►
Description
🙏 Thank you for watching! Hit 👍 and subscribe 🚩 to support this work
🌱Join the Community🌱
on Discord https://discord.gg/uM4ZWDjNfK
or say hello on Telegram https://t.me/tecommons
Join the conversation https://forum.tecommons.org/
Follow us on Twitter: https://twitter.com/tecmns
Learn more http://tecommons.org/
A
C
B
Yeah,
I
think,
there's
some
patterns
that
cross
over
between
disciplines.
I
know
there's
a
there's:
a
famous
book
for
software
engineering
called
design
patterns
and
I
think
it
was
it
all
those
those
patterns
come
from
like
civil
engineering
and
mechanical
engineering.
It's
the
same
kind
of
abstraction
concepts
to
like
architect,
code
as
it
is
to
actually
architect
machines
and
buildings
and
stuff.
B
Yeah,
it
could
be
merged.
I
mean,
I
think,
a
lot
of
the
other
code.
That's
in
there
now
is
kind
of
like
deprecated
I
mean
it
could
probably
just
be
deleted,
like
the
other,
because
I
merged
the
hatch
and
the
impact
hours
and
I
think
the
merged
the
mur.
The
combined
version
is
better.
B
B
Yeah,
so
are
you
presenting
that
next
week
in
the
community
call.
B
I
think
it's
good
timing,
because
I
think
all
the
contracts
are,
I
know
they've
been
saying
they're
having
some
big
bug
or
or
something
that
they're
trying
to
iron
out
in
the
contracts,
but
I
think
that's
gonna
be
resolved
soon.
I
think
everything's
going
to
be
ready,
like
it's
really
good
timing,
that
we
have
this
hatch
dashboard
with
the
voting
all
ready
to
go,
and
I
think
the
smart
contracts
are
all
ready
to
go
so
everything's
like
lined
up
in
parallel.
A
C
B
Yeah,
I
keep
it's
a
perpetual
struggle,
I
mean
I'm
always
thinking
I'm
getting
ahead.
Like
okay,
I've
carved
out
an
extra
three
hours
in
my
week.
You
know
by
handing
off
this
one
project,
and
then
I
think
I'm
gonna
have
so
much
available
time
and
then
something
else
comes
up
like
this
whole.
Get
this
whole
so
right
now
it's
the
office
hours
basis.
This
is
a
different
kind
of
lab,
it's
gonna
be
more
open-ended
and
anyone
can
just
bring
topics.
It's
going
to
be
about
the
git
coin.
B
C
Yeah,
but
I
I
as
I
I
I,
I
watched
a
recording,
an
earlier
recording
of
you
presenting
the
reinforcement
learning
agent
in
the
academic
session.
I
recall
and
zargham
was
saying:
oh
yeah.
Well,
basically,
you
are,
you
are
doing
some
sort
of
an
attack,
an
attack
factor
or
an
attacking
agent,
and
I
think
well
yeah.
C
This
is
a
really
nice
aspect
of
of
doing
open
science,
because
maybe
these
guys
haven't
thought
thought
about.
You
know
getting
getting
your
research
also
on
the
other
half
of
the
the
spectrum
and
the
bad
guys.
How
are
they
because
you?
You
are
now
trying
to
to
build
up
your
defense
in
in
maybe
tweaking
around
with
equaderic
funding
algorithms,
but
on
the
other
hand
there
also
also
all
sorts
of
other
guys
trying
to
attack
this
equaderic
funding
algorithm
in
the
first
place
or
playing
around
with
it.
B
Yeah,
I'm
sure
you're
right
I
mean
that's.
That's
the
whole
point
I
mean
I'm
sure,
there's
lots
of
people
looking
at
it
thinking
about
how
they
could
what
are
how
they
can
game
it,
and
it
is
this
kind
of
two-edged
sword
like
maybe
playing
around
with
adversarial
models
in
public
in
open,
because
it
allows
us
and
the
git
coin
team
to
explore
what
are
the
vulnerabilities
but
you're
right.
It's
also
publishing
all
the
vulnerabilities
right
away
right,
so
I
you
know
it's.
B
It
deserves
some
consideration,
and
I've
been
thinking
about
this
for
some
time
like
in
the
work
that
that
I'm
really
interested
in
and
mark
you've
shown
a
lot
of
interest.
Is
this
intersection
of
ai
and
token
engineering?
It's
almost
like
just
too
powerful,
and
I
think
it
deserves
serious,
like
guard
rails
from
sort
of.
D
D
B
Yeah,
I
think
so,
I
think
the
work
that
sepnam
is
doing
with
spinning
up
the
working
group
on
omega.
It's
the
omega
working
group,
which
is
sort
of
the
ethics
and
assumptions
that
we
have
about
token
engineering,
and
I
think
this
kind
of
work
is
really
important
when
we
start
to
yield
these
very
powerful
tools.
B
So
mark
was
just
saying,
like
it's
nice,
to
have
this
open
science
venue
that
we
can
explore
these
sort
of
adversarial
models,
but
we
have
to
also
be
aware
that
well,
if
there
are,
you
know
people
out
there
in
the
world
who
are
looking
for
how
to
collude
and
they
start
watching
these
research
track
videos
and
they
see
the
repo
with
reinforcement
learning.
Well,
it's
a
two
edged
sword
right
because
we
can
use
it
to
explore
the
vulnerabilities,
but
also
people
can
grab
it
and
use
it
as
well
to
actually
utilize
vulnerabilities.
B
So
it's
like
yeah,
it's
kind
of
the
pros
and
cons
of
the
open
research
aspect,
yeah
the
openness.
D
Yeah
yeah,
I
think
there
are
many
pros
and
cons
in
this
openness.
Actually
I
haven't
actually
that's
a
good
fit
hint.
I
haven't
looked
at
it
from
from
that
angle,
I
I
mean
when
it
comes
to
these
tools,
reinforcement,
learning
and
agent,
I
just
wonder
if
people
who
would
be
who
are
able
to
utilize
reinforcement
learning
to
attack,
for
example,
bitcoin
grants,
if
if
they
would
really
need
our
sessions,
you
know
what
I
mean,
of
course,
the
other
way
around.
D
Okay,
I
mean
this,
this
okay,
I'm
building
this
and
and
then
I
will
walk
away
with
this
profit.
I
I'm
not
sure
if
we
can
provide
that
kind
of
clear
instructions
in
in
the
sessions.
You
know
what
I
mean:
yeah.
B
D
Yeah
I
mean
for
the
good
and
the
bad
of
it,
I'm
I
would,
of
course
it
would
be
awesome
to
have
such
a
clearly
shaped
end
result
in
this
research
group,
however,
not
sure
if
we're
really
able
to
to
build
this,
but
yeah
and
in
general,
I
think
it's
a
valuable
thought.
C
D
C
Okay
but
then,
then
again
you
are
really
you
know
hacking
into
the
algorithm
itself
and
I'm
not
sure
if
that's
that's
the
whole
idea
of
the
the
get
going
philosophy.
C
You
know-
maybe
maybe
sean
you,
I
think
you
you,
you
mentioned
some
sort
of
picture
of
the.
I
think
the
get
the
the
get
going
goals
or
something
like
that
or
like
the
united
nations
goals,.
D
B
So
some
of
my
work
there
was
directly
working
with
the
un
data
and
the
sdgs
and
all
the
metrics
and
indicators
that
they
have
for
those,
and
so
I'm
familiar
with
the
sort
of
architecture
that
they
have
of
having
like
it's
like
hierarchical
right,
there's,
seven
goals
and
then
you
for
each
goal.
You
break
it
up
into
metrics
that
we
can
track
and
then
for
each
metric.
You
break
that
up
into
indicators
that
are
the
actual
sort
of
data
feeds
that
we're
tracking
and
we
can
almost
do
them.
B
You
know
it's
not
like
all
the
indicators.
We
need
them
to
go
up
some
of
them.
We
need
to
go
up
some.
We
need
to
go
down,
so
it
kind
of
becomes
like
a
machine
learning
problem,
but
I
like
this
architecture
and
I've,
I
brought
it
up
in
one
of
the
blog
posts
on
the
tec
for
tracking.
I
forget
what
the
exact
context
was.
I
think
it
was
like.
People
were
discussing,
source
cred
and
I
was
thinking
about
okay,
what
are
our
highest
level
missions
and
goals?
B
You
know
someone
found
that
really
interesting
and
expanded
on
it,
and
some
people
started
kind
of
fleshing
out
that
thread,
but
I
think
that
works,
because
it's
a
really
good
model
and
then,
when
I
saw
those
the
high
level
missions
sort
of
mission
objectives
of
git
coin,
I
realized
yep,
that's
the
same
kind
of
architecture
like
for
each
of
those.
You
could
flesh
them
out
into
metrics
and
then
have
actual
data
indicators,
so
that
could
be
a
whole
another
project.
I
guess.
D
Oh
yeah
yeah,
and
I
think
actually
this
this
also
rang
a
bell
when,
when
I
looked
at
it
in
your
presentation
on
tuesday-
since
I
think
I
mean
I
was
here
and
there
I
was
trying
to
collect
okay,
what
my
observations,
what
happened
over
the
rounds
of
bitcoin
grants?
We
had
with
token
engineering
and
ge
academy-
and
I
haven't
put
it
down,
but
maybe
I
would
be
able
to
I've,
tried
to
remember
the
okay.
D
What's
what
are
the
patterns
in
activities
like
okay,
somebody,
somebody
owning
an
an
owner
of
a
grant,
is
contributing
to
our
grant.
Okay,
I
take
a
look
at
it.
Okay,
his
contribution,
of
course,
is
a
nice
reminder
to
do
it.
Vice
versa.
D
Bang
you
have
this
pairwise
funding,
then
I'll
send
out
emails
to
our
network
and
ask
for
can
you
please
contribute
and
of
course
those
people
will
click
a
link
with
the
collection
and
then
all
perhaps
don't
change
the
values
in
in
the
in
the
cart
and
just
click
donate
and
it's
fine.
But
then
there
you
have
a
pattern.
It
looks
like
okay
is
this
collusion
and
then
then
I
got
this
overview
on.
Okay.
What
are
your
actual
goals
defined
in
okay?
D
It's
about
giving
it's
about
supporting
tiny
projects
that
might
not
be
able
to
attract.
D
You
see
funding
it's
about
this
ecosystem
purpose
of
bitcoin
grants
and
there
I
found
perhaps
it's
much
more
relevant
to
define
to
break
down
these
objectives,
to
the
git
coin,
grants
area
and
and
case
and
then
c
for
metrics,
because
I
mean,
of
course
you
might
want
to
let's
say
we
start
project
starts
very
early,
just
three
developers,
no
community
over
two
three
four
five
bitcoin
grants
round:
they
develop
their
community,
they
professionalize
their
communications
and
suddenly
are
able
to
activate
100
200
people
to
donate,
and
since
they
are
lazy,
they
just
click
and
we
have
very
similar
donation
patterns.
D
Now
I
wouldn't
say
this
is
collusion.
I
would
rather
say
this
is
not
the
area
of
this
is
not
a
pattern
you
want
to
support,
since
those
projects
are
already
grown
up
and
they
should
maybe
acquire
different
sources
of
funding.
But
this
is
then
is
not
collusion
prevention.
It's
rather
defining
your
your
key
focus
area
of
projects
and
still
it's
about
metrics,
and
still
it's
about
optimizing.
D
Quadratic
funding,
perhaps
not
sure
if
this
could
be
the
algorithm
it's
rather
about
implementing,
then
fostering
particular
or
supporting
particular
patterns
and
reducing
the
the
effectiveness
of
other
patterns,
but
I
feel
well
taking
a
step
back.
I
feel
this
is
a
actually
an
area
where
I
see
optimization
for
bitcoin
grants.
C
D
B
So
what
exactly
is
the
matthew
effect
is
this?
Is
this
sort
of
like
a
community
has
outgrown
the?
B
They
don't
require
the
git
coin
grants
or
or
they
have,
the
ability
kind
of
how
angela
was
saying
they've
professionalized
their
community,
so
they
have
the
ability
to
recruit,
200
people
to
vote
on
their
grant
and
it's
sort
of
not
aligned
with
the
purpose
of
the
git
coin,
grants
which
is
to
fund
these
smaller
projects
that
can't
quite
acquire
funding
elsewhere.
Is
this
sort
of
the
matthew
factor.
B
D
Yeah,
if
you
already
see
for
one
example,
would
be
you
already,
since
the
contributions
that
has
been
already
collected
in
in
the
grants
period
are
visible
to
everyone
and
the
number,
and
also
yet
the
size
of
contributions.
You
can,
you
can
quickly
see
okay.
Obviously,
here
this
grant
has
a
lot
of
contributors,
it
seems
to
be
very
valuable
or
prominent
or
popular
and
and
those
according
to
the
matthew
effect.
Those
grants
are
able
to
then
collect
even
more
contribution.
B
So
I
was
thinking
as
you
were,
describing
that
angela,
how
that
might
unfold
and
sort
of
one
organization?
Well,
I
guess
the
matthew
effect
how
one
organization
could
change
where
it
might
fit.
The
purpose
of
bitcoin
grants
really
well,
but
sort
of
as
it
grows
and
as
it
scales
and
based
on
the
effectiveness
of
the
get
coin
grants.
I
guess
that's
what
we
would
want
so
one
way
that
I
see
that
is
having
git
coin
plug
into
the
greater
ecosystem.
B
B
One
and
initialize
its,
like
you,
said
communications
and
organization
enough
to
then
get
go
for
the
git
coin
platform,
where
that's
like
phase
number
two
and
then
it
can
go
through
get
coin,
to
get
funded
and
build
some
momentum
and
get
more
visibility
in
the
ecosystem.
B
But
then
there
might
be
this
sort
of
phase
three
like
this.
Maybe
we
should
have
a
well-defined
path:
post
bitcoin
grants
when
a
project
is
self-sustaining
and
it
wan-
and
maybe
this
is
like
a
vc
world
or
maybe
this
is
a
token
launch
like
a
fair
launch
or
something
along
these
lines,
but
maybe
having
a
well-defined
like
exit
paths,
post,
successful,
bitcoin
grants
or
maybe
projects,
don't
necessarily
need
funding.
B
Maybe
they
will
be
self-sustaining
after
that,
or
maybe
they
will
launch
like
a
kind
of
token
launch
that
produces
the
economics
for
the
system
and
and
maybe
that
I
guess
we
live
in
the
crypto
economy
right.
So
maybe
that
would
make
the
most
sense,
but
I
guess
just
having
some
sort
of
guidelines
of
like
well
that's
hard
to
define
because
we're
in
this
crypto
world.
It's
like
there's,
infinite
creativity
and
possibilities.
So.
D
I
think
this
is
this
kind
of
an
emerging
behavior,
also
based
on
the
range
of
opportunities
you
have
in
git
congress,
and
I
just
think
that
how
is
this
relevant
for
the
algorithm
itself?
Because,
as
you
say,
this
is
rather
information
around
it
and
a
definition
of
what?
What
is
the
focus
of
bitcoin
grants
and
not
that
much
can't
be
steered
that
much
by
the
algorithm.
B
B
D
B
Yeah
yeah
there's,
I
think,
there's
so
much.
I
mean
this
is
so
early
right.
This
is
really
a
new
experiment
and
it's
so
good
to
have
this
research
group
just
to
get
so
much
exposure
and
so
many
different
people
with
different
backgrounds,
eyeballs
reading
the
paper
and
participating
on
the
platform
and
yeah
there's,
you
know,
there's
so
much
happening
out
there,
there's
so
much
so
many
things
to
sort
of
divide
our
attention
in
a
good
way
like
in
crypto.
B
Generally,
it's
cool
all
the
new
projects
coming
out,
but
it's
nice
to
have
this
time
to
like
bring
the
focus
back
and
just
stare
at
this
one
thing
that
is
very
powerful
and
a
very
useful
solution
for
the
whole
ecosystem
and
just
stare
at
it.
Long
enough
that
we
start
to
see
like
the
additional
details
like
I
was
just
going
through
on
git
co
on
the
platform
the
other
day
they
have
these
quests
sort
of
learning
quests,
and
I
think
it's
so
cool.
I
think
they
did
a
really
really
good
job.
B
I
mean
I
was
having
so
much
fun,
just
learning
about
solidity
contracts
and
defy
mechanisms
and
roll
ups
and
all
this
kind
of
stuff,
so
yeah,
it's
easy
to
overlook
like
when
there's
brand
new
things
coming
out
every
day.
It's
easy
to
kind
of
overlook
some
things,
but
the
git
coin
platform
alone
has
so
many
great
aspects
to
it.
I
mean
it
can
be
used
as
like
a
social
platform.
B
It's
got
the
whole
bounty
bounty
system
and
you
know
I
I'm
me
personally.
I'm
I've
I've
kind
of
touched
all
these
points
a
little
bit,
but
I
think
it'll
take
a
long
time
to
get
really
to
ingrain
the
intuition
around
what
this
enables
and
what
this
unlocks
and
then
really
utilizing
it
like
using
the
bounty
systems
nico
and
I
were
working
on
the
spirit
foundation
project
and
we
had
a
bug
that
we
were
trying
to
solve
and
we
thought
probably
someone
out
there
could
solve
this
really
fast.
B
A
A
Cool
so
they're
not
planning
any
like
side
chains
right
now.
Are
they
like
any
side
chain
deployments
for
the
contracts.
D
I
mean
I
haven't
better
if
you
want
to
have
the
full
picture
better
oscar
bucky,
but
I
haven't
seen
anything
in
that
regards
and
well
considering.
Okay,
they
want
to
enable
all
kinds
of
erc20
token
contributions.
I
wouldn't
assume
that
they
take
side
chains
rather
layer,
two
optimization
gas
cost
reductions.
D
Yeah,
but
I
mean
get
congruent,
is
also
a
good
example
for
this.
This
okay,
you
might
be
able
to
build
a
really
a
great
range
of
mechanisms
around
your
ecosystem
purpose
and
then
okay
as
a
token
engineer
how
to
orchestrate
and
how
to
make
sure
this
is
all
going
into
the
direction
you
have
in
mind
as
your
purpose.
D
So
this
is
in
a
way.
It's
a
it's
a
great
exercise.
I
wonder
how
this
is
relevant,
also
at
bozone,
since
you
I
mean
there
is
a
strong
purpose,
and
now
you
can
go
with
all
kinds
of
measures
how
to
how
to
steer
your
ecosystem
in
this
direction.
Using
your
tokens
and
it's
it's
really,
it's
becoming
complex,
very,
very
quickly
and
then
yeah
how
to
handle
this
complexity.
A
Thinking
about
this
last
night
so
because
there's
like
a
boson
doubt
right,
just
like
there's
like
a
you
know,
there's
like
an
ocean
dao-
and
I
was
thinking
like
whoa
like
it-
would
be
really
cool
to
use
like
quadratic
funding.
For
for
for
this
dow
I
mean
yeah,
but
but
then
that's
like
it
brings
up
like
this
question
with
like
other
chains
and
stuff,
and
then
I
started
thinking
like
okay
like
is
there
a
standardized
way
of
doing
this
like
or
is
it
all
through
get
coin?
So
yeah.
C
Yeah
well
actually,
when
I
I
enrolled
in
this
in
this
bitcoin
course
I
was
thinking
about
leveraging
the
stuff.
I
would
be
learning
to
my
business
for
local
government,
because
at
local
government
there
are
really
struggling
with,
let's
say,
tendering
processes
and
keeping
out
the
usual
suspects
may
sound
familiar
because
local
government
is
being
haunted
by
you
know
big
companies
doing
their
I.t
and
I'm
trying
to
leverage
the
git
coin
model
in
order
to
well,
let's
say,
build
up
a
new
ecosystem
around
their
I.t
suppliers
in
the
git
coin
way.
C
So
maybe
you
know
funding
as
angela
said,
maybe
phase
three
kind
of
communities
or
projects
or
businesses
that
have
some
components
that
together
would
be
would
be
nice
to
have
in
a
local
government
ecosystem
or
supporting
their
I.t.
For
for
that
matter,
so
I'm
trying
to
leverage
the
gitcoin
model
to
local
government
in
a
way
all
all
these
kinds
of
of
issues
we
are.
You
are
mentioning
right
now,
like
basically
the
quader
quadratic
funding
thing
is
a
resource
allocation
problem,
but
it
starts
with
the
purpose
of
the
ecosystem.
C
So
my
my
endeavor,
I
would
say
my
quest-
is
to
see
how
can
we
leverage
this
this?
These
thoughts,
which
became
from
which
came
from
some
sort
of
a
social
good
kind
of
movement
into
the
real
world
of
doing
business
and
government,
is
a
as
some
sort
of
a
nice
playground.
I
think,
because
they
need
to
well.
Basically,
they
are.
They
have
a
purpose
of
of,
let's
say,
social
impact
or
social,
good
or
public
good.
C
C
Yeah,
the
contributors
are,
are
government
grants
or
no,
the
the
matchers
are
the
government
grants
and
the
contributors
are
yeah,
basically
the
local
government
itself,
because
they
have
funding
but
also
other
other
people
that
are
seeing
basically
they're
seeing
stuff
happening
in
their
community
in
their
in
their
local
community,
and
they
want
to
fix
it
like.
Okay,
like
a
mini
bill,
gates,
kind
of
type,
which
is
saying
okay.
I
want
this
neighborhood
fixed
up
and
I
will
donate
a
certain
amount
of
money
to
the
government
to
fix
that.
B
So
it's
like
a
quadratic
democracy
kind
of
it's,
which
is
the
way
I
see
get
coin
kind
anyways,
but
it's
like
the
government
might
have
some.
B
You
know
the
the
people
who
rep
represent
representatives
right.
They
have
the
representatives
and
maybe
each
representative
can
vote
on
the
allocation.
B
B
Overhead
and
yeah
I
mean,
I
wonder
if
the
git
coin
platform
could
be
cloned
and
and
maintained,
like
a
fork
maintained
by
a
sort
of
government
body
that
uses
this
system
and
and
it's
the
people
the
inhabitants
can
authenticate
with
web3
I
mean
it's
so.
C
B
C
Actually,
this
was
one
of
this
was
one
of
the
elements
I
won
the
hackathon
from
odyssey
with,
because
I
I
I
pitched
this
model
to
the
to
the
jury
and
that
they
found
it
far
very
far
fetched
but
interesting.
B
D
B
Yeah,
if
you
tell
them
hey,
all
we
got
to
do
is
fork
this.
It's
already.
It's
ready
to
go
yeah
well,
usually
there's
the
response
where
they
can't
believe
it
right.
They
don't
think
they
can.
B
B
C
Well,
except
this
is
exactly
my
business
model,
what
I'm
trying
to
to
achieve
here,
because
they're
they're,
really
you
know,
and
that
they
like
these
kinds
of
innovations
and
they
need
consultants
to
help
them
with
this,
and
I
am
such
a
consultant.
B
C
Yeah,
well,
I
will
send
you
over
a
presentation,
excellent
matter.
B
B
Yeah,
it's
a
great
innovation.
It's
a
great
way
that
that
trent
laid
that
out
yeah.
C
B
Yeah
ocean
protocol
I
mean
trent
and
the
whole
ocean
protocol
they
laid
they
made
all
the
maps
for
how
to
make
this.
I
mean
they
put
out
so
much
like.
If
you
look
at
their
medium
page,
they
have
like
over
100
articles
right
and
they've
just
been
doing
this
for
years,
so
there's
so
much
content
there
to
mine
through
and
reuse
for
making
a
dow.
With
this
sustainability
model,
I
mean
there's
other
ways
you
could
make
a
dao,
but
ocean
protocol
has
fleshed
out
so
many
concepts
across
the
board.
B
Token
engineering
and
sort
of
like
fiscal
policy
and
community
growth
and
defy
mechanisms
like
it's
really
nice
to
have
that
foundation
to
to
build
on
top
of
with
boson,
because
it
maps
really
well
it's
a
very
similar
product.
It's
just
kind
of
expanding.
Instead
of
data
marketplaces,
it's
thing
marketplaces
and
they
they
have.
D
B
B
But
yeah
that
I
think
I
don't
know
if
nico
wants
to
comment
on
his
article,
but
it's
just
under
review
by
the
team
now,
so
I
think
it
should
come
out
pretty
soon.
B
A
B
A
B
B
C
C
So
sean
any
updates
on
your
reinforcement,
learning
agent
coding.
B
Well,
I
think
I'm
really
happy
to
have
that
very
clean
example
of
like
a
minimal,
minimal
working
custom
environment.
So
I
think
we
can
take
that
and
basically
splice
it
into
wherever
we
need
reinforcement
learning,
so
we
can
open
up
the
energy
web
simulator.
I
was
thinking,
but
you
know
what
I
did.
You
push
your
latest
code
because
I
it
seems
like
yeah.
I'm
just
wondering.
Did
you
push
to
your
energy
web
branch?
There.
C
No,
I
haven't,
I
haven't,
pushed
any
any
of
the
the
reinforcement
learning
agent
stuff
there
yet
because
I
haven't
got
time
yet
to
to
fix
that.
B
So
I
was
thinking
we
could
just
book,
another
tec,
lab
and
and
now
that
we
have
this
clean,
stable
baselines.
Three
example:
we
could
probably
in
an
in
a
lab
session.
We
could
splice
that
in
and
well,
we
might
want
to
do
a
little
bit
of
we
might
have
to
sort
of
think
about
refining
the,
whether
it's
the
action
space
or
the
observation
space.
B
Yeah,
so
so
that's
what
I
wanted
to
do.
There's
that
piece
that
it's
like.
I
really
want
to
have
make
an
agent-
and
I
might
just
do
this
similar
to
what
I
did
with
this
idea
of
like
a
encapsulated
toy
example.
So
I'll
just
make
it
outside
of
the
simulator.
But
I
want
to
make
a
simple
machine
learning
system
that
just
reads
in
the
energy
data
and
makes
a
reasonable
forecast
and
then
yeah.
So
I
think
we
can
actually
implement
all
the
components
right.
So
we
can
have
the
optimizer.
We
can
make
an
optimizer
agent.
D
B
The
energy
data
and
makes
a
forecast,
and
then
that
forecast
goes
over
to
the
staker
agent,
which
is
can
actually
read
that
forecast
and
do
again
some
some.
Actually
that's
the
observation
space
right.
It
reads
the
optimization
data
and
well
it
we
need
to
think
about
what
it
needs
to
know
relatively
like
it
needs
to
have
a
consistent
idea
of
the
energy
out
of
the
energy
producers
and
the
energy
optimizers,
so
that
it
can
maintain
a
sort
of
memory
of
the
performance
of
them.
C
C
B
Of
thing
yeah,
so
I
think
on
the
week
like
we
can
pick
a
weekend,
maybe
and
do
put
a
few
hours
in
and
probably
we
could
get
everything
working
and
then
maybe
we
could
take
like
a
final
lab
session
to
present
it.
C
Yeah
sound
good
sounds
good
to
me
yeah.
I
I
just
noticed
that
yeah,
the
the
simulation
of
the
data
is
still
running.
So
we
have.
We
have
a
data
set
available
for
that
yeah.
B
C
No
spoken
key
yeah.
I
was
talking
to
the
energy
web
guys
that
they
really
see
some
potential,
but
they
need
a
real.
You
know
some
sort
of
a
concrete
use
case
for
that,
and
then
I
was
hooking
up
with
lior
from
data
dao
to
put
put
something
together,
but
yeah
resource
issues.
You
know
time,
I
know.
C
Found
my
I
found
my
slide
deck,
but
I
need
to
anonymize
something.
But
then,
if
I,
if
I've
done
that,
then
I
will,
I
will
put
it
in
the
the
tech
labs.
C
B
C
Yeah
yeah
yeah.
I
have
to
go
as
well
so
well,
thanks
sean
for
this
really
nice
well,
bonfire
talk
yeah,
and
so
we
need
to
yeah
need
to
get
some
work
done.
That's
true,.
B
Yeah,
so,
let's
book
a
weekend
mark,
maybe
the
following
weekend
after
this
one,
we
can
do
yep
that
actually
works
for
me.
If
we
can
do
it
on
a
saturday
and
maybe
see
if
lee
or
if
we
can
loop
in
lior
as
well.
C
D
B
A
A
Thanks
sean
yo
sean
yep,
I
was
thinking
like.
A
Bye
victor,
I
was
thinking
like
how
could
we
like
load
data
into
these
like
because
bro,
it's
literally
the
intersection
between
like
the
bots,
because
we
essentially
have
to
like
dude.
This
is
such
a
mind-blowing
thing.
We
have
to
like
replay
the
data
just
like
we
did
with
the
training
like
cryptobots.
You
know
what
I
mean,
because
it's
like
it's
like
going
through
time
like
like.
A
Is
like
yeah,
like
people
are
staking
and
they're
trying
to
figure
out
how
to
stay
better,
and
I
feel
like
the
only
way
that
we
can
do
this
is
with
amazon
data
or
with
ebay
data
where
we
can
like
put
in
like
review
after
review,
depending
on
what
time
step
it's
on
and
then
like.
The
reviews
like
are
like
what
impacts.
Like
the
actual
you
know,
the
actual
like
ratios
right
like
like.
A
Yeah,
exactly,
and
also
like,
like
also
by
doing
that,
like
trying
to
hack
the
system
almost
and
like,
like
you
know
what
I'm
saying
like
trying
to
find
those
like
glitches,
because
dude
like
after
you
sent
me
that
podcast,
like
I've,
just
been
thinking
like
okay,
like
our
mission
with
these
simulations,
is
to
break
boson.
Essentially,
you
know
what
I'm
saying
so
I
think
that
might
be
the
best
way.
B
B
They
they
can
go
to
work
right
and
make
cause.
As
of
now
like.
If
everything
goes
along
the
happy
path
forever,
then
like
boson's,
already
successful
right,
it's
gonna
be
a
great
token
launch.
It's
gonna,
be
a
great
product
like
everyone's
gonna,
be
happy
forever,
but
the
you
know-
and
they
don't
need
us
in
that
case,
but
the
reason
we're
there.
The
reason
they
hire
us
is
to
like
find
all
the
possibilities
where
it's
not
a
happy
case
and
like.
A
B
A
Yeah
yeah
yeah.
No,
I
definitely
feel
you
like,
I
guess
like
all
I
was
thinking,
was
like
okay,
like
with
this
like
energy
web
stuff,
which
is
like
that's
what
that's
what
you
guys
are
doing
right.
This
is
like
making
this
like
reinforcement,
learning,
yeah
agents
off
of
this,
like
energy
data,
like
from
where.
B
It's
actually
from
energy
web,
so
energy
web
is
a
way.
That's
like
the
infrastructure
for
energy
producers
to
like
put
their
data
on
blockchain
yeah.
It's
basically.
A
B
So
then
it's
these
optimizers
who
are
kind
of
predicting
the
future
and
saying
like
how
the
energy
should
be
routed
now
and
then
it's
the
stakers
jobs
to
stake
on
the
optimizers
and
give
them
stake
on
the
ones
that
are
more
accurate
and
more
effective
so
that
they
everyone
in
its
incentive
alignment,
so
that
the
optim,
the
optimizers
who
perform
the
best
are
getting
rewarded
and
the
stakers
who
stake
on
the
best
optimizers
are
also
getting
rewarded
so
that
we
get
this
like
really
high,
optimization
output.
That
kind
of
knows.
A
A
B
A
A
B
Let's
we
can,
we
can
hack
on
this
right
away.
I
mean
I've
always
wanted
to
open
up
the
amazon
api.
I
know
that
there's
a
lot
of
data
you
can
get
from
there.
I
don't
know
what
exactly,
but
I
think
we
have
to
do
that
like
we
should
do
that
now.
Basically,
because
it's
all
it's
it's
what
I
say
like
nothing
gets
left
behind
right.
If
we
always
just
tackle
these
things
as
soon
as
we
think
of
them,
then
we
have
this
ever-growing
collection
repertoire
of
like
components
like
if
we
need.
B
A
A
B
A
Yeah,
okay,
so
yeah.
I
was
hacking
on
token
spice
and
I
was
like
bro.
This
is
gonna
like
melt
anybody's,
mind
who
like
looks
at
this,
so
I
just
like
I
just
like
ripped
out
everything
and
I
just
put
like
agents
reward,
giving
agent
tokens,
commitment,
token
thing,
token
settings
constants
and
some
state
and
then
everything
else
I
just
left
in
core.
A
Cool,
so
I
just
took
token
spice
yeah
and
I
took
like
I
just
said:
okay,
we're
gonna
need
like
tokens
like
commitment
tokens
thing
tokens
like
I'm
just
making
those
like
as
easy
as
possible
for
like
peter
and
like
rio
and
stuff,
and
so
I
in
settings
I
put
like
constance
and
sim
stand
and
for
myself
and
anybody
else
who
wants
to
use
it
but
and
then
in
agents.
A
I
just
put
reward
giving
agent
so
like
my
idea,
is
that,
like
whenever
we
need
to
kind
of
like
wire
something
up,
we
can
just
go
in
the
sim
state,
and
so
yeah
like
I've,
been
like
refactoring
all
the
imports
and
stuff
and
anything
else
that
like
we
might
not
need.
That's
just
like
in
the
nitty
gritty
of
token
spice,
like
it's
just
left
in
this
core
folder.
So
all
we
have
to
worry
about.
Are
these
like
three
other
photos.
A
A
A
A
But
yeah
we
can
just
hack
on
like
regular
token
spice.
I
already
had
code
there.
It's
just
like.
I
got
to
this
point
where
I
was
like
man
like
like.
If
I
don't
get
like
organized
with
this
token
surprise
stuff
like
it's
just
gonna
eat
me,
yeah.
A
Yeah
yeah
cause
like
I'm,
I
kind
of
started
off
by
like
just
like
making
this
reward,
giving
agent
which
just
like
loops
iterates
through
the
stakers
and
and
gives
them
their
assigned
piece.
But
we.
D
B
Token
engineering
and
just
says:
how
do
we
have
a
system
that
we
can
upgrade
and
reuse,
and
I
like
this
idea
of
like
upgradable
apps,
so
it's
like
you
take
out
if
we
want
to
make
what
do
we
want
to
make?
We
want
to
make
like
this
amazon,
someone
who
stakes
on
products,
so
this
would
be
a
token
spice
app.
So
it's
it's
like
it
would
be
its
own
github
repository
with
like
an
agent,
and
you
know
this.
This
is
one
way
to
do
it.
I
don't
know
if
this
is
optimal.
A
B
Whole
team
working
on
this
in
parallel
to
us
actually
getting
done.
You
know
it's
like
we.
We
almost
need
to
it's
like
you
and
I
would
do
best
if
we
literally
just
hack
like
how
you
have
it
now
and
we
just
keep
hacking
on
it.
Hacking,
hacking,
hacking
and
getting
results
like,
but
then,
if
we
need
like
another
team
in
parallel
who's,
just
watching
all
this
and
and
reading
like
design
pattern
books
and
like
thinking
about
how
to
optimally
enable
us
right,
almost
like
a
devops
team,
just.
B
A
Dude
I'm
down
man
I
was
like
just
telling
like
I
was
talking
to
my
parents
about
it
yesterday,
I'm
like
yeah,
like
nobody
knows
it.
Therefore,
it
like
like
it's
like
so
rare
for
token
engineers
right
now,
but
at
the
same
time
like
by
us
teaching
it
to
people
like
we're,
generating
value
and
people
like
will,
like
literally
hire
us
just
because
we're
generating
value.
A
You
know
it's
like
it's
like
principle
of
reciprocity,
right,
yeah
yeah,
so
like
the
person
who,
like
teaches
the
complex
subject,
is
not
like
somehow
like
devaluing
themselves,
they're
actually
valuing
themselves
even
more
because,
like
the
teacher
is
like
almost
like,
almost
like
the
most
valuable
in
the
subject.
Well,.
B
Created
and
he
taught
people
and
then
those
people
taught
people,
and
then
those
people
taught
people,
and
then
they
made
a
course
out
of
it
like
he
got
to
define
the
paradigms
of
how
to
look
at
cad
cat
and
how
to
approach
it
and
how
to
use
it
because
he
created
it.
Obviously,
and
but
it's
the
educational
piece
that
I'm
really
talking
about
like
you
know,
you
can
look
on
youtube
and
literally
the
first
year
of
cad
cad
2018,
there's
like
dozens
of
talks
of
zargham
just
going
around
and
giving
talks
right
about
cad
cad.
B
Behind
it
and
why
it's
useful
and
that's
what
it
takes
like
now,
cad
cat
is
famous
like
if
everyone,
every
token
engineer
knows
cad
cat,
but
not
because
it's
the
best
piece
of
software
in
the
world.
It's
because
of
the
momentum
behind
like
the
education
and
the
awareness
and
the
community
and
and
all
that
stuff.
So
this
is
what
token
spice
could
use.
B
B
C
B
Actually
import
that
into
tokens,
so
you
can
go
like
import
boson
protocol
into
token
spice
and
use
the
so
so
if
we
make
like
this
aws
data
consumer
that
gets
nice
clean,
aws
data,
then
we
can
just.
I
can
build
that
out
in
the
bozon
tokenomics
repository
and
then
in
here
in
in
spicy
burrito.
We
can
just
import
boson
protocol
tokenomics
and
then
like
from
bozon
protocol,
import,
aws,
fetcher
or
whatever
yeah.
A
Yeah
yeah
yeah.
I
I
like,
I
totally
agree
but
like
when
we
no,
I
guess
we
just
have
to
send
the
state
yeah
so
like.
Let's
just
look
at
some
asians
yeah
yeah,
I
guess
you're
right
dude
wait.
Would
we
have
to
have
like
the
base
agent
like?
A
Would
we
just
like
how
would
that
work?
If
we
made
a
module
without
the
base
agent
or.
B
B
So
the
more
the
more
disconnected
components
we
have,
the
better
like.
We
want
everything
to
run
by
itself
like
any
sort
of
utility
like
this.
Like
grabbing
data
processing
data
charting
data,
we
can
have
it
all
independent
and
then,
when
we're
wiring
up
agents,
they
can
just
easily
grab.
They
have
basically
have
access
to
that
functionality
that
we've
that
we've
made.