►
From YouTube: HTM Hackers' Hangout - May 6, 2016
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
A
D
E
A
Of
it,
I'm
just
I'm
gonna
dive
right
in
my
side
looks
fine.
So
hopefully
the
recording
is
is
fine.
Now,
if
you
continue
to
have
issues
wave,
your
hands
are
around
or
something
so
anyway.
Welcome
to
the
HTM
ackers
hangout
is
May.
Sixth
I
have
a
few
things,
probably
on
my
agenda
to
talk
about
and
then
we'll
just
open
up
the
floor
to
who
or
whatever
we
want
to
talk
about.
A
So
the
first
thing
I
wanted
to
mention
was
of
events
this
year
as
far
as
new
meta
sponsored
events
or
events
that
we
put
on
so,
as
you
probably
know,
to
hackathons
a
year,
we've
done
that
for
three
years
in
a
row,
so
I
think
we've
done
six
or
seven
paths
online
events
this
year
we're
taking
a
hiatus
I
was
we're
not
going
to
have
a
workshop
or
a
hackathon
this
year.
I
know
that
might
disappoint
some
of
you
and
I
apologize
for
that,
but
the
the
emphasis
is
changing
a
bit.
A
There
can
you
hear
me:
okay,
wow
I
can
I
can
alright,
so
I
was
seeing
right
we're
trying
to
sponsor
a
big
hackathon
in
the
Midwest
coming
up
soon,
which
would
get
us
in
front
of
a
lot
of
people
which
would
be
great
I.
Don't
know
how
that's
going
to
go!
Cuz
it's
expensive!
When
you
have
packet
bonds
with
like
thousands
of
people,
it
lost
tens
of
thousands
of
dollars
sponsor
that's
starting
out
like
that's.
That's
the
entry
point
to
sponsor
and
I.
A
Richard
is
an
event,
so
the
good
news
is,
we
probably
will
still
hold
an
event
every
other
year
or
so,
but
not
so
often
as
we
as
we
have
been
so
sorry
to
say
that,
but
I
still
want
to
get
the
community
together.
I'll
probably
be
having
more
smaller
meetups
trying
to
encourage
people
to
start
meetups
around
the
globe
of
this
coming
year
and.
B
D
A
That
sucks
I
can't
tell
it
all
on
my
end.
Maybe
the
video
catches
it
when
it
happens.
I'd
hope
so
all
right
well
I
was
trying
to
explain
that
we're
not
we're
not
having
a
workshop
or
act
on
this
year.
You
hear
that.
Okay,
all
right
good
and
if
you
want
to
complain
about
it
or
I'm,
happy
to
discuss
it
with
you
on
the
mailing
list
or
wherever
so
that's
sort
of
a
bummer.
A
But
it's
also
it
gives
us
opportunity
to
work
on
other
things
and
try
and
offload
the
work
and
logistics
it
takes
to
put
on
events
people
so,
but
we
can
still
try
and
take
advantage
of
the
exposure
we
would
get
in
a
bit
like
that
and
honestly
after
having
six
or
seven
hackathons,
you
know
we
get
the
same.
Typically
the
same
core
group
of
people,
which
is
it's
great
because
that
gets
the
community
sort
of.
A
E
A
A
Okay,
so
I'm
done
talking
about
events,
I
also
want
to
talk
about
just,
and
this
is
not
to
clear
the
air.
I
just
wanted
this
as
a
status
update,
so
which
we
want
to
get
a
in
the
anomaly
benchmark
or
NAB,
as
we
always
call
it.
We
want
to
get
a
presence
there
from
HTM
Java
as
well.
We
have
new
pic
running
the
evil
hold
temporal
pooling.
We
don't
have
one
for
temporal
memory,
yet
HTM
java,
if
you
remember,
has
the
temporal
memory,
algorithms
and
so
Louis.
A
Ok,
so
we
what
we
really
the
right
way
to
do.
This
I
think
is
to
at
some
point
get
a
temporal
memory
algorithm
inside
of
nab,
and
then
we
could
probably
just
copy
and
paste
the
model
parameters
or
the
network
parameters
whatever
from
from
that
implementation
over
into
to
set
up
the
HTM
gotta.
So
all
the
infrastructure
for
getting
that
NAB
detector,
recall
them
detectors
in
place-
is
done.
A
What
context
every
conversation
is
being
discussed
within
and
honestly,
I'm,
sick
and
tired
of
mailing
lists
of
this
mailman,
and
it's
just
I
mean
this
is
1970s.
80S
technology
were
using
for
communication
and
and
once
actually
talked
her
email
back
and
forth
with
with
Jeff
Atwood
and
his
team
about
this
discourse,
and
they
gave
us
an
instance
for
free
because
we're
open
source
and
I
had
a
chance
to
try
it
out
over
the
past
week,
I've
had
it
for
a
few
months.
A
A
F
F
A
So
I
plan
on
setting
him
up
so
he's
only
watching
the
new
pic
theory
discussion.
The
new
pic
theory
category
will
be
called
a
category
in
this
and
hopefully
we'll
get
them
set
up
next
week
and
he'll
be
happy
honestly,
I
hope
I
convince
them
to
use
the
interface,
because
the
interface
is
really
nice.
Well,.
E
Can
it
be
set
up
so
that
you
have
a
privileged
submitter
that
can
you
know,
take
a
comments
on
a
topic
and
just
send
an
email
to
the
mail
look
mailer
in
the
background,
just
your
archive
them
and
then
that
way,
and
you
know
no
one
could
have
access
to
it
except
you
know
this
whatever
privileged
process,
and
then
you
know,
if
anything
ever
happens,
you
could
always
just
refer
to
that.
I.
E
A
No,
no
I,
don't
I,
don't
want
to
do
that.
Well,
I
probably
could
but
I
don't
want.
Do
I.
Think
well,
I'll,
probably
just
freeze
the
an
archive,
the
mailing
lists
as
we
as
we
transfer
over.
So
those
will
stick
around
they'll
be
archived
forever
and
then
I
can
get
everybody's
email
off
of
those
and
I
will
send
them
all
invitations
to
join
the
new
forum
and
they
will
have
to
go
to
the
forum
and
sign
up
and
then
I'll
probably
set
default.
So
they
get
emails.
E
A
I'll
confirm
with
them
what
their
I
mean
in
the
admin
UI,
they
have
an
interface
for
backup.
So
maybe
it
may
be
my
responsibility
to
set
up
a
cups
to
start
happening,
but
right
now
it
is
all
running
on
their
servers
for
up
for
free
for
us,
but
we
should
all
take
a
note
right
now
and
make
sure
that
I
get
backup
set
up
properly.
A
All
right
anything
else
about
forum
mailing
list
of
stuff,
okay,
so
a
couple
things
about
well,
one
thing
about
my
priorities:
I've
been
thinking
a
lot
about
what
my
priorities
in
my
role
is
for
the
community
for
right
now.
What's
the
best
thing
for
me
to
do
right
now
for
new
big
HTM
job
of
HTM
period
and
the
community
I'm
going
to
be
backing
away
from
sample
applications
for
a
while.
A
If
you
know
me
for
the
past
four
years,
I'm
always
working
on
at
least
one
sample
application
of
HTM,
putting
together
some
tutorial
or
something
all
right
now,
I
have
two
or
three
and
I'm,
I'm
going
to
put
those
all
on
on
hold
and
and
focus
on,
project
management
and
processes.
That's
something
that
I've
been
neglecting
for
a
while,
and
what
I'm
talking
about
github
issues
is
retracting
organization,
making
sure
we
have
a
decent
roadmap
and
a
pull
request,
monitoring
that
sort
of
stuff
I.
A
That
was
a
long
time
ago,
so,
okay,
so
process
stuff
is
what
I
was
talking
about:
getting
a
project
management
and
process
stuff
for
for
new
pic
and
associated
HTM
projects,
keeping
better
track
of.
What's
going
on
right
now,
community
making
sure
the
issues
you're
getting
addressed,
making
sure
that
pull
requests
are
being
addressed.
All
that
stuff
I
haven't
put
as
much
effort
as
I
needed
to
over
the
past,
probably
three
to
six
months
on
that.
A
So
I'm
going
to
put
back
effort
in
that
direction
this
year
and
the
other
thing
is
I'm
going
to
put
a
lot
of
emphasis
this
year
on
public
education,
about
HTM,
you
tube
by
making
these
videos
the
HTML
stuff
and
and
there's
going
to
be
more
YouTube
videos
I
have
a
schedule
of
videos
we're
going
to
be
put
putting
on
our
on
our
YouTube
channel
over
the
next
three
to
six
months
includes
not
only
HTM
school,
but
a
couple
of
other
types
of
series.
I'm
going
to
be
interviewing.
A
You
David
Ray
next
week
a
week
from
today
and
we're
still
having
these
HTM
hacker
hangouts
once
a
month.
So
these
are
all
sort
of
parts
of
our
our
YouTube
offering
on
our
YouTube
channel
and
the
so
I'm
going
to
continue
working
on
those
because
they're
getting
a
good
response
that
people
are
asking
for
for
more
and
they
take
a
considerable
amount
of
time.
Honestly,
not
just
editing
and
recording,
but
setting
up
all
the
script
and
me
doing
the
research
and
figure
out
making
sure
that
I'm
not
saying
things
that
are
stupid.
A
E
A
D
Just
so
you
know
everyone
here
who
is
like
Alex
of
bham
Markel's.
This
is
scott,
my
own
and
then
awesome.
A
So
it's
it's
funny.
I've
told
you
guys
this,
but
so
I
went
through
and
did
a
thumbnail
treatment
on
our
YouTube
videos,
because
I
went
through
like
some
of
the
YouTube
creators
series
and
they
said
they
told
me
the
right
way
to
do.
Thumbnails
and
I
changed
all
of
our
hacker
hang
out
thumbnails
two
things.
I
thought
were
more
interesting
and
one
of
them
just
has
a
picture
of
Austin's
bearded
face
and
it
says
Austin
is
mad,
and
people
really
watched
that
a
lot
to
figure
out
why
this
doping,
bearded
guy,
was
mad.
G
A
G
A
Just
so
here's
another
reason
Austin's
mad.
By
the
way
we
keep
a
list
of
fittings
on
our
internal
wiki
of
things
that
Austin
disapproves
of
and
and
one
of
the
highest
ones,
is
that
it
on
my
in
my
that
I
know
of
right
now
is
the
term
sparsity,
which
is
really
refers
to
density.
How
many
bits
are
on
in
an
array
win.
A
A
D
Is
so
I
don't
know,
Scott
might
be
interested
this
conversation
as
well
move
so
he's
bringing
me
up
because
I
just
changed
the
template
c++
temporal
memory
to
be
like
faster.
It
uses
a
different
way.
It
walks
the
segments
and
everything.
So
it's
a
lot
faster
and
we
haven't
made
that
change
to
the
Python
version
yet
and
that's
why
David
is
bringing
this
up.
So
our
plan
is
to
eventually
do
that,
but
I'm
not
currently.
This
is
not
what
I'm
doing
today.
E
Well,
I
mean
not
much
I
know
you
guys
have
your
priorities
and
and
I
probably
have
been
heard
from
enough
this
week.
So
as
I
just
want
to
see
if
I
could
get
Scott
to
grid.
E
E
D
Just
the
question
of
ooh,
because
right
now
right
now,
our
new
pic
or
temporal
memory,
the
c++
one
hits
the
random
number
generator
in
a
different
order
than
the
Python
one
and
we're
not
treating
it
as
like
the
priority.
One
thing
to
solve
right
now
and
that's
just
kind
of
miss
gunshot
of
today
Oh
basics
right.
E
Right
well,
you
know,
however,
it
plays
out
I'm
in
favor
of
it,
because
it'll
be
more
efficient
and,
like
I
said
before,
when
I
first
ported,
the
TEM
I
noticed
that
you
know
there
are
opportunities
to
do
something
like
this
myself
mom.
So
and
as
long
as
you
guys
vet
it,
then
you
know
I.
All
I
do
is
make
sure
that
the
test
produce
absolutely
the
same
output
except
you
know.
You
know
the
rng
things.
E
I
have
to
do
some
extra
stuff
that
I
can't
have
in
the
test
over
a
long
period,
but
I
can
play
around
with
them
to
make
sure
that
I'm
getting
the
same
output
and
in
those
places
where
the
rng
is
in
an
impact.
You
know,
I
do
have
the
same
outfit.
So
as
long
as
you
guys
get
the
Python
version,
you
know
as
long
as
it's
it's
it's
you
know
is
up
to
your
standards
standards
and
you
guys
can
get
it.
E
E
E
What
I'm
saying
is,
is
that
do
you
I
mean
you
had
the
Mersenne
algorithm
and
it's
supposed
to
be
the
same,
but
even
when
you
use
the
get
int
and
not
any
of
the
more
obscure
methods
in
the
different
languages
there
are,
there
are
floating
floating
bit
opportunities
for
differences
in
the
output
too.
I
noticed
that
even
even
when
I
could
get
it
to
work
exactly
like
the
Python
version.
There
are
points
where
you
know:
I
just
wasn't
normalizing
the
bits
and
you
know
I
get
started
to
get
variants.
So
it's
it's
just
weird.
E
You
can't
really
do
it
cross
unless
I
could
I,
don't
know,
get
a
dll
or
something
and
do
it
natively
and
just
use
yours
I,
don't
know.