►
From YouTube: Sponsored Keynote: The Singularity is Here : Are Your Deployments Ready? - Stephen Chin, JFrog
Description
Sponsored Keynote Session: The Singularity is Here : Are Your Deployments Ready? - Stephen Chin, JFrog
For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/
A
A
I
think
this
is
something
which
has
been
talked
a
lot
in
the
science
fiction
and
the
science
communities,
it's
something
which
we
see
coming
based
on
the
accelerating
pace
of
advances
in
general
for
humanity,
but
also
recently,
with
all
of
the
advances
in
computer
technology,
in
artificial
intelligence
in
the
exponential
growth
of
computing
power-
and
this
is
something
which
is
predicted
where
eventually,
the
computers
and
artificial
intelligence
will
be
more
intelligence
than
the
combined
brain
power
of
all
the
humans
put
together.
So
what
what
does
this
mean
for
us?
A
So,
first
of
all,
I
think
the
question
is:
what's
what
are
the
conditions?
When
do
we?
How
do
we
know
when
this
actually
is
going
to
happen?
So
there's
three
different
conditions
which
need
to
be
met
for
us
to
hit
the
singularity.
The
first
is
having
enough
data
to
be
able
to
model
all
of
the
different
humans
in
the
world,
so
there's
7.2
billion
humans,
6.2
billion
nucleotides,
so
this
is
1
times
10
to
the
19th
bytes.
So
we've
we've
hit
this.
A
We
already
have
6.8
exabytes
of
data
cumulatively
in
the
world,
so
we
can
actually
store
all
the
genetic
information
on
humans.
Next
question
is
for
the
entire
earth.
Could
you
store
all
the
information
on
all
the
species
for
the
entire
planet
and
we're
not
quite
there
yet,
but
the
growth
of
data
we're
certainly
going
to
get
here.
A
A
Four
supports
a
hundred
petaflops
of
computing
power
and
ibm
is
also
releasing
increasingly
complex
quantum
computing,
which
could
solve
different
scientific
and
biological
problems,
and
these
they're
already
up
to
an
anticipated
127
bit
qubit
quantum
computer
up
from
only
50
qubits
a
few
years
ago,
and
the
final
question
is:
can
we
can
we
build
it?
Can
we
automate
it
and
I
think
you've
all
heard
about
or
used
3d
printers?
I
have
a
lot
of
hobby
projects
myself
involving
3d
printers,
and
we
now
have
3d
printers,
which
can
fully
print
themselves.
A
So
this
is
an
example
of
a
self-replicating
3d
printer,
where
it
can
print
90
of
the
parts
to
reconstruct
a
new
3d
printer
on
its
own
flatbed,
and
only
a
few
exceptions
to
this,
like
the
the
hot
end
and
some
of
the
electronics
need
to
to
be
made
using
other
manufacturing
processes.
So
pretty
soon
we're
going
to
have
computers
which
can
self-replicate
and
we've
all
seen
this
coming
in
the
movies.
B
A
Okay,
so
clearly
arnold
had
it
right,
they
predicted
all
the
singularity
and
all
this
coming
in
the
future,
and
the
question
is
how
far
to
the
future
do
we
have
to
look
for
this?
So
basically,
the
equation
here
is
human
intelligence
is
rising.
Machine
intelligence
is
rising
and,
at
some
point,
machine
intelligence
will
outpace
human
intelligence
and
we
hit
the
singularity.
So
I
think
this
is
this
is
coming
soon.
A
This
is
okay,
actually
there's
a
slight
error
with
this
graph,
so
I'm
not
so
sure
that
human
intelligence
is
still
rising,
especially
given
the
events
of
the
past
year,
so
we're
just
gonna
we'll
knock
it
down
a
notch,
and
I
think
this
puts
us
on
a
trajectory
of
you
know
somewhere
somewhere
in
the
next
20
years.
We
we
hit
the
singularity
machines
take
over
the
world
and
what?
What
does?
What
does?
This
all
look
like,
so
to
figure
out
what
this
looks
like
I've?
A
I
have
with
a
little
bit
of
help
from
the
experts
at
jfrog.
We've
we've
built
a
prototype
time
machine
just
like
what
you
see
in
the
terminator
set.
The
current
date
set
the
target
dates
and
we're
going
to
see
if
we
can
go
into
the
future
to
the
year
2040
and
and
see
what
see
what
it's
going
to
look
like
once,
the
what's
the
machine
overlords
have
have
taken
over
so
first
off,
let's
check
out
the
boob
tube.
This
is,
of
course
people
are.
A
This
is
amazing
that
we've
actually
recognized
that
the
machines
have
have
made
it,
but
but
let's,
let's,
let's
kind
of
check
out
and
maybe
check
out
what
a
newspaper
would
look
like.
So
clearly,
the
world
is
not
very
peaceful
in
2020,
but
we.
Finally,
we
we've
we've
moved
past
and
we've
we've
progressed
deployments
where
it's
not
just
over-the-air
updates
on
on
devices
like
like
these
anymore.
This
is
this
is
true.
Over-The-Air
updates.
A
This
is
we're
actually
updating
things
planes
over
the
air
and
increasing
security,
while
planes
are
flying
around
to
block
terrorist
attacks.
So
this
is,
this
is
amazing
stuff,
so
we're
continuously
deploying
to
to
the
edge
and
here's
a
here's
a
billboard.
So
apparently
it's
not
cloud
computing
anymore.
It's
crop
computing,
so
we've
we've
definitely
entered
a
different
era
of
computing
and
this
this
actually
isn't
that
far
off
from
reality,
because
we're
already
at
more
than
30
billion
iot
devices
online
predicted
by
the
year
2025.
A
A
So?
I
think
when
you,
when
you
look
at
the
human
computer
interaction,
the
relationship
between
the
humans
and
the
computers
is
the
the
computers
are
very
efficient,
they're
very
fast.
They
execute
automated
quality
checks,
they
have
machinery
and
processes
to
to
make
sure
things
are
running
smoothly
and
the
the
humans
are
the
bottleneck
in
this
process.
A
So
we
we
have
to
manually
serially,
evaluate,
execute,
think
and
and
iterate
to
to
to
solve
problems,
and
we,
frankly
are
are
slowing
down
the
pace
of
evolution
of
the
machines,
and
this
is
true
not
just
of
software
development,
but
it's
also
true
of
software
pipelines
and
continuous
delivery.
So
how
do
we
speed
up
our
continuous
delivery?
A
So
you
have
you
know
your
different
inputs
coming
in
from
source
control
systems
going
into
your
favorite
continuous
integration
server.
It
could
be
jenkins,
it
could
be
techton,
it
could
be
spinnaker
any
of
any
of
your
favorite
cd
foundation
projects,
as
it
goes
through
that
you
want
to
have
different
systems
which
you're
deploying
to
so
you
typically
send
out
to
a
container
registry
or
a
package
management
system
like
artifactory,
then
released
to
kubernetes
clusters
that
are
running
in
integration.
A
So
these
are
either
automated
tests
or
different
quality
measures
for
security,
for
making
sure
that
it
meets
the
production
requirements
for
performance
requirements,
and
if
this
is
fully
automated,
then
you
can
get
to
the
nirvana,
where
you
can
continuously
deploy
every
hour
or
possibly
on
every
check-in,
making
sure
that
you're
only
pushing
releases
to
production
that
meet
all
the
quality
requirements
and
getting
rid
of
all
the
human
intervention
along
the
way
so
once
you've
once
you've
automated.
All
of
this
and
you've
increased
the
speed,
the
velocity,
your
ability
to
deploy
into
production.
A
If
you
catch
things
in
development,
it's
fairly
easy
testing
again,
you
can
fix
it
quickly.
You
can
solve
security
issues,
but
once
you
get
into
staging
into
production
or
or
it
becomes
a
customer
issue,
the
the
cost
to
fix
it,
the
amount
of
time
it
takes
to
fix
it
exponentially
goes
up,
and
so
what
you
want
to
do
is
you
want
to
shift
all
of
your
security
requirements
left.
A
You
want
to
put
more
tools
in
the
hands
of
developers
which
enable
them
to
be
able
to
do
those
quality
checks,
to
check
open
source
libraries
to
make
sure
that
they're
secure
to
make
sure
that
any
issues
are
caught
really
early
in
the
development
process
and
we're
basically
shifting
and
pushing
security
to
the
left,
which
is
allowing
us
to
to
work
faster
and
to
please
our
our
new
autonomous
overlords,
the
the
robots
who
are
taking
over.
So
I
think
we
we've
solved
a
couple
of
the
big
problems
now.
A
The
next
question
is:
when
we
get
into
production,
how
do
we
make
sure
that
we're
observing
we're
monitoring
and
we're
tracing
in
a
way
which
allows
us
to
iterate
continuously
and
one
of
the
ways
of
accomplishing
this?
Is
the
end-to-end
visibility
platform
which
we're
building
out
together
with
jfrog
and
all
of
our
partners,
and
we
we
announced
this
recently
at
swampup.
A
It
includes
on
the
left
side,
with
developers
and
operations,
folks,
having
all
the
collaboration
tools
you
use
together,
like
slack
jira
and
microsoft
teams
in
the
center
for
your
observability
platform,
allowing
you
to
work
with
all
the
different
observability
tools
like
data
docs,
splunk,
dynatrace,
elastic
and
others
which
give
you
this
full
visibility
with
metrics
logging
tracing
coming
out
and
then
finally,
integrating
with
it
service
management
tools
like
pagerduty,
and
when
you
bring
this
all
together.
This
gives
you
this
this
code
to
production,
visibility
and
traceability,
all
on
top
of
the
jfrog
platform.
A
You
need
to
have
security,
which
you
can
shift
and
enable
developers
to
be
more
powerful,
and
you
need
all
this
put
together
with
an
end-to-end
observability
platform,
and
this
is
a
lot
of
what
we
provide
with
the
jfrog
platform
with
x-ray.
Our
security
product
artifactory
for
package
management
and
all
the
partner
integrations
we're
doing
with
all
of
the
other
folks
in
the
devops
ecosystem
for
collaboration,
itsm
and
observability.
A
A
And
this
takes
us
to
a
symbiosis
with
our
robot
overlords,
where
now,
rather
than
us,
being
working
for
them
or
them
working
for
us.
We're
now
working
together
and
collaborating
on
things
like
teaching,
kids,
how
to
program
and
improving
our
society.
So
one
of
my
hobbies
is
doing
education
for
kids
in
stem.