►
From YouTube: In the Clouds (E27) ft. Matt Butcher
Description
Across industries and business sectors, senior leadership has to execute at an ever-increasing pace, and today's technology decisions must balance short-term risk with long-term gains. Join host Stu Miniman as he leads insightful and candid discussions with a broad spectrum of experts, executives, and decision-makers on diverse and hot technology topics.
In this episode of In the Clouds, join your host Stu Miniman as he sits down with Fermyon CEO Matt Butcher, amidst their release of their new NoOps Database for developers building Serverless Applications; with WebAssembly. The Cloud Tech revolutionary is also open source and partners with Red Hat often in the field.
Matt Butcher, co-founder, is one of the original creators of Helm, Brigade, CNAB, OAM, Glide, and Krustlet. He has written or co-written many books, including Learning Helm and Go in Practice. He is a co-creator of the Illustrated Children’s Guide to Kubernetes series. These days, he works mostly on WebAssembly projects such as Spin, Fermyon Cloud and Bartholomew. He holds a Ph.D. in Philosophy. He lives in Colorado, where he drinks lots of coffee.
Join us live for this insightful episode, and feel free to ask all your burning questions, as we deep dive into all things tech and innovation, with a special focus on serverless application development.
Want to learn more?
Try Red Hat OpenShift: https://red.ht/3Haxapr
Follow us:
Stu Miniman https://twitter.com/stu
Matt Butcher https://twitter.com/technosophos
OpenShift https://twitter.com/openshift
Fermyon https://twitter.com/fermyontech
Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_confirmation=1
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3JBaOia
OpenShift on GitHub https://github.com/openshift/
#RedHat #OpenShift #HybridCloud
B
A
Welcome
to
another
episode
of
in
the
clouds:
I
am
your
host
Stu,
miniman
and
wow.
We
have
reached
August
2023.
before
we
get
to
our
wonderful
guest
this
week.
I
just
want
to
a
big
thanks
to
the
technology
Community.
This
last
weekend,
I
participated
in
a
charity
bike
ride,
it's
known
as
The
panmass
Challenge,
you
might
notice.
You
know
up
on
my
wall.
There's
the
PMC
flag,
it's
the
fifth
year.
I've
done
this
event.
A
A
It
was
inspiring,
it
was
tiring
and
it
was
really
good.
I've
raised
about
eight
thousand
dollars
want
to
give
a
a
huge
thank
out
to
plenty
of
my
work.
Colleagues,
people
from
red
hat
from
VMware
from
AMD
who
donated
those
companies
do
company
matches
and
it's
awesome
there
and
you
know
hey
next
year,
if
you're
interested
in
finding
out
more
feel
free
to
reach
out
to
me,
we'll
put
a
link
in
chat
here.
Anybody
that
wants
to
give
any
last
minute
donations.
Those
are
open.
A
100
of
all
the
donations
go
to
the
Dana
Farber
Institute,
which
is
a
wonderful
organization,
the
nonprofit
that
does
cancer
research
cancer
support
so
really
really
important
event.
That
I
am,
you
know,
really
happy
to
be
able
to
participate
in.
So
thank
you
for
that.
Let's
switch
gears.
We
have
have
a
wonderful
guest,
this
time,
I'm
happy
to
bring
on
the
program
Matt
butcher.
A
He
is
the
co-founder
and
CEO
of
fermion
I'm
sure
many
of
you
in
the
cloud
and
kubernetes
space,
Have
Heard
lots
of
Buzz
about
webassembly
wasm
technology,
and
those
of
you
in
the
container
industry
might
remember.
Matt
from
some
of
his
previous
work
he's
been
pretty
crucial
in
some
some
awesome
technology
that
we
still
leverage
you
know
in
the
in
the
containers
in
kubernetes
space,
so
Matt.
Thank
you.
So
much
for
joining
me,
it's
it's
great
to
see
you
yeah.
A
B
When,
when
my
career
started,
I
was
really
more
focused
on
content,
Management,
Systems
and
and
ended
up
working
at
HP
in
a
brand
new
unit,
there
called
HP
Cloud
originally
to
do
the
CMS
system
that
ran
the
developer,
docs
and
I.
As
soon
as
I
got
a
taste
of
cloud
computing
I
was
just
all
in
so
switch
from
CMS
into
sort
of
the
early
openstack
work
worked
on
platform
as
a
service
and
openstack,
and
that
really
sort
of
redefined.
B
My
career
and
I
got
going
right
into
the
cloud
world
ended
up
a
few
years
later
at
Google
and
then
after
a
stint,
there
started
it
at
Deus,
which
is
a
small
open
source
oriented
company.
In
Boulder
we
were
building
a
platform
as
a
service
on
you
know
at
the
time
this
brand
new
technology
that
was
so
exciting.
It
was
containers
right
and
nobody
else
was
building
a
paths
on
that
yet
and
then
Along
Came
kubernetes.
B
That
caught
the
attention
of
Microsoft
Brandon
Burns
had
recently
gone
from
Google
to
Microsoft
and
was
interested
in
building
really
a
top-tier
kubernetes
team
and
so
Deus
got
acquired
by
Microsoft.
A
bunch
of
us
became
sort
of
like
the
open
source,
Deus
Labs
team
inside
of
Microsoft,
where
we
were
I
I,
had
what
I
think
was
the
best
job
ever
right.
We
were
doing
Upstream
open
source
all
the
time
going
to
all
the
conferences.
B
You
know
go
from
start
to
finish
in
milliseconds
instead
of
running
for
hours,
days
or
months,
and
that
ultimately
led
us
toward
webassembly.
A
group
of
us
left
Microsoft
back
in
2021
to
found
fermion
really
focused
on
building
this
kind
of
next
wave
of
cloud
compute
and
our
team's
grown
over
the
last
couple
of
years
we're
about
30.
Now
we
do
a
lot
of
Open
Source
work,
Upstream
inside
of
bytecode
Alliance
and
cncf
and
W3,
and
then
you
know
build
in
Fermi
on
the
company
yeah.
A
It's
awesome
stuff,
you
know,
I
I,
remember
boy,
the
the
the
the
the
excitement
around
the
community
around
Helm
charts.
It
was
funny.
It
was
one
of
those
things
that
just
sticks
with
you.
Matt
is
certain
announcements,
get
really
excited
and
it
was
the
first
time
that
I
remembered
something
getting
removed
that
got
everybody
excited
when
they
got
rid
of
the
tiller
yeah.
A
Yeah
somebody
built
that
originally
and
we
thought
that
was
the
right
architecture,
but
absolutely
we
need
to
make
changes,
but
you
know
it's
hugely
useful.
You
know
gosh,
you
know
I
I,
think
about
you
know
we
we
still
leverage.
You
know
on
the
red
hat
openshift
side,
things
like
hashicorp
Vault,
you
know,
or
you
know,
use
help
charts
and
that's
how
we
put
them
into
the
openshift
ecosystem
there
so
hugely
important.
Thank
you
to
that
crew.
A
So,
if
you
could
2021
you
started,
fermion
webassembly
has
been
around
since
you
know
it
was
announced
in
2015
it
launched
in
2017..
Could
you
for
our
audience
just
give
us
a
snapshot?
You
talked
a
little
bit
about
Services,
you
know
things,
but
what
is
this
web
assembly
thing?
How
does
this
fit
into
this
ecosystem
of
kind
of
the
apps?
And
you
know,
containers
and
VMS
and
everything
what
you
know
the
web
browsers?
You
know
webassembly
itself,
yeah.
B
B
When
the
Mozilla
team
in
2015
decided
to
take
on
this
problem
Anew,
they
basically
said:
we've
made
a
couple
of
mistakes
along
the
way
in
this
idea
of
embedding
a
programming
language
environment
inside
of
the
browser.
The
first
one
is
every
single
time
it
was
proprietary
or
supported
by
just
one
or
two
vendors
and
the
second
one
was.
B
We
were
always
bringing
a
very
specific
language
and
saying
you
know
this
is
the
language
you
can
use
in
the
browser
alongside
JavaScript
and
and
so
this
effort
began
first
with
Mozilla,
and
then
they
quickly
joined
forces
with
you
know
the
Chrome
team
from
Google
and
the
at
the
time
the
IE
team
from
Microsoft
the
Safari
team
from
Apple
and
said:
let's
build
a
standard
platform
agnostic.
You
know
OS
architecture,
agnostic,
language,
agnostic,
runtime
that
can
go
in
the
browser
and
that
can
just
instead
of
doing
all
the
UI
stuff.
B
Initially,
let's
just
figure
out
how
to
solve
the
core
compute
stories.
So
we
can,
you
know,
give
those
old,
C
libraries,
New
Life
by
compiling
the
web
assembly
and
running
them
in
the
browser.
We
can
add
some
high
performance
or
support
some
high
performance
languages
like
C
and
rest
and
and
go
and
be
able
to
have
those
running
inside
the
browser
to
complement
JavaScript
a
great
example
of
this
in
practice
right
both
Adobe
and
figma
use
webassembly
in
their
online
graphical
tools.
B
You
know
figma
being
a
good
example
of
a
high
speed,
Vector
drawing
program
that
you
have
in
the
browser.
When
you
can
compile
C
plus
plus
to
webassembly,
you
can
do
very
quick,
Vector
math
and
then
be
able
to
interface
with
those
libraries
from
JavaScript,
so
that
was
really
kind
of
the
original
vision
us
you
know
coming
from
at
that
time,
Microsoft
and
looking
at
serverless
as
the
big
problem
we're
we're.
Looking
we
were
looking
at
something
very
differently
right.
B
We
did
not
start
with
webassembly
and
say:
is
there
an
application
in
the
cloud
we
started
with
this
problem
statement
and
the
problem
statement
went
something
like
this:
we
need
a
cloud
runtime
that
can
run
serverless
functions
really
well.
What
are
the
problems
we
hear
from
developers
that
they
want
solved
with
serverless
functions?
First,
one
is
startup
time.
You
know,
Amazon
Lambda
is
still
even
today,
200
to
500,
millisecond
cold
start
time.
We
were
looking
for
something
in
the
10
to
20,
millisecond
cold
start
time.
You
know
we
needed
the
cross-platform
cross
architecture.
B
You
know
to
really
realize
the
idea
of
serverless.
You
can't
really
be
telling
developers
if
you
develop
it
on
Windows,
then
you're
going
to
have
to
run
it
in
a
Windows,
serverless
environment.
If
you
develop
it
on
Linux
you'll
have
to
run
it
in
the
Linux
serverless
environment.
We
really
thought
we
could
do.
You
know
find
an
environment
that
would
be
more
amenable
to
this
kind
of
cross-platform
thing
and
then
developer.
Experience
is
a
big
deal
to
us
right.
We
just
felt
like
the
way
and
we
were
at
Microsoft.
So
we
were.
B
This
was
an
act
of
self-criticism
at
the
time
where
we
were
going
we're
asking
developers
to
keep
cobbling
together
very
sophisticated
systems
and
we're
just
inserting
functions
as
a
service
as
one
more
piece
that
they
have
to
you
know,
string
together
all
the
way
through
all
the
services.
We
need
to
figure
out
a
way
to
make
this
developer
experience
really
really
streamlined.
B
So
we
started
looking
at
technologies
that
we
felt
like
would
be
good
for
that
and-
and
you
know
we
played
around
with
you-
know-
various
stripped
back
virtual
machines
and
unit
kernels
and
and
and
we
tried
out
a
number
of
kind
of
low-level
runtime,
Technologies
But,
ultimately,
webassembly
really
bubbled
up
to
the
top.
We
looked
at
it
and
said
the
security
profile
that
that
you
need
to
run
webassembly
in
the
browser
is
almost
identical
to
the
security
profile.
You
need
on
the
cloud
side
right.
It
needs
a
strong,
sandbox
isolation
model.
B
That's
really
I
mean
that's
really
important.
We
think
think
about
how
much
you
trust
your
web
browser
to
protect
you
from
bastardly
things
that
could
be
done
by
Bad
actors
on
the
internet
right
that
level
of
security
was
the
same
kind
of
thing,
so
that
check
the
biggest
most
important
box.
But
then,
as
we
got
into
webassembly
and
the
cross-platform
cross
architecture
in
nearly
instant
startup
goals
that
the
project
had
it
was
just
one
after
another,
we
were
checking
off
the
check
boxes
on
our
wish
list
for
this
kind
of
cloud.
B
A
Yeah
so
so
Matt,
it's
interesting,
I.
Think
of
some
of
the
parallels.
You
know,
Docker
didn't
create
containers,
but
they
really
helped.
You
know,
bring
it
out
to
the
masses
and
make
it
there.
Your
team
didn't
create
web
assembly.
Helped
me
understand,
you
know
what
what
is
the,
why
and
kind
of
the
the
the
the
core
capabilities
that
you
know
your
team
have
and
kind
of
the
vision
that
you
have
of
you
know
what
you
take
that
existed
and
you
know
where
it
will
go
from
there.
B
Yeah,
the
docker
analogy
is
so
good
because
I
think
if
Docker
Docker
managed
to
cut
a
couple
of
years
off
of
the
adoption
curve,
that's
the
standard
adoption
curve
and
they
did
it
in
a
fascinating
way.
They
had
an
operations,
oriented
technology
and
they
convinced
developers
that
developers
that
they
wanted
to
use
it
right
and-
and
you
know,
operations
teams
are
much
happier
when
the
developers
come
to
them
and
say
we'd
like
to
make
you
know
the
operational
experience
easier
for
you,
because
it's
easier
for
us
on
the
development
side.
B
So
we
really,
you
know
we
borrowed
liberally
from
the
docker
playbook.
In
that
sense
and
said
you
know
the
first
time
I
ran
through
the
the
Lambda
hello
world
example.
It
took
me
47
minutes,
I,
actually
timed
it,
and
we
thought
that
is
just
that
is
a
hard
developer
experience
to
sell.
You
know
you
want
a
developer
to
be
able
to
sit
down
at
the
end
of
the
day
on
Friday
and
say,
I'm
I've
got
some
time
to
learn
something
new.
What
you
know
what's
out
there,
that
I
can
learn.
B
The
last
thing
you
want
to
do
is
have
them
frustrated
because
of
all
the
little
steps
they
need
to
take
to
get
Hello
World
running
before
they
can
really
experiment
with
it,
and
we
we
said:
okay,
that's
that's
a
hurdle
that
we
can
take
on
right.
Can
we
can
we
make
the
developer
experience
so
good
for
serverless
applications
that
they
can
do
it
in?
B
Oh
I,
don't
know
two
minutes:
let's
get
them
to
hello
world
in
two
minutes,
so
that
became
sort
of
like
a
a
crucial
story
for
us
for
our
first
year
as
fermion
was
as
a
developer.
I
can
get
from
a
blinking
cursor
to
my
first
deployed
application
in
two
minutes
or
less,
and
that
that
we
thought
okay,
then
we
have
a
chance
at
articulating
our
actual
value
prop,
which
is
we
can
run
faster.
We
can
remove
a
bunch
of
Road
speed
bumps
in
the
in
the
road
to
getting
your
application
deployed.
B
We
can
run
cheaper.
We
can
run
cross-architecture
across
platform.
All
of
those
stories
really
take
a
little
bit
of
experience
before
the
developer
understands
the
benefit
of
them,
but
if
we
can
just
make
the
developer
experience
excellent
right,
really
awesome,
then
people
will
be
we'll,
try
it
out
and
say:
oh
well,
just
for
this
I
would
like
to
use
it.
That
I
think
is
essentially
what
Docker
did
right.
Yes,
you
could
build
containers
if
you
had.
A
Absolutely
so,
quick
aside,
we
actually
got
a
a
comment
from
the
audience
they,
after
your
thoughts
on
Gollum
Cloud,
they
said
was
it
wasn't,
makes
it
possible
I'm,
not
real,
familiar
with
it.
It's
something
about.
You
know
decentralized.
You
know
Computing
and
the
like,
but
if
that's
that's
one,
you
can
answer.
B
Yeah
I
mean
and
you'll
see,
you'll
see
a
number
of
of
attempts
to
figure
out
what
parts
of
webassembly
we
can
leverage
Golem
hasn't
been
totally.
You
know,
they've
got
a
couple
of
blog
posts
and
and
I
think
it's
still
a
closed
Source
solution
at
this
point,
but
you
know
part
of
what
they're
looking
into
is,
whether
or
not
you
can
freeze
and
unfreeze
webassembly
modules,
and
this
actually
gets
to
a
very
interesting
point
about
webassembly
itself.
You
know,
as
a
byte
code,
a
stack,
oriented,
byte
code
format.
B
You
can
do
some
very
interesting
things,
because
you've
got
the
encapsulated
sort
of
the
entire
running
unit
in
one
one
nicely
snapshotable
stack.
We
played
around
with
this
at
Docker
with
Docker
for
a
long
time
too,
because
Docker
is
very
similar
and
there's
been
this
idea
for
a
long
time
that
it
might
be
possible
to
freeze
move
and
unfreeze
a
Docker
container
right
and
I.
B
Think
there's
a
lot
of
interest
in
the
webassembly
community
in
general
and
that
same
idea
can
we
freeze,
move
and
unfreeze
a
webassembly
module
and
the
same
issues
that
came
up
with
Docker
are
going
to
be
the
same
issues
we
run
into
with
webassembly.
If
you
open
a
socket
and
then
move
something
over
to
another
server,
you
might
move
the
binary,
but
you
you
also
have
to
figure
out
a
way
to
move
to
make
it
look
transparent
to
the
binary.
B
The
first
stage
of
any
interpreted
language
is
to
sort
of
load
all
the
source
into
memory.
You
know
Parts
everything
load
everything
and
then
you
start
executing.
So
when
we
looked
at
this
there's
a
great
project
from
from
the
bytecode
alliance
called
wiser
w-I-z-e-r.
That
allows
you
to
sort
of
make
an
init
step
and
then
freeze
your
webassembly
binary
back
out,
and
so
you
know
with
something
like
a
JavaScript
runtime.
You
can
load
all
the
JavaScript
into
memory,
snapshot
it
as
a
webassembly,
binary
and
shipped
the
pre-initialized
one.
B
So
for
some
of
these
systems
like
net
Python
and
JavaScript,
you
can
actually
typescript.
For
that
matter
too.
You
can
actually
speed
up
the
startup
time
over
a
sort
of
native
implementation,
because
you
can
pre-load
a
bunch
of
stuff
right
up
front.
So
you
know
I,
think
Gollum.
Several
other
people
were
looking
into
US
included,
we're
all
kind
of
looking
into.
What
can
we
do
with
that
kind
of
snapshotting
ability?
What
you
know?
What
are
the
limits
to
which
we
can
push
it
Golem's
got
very
aggressive
limits.
B
We
we
tend
to
take
a
very
pragmatic
approach,
because
serverless
functions
are
our
primary
focus
right
now,
and
we
know
those
aren't
going
to
last
for
days
months
days,
weeks
months,
right,
those
are
going
to
execute
in
set
milliseconds
to
minutes
and
so
snapshotting
a
running
application
is
less
attractive
to
us
than
some
others.
But
it's
cool
to
see
that
work
kind
of
happening
in
the
ecosystem
at
large,
okay,.
A
So
so
Matt
help
us
understand
a
little
bit.
You
know
we're
talking
a
lot
at
kind
of
the
the
application
run
time
you
mentioned
Lambda.
Does
it
does
what
you're
working
on
work
with
any
serverless?
Does
that
bleed
into
kind
of
the
container
in
container
serverless
things
you
know
or
VMS
left
out
in
the
dust
you
know
what?
What?
Where
do
you
fit
with
kind
of
the
that
infrastructure
layer
that
many
of
us
deal
with.
B
Yeah,
so
at
the
core,
what
we
wanted
to
build
was
a
was
a
set
of
developer,
tooling
and
a
runtime
for
serverless
applications.
So
what
does
a
runtime
look
like
for
a
serverless
application?
A
serverless
application
is
an
application
in
which
you,
the
software
developer,
do
not
write
the
server
code
right.
The
thing
that
stands
up
a
a
process
listens
on
a
socket
manages
the
interrupts
all
of
that
kind
of
stuff.
Instead,
a
serverless
application
is
one
where
your
application
code
takes
a
request
or
an
event
handles.
B
It
returns
a
response
and
shuts
down
so
again,
you're.
Looking
at
the
lifespan
of
a
serverless
function,
being
you
know,
milliseconds
to
maybe
15
20
minutes
at
the
top
end,
right,
so
very
short,
running
piece
of
code
compared
to
a
server
which
is
designed
to
start
up
and
run
for
hours.
Days
months,
you
know,
and
so
on.
So
by
focusing
on
just
building
that
piece
and
that
spin
and
fermions
spin
open
source
project,
we
can
provide
the
run
time
for
a
serverless
environment.
B
Webassembly
happens
to
be
the
best
binary
format
to
execute
these
kinds
of
things,
but
that
runtime
can
then
be
portable
across
a
wide
variety
of
environments.
So
for
me,
on
Cloud,
for
example,
runs
inside
of
AWS.
We
use
the
entire
hashicorp
Nomad
stack
to
to
run
everything
nomad's
such
a
flexible
scheduler
that
we
had.
B
You
can
run
it
inside
of
Docker.
Desktop
Docker
desktop
supports
this.
In
each
of
these
cases,
really
all
you're
doing
is
saying.
Okay,
all
we
need
to
do
is
is
map
the
idea
of
executing
a
very
quick
serverless
runtime
to
the
idea
of
being
able
to
schedule
and
Route
traffic
into
this
particular
environment,
and
so
to
that
end,
our
goal
is
to
make
it
possible
to
run
this
in
just
about
any
kind
of
cloud
environment,
any
type
of
type
of
iot,
environment
and
locally.
Okay,.
B
And
it
breaks
easily
into
a
two-part
story
right.
First
of
all,
how
do
we
build
these
kinds
of
applications?
Spin
is
an
open
source
tool
that
allows
you
to.
You
know
very,
very
rapidly
build
a
new
webassembly
based
serverless
function,
so
you
basically
type
spin
new
it'll
scaffold
things
out
for
you
spin
build
it'll,
build
it
into
webassembly
spin
up,
you
can
test
it
locally
spin,
deploy
it'll
push
it
somewhere
else
to
run
it
and
we
support.
We
have
top
level
sdks
for
JavaScript.
B
Typescript
python
go
rust
and
we're
adding
more
languages
as
we
go,
but
basically
any
language
that
supports
the
standard
called
Wazi,
which
is
one
of
the
W3
standards
for
webassembly.
We
can
execute
those
inside
of
spin.
So
spin
is
the
open
source
really
the
core
of
everything
that
we
do
and-
and
you
can
actually
go-
take
a
look
at
the
at
the
source
code
at
github.com,
Spin
and
kind
of
dive
into
your
heart's
content.
It's
all
written
in
Rust,
but
we
needed
a
place
to
run
these
kinds
of
applications.
B
So
we
built
a
reference
implementation
called
fermion
platform
that
just
takes
a
terraform
and
installs.
It
can
install
into
clouds,
ranging
from
you
know,
Azure
and
AWS,
to
sivo
and
Equinox
and
wherever
you
kind
of
want
to
run
your
thing,
that
that
was
kind
of
a
reference
implementation
to
show
how
one
might
run
it
in
production.
But
we
wanted
to
provide
you
know
one
layer
on
top
of
that
and
say
developers
want
the
easiest
possible
way,
and
this
is
a
big
thing
for
us.
B
B
Fermion
cloud
is
that
right,
it's
a
developer,
oriented
hosting
platform
that
allows
you
to
with
one
command,
deploy
out
there,
and
then
you
can
log
in
there
and
map
your
domain
to
it
and
wire
up
some
databases
and,
and
whatever
you
need
to
do
to
take
that
thing.
From
sort
of
like
the
idea.
You've
got
locally
into
a
production
environment
where
you
can
run
it
for
everyone
to
see.
A
A
So
you
know
Seattle
a
great
place
to
go
for
these
things
and
yeah
it's
one
of
those
because
webassembly
it's
been
around,
for
you
know
a
number
of
years,
but
boy
I
remember
like
a
year
ago,
the
buzz
on
it
was
really
really
high
expect
you
know
a
lot
of
companies
there
and
it's
not.
You
know
one
company
that
that's
driving
this
technology,
there's
a
pretty
big
Consortium.
You
know
you
mentioned
a
couple.
You
know
we've
definitely
I'm
sure.
We've
got
folks
from
Red
Hat
side
I've
been
hearing.
A
B
Red
hat
Red
Hat
was
one
of
the
founding
members
of
The
bytecode
Alliance.
They
have
been
a
tremendous
influence
in
the
webassembly
ecosystem.
Now,
spanning
years,
right
and-
and
it
is
Awesome
Con
is
really
cool.
They
should
give
a
shout
out
to
the
fact
that
cncf
has
gone
from.
You
know
a
kind
of
pre-day
thing,
the
awesome
day
that
was
really
oriented
around.
How
does
webassembly
work
with
with
existing
Cloud
native
Technologies
to
now
a
multi-day
conference.
B
That's
just
designed
to
be
open
and
pull
in
a
lot
of
people
from
the
other
other
areas
of
webassembly
right,
there's
still
a
huge
number
of
people
who
write,
browse
browser-based
applications,
there's
an
increasingly
or
very
rapidly
growing,
tooling,
environment,
and
when
you
think
about
what
would
be
the
weak
Link
in
the
web
assembly
plan
right
from
the
beginning.
Well,
the
weak
Link
in
the
plan
is
yeah.
B
It's
nice
that
you
can
have
a
runtime
that
can
run
any
language,
but
you
actually
have
to
get
every
single
language
Community
out
there
to
support
your
binary
format
and
an
event
like
wasmcon
is
a
great
place
to
bring
like
you
know,
representatives
from
the
python
Community,
the
kotlin
community.
You
know
the
Swift
Community
all
these
different
programming,
language
communities.
You
know
into
the
same
room
to
talk
about
the
same
problems
and-
and
you.
A
Know
I
remember
right,
I,
believe
Russ
Khan
is
also
co-located.
You
know
with
the
event
so
that
that'll
definitely
you
know
bring
in
people
there,
so
definitely
check
that
one
out
yeah,
cncf,
lots
of
event,
really
cool
stuff.
Of
course,
I'm
sure
there'll
be
a
lot
of
that
bleeding
into
the
kubecon
show
happening
at
the
beginning
of
November.
Also,
you
know
I
know
fermion
was
there
you
and
I
got
to
speak.
A
You
know
a
year
ago,
at
the
kubecon
North
America
show
awesome,
hey
Matt,
you've
got
for
me
on
Beyond
fermion
cloud.
Spin
of
the
open
source
project
recently
had
an
announcement
about
a
no
Ops
database.
I
believe
key
value.
It
was
that
was,
was
the
discussion
so
help
help
our
audience
explain
a
little
bit
about
the
recent
announcements
and
you
know
where
you're
expanding
your
offerings,
yeah.
B
Absolutely
for
us
again
going
back
to
that
as
a
developer.
I
can
go
from
blinking
cursor
to
deployed
application
in
two
minute
story.
You
know
that
was
our
first
year
we
got
to
the
end
of
that
year
and
said:
okay,
we
can
do
that
right.
Our
CTO
Roddy
was
very
proud
of
saying
it's
actually
66
seconds
now.
I
can
go
from
blinking
cursor
to
the
buttons,
but
we
started
saying:
okay.
Well,
what
are
the
other
kind
of
speed
bumps
that
are
frustrating
or
annoying,
or
that
developers
could
self-report?
B
They
could
live
without
right
and
one
of
them
was
setting
up
a
local
environment
that
matched
the
production
environment
and
another
one
was
wiring
up
services
right,
essentially
an
operational
concern
that
many
developers
find
themselves
having
to
take
on
and
become
experts
in
something,
or
at
least
you
know
proficient
in
something
that
they
really
don't
have
any
interest
in.
So
we
listen
to
that
feedback,
and
we
said
okay
is
it?
B
Would
it
be
possible
to
just
sort
of
remove
some
speed
bumps
out
of
that
developer
path
and
and
what
would
be
the
first
and
easiest
one
to
do
and
decided
that
the
easiest
one
was
one
that
we
learned
firsthand
in
the
early
kubernetes
days
right,
I
I.
Remember
as
I
was
writing
my
first
kubernetes
apps
right
I
was
one
of
those
people
that
was
like
Hey
everybody.
The
future
is
stateless.
Your
application
should
not
have
any
state
in
it.
B
You
know-
and
you
know
shouting
from
the
rooftops
along
with
many
of
us
who
are
early
in
that
in
that
ecosystem.
At
the
same
time,
wondering
in
the
back
of
my
mind,
yeah,
but
I
need
the
store
State
somewhere.
Where
am
I
going
to
store,
State
I
mean
God,
it's
got
a
persist,
and
so
you
know
and
of
course,
in
the
kubernetes
ecosystem.
B
Storage
system
was
the
first
and
easiest
step
for
us.
Let's
just
make
it
easy
to
store
key
value.
The
API
surface
is
Tiny,
so
we
built
a
local
version
into
spin
where,
if
you're
running
spin,
you
declare
I
need
key
value,
storage
and
instantly.
It's
just
that.
You
don't
install
anything.
There's
no
processes
to
manage
it's
just
part
of
spin.
B
My
API
calls
to
get
and
set,
and
things
like
that
that
worked
out
really
well,
we
were,
we
were
surprised
at
how
quickly
and
how,
how
quickly
users
adopted
it
and
what
positive
feedback
we
got
back
about
it
because
it
was
just
one
thing:
they
didn't
have
to
worry
about.
Writing
stateless
apps
felt
like
writing,
stateful
apps,
because
everything
was
just
happening
at
the
API
layer.
B
It's
just
a
database!
That's
there
and
you
can
start
writing
selects
and
inserts.
So
we
partnered
with
turso,
who
had
been
working
on
a
distributed,
sqlite,
flavored,
database
and
and
rolled
that
out
only
a
couple
of
weeks
ago
and
again,
you
know
we're
really
early
on
that
one,
but
still
the
feedback.
We're
getting
has
been
really
overwhelmingly
positive,
that
this
is
just
one
less
thing.
The
developer
has
to
worry
about
now,
very
quickly.
B
B
Here's
the
redis
database
over
here
and
inject
that
as
config
into
the
application,
so
that
they
have
an
opt-out
same
is
true.
If
you
want
to
use
postgres
or
MySQL
instead
of
SQL
Lite,
and
we're
excited
about
that,
because
it
feels
like
we're
starting
to
make
the
right
division
between
the
developer
concern
help
me
just
write
my
code
as
fast
as
possible,
test
it
as
fast
as
possible
and
get
it
deployed
without
a
headache
and
the
operational
concern
which
is
hey
from
day.
B
A
Well,
Matt,
you
know
I'm
a
little
surprised,
we're
30
minutes
into
the
call.
You
were
just
talking
about
some
of
the
challenges
that
we're
seeing
on
the
developer
side.
We
have
not
talked
about
generative
AI
at
all.
You
know.
You've
got
series
a
in
I'm
sure
by
the
time
you
go
to
the
next
time.
You've
got
to
have
that
mentioned.
At
least
you
know
every
paragraph
that
you
mentioned
it
in
here,
but
no
in
all
seriousness.
You
know
I,
remember
back.
A
You
know
a
number
of
years
ago
it
was
like
Docker
Docker
Docker,
like
everything
that
we
talked
about
you
know.
Docker
was
in
the
tech
space
right
now.
Generative
AI
is
all
of
the
discussion,
not
only
in
the
Tech
Community
Matt
I
said
it's
one
of
those
few
things
that,
like
normally
when
I
you
know,
get
family
gatherings
and
they're
like
oh
yeah,
Stu
works
on
computer
stuff.
A
B
Yeah
and
I
I
think
you
were
totally
right
about
the
the
magnitude
of
the
impact
of
generating
I
mean
it
reminds
me
of
when
my
you
know:
parents
discovered
the
first
websites
right
and
they're
going.
Hey,
can
I
go
to
this
WebMD
place
and
look
up?
Is
it
reliable,
am
I
going
to
get
good
information
out
of
this,
and
and
now
we're
you
know
we're
seeing
the
same
thing
with
early
AI,
where
my
my.
B
To
use
it,
my
parents
know
how
to
use
it.
Everybody
asks
me
when
is
it
taking
over
the
world,
and
sometimes
they
mean
that
in
a
good
way
and
sometimes
in
a
bad
way,
but
it's
a
it's
a
really
exciting
moment
right
and
once
more,
it's
it's
super
exciting
to
be
in
an
industry
that,
at
this
particular
Epoch
of
human
history
is
like
the
the
The
Innovation
curve
is
just
so
fast
so
and
the
bottom
line
there
is
yeah.
B
B
That
need
to
be
easy
enough
to
pick
up
and
include
in
your
tool
belt,
alongside
knowing
how
to
write
SQL,
knowing
how
to
use
a
key
value
storage,
knowing
how
to
use
a
cache.
Knowing
what
a
proxy
is.
You
know
that
kind
of
thing
webassembly
seems
to
be
peculiarly
well,
positioned
in
this
case
right,
because
the
core
value
prop
one
of
those
first
things
that
Drew
us
to
webassembly
back
when
we
were
looking
for
what
this
candidate
next
wave
of
cloud
computing
would
be.
B
B
So
I
think
webassembly
is
going
to
be
one
of
those
core
technologies
that
in
many
ways
blossoms
because
by
pure
happenstance
it
was
addressing
one
of
the
problems
that
that
you
know
the
the
sudden
burst
of
energy
in
the
generative
AI
space
has
indicated
is
actually
a
really
important
problem
to
solve
so
I'm.
Looking
forward
to
the
next,
you
know
year
as
I
think
we'll
see
a
lot
of
generate
AI,
oriented
tools
that
come
out
of
the
web
assembly
space
and
and
think
it'll
be
a
really
good
area
of
growth.
A
B
B
Webassembly
is
a
neutral
format
that
runs
in
all
these
places,
but
it
runs
inside
of
a
host
runtime
and
the
host
runtime
does
two
things:
it
executes
the
bytecode
instructions
right,
but
the
other
thing
is:
it
can
inject
essentially
inject
apis
directly
into
into
userline
and
say
Hey.
You
can
run
this
call
and
it'll
do
this
kind
of
thing.
So
that's
how
we
do
database
right.
B
We
expose
the
database
layer
through
the
hosts
through
the
host
and
then
when
you're
writing
code,
you're
writing
against
an
API
in
webassembly,
and
the
host
is
doing
the
work
of
translating
your
inside
code
to
whatever
needs
to
run
on
the
outside
environment.
So
in
that
way,
then,
as
long
as
you
can
standardize
the
interfaces
along
that
host
boundary,
then
everything
becomes
portable,
so
wazzy
the
webassembly
system
interface.
B
B
The
practicality
of
that
means
that
when
you
take
a
host
a
webassembly
host
like
spin,
you
can
plug
in
different
back
ends
to
that.
In
some
cases
it's
a
configuration.
You
know
if
it's
something
that
we
thought
about
and
support
already
like
databases,
then,
yes,
you
can
just
plug
in
your
your
mySQL
database
configuration
here
or
your
postgres
database
information
there.
In
other
cases,
for
more
exotic
things
is
the
kind
of
thing
where
you
actually
have
to
write
the
the
host
plug-in.
B
There
is
a
big
specification
working
its
way
through
W3
under
the
auspices
of
bytecode
Alliance,
who
does
all
of
the
writing
the
specs
and
then
doing
the
reference
implementation.
This
spec
is
called
the
component
model,
and
the
component
model
describes
a
very
a
very
clear
way,
almost
like
grpc
style,
like
a
way
of
articulating
what
a
host
provides
and
articulating
what
webassembly
modules
provide,
and
it
allows
you
to
sort
of
build
up
and
expose
different
services
and
in
new
and
interesting
ways
that
will
streamline
this
whole
process.
B
Wazam
Khan
I
expect
this
to
be
the
number
one
story
at
wasam.
Khan
is
what
are
people
doing
with
the
component
model
and
where
is
it
going
from
here,
because
it's
such
a
powerful
unlocking
technology
for
us
to
building
applications
that,
frankly,
we've
not
been
able
to
build
before,
because
there
just
wasn't
a
technology
analogous
to
this
yeah.
A
B
So
there
are
about
three
different
work
streams
that
I
think
will
really
come
to
fruition.
The
first
one
is,
there
are
a
whole
bunch
of
add-on
specifications
to
webassembly
that
are
working
their
way
through
W3.
Garbage
collection
is
one
when
the
garbage
collection
thing
lends
then
a
whole
bunch
of
other
languages
will
be
able
to
Port
over
very
easily
because
they
won't
have
to
bring
their
own
garbage
collectors
they're
all
kinds
of
like
low
level.
B
And
then
you
can
write
a
JavaScript
function
that
calls
into
the
Rest
by
library,
and
you
can
start
the
language
is
no
longer
relevant
because
they're
all
byte
code
they're
all
the
same
bytecode
format,
and
they
have
this
kind
of
standard
way
to
expose
imports
and
exports.
And
when
that
happens,
I
think
then
we
really
start
to
open
up
some
potential
where
things
like
you
can
use
numpy
from
typescript,
and
you
can,
you
know,
pick
your
favorite
parts
of
go
and
use
them
from
inside
of
rust.
B
Just
by
exposing
these
different
components,
that'll
be
a
big
game,
changer
and
then
I
think
that
the
third
and
last
one
is,
you
know:
we're
gonna,
see
a
lot
of
languages.
Moving
I've
been
trying
to
sort
of
diligently
track.
The
language,
ecosystem,
rust
and
C
moved
right
away
in
part,
because
rust
moved
alongside
the
specification
and
C
was
the
target
language,
but
now
we're
saying
JavaScript,
Ruby
python.net.
B
A
A
B
B
Then
he
would
go
back
to
give
her
the
same
thing
later
in
the
day
and
she
would
not
eat
it
at
all
and
so
he'd
take
another,
can
out
and
open
it
up,
and
this
cat
was
just
so
finicky,
and
so
he
kept
a
spreadsheet
on
his
fridge
just
trying
to
track
and
see
if
there
were
any
patterns.
At
one
point
he
I
called
him.
You
know,
in
our
run
on
one
he
said:
yeah
I
went
to
the
pet
food
store
and
bought
one
of
every
kind
of
canned
cat
food
to
see.
B
If
there's
any
pattern
at
all
to
what
she
wants
to
eat
and
what
she
didn't.
So
when
we
created
that
finicky
whiskers
game,
you
know
it
was
originally
sort
of
a
joke
on
on
slats
or
Ivan's
catch,
and
then
slots
has
sort
of
like
taken
on
a
life
of
her
own
and
we
all
wear
the
t-shirts
and
and
love
this
last
character
and
finicky
though
she
may
be
she's,
also
really
adorable.
A
Awesome:
well,
hey
a
cute
story,
always
love!
You
know,
bringing
the
real
life
into
what
you're
doing
so
Matt
butcher
yeah!
There's!
If
you
see
this
last
comment,
thanks
guys
for
the
amazing
job
you
and
your
team's
doing
really
great
to
catch
up
with
you
appreciate
you
sharing
with
our
community
what's
happening,
wish
you
the
best
luck
at
waslam,
Khan,
I'm,
sure
our
teams
will
be
crossing
paths
with
you.
So
thank
you
so
much
for
joining
us
yeah.
A
Right
and
thank
you
to
the
audience,
as
always,
please
feel
free
to
reach
out
if
you've
got
guests
topics
that
you
want
to
cover
Summer's
a
little
slow,
a
lot
of
people
on
vacation,
so
apologize
for
not
hitting
quite
the
regular
bi-weekly
Cadence,
but
definitely
lots
of
cool
things.
I'd
like
to
do
in
the
second
half
of
the
year.
So
until
next
time,
I
want
to
thank
you,
I'm
Stu
miniman,
and
thank
you
for
joining
us
on
your
journey
in
the
clouds.