
label
as
caused
by
the
dark
energy
and
we're
trying
to
find
out
more
about
it.
And
as
you
go
back
that
that
period
say
roughly
a
third
of
the
way
back,
it
no
longer
is
accelerating
it's
slowing
down.
B
That's
when
the
structure
form
when
the
first
stars
and
first
galaxies
form
and
they
merge
in
the
large
scale
structure
form
and
eventually
you
got
the
modern
galaxies.
Our
solar
system
had
to
form
at
roughly
this
period,
but
that's
partly
accidental,
because
once
you
start
making
galaxies,
you
start
making
solar
systems.
B
So
we
can
see
and
then
we're
worried
these
days,
so
we're
looking
back
here
to
before
through
the
period
we
call
the
dark
ages
before
there
were
stars
back
to
the
time
when
the
universe
was
as
hot
as
the
sun,
but
it's
expanded
a
thousand
times
since
then.
So
this
is
not
quite
the
scale,
but
you
know
even
log
scale
that,
but
this
is
what
the
universe
looked
like
when
it
was
400,
000
years
old
compared
to
13.7
billion
years.
B
So
it's
roughly
the
difference
between
you
now
and
you
12
hours
after
conception
in
terms
of
time
scales,
and
if
we
go
back
further,
the
universe
gets
much
smaller
until
we
think
there's
a
period
way
back
here
when
there
was
inflation
a
period
when
the
universe
went
through
as
much
expansion
in
a
fraction
of
a
second
that
has,
as
it
has
in
the
14
billion
years
since
then,
and
at
that
point
all
hundred
billion
galaxies.
That's
in
the
sphere.
Here
they
were
back
in
a
size,
much
smaller
than
atom
and
quantum
fluctuations
became
important
right.
B
So
it's
a
simple
model
and
we
can
calculate
that
because
we
have
powerful
computers
and
simple
algorithms
and
in
fact
one
of
the
people
who
helped
do
that
is
one
of
the
professors
here.
So
he
was
at
that
time
a
grad
student
postdoc
was
euros,
shell
jack,
and
he
and
matthias
zelda
made
some
of
the
code
that
allows
us
to
simulate
billions
of
millions
of
universes
in
short
times
in
terms
of
what
goes
on.
B
They
make
some
approximations
but
they're,
not
bad,
so
the
cmb
angular
power
spectrum
now
measured
for
and
we
put
in
our
review.
Oh
you
can
see
it
with
the
lights
on
here.
This
is
surprising.
Usually
we
try
to
make
a
very
faint
yellow
for
our
best
fit
model,
but
I
didn't
point
to
you
out
to
you
before
on
that
plot,
that
I
showed
you
before.
Even
in
2001,
we
still
had
a
what
we
call
the
cold
dark
matter
universe.
B
We
didn't
have
the
dark
energy
in
it,
even
though
it
had
been
discovered
by
then
it
took
another
couple
of
years
before
we
started
having
what
we
call
lambda
cdm
or
dark
energy
dark
matter
universes
right,
but
this
is
the
best
fit
sort
of
lambda
cdm
universe
so
including
the
dark
energy,
and
now
you
can
see
with
wmap
and
with
the
new
measurements
that
are
coming
in
you're.
Seeing
the
second
third
fourth
fifth
heads
of
the
sixth,
but
not
quite
the
seventh
peak,
yet
so
we're
making
progress
well,
why?
Why
do
you
care
about
this?
B
B
Well,
the
first
thing
you
can
do
is
look
at
the
total
amount
of
energy
which
is
equivalent
to
looking
at
the
curvature
of
space-time
I'll,
come
back
to
it,
and
you
find
out
that
shifting
that
around
basically
just
shifts
this
power
spectrum
left
and
right
and
if
the
universe
is,
is
more
open.
It
shifts
this
way,
it's
more
close.
It
shifts
that
way
and
there's
a
little
bit
of
stuff
that
happens
down
in
the
really
large
angular
scales.
B
If
you
put
in
dark
energy,
it
looks
like
you
can't
tell
much,
because
it
only
happens
down
here
where
the
this
kind
of
activity
goes
on,
but
now
our
cmb
measurement
getting
so
precise.
There
are
even
small
variations
here
that
we
have
that
we
can
begin
to
pick
out
and
make
some
summer.
But
baryons
is
the
physicist's
word
for
like
protons
and
neutrons.
B
C
B
All
matter
it
pushes
all
the
peaks
up,
so
you
can
actually
by
looking
at
the
ratio
of
the
second
peak
to
the
first
and
the
third
peak
you
can
tell
how
much
is
dark
matter
versus
how
much
is
ordinary
matter
and
so
on.
So
there's
a
whole
bunch
of
stuff
there's
six
parameters.
You
can
fit
for
the
six
parameters
very
very
carefully,
so
the
issue
we
have
now
is:
are
we
gonna
have
to
add
more
parameters
or
we're
gonna
in
that?
So
I'm
gonna.
B
I
promise
I
would
talk
to
you
about,
is
the
universe
flat
and
the
answer
is
very
simple.
We
look
here
with
our
eyes
and
we
judge
how
big
things
are
by
projecting
back
in
straight
lines
and
if
the
universe
is
curved,
the
light
will
travel
a
curved
path.
Otherwise
we
get
this
nice
euclidean
geometry
and
an
the
universe
which
is
a
little
trickier
or
if
it's,
if
it's
open
you,
you
have
the
other
thing.
So
I
have
this
nice
video
this.
This
tells
you
the
same
pattern
on
the
sky.
B
B
Here
these
are
made
to
look
to
be
the
same
levels.
The
levels
change
a
little
bit
in
those
situations.
You
that's
how
you
can
begin
to
recognize.
The
universe
is
close
to
flat.
So
I
don't
say
it's
exactly
flat,
but
it's
really
close
to
flat.
So
that's
that's
interesting.
That's
an
example!
What
you
can
learn
you
can
do
the
same
thing
for
how
much
ordinary
matter
is.
How
much
dark
matter
is
you
can
compare
it
to
the
actual
observations
or
you
can
compare
to
the
power
spectrum?
You
can
you
get
some
results?
B
So
we've
made
some
progress
in
this
field.
We
went
from
kobe
with
6144
pixels
by
6144
pixels
and
then
the
thing
that
was
launched
in
89
our
results
came
out
in
92
made
a
big
splash
and
in
2000
wmap
was
launched.
The
results
came
out
in
about
2003
the
main
results.
It
had
many
frequencies
that
in
principle
had
revolution
to
here,
but
when
you
had
to
take
out
the
galaxy
you
have
to
use
the
largest
wavelength
scale,
and
that
gives
you
poor
resolution.
B
So
you
got
pixels
that
look
like
that
and
in
2000
may
2009
we
launched
the
planck
satellite,
which
was
very
interesting
because
it
was
you
know,
a
rainstorm
before
and
a
rainstorm
after,
and
clouds
and
stuff
going
on.
But
this
is
from
from
kuru
and
so
they're
used
to.
C
B
In
french
guyana,
they're
used
to
having
to
launch
through
rainy
season
and
they
managed
to
launch
it
very
spectacularly
successfully,
and
it
has
nine
different,
wavelengths
and
and
more
sensitivity,
but
also
more
wavelengths
and
more
angle
resolution.
So
you
get
this
much
higher
resolution
map
and
this
part's,
the
smudgy
part
where
you
get
reasonable
polarization
data.
These
are
to
represent
polarization
vectors.
B
So
here's
a
picture
just
to
show
you
the
plant
can
do
it.
So
sorry,
I
was
giving.
I
used
pulled
a
slide
from
a
talk
I
just
gave
where
they
like
russian.
So
here
are
the
antennas.
Here's
the
secondary
the
primary.
This
whole
thing
is
out
at
l2.
It's
a
warm
on
one
side
and
cool
there's
a
set
of
radiators,
so
it
cools
down
to
quite
cold
temperature,
and
then
we
have
refrigerators
on
board
that
bring
some
of
the
detectors
down
below
a
tenth
of
a
kelvin.
B
B
Now
it's
going
to
unfold
and
rotate
because
the
magic
of
computers
is,
you
can
put
it
in
a
nice,
a
map
and
a
nice
representation
that
you
can
use.
So
here's
the
plane
of
a
galaxy,
the
the
two
spiral
arms
that
we're
in
it's
what
it
looks
like
in
the
optical.
Sorry,
it
was
too
fast
for
my
and
the
resolution
isn't
quite
good
enough.
So
I'll
show
you
so
then
you
can
take
your
different
wavelengths
fit
to
the
three
components
of
the
galaxy
in
a
some
mystery
component.
B
If
you
want
to
allow
for
it
or
marginalize
over
it
and
then
for
the
cosmic
microwave
background,
which
shows
pretty
clearly
here
to
begin
with,
and
you
get
a
map
that
looks
like
this.
So
this
looks
pretty
much
like
the
wmap
map,
except
now,
there's
even
more
speckles
in
spots,
but
it's
still
dominated
by
the
stuff.
That's
on
the
0.9
degree
scale
you.
F
B
The
the
huge
the
big
peak
is
on
the
structural
scale.
That's
the
the
the
acoustic
horizon,
the
sound
horizon
at
the
time.
The
universe
gets
to
be
about
the
same
temperature
as
the
surface
of
the
sun.
So
there's
planck
doing
its
job
right
out
in
the
universe
and
what
we
have
is
we
have
a
series
approximately
on
10
year
increments.
I
don't
think
the
next
one
will.
C
B
In
tenure,
I
think
we're
going
to
fall
off
the
moors
curve,
but
we'll
see
what
happens
so
cope
is
the
same
piece
of
the
sky.
Kobe
barely
detected
stuff
on
large
scales,
but
it
detected
it
and
saw
that
was
right
and
motivated
people
to
go
on
w
map
and
that
same
part
of
the
sky
sees
much
more
structure
and
planck
sees
a
lot
more
structure.
B
And
now
we
have
you
know
the
south
pole,
telescope,
the
anaconda,
cosmology
telescope
and
polar
bear,
which
is
centric
one,
making
even
higher
angular
solutions
sort
of
measurements
in
this
kind
of
regime.
I
didn't
happen
to
have
a
beautiful
plot
of
that,
so
this
this
map
that
we
have
this
this
concept.
B
Well,
you
know
it
was
pretty
good
at
the
time.
So
one
of
the
things
we
did
was
we
made
this
angular
power
spectrum,
and
this
is
our
best
fit
model
right.
This
green
curve
is
the
best
fit
model.
Now
you
can
actually
count
seven
peaks,
we've
gotten
gotten
quite
far,
so
this
this
is
up
to
date
in
terms
of
what's
going
on,
and
now
we're
worrying
about
little
things
these
anomalies,
this
data
point's
a
little
low.
These
data
points
are
low.
That
was
a
little
high.
B
We're
worried
about
little
things
now
that
are
deviations
from
our
model,
but
in
terms
of
of
our
our
overall
fit,
we
have
things
in
in
an
amazing
shape,
considering
that
that
the
first
discovery
of
these
fluctuations-
the
few
points
down
in
here
that
was
23
years
ago-
and
now
we
have
mapped
the
entire
curve
out
with
high
precision,
but
we've
done
even
more
than
that,
and
so
I
have
to
since
I'm
running
close
to
the
end
of
my
time.
I
have
to
skip
over
a
couple
things
and
go
straight
to
bicep.
B
So
here's
a
slide
I
pulled
out
of
a
talk
I
gave
in
2005
of
the
bicep
instrument,
as
it
was
being
deployed
to
the
south
pole.
Now
some
of
you
guys
heard
about
them
last
month,
claiming
they've
seen
signals
for
the
evidence
for
gravity
waves
in
the
early
universe,
which
is
really
fairly
much
directly
evidence
for
the.
B
B
You
know,
1000
element,
arrays
and
so
forth
that
that
that
are
going
and
many
thousands
soon
to
be
many
thousand
element,
arrays
that
are
going
on
the
next
generation
of
experiments
that
some
of
them
are
actually
operating,
and
so,
if
these
guys
are
right-
and
that's
part
of
the
debate
what's
going
on,
if
these
guys
are
right,
we're
in
a
regime
where
we're
we're
entering
in
in
the
fourth
and
final
phase
of
looking
at
anastasia,
how
can
I
say
that?
B
Well,
you
can
make
a
prediction
if
our
model
is
correct,
so
I've
been
talking
mostly
about
the
temperature
power
spectrum,
and
this
is
the
temperature
power
spectrum
that
comes
from
scalars.
That's
our
technical
term.
What
scalars
mean
is
energy
density
fluctuations.
You
can
also
have
metric
fluctuations
that
is,
fluctuations
in
the
fabric
of
space
and
time.
B
Those
are
tensor
fluctuations
and
when
they
come
inside
the
horizon,
they
can
be
propagated
and
freely
and
become
gravity
waves,
at
least
the
the
space
space
part
of
them
can
so
this
as
soon
as
you
measure
this
part,
you
predict
all
the
rest
of
everything
on
this
left-hand
side
of
the
plot.
You
predict
the
whole
power
spectrum
that
you
should
see
here.
B
You
predict
the
cross-correlation
between
the
temperature
and
what
we
call
the
e-mode
polarization,
the
polarization
that
that
has
a
simple
patterns
around
the
hot
gold
spots
caused
by
the
fact,
if
there's
an
over
density
of
matter,
there's
an
over
density
of
photons,
and
that
means
when
the
photons
are
stringing
out,
they
scatter
and
they
make
a
certain
polarization
pattern,
and
that
also
predicts
the
e.
What
the
mode
polarization
pattern
has
this
stuff.
B
That's
on
our
way
from
the
beginning
of
the
universe
to
us,
and
they
make
a
slight
distortion
of
the
pattern
and
that
that
that
that
feeds
down
to
the
b
moments
right,
but
we
had
no
clue
of
what
the
tensor
variation
is
going
to
be,
because
you
have
a
degree
of
freedom
that
lets
you
change
the
energy
scale
of
inflation.
This
is
set
by
the
energy
scale
inflation.
This
is
set
by
the
equation
of
state
of
inflation.
B
There's
one
extra
parameter,
so
you
can
adjust
the
two
roll
of
each
other,
but
as
soon
as
you
make
one
measurement,
and
now
it's
very
hard
because
the
temperature
is
buried
in
this
temperature
and
the
te
everything
is
buried
in
the
other
things.
But
if
you
measure
the
b
modes
and
and
what
bicep
has
done
is
tried
to
measure
what
they
think
as
a
primary
b
modes
in
this
region,
if
you
measure
that
it
predicts
everything
here
and
so
now,
we
have
to
see
if
it
all
fits
together,
right
and
so
we'll
see.
B
What's
going
on,
so
it's
more
complicated
and
more
complicated
means
more
computation.
So
here's
the
pattern
that
bicep
claims
to
have
seen
notice.
It's
swirly!
That's
why
you
saw
some
advertisements
as
well
as
vorticity,
see
so
here's
a
hot
spot.
It
goes
around
it,
whereas
the
e
modes
only
go
out
radially
or
tangentially,
they
don't
go
in
a
swirl
right
and.
A
B
B
A
Thank
you
very
much.
These
have
all
been
well
attended,
but
this
has
definitely
set
the
record
there's.
I
guess
they're
standing
room
only
so
since
this
is
being
recorded,
we
have
time
for
questions
but
I'll
hand.
The
mic
to
you
so
that
that
people
online
can
hear
so
questions.
B
F
The
fact
that
the
universe
universe
is
flat
does
that
imply
the
universe
is
also
infinite.
B
Okay,
so
I
said
it's
flat,
but
I
used
the
term
a
little
loosely.
I
said
if
you
heard
a
couple
of
times.
I
was
very
careful,
and
I
said
it's
very
close
to
flat,
and
so
it
depends
on
your
view.
There
are,
there
are
actually
multiple
possibilities,
one
of
them
which
I
actually
had
a
grad
student,
get
her
thesis
on
and
was
you
can
think
of
the
universe
as
as
being
simply
connected
so
that
it
has
a
finite
size
and
then
the
edges
are
identified.
B
So
I
don't,
if
you
remember
the
old
days
of
pong,
there
was
a
there
was
a
war.
You
know
a
space
war
thing
where
there
was
a
sun
that
you
could
get
pulled
into
by
gravity,
but
when
the
space
ship
went
off
one
side
it
appeared
on
the
other.
The
universe
could
be
like
that,
but
it
is
a
simulation
right.
You
know,
and
so
well
you
can't
prove
it's
not
actually,
but
the
you
can
imagine
a
universe.
A
B
Large
there's
a
lot
of
things
you
can
do
and
if
you're
on
a
budget,
you
might
think
that
it's
like
putting
mirrors
up
in
a
in
a
tiny
apartment
right
or
something
like
that,
and
but
the
trick
you
can
use
is
if
there's
a
carpet
and
the
carpet's
not
perfect,
it
will
have
some
ripples
in
it
and
there's
a
ripple
that's
way
longer
than
what
you
think
the
cell
size
is
then
you're
wrong.
The
cell
size
is
bigger
right,
so.
B
Right
in
the
universe-
and
so
you
can
do
it
from
the
galaxy
surveys,
you
can
do
it
from
the
cosmic
microwave
background.
So
that's
that's
one
way
so,
but
that
still
could
be
outside
of
horizon
and
we
haven't
got
the
data.
Yet
right
I
mean
it's,
you
don't
know
what
the
dragons
could
be
out
there,
and
so
we
don't
know
that,
but
the
other
is,
if
you
actually
look
at
making
the
universe.
B
There
are
different
approaches
that
people
have
of
creating
universe
from
nothing
or
from
almost
nothing
or
from
some
other
kind
of
universe
and
butting.
Our
part
off
and
many
of
those
wanted,
universes
open,
which
in
principle
should
be
infinite.
Many
of
those
want
a
universe,
that's
closed.
Some
of
them
do,
but
if
the
dark
energy
stays
the
wrong
kind,
you
can
actually
be
a
closed
universe,
but
expand
forever
and
eventually
become
infinite.
So
we
don't
know
the
answer
to
that
question.
We
know
that
it's
still
a
question.
C
C
B
Of
the
region,
where
we
look
for
dragons,
and
so
it
it,
it
is
a
good
working
hypothesis
that
the
universe
is
effectively
infinite
but
in
fact
it
might
well
be
not
so
much
bigger
than
the
horizon
that
we
see,
although
it's
unlikely,
we're
in
a
very
special
place,
that
the
universe
is
substantially
bigger
than
what
we
see
but
not
incredibly
bigger,
but
it's
also
possible.
It's
way
bigger,
I
mean
so
we.
A
Don't
know
the
answer,
so
I
had
a
question
on
on
how
long
does
the
the
data
that
you
collected
say
decades
ago
remain
valuable?
Do
the
new
instruments
with
the
better
data?
Does
it
does
it
kind
of
make
everything
that
was
collected
before
outdated
or
are
there
still
things
to
be
learned
by
going
back
and
looking
at
things
that
were
observed
a
while
ago.
B
There
is
some
potential
stuff
to
learn,
except
the
sensitivity
of
their
instruments.
We've
been
on
the
moore's
law
for
sensitivity
of
detectors
of
the
cmb.
Their
sensitivity
has
been
improving
by
more
than
a
factor
of
two
ever
approximately
every
18
months.
That's
why
we've
been
able
to
go
from
measuring
you
know
30
micro,
kelvin
down
to
30
nano
kelvin
in
in
the
20
years
right.
I.
B
B
C
Here
so
it
seems,
like
your
map,
of
the
universe,
has
gotten
more
precise
clearly
over
the
years,
and
that
your
future
goal
is
to
make
it
even
more
precise
and
to
be
able
to
fit
data
points
to
you
know
the
theoretical
curve,
even
better,
what
gain
to
science
we
made
by
making
the
map
even
more
precise
than
it
already
is,
and
fitting
the
data
to
to
where
it.
You
know
it
should
already
be.
B
Okay,
so
that's
a
question
and
that's
why
I
was
saying
I
think
we're
entering
the
sort
of
more
final
phase
of
measuring
anticipates
at
some
point,
even
though
you
make
your
detectors
better,
you
get
down
to
where
your
photon
limited,
and
it's
just
you
know.
I
got
to
measure
10
of
the
12
photons
to
measure
apart
and
10
of
the
you.
B
6
because
of
statistical
fluctuations,
there
comes
a
point
in
which
the
sample
of
the
universe
you
have
is
limiting
you.
So
if
you're
looking
at
a
really
big
scale,
so
you're
so
you're
looking
at
stuff,
that's
on
a
90
degree
scale.
You
only
have
five
samples.
Five
five
independent
same.
You
can
only
expose
in
five.
So
your
errors,
if
you're,
if
you
believe
that
the
thing's
a
stochastic
process,
it's
quantum
mechanical
fluctuations,
it's
a
stochastic
process.
You
have
a
sample
of
five
that
looks
like
going
out
and
asked
five
random
people.
B
E
B
Out
to
a
reasonable
large
sample,
but
when
you
you
know,
when
you
get
out
to
really
large
scales,
then
you
have
other
systematics
affecting
you
and
so
we're
at
a
point
we're
reaching
the
point
where
in
another
generation
we
should
get
to
the
point
where
we're
kind
of
fundamentally
limited
in
terms
of
how
accurate
we
can
go,
what
we're
needing
to
do
is
getting
some
hints
of
what
the
physics
is
beyond
what
we
have.
That's
why
I
was
focusing
back
here
on
these
anomalies.
This
is
one
of
the
things
that
I'm
working
on.
B
So
you
know
this
gave
me
a
lot
of
sleepless
nights
a
couple
of
things.
This
is
a
little
low.
That's
why
that
plot.
This
guy
is
really
low,
really
slimy.
There's
several
in
a
row
here
that
are
low,
they're
averaged
together.
This
guy
is
significantly
high.
They
could
be
real.
There
are
features
in
here
that
you
don't
know
if
they're
a
random
sample
or
not,
but
because
you
have
this
prediction
that
it
that
if
you
know
one
whoops,
if
you
know
one,
you
should
see
it
several
places.
B
If
you
see
some
anomaly
here,
you
should
also
see
it
here
and
if
you
see
it
at
the
at
the
l
equals
20,
you
should
also
see
an
e
at
20
and
you
should
see
it
the
lens,
but
that's
so
tiny
to
see,
but
you
should
see
it
there,
but
you
should
also
see
it
in
the
primary
of
power
spectrum.
That
is
the
these.
Are
these
things
there's
inflation
creating
stuff?
There
are
primary
fluctuations.
The
c
and
b
is
just
a
snapshot
of
them
right.
B
It
just
takes
it's
just
a
circular
snapshot
around
us
of
what
they
look
like
if
you
go
out
measure
where
all
the
galaxies
are
and
all
the
clusters
and
the
liquids
are
same
thing
really
precisely
you
can
map
these
out
in
3d.
You
can
look
to
see
or
is
that
is
that
little
structure
I
see
there?
Does
it
show
up
in
the
in
the
polarization?
Does
it
show
up
in
the
so?
So
I
think
it's
going
to
be
combined
measurements
and
looking
to
see
if
this
model
we
have
that
has
only
six
parameters.
B
Is
it
able
to
describe
the
universe
with
incredible
precision
and
therefore
we
think
there's
something
beyond
it,
but
we
have
no
clue.
Are
we
going
to
actually
get
a
clue
and
I
think
we're
getting
a
clue
already?
That's
what
was
exciting
for
me
about
this,
this
swirly
map,
because
as
soon
as
you
have
this,
you
only
have
one
point.
B
C
B
By
with
a
year
from
now,
we'll
have
four
new
sets
of
results
in
and
we'll
have
some
good
idea
whether
this
is
true
or
not,
and
but
it'll
be
interesting
to
see.
But
I
don't
want
to
go
on
forever
on
this,
while
I'm
getting
older
according
to
horst
and
but
also,
I
think,
there's
a
limit
diminishing
return.
There
comes
a
point
where
it's
diminishing
returns
and
you
have
to
look
at
combined.
B
C
B
B
There's
the
physics
numbers
and
then
there's
the
astronomy
numbers
and
then
the
cosmology
guesses,
and
now
it's
not.
I
mean
it's
we're
we're
computing
things
in
a
real
way
and
it's
not
so
simple,
because
even
though
every
step
is
simple,
there's
a
lot
of
steps,
then
it
adds
up
to
a
complicated
university.
So
we
have
another
question.
B
Because
because
it's
gossip
I
mean,
I
know
the
the
the
next,
you
know,
if
you
look
at
the
fields
and
so
forth,
there
are
many
fields
in
physics
for
talking
about
nobel
prize
in
physics.
There
are
many
fields
in
physics
and
there's
a
lot
of
excitement.
Cosmology
lately,
so
cosmology
has
actually
been
getting
nobel
prizes
at
slightly
higher
than
the
rate
you
would
anticipate
based
on
the
population
profile.
B
And
so
they
made
the
claim.
But
then
their
claim
people
are
saying
is
not
enough
standard
deviations
to
count
as
a
real
claim
anymore,
and
so
that's
one
of
the
issues.
However,
the
other
thing
the
confused
thing
was
just
just
five
days
ago,
four
or
five
days
ago,
the
cavalier
prize
was
given
to
people
for
the
theory
of
inflation,
so
the
cavalier
prize
is
giving
out
by
the
norwegian
king
right,
and
so
it
was
given
to
alan
guth
andrei
linde
and
alexis
starovinsky.
B
Now
some
of
us
have
been
speculating.
If
a
theory
prize
was
going
to
give
out,
it
would
probably
be
given
out
to
star
bensky,
who
would
have
had
the
very
first
paper
gutha,
the
second
paper
and
sato,
who
had
the
third
paper.
Even
though
there
are
other
people
you
could
consider,
but
you
know
how
I
actually
think
inflation's
on
a
slightly
firmer
footing
than
this
discovery.
This
discovery
would
kind
of
nail
it,
but
it
was
it's
it's
you.
C
B
There's
an
awful
lot
of
things
that
make
inflation
look
right
and
it
certainly
has
influenced
our
view
of
the
field
in
a
tremendous
way.
So,
but
you
know
now,
there's
a
nobel
committee
who
would
be
making
the
decision
or
they're
going
to
be,
you
know:
do
they
have
a
problem?
They
want
to
give
it
somebody
slightly
different
than
the
cavalier
price
or
whatever
it's
a
so
now.
I
think
the
the
field
is
confused,
but
then
the
next
cycle
is
five
more
years
later
or
something
is
sort
of
in.
B
Kind
of
times
period
and
if
you're
really
smart,
you
can
kind
of
guess
what
areas,
because
they
rotate
people
on
and
off.
The
nobel
committee
there's
many
stages,
but
there's
a
final
selection
committee
and
there's
certain
the
people
who
have
expertise
and
cosmology
got
rotated
off
shortly
after
well.
Basically
about
the
time
saul
got
the
prize,
and-
and
it's
there's
only
one
left
on
so
you
can,
you
know
they
have
to
they'd-
have
to
pull
in
their
friends.
You
know
it's
not
I
mean,
and
there
are
stuff
there's
there's
huge
levels.
I
mean
there's
nominations.
B
There
are
committees
that
do
reviews
and
there
are
committees
that
actually
prepare
cases
it.
So
it's
not
like
they're,
but
you
have
to
realize
there's
a
lot
more
people
and
results
deserving
prices
than
there
are
prices,
so
there
are
going
to
be
situations
where
they
aren't
going
to
come
out,
and
so
that's
why
I
say
the
cavalier
prize
probably
confuses
the
issue.
E
I
have
a
question
regarding
the
flat
universe:
let's
can
one
assume
that
photons
travel
along
the
gravitational
overall
gravitation
field
along
the
tangential
to
the
gravitation
field
along
the
iso
surface
and
as
a
result,
you,
you
see
the
flat
universe
due
to
the
photon
photon
traveling
pathway,.
B
B
Okay,
so
the
the
point
is
that
the
universe
may
have
a
geometry,
that's
flat
or
not
quite
flat,
but
in
it
the
that
would
be
the
case
if
it
had
a
uniform
distribution
of
matter
and
energy.
The
madame
energy
are
lumped
up
because
of
that
we
have
perturbations
so
that
light
that
would
normally
be
traveling
through
an
expanding
universe,
so
it
actually
has
a
depending
on
what
frame
you
look.
B
If
you
look
in
the
co-moving
frame,
which
is
a
frame
that
makes
light
travel
on
45
degrees,
then
it
looks
like
nice
straight
line,
even
then,
even
in
that
frame.
If
you
go
near
a
cluster
of
galaxies
or
close,
very
close
to
a
galaxy,
the
trajectory
will
be
slightly
disturbed
and
you
do
lensing
and
in
fact
I
didn't
have
it
here
in
the
talk,
but
one
of
the
calculations
that
blank
has
done
is
to
actually
reconstruct
the
lensing
map.
That
is
the
projection.
B
There's
a
you
know:
here's
your
source,
which
is
the
cosmic
microwave
background,
here's
some
structure
in
the
middle,
and
here
you
are
looking
at
it
and
the
light
travels
and
it
gets
just.
It
gets
slightly
disturbed
in
its
path
along
the
way
and
you
can
reconstruct
how
much
that
is.
If
you
think
you
understand
the
theory
of
the
universe.
Well
enough,
you
can
then
fit
the
what
you
think
the
original
source
looked
like
on
average
and
then
make
a
map
of
what
the
distortion.
B
You
know
what
the
lensing
distortion
looked
like,
and
so
blank
has
done
that
and
we
published
that
in
a
paper
I
don't
know
in
the
spring
I
mean
I'm
sorry
late
last
fall
and
it's
I'm
trying
to
remember
when
it
was,
but.
B
The
last
year
and
that's
the
first
half
of
the
day,
the
second
half
of
the
day,
though,
will
be
coming
out
in
in
october
and
then
a
little
bit
more
after
that
and
it
you
you
can
do
that
and
now
you
have
to
say
well.
Did
I
make
a
mistake?
Is
the
universe
much
more
comp?
There
are
other
possible
geometries
that
tend
towards
the
kind
of
geometry.
We
know
we're
in
the
present
day
that
you
can
start
out
with
they're
called
bianchi
spaces.
B
You
can
start
out
with
a
more
distorted
kind
of
thing,
but
there's
certain
ones
that
then
evolve
towards
our
situation,
and
so
it
is
possible
that
you
have
a
more
complicated
universe.
It's
not
very
likely
because
the
face
base
for
that
is
very
small,
but
it's
still
possible
right,
and
so
there
was
an
attempt
to
look
for
for
some
evidence
of
that
kind
of
thing
in
the
cmb
data.
B
So
far,
there's
no
strong
evidence
for
that,
but
it
is
possible
that
on
a
huge
scale
or
back
in
time,
scale
a
huge
scale
back
in
time
that
the
universe
wasn't
quite
so
flat.
But
it's
not
I'm
not
guaranteeing
it's
perfectly
flat
anyway.
What
I'm
saying
is
for
the
for
a
reasonably
long
period
in
really
big
spaces,
it's
pretty
flat,
but
you've
got
to
average
over
a
cluster.
B
B
Across
it
doesn't
go
in
a
straight,
it
gets
bent
slightly,
so
there
is
bending
going
on
all
the
time.
It's
just
usually
very
minor,
and
it's,
and
so
I
believe
that
we've
treated
that
reasonably
well.
But
there
is
always
prejudice
and
usually
people
do
the
fits
from
the
simplest
possible
model,
because
in
the
old
days
we
had
no
data.
C
B
B
F
Okay,
suppose
we
have
an
efficient
detector
for
higgs
bosons
in
couple
of
decades
from
now,
and
we
are
mapping
the
universe
on
higgs
bosons
and
we
have
angular
distribution
of
hicks.
Are
we
going
to
finally
get
the
shape
of
the
universe
or
not
yet.
B
Well,
I
think
our
map
and
the
photons
is
going
to
be
much
better.
The
higgs
is
everywhere.
This
is
supposed
to
be
here
in
this
room
right.
It's
supposed
to
be
this
continuum
condensate
everywhere,
and
so
it's
not
propagating
to
us
from
a
long
distance
to
provide
like
the
photons
are
to
provide
us
the
information
it
is.
It
is
here
already
it's
a
question
of
the
disturbances
in
the
field.
B
It's
it's
like
you
know,
there's
a
trampoline
and
people
over
there
jumping
up
and
down
and
you're
at
the
edge
feeling
it
that's
it's
a
different
than
when
you,
when
you
get
to
watch
them
by
light,
so
it'll
be
different.
Now,
if
we
could
use
gravity
waves,
then
we
could
see
back
to
really
beginning
time
so
someday.
We
may
be
doing
that.
There's
proposal.
B
Well,
there's
a
pre-proposal
thought
to
make
a
thing
called
the
big
bang
observer
right,
which
is
a
really
big
version
of
lisa,
which
is
the
laser
interferometer
in
space.
That's
that
that
the
europeans,
somehow
the
us,
has
dropped
out
of
for
a
while.
The
european
space
agency
is
pushing
forward
and
that's
to
measure,
gravity
waves
and
we'll
see
black
hole,
mergers
in
in
galaxies
that
have
murders
and
so
forth.
But
if
you
build
one
on
a
much
bigger
scale,
you
can
see
longer
wavelength
gravity
waves.
B
You
can
try
and
look
to
see
the
gravity
waves
that
came
from
the
beginning
of
the
universe,
and
you
can
also
image
things
back
to
much
earlier
time
because
they
did
a
couple
much
earlier
and
likewise,
if
you
could
figure
out
how
to
image
the
neutrinos
you
could
you
could
the
neutrinos
decouple
in
a
much
earlier
time.
So
those
are
those
are
messengers
that
come
moderately
directly
and
carry
the
information
fairly
preserved.
The
higgs
is
not
likely
to
be
that
case.
That's
so
the
higgs
by
itself
decays.
D
B
Yeah
the
the
answer
is,
I
I
don't
think
that
I've
ever
seen
it
it's
a
you
know
I
have.
It
was
pointed
out
to
me
when
I'm
publishing
holes
in
time
that
that
there
was
such
a
book,
and
I
found
that
it's
a
very
delightful
children's
book.
It's
quite
it's
the
pre-harry
potter
kind
of
kind
of
book
and
she
didn't
make
as
much
money.
But
but
anyway,
I
recommend
the
book.
B
But
there
was
no
direct
relationship,
the
wrinkles
in
time
I
was
looking
for
an
interesting
title,
and
I
was
trying
to
make
sure
that
I
made
the
point
that
the
real
things
we
were
seeing
were.
The
the
the
temperature
fluctuations
were
what
I
call
the
scalars,
but
that's
the
time
time
component,
it's
the
it's
the
wrinkle
and
the
time
component
of
the
metric
more
than
the
wrinkle
in
the
space
component.
B
Even
though
we
talk
about
gravitational
potential,
being
distortion
of
space
is
distortion
of
spine
time
too,
and
in
the
really
low
field,
the
time
one
actually
dominates
or
something,
but
it
was
just
to
get
people
to
think
a
little
more
broadly
right
now.
The
problem
is
people
thought
it
was
a
cosmetic
book.