Do LLMs Understand? AI Pioneer Yann LeCun Spars with DeepMind’s Adam Brown. - part 2/15
2025-12-12_17-05 • 1h 15m 39s
Yann LeCun (Chief AI Scientist)
00:00.980
They
didn't
really
find
until
the
1980s,
nobody
found
really
a
good
way
to
train
those
those
multi-layer
systems
mostly
because
the
neurons
that
that
they
had
at
the
time
were
the
wrong
type.
They
had
neurons
that
were
binary.
So
neurons
in
the
brain
are
binary.
They
they
either
Yann LeCun (Chief AI Scientist)
00:19.220
fire
or
they
don't
fire
and
people
wanted
Yann LeCun (Chief AI Scientist)
00:23.020
to
reproduce
that.
So
they
they
build
simulating
neurons
that
would
either
be
active
or
inactive.
And
it
turns
out
for
the
modern
learning
algorithms
to
work,
we
call
them
back
we
call
it
back
propagation.
You
need
to
have
neurons
that
have
sort
of
graded
responses.
Um,
and
uh
Yann LeCun (Chief AI Scientist)
00:40.100
that
only
became
practical
possible
or
people
realized
it
could
work
in
the
1980s.
People
Yann LeCun (Chief AI Scientist)
00:46.220
had
the
idea
before,
but
they
never
could
really
make
it
work.
And
so
that
caused
um
a
renewal
of
interest
in
neural
nets
in
the
1980s.
They
had
been
largely
abandoned
in
the
late
60s.
And
then
they
came
to
the
fore
again
in
the
mid-to-late
80s.
That's
when
I
started
kind
of
my
Yann LeCun (Chief AI Scientist)
01:04.740
graduate
school
basically
in
1983.
Yann LeCun (Chief AI Scientist)
01:07.500
And
uh
there
was
a
wave
of
interest
that
lasted
about
10
years.
And
then
interest
went
again
um
in
the
mid-90s
until
the
late
2000
when
we
rebranded
it
into
deep
learning.
Neural
net
had
kind
of
a
bad
rep.
Um
people
in
computer
science
and
engineering
thought
neural
nets
were
Yann LeCun (Chief AI Scientist)
01:28.900
kind
of
a
bad
thing
had
a
bad
reputation.
Yann LeCun (Chief AI Scientist)
01:32.060
And
so
we
we
branded
it
into
deep
learning
and
sort
of
brought
it
back
to
the
floor
and
then
the
results
were
were
there
in
computer
vision,
in
natural
language
understanding,
speech
recognition
to
really
convince
people
that
this
was
uh
a
good
Janna Levin (Professor of Physics and Astronomy)
01:46.100
thing.
Now,
Adam,
you
at
at
a
very
young
age
were
interested
in
theoretical
physics,
not
specifically
computer
science.
And
you're
watching
some
of
this
unfold
in
some
sense
from
afar.
What's
the
catalyst
that
sweeps
up
so
many
people
decades
later.
There's
Janna Levin (Professor of Physics and Astronomy)
02:03.460
There's
this
time
where
it's
of
great
interest,
there's
great
success
in
handwriting
recognition
or
in
uh
visual
recognition
and
these
things,
but
it's
not
sweeping
up
the
world.
What
What
happens
that
brings
us
to
this
point
where
we're
all
now
talking
about
large
language
Janna Levin (Professor of Physics and Astronomy)
02:18.300
models.
Adam Brown (Research Scientist)
02:19.500
So,
many
physicists
in
the
last
years
have
pivoted,
should
we
say,
from
working
on
physics
to
working
on
AI,
and
it
really
traces
back
to
some
of
the
work
that
Jan
and
others
did
to
prove
that
it
works.
Adam Brown (Research Scientist)
02:34.300
Like
when
it
wasn't
working,
it
was
just
this
this
thing
that's
over
there
in
computer
science
and
like
of
many
things
in
the
world
that
are
not
particularly
uh
may
be
interesting
but
not
many
physicists
are
paying
attention
to
it.
But
then
after
Adam Brown (Research Scientist)
02:46.580
you
know
Jan
and
some
of
the
other
pioneers
of
this
field
proved
that
it
would
work,
it
became
a
totally
fascinating
subject
for
physics.
That
you
link
up
these
neurons
together
in
a
certain
way
and
suddenly
you
get
emergent
behaviour
that
didn't
exist
at
the
individual
neuron
Adam Brown (Research Scientist)
03:02.740
level.
That
seems
like
a
a
subject
Adam Brown (Research Scientist)
03:05.300
that
physicists
who
spend
their
life
imagining
how
the
sort
of
rich
pageantry
of
the
world
could
emerge
from
simple
laws
that
immediately
attracted
the
attention
of
many
physicists.
And
nowadays
it's
a
a
a
very
common
career
path
to
do
a
PhD
in
physics
and
then
apply
it
to
a
Adam Brown (Research Scientist)
03:20.020
emergent
system.
But
the
emergent
system
is
an
emergent
network
of
neurons
that
collectively
give
rise
to
intelligence.
Mhm.
Janna Levin (Professor of Physics and Astronomy)
03:27.580
Now,
mhm,
let's
do
a
lightning
round
cos
you
raised
the
dreaded
word
intelligence.
Everybody
in
this
room
very
likely
has
interacted
with
something
that
we're
now
calling
an
AI.
These
are
all
large
language
models
and
before
I
ask
you
to
define
those
for
us,
I
just
want
to
kind
Janna Levin (Professor of Physics and Astronomy)
03:43.380
of
do
a
lightning
round
of
of
what's
your
yes
or
no
response
to
certain
things.
So
Adam,
Janna Levin (Professor of Physics and Astronomy)
03:53.420
yes
or
no
are
these
AIs,
these
large
language
models
understanding
the
meaning
of
the
conversation
Yes.
Yann LeCun (Chief AI Scientist)
04:05.140
Sort
of.
Janna Levin (Professor of Physics and Astronomy)
04:08.980
Perfect.
Janna Levin (Professor of Physics and Astronomy)
04:14.260
Right,
exactly.
It
was
my
fault
for
giving
you
a
binary
choice.
Okay,
so
that
allows
me
to
ask
the
next
question
because
it's
not
a
foregone
conclusion.
If
you
don't
say
yes
to
that,
it's
going
to
be
interesting
what
you
say
to
this.
Are
these
AIs
conscious?
Yann LeCun (Chief AI Scientist)
04:30.500
Absolutely
not.
Janna Levin (Professor of Physics and Astronomy)
04:32.300
Adam.
Adam Brown (Research Scientist)
04:33.820
Probably
not.
Okay.
Janna Levin (Professor of Physics and Astronomy)
04:36.260
Um,
will
they
soon
be?
Adam Brown (Research Scientist)
04:39.660
I
think
they'll
one
day
be
conscious
if
if
progress
continues
in
the
way
that
we're
we're
continuing.
When
is
hard
to
say,
but
Mhm.
Yann LeCun (Chief AI Scientist)
04:49.060
John.
Yes,
for
appropriate
definitions
of
consciousness.
Janna Levin (Professor of Physics and Astronomy)
04:52.100
Yes,
okay.
Well,
we
do
have
some
philosophers
in
the
house
and
um
we're
we're
not
going
to
indulge
in
philosophical
definitions
of
consciousness
or
there
our
hour
would
go.
And
we'd
still
be
here.
Oh,
I
just
heard
that
groan
I
think
from
our
friends
up
in
the
balcony.
Autoscroll