►
From YouTube: Fission & The WebNative SDK - Boris Mann
Description
Learn about Fission’s process of going from protocol to product, while creating composable specs and libraries that help with interop and re-use across the ecosystem.
Demo the Webnative SDK, an edge computing stack.
Thoughts on interop & the magic of UCANs.
Published slides: https://noti.st/bmann/Kuk5iE/fission-the-webnative-sdk
A
So
I'm
Boris
man,
CEO
and
co-founder
of
fission
and
I'm
going
to
talk
to
you
about
fission,
the
company
and
how
we
got
here
and
the
web
native
SDK,
which
is
our
main
product
today,
so
Brooke,
Vision
CTO,
my
co-founder
and
I,
set
out
and
said.
Apps
are
too
damn
hard.
We've
been
working
in
the
ethereum
space
in
the
early
days
and
people
weren't
were
learning
how
to
do
smart
contracts,
and
it
was
all
new.
A
You
know
he
was
as
kind
of
back
end
as
a
service,
but
then
there
was
this
JavaScript,
tooling
and
and
in
general,
front-end
developers
had
to
become
full
stack
developers
and
do
devops
and
from
our
experience
in
the
emerging
Technologies
around
distributed
systems
around
cryptography
on
some
of
these
modes,
we
said
we
think
we
can
take
some
of
these
techniques
and
make
it
easier
for
developers.
A
So
that
was
our
plan
is
to
make
these
far
Edge
apps
and
that
we
would
hand
front-end
developers
an
entire
stack
in
the
front
end
of
identity
data
and
compute.
A
Let's
make
it
more
secure
and
private
that's
hard
to
get
right.
Let's
do
passwordless,
logins
and
end-end
encryption
very
hard
for
developers
to
take
in
any
system.
If
you
go
browsing
in
the
Amazon
console,
you're,
basically
forced
to
pick
and
choose
plug
things
together.
There's
a
good
analogy
that
says:
you're
trying
to
build
a
house
and
going
to
the
Amazon
console
is
like
Wheeling
into
Home
Depot,
and
you
hope
that
the
parts
you
put
together
correctly
that
they
don't
cost
you
too
much
and
they're
secure
end
to
end.
A
We
feel
it's
very
important
to
not
do
account
lock-in,
so
we
used
decentralized
identifiers
as
the
basis
for
account
identifiers.
A
user
could
bring
their
own.
It
was
centered
Central
to
the
user
rather
than
being
managed
by
a
server
somewhere
and
and
portable
between
systems.
Also.
Let's
make
user
data
portable
use,
ipfs
for
Content
addressing.
We
were
ipfs
pilled
very
early
on
and
believed
that
content
addressing
is
a
sustaining
Innovation.
A
It's
amazing,
it
actually
is
interplanetary
and
then
this
other
thing,
as
we
saw
serverless
becoming
the
dominant
thing
that
eventually
is
going
to
eat
everything.
A
I've
been
building
open
source
software
for
a
long
time,
we've
been
building
lamp
stack
software
for
30
40
years,
I
think
it's
basically
over,
because
it's
not
decoupled
enough
and
the
computation
cycle
of
taking
down
the
transaction
that
we
see
in
everything
from
serverless
to
Smart
contract
is
the
kind
of
thing
that
you
get
scale
out
of
and
we
need
to
rebuild
things
in
this
serverless
model,
but
we
can't
let
it
get
locked
up
in
one
of
three
U.S
corporations
that
are
currently
controlling
the
future
of
computing.
A
The
first
thing
that
we
did
well,
we
said
ipfs
is
great,
but
we've
got
to
store
private
data
on
it.
How
do
we
do
that?
Well,
it
will
encrypt
it.
A
If
you
do
encryption,
you
need
identity
and
Key
Management,
and
so
you
can
solves
the
problem
of
decentralized
auth,
based
on
dids
kind
of
like
oauth
at
the
beginning
of
web
2
needed
to
be
invented
to
securely
integrate
between
systems.
That's
where
you
can
sits
in
a
lot
of
different
ways.
A
It
helps
us
pass
around
keys
and
give
access
and
differential
access
to
winifests
So
Daniel
mentioned
a
file
system.
We
have
built
in
to
find
a
standard,
a
protocol
for
a
second
file
system
that
sits
on
top
of
ipfs.
It's
end
and
encrypted
it's
encrypted
at
rest,
and
it
has
capabilities,
versioning
and
a
few
other
features
in
it.
A
We
have
other
protocols
in
the
hopper
like
carpool
dialogue,
a
database
system,
and
some
of
you
may
have
seen
some
of
the
recent
interest
in
ipvm
and
ipvm
is
going
to
design
to
fit
that
serverless
use
case
where
we
can
displace
Amazon,
Lambda
and
similar
systems
by
running
wasm.
Wasm
is
the
thing
that
that
fission
will
will
focus
on
in
this
area.
A
Research
make
a
spec
so
write
it
down,
write
down
the
rules
not
with
a
formal,
centralized
entity,
but
with
other
people
in
the
community
who
are
working
on
it,
make
a
code
implementation
and
then
compose
that
Library
into
a
product,
something
that
I
want
to
start.
Seeing
more
people
think
about
is
the
right
balance
towards
a
web,
2
style
I'm
just
going
to
build
my
company
to
an
open
source
style
where
we
can
work
on
shared
libraries.
A
A
A
It's
people
coming
together
on
community
calls
using
a
GitHub
repo
to
maintain
a
spec
versioning
it
and
then
implementing
that
code
or
solving
certain
problems,
use
cases
and
say
hey.
You
know
what
I
think
this
is
a
real
use
case.
This
is
a
problem.
What
works
for
you
and
going
through
that
and
then
having
separate
implementations.
I've
come
to
the
realization.
I
have
a
long
background
in
open
source
that
even
more
valuable
than
any
chunk
of
Library
code.
A
That
you've
written
is,
in
fact
a
shared
spec,
because
then
someone
else
can
write
an
interoperable
piece
of
code
rather
than
just
having
an
implementation
that
you
need
to
be
bug
compatible
with
ucan
is
being
used
by
folks
like
web3
storage
and
new
sphere
new
spear.
Recently
I
was
one
of
the
impact
Grant
recipients.
A
Actually,
maybe
I
need
to
put
capiloon
in
in
that
area
as
well.
There's
a
few
other
folks
like
Bluesky,
who
are
looking
at
pieces
of
ucan
and
we've
worked
very
closely
with
and
we
hope
to
see.
Interop
with
the
web
native
file
system
is
being
looked
at
or
will
actively
be
implemented
in
buy
and
xero
capiloon
web
recorder,
ipfs
and
filecoin,
and
we're
constantly
having
more
discussions
of
people
being
interested
in
in
solving
this
and
interoperating,
and
we
make
fun
mascots.
A
If
you
write
an
implementation
of
any
of
our
specs
I
will
Custom
Design
you
an
illustration
that
you
get
to
pick
and
I
will
send
you
as
many
stickers
as
you
want.
Yeah.
A
And
now
we
come
to
all
of
that
protocol
stuff,
which
we
compose
into
a
product.
Web3
storage
has
built
other
products
with
it
now
we're
back
to
only
be
starting
being
able
to
do
the
thing
that
we
want
to
do
at
the
beginning
after
we've
gone
through
protocol
land,
let's
Empower
a
front-end
Dev
to
make
entire
apps
without
having
to
become
a
full
stack
developer
and
without
devops
for
various
versions
of
without
Works
in
a
browser,
including
mobile.
A
It
is
very
important
for
me
that
this
community
internalizes,
that
of
our
stuff,
doesn't
work
on
mobile.
We
are
not
serving
Humanity
right.
No
browser
extensions
lots
of
interesting
stuff
happening
in
a
lot
of
different
ecosystems.
With
browser
extensions,
we
don't
get
to
hundreds
of
Millions
with
browser
extensions
we
might
get
there
with
Native
mobile
apps.
A
Local
first
is
an
interesting
new
capability
that
I
think
ipfs
and
content.
Addressing
can
really
shine
at
I.
Think
of
things
like
World
of
Warcraft
using
BitTorrent.
What
BitTorrent
it
saved
them
a
ton
of
money
in
bandwidth
costs,
because
all
the
people
downloading
updates
shared
that
with
them.
A
Some
of
these
same
things
can
happen
with
Edge
apps
and
we're
going
to
start
being
ready
to
put
this
into
native
mobile,
apps
and
desktops
and
and
solve
problems
in
this
area
and
now
I'm
going
to
show
you
an
app
template
and
do
a
live
demo.
So
this
is
you:
can
it's
winifests
it's
put
together
in
a
JavaScript
template
that
developers
can
use
to
build
these
types
of
applications.
It
features
both
public
and
private
files
and
device
linking
basically,
what
we're
going
to
do
here
is
the
web.
A
Crypto
API
is
a
browser
API
that
we
rely
on
and
that's
how
we're
able
to
do
what
we
do
without
a
browser
extension.
This
was
the
first
thing
to
figure
out.
How
do
you
do
it?
How
do
you
secure
something
so
the
web
crypto
API
is
everywhere,
including
on
mobile
browsers,
and
it
keeps
a
non-exportable
private
key
inside
the
browser
so
not
exposed
where
it
can
be
read
by
any
JavaScript
program
or
browser
extension
or
or
anything
else.
A
So
that's
what
routes
us
to
this
device
and
what
we
derive
a
did
from
you
know
this.
This
says:
Connect
I'm,
not
sure.
If
we're
going
to
use
portable
data
in
user
local
identities,
what
it
should
say,
because
it's
not
really
log
in
other
communities,
say
things
like
connect
wallet,
I've,
seen
some
communities.
That
say
the
word
sync
there's
active
areas
of
research
that
we
have
to
do
more
here,
so
we
click,
connect
and
I
can
create
a
new
account
or
I.
A
Oh
yeah,
there's
developer
laughter
in
the
room,
amazing,
so
I'm
going
to
hit
register
initializing
file
system,
so
this
is
actually
booting
up.
Jsi
PFS
node
in
the
browser
shout
out
to
the
maintainer
over
here.
Thank
you
very
much.
A
We've
had
a
lot
of
people
say
like
look
under
their
computer
check,
their
Network
traffic
and
they're
like
how
is
how
this
work
content
addressing
is
pretty
magical.
So,
let's
create
an
account.
It's
extantiated
a
winifest
file
system.
In
the
background,
wait
what's
my
password,
it
uses
public
key
cryptography.
All
of
this
needs
wording
help
continue.
A
It's
now
saying
you
know,
like
literally
your
account
only
lives
on
this
device.
You've
handshake
with
some
persistent
services
at
that
Vision
runs,
but
your
account
is
just
here.
A
So
it's
saying
you
know
you
should
create
a
backup
device.
I'll
click
on
it
and
it'll
say:
connect
to
backup
device.
So
if
any
of
you
scan
this,
we
could
actually
handshake
into
your
device
we're
not
going
to
do
that
and
I
there's
ways
and
now
we're
like.
You
really
should
back
this
up.
If
you
don't
want
to
lose
your
stuff
right,
so
this
is.
This.
A
Is
public
cryptography,
so
some
of
this
key
management
stuff
is
a
little
hard,
yellow
I'll
risk
just
one
device
for
now,
depending
on
the
app
You're
Building
as
an
app
developer,
we
wanted
to
show
some
flows,
including,
if
you
think
right
back.
There
was
a
check
mark
that
says
this
is
a
shared
computer,
so
those
are
use
cases
that
we
can
do
with
ucan
to
say
this
account
will
live
here
for
an
hour
or
for
half
an
hour
and
then
that
you
can
token
will
actually
automatically
expire.
A
I've
got
a
username.
It's
bugging
me
that
backup
is
recommended,
but
I'm
going
to
just
go
ahead
and
do
a
photo
gallery
demo.
So
this
is
the
public
side.
You
know
go
ahead
and
click
upload,
I,
don't
know
go
into
here
and.
A
Let's
pick
this
it's
pretty
fast
because
it's
just
doing
it
locally,
so
it's
added
it
to
the
local
JSI
PFS
node
and
in
the
background,
it's
communicating
with
fission
servers
to
persist.
It
I
can
now
click
on
view
on
ipfs.
A
A
Now
I'll
go
ahead
and
put
something
on
the
private
side:
let's
use
Wally
the
wallet
off,
mascot
again
same
thing:
it's
locally
encrypted
I've
got
the
local
Keys
encryption
keys,
so
I
can
see
it
and
read
it.
We
haven't
fixed
the
screenshot
problem
if
I
click
view
on
ipfs
it'll
actually
give
me
an
error,
because
we
need
to
update
this
template
and
it's
an
encrypted
blob
that
you
couldn't
read
anyway.
A
The
real
test
would
be
connecting
to
a
second
device
proving
that
you
could
read
it
in
that
second
device,
even
though
it's
an
encrypted
blob,
we
are
working
on
some
other
key
management
things,
including
so
this
looks
like
the
exact
same
thing
where,
instead
of
having
this
this
local
account,
what
you
can
do
is
hit
connect
and
in
fact,
look
look
at
that
beautiful
you
can
being
signed
by
your
metamask
ethereum
browser.
A
It's
now,
I'm
keeping
the
decryption
key
in
part
of
this
chain
in
metamask
and
being
asked
to
decrypt.
It
and
I
have
have
access
to
a
synced
file
system.
I've
used
this
before
on
other
machines.
A
Here
is
the
ucan
website
screenshot
and
here's
a
fun
photo
of
an
advisor
of
fission
Randall,
and
this
is
a
lovely
screenshot
that
you
wouldn't
want
shared,
which
is
why
we've
put
it
in
the
private
side
of
our
system,
not
quite
ready
to
be
released,
need
some
more
documentation
and
a
few
other
things
like
that,
the
code
is
there
you're
welcome
to
look
at
it.
You're
welcome
to
start
hacking
on
it.
It's
all
open
source
and
ready
for
you
all
to
play
with.
A
We
really
need
to
work
on
interop
we're
very
small
compared
to
the
entire
web
2
centralized
world.
We
have
no
choice
but
to
interrupt,
and
we
need
to
do
more
of
it.
Blockchains
talk
a
lot
about
this
a
lot
because
a
bunch
of
stuff
is
just
completely
public
data.
That's
not
ideal.
Either
we
do
need
encrypted
data.
We
need
decentralized,
Edge,
apps,
to
work
with
blockchains,
to
extend
the
utility
and
usability
to
do
things
like
ad
encryption,
large
file,
storage
and
computation.
A
That
doesn't
make
sense
to
do
the
more
we
work
on
many
intropical
accounts
and
portable
data
equals
a
bigger
web
3
Universe.
We
can
work
together
because
all
of
this
stuff
is
Central
to
the
user.
If
we
each
build
our
own
encryption
scheme,
we've
just
replicated
web
2
silos,
which
is
why
we
have
to
work
on
protocols
and
standards.
A
There
are
a
ton
of
ux
challenges.
If
someone
has
a
web,
3
storage
account
and
they
come
to
something
that
was
built
on
the
web
native
stack.
They
already
have
an
account.
How
do
we
tell
them
what
to
type
in
this
is
harder
than
it
sounds,
and
there's
some
interesting
challenges
around
this.
How
do
we
tell
people
that
their
data
isn't
locked
in
a
silo
that
they're
bringing
their
own
and
can
reuse
it?
A
So
this
is
an
XKCD
that
Brooke
sent
me
and
I
put
into
the
The
Talk
this
morning
it
says
my
new
secure
texting
app
only
allows
people
named
Alice
to
send
messages
to
people
named
Bob.
Clearly,
this
is
an
anti-pattern.
We
need
to
work
together,
a
few
future
directions
that
I'll
speed
run
through
I
showcased
wallet
auth.
So,
instead
of
managing
keys
with
web
native
use,
any
blockchain
account
right
now
proof
of
concept
supporting
ethereum
and
evm
chains
on
metamask
filecoin
will
be
next.
We
can
build
a
web3
iCloud,
not
a
joke.
A
This
is
a
real
thing.
We
can
do
it.
There
are
literally
no
blockers
to
this
asterisks
other
than
standards
and
interop.
It
also
gets
all
of
us
building
on
ipfs
50
million
ish
existing
accounts,
not
only
with
browser
extensions.
There's
things
like
wallet
connect
that
lets
us
use.
Mobile
apps
as
well.
A
50
million
apps
accounts
is
cool.
You
know,
what's
really
cool
500
million
accounts.
This
is
a
new
spec.
It's
in
iOS
16.1
and
basically
Apple
iOS,
Mac,
OS,
Google,
Android
and
chrome.
Microsoft
Amazon
are
all
going
to
support
this.
It's
from
the
Fido
Alliance
and
they're
like
we
really
should
kill
passwords
amazing.
There
are
a
ton
of
people
outside
the
web3
community
who
are
aligned
in
using
cryptography
to
kill
passwords
we're
on
the
same
team.
A
A
This
is
on
the
roadmap,
we'll
work
on
it
with
the
ucan
crew,
so
web3
storage
and
others
will
be
able
to
use
this
and
we'll
integrate
it
in
a
web
negative
so
that
any
developer
can
easily
offer
passwordless
passkey-based
authentication
I'm
very
excited
about
this,
because
it
has
the
potential
to
solve
certain
aspects
of
Key
Management
in
a
very
user-friendly
way.
A
To
give
you
an
idea,
it
was
amazing
to
hear
Jay
Chris
talk
about.
You
can
I
saw
the
cathedral
and
Bazaar
again
so
I
mentioned
before
you
can
takes
authentication
outside
of
a
single
siloized
system
and
means
that
you
can
remix
arbitrary
systems
just
like
you're
picking
and
choosing
from
the
Amazon
console.
We
can
now
do
this
on
the
open
web.
If
you
you
can
enable
your
app
let
that
sink
in,
we
literally
are
on
the
way
to
building
a
suite
of
services
that
interoperate
from
different
companies.
A
Yay
fission
intends
to
work
with
web
3
storage
so
that,
instead
of
having
a
single
integrated
system,
where
fission
does
all
the
persistence
you'll
be
able
to
put
in
your
web3
storage
credentials
and
there's
a
lot
of
ways
to
mix
and
match
that
bring
your
own
as
long
as
you
can
enable
it
you'll
be
able
to
plug
that
into
webnative,
SDK
PR's
welcome
doing
the
login
system
because
we're
using
compatible
did
based
accounts.
People
won't
constantly
have
to
create
new
new
accounts.
They
can
recreate
them
all
of
a
sudden.
A
A
A
Any
of
you
again
not
just
Vision,
could
create
a
wholesale
account
on
web3
storage,
either
as
a
developer,
with
a
single
app,
which
is
their
main
use
case
or
a
service
provider.
Writing
tooling.
That
is
amazing.
Let's
start
building
the
Amazon
of
web
3..
Thank
you.
Web3
storage
I
mentioned
it
at
the
beginning.
A
Identity
and
access
management
is
what
controls
the
hyper
clouds.
They
keep
building
more
services,
but
you're
inside
the
walls
of
that
with
ucan.
We
have
the
opportunity
to
build
something
that
is
in
fact
better
than
that
you
can
use
the
same
system.
Whatever
the
arbitrary
systems
are,
you
can
have
components
of
your
own
system
that
you
talk
to,
but
that's
the
same
as
talking
to
a
quote:
unquote
third-party
system
I'll
hand
wave
over
the
billing
and
business
agreements
that
you
still
need
to
do,
but
it
means
services
on
the
open
web
being
remixed
and
combined.
A
The
other
thing
with
those
other
systems
is
that
you
have
to
pre-provision.
We've
all
had
this
experience
as
developers
you
go
somewhere,
you
create
a
new
account,
you
get
a
developer
token,
and
then
you
have
to
come
back
and
can
can
do
your
system
and
also
you
have
to
figure
out.
Oh
I
have
to
run
a
server
because
I
can't
put
secrets
in
the
front
end.
A
You
can
put
secrets
in
the
front
end
because
you're
sending
signed
messages
with
you
can
and
that's
what
you're
authorizing
it
to
so.
Instead
of
the
ucan
connection
check
or
the
checking
checking,
if
someone
is
allowed
to
have
access,
they
can
talk
to
storage
directly
because
it's
signed
and
it's
been
delegated
from
your
wholesale
account.
There's
an
interesting
quote:
I'll
share
the
link
later
in
the
next
10
to
20
years.
I
believe
we
will
see
a
massive
disruption
in
the
developer-led
landscape,
driven
by
the
emergence
of
autonomous
software
development.