►
From YouTube: Block.one Plans to Integrate dfuse After Team Announces Open Sourcing Their Blockchain Data Stack
Description
Everything EOS is the longest running EOS podcast. In this episode, dfuse CTO Alex Bourget joins Zack Gall (LiquidApps) to discuss what led dfuse to recently announce they are open sourcing their blockchain data stack. They also discuss where the idea for dfuse came from, their robust blockchain APIs, and their plans moving forward to support more blockchains such as Ethereum and Solana.
dfuse Open Sources Blockchain Development Stack to Provide Higher-Order Blockchain Data Services
https://www.dfuse.io/en/blog/dfuse-open-sources-its-blockchain-development-stack
* dfuse
* https://www.dfuse.io
* https://twitter.com/dfuseio
* https://t.me/dfuseAPI
****Everything EOS****
https://www.twitter.com/EverythingEOSio
https://youtube.com/c/Everything_EOS
http://t.me/Everything_EOS
https://anchor.fm/everythingeos
https://itunes.apple.com/us/podcast/everything-eos/id1434560651?mt=2
For information on sponsorship opportunities or content partnerships, please contact zack at everythingeos.io or @BlockchainZack on Twitter and Telegram.
******************************************************************
LiquidApps
https://liquidapps.io
https://twitter.com/LiquidAppsIO
https://t.me/LiquidAppsOfficial
A
Welcome
back
to
everything
AOS
the
longest-running
yos
podcast
I
am
here
with
Alex
for
Jeff,
from
es
Canada
and
defuse
and
defuse.
If
you
missed
the
news
last
week,
they
made
a
very
very
major
announcement.
They
put
out
a
press
release
that
quotes
from
Dan
Larimer
at
everything.
Mind-Blowing
community,
the
first
community
project
to
I
guess
be
mentioned
by
block
one
as
far
as
using
the
services
we'll
get
into
it
later.
I'll
just
skip
it
for
now
a
little
bit
about
you,
though
you're
quite
a
beast.
A
With
your
background,
so
you
had
a
prior
startup
you
co-founded
vid
credits,
which
was
the
very
first
Bitcoin
payments
processor
in
Canada
and
then
prior
to
diffuse,
any
owes
Canada
you're
working
at
password
box,
which
was
acquired
by
Intel.
So
were
you
the
lead
data
scientist
for
Intel
or
for
password
box?
What
was
that
password.
B
Box,
though,
you
know
my
co-founder
here
in
diffuse
marc-antoine,
he
founded
password
box
and
he
brought
me
in
to
do
all
the
data
sciency
stuff
and
when
we
were
acquired
by
Intel,
we
just
continued
and
become
an
important
part
of
Intel
security.
You
know
behavior
elastics
tracking
of
their
products
that
deployed
on
200
million
machines
per
year,
so
we've
learned
quite
a
few
of
the
ropes
to
to
build
data
system.
So
I
think
that
leads
into
you
know
what
we're
doing
today
pretty
well
I.
Think
that's
what.
A
So
after
password
box
was
acquired
by
Intel
you're
like
an
Intel
employee,
essentially
and
you're
the
lead
data
scientist,
it
says
Intel
security,
consumer
division,
right
and
I
I
have
this
later
on
I'm,
going
all
out
of
order.
It
doesn't
matter
I
just
wanted
to
call
out
that
one
of
your
investors
in
diffuse
is
actually
Intel
Capital,
V
Intel.
Does
that
have
anything
to
do
with
you
being
a
prior
employee?
There
is
that
something
that
happened.
B
Like
people
there
are
common
there
that
you
know
was
there
then,
or
there
now
helped
out
like
math
counseling,
would
know
best
about
that,
but
yeah
I
mean
they
appreciated
the
the
the
drive
and
the
entrepreneurial.
You
know
leadership
of
mark
and
know
some
of
the
members
in
our
company
today,
so
they
found
it
was
worth
it
to
invest
and
they
had
a
blockchain
a
stick.
They
wanted
to
tackle
too,
and
the
I
mean
right.
We're
smelling.
A
I
mean
its
Intel
man,
but
this
is
Krypton
and
Intel's
huge
to
have
part
anything
to
do
with
your
company.
That's
why
I
wanted
to
point
it
out,
but
we're
here
to
talk
about
defuse
we'll
talk
a
little
bit
about
iou's
Canada,
because
that's
kind
of
how
you
got
here.
So
why
don't
we
just
start
off?
Who
is
zeroes
Canada
and
what
is
defuse
right.
B
Yes,
Canada
is
the
jumping
board,
toplink
wing
wing
wing
right.
We
used
to
get
into
this
space
fast
to
be
known
and
it's
always
been
a
dream
to
build
their
product,
but
you
came
from
a
product
company.
We
wanted
to
build
something
great
and
we
launched
it
off
Canada.
We
saw
the
needs
there,
so
we
did
write
a
bunch
of
software.
B
We
like
crawl
into
the
source
code
and
dawn
three
and
Dawn
four
on
no
díos
right,
so
we
we
really
springboard
on
iOS's
launched
and
so
from
there
we
observed
there
was
a
need
to
orchestrate.
You
know
a
launch.
So
I
don't
know
if
you
guys
remember
the
zoom
cause
of
200
people,
their
age
launches
and
based
on
your
looking
for
the
best
database
or
actually
consensus
right.
B
A
Was
a
big
deal
for
the
longest
time?
If
you
ask
someone
what
they
knew
about
iou's
Canada,
they
would
be
like.
Aren't
those
the
guys
that
that
launched
the
EOS
BIOS,
and
that
was
like
the
thing
you
guys
were
known
for
I
knew
you
guys
since
before
the
main
that
launch
I
was
talking
to
Josh
for
it
for
a
long
time,
he
was
one
of
the
big
thought
leaders.
A
B
When
the
chain
unlocked
cos,
Canada
was
number
one
right.
It
might
be
just
a
random
event,
but
you
know
I
like
to
think
it's
not,
but
anyway,
a
lot
of
people
have
moves
like
that,
and
you
know
today
Sokol
solely
so
this
the
the
stories
that
slowly
we
shifted
our
focus
like
as
we
wrote
Yas
by
us.
We
actually
wrote
the
first
go.
B
Libraries
for
us,
but
didn't
exist
at
that
point
and
we
use
those
things
to
slowly
build
up
tool
set
toolkits
and
things
that
were
bring
value,
and
we
have
identified
issues
like
first
of
all,
reading
the
chain
right
and
having
all
the
DNA
exfiltrated
it
and
mind
you.
That
was
a
data
problem,
something
we
were
very
good
at
and
wanted
to
tackle
in
the
blockchain
spaces
like
a
brand-new
space
words,
that's
data
is
actually
more
valuable.
Often,
maybe
not
all
the
value
right
these
days
and
meinen,
but
you
know
usually
on
the
blockchain.
B
You
have
a
high
value
data
flowing,
so
it
being
in
that
space
was
very
second
so
and
the
first
things
we
did
I
was
I.
Remember
asking
Dan:
when
are
you
going
to
exfiltrate?
You
know
action
deltas,
because
I
was
very
interested
in
the
database
and
to
me
you
know,
block
chains
are
databases
right
to
me.
You
choose
a
database
based
on
the
data
model.
C
B
Choose
because
you
want,
you
know,
collections
and
you
you
don't
need
joins.
For
example,
you
have
objects
that
are
nested
within
a
document.
You
choose
our
DBMS.
There
are
relational
databases
because
you
want
to
do
joins
and
you
want
to
have
those.
You
know
sorts
of
query.
The
data
model
of
databases
influences
why
you
choose
it
and
I
appreciate
that.
For
me,
yes,
I
do
a
lot
right.
B
The
fact
that
we
had
tables
with
a
bi
Abul,
you
know
schema
full-on
chain
schemas
with
the
data
that
you
could
externally
have
decoded
in
JSON
simplifies
great
deal
of
reading
aspect.
I
wanted
to
have
access
to
that
data,
and
we
slowly
built
diffuse.
We
we
made
our
instruments
at
our
nose
and
built
ourselves
that
we
needed
to
be
able
to
answer
this
question.
How
would
we
want
to
query
a
blockchain?
Have
we
not
been
influenced
by
you
know
original
JSON,
RPC
methods
or
other
ways?
B
Nodes
typically
offer
access
to
the
data
within
because
there's
a
lot
of
things
that
couldn't
give
you
the
level
of
guarantees
that
me
as
a
developer,
I'd
like
to
have
from
a
database
right
closer
to
a
city
closer
to
immutability
and
not
being
you
know,
fooled
by
querying
two
nodes
that
are
not
exactly
the
same
block
height.
One
says
one
thing,
the
other,
the
other
one
you
know,
I,
don't
want
I,
don't
want
to
have
to
deal
with
that.
I
want
my
code
to
bull
as
a
developer.
I
think
that
brings
a
lot
of
value.
B
B
Anything
to
build
while
we
were
you,
know
top
top
of
the
chain
there
and
something
we
felt
that
everyone
we
saw.
Also.
Everyone
was
writing,
because
everyone
wanted
to
reliable
way
to
query
the
database
and
we
can
get
into
the
different
products
we
have
there
after.
They
know
that
the
levels
of
guarantees
we
wanted
and
people
also
want
it
and
they
worked
hard.
So
it
really
diffuses
replaces
the
read
stack
every
company
using
a
blockchain
needs
and
that's
40oz,
but
it's
also
true
for
theory
I'm,
another
change.
B
We
can
see
cropping
up
these
days
because
the
protocol
teams
they're
really
working
hard
to
solve
scalability
issue
and
the
scalability
issue
is
right:
throughput,
WR,
ite,
so
writing
throughput,
but
that
doesn't
leave
much
space
for
reading
right.
If
you're
concentrating
all
your
effort
to
put
as
much
data
that
you
can
on
you
know
a
set
up,
a
master
to
master
replicated
nodes,
that's
that's
what
we
had
today,
then
you
hop.
Then
you
have
those
nodes.
You
need
to
query.
Also.
We
figure
that
doesn't
work
like
that.
B
B
The
node
does
that
so
our
and
then
up
we'll
link
into
other
question
I'm
getting
ahead
of
myself
here,
the
core
design
principle
we
brought
there
was
that
we
exfiltrate
the
data
as
fast
as
possible,
bring
into
systems
that
would
allow
us
to
consume
the
different
facets
of
information,
the
tables,
a
transaction,
the
history
and
what
happens
here
in
there
you
know
and
then
yeah
into
systems
that
would
optimize
for
that
and
scale
well
for
those
purposes
and
give
higher
guarantees,
because
you
know
would
give
API
wise,
better
semantics.
Does
that
make
sense?
It
does.
A
I'm
just
trying
to
get
wrap
my
mind
around
time
frames
because
I
didn't
research
that
piece
before
this
call
what
when
did
diffuse?
When
did
you
start
working
on
it?
Was
it
prior
to
the
main
that
launched
after
the
main
that
launched
and
then
when
did
you
release
like
the
first
usable
product
for
a
developer,
good.
B
Question
so
I
think
it's
always
been
in
German
right
since
the
awesome,
a
net
loss
we've
seen
we've
built
libraries
that
allows
us
to
allow
us
to
do
one
thing.
Then
we
built
a
search
engine
and
we
built
the
other
engine,
we're
gonna,
open
source,
a
fuse
called
flux,
TV,
that's
a
code
named
flux,
TV
for
state
snapshots,
and
then
when
was
that,
maybe
in
October
of
2018
we
then
changed
company
name.
We
started
offering
the
platform
as
a
service
for
main
net,
and
many
people
could
subscribe
at
that
point
and
then
we've
increased.
B
We
went
also
across
change.
So
during
is
the
last
I
would
say.
Six
eight
months
we've
been
working
on
a
theory
on
bringing
the
same
sort
of
radical.
You
know,
thinking
to
other
chains,
and
now
the
platform
diffuse
platform
is
ready
to
take
on
some
new
challenges.
For
example
the
Solana
guys
just
announced
you
know
we
working
it
with
that
with
them.
Don't.
B
Another
protocol
there's
many
protocols.
We
think
the
future
is
going
to
be
multiple
protocols,
do
appreciate,
Yas
IO
a
lot
and
there's
many
things
to
be,
as
I
could
go
ramble
about
all
the
good
things
in
there.
But
the
realities
is
many
chains
in
the
vision
we
have
for
the
future
of
blockchain.
Is
it
okay,
my
head
of
myself
right
now?
No.
B
Know
the
vision
is
that
diffuse
would
become
is
becoming
the
apache
spark
for
blockchain
data
flows
right
patchy
spark
is
that
you
know
incredible
tool
that
allows
you
to
manipulate
data
and
in
such
and
such
a
way,
and
we
see
diffuse
with
this
radical
approach
to
taking
data
out
of
block
chains,
building
a
nice
fully
and
complete
model
of
the
data
and
as
we're
open
sourcing,
we're
revealing
these
things
right.
It's
a
protobuf
model
of
every
piece
of
data,
that's
existing
in
chain
that
we
could
eventually
replicate
another
blockchain
with
that
data.
B
It's
fully
you
know
defined
and
then
some
streaming
machine
that
are
then
multi-protocol
aware.
So
imagine
if
you
have
the
best
data
model
for
ESI
of
chains,
the
best
data
model
for
aetherium,
which
I
think
that's
what
we
have
today
and
then
so
a
lotta
another
chain,
and
then
you
can
have
processes
are
listening
to
different
chains
taking
their
data
using
that
building
higher
order
services,
because
you
have
the
basic.
B
A
A
lot
of
data,
a
blockchain,
has
a
lot
of
data
and
diffuse
helps
you
sift
through
it.
So
before
without
diffuse,
let's
talk
Osias
pacifically,
since
that's
the
the
my
core
audience.
What
alternatives
even
are
there
for
diffuse,
because
I
know
projects
I
heard
like
karma
when
they
were
leaving
the
main
net
and
they
had
to
decide
between
Talos
and
wax
and
I.
Don't
know
if
I'm
pretty
sure
this
is
true,
but
they
chose
wax
for
one
main
reason,
because
they
needed
diffuse
and
there's
really
no
other
good
alternative
about
maybe
building
something
yourself.
B
B
You
know
some
of
the
things
that
that
we
hard
to
deliver
are
there
sort
of
guarantees,
but
let's
take
an
example.
If
you
want
to
query
a
Neos
table,
let's
say
the
voters
table,
there's
a
lot
of
voters,
a
million
voters.
More
than
that.
If
you
want
to
query
it
your
query
when
you
create
either
no,
do
you
get
a
thousand
rows
right
max
and
then
you
need
to
paginate
get
the
next
thousand
rows
and
then
another
thousand
rows
right.
B
But
by
the
time
you
reach
like
the
fifth
page,
maybe
someone
modified
the
first
row
and
then
how
do
you
know?
Maybe
there's
a
transfer
from
the
air
here
and
there
you
don't
get
a
consistent
view
and
it's
you
can't
expect
the
node
itself
to
keep
a
consistent
view
of
all
the
blocks
forever
right
mm-hmm.
B
So
we
built
for
that
and
dedicated
custom
solution.
That's
the
thing
called
flux,
DB
that
allows
you
to
get
consistent
views
of
tables
and
return
that
at
a
particular
block
ID
and
we've
designed
the
whole
platform
to
always
be
aware
of
all
Forks
and
keep
them
so
that
if
we,
you
are
left
off
in
some
random
before
that
doesn't
exist
anymore,
you're
at
least
always
able
to
understand
what
you
need
to
do
to
get
back
to
the
canonical
chain.
B
Right
so
any
case,
all
the
systems
have
been
designed
for
wear
throughout,
but
that
particular
problem
is
that
we
can
then
query
the
state
of
a
table
in
a
consistent
way
and
you're
gonna
download,
500
Meg's,
maybe
but
you're
sure
that
you're
not
gonna,
see
anything
moving
while
you're
querying
it
and
then
you're
gonna
get.
The
reply
saying
is
that
this
particular
block
ID
a
hash
right,
not
a
block,
none
that
you
not
sure
what
happens
next
and
then
you
can
get
streaming
deltas.
B
So
that's
the
sort
of
reliability
that,
if
you're,
if
you
don't
have
it
it
happen
happens,
is
this.
What
happens
is
what
the
folks
on
aetherium
have
done
when
we
they
told
us
when,
when
we
rolled
out
to
aetherium
is
that
they
need
to
query
one
node,
another
node
repeat
and
then
because
some
often
times
you
clear,
load-balanced
knows
right
and
you're,
not
even
sure
who
you're
querying.
So
when
you're
clearing
for
one
table
are
you
sure
which
block
ID?
C
B
You
need
to
do
well,
you
need
to
build
some
sort
of
database
that
abstracts
and
gives
you
that
consistency.
That's
what
we
built
a
lot
of
people
need
that
and
that's
just
for
one
aspect
querying
state
right
and
we
wanted
to
do
the
same
thing
for
all
the
facets.
Like
history
state,
you
know
when
you're
writing
to
the
chain.
Although
I
see
what
I
didn't
yeah
would.
A
You
yeah
some
of
this
stuff
is
above
I,
hope,
I,
hope
some
people
watching
get
this
I
know
the
technical
people
watching
will
hear
some
of
the
terminology
you're
using
and
get
it,
but
I
think
other
people
just
will
understand
your
excitement
here
and
just
look.
They
could
google
it
if
I
wasn't
in
the
middle
of
a
conversation,
I
would
have
probably
googled
a
couple
of
things,
but
well.
B
I
leave
that,
to
those
who
understand
detective,
please,
when
we
rolled
out
young
customers
are
saying
look:
I
can
kill
80%
of
my
code,
90%
of
my
code,
because
all
the
things
that
you're
giving
me
a
simple
API
with
now
I
was
doing
with
retries
incurred.
Many
many
many
problems
and
latency
I
need
to
reconcile
that
on
my
own.
That's
the
client
right!
That's
annoying!
If
you're
using
my
sequel,
you
don't
want
to
be
doing
all
that
in
your
own
code.
You
want
don't
delay
it.
A
B
Needs
to
do
it
because
everyone
wants
to
reliable
access
to
the
chain.
Data
right,
let's
play,
makes
so
much
sense
to
take
that
and
I
think
that
might
be.
You
know
why
block
one
is
interested
in
these
things
is
that
it
pairs
so
well
with
them
fiddling.
You
know,
write
throughput
where
diffuse
is
really
solving
the
read
aspect
on
many
ask
many
facets:
I'm.
A
Sure
you're
very
well
aware
of
since
probably
November
since
the
blocksberg
summit
block
1
the
the
the
way
they've
been
describing
yes,
I/o
is
like
a
next-gen
database
and
Dan.
Really,
when
he's
describing
the
advantages
of
Eos
I/o
over,
like
a
sequel
database,
he's
always
talking
about
the
right
speed,
he
said
a
sequel
database.
It
could
read
super
fast,
but
it's
really
hard
to
have
deterministic
writing
and
that's
what
you're
saying
here
is.
B
B
It
would
make
sense,
it
would
make
sense
sick
and
they
need
to
handle
the
write.
Throughput
I
mean
and
there's
so
much
time
in
a
day
that
you
can't
solve
everything
all
the
time
and
then
you
can
get
ideas
brewing.
So
that's
our
take
on
you
know.
The
API
is
I.
Think
it's
a
great
option
in
this
space.
I
think
it's
currently
the
best
option
in
the
space.
Also,
but
no
one
I'll.
A
Go
to
the
press
release
that
you
guys
later,
the
blog
article,
whatever
you
want
to
call
it
and
I'll
quote:
Dan
Larimer,
he
said
block
one
welcomes
the
significant
technical
contribution
from
diffuse
and
we
plan
to
closely
integrate
it
with
ESI,
oh
and
other
products.
We
are
also
proud
that
the
fuse
emerged
from
the
EOS
I/o
ecosystem
and
that
they
share
our
vision
to
accelerate
the
global
adoption
of
Bach
chain.
A
That's
huge
that
that
is
a
huge
endorsement
right
there
and
I
think
it's
the
biggest
one
block
one's
ever
made
for
a
project
up
up
until
this
point
so
when
he
says
closely
integrated
with
ESI
owed.
This
deceive
me:
does
that
mean
like
into
the
base
layer,
infrastructure
or
or
what
this?
What
do
you
think
he
means
by
that?
So.
B
A
base
layer,
no
the
diffuse
platform,
is
all
written
in,
go
it's
a
set
of
services,
but
we're
shipping
diffuse
boxes.
You
know
you
just
plug
it
on
your
machine,
you
boot
and
you
get
the
Explorer.
You
ask
you
there
you
get
the
search
system,
which
is
the
log
real-time
flux,
DB
you
get
you'll
get
in
there
right,
so
it's
really
booting,
very
simply
for
you.
But
this
thing
also
scales
out.
Why
am
I
saying
that
you're
saying
how
integrated
the
system
is
not
gonna
be
bolted
inside
no
do
I
give
the
reason.
B
I
I
think
it
doesn't
make
sense.
You
don't
want
to
know
to
do
that.
Like
I've
seen
other
protocols
I
think
it's
Clayton
another
watch
protocol
for
out
of
Korea
they
actually
designed
their
systems
from
the
get-go
separate.
The
right
system
gets
the
throughput
for
writing
and
it
pushes
any
data
to
another
system.
Is
there
for
reading
right?
They
design
it
from
the
get-go
I.
Imagine
this
right,
I!
Imagine
nobody
else
and
you'll
see
right
now,
as
of
today,
there's
a
patch,
the
the
famous
diffuse
instrumentation
is
out
there.
B
Now
it's
public
because
we're
open
sourcing.
If
we
decide
to
do
that,
so
no,
no
turning
not
so
you
know
what
we're
doing
is
taking
the
data
out
of
node
EOS,
bringing
it
in
in
this
data
model
new
data
model
and
then
the
streaming
machinery,
that's
the
sort
of
you
know
integration.
You
could
imagine
right,
it's
not
that
we're
gonna
rewrite
everything
in
C++,
put
it
inside
node,
EOS
and
node
EOS
becomes.
That
became
I.
Don't
think
that
you
really
never
want
to
run
right.
No,
so.
A
C
B
A
B
Saying
is
half
as
a
search,
you
know
what
we'd
be
available
for
people
who
want
to
build
search
and
at
some
point
I
know
nothing
about
voice
really
I'm,
not
even
on
voice.
You
know
if
some
people
can
eventually
read
that
chain
and
then
get
the
data
out.
Somehow
people
could
build
search
engines
here
and
there
so
I,
don't
know
where.
A
B
Yeah,
did
you
try
to
make
sure
that
at
least
you
know
bought
one
was
aware,
because
it's
coming
their
way
and
it's
actually
the
feedback
I
got
is
that's
the
first
time
that
really
they
they
can
even
consider,
except
you
know,
public
libraries
like
I,
don't
know
you
know
open
SSL
or
libraries
that
everyone
uses,
but
it's
the
first
time
a
major
contribution
arrives.
There
wases
know
they.
They
ask
questions
and
they're
I
and
we
presented
to
a
few
a
block
on
folks.
B
You
know
what's
the
diffusor
architecture
and
all
that,
and
we
want
to
do
that
with
the
community
they're
very
soon.
This
is
streamed,
wet
I,
don't
know
what
but
I
don't
like
to
have
a
you
know:
public
zoom,
where
we're
gonna
go
through
the
architecture
so
that
people
understand
what
are
the
you
know,
nuts
and
bolts
and
what's
come
they're
way
too,
because
he
may
be
open
to
anyone.
It
works
very
well
with
aetherium.
Oh
sorry,
perfect
videos!
Well!
A
Okay,
it's
gonna
be
an
open
product
people.
Good
luck,
so
I'm
sure
that
that's
III
noticed
some
of
the
guys
on
Talos
I
that
were
basically
trying
to
re,
build
their
own
version
of
the
fuse
out
of
necessity
because
they
didn't
have
it.
No
one.
No
one
saw
this
coming.
I.
Think
that's
why
I
like
hit
everyone
like
a
freight
train.
A
Last
week,
I
saw
the
the
the
tweet
itself
got
like
200
likes,
which
is
like
insane
and
like
this
climate,
like
engagements
really
really
down
since
since
the
Corona
hit,
and
probably
everyone
lost
half
their
bitcoins
yeah.
So
you
have
a
couple
products
we'll
get
into
this.
There's
diffuse,
search,
diffuse
lifecycle
and
diffuse
state.
You
want
to
kind
of
give
a
tldr
on
like
what
each
one
of
those
are
and
what
purpose
they
serve
of.
A
Am
it
so
in
your
and
let
me
let
me
point
out
you're
on
not
you,
but
the
fuse
is
used
on
basically
every
block
Explorer
and
it's
almost
like
a
necessity
on
it
and
I.
Don't
think
like
on
blocks
IO,
for
example,
whenever
you
start
typing
in
an
account
name,
how
it
autocompletes
it
that's
all
possible,
specifically
because
of
diffuse
search,
correct
right.
B
I
don't
know
who
he
built
it
I
know
Syed
so
from
blocks.
He
uses
the
fuse
there.
A
lot
of
people
use
like
most
major
projects
that
want
reliable
access
at
least
have
a
diffuse
option.
Even
though,
but
now
we're
open
sourcing,
we've
seen
a
lot
of
people
opening
up,
say:
okay,
maybe
I
can
make
that
my
pram
primary,
because
there
was
concern,
maybe
at
some
point
that
you
know
being
closed
stores.
A
B
Diffuse
search
came
about
because
what
are
the
first
thing
that
happens,
that
when
a
transaction
is
sent
to
the
chain,
okay,
lady
posit
thinks
this
way.
The
transaction
in
a
blockchain
is
the
intent
right.
It's
like,
please,
you
know,
add
one
to
my
row
like
an
update
statement
in
my
sequel,
equivalent.
It's
the
intent.
But
what
really
happens
when
it's
executed
could
be
very
different
from
what
you
think
happens.
Like
you'd,
say
intent,
withdraw
my
money,
but
you
don't
tell
the
amount.
So
what
happens?
B
Is
that
once
this
executor,
you
get
a
boatload
of
data
called
the
traces
of
execution
right
and
it's
what
actually
happened.
This
is
actually
the
ledger
because,
if
you
say
withdraw
you're
withdrawing
a
million
dollars
or
five
bucks,
you
know
the
reality.
Is
that
thing,
which
is
the
side
effect
of
the
intent,
the
intent
being
withdraw
please.
But
now,
when
you
get
into
the
nuts
and
bolts
and
notice
at
the
beginning,
does
not
like
it
provided
a
few
tools
like
the
history
tools,
but
clearly
not
you
know
design.
B
It
was
all
in
RAM,
you
clearly
not
designed
to
scale
or
to
know
to
just
survive.
In
this
form
and
way,
so
we
wanted
to
build
a
search,
but
then
again
from
first
principles.
How
would
we
want
to
query
a
history
knowing
that
we
have
a
chain
with
the
moving
tip
it
could
fork
out?
You
know
if
we're
very
well
aware
these
things
in
enemy
and
eyes.
A
lot
of
these
things
remember
the
graphic
Grapher
from
miios
canada.
B
You
could
see
there
that
for
the
nasty
four
that
were
happening
from
time
to
time
unit
from
graphically,
and
so
we
designed
search
that
is
aware
of
the
consensus
algorithm.
So
you
can
search
the
history
of
transaction,
including
the
trace
Lee.
If
you
no
did
I
have
a
million
dollars
transfer
out
of
that
withdrawal,
then
you
will
know
yes,
I
gave
me
the
transfer
a
list
that
transaction
has
having
existed
and
it'll
point
you
to
that
original
intent
right.
So
the
data
model
there
is
very
complete,
it
represents.
B
B
There's
rows
mutated
right,
so,
okay,
it
allows
you
to
go
through
the
history
backwards
and
forward,
but
also
when
you're
going
forward,
you
can
go
and
stream
into
eternity.
You
can
continue
listening
to
the
chain
for
that
particular
query,
because
mi
transfers
in
and
out
whatever
along
this
particular
contract
and
we've
expose
also
a
small
language
that
allows
you
to
craft
queries
that
are
more
rich
than
what
we
originally
had
with
the
history
API.
So
you
can
select
this
contract
or
that
contract
or
a
list
of
contracts.
B
Now
from
me
to
me
what
it,
whatever
right
bunch
of
good
and
interesting
facets,
you
can
query
on
and
when
you're
streaming,
that
into
eternity
you're
actually
getting
as
fast
as
possible
to
results.
You
know
video
they're
executed,
proof,
they're
index
and
then
you're
querying
that.
So
you
can
see
what
there's
a
new
transaction
but
hey
it's
in
a
block,
but
it
might
not
stick
there
forever,
because
you
know
some
force
I'll
cure
from
time
to
time.
David
Allen
is
no
producer
handoffs,
and
so
we
just
understood
so
that
it
would
inform
you
of
that.
B
So
maybe
it's
a
technicality
but
I
think
it's
important.
We
provide
upon
search
results,
a
cursor,
that's
a
small,
opaque
piece
of
information
that,
if
you're
disconnected
or
you
shut
down,
whether
your
eye
or
two
device
turns
off
for
an
hour
or
whatever
you
can
bring
it
back
and
when
you
say
continue
my
search.
B
It
will
then,
if
it's
fork
aware
it
will
know
that
it
left
you
on
something,
that's
potentially
forked
and
if
it
was
then
after
the
fact
it
will
give
you
like,
okay
undo
this
thing,
this
is
not
true
anymore
and
then
reen
advocate
the
the
canonical
chain.
Maybe
that
transaction
was
executed
separately
after
that
and
had
a
different
outcome.
Maybe
it's
not!
B
One
million
is
now
true,
meaning
that
you
have
withdrawn
right
and
then,
but
as
a
user,
you
want
to
know
that,
but
you
don't
want
to
pull
what
you
seeing
people's
do
is
that
according
history,
they
always
get.
You
know
a
certain
time
of
the
chamber
that
when
there's
a
reorganization,
they
get
a
different
piece
of
the
chain,
but
then
they
need
to
reconcile
these
two
things
together.
Super
annoying
right
all
in
coupon
code,
so
we
didn't
want
that.
No.
A
Fur
are
you
this
is
that
sounds
complicated,
especially
whenever
you're,
like
the
end
goal
here,
is
like
there's
a
lot
of
like
battling
like
back
and
forth
between
other
blockchains,
even
like
Josiah
chains
against
each
other,
a
theory
many
O's,
but
really
like
for
this
to
work.
You
have
to
on
board
just
developers,
developers.
A
B
Api
becomes
so
simple:
they
still
need
to
understand
that
something
they
received
could
be
undone,
although
we
do
have
a
flag
saying:
I
want
irreversible
only
right.
So
in
that
case
you
query
that
we
will
never
give
you
things
that
are,
you
know
could
be
undone
so
in
that
case
it
getting
there
guaranteed,
but
people.
You
know
they
already
need
to
learn
that
this
exists
right
when
their
uses
in
my
sequel
or
whatever
other
database
is
transactional,
and
if
you
ask
them
to
even
do
all
the
retry
themselves,
it's
crazy.
B
B
Let's
bring
about
the
best
that
the
underlying
chain
protocol
can
provide,
make
sure
that
they
did
an
API.
We
offer
like
never
yeah
tons,
I,
never,
never
impedes
it
in
any
way.
Right
makes
it
shine.
I
think
that's
why
people
like
diffuse
right,
because
this
brings
the
power
of
the
chain
underneath
and
the
easy
way.
A
Finally,
mmm-hmm
I
remember
Rob
Finch
when
four
years
named
service
I
guess
he
wasn't
using
where
his
contractors
weren't
using
diffuse
right
away
and
then
at
some
point
they
must
have
talked
to
you
guys
and
got
started
he's
like
what
the
heck
was
I
doing
it.
The
old
way
for
like
he's
like
this,
like
cut
the
development
time
I'm
paying
for
by
like
two-thirds,
because
it
just
made
it
so
much
easier
for
I,
don't
even
know
what
he
was
even
doing
for
years
name
service,
but
I,
remember
like
distinctly
whenever
that
happened.
B
Like
the
best
there's
a
few
things
that
also
we've
had
like
we're
open
sourcing
now
as
part
of
the
patch,
we
have
an
odious
which
going
into
core,
and
you
know
that
diffuse
platform
allows
you
to
do
that.
Didn't
that
didn't
exist
elsewhere,
for
example,
to
query
the
deferred
transaction
like
that
was
literally
not
available
anywhere.
B
But
you
know
the
diffuse
stream
would
bring
you
a
small
streaming
and
you
want
to
know
that
the
stage
of
that
transaction-
that's
a
life
cycle,
maybe
we're
getting
a
little
ahead
there,
but
the
life
cycle
endpoint
allowed
us
to
give
one
streaming,
endpoint
telling
you
what
that
transaction
has
been
created,
but
it's
deferred.
Here's
the
content,
it's
in
this
block.
If
this
block
is
undone,
maybe
that
disappears-
and
it's
not
executed-
never
but
then
movie,
there's
another
block
that
executes
sit
or
another
transaction
that
cancels
it
all
that
life
cycle
I.
A
Think
that's
the
problem.
You
guys
solved
with
the
name
service
they
they
found.
There
is
a
vulnerability
with
different
transactions
where
they
could
basically
imagine
I
sold
you
my
account
Anya's
name
service,
but
I
had
a
deferred
transaction
that
was
going
to
switch.
It's
not
like
after
I
sold
it
to
you
and
III
think
diffuse,
and
even
we
didn't
really
talk
about
it,
but
it
I
don't
even
know
how
you
guys
caught
I
called
esq,
but
you
guys
have
the
highest
precision
block
Explorer.
A
B
The
diffuse
that
we
set
out
to
provide
developers
because
we
thought
you
know
you're
a
developer
and
you
might
notice
that
EOS
queues
not
there
with
the
token
prices
right,
you
wanted
to
build
a
data
product.
So
that's
the
data
product,
but
it's
the
most
precise.
In
that
sense,
you
get
all
the
RAM
operations
per
action,
yet
all
the
data
base
operation.
So
you
can
know
what
was
changed
by
which
action,
which
is
the
basic
unit
of
execution
in
eoeo,
and
we
didn't
want
to
compromise
on
that.
B
That's
why
we
went
straight
down
to
the
source
and
instrument
to
that,
because
that's
the
level
of
data
we
wanted
to
bring
people
and,
of
course,
now
only
blocks
and
a
few
X
bar.
Is
they
feed
from
that
because
that
doesn't
exist
in
the
EOS,
no
Dios
land
at
this
point
until
until
now,
which
I
hope
you
know
if
this
does
that
patch?
But
you
know
it
also
requires
some
of
the
diffuse
machinery
so
to
make
sense
out
of
that
and
but
is
gonna
be
in
developers,
hands
right,
I'm,
really
happy
to
so.
A
B
One
source
thought
since
a
long
time
about
open
sourcing
and
we
went
to
like
took
it.
Let's
say
year
for
opening
and
then
deciding
a
year
and
a
half,
maybe
and
before
you
know,
opening
for
service
and
people
could
use
it.
There's
a
lot
of
learnings
in
there
and
it's
been
a
fine.
It's
been
a
fine
route
because
we've
been
able
to
go
very
fast,
you
know,
search
engine
is
that's
it.
It's
third
iteration!
It's
been
started
at
a
simple.
B
You
know
one
one
program
and
then
today
the
search
engine
is
they
actually
a
distributed
system
with
the
router
similar
to
s1,
and
you
know,
shards
and
then
live
segments
that
can
scale
like
we
put
a
lot
of
stress
into
that
and
we've
redesigned
a
few
things.
Maybe
it
would
have
slowed
us
down
to
have
that
open
system.
The
get-go
I
think
today
is
a
good
time,
because
we
have
good
nice
interesting
pieces
to
provide
and-
and
we
were
hesitating
at
some
point-
where
should
we
do
that?
B
And
then,
if
you
looked
at
the
incentives,
you
say:
okay,
what
is
that
gonna
mean?
And
then
we
wanted
to
try.
Also
to
learn
you
know,
as
a
start-up.
First
thing
we
need
to
do
is
learn,
learn
about
the
market,
learn
what
people
are
willing
to
pay
where
people
are
willing
to
go
right.
What's
the
pockets
of
money
you
like?
Is
that
like
on
main
yet
for
exams
a
lot
of
developers,
single
developers
want
to
use
maintenance
efficiently?
B
What's
the
price
is
30
bucks,
100
bucks,
a
thousand
bucks,
you
know,
and
enterprise
customers
those
who
build
their
business
on
top?
What's
the
size
that
we
wanted
to
learn
these
things?
Could
we
have
done
that
by
being
open
source?
Maybe,
but
we
wanted
to
do
that
and
we
we
went
sufficiently
deep
in
there
that
at
some
point,
looking
at
the
market,
we
felt
oK
we've
kept
something
that
could
empower
people.
B
How
can
we
empower
people
more
and
that
was
really
was
the
driving
factor
like
I
remember
when
I
started
into
Bitcoin
read
the
white
paper
read
the
Bitcoin
labor
seven
page
at
the
end,
I
was
like
wow.
This
me
me
I'm
a
developer.
This
is
now
in
my
hands.
I
can
now
program
money.
Imagine
that
I
remember
that
feeling
and
I
was
feeling
that
with
defuses
we're
going,
there
was
something
missing:
it's
a
hosted
version.
People
are
not
empowered
enough.
They
can't
build
on
this.
The
internal
building
blocks
we
try.
A
B
The
the
code
is
not
open
fully
yet
so
what
we
want
to
do
is
build
a
great
onboarding
experience
so
in
the
next
week
or
so
we're
gonna
start
like
send
in
a
few
a
few
cherished.
You
know:
Yoshi
ins
there
and
oh
the
diffuse
box,
the
binary
version
so
that
you
can
try
it
see
a
good,
have
a
good
experience
just
like
when
you
boot,
elasticsearch
right,
elastic
people's.
You
see
Wow,
so
we
want
to
give
them
that
experience.
Then
everything
that's
in
there.
A
B
A
B
B
One
thing
name
is
Pete
Rios
and
some
people
didn't
know
that
so,
first
of
all
the
EOS
door,
library,
is
the
library
of
reference
I
I,
see
today
for
interacting
with
Ito's
in
go
and
I've
seen
also
plenty
of
people,
even
a
blot
when
using
EOS,
C
or
a
command-line
tool
which
replaces
cleo's,
which
has
a
bunch
of
facilities.
People
can
contribute
I.
Think
it's
easier
for
people
to
jump
in
when
it's
in
go
so
you
know,
that's
all
reopen
source
EOS
go.
B
B
A
B
Eyes,
I,
don't
remember,
maybe
it
was
used.
No,
no,
it's
really
something
we
piece
be
published,
but
me
at
this
time
people
were
not
as
aware
and
I
was
just
surprised.
I
see
people,
even
at
blog.
When
didn't
know
it.
We're
like
impressed.
Oh,
we
can
restore
snapshots
in
two
minutes.
Yes,
you
know
if
you
optimize
for
write,
append
only
things
and
sparse
files
you
get
to
that,
and
the
software
is
there
what.
B
A
B
Were
into
that
problem
earlier,
other
I
needed
to
snapshot
and
for
backup
and
recovery
purposes,
you
always
want
to
backup.
You
know
to
some
ways
right:
I,
don't
I
just
thought
we
could
put
it
out
and
some
people
more
people
know
about
it,
and
maybe
that
becomes
also
in
the
lists
of
developer.
Tooling.
More
higher-level.
You
know,
that's
the
sort
of
thing
I
think.
Maybe
if
you
wanted
to
be
interesting
to
have
a
more
reference
more
direct,
so.
A
B
A
And
did
the
workshops
there
I
saw
that
because
I
we're
for
liquid
apps
we're
planning
our
hackathon
around
that
same
time.
So
I
was
like
basically
spying
on
all
of
the
etherium
hackathons,
because
they
look
so
well
organized
and
so
many
workshops
and
ways
to
like
touch
those
develop
like
do
you
have
the
all
hundreds
of
developers
in
one
place
for
that
weekend
and
I
I
think
aetherium
in
consensus.
I
guess
who
organized
is
that
is
that
are
the
global
hackathons,
a
consensus
thing
I,
don't
know.
A
B
Saying
that
we
had
interesting
discussion,
but
you
know
it's
like
anything
like
if
people
don't
know
about
something,
they've
heard
a
thing
to
read
it
and
that's
their
opinion,
but
at
the
same
time
it's
swayed
away
very
fast.
When
you
get
into
some.
You
know
the
details.
I
had
a
very
interesting
discussion
of
it
was
vitalik
and
some
other
core
devs
there
about
how
proof
of
stake
might
just
recreate
exactly
what
O'neal's
main
net?
In
that
you
know.
What
are
the
mechanism
just
to
go
against?
B
Ask
sometime
they're,
saying:
oh,
no,
there's
gonna
be
slashing,
but
you
know
if
you're,
if
you're
going
to
set
up
where
you
have
a
you
know
a
group
of
people
who
are
controlling
a
chain,
they
don't
make
their
block
right.
It's
part
of
the
algo
they'll
make
their
block
and
that's
not
gonna
anyway.
So
we
had
some
discussions
there.
Why
am
I
saying
that
it's
not
I'm.
B
To
say
when
we
came
into
the
the
spaces,
it's
large
its
diverse
there's
a
lot
of
people
with
different
interests.
That's
you
know.
It's
a
large
community
and
I
have
to
say
that
it's
impressive
and
I
I
think
it's
has
a
stronger
network
effect
than
I
thought.
You
know
at
some
point,
I
said
like
best
tech
wins
and
all
that
I.
Don't
think.
A
Iii
think
one
of
the
biggest
differences
between
definitely
a
CEO
and
aetherium,
but
a
lot
of
other
blockchains
in
aetherium
is
aetherium.
There's
a
lot
of
people
in
the
ecosystem
that
made
a
few
money
at
some
point.
So
there's
a
lot
of
people
and
oh
yeah.
If
you
basically
made
all
of
your
well
a
lot
majority
of
your
wealth
off
of
this
ecosystem,
then
it
doesn't
matter
if
there's
something
better
out
there,
because
you're
gonna
be
loyal
to
how
you
is
the
thing
that
got
you
there
sometimes
and
it's
not
a
shot.
C
B
It's
true
it's
true
that
shows
also
and
there's
also
a
lot
of
you
know
price
talk
about.
You
know
a
CEOs
candidate.
We
never
wanted
to
have.
Let
me
have
a
a
rule
at
a
company.
We
don't
talk
about
the
price,
it's
not
the
goal,
it's
not
that
you
know
we
want
actually,
as
a
company,
we
want
to
build
a
product
people
value
for
what
it
does
and
how
it
helps
them
right.
That's
sort
of
a
what
drives
us
cuz!
We
want
it.
We
didn't
want
to
build
a
token.
B
We
didn't
want
to
build
such
things
because
we
want
to
have
some
solid
meat
and
I
think
we're
getting
in
there
right
we're
building
something
that
is
necessary
needed
want
to
desire.
You
know
helping
you,
so
that
puts
us
in
a
good
position,
but
it
does
have
an
influence
on
what
what
how
communities
move?
That's
for
sure
and
and
I,
perhaps
I'd
say
you
know,
EOS
didn't
make
a
big
swarm
of
people
rich.
A
hundred.
You
know,
X
they've
they've
had
a
good
timing.
B
They've
had
you
know
a
good
story,
they
were
the
first
first
mover
advantage.
You
know
the
the
system,
you
know
bootstrapped
money-wise,
an
echo
system,
people
are
excited
and
then-
and
then
you
know
it's
also
very
hard
I,
don't
know,
I
sometimes
use
a
comparison
when
I
switched
from
SVM
subversion
of
version
control
system
to
get
or
from
CVS
to
subversion.
It's
sort
of
the
same
thing,
but
it's
so
different.
It
uses
the
same
words
for
different
concepts
and
same
words,
but
all
your
assumptions
from
one
system
don't
transpose
immediately.
B
It's
very
difficult
for
a
mind
to
shift
from
one
chain
to
the
other.
So
it's
understandable
that
when
you
understand
a
chain
going
to
another
one,
it's
hard
to
knitted
Lee,
it's
very
hard,
I
think
it's
very
rewarding,
but
there's
an
there's,
a
big
cost.
So
it's
normal
people
get
into
a
chain
and
they
say
they
want
it's
like
like
it's
so
hard,
it's
easier
to
say:
look
they
have
this
flaw.
It
has
this
flaw.
You
know
I
diffuse.
B
We
want
to
bring
all
the
blockchain
data
and
because
we're
at
that
level
right,
we're
sort
of
agnostic
to
the
token
economics.
We
don't
really
care
about
that.
We
care
about
the
tech.
The
data
that
doesn't
lie
that
doesn't
you
know,
try
to
screw
you
whatever
and
we'll
have
the
best
data
for
all
these
chains
in
a
radical
way,
so
that
no
one
can
undercut
right,
because
when
you
have
all
the
data,
there's
nothing
left
right,
you
have
it
all,
and
then
you
have
good
systems
to
process
it.
A
We
talked
a
little
bit
about
what
your
business
model
with
diffused
was
like
prior
to
open
source,
and
actually
you
had
different
packages.
A
lot
of
people
were
probably
on
the
free
tier,
but
then
you
had
that
enterprise
tier
with
everything
being
open
sourced.
How
does
that
affect
your
business
model
and
how
does
diffuse
grow
into
this
giant
enterprise
that
you
envision
it
to
be
right?.
B
So
yeah
we
do
have
customers
and
all
these
tiers-
and
you
know
we're
still
figuring
out
I'm,
not
saying
we're
gonna
change
pricing
right
now,
but
it
was
still
figuring
out
the
best
way
for
such
machinery
such
services
to
be
priced.
You
know,
the
whole
ecosystem
is
learning
at
this
point.
I'm
thinking
about
the
inferiors
of
this
world
for
consensus,
you
know
figuring
out
the
model,
so
that's
still
ongoing.
B
Something
that's
very
helpful,
for
us
is
all
these
on-demand
networks,
people
running
private
chains,
private
networks,
test
nets
and
all
these
things
I
mean
this
is
something
people
appreciate
having
the
diffuse
platform
around
plus
and
manage
nodes,
and
all
that
so
that's
been
very
useful
for
successful
of
ours
and
the
other
thing
is
how
we
grow
the
business
from
there.
I
mean
it's:
it's
not
complicated.
Like
we've
analyzed
elastic
or
you
know,
Apache
spark
these
guys
have
taken
the
tickle
or
even
Hoshi
Corp,
another
nice
company
right
open-source
I,
don't
know
what
they
they
raise.
B
But
no
we
see
that
easily
and
especially
in
watching
space,
wherever
it's
always
mission-critical,
you
don't
have
a
blockchain
for
non
mission-critical.
Things
like
you'd
have
on
my
sequel
database,
or
you
know
elasticsearch
cluster,
sometimes
for
your
dolls
or
whatever.
You
know
bunch
of
books.
So
you
know
I,
think
it's
natural.
How
things
go
we
work
on
the
product
and
make
it
the
best
product
like
similar
to
the
approach
elastic
like
the
product
is
the
baby
in
the
rest
comes
about
and
we're
seeing
that
already
right.
B
First
of
all,
open
sourcing
opens
some
new
doors,
but
already
those
doors
even
before
open
sourcing,
we're
already
opening,
because
we
had
solid
support
because
we
had
a
real
company
behind
willing
to
put
you
know
razor
in
the
morning
at
night.
If
there's
any
issue
of
the
network
and
that's
still
valued
I
mean
that's
the
real
world,
so
we
see
fit
that
you
see
that
fit.
It's
not
so
it's.
A
A
So
it
would
be
like
a
consulting
thing
and
then
you
guys
are
still
always
even
with
the
open
source
you're
still
going
to
run
infrastructure
and
still
let
people
to
like
pay
you
in
fiat,
which
I
think
is
one
of
the
like
things
that
stands
out
with
diffuse
is
like
you
said
you
don't
you
never
went
to
token
route,
so
you
just
if
you're
working
with
an
enterprise
who
could
care
less
about
anyone's
token
aetherium
Bitcoin,
it
doesn't
matter
to
them.
We
do
accept
payments.
B
And
die
and
I
don't
remember,
it
will
cause
DT
or
something
like
that
and
the
freedom
folks.
So
when
you
accept
those
things
and
that
in
that
sense,
except
token
famous,
but
we
didn't
want
to
have
our
token
to
put
the
whole
platform,
you
know
at
a
systemic
risk
of
speculation
right,
that's
that's
the
decision
we
make.
So,
yes,
you
can
paint
via
and
and
and
tokens,
but
we
don't
want
to
roll
on
yet
another.
B
A
B
Services,
perhaps
you
know,
search
lifecycle,
talk
about
with
tracking
the
lifecycle.
There's
one
thing:
I
think
people
really
appreciate
with
platforms
coming
your
way
too.
It's
we
call
it
the
push
guarantee,
and
one
thing
there
is.
You
know
the
fact
that
people
when
they
submit
a
transaction
to
the
network-
and
we
see
that
our
theorem
is
crazy,
because
it's
12
seconds
so
we
have
it
yeah
yeah
a
full-blown
service
for
that
on
a
theory
about
audios
when
you
send
a
transaction.
B
What
happens
is
that
you
get
an
execution
from
the
note
you
hit,
but
we
call
that
a
speculative
execution-
it's
not
the
one,
that's
gonna
make
it
into
the
block.
It's
just
for
him.
For
that
note
to
say:
yes,
it
worked
I'm
gonna
propagate
that
transaction
and
at
some
point
the
block
producer
takes
it
and
execute
it.
B
Definitely
now
the
problem
is
that
the
users,
when
they
send
that
they
get
an
immediate
response,
but
they're
not
sure
if
it
made
it
to
a
block
if
it
made
it
over
a
fork
if
it
made
it
into
reversibility,
if
it's
final,
whatever
they
don't
know,
that's
super
scary.
Right
and
again,
they
need
to
have
all
that
code
on
their
own
run
on
their
own
code.
The
query
knows:
do
you
have
that
transaction?
Give
me
that
block.
B
Do
I,
see
my
transaction
that
block
where
there's
so
many
polling
going
on,
so
we
want
it
again
to
have
an
easy
way.
So
what
we
did
is
we
have
that
push
transaction
endpoint,
which
is
the
drop
in
replacement
for
no
Dioses.
You
know
push
transaction
when
you
submit
transactions,
it's
chained,
but
it
stops
there
and
it
listens
to
the
network
and
it
ditches
the
speculative
execution.
Unless
it
fails.
If
it
fails,
it
returns
right
away.
So
you
know
you'll
use
it
fails.
B
It
won't
work,
it's
not
even
gonna
propagate,
but
if
it
doesn't
fail,
then
we're
gonna
wait
for
that
one's
actually
propagate
to
a
block.
Producers,
see
the
execution
and
return
the
actual
execution
from
the
real
block.
So
then
you
can
give
that
to
your
user,
it's
more
certain
and
you
use
just
a
small
header
in
the
query
and
then
you
can
have
also
handoffs
one
handoffs
two
means
I
want
to
level
a
higher
level
of
guarantee
that
my
transaction
is
gonna.
Stick
you
know
hand
Goss
is
when
Forks
often
happen.
B
So
if
you
go
over
a
handoff
and
your
transaction
is
still
there,
then
you
have
greater
guarantees,
gonna
stick
and
then
you
can
have
also
irreversible.
In
that
case,
it's
gonna
wait
for
a
longer
time
right
like
two
minutes,
but
it's
gonna
return
to
you,
your
user.
Only
when
it's
irreversible,
so
you
get
that
guaranteed.
That
plus
in
a
streaming
transaction
lifecycle
allows
users
to
build
web.
B
You
eyes
with
you,
know
just
hooking
a
small
piece
of
code
and
it
reacts
to
what's
happening
on
the
chain
right
without
them
pulling
you
and
sauce
those
crazy
things.
A
simple
API
hooks
directly
to
a
web
widget,
so
your
users
feel
comfortable
right
on
a
theorem.
It's
crazy
because
you
set
a
transaction
most
of
the
time.
You
don't
see
it.
You
don't
know,
and
it's
there
for
12
15
30
seconds
and
then
a
million
dollars
isn't
really
telling
no
feedback
the
crushing.
A
B
Different,
you
know
the
way
aetherium
lays
out
his
state
is
more
opaque,
so
people
want
to
execute
code
and
be
one
might
be
working
on
some
of
these
things
right
to
have
an
abstraction
where
you
have
a
read
function.
I
think
they
have
that
in
their
proposal
somewhere,
but
their
state
is
pretty
opaque,
whereas
you
know,
as
I
said
earlier,
when
you're
querying
the
voters
table
you're
happy
to
see
voter,
you
have
the
user,
you
have.
B
You
know
the
balance
and
all
that,
it's
very
let's
say
transparent
and
readable,
so
the
state
system
we
built
for
EOS
is
not
applicable
to
two
to
aetherium.
Although
the
because
the
block
cam
is
so
fast
on
ESI
yo,
we
rarely
need
to
query
the
mempool
because
it's
in
a
block
a
point
five
seconds
later,
most
of
the
time.
B
So
we
don't
have
that
knee
as
much,
whereas
on
aetherium
we
needed
to
have
a
dedicated
mental
service
so
that
we
can
track
the
flows
from
the
mempool
to
blocks
and
all
that,
so
different
chains
have
different
requirements
that
the
same
underlying
screaming
machinery
same
underlying.
You
know
you
know
a
model
ization
of
the
chain
and
radical
first
principle
design
right.
No,
so.
A
B
A
Suspect
every
I
speculate
everything
you're
talking
about
specifications,
I'm
talking
about
speculation
but
same
spec.
So
where
are
they
going
here?
There's
been
a
lot
of
like
discussion
with,
so
they
wrote
off
the
eos
bc
grants
and
then
there
is
this
thing
that
said
in
the
fine
print
like
if
you've
been
a
block
producer
in
the
last
six
months
like
we
can't
fund
you
and
not
that
call
some
controversy,
bloomer
came
out
and
said
spin
up
a
second
company,
so
you
guys
spun
up
a
second
company
a
long
time
ago.
B
The
structure
and
the
not
structure,
let's
say
let's
say
it,
probably-
has
slowed
down
the
potential
collaboration
in
the
past
years
right.
The
fact
that
we
were
a
block
producer
that
might
be
changing
a
little
bit.
You
know
they
settle
to
the
sec.
Maybe
this
is
a
small
change.
I
don't
know
if
they
they
said
that
quote.
Maybe
I
don't
know
you
know,
does
that
mean
anything
you
could
speculate,
but
maybe,
since
the
sec
settled
that
means
some
new.
I
know.
A
It's
it's
weird
and,
and
some
of
the
people
that
work
there
will
even
admit
that
it's
weird,
like
I
think
as
part
of
their
job
interviews,
they'll
ask
the
people
if
they've
ever
been
a
block
producer.
It's
like.
What's
it
matter
like
all
of
the
top
talent
in
this
ecosystem.
As
far
as
tech,
talent
has
all
been
a
block
producer,
there's
very
very
few
like
nathan,
james
is
one
of
the
last
holdouts
and
scatter
has
a
block
producer.
So
III
don't
know
many
people
that
wouldn't
fit
that.
B
My
perspective,
I
get
that
feeling
right.
So
how
will
that
be
concrete,
we'll
see,
but
I
get
that
feeling
that
there's
a
certain
more
opening
and
just
by
the
fact
that
they're
now
taking
a
look,
you
know
how
that
diffuse
thing
that
came
about
now:
it's
use
it
to
their
defense.
Maybe
what
was
where
defense?
Maybe
we
should
have
presented
that
more
clearly
to
block
one
employees
sooner,
but
then
again.
A
A
B
The
best
way
we
have
a
telegram
support
channel.
If
you
go
to
telegram
or
website
diffuser
you
and
then
the
telegram,
but
there's
a
lot
of
help
most
of
the
engineers
on
there.
You
can
help
out
now.
The
fact
that
we're
open
sourcing
will
change
dynamics
that
guess
a
little
bit
cuz
we're
gonna,
be
you
know,
helping
out
also
the
core
developers,
people
who
want
to
contribute
the
code
not
just
use
the
api.
B
So
that's
gonna,
open
a
new
facet
and
we're
going
to
be
super
happy
to
have
you
guys
interested
in
having
an
awesome
solution
for
both
history,
state,
reliable
querying
and
I.
Didn't
mention
one
thing:
I'm
just
skew
Annette
in,
but
does
not
all
the
system
to
be
highly
available.
Oh
yeah,
let
me
push
that
in
though
go.
B
I
think
it's
the
distinguisher
from
what
we
have
out
there.
We've
designed
all
the
systems
to
be
highly
available.
That
means
you'll
run
to
three
nodes,
potentially
different
regions,
seeing
different
force,
but
all
these
things
get
cumulated
in
what
we
call
the
platform,
the
diffuse
platform
view
of
the
world,
and
so
it's
built,
for
you
know,
enterprise.
B
So
right
now
we
can
reprocess
the
chain
the
whole
chain
in
a
matter
of
one
or
two
hours
and
that's
all
components
we
built
with
this
in
mind
so
that
we
can
always,
if
anything,
breaks,
come
back
on
our
feet
in
a
matter
of
seconds
minutes,
rather
than
hours
of
days,
we've
seen
some
competitor
in
aetherium.
Let
down
for
three
weeks
saying
what
we
need
to
reprocess
from
Genesis.
B
Never
we
want
to
have
you
know
this
situation
come
about
us
we've
designed
things
so
that
this
does
never
need
to
happen,
not
that
we
never
process
from
Genesis.
If
there's
a
need,
you
will
do
it
literally,
but
not
in
the
midst
of
you
know,
operations
so,
and
so
we
for
a
customer
recently
process.
You
know
a
hundred
ten
million
blocks.
It
took
twenty
minutes
right
and
then
we
gave
them
so
some
of
the
interesting
customers
in
this
space.
You
could
say
so,
okay.
Why
was
I
saying
that?
B
If
you
want
to
laptop-
and
eventually
this
you
know,
you
can
run
on
it
on
larger
servers
and
they
you
could
deploy.
All
of
the
app
is
kubernetes
native,
somehow
G
RPC
protocol
in
written
in
go
it's
gonna,
be
an
amazing
ride
for
all
these.
We're
gonna
be
contributing
to
that,
because
it's
gonna
be
to
the
moon
for
all
chains.
It's
a
great
opportunity
to
learn
also
about
many
protocols.
Things
are
similar
dissimilar,
okay,
it's.
A
Exciting
man
so
I
I
know
if
Josh
was
on
JA.
Josh
was
on
here
a
little
over
a
year
ago,
less
than
a
year
ago,
I
don't
know
when
josh
is
on.
Here
we
talked
about
WPS,
so
it's
like
right
after
main
that
launch
and
I
know
he
would
know
how
to
close
out
the
show.
But
I
don't
know
if
you
know
how
we
closed
off.
This
show
whoo.