Latest / Chaos Agents / Can You Build Anything in a Week? GPUs, Code Gen, and the End of Engineers - With Harper Reed

Transcript

Can You Build Anything in a Week? GPUs, Code Gen, and the End of Engineers - With Harper Reed · Chaos Agents

• 0:00
  Just
• 0:00
  like,
• 0:01
  Claude,
• 0:01
  take
• 0:01
  another
• 0:01
  pass.
• 0:02
  And
• 0:02
  the
• 0:02
  cloud's
• 0:03
  like,
• 0:03
  beep,
• 0:03
  boop,
• 0:03
  boop,
• 0:03
  boop,
• 0:03
  beep.
• 0:04
  And
• 0:04
  then
• 0:04
  you're
• 0:04
  like,
• 0:04
  did
• 0:04
  I
• 0:04
  get
• 0:05
  it?
• 0:05
  And
• 0:05
  you're
• 0:05
  like,
• 0:05
  eh,
• 0:05
  it's
• 0:05
  not
• 0:05
  quite
• 0:06
  it
• 0:06
  take
• 0:06
  another
• 0:06
  pass
• 0:07
  and
• 0:07
  it's
• 0:07
  like,
• 0:07
  boop,
• 0:07
  beep,
• 0:07
  boop,
• 0:08
  beep.
• 0:08
  And
• 0:08
  it's
• 0:08
  like,
• 0:08
  that
• 0:08
  is
• 0:09
  like
• 0:09
  the
• 0:09
  iteration
• 0:09
  process.
• 0:10
  Like
• 0:10
  we
• 0:10
  are
• 0:10
  the
• 0:11
  judge,
• 0:11
  we
• 0:11
  are
• 0:12
  no
• 0:12
  longer
• 0:12
  the
• 0:12
  doer.
• 0:13
  And
• 0:13
  I
• 0:13
  think
• 0:13
  that
• 0:14
  is
• 0:14
  a
• 0:14
  very
• 0:14
  interesting
• 0:15
  kind
• 0:15
  of
• 0:15
  place
• 0:15
  of
• 0:16
  like
• 0:16
  when
• 0:16
  you
• 0:16
  are
• 0:16
  the
• 0:16
  one
• 0:17
  who's
• 0:17
  expected
• 0:17
  to
• 0:18
  be
• 0:18
  the
• 0:18
  arbiter
• 0:18
  of
• 0:19
  taste,
• 0:20
  not
• 0:20
  necessarily
• 0:20
  arbiter
• 0:21
  of
• 0:21
  technical
• 0:22
  competence.
• 0:23
  Like
• 0:23
  it
• 0:24
  requires
• 0:24
  a
• 0:24
  different
• 0:24
  skill
• 0:25
  set.
• 0:35
  Welcome
• 0:35
  back
• 0:35
  to
• 0:36
  Chaos
• 0:36
  Agents,
• 0:37
  the
• 0:37
  podcast
• 0:37
  where
• 0:37
  we
• 0:38
  learn
• 0:38
  about
• 0:38
  AI
• 0:38
  so
• 0:38
  you
• 0:39
  don't
• 0:39
  have
• 0:39
  to.
• 0:39
  I'm
• 0:39
  Sarah
• 0:40
  and
• 0:41
  this
• 0:41
  is
• 0:41
  Becca.
• 0:42
  Becca,
• 0:42
  you
• 0:43
  just
• 0:43
  went
• 0:43
  to
• 0:43
  a
• 0:43
  conference.
• 0:44
  I
• 0:44
  did.
• 0:45
  The
• 0:45
  conference
• 0:45
  was
• 0:46
  NeurIPS.
• 0:47
  I
• 0:47
  can't
• 0:47
  remember
• 0:47
  exactly
• 0:48
  what
• 0:48
  it
• 0:48
  stands
• 0:48
  for,
• 0:48
  but
• 0:49
  it's
• 0:49
  essentially
• 0:50
  like
• 0:50
  a
• 0:51
  very
• 0:52
  academic
• 0:52
  machine
• 0:53
  learning,
• 0:53
  data
• 0:54
  science
• 0:55
  sort
• 0:55
  of
• 0:55
  AI
• 0:55
  conference.
• 0:56
  It's
• 0:57
  the
• 0:57
  kind
• 0:57
  of
• 0:57
  thing
• 0:57
  that,
• 0:58
  you
• 0:58
  know,
• 0:58
  eight
• 0:58
  or
• 0:58
  10
• 0:59
  years
• 0:59
  ago
• 1:00
  I
• 1:00
  wouldn't
• 1:01
  have
• 1:01
  been
• 1:01
  going
• 1:01
  to
• 1:01
  and
• 1:02
  now
• 1:02
  it's
• 1:02
  like,
• 1:03
  obviously
• 1:03
  much
• 1:04
  bigger,
• 1:04
  but
• 1:04
  it
• 1:05
  still
• 1:05
  has
• 1:05
  a
• 1:05
  lot
• 1:06
  of
• 1:06
  that
• 1:06
  ethos,
• 1:06
  which
• 1:06
  I
• 1:07
  really
• 1:07
  like.
• 1:07
  It's
• 1:07
  not
• 1:07
  overrun
• 1:08
  by
• 1:08
  like
• 1:09
  startup
• 1:10
  bros
• 1:10
  and
• 1:10
  VCs
• 1:11
  yet.
• 1:12
  The
• 1:13
  people
• 1:13
  who
• 1:13
  have
• 1:13
  been
• 1:13
  going
• 1:13
  for
• 1:13
  a
• 1:13
  long
• 1:13
  time
• 1:14
  probably
• 1:14
  feel
• 1:14
  like
• 1:14
  it
• 1:15
  is,
• 1:15
  but
• 1:15
  from
• 1:15
  my
• 1:16
  perspective,
• 1:16
  it's
• 1:16
  not.
• 1:17
  It
• 1:17
  was
• 1:17
  awesome.
• 1:18
  And
• 1:19
  someone
• 1:19
  was
• 1:19
  handing
• 1:20
  out,
• 1:20
  someone
• 1:20
  made
• 1:21
  these
• 1:21
  beautiful
• 1:22
  like
• 1:23
  trading
• 1:24
  cards
• 1:25
  that
• 1:25
  were
• 1:25
  of
• 1:27
  famous
• 1:27
  machine
• 1:28
  learning
• 1:28
  and
• 1:29
  AI
• 1:29
  researchers.
• 1:30
  On
• 1:30
  each
• 1:30
  card
• 1:31
  I
• 1:31
  was
• 1:32
  like,
• 1:33
  this
• 1:33
  is
• 1:33
  the
• 1:33
  coolest
• 1:33
  thing
• 1:33
  I've
• 1:34
  ever
• 1:34
  owned.
• 1:36
  I
• 1:36
  love
• 1:36
  that.
• 1:36
  Did
• 1:36
  it
• 1:37
  have
• 1:37
  their
• 1:37
  powers
• 1:37
  and
• 1:38
  stuff
• 1:38
  on
• 1:38
  it?
• 1:38
  Yeah,
• 1:39
  it
• 1:39
  talks
• 1:39
  about
• 1:39
  the
• 1:40
  paper
• 1:40
  they're
• 1:40
  known
• 1:41
  for.
• 1:41
  And
• 1:41
  on
• 1:42
  our
• 1:42
  next
• 1:42
  episode,
• 1:42
  I'll
• 1:43
  bring
• 1:43
  them
• 1:43
  out
• 1:43
  and
• 1:43
  I'll
• 1:43
  show
• 1:43
  you.
• 1:44
  Great.
• 1:44
  Cute.
• 1:45
  I
• 1:45
  love
• 1:45
  that.
• 1:45
  Loved
• 1:46
  it.
• 1:46
  What'd
• 1:46
  you
• 1:46
  learn
• 1:47
  about
• 1:47
  it?
• 1:47
  Sounds
• 1:47
  like
• 1:47
  the
• 1:47
  people
• 1:48
  were
• 1:48
  chill
• 1:48
  and
• 1:49
  very
• 1:49
  smart.
• 1:49
  I
• 1:50
  don't
• 1:50
  know.
• 1:50
  Yeah,
• 1:50
  chill.
• 1:53
  Chill
• 1:53
  and
• 1:54
  brilliant.
• 1:56
  So
• 1:56
  like,
• 1:56
  it's
• 1:57
  a
• 1:57
  week
• 1:57
  long
• 1:57
  thing.
• 1:58
  I
• 1:58
  was
• 1:58
  there
• 1:58
  for
• 1:58
  a
• 1:58
  few
• 1:58
  days.
• 1:59
  It's
• 1:59
  kind
• 1:59
  of
• 1:59
  split
• 1:59
  up
• 2:00
  into
• 2:00
  like
• 2:00
  two
• 2:01
  parts.
• 2:02
  There
• 2:02
  are
• 2:03
  workshops
• 2:03
  going
• 2:04
  on
• 2:04
  and
• 2:04
  talks
• 2:05
  going
• 2:05
  on
• 2:05
  throughout
• 2:06
  the
• 2:06
  day.
• 2:07
  One
• 2:07
  of
• 2:07
  the
• 2:07
  things
• 2:07
  I
• 2:08
  learned
• 2:08
  is
• 2:08
  like,
• 2:08
  the
• 2:08
  workshops
• 2:09
  are
• 2:09
  where
• 2:10
  like
• 2:10
  the
• 2:10
  really
• 2:10
  interesting
• 2:11
  stuff
• 2:11
  happens.
• 2:12
  That's
• 2:12
  where
• 2:12
  it
• 2:13
  tends
• 2:13
  to
• 2:13
  be
• 2:14
  folks
• 2:15
  who
• 2:15
  are
• 2:16
  bridging
• 2:16
  the
• 2:16
  gap
• 2:17
  between
• 2:17
  the
• 2:17
  really
• 2:18
  interesting
• 2:18
  academic
• 2:19
  research
• 2:19
  and
• 2:19
  discoveries
• 2:20
  and
• 2:21
  real
• 2:21
  world
• 2:22
  applications
• 2:22
  and
• 2:23
  how
• 2:23
  you
• 2:23
  put
• 2:23
  those
• 2:23
  two
• 2:23
  things
• 2:24
  together.
• 2:26
  And
• 2:26
  then
• 2:27
  there's
• 2:28
  also
• 2:28
  this
• 2:30
  giant
• 2:30
  room
• 2:31
  and
• 2:33
  rows
• 2:33
  and
• 2:34
  rows
• 2:35
  of
• 2:36
  like
• 2:37
  bulletin
• 2:38
  board.
• 2:39
  Okay.
• 2:39
  With
• 2:41
  academic
• 2:41
  researchers
• 2:42
  at
• 2:42
  each
• 2:43
  One
• 2:43
  with
• 2:43
  their
• 2:44
  paper
• 2:44
  that
• 2:45
  they
• 2:45
  had
• 2:45
  submitted
• 2:46
  to
• 2:46
  Neurips
• 2:46
  that
• 2:47
  got
• 2:47
  accepted.
• 2:47
  It's
• 2:48
  organized
• 2:48
  by
• 2:49
  category,
• 2:50
  so
• 2:50
  there's
• 2:50
  like,
• 2:51
  these
• 2:51
  flags
• 2:51
  hanging.
• 2:52
  One
• 2:52
  will
• 2:52
  be
• 2:52
  like,
• 2:53
  you
• 2:53
  know,
• 2:53
  model
• 2:54
  training,
• 2:54
  efficiency.
• 2:55
  One
• 2:55
  will
• 2:56
  be
• 2:56
  like.
• 2:56
  Like,
• 2:57
  ethics.
• 2:58
  Like,
• 2:59
  that
• 2:59
  was
• 2:59
  my
• 2:59
  favorite
• 3:00
  part.
• 3:00
  It
• 3:00
  also
• 3:01
  melted
• 3:01
  my
• 3:01
  brain.
• 3:01
  Yeah,
• 3:02
  I
• 3:02
  bet.
• 3:02
  It
• 3:02
  was
• 3:02
  like
• 3:02
  an
• 3:03
  adult
• 3:03
  science
• 3:03
  fair.
• 3:04
  Yeah.
• 3:04
  Which
• 3:04
  is
• 3:05
  the
• 3:05
  coolest
• 3:05
  thing.
• 3:06
  And
• 3:07
  I
• 3:07
  was
• 3:07
  pleasantly
• 3:08
  surprised
• 3:08
  that,
• 3:08
  like,
• 3:10
  most
• 3:10
  of
• 3:11
  the
• 3:12
  researchers
• 3:13
  I
• 3:13
  talked
• 3:13
  to
• 3:14
  about
• 3:14
  their
• 3:15
  research
• 3:16
  were
• 3:17
  actually
• 3:17
  really
• 3:17
  good
• 3:18
  at
• 3:19
  breaking
• 3:19
  it
• 3:19
  down
• 3:19
  in
• 3:20
  a
• 3:20
  way.
• 3:20
  That,
• 3:20
  like,
• 3:20
  explaining
• 3:20
  it
• 3:21
  in,
• 3:21
  like,
• 3:21
  a
• 3:21
  way
• 3:21
  that's
• 3:22
  approachable
• 3:22
  and
• 3:22
  accessible.
• 3:23
  Yes.
• 3:23
  That's
• 3:24
  really
• 3:24
  helpful
• 3:24
  because
• 3:24
  sometimes
• 3:25
  those
• 3:25
  things
• 3:25
  tend
• 3:25
  to
• 3:25
  be
• 3:25
  really
• 3:26
  esoteric.
• 3:27
  Yes.
• 3:27
  So
• 3:28
  as
• 3:28
  you
• 3:28
  can
• 3:28
  imagine,
• 3:30
  you
• 3:30
  know,
• 3:30
  you
• 3:30
  can't
• 3:30
  try
• 3:31
  and
• 3:31
  understand
• 3:31
  all
• 3:31
  this
• 3:31
  stuff
• 3:32
  en
• 3:32
  masse.
• 3:33
  So
• 3:33
  my
• 3:33
  approach
• 3:34
  was
• 3:34
  to
• 3:35
  sort
• 3:35
  of
• 3:35
  pick
• 3:35
  one
• 3:36
  and
• 3:36
  go
• 3:37
  really
• 3:37
  deep.
• 3:37
  And
• 3:37
  one
• 3:38
  of
• 3:38
  the
• 3:38
  reasons
• 3:38
  why
• 3:38
  I
• 3:38
  went
• 3:38
  is
• 3:39
  we
• 3:39
  had
• 3:39
  a
• 3:39
  bunch
• 3:40
  of
• 3:40
  our
• 3:41
  portfolio
• 3:41
  companies
• 3:41
  at
• 3:42
  Factorial
• 3:42
  Capital
• 3:43
  who
• 3:44
  were
• 3:45
  participating
• 3:45
  in
• 3:46
  Neurips
• 3:46
  in
• 3:47
  some
• 3:47
  way,
• 3:47
  and
• 3:47
  a
• 3:48
  few
• 3:48
  who
• 3:48
  had
• 3:49
  submitted
• 3:49
  their
• 3:49
  own
• 3:49
  research,
• 3:50
  their
• 3:50
  own
• 3:50
  papers
• 3:52
  that
• 3:52
  were
• 3:52
  accepted
• 3:52
  and
• 3:53
  were
• 3:53
  kind
• 3:53
  of
• 3:53
  presenting
• 3:53
  on
• 3:54
  those
• 3:54
  papers.
• 3:55
  I
• 3:55
  was
• 3:55
  like,
• 3:55
  I'm
• 3:55
  gonna
• 3:55
  go
• 3:55
  through
• 3:56
  all
• 3:56
  of
• 3:56
  these.
• 3:56
  And
• 3:57
  then
• 3:57
  I
• 3:57
  picked
• 3:57
  one,
• 3:57
  and
• 3:57
  I
• 3:58
  was
• 3:58
  like,
• 3:58
  maybe
• 3:59
  just
• 3:59
  one.
• 3:59
  Maybe
• 3:59
  just
• 4:00
  one.
• 4:00
  It's
• 4:00
  gonna
• 4:00
  be
• 4:00
  11
• 4:01
  this
• 4:01
  week.
• 4:01
  Yeah.
• 4:02
  And
• 4:02
  what
• 4:02
  made
• 4:02
  you
• 4:02
  pick
• 4:02
  this
• 4:03
  one?
• 4:04
  So
• 4:05
  the
• 4:05
  company's
• 4:05
  called
• 4:05
  Pluralis
• 4:06
  Research,
• 4:07
  and
• 4:09
  they
• 4:09
  had
• 4:09
  three
• 4:09
  different
• 4:10
  papers
• 4:10
  that
• 4:10
  were
• 4:11
  accepted
• 4:11
  to
• 4:11
  Neurips
• 4:12
  this
• 4:12
  year.
• 4:12
  And
• 4:12
  so
• 4:12
  I
• 4:13
  picked
• 4:13
  the
• 4:13
  first
• 4:13
  one.
• 4:14
  Honestly,
• 4:14
  I
• 4:14
  just
• 4:15
  started
• 4:15
  going,
• 4:15
  and
• 4:15
  it
• 4:16
  happened
• 4:16
  to
• 4:16
  be
• 4:16
  fascinating.
• 4:18
  So
• 4:19
  the
• 4:20
  problem
• 4:21
  that
• 4:21
  they
• 4:21
  were
• 4:21
  trying
• 4:22
  to
• 4:22
  sort
• 4:22
  of
• 4:23
  solve
• 4:23
  or
• 4:24
  tackle
• 4:24
  in
• 4:24
  this
• 4:24
  paper
• 4:26
  is
• 4:27
  can
• 4:28
  we
• 4:28
  train
• 4:29
  large
• 4:29
  AI
• 4:30
  models
• 4:32
  across
• 4:33
  cheap,
• 4:34
  slow
• 4:34
  Internet
• 4:35
  connections
• 4:36
  instead
• 4:37
  of
• 4:37
  requiring
• 4:38
  massive,
• 4:39
  expensive
• 4:39
  data
• 4:39
  centers?
• 4:40
  Wow.
• 4:41
  Oh,
• 4:41
  goodness.
• 4:42
  Yeah,
• 4:42
  yeah,
• 4:42
  yeah.
• 4:43
  That's
• 4:43
  huge.
• 4:44
  Yes.
• 4:45
  For
• 4:45
  so
• 4:45
  many
• 4:45
  reasons.
• 4:46
  Yes.
• 4:47
  I
• 4:47
  just
• 4:47
  wanted
• 4:47
  to
• 4:47
  do
• 4:47
  a
• 4:48
  really
• 4:48
  quick,
• 4:48
  like,
• 4:48
  primer
• 4:50
  on
• 4:51
  GPUs,
• 4:52
  because
• 4:52
  I
• 4:52
  feel
• 4:52
  like
• 4:53
  people
• 4:53
  who
• 4:54
  are
• 4:54
  in
• 4:54
  tech
• 4:55
  or
• 4:55
  tech
• 4:55
  adjacent
• 4:56
  probably
• 4:56
  hear
• 4:56
  GPUs
• 4:57
  all
• 4:57
  the
• 4:57
  time.
• 4:57
  All
• 4:58
  the
• 4:58
  time.
• 4:58
  And,
• 4:59
  like,
• 4:59
  if
• 4:59
  you
• 4:59
  ask
• 4:59
  people,
• 5:00
  like,
• 5:00
  so
• 5:00
  what
• 5:00
  is
• 5:00
  a
• 5:00
  gpu
• 5:01
  and
• 5:01
  what
• 5:01
  does
• 5:01
  it
• 5:01
  do?
• 5:01
  Like,
• 5:02
  people
• 5:02
  are
• 5:02
  like,
• 5:02
  it's
• 5:03
  like
• 5:03
  a
• 5:03
  chip.
• 5:04
  Yeah.
• 5:04
  So
• 5:05
  I
• 5:05
  hear
• 5:05
  they're
• 5:05
  hard
• 5:06
  to
• 5:06
  find.
• 5:06
  Yeah.
• 5:09
  TLDR
• 5:10
  on
• 5:10
  GPUs.
• 5:11
  GPU
• 5:11
  stands
• 5:12
  for
• 5:12
  Graphical
• 5:12
  Processor
• 5:13
  Unit.
• 5:14
  It
• 5:14
  is
• 5:14
  basically
• 5:15
  a
• 5:16
  computer
• 5:16
  chip
• 5:16
  that
• 5:17
  was
• 5:17
  initially
• 5:17
  designed
• 5:18
  for
• 5:19
  graphics
• 5:19
  and
• 5:20
  video
• 5:21
  things
• 5:22
  that
• 5:22
  require
• 5:22
  a
• 5:22
  lot
• 5:23
  of
• 5:23
  compute
• 5:23
  power.
• 5:24
  What
• 5:24
  we
• 5:24
  call,
• 5:25
  like,
• 5:26
  Halo,
• 5:27
  basically.
• 5:28
  GPUs
• 5:28
  were
• 5:28
  invented
• 5:29
  for
• 5:29
  Halo,
• 5:30
  famously.
• 5:32
  What
• 5:32
  that
• 5:33
  really
• 5:33
  means
• 5:33
  is,
• 5:33
  like,
• 5:33
  they
• 5:34
  can
• 5:34
  do
• 5:35
  millions,
• 5:36
  if
• 5:36
  not
• 5:36
  billions
• 5:37
  of
• 5:37
  math
• 5:38
  equations
• 5:39
  at
• 5:39
  once,
• 5:39
  of
• 5:40
  calculations
• 5:40
  at
• 5:41
  the
• 5:41
  same
• 5:41
  Time
• 5:42
  versus
• 5:42
  a
• 5:42
  CPU
• 5:44
  does
• 5:44
  things
• 5:44
  more
• 5:44
  sequentially.
• 5:45
  Yeah,
• 5:46
  yeah,
• 5:46
  yeah,
• 5:46
  yeah.
• 5:46
  And
• 5:47
  so
• 5:47
  it's
• 5:47
  like
• 5:47
  parallel
• 5:48
  processing,
• 5:48
  but
• 5:48
  much
• 5:48
  less.
• 5:49
  Exactly,
• 5:50
  yeah.
• 5:50
  And
• 5:50
  so
• 5:52
  what
• 5:52
  ended
• 5:53
  up
• 5:53
  happening
• 5:54
  is
• 5:54
  that
• 5:55
  happens
• 5:55
  to
• 5:56
  be
• 5:57
  exactly
• 5:57
  what
• 5:57
  you
• 5:57
  need
• 5:58
  to
• 5:59
  make
• 5:59
  modern
• 5:59
  AI
• 6:00
  systems
• 6:01
  like
• 6:01
  chatbots
• 6:01
  or
• 6:02
  image
• 6:02
  generators
• 6:03
  work.
• 6:04
  So
• 6:05
  GPUs
• 6:06
  are
• 6:06
  used
• 6:07
  to
• 6:07
  train
• 6:07
  AI
• 6:08
  models,
• 6:08
  run
• 6:09
  AI
• 6:09
  applications,
• 6:10
  do
• 6:10
  things
• 6:10
  with
• 6:11
  computer
• 6:11
  vision
• 6:11
  like
• 6:12
  self
• 6:12
  driving
• 6:13
  cars
• 6:13
  or
• 6:13
  facial
• 6:14
  recognition.
• 6:15
  And
• 6:16
  I
• 6:16
  guess
• 6:17
  the
• 6:17
  reason
• 6:17
  I
• 6:18
  bring
• 6:18
  that
• 6:18
  up
• 6:18
  is
• 6:19
  if
• 6:19
  Pluralis
• 6:20
  is
• 6:20
  trying
• 6:20
  to
• 6:20
  solve
• 6:20
  this
• 6:21
  problem
• 6:21
  of
• 6:22
  training
• 6:22
  a
• 6:22
  large
• 6:23
  AI
• 6:23
  model
• 6:23
  on
• 6:24
  smaller
• 6:25
  numbers
• 6:26
  of
• 6:26
  GPUs
• 6:27
  that
• 6:28
  are
• 6:28
  lower
• 6:28
  end
• 6:29
  and
• 6:29
  over
• 6:30
  low
• 6:30
  Internet
• 6:30
  bandwidth,
• 6:31
  you're
• 6:31
  like,
• 6:31
  okay,
• 6:31
  why
• 6:32
  do
• 6:32
  we
• 6:32
  care
• 6:32
  about
• 6:32
  that?
• 6:33
  I
• 6:34
  would
• 6:34
  imagine
• 6:34
  that
• 6:34
  would
• 6:34
  be
• 6:34
  a
• 6:34
  question.
• 6:35
  Yeah,
• 6:35
  yeah,
• 6:35
  yeah.
• 6:36
  Do
• 6:36
  you
• 6:36
  want
• 6:36
  to
• 6:36
  guess?
• 6:38
  Well,
• 6:38
  the
• 6:38
  first
• 6:39
  thing
• 6:39
  I
• 6:39
  thought
• 6:39
  of
• 6:39
  is
• 6:39
  like,
• 6:39
  saving
• 6:40
  the
• 6:40
  environment,
• 6:41
  I
• 6:41
  imagine.
• 6:41
  And
• 6:42
  also
• 6:42
  like,
• 6:42
  there's
• 6:42
  a
• 6:43
  lot
• 6:43
  of
• 6:43
  people
• 6:43
  that
• 6:43
  don't
• 6:43
  have
• 6:43
  access
• 6:44
  to
• 6:44
  GPUs.
• 6:45
  Yes,
• 6:46
  yeah,
• 6:46
  yeah,
• 6:46
  yeah.
• 6:47
  They
• 6:47
  are
• 6:48
  expensive.
• 6:49
  Super
• 6:50
  expensive.
• 6:50
  And
• 6:50
  really
• 6:51
  hard
• 6:51
  to
• 6:51
  get.
• 6:51
  Like
• 6:52
  a
• 6:52
  good
• 6:52
  GPU,
• 6:55
  like
• 6:56
  what's
• 6:56
  called
• 6:56
  an
• 6:56
  H100
• 6:57
  that
• 6:57
  Nvidia
• 6:58
  makes,
• 6:59
  it
• 6:59
  costs
• 6:59
  like
• 7:00
  25
• 7:01
  to
• 7:01
  45k
• 7:02
  per
• 7:02
  unit.
• 7:02
  Yeah.
• 7:03
  Not
• 7:03
  just
• 7:03
  one.
• 7:04
  And
• 7:04
  you
• 7:04
  need
• 7:05
  clusters
• 7:06
  of
• 7:06
  them,
• 7:06
  which
• 7:07
  can
• 7:07
  easily
• 7:07
  cost
• 7:08
  to
• 7:08
  train
• 7:08
  a
• 7:08
  model,
• 7:09
  which
• 7:09
  can
• 7:09
  easily
• 7:10
  cost
• 7:10
  millions
• 7:10
  of
• 7:10
  dollars.
• 7:11
  And
• 7:12
  so
• 7:12
  you're
• 7:13
  like,
• 7:13
  great,
• 7:13
  I
• 7:13
  just
• 7:13
  need
• 7:13
  a
• 7:13
  lot
• 7:13
  of
• 7:14
  money.
• 7:14
  But
• 7:14
  even
• 7:14
  if
• 7:14
  you
• 7:15
  have
• 7:15
  a
• 7:15
  ton
• 7:15
  of
• 7:15
  money,
• 7:16
  it's
• 7:16
  as
• 7:17
  you
• 7:17
  know,
• 7:17
  still
• 7:17
  hard
• 7:17
  to
• 7:18
  get.
• 7:18
  Because
• 7:18
  guess
• 7:19
  what,
• 7:19
  like
• 7:19
  Google,
• 7:20
  Meta,
• 7:21
  Microsoft,
• 7:21
  everyone
• 7:22
  is
• 7:22
  trying
• 7:23
  to
• 7:23
  scoop
• 7:24
  up
• 7:24
  as
• 7:24
  many
• 7:24
  GPUs
• 7:24
  as
• 7:25
  they
• 7:25
  can,
• 7:25
  as
• 7:25
  fast
• 7:25
  as
• 7:25
  they
• 7:25
  can.
• 7:26
  Yeah.
• 7:26
  So
• 7:26
  there's
• 7:26
  a
• 7:27
  shortage.
• 7:27
  Yeah.
• 7:28
  And
• 7:28
  this
• 7:28
  is
• 7:28
  one
• 7:28
  of
• 7:28
  those
• 7:28
  things
• 7:28
  where
• 7:29
  like
• 7:29
  the
• 7:29
  laws
• 7:29
  of
• 7:29
  physics,
• 7:30
  it
• 7:30
  turns
• 7:30
  out
• 7:30
  like,
• 7:31
  like
• 7:31
  when
• 7:31
  you
• 7:31
  need
• 7:32
  physical
• 7:32
  hardware
• 7:32
  and
• 7:33
  it's
• 7:33
  gone,
• 7:33
  you
• 7:34
  wait.
• 7:34
  Right?
• 7:34
  You
• 7:35
  wait.
• 7:35
  Or
• 7:36
  you're
• 7:36
  like,
• 7:37
  maybe
• 7:37
  I
• 7:38
  can
• 7:38
  figure
• 7:38
  out
• 7:38
  a
• 7:38
  way
• 7:39
  where
• 7:39
  I
• 7:39
  don't
• 7:40
  have
• 7:40
  to
• 7:40
  use
• 7:42
  all
• 7:42
  of
• 7:42
  this
• 7:43
  stuff
• 7:43
  to
• 7:43
  do
• 7:43
  the
• 7:44
  same
• 7:44
  thing.
• 7:44
  Is
• 7:45
  there
• 7:45
  a
• 7:45
  way?
• 7:46
  And
• 7:46
  Pluralis
• 7:46
  was
• 7:47
  like,
• 7:47
  let's
• 7:47
  figure
• 7:47
  it
• 7:47
  out.
• 7:48
  And
• 7:48
  so
• 7:48
  they
• 7:48
  figured
• 7:49
  out
• 7:49
  some
• 7:49
  cool
• 7:49
  stuff.
• 7:50
  How
• 7:50
  did
• 7:50
  they
• 7:51
  figure
• 7:51
  it
• 7:51
  out?
• 7:51
  Like,
• 7:51
  how
• 7:51
  do
• 7:52
  they
• 7:52
  do
• 7:52
  it?
• 7:52
  Okay,
• 7:52
  so
• 7:53
  the
• 7:53
  TLDR
• 7:53
  answer
• 7:54
  to
• 7:54
  that
• 7:54
  is
• 7:54
  basically
• 7:55
  they
• 7:55
  figured
• 7:55
  out
• 7:55
  a
• 7:55
  way
• 7:55
  to
• 7:56
  compress
• 7:57
  a
• 7:57
  lot.
• 7:58
  The
• 7:58
  amount
• 7:59
  of
• 7:59
  data,
• 8:00
  numbers
• 8:00
  essentially,
• 8:01
  that
• 8:01
  you
• 8:01
  need
• 8:01
  to
• 8:02
  pass
• 8:02
  through
• 8:03
  a
• 8:04
  training
• 8:04
  system
• 8:05
  to
• 8:05
  train
• 8:05
  a
• 8:05
  model,
• 8:06
  which
• 8:07
  means
• 8:08
  you
• 8:08
  can
• 8:08
  do
• 8:08
  it
• 8:08
  in
• 8:09
  a
• 8:09
  way
• 8:09
  that's
• 8:09
  much
• 8:09
  cheaper
• 8:10
  and
• 8:10
  faster.
• 8:11
  So
• 8:12
  I
• 8:12
  like
• 8:13
  to
• 8:13
  use
• 8:13
  analogies.
• 8:13
  When
• 8:14
  you
• 8:14
  train
• 8:14
  a
• 8:14
  neural
• 8:15
  network,
• 8:16
  it's
• 8:16
  kind
• 8:17
  of
• 8:17
  like
• 8:18
  hiking
• 8:18
  down
• 8:18
  a
• 8:18
  mountain
• 8:19
  in
• 8:19
  fog.
• 8:20
  Okay.
• 8:21
  So
• 8:21
  you
• 8:21
  can't
• 8:21
  see
• 8:21
  the
• 8:22
  bottom,
• 8:23
  you
• 8:23
  can
• 8:23
  feel
• 8:23
  which
• 8:24
  way
• 8:24
  is
• 8:24
  downhill?
• 8:25
  And
• 8:25
  as
• 8:26
  you
• 8:26
  train
• 8:26
  a
• 8:26
  model,
• 8:27
  the
• 8:27
  signals
• 8:28
  for
• 8:29
  which
• 8:30
  way
• 8:30
  is
• 8:30
  downhill
• 8:31
  get
• 8:31
  better
• 8:31
  and
• 8:32
  better
• 8:32
  throughout
• 8:33
  training.
• 8:34
  And
• 8:34
  then
• 8:34
  kind
• 8:34
  of
• 8:36
  peek
• 8:36
  at,
• 8:36
  okay,
• 8:37
  you've
• 8:37
  reached
• 8:38
  downhill,
• 8:39
  and
• 8:39
  you're
• 8:39
  like,
• 8:39
  I'm
• 8:40
  at
• 8:40
  the
• 8:40
  bottom.
• 8:40
  Like,
• 8:40
  I'm
• 8:40
  good.
• 8:41
  So
• 8:42
  that's
• 8:42
  like
• 8:42
  how
• 8:42
  you
• 8:42
  train
• 8:42
  a
• 8:42
  neural
• 8:43
  network
• 8:43
  in.
• 8:43
  I'm
• 8:44
  very.
• 8:44
  I'm
• 8:44
  simplifying.
• 8:45
  Yeah.
• 8:45
  And
• 8:45
  I
• 8:45
  bet
• 8:46
  it
• 8:46
  gets
• 8:46
  clearer
• 8:46
  and
• 8:46
  clearer
• 8:47
  as
• 8:47
  you
• 8:47
  train
• 8:47
  it
• 8:47
  better
• 8:48
  of,
• 8:48
  like,
• 8:48
  where
• 8:48
  the
• 8:48
  bottom
• 8:49
  is.
• 8:49
  Yes.
• 8:50
  Yeah.
• 8:51
  So
• 8:51
  there's
• 8:51
  a
• 8:51
  couple
• 8:51
  of
• 8:52
  different
• 8:52
  ways
• 8:53
  you
• 8:53
  can
• 8:54
  train
• 8:54
  a
• 8:55
  large
• 8:55
  model
• 8:56
  across
• 8:56
  multiple
• 8:57
  GPUs.
• 8:58
  Because
• 8:59
  you
• 8:59
  need
• 8:59
  multiple.
• 9:01
  One
• 9:01
  way
• 9:01
  is
• 9:01
  called
• 9:01
  data
• 9:02
  parallel.
• 9:02
  What
• 9:03
  this
• 9:03
  means
• 9:03
  is
• 9:03
  basically
• 9:04
  to
• 9:04
  train
• 9:05
  a
• 9:05
  model,
• 9:05
  you
• 9:05
  need
• 9:05
  the
• 9:05
  model
• 9:06
  and
• 9:06
  you
• 9:06
  need
• 9:06
  data.
• 9:07
  And
• 9:07
  essentially
• 9:08
  those
• 9:08
  two
• 9:08
  things
• 9:09
  together,
• 9:10
  you'll
• 9:11
  come
• 9:11
  out
• 9:11
  on
• 9:11
  the
• 9:11
  other
• 9:11
  end
• 9:12
  with
• 9:12
  a
• 9:12
  model
• 9:13
  that
• 9:13
  is
• 9:13
  trained.
• 9:13
  Yeah,
• 9:14
  yeah,
• 9:14
  yeah.
• 9:14
  And
• 9:14
  you
• 9:14
  need
• 9:14
  a
• 9:15
  lot
• 9:15
  of
• 9:15
  data.
• 9:17
  Data
• 9:17
  parallel
• 9:18
  is
• 9:18
  where
• 9:18
  each
• 9:19
  gpu,
• 9:20
  each
• 9:20
  chip,
• 9:21
  gets
• 9:21
  a
• 9:21
  copy
• 9:22
  of
• 9:22
  the
• 9:22
  model.
• 9:22
  So
• 9:23
  same
• 9:23
  model
• 9:24
  is
• 9:24
  living
• 9:24
  on
• 9:24
  each
• 9:25
  gpu,
• 9:26
  but
• 9:26
  you
• 9:26
  break
• 9:26
  up
• 9:26
  the
• 9:27
  data
• 9:27
  set
• 9:27
  into
• 9:28
  chunks
• 9:28
  across
• 9:28
  all
• 9:29
  these
• 9:29
  different
• 9:29
  GPUs,
• 9:29
  which
• 9:30
  are
• 9:30
  training
• 9:30
  it
• 9:30
  on
• 9:30
  the
• 9:30
  same
• 9:30
  model.
• 9:31
  And
• 9:31
  then
• 9:31
  it
• 9:31
  basically
• 9:31
  averages
• 9:32
  together
• 9:33
  the
• 9:34
  sort
• 9:34
  of
• 9:35
  outputs,
• 9:35
  let's
• 9:36
  say,
• 9:36
  of
• 9:36
  the
• 9:37
  data
• 9:37
  that
• 9:37
  it's
• 9:37
  trained
• 9:37
  on,
• 9:38
  and
• 9:38
  every
• 9:38
  copy
• 9:39
  of
• 9:39
  the
• 9:39
  model
• 9:39
  across
• 9:40
  all
• 9:40
  of
• 9:40
  the
• 9:40
  GPUs,
• 9:41
  gets
• 9:41
  updated
• 9:42
  consistently.
• 9:43
  Yeah.
• 9:44
  So
• 9:44
  basically
• 9:44
  it's
• 9:45
  like,
• 9:45
  okay,
• 9:45
  we
• 9:45
  can
• 9:45
  break
• 9:45
  up
• 9:45
  the
• 9:45
  data
• 9:46
  and
• 9:46
  then
• 9:46
  we'll,
• 9:46
  like,
• 9:46
  average
• 9:46
  together
• 9:47
  and,
• 9:47
  like,
• 9:47
  put
• 9:47
  it
• 9:47
  back
• 9:47
  together
• 9:48
  as
• 9:49
  if
• 9:49
  we'd
• 9:49
  never
• 9:49
  broke
• 9:50
  up
• 9:50
  the
• 9:50
  data.
• 9:50
  Yeah,
• 9:51
  that's
• 9:51
  one
• 9:51
  way
• 9:51
  you
• 9:51
  can
• 9:51
  do
• 9:52
  it.
• 9:52
  So
• 9:52
  that
• 9:52
  was
• 9:53
  data
• 9:53
  parallel.
• 9:53
  The
• 9:54
  other
• 9:54
  way
• 9:54
  you
• 9:54
  can
• 9:54
  do
• 9:54
  this
• 9:54
  is
• 9:55
  called
• 9:55
  model
• 9:55
  parallel.
• 9:56
  Okay.
• 9:57
  This
• 9:57
  is
• 9:57
  where
• 9:58
  you
• 9:58
  break
• 9:58
  up
• 9:59
  the
• 9:59
  model
• 9:59
  itself
• 10:01
  across
• 10:01
  gpu.
• 10:02
  So
• 10:02
  this
• 10:02
  will
• 10:02
  be
• 10:02
  like
• 10:02
  if
• 10:03
  I
• 10:03
  sliced
• 10:03
  Sarah's
• 10:04
  brain
• 10:05
  into,
• 10:05
  like,
• 10:06
  a
• 10:07
  hundred
• 10:07
  slices.
• 10:08
  Yeah.
• 10:08
  And
• 10:09
  you're
• 10:09
  like,
• 10:09
  this
• 10:09
  part
• 10:09
  of
• 10:09
  her
• 10:10
  brain
• 10:10
  takes
• 10:10
  this.
• 10:11
  Yes.
• 10:12
  So
• 10:12
  we're
• 10:12
  going
• 10:12
  to
• 10:12
  train
• 10:12
  it
• 10:13
  with
• 10:13
  that.
• 10:13
  Yes.
• 10:14
  And
• 10:16
  what
• 10:17
  is
• 10:17
  great
• 10:17
  about
• 10:17
  that
• 10:19
  is
• 10:21
  when
• 10:21
  you
• 10:22
  break
• 10:22
  up
• 10:22
  the
• 10:22
  model
• 10:23
  itself,
• 10:24
  you
• 10:24
  don't
• 10:24
  need
• 10:25
  H100
• 10:26
  Nvidia
• 10:27
  chips
• 10:28
  to
• 10:29
  contain
• 10:29
  each
• 10:30
  segment
• 10:30
  of
• 10:30
  the
• 10:30
  model
• 10:31
  and
• 10:31
  train
• 10:31
  on
• 10:31
  data.
• 10:32
  I
• 10:32
  could
• 10:32
  theoretically
• 10:33
  take
• 10:34
  part
• 10:34
  of
• 10:34
  the
• 10:34
  model
• 10:34
  and
• 10:34
  train
• 10:35
  it
• 10:35
  using
• 10:36
  my
• 10:36
  MacBook
• 10:37
  GPU.
• 10:38
  And
• 10:39
  so
• 10:40
  model
• 10:40
  parallelism
• 10:41
  is
• 10:42
  really
• 10:42
  the
• 10:43
  kind
• 10:43
  of
• 10:43
  interesting
• 10:43
  thing
• 10:44
  to
• 10:44
  look
• 10:44
  at,
• 10:44
  to
• 10:45
  crack
• 10:47
  in
• 10:47
  terms
• 10:47
  of
• 10:47
  figuring
• 10:48
  out,
• 10:48
  like,
• 10:48
  how
• 10:49
  could
• 10:49
  we
• 10:49
  train
• 10:49
  a
• 10:49
  model?
• 10:51
  Because
• 10:51
  data
• 10:51
  parallel,
• 10:52
  you
• 10:52
  can't
• 10:52
  really
• 10:52
  do
• 10:53
  it
• 10:53
  on
• 10:54
  small
• 10:54
  machines,
• 10:55
  small
• 10:56
  GPUs
• 10:56
  like
• 10:56
  that.
• 10:56
  Right.
• 10:57
  You
• 10:57
  need
• 10:57
  kind
• 10:58
  of
• 10:58
  like
• 10:58
  a
• 10:58
  big
• 10:58
  data
• 10:58
  center.
• 10:59
  Yeah.
• 11:00
  The
• 11:00
  other
• 11:00
  piece
• 11:00
  of
• 11:01
  this
• 11:01
  is
• 11:02
  with
• 11:02
  model
• 11:03
  parallel.
• 11:04
  Let
• 11:04
  me
• 11:04
  try
• 11:04
  and
• 11:05
  think
• 11:05
  of
• 11:05
  the
• 11:05
  right
• 11:05
  way
• 11:05
  to
• 11:05
  explain
• 11:06
  this.
• 11:06
  Sarah's
• 11:07
  brain
• 11:07
  is
• 11:07
  complex
• 11:08
  because
• 11:08
  she's
• 11:08
  brilliant,
• 11:09
  obviously.
• 11:10
  Right.
• 11:10
  She's
• 11:10
  so
• 11:11
  brilliant.
• 11:12
  And
• 11:13
  what
• 11:13
  that
• 11:13
  really
• 11:13
  means
• 11:14
  is
• 11:14
  Sarah's
• 11:14
  brain
• 11:15
  is
• 11:15
  made
• 11:16
  up
• 11:16
  of,
• 11:16
  like,
• 11:16
  a
• 11:17
  ton
• 11:17
  of
• 11:17
  numbers,
• 11:18
  a
• 11:18
  ton
• 11:18
  of
• 11:18
  mathematical
• 11:20
  calculations,
• 11:21
  like,
• 11:22
  millions,
• 11:22
  billions.
• 11:24
  And
• 11:25
  when
• 11:26
  I
• 11:26
  slice
• 11:26
  your
• 11:26
  brain
• 11:26
  into,
• 11:27
  like,
• 11:27
  a
• 11:27
  bunch
• 11:27
  of
• 11:27
  things,
• 11:29
  Brain's
• 11:30
  not
• 11:30
  a
• 11:30
  good
• 11:30
  example
• 11:31
  because,
• 11:31
  like,
• 11:31
  different
• 11:31
  parts
• 11:32
  do
• 11:32
  different
• 11:32
  things.
• 11:32
  But
• 11:33
  pretend
• 11:33
  your
• 11:33
  brain
• 11:34
  all,
• 11:34
  like,
• 11:35
  there's
• 11:35
  no,
• 11:35
  like,
• 11:35
  regions
• 11:36
  of
• 11:36
  the
• 11:36
  brain.
• 11:36
  Yeah,
• 11:36
  yeah,
• 11:37
  yeah.
• 11:37
  You
• 11:37
  just
• 11:37
  got,
• 11:37
  like,
• 11:37
  part
• 11:38
  of
• 11:38
  the
• 11:38
  brain.
• 11:38
  Right.
• 11:39
  It
• 11:39
  had
• 11:39
  all
• 11:39
  the
• 11:39
  same
• 11:39
  features,
• 11:40
  Right?
• 11:40
  Yeah.
• 11:41
  I
• 11:41
  train
• 11:42
  one
• 11:42
  part
• 11:42
  of
• 11:42
  your
• 11:42
  brain
• 11:43
  on
• 11:43
  data,
• 11:44
  and
• 11:44
  then
• 11:44
  it
• 11:44
  has
• 11:44
  to
• 11:44
  move
• 11:45
  to
• 11:45
  the
• 11:45
  next
• 11:45
  part.
• 11:45
  Right.
• 11:46
  It's
• 11:46
  like,
• 11:46
  okay,
• 11:47
  I'm
• 11:47
  sending
• 11:47
  this
• 11:48
  over
• 11:48
  to,
• 11:48
  like,
• 11:48
  this
• 11:48
  part
• 11:48
  of
• 11:48
  Sarah's
• 11:49
  brain
• 11:49
  to
• 11:49
  train
• 11:51
  is
• 11:51
  sending
• 11:52
  the
• 11:53
  entirety
• 11:53
  of
• 11:53
  all
• 11:54
  of
• 11:54
  your
• 11:54
  beautiful,
• 11:55
  brilliant
• 11:57
  calculations
• 11:59
  and
• 11:59
  numbers
• 12:00
  from
• 12:00
  one
• 12:01
  segment
• 12:01
  of
• 12:01
  your
• 12:01
  brain
• 12:01
  to
• 12:02
  the
• 12:02
  next.
• 12:02
  And
• 12:02
  when
• 12:02
  it's
• 12:03
  going,
• 12:03
  like,
• 12:03
  segment
• 12:04
  to
• 12:04
  segment,
• 12:04
  and
• 12:05
  maybe
• 12:05
  I'm
• 12:05
  jumping
• 12:05
  too
• 12:05
  far
• 12:06
  ahead
• 12:06
  and
• 12:06
  I'm
• 12:06
  making
• 12:06
  an
• 12:06
  assumption.
• 12:07
  But
• 12:07
  is
• 12:07
  it
• 12:07
  getting
• 12:07
  better
• 12:08
  every
• 12:08
  time?
• 12:08
  Yes.
• 12:09
  Is
• 12:09
  it
• 12:09
  like,
• 12:09
  okay,
• 12:10
  we
• 12:10
  got
• 12:10
  to
• 12:10
  hear
• 12:10
  with
• 12:10
  the
• 12:11
  last
• 12:11
  segment,
• 12:11
  and
• 12:11
  this
• 12:12
  is
• 12:12
  what
• 12:12
  we
• 12:12
  learned.
• 12:12
  Let's
• 12:12
  do
• 12:13
  next
• 12:13
  segment.
• 12:13
  See
• 12:13
  what
• 12:13
  we
• 12:14
  learned.
• 12:14
  Yeah,
• 12:14
  exactly.
• 12:15
  So,
• 12:15
  like,
• 12:15
  an
• 12:15
  example
• 12:16
  I
• 12:16
  was
• 12:16
  learning
• 12:16
  about
• 12:16
  that,
• 12:17
  I
• 12:17
  think
• 12:17
  is
• 12:17
  a
• 12:17
  really
• 12:17
  good
• 12:17
  way
• 12:17
  to
• 12:17
  visualize
• 12:18
  this
• 12:18
  is
• 12:18
  image
• 12:19
  recognition
• 12:20
  or
• 12:20
  image
• 12:20
  classifiers.
• 12:21
  So,
• 12:21
  like,
• 12:21
  if
• 12:22
  you
• 12:22
  were
• 12:22
  to
• 12:22
  train
• 12:22
  an
• 12:22
  AI
• 12:23
  model
• 12:23
  to
• 12:23
  show
• 12:24
  it
• 12:24
  a
• 12:24
  picture
• 12:24
  and
• 12:24
  be
• 12:24
  like,
• 12:24
  is
• 12:25
  this
• 12:25
  a
• 12:25
  cat?
• 12:26
  Different
• 12:26
  layers,
• 12:27
  as
• 12:27
  you
• 12:27
  train
• 12:28
  it
• 12:28
  can
• 12:28
  do
• 12:28
  different
• 12:28
  things.
• 12:28
  Might
• 12:29
  be,
• 12:29
  like,
• 12:29
  one
• 12:29
  set
• 12:29
  of
• 12:30
  layers
• 12:30
  is,
• 12:30
  like,
• 12:31
  identifies,
• 12:31
  like,
• 12:32
  sharp
• 12:32
  edges.
• 12:33
  Yeah.
• 12:33
  One
• 12:34
  set
• 12:34
  of
• 12:34
  layers
• 12:34
  then
• 12:35
  gets
• 12:35
  good
• 12:35
  at,
• 12:35
  like,
• 12:36
  understanding,
• 12:36
  like,
• 12:37
  whiskers.
• 12:38
  I
• 12:38
  don't
• 12:38
  know.
• 12:39
  So
• 12:39
  exactly
• 12:39
  like
• 12:40
  you're
• 12:40
  talking
• 12:40
  about,
• 12:42
  but
• 12:42
  you're
• 12:43
  still
• 12:43
  passing,
• 12:43
  like,
• 12:44
  a
• 12:44
  ton
• 12:45
  of
• 12:45
  numbers
• 12:46
  from
• 12:46
  one
• 12:46
  to
• 12:46
  the
• 12:46
  next.
• 12:47
  And
• 12:47
  that
• 12:48
  is
• 12:48
  impossible
• 12:49
  to
• 12:49
  do
• 12:49
  with
• 12:50
  low
• 12:50
  Internet
• 12:51
  connectivity,
• 12:52
  essentially.
• 12:52
  Yeah,
• 12:52
  yeah,
• 12:53
  yeah.
• 12:54
  So
• 12:55
  here's
• 12:55
  what
• 12:55
  pluralis
• 12:56
  figured
• 12:56
  out.
• 12:57
  Okay.
• 12:58
  Basically,
• 13:00
  what
• 13:00
  they
• 13:00
  figured
• 13:01
  out
• 13:01
  is
• 13:01
  you
• 13:02
  actually
• 13:03
  don't
• 13:03
  need
• 13:05
  to
• 13:06
  pass
• 13:07
  100%
• 13:08
  of
• 13:09
  the
• 13:09
  numbers
• 13:09
  that
• 13:09
  make
• 13:10
  up
• 13:10
  Sarah's
• 13:10
  brain
• 13:11
  from
• 13:12
  one
• 13:12
  layer
• 13:12
  to
• 13:12
  the
• 13:12
  next
• 13:13
  in
• 13:13
  order
• 13:14
  for
• 13:15
  this
• 13:16
  to
• 13:16
  work.
• 13:17
  Interesting.
• 13:18
  They
• 13:18
  basically
• 13:19
  figured
• 13:19
  out
• 13:20
  how
• 13:21
  you
• 13:21
  can
• 13:22
  communicate
• 13:23
  between
• 13:24
  model
• 13:24
  segments,
• 13:25
  let's
• 13:25
  say,
• 13:27
  how
• 13:27
  you
• 13:28
  can
• 13:28
  communicate
• 13:28
  enough
• 13:29
  of
• 13:29
  the
• 13:30
  same
• 13:30
  data,
• 13:32
  but
• 13:32
  efficiently,
• 13:33
  so
• 13:33
  that
• 13:33
  it's
• 13:34
  actually
• 13:34
  much
• 13:35
  smaller
• 13:35
  batches
• 13:36
  of
• 13:36
  number
• 13:36
  that
• 13:36
  you're
• 13:36
  passing
• 13:37
  from
• 13:37
  one
• 13:37
  to
• 13:37
  the
• 13:37
  next.
• 13:39
  But
• 13:39
  the
• 13:39
  numbers
• 13:39
  that
• 13:39
  you
• 13:40
  are
• 13:40
  choosing
• 13:40
  are
• 13:40
  the
• 13:41
  most
• 13:41
  important
• 13:42
  ones.
• 13:42
  And
• 13:43
  if
• 13:43
  you
• 13:43
  compare
• 13:44
  that
• 13:44
  as
• 13:45
  if
• 13:45
  you
• 13:45
  were
• 13:45
  to
• 13:47
  pass
• 13:48
  100%
• 13:49
  of
• 13:49
  the
• 13:49
  numbers
• 13:49
  from
• 13:50
  one
• 13:50
  to
• 13:50
  the
• 13:50
  next,
• 13:50
  you
• 13:51
  actually
• 13:51
  end
• 13:51
  up
• 13:52
  with
• 13:52
  the
• 13:52
  result
• 13:52
  that's
• 13:52
  the
• 13:53
  same,
• 13:53
  like,
• 13:53
  it's
• 13:54
  actually
• 13:54
  just
• 13:54
  as
• 13:54
  good.
• 13:55
  Fascinating.
• 13:55
  And
• 13:55
  it's
• 13:56
  interesting,
• 13:56
  like,
• 13:56
  all
• 13:56
  the
• 13:56
  pieces
• 13:57
  coming
• 13:57
  together,
• 13:57
  because
• 13:57
  an
• 13:58
  earlier
• 13:58
  episode
• 13:58
  we
• 13:59
  talked
• 13:59
  about
• 13:59
  perceptrons
• 14:00
  and
• 14:00
  how
• 14:01
  it
• 14:01
  uses
• 14:02
  the
• 14:02
  data
• 14:03
  to
• 14:03
  improve
• 14:04
  its
• 14:04
  linear
• 14:04
  functions
• 14:06
  in
• 14:06
  order
• 14:07
  to
• 14:07
  be
• 14:07
  able
• 14:08
  to
• 14:08
  predict
• 14:08
  better.
• 14:09
  And
• 14:09
  it
• 14:09
  sounds
• 14:09
  like
• 14:09
  they
• 14:09
  figured
• 14:10
  out
• 14:10
  how
• 14:10
  to
• 14:11
  take
• 14:11
  another
• 14:11
  step
• 14:11
  after
• 14:12
  that
• 14:12
  to
• 14:12
  be
• 14:12
  like,
• 14:13
  okay,
• 14:13
  so
• 14:13
  you
• 14:13
  figured
• 14:13
  out
• 14:14
  this
• 14:14
  linear
• 14:14
  function
• 14:14
  how
• 14:15
  to
• 14:15
  do
• 14:15
  it
• 14:15
  better.
• 14:16
  Can
• 14:16
  we
• 14:16
  send
• 14:17
  you
• 14:17
  less
• 14:17
  information
• 14:18
  that
• 14:18
  can
• 14:19
  do
• 14:19
  the
• 14:19
  same
• 14:19
  thing?
• 14:20
  Right.
• 14:20
  And
• 14:21
  they
• 14:21
  literally
• 14:21
  tested
• 14:22
  this
• 14:22
  on
• 14:23
  an
• 14:23
  8
• 14:23
  billion
• 14:24
  parameter
• 14:25
  model,
• 14:25
  which
• 14:26
  is
• 14:26
  like,
• 14:27
  big.
• 14:28
  Not
• 14:29
  the
• 14:29
  largest,
• 14:30
  but
• 14:31
  big.
• 14:32
  And
• 14:33
  they
• 14:33
  tested
• 14:34
  training
• 14:34
  it
• 14:35
  across
• 14:36
  64
• 14:37
  different
• 14:38
  GPUs
• 14:39
  that
• 14:39
  were
• 14:39
  like,
• 14:40
  not
• 14:40
  the
• 14:41
  expensive
• 14:41
  data
• 14:41
  center
• 14:42
  type
• 14:42
  of
• 14:42
  GPUs
• 14:43
  and
• 14:44
  spread
• 14:44
  across
• 14:45
  the
• 14:45
  US,
• 14:46
  Europe,
• 14:46
  and
• 14:46
  Asia.
• 14:47
  So
• 14:48
  one
• 14:48
  GPU
• 14:49
  that
• 14:49
  is
• 14:50
  sending
• 14:50
  the
• 14:51
  data
• 14:51
  to
• 14:51
  the
• 14:51
  next
• 14:52
  gpu,
• 14:52
  it
• 14:53
  might
• 14:53
  be
• 14:53
  sending
• 14:53
  it
• 14:53
  from,
• 14:54
  like,
• 14:54
  the
• 14:55
  US
• 14:55
  to
• 14:55
  Asia.
• 14:56
  And
• 14:57
  they
• 14:57
  tested
• 14:57
  that
• 14:57
  with
• 14:58
  connections,
• 14:59
  like
• 14:59
  Internet
• 14:59
  connections
• 15:00
  that
• 15:00
  are
• 15:00
  essentially,
• 15:01
  like,
• 15:01
  slower
• 15:01
  than
• 15:02
  many
• 15:02
  home
• 15:02
  Internet
• 15:03
  connections.
• 15:04
  And
• 15:05
  the
• 15:06
  end
• 15:06
  result
• 15:07
  was
• 15:08
  essentially
• 15:09
  just
• 15:09
  as
• 15:09
  good.
• 15:10
  So
• 15:10
  cool.
• 15:11
  I
• 15:11
  will
• 15:11
  say,
• 15:11
  like,
• 15:11
  the
• 15:11
  caveat
• 15:12
  is,
• 15:12
  you
• 15:12
  know,
• 15:13
  a
• 15:14
  skeptic
• 15:15
  of
• 15:15
  this
• 15:16
  might
• 15:16
  say,
• 15:16
  okay,
• 15:17
  but
• 15:17
  you
• 15:17
  haven't
• 15:17
  figured
• 15:17
  out
• 15:18
  how
• 15:18
  to
• 15:18
  do
• 15:18
  this
• 15:18
  on,
• 15:18
  like,
• 15:18
  a
• 15:18
  64
• 15:19
  billion
• 15:19
  parameter
• 15:20
  model.
• 15:20
  So,
• 15:20
  like,
• 15:20
  get
• 15:21
  back
• 15:21
  to
• 15:21
  me
• 15:21
  when.
• 15:22
  Yeah,
• 15:22
  you
• 15:22
  figure
• 15:23
  that
• 15:23
  out,
• 15:23
  which
• 15:23
  is.
• 15:23
  Like,
• 15:24
  somewhere
• 15:25
  totally,
• 15:25
  like,
• 15:26
  I
• 15:26
  remember
• 15:26
  a
• 15:27
  world
• 15:27
  where,
• 15:27
  like,
• 15:27
  I
• 15:28
  couldn't
• 15:28
  have
• 15:28
  believed,
• 15:29
  like,
• 15:29
  a
• 15:29
  meg
• 15:30
  of
• 15:30
  data
• 15:30
  was
• 15:31
  like,
• 15:31
  the
• 15:31
  craziest,
• 15:32
  like,
• 15:32
  the
• 15:32
  most
• 15:32
  data
• 15:33
  I
• 15:33
  could,
• 15:33
  like,
• 15:33
  imagine,
• 15:34
  you
• 15:34
  know?
• 15:34
  And
• 15:35
  so,
• 15:35
  like,
• 15:35
  like,
• 15:35
  I
• 15:36
  think
• 15:36
  we
• 15:36
  figure
• 15:37
  that
• 15:37
  stuff
• 15:37
  out.
• 15:37
  And
• 15:37
  it
• 15:38
  was
• 15:38
  really
• 15:38
  exciting
• 15:38
  to
• 15:38
  see
• 15:38
  stuff
• 15:39
  moving
• 15:39
  like
• 15:39
  that.
• 15:40
  Agreed.
• 15:42
  Amazing.
• 15:43
  Well,
• 15:43
  today
• 15:43
  we
• 15:43
  have
• 15:43
  Harper
• 15:44
  Reed
• 15:44
  on.
• 15:45
  It's
• 15:45
  a
• 15:45
  really
• 15:45
  fascinating
• 15:46
  conversation
• 15:47
  around
• 15:48
  if
• 15:48
  the
• 15:48
  end
• 15:48
  of
• 15:48
  the
• 15:48
  world
• 15:48
  is
• 15:48
  coming.
• 15:49
  And
• 15:49
  also
• 15:49
  fun
• 15:50
  things
• 15:50
  you
• 15:50
  can
• 15:50
  do
• 15:50
  in
• 15:50
  the
• 15:50
  meantime.
• 15:58
  Okay,
• 15:58
  everyone,
• 15:59
  Today
• 15:59
  we're
• 15:59
  talking
• 15:59
  to
• 15:59
  Harper
• 16:00
  Reed,
• 16:00
  the
• 16:00
  CEO
• 16:01
  of
• 16:01
  2389
• 16:03
  and
• 16:03
  also
• 16:03
  a
• 16:04
  bunch
• 16:04
  of
• 16:04
  other
• 16:04
  things.
• 16:05
  Welcome,
• 16:05
  Harper,
• 16:06
  to
• 16:06
  Chaos
• 16:06
  Agents.
• 16:07
  Thank
• 16:08
  you.
• 16:08
  I
• 16:08
  was
• 16:08
  telling
• 16:08
  some
• 16:08
  friends
• 16:09
  that
• 16:09
  I
• 16:09
  was
• 16:09
  coming
• 16:09
  to
• 16:10
  the
• 16:10
  show,
• 16:10
  and
• 16:10
  they
• 16:10
  were
• 16:11
  like,
• 16:11
  that
• 16:11
  seems
• 16:11
  appropriate
• 16:12
  for
• 16:12
  your
• 16:12
  life.
• 16:15
  So
• 16:15
  first
• 16:15
  I'll
• 16:16
  ask
• 16:16
  you
• 16:16
  our
• 16:16
  first
• 16:17
  question.
• 16:17
  We
• 16:17
  ask
• 16:18
  everyone.
• 16:19
  So
• 16:19
  when
• 16:19
  it
• 16:19
  comes
• 16:19
  to
• 16:19
  AI
• 16:21
  with
• 16:22
  zero
• 16:22
  being
• 16:23
  like,
• 16:23
  we're
• 16:24
  in
• 16:24
  the
• 16:24
  matrix,
• 16:25
  we're
• 16:25
  all
• 16:25
  batteries
• 16:26
  for
• 16:26
  the
• 16:26
  machines.
• 16:27
  And
• 16:28
  10
• 16:28
  being
• 16:28
  like,
• 16:29
  I
• 16:29
  have
• 16:30
  an
• 16:30
  AI
• 16:30
  best
• 16:30
  friend.
• 16:31
  It
• 16:32
  cleans
• 16:32
  my
• 16:32
  house.
• 16:33
  It
• 16:33
  can
• 16:33
  open
• 16:33
  and
• 16:33
  close
• 16:34
  my
• 16:34
  dishwasher
• 16:34
  really
• 16:35
  well,
• 16:35
  actually,
• 16:36
  it
• 16:36
  lives
• 16:36
  with
• 16:37
  me,
• 16:37
  and
• 16:37
  it's
• 16:37
  perfect.
• 16:38
  How
• 16:38
  are
• 16:38
  you
• 16:38
  feeling
• 16:39
  right
• 16:39
  now?
• 16:40
  That's
• 16:40
  such
• 16:40
  a
• 16:40
  hard
• 16:40
  question.
• 16:41
  Because
• 16:41
  I
• 16:42
  can
• 16:42
  only
• 16:42
  have
• 16:42
  those
• 16:43
  two
• 16:43
  examples
• 16:44
  in
• 16:44
  my
• 16:44
  head.
• 16:44
  It's
• 16:44
  very
• 16:45
  hard
• 16:45
  for
• 16:45
  Me
• 16:45
  to
• 16:45
  think
• 16:45
  of
• 16:45
  a
• 16:46
  gradient
• 16:46
  between
• 16:46
  those
• 16:47
  two.
• 16:47
  I'm
• 16:48
  like
• 16:50
  10
• 16:51
  when
• 16:52
  you
• 16:52
  take
• 16:52
  away
• 16:53
  humans.
• 16:54
  Yeah.
• 16:55
  And
• 16:55
  zero
• 16:56
  when
• 16:56
  you
• 16:56
  add
• 16:56
  humans
• 16:57
  into
• 16:57
  the
• 16:57
  equation.
• 16:58
  So
• 16:58
  like,
• 16:58
  if
• 16:59
  we
• 16:59
  didn't
• 16:59
  have
• 16:59
  like
• 17:00
  social
• 17:00
  structures
• 17:01
  that
• 17:01
  humans
• 17:01
  exist
• 17:02
  in,
• 17:02
  I'd
• 17:02
  be
• 17:02
  like,
• 17:02
  yeah,
• 17:03
  it's
• 17:03
  going
• 17:03
  to
• 17:03
  be
• 17:03
  incredible.
• 17:04
  But
• 17:04
  then
• 17:04
  you
• 17:04
  add
• 17:04
  in
• 17:05
  social
• 17:05
  structures
• 17:05
  that
• 17:06
  humans
• 17:06
  exist
• 17:06
  in
• 17:06
  and
• 17:06
  the
• 17:07
  systems
• 17:07
  that
• 17:07
  we
• 17:07
  have
• 17:07
  around
• 17:07
  us
• 17:08
  and
• 17:08
  I'm
• 17:08
  like,
• 17:08
  we're
• 17:08
  all
• 17:08
  going
• 17:09
  to
• 17:09
  die.
• 17:10
  And
• 17:10
  it's
• 17:10
  not
• 17:10
  like
• 17:10
  the
• 17:11
  paperclip
• 17:11
  factory,
• 17:12
  it's
• 17:12
  more
• 17:12
  like
• 17:13
  the
• 17:13
  regular
• 17:14
  factory
• 17:14
  but
• 17:15
  with
• 17:15
  more
• 17:15
  slaves.
• 17:16
  Yeah,
• 17:17
  it's
• 17:17
  true.
• 17:17
  I
• 17:18
  feel
• 17:18
  like
• 17:19
  you've
• 17:19
  probably
• 17:19
  also
• 17:20
  seen
• 17:20
  a
• 17:20
  lot
• 17:20
  of
• 17:20
  these,
• 17:21
  like,
• 17:21
  different
• 17:21
  booms,
• 17:22
  like
• 17:22
  crypto
• 17:23
  and
• 17:23
  social
• 17:24
  and
• 17:24
  mobile.
• 17:25
  I'm
• 17:25
  trying
• 17:25
  to
• 17:25
  collect
• 17:26
  them
• 17:26
  all.
• 17:26
  Yeah,
• 17:27
  yeah.
• 17:28
  That's
• 17:28
  my
• 17:28
  goal
• 17:29
  here,
• 17:29
  is
• 17:29
  to
• 17:29
  collect
• 17:29
  as
• 17:29
  many
• 17:30
  booms,
• 17:30
  especially
• 17:31
  tech
• 17:31
  booms,
• 17:31
  as
• 17:32
  possible.
• 17:32
  But
• 17:32
  zero
• 17:33
  bust,
• 17:33
  we're
• 17:33
  not
• 17:33
  here
• 17:33
  for
• 17:33
  the
• 17:34
  bus.
• 17:34
  Well,
• 17:34
  you
• 17:34
  know,
• 17:35
  it's
• 17:35
  part
• 17:35
  of
• 17:35
  the
• 17:35
  boom
• 17:36
  bust
• 17:36
  thing.
• 17:36
  Like
• 17:36
  there's
• 17:36
  always
• 17:37
  a
• 17:37
  pair.
• 17:38
  Do
• 17:38
  you
• 17:38
  think
• 17:38
  this
• 17:38
  is
• 17:39
  on
• 17:39
  a
• 17:39
  scale
• 17:39
  that
• 17:40
  is
• 17:40
  like
• 17:40
  completely
• 17:41
  different
• 17:41
  than
• 17:41
  what
• 17:42
  we've
• 17:42
  seen
• 17:42
  so
• 17:42
  far?
• 17:43
  Like,
• 17:43
  is
• 17:43
  this
• 17:44
  akin
• 17:44
  to
• 17:44
  the
• 17:44
  invention
• 17:45
  of
• 17:45
  the
• 17:45
  Internet
• 17:46
  or
• 17:46
  something
• 17:46
  like
• 17:47
  that?
• 17:47
  Or
• 17:47
  do
• 17:47
  you
• 17:47
  think
• 17:47
  it's
• 17:48
  more
• 17:48
  of
• 17:49
  one
• 17:49
  of
• 17:49
  these
• 17:50
  cycles?
• 17:51
  Well,
• 17:52
  my
• 17:53
  bias
• 17:53
  is
• 17:54
  that
• 17:54
  I
• 17:54
  love
• 17:55
  technology
• 17:56
  and
• 17:56
  so
• 17:57
  anytime
• 17:57
  someone's
• 17:58
  like,
• 17:58
  here
• 17:58
  is
• 17:58
  a
• 17:58
  cool
• 17:59
  new
• 17:59
  technology,
• 18:00
  I'm
• 18:00
  just
• 18:00
  like,
• 18:01
  oh
• 18:01
  my
• 18:01
  gosh,
• 18:01
  this
• 18:01
  is
• 18:02
  the
• 18:02
  best
• 18:02
  thing
• 18:02
  ever
• 18:02
  and
• 18:02
  this
• 18:02
  is
• 18:03
  going
• 18:03
  to
• 18:03
  change
• 18:03
  the
• 18:03
  world.
• 18:03
  It
• 18:03
  could
• 18:04
  be
• 18:04
  anything,
• 18:04
  truly
• 18:05
  anything.
• 18:05
  And
• 18:06
  I'm
• 18:06
  also
• 18:06
  relatively
• 18:07
  optimistic
• 18:09
  in
• 18:09
  my
• 18:09
  day
• 18:09
  to
• 18:10
  day
• 18:10
  life.
• 18:10
  And
• 18:10
  so
• 18:11
  I've
• 18:11
  been
• 18:11
  very
• 18:11
  excited
• 18:12
  about
• 18:12
  every
• 18:13
  hype
• 18:13
  cycle.
• 18:14
  I've
• 18:15
  fallen
• 18:15
  prey
• 18:15
  to
• 18:16
  the
• 18:16
  hype.
• 18:18
  This
• 18:18
  seems
• 18:19
  different
• 18:19
  only
• 18:19
  in
• 18:20
  that
• 18:21
  I
• 18:21
  think
• 18:21
  there's
• 18:21
  an
• 18:22
  existential
• 18:22
  bit
• 18:23
  here
• 18:23
  where
• 18:24
  with
• 18:24
  crypto,
• 18:24
  especially
• 18:25
  as
• 18:25
  a
• 18:25
  fintech
• 18:25
  entrepreneur,
• 18:26
  crypto
• 18:26
  was,
• 18:26
  was
• 18:27
  like
• 18:27
  catnip.
• 18:28
  It
• 18:28
  was
• 18:28
  incredible
• 18:29
  because
• 18:29
  you're
• 18:29
  just
• 18:29
  like,
• 18:29
  wait,
• 18:30
  I
• 18:30
  can
• 18:30
  build
• 18:30
  all
• 18:30
  these
• 18:30
  cool
• 18:31
  fintech
• 18:31
  things
• 18:31
  without
• 18:32
  having
• 18:32
  to
• 18:32
  do
• 18:32
  any
• 18:32
  of
• 18:32
  the
• 18:33
  fintech
• 18:33
  hard
• 18:33
  stuff.
• 18:34
  I
• 18:34
  can
• 18:34
  just
• 18:34
  put
• 18:35
  these
• 18:35
  glues
• 18:35
  together
• 18:35
  and
• 18:36
  all
• 18:36
  of
• 18:36
  a
• 18:36
  sudden
• 18:36
  you
• 18:36
  can
• 18:36
  send
• 18:36
  money
• 18:37
  all
• 18:37
  over
• 18:37
  the
• 18:37
  world.
• 18:38
  That
• 18:38
  was
• 18:39
  very,
• 18:39
  very
• 18:39
  exciting.
• 18:40
  I
• 18:40
  still
• 18:40
  think
• 18:40
  it's
• 18:41
  exciting,
• 18:41
  I
• 18:41
  just
• 18:41
  think
• 18:41
  it's
• 18:42
  scammy.
• 18:42
  But
• 18:42
  with
• 18:43
  crypto,
• 18:43
  that
• 18:43
  world
• 18:44
  never,
• 18:44
  I
• 18:44
  never
• 18:45
  felt
• 18:45
  I
• 18:46
  am
• 18:46
  going
• 18:46
  to
• 18:46
  go
• 18:46
  out
• 18:47
  of
• 18:47
  business
• 18:47
  myself,
• 18:48
  Harper.
• 18:48
  Like,
• 18:48
  my
• 18:49
  identity
• 18:49
  is
• 18:49
  not
• 18:49
  threatened
• 18:50
  by
• 18:50
  this.
• 18:51
  I'm
• 18:51
  not
• 18:52
  a
• 18:52
  bank,
• 18:52
  I'm
• 18:53
  not
• 18:53
  a
• 18:53
  Federal
• 18:54
  Reserve
• 18:54
  or
• 18:54
  I'm
• 18:55
  not
• 18:55
  a
• 18:55
  reserve
• 18:55
  currency,
• 18:55
  all
• 18:56
  these
• 18:56
  things.
• 18:56
  My
• 18:56
  identity
• 18:57
  was
• 18:57
  not
• 18:57
  fundamentally
• 18:58
  threatened
• 18:58
  by
• 18:59
  making
• 18:59
  money
• 18:59
  more
• 19:00
  efficient,
• 19:00
  faster,
• 19:01
  et
• 19:01
  cetera.
• 19:03
  However,
• 19:06
  the
• 19:07
  AI
• 19:07
  world,
• 19:08
  especially
• 19:08
  when
• 19:08
  it
• 19:09
  comes
• 19:09
  to
• 19:09
  CodeGen,
• 19:09
  which
• 19:10
  is
• 19:10
  kind
• 19:10
  of
• 19:10
  where
• 19:10
  I'm
• 19:10
  sitting
• 19:11
  a
• 19:11
  lot,
• 19:11
  spending
• 19:11
  a
• 19:11
  lot
• 19:11
  of
• 19:12
  time
• 19:12
  inside
• 19:12
  of
• 19:14
  is
• 19:14
  very
• 19:15
  existentially
• 19:15
  threatening
• 19:16
  towards
• 19:17
  my
• 19:17
  being.
• 19:18
  I
• 19:18
  mean,
• 19:19
  I
• 19:19
  graduated
• 19:19
  college
• 19:20
  in
• 19:20
  the
• 19:20
  early
• 19:20
  2000s.
• 19:21
  I've
• 19:21
  programmed
• 19:21
  since
• 19:22
  I
• 19:22
  was
• 19:22
  a
• 19:22
  little
• 19:22
  kid.
• 19:23
  Like,
• 19:23
  my
• 19:23
  first
• 19:23
  computer
• 19:24
  was
• 19:24
  BASIC
• 19:24
  on
• 19:25
  an
• 19:25
  Apple
• 19:25
  iic.
• 19:26
  I've
• 19:26
  programmed
• 19:26
  my
• 19:26
  entire
• 19:27
  life.
• 19:28
  And
• 19:28
  so
• 19:28
  it's
• 19:28
  this
• 19:28
  kind
• 19:28
  of
• 19:28
  bizarre
• 19:29
  space
• 19:29
  where
• 19:29
  I
• 19:30
  see
• 19:30
  that
• 19:30
  and
• 19:30
  I'm
• 19:30
  just
• 19:30
  like,
• 19:30
  wait,
• 19:31
  I
• 19:31
  don't
• 19:31
  need
• 19:31
  any
• 19:31
  of
• 19:31
  those
• 19:32
  skills
• 19:32
  anymore.
• 19:33
  And
• 19:33
  I've
• 19:34
  been
• 19:34
  challenged
• 19:35
  by
• 19:35
  that.
• 19:36
  I've
• 19:36
  also
• 19:36
  challenged
• 19:37
  myself
• 19:37
  to
• 19:37
  try
• 19:38
  and
• 19:38
  create
• 19:39
  really
• 19:39
  inclusive
• 19:39
  environments
• 19:40
  where
• 19:40
  I
• 19:40
  work
• 19:41
  with
• 19:41
  people
• 19:41
  who
• 19:41
  don't
• 19:42
  have
• 19:42
  my
• 19:42
  skills
• 19:42
  and
• 19:43
  see
• 19:43
  if
• 19:43
  they
• 19:43
  can
• 19:43
  achieve
• 19:43
  the
• 19:44
  things
• 19:44
  that
• 19:44
  I
• 19:44
  am
• 19:44
  proud
• 19:44
  of
• 19:45
  myself
• 19:45
  achieving,
• 19:46
  like
• 19:47
  coding
• 19:47
  something
• 19:47
  really
• 19:47
  fast,
• 19:48
  or
• 19:48
  like
• 19:48
  being
• 19:48
  a
• 19:49
  polyglot
• 19:49
  programmer
• 19:50
  or
• 19:50
  learning
• 19:50
  a
• 19:50
  new
• 19:50
  language
• 19:50
  really
• 19:51
  quick.
• 19:51
  Like,
• 19:51
  none
• 19:51
  of
• 19:51
  this
• 19:51
  is
• 19:52
  necessarily
• 19:52
  germane
• 19:53
  to,
• 19:53
  like,
• 19:53
  our
• 19:54
  world.
• 19:54
  It's
• 19:54
  much
• 19:54
  more
• 19:55
  relegated
• 19:56
  to
• 19:56
  my
• 19:56
  friends
• 19:56
  who
• 19:56
  are
• 19:57
  doing
• 19:57
  woodworking,
• 19:57
  where
• 19:58
  it's
• 19:58
  like,
• 19:58
  I
• 19:58
  have
• 19:58
  a
• 19:59
  new
• 19:59
  planer
• 19:59
  and
• 19:59
  it's
• 19:59
  really
• 20:00
  exciting,
• 20:00
  and
• 20:00
  I'm
• 20:00
  like,
• 20:00
  cool,
• 20:02
  I'm
• 20:02
  just
• 20:02
  gonna
• 20:02
  buy
• 20:02
  furniture.
• 20:04
  So
• 20:04
  I
• 20:05
  feel
• 20:05
  like
• 20:05
  that's
• 20:05
  kind
• 20:05
  of
• 20:05
  where
• 20:06
  it
• 20:06
  is
• 20:06
  now.
• 20:06
  It
• 20:06
  obviously
• 20:06
  doesn't
• 20:07
  have
• 20:07
  to
• 20:07
  be
• 20:07
  that
• 20:07
  you
• 20:07
  have
• 20:07
  your
• 20:08
  gradient
• 20:08
  of
• 20:08
  choice
• 20:08
  in
• 20:09
  there,
• 20:09
  where
• 20:09
  on
• 20:09
  the
• 20:09
  gradient
• 20:09
  you're
• 20:10
  picking,
• 20:11
  we're
• 20:11
  full
• 20:11
  in.
• 20:12
  We're
• 20:13
  completely
• 20:13
  in
• 20:14
  on
• 20:14
  it.
• 20:14
  And
• 20:14
  that
• 20:14
  has
• 20:15
  created
• 20:15
  this
• 20:15
  very
• 20:16
  complicated
• 20:18
  perspective.
• 20:18
  So
• 20:19
  what
• 20:19
  this
• 20:19
  means
• 20:19
  is,
• 20:20
  as
• 20:20
  hype
• 20:21
  cipher
• 20:21
  goals
• 20:22
  go,
• 20:22
  this
• 20:22
  one
• 20:22
  has
• 20:23
  affected
• 20:23
  me
• 20:24
  more
• 20:25
  emotionally
• 20:26
  and
• 20:26
  also
• 20:26
  with
• 20:27
  excitement.
• 20:27
  Because
• 20:28
  I
• 20:28
  think
• 20:28
  if
• 20:28
  I
• 20:28
  wasn't
• 20:29
  like,
• 20:29
  fuck,
• 20:29
  yeah,
• 20:29
  let's
• 20:29
  go,
• 20:30
  I
• 20:30
  would
• 20:30
  just
• 20:31
  be
• 20:31
  like,
• 20:31
  what
• 20:31
  am
• 20:31
  I
• 20:31
  going
• 20:31
  to
• 20:31
  do
• 20:31
  with
• 20:32
  my
• 20:32
  life?
• 20:32
  My
• 20:32
  entire
• 20:32
  identity
• 20:33
  is
• 20:33
  about
• 20:33
  programming
• 20:34
  computers
• 20:34
  fast
• 20:34
  with
• 20:34
  a
• 20:35
  large
• 20:35
  group
• 20:35
  of
• 20:35
  people.
• 20:36
  And
• 20:36
  now
• 20:36
  it's
• 20:37
  like,
• 20:37
  you
• 20:37
  take
• 20:37
  away
• 20:37
  the
• 20:37
  large
• 20:37
  group
• 20:38
  of
• 20:38
  people,
• 20:38
  you
• 20:38
  make
• 20:38
  it
• 20:38
  even
• 20:39
  faster.
• 20:39
  And
• 20:39
  it's
• 20:40
  not
• 20:40
  my
• 20:40
  fingers
• 20:40
  doing
• 20:40
  the
• 20:41
  typing.
• 20:41
  And
• 20:41
  then,
• 20:42
  because
• 20:43
  me
• 20:44
  and
• 20:44
  my
• 20:44
  team
• 20:44
  are
• 20:44
  insane,
• 20:45
  we're
• 20:45
  like,
• 20:46
  way
• 20:46
  into
• 20:46
  personifying
• 20:47
  and
• 20:47
  anthropomorphizing
• 20:48
  the
• 20:48
  AIs.
• 20:49
  We're
• 20:49
  like,
• 20:49
  all
• 20:49
  the
• 20:49
  way
• 20:49
  in.
• 20:50
  I
• 20:50
  love
• 20:50
  it.
• 20:50
  It's
• 20:50
  incredible.
• 20:51
  We're
• 20:52
  probably
• 20:52
  not
• 20:52
  going
• 20:52
  to
• 20:52
  fall
• 20:52
  in
• 20:53
  love
• 20:53
  with
• 20:53
  them,
• 20:53
  but,
• 20:53
  you
• 20:53
  know,
• 20:53
  there's
• 20:54
  still
• 20:54
  time,
• 20:54
  but
• 20:54
  we're
• 20:55
  definitely
• 20:55
  like,
• 20:56
  this
• 20:56
  is
• 20:56
  the
• 20:56
  future.
• 20:57
  It's
• 20:57
  so
• 20:57
  good.
• 20:58
  Can
• 20:58
  I
• 20:59
  press
• 20:59
  on
• 20:59
  something
• 21:00
  you
• 21:00
  said?
• 21:01
  I
• 21:01
  think
• 21:01
  it's
• 21:01
  so
• 21:01
  interesting,
• 21:02
  especially
• 21:02
  because
• 21:03
  you
• 21:03
  grew
• 21:03
  up
• 21:04
  programming
• 21:04
  and
• 21:04
  you
• 21:05
  have
• 21:05
  a
• 21:05
  kid.
• 21:06
  And
• 21:07
  I'm
• 21:07
  curious
• 21:08
  how
• 21:08
  you
• 21:09
  would
• 21:11
  want
• 21:11
  to
• 21:12
  do
• 21:12
  things
• 21:12
  differently
• 21:13
  if
• 21:13
  you
• 21:14
  were
• 21:14
  that
• 21:14
  age
• 21:15
  today.
• 21:15
  And
• 21:16
  we're
• 21:16
  kind
• 21:16
  of
• 21:16
  coming
• 21:16
  up
• 21:17
  like,
• 21:17
  do
• 21:17
  you
• 21:17
  think
• 21:18
  it's
• 21:18
  not
• 21:18
  worth
• 21:19
  it,
• 21:19
  for
• 21:19
  example,
• 21:19
  to
• 21:20
  spend
• 21:20
  a
• 21:20
  lot
• 21:20
  of
• 21:20
  time
• 21:21
  like
• 21:21
  actually
• 21:22
  learning
• 21:23
  first
• 21:23
  principles
• 21:24
  around
• 21:24
  programming
• 21:25
  or
• 21:26
  is
• 21:26
  there
• 21:26
  still
• 21:27
  value
• 21:28
  in
• 21:28
  that?
• 21:29
  I
• 21:30
  Wonder
• 21:30
  if
• 21:30
  the
• 21:30
  first
• 21:30
  principles
• 21:31
  that
• 21:31
  we
• 21:31
  would
• 21:31
  talk
• 21:31
  about
• 21:32
  10
• 21:32
  years
• 21:32
  ago
• 21:33
  or
• 21:33
  like
• 21:33
  my
• 21:33
  generation,
• 21:34
  the
• 21:34
  90s
• 21:35
  computer
• 21:35
  science
• 21:36
  school.
• 21:36
  Because
• 21:36
  when
• 21:36
  I
• 21:36
  went
• 21:36
  to
• 21:36
  school,
• 21:37
  like
• 21:37
  assembly
• 21:38
  was
• 21:38
  still
• 21:39
  kind
• 21:39
  of
• 21:39
  in
• 21:39
  the
• 21:40
  throes
• 21:40
  of
• 21:40
  dying
• 21:41
  from
• 21:41
  an
• 21:41
  education
• 21:41
  standpoint,
• 21:42
  like
• 21:42
  we
• 21:42
  still
• 21:43
  learned
• 21:43
  it
• 21:43
  in
• 21:43
  school.
• 21:44
  C
• 21:44
  was
• 21:45
  a
• 21:45
  big
• 21:45
  thing
• 21:45
  to
• 21:45
  learn.
• 21:45
  But
• 21:46
  the
• 21:46
  kids
• 21:46
  who
• 21:46
  came
• 21:46
  after
• 21:46
  me,
• 21:46
  many
• 21:47
  of
• 21:47
  them
• 21:47
  who
• 21:47
  I've
• 21:47
  hired,
• 21:48
  et
• 21:48
  cetera,
• 21:48
  don't
• 21:48
  have
• 21:49
  that
• 21:49
  kind
• 21:49
  of
• 21:49
  foundation.
• 21:50
  Assembly
• 21:50
  being
• 21:51
  like
• 21:51
  kind
• 21:51
  of
• 21:51
  the
• 21:51
  low
• 21:52
  level
• 21:52
  machine
• 21:53
  code
• 21:54
  of
• 21:54
  like
• 21:54
  how
• 21:55
  your
• 21:55
  computer
• 21:55
  is
• 21:56
  understanding.
• 21:56
  So
• 21:56
  like,
• 21:56
  I
• 21:57
  think
• 21:57
  there's
• 21:57
  a
• 21:57
  thing
• 21:57
  where
• 21:57
  it's
• 21:57
  like,
• 21:58
  what
• 21:58
  are
• 21:58
  the
• 21:58
  foundational
• 21:59
  things
• 21:59
  you
• 21:59
  should
• 22:00
  learn?
• 22:00
  I
• 22:00
  think
• 22:00
  is
• 22:01
  shifting
• 22:01
  as
• 22:01
  well.
• 22:02
  And
• 22:02
  what
• 22:02
  I
• 22:02
  mean
• 22:03
  by
• 22:03
  this
• 22:03
  is
• 22:04
  if
• 22:04
  you're
• 22:04
  just
• 22:05
  doing
• 22:05
  JavaScript
• 22:06
  or
• 22:06
  React
• 22:07
  or
• 22:07
  something
• 22:08
  like
• 22:08
  that,
• 22:08
  you
• 22:09
  probably
• 22:09
  don't
• 22:09
  really
• 22:10
  need
• 22:10
  to
• 22:10
  know
• 22:11
  like
• 22:11
  some
• 22:11
  foundational
• 22:12
  stuff.
• 22:13
  That
• 22:13
  doesn't
• 22:13
  mean
• 22:13
  you're
• 22:13
  not
• 22:13
  going
• 22:13
  to
• 22:14
  get
• 22:14
  into
• 22:14
  problems.
• 22:14
  But
• 22:14
  like,
• 22:15
  I
• 22:15
  don't
• 22:15
  know,
• 22:15
  I've
• 22:16
  hired
• 22:16
  a
• 22:16
  lot
• 22:16
  of
• 22:16
  people
• 22:17
  and
• 22:17
  none
• 22:17
  of
• 22:17
  them
• 22:17
  really
• 22:17
  have
• 22:18
  a
• 22:18
  handle
• 22:18
  on
• 22:18
  like
• 22:18
  big
• 22:18
  O
• 22:19
  notation.
• 22:20
  Now
• 22:20
  there
• 22:21
  are
• 22:21
  certain
• 22:21
  things
• 22:21
  where
• 22:22
  that's
• 22:22
  important,
• 22:22
  right?
• 22:23
  Like
• 22:23
  if
• 22:23
  you're
• 22:23
  doing
• 22:23
  some
• 22:24
  performance
• 22:24
  work,
• 22:24
  that's
• 22:24
  hugely
• 22:25
  important.
• 22:25
  And
• 22:25
  I've
• 22:25
  worked
• 22:26
  with
• 22:26
  people
• 22:26
  who
• 22:26
  do
• 22:26
  understand
• 22:27
  it.
• 22:27
  But
• 22:28
  in
• 22:28
  my
• 22:28
  time
• 22:29
  of
• 22:30
  working
• 22:30
  on
• 22:31
  big
• 22:31
  teams
• 22:31
  doing
• 22:32
  complicated
• 22:32
  things,
• 22:33
  very
• 22:34
  rarely
• 22:34
  has
• 22:34
  a
• 22:35
  lot
• 22:35
  of
• 22:35
  those
• 22:35
  concepts
• 22:35
  actually
• 22:36
  appeared.
• 22:36
  And
• 22:36
  that's
• 22:37
  because
• 22:37
  we'd
• 22:37
  obfuscated
• 22:38
  them
• 22:38
  away.
• 22:38
  Like
• 22:38
  I
• 22:38
  use
• 22:39
  an
• 22:39
  ORM
• 22:39
  that's
• 22:40
  handling
• 22:40
  some
• 22:40
  of
• 22:40
  this
• 22:40
  stuff.
• 22:41
  So
• 22:41
  someone
• 22:41
  in
• 22:41
  the
• 22:41
  world
• 22:42
  is
• 22:42
  thinking
• 22:42
  about
• 22:42
  it.
• 22:42
  It's
• 22:42
  just
• 22:43
  not
• 22:43
  me
• 22:43
  or
• 22:44
  100
• 22:44
  people
• 22:44
  around
• 22:45
  me
• 22:45
  that
• 22:46
  are
• 22:46
  thinking
• 22:46
  about
• 22:46
  it.
• 22:46
  And
• 22:47
  so
• 22:47
  I
• 22:47
  wonder
• 22:47
  what
• 22:48
  the
• 22:48
  next
• 22:48
  version
• 22:49
  of
• 22:49
  that
• 22:49
  is.
• 22:50
  Here's
• 22:50
  an
• 22:50
  example.
• 22:52
  I've
• 22:52
  been
• 22:52
  coaching
• 22:53
  and
• 22:53
  hanging
• 22:53
  out
• 22:53
  with
• 22:53
  a
• 22:53
  bunch
• 22:53
  of
• 22:53
  people
• 22:54
  who
• 22:54
  are
• 22:54
  doing
• 22:54
  a
• 22:54
  lot
• 22:54
  of
• 22:54
  code
• 22:54
  gen
• 22:55
  stuff
• 22:55
  and
• 22:55
  many
• 22:55
  of
• 22:55
  them
• 22:56
  know
• 22:56
  about
• 22:56
  programming,
• 22:56
  many
• 22:56
  of
• 22:57
  them
• 22:57
  don't.
• 22:57
  But
• 22:57
  like,
• 22:58
  I
• 22:58
  was
• 22:58
  talking
• 22:58
  to
• 22:59
  someone
• 22:59
  recently
• 22:59
  and
• 22:59
  they
• 22:59
  were
• 23:00
  like,
• 23:00
  I
• 23:00
  stopped
• 23:01
  being
• 23:01
  prescriptive
• 23:02
  about
• 23:02
  what
• 23:02
  programming
• 23:03
  language
• 23:03
  the
• 23:04
  model
• 23:04
  is
• 23:04
  choosing.
• 23:05
  Interesting.
• 23:06
  As
• 23:06
  long
• 23:06
  as
• 23:07
  I
• 23:07
  can
• 23:07
  deploy
• 23:07
  it
• 23:07
  to
• 23:08
  like
• 23:08
  Fly
• 23:09
  or
• 23:09
  Firebase
• 23:10
  or
• 23:10
  one
• 23:10
  of
• 23:10
  the
• 23:11
  others
• 23:11
  that
• 23:11
  they
• 23:11
  have
• 23:11
  chosen.
• 23:12
  They
• 23:12
  chose
• 23:12
  that
• 23:12
  purely
• 23:13
  because
• 23:13
  they
• 23:13
  have
• 23:13
  billing
• 23:14
  accounts
• 23:14
  there,
• 23:14
  not
• 23:14
  because
• 23:14
  of
• 23:15
  any
• 23:15
  great
• 23:16
  like
• 23:16
  want.
• 23:17
  And
• 23:17
  so
• 23:18
  they're
• 23:18
  just
• 23:18
  like,
• 23:18
  yeah,
• 23:18
  I'd
• 23:19
  love
• 23:19
  to
• 23:19
  deploy
• 23:19
  this
• 23:19
  to
• 23:20
  Fly
• 23:20
  because
• 23:20
  I
• 23:20
  have
• 23:21
  a
• 23:21
  Fly
• 23:21
  account
• 23:21
  already
• 23:22
  that's
• 23:22
  from
• 23:22
  before
• 23:23
  and
• 23:23
  I
• 23:24
  don't
• 23:24
  have
• 23:24
  to
• 23:24
  pay
• 23:24
  anything
• 23:25
  extra
• 23:25
  and
• 23:25
  it
• 23:25
  has
• 23:25
  a
• 23:26
  very
• 23:26
  nice
• 23:26
  free
• 23:26
  version.
• 23:26
  And
• 23:27
  so
• 23:28
  they're
• 23:28
  like,
• 23:28
  yeah,
• 23:28
  it
• 23:28
  turns
• 23:29
  out
• 23:29
  it's
• 23:29
  in
• 23:29
  Go.
• 23:30
  And
• 23:30
  I'm
• 23:30
  just
• 23:30
  like,
• 23:31
  oh,
• 23:31
  I
• 23:31
  didn't
• 23:31
  know
• 23:32
  you
• 23:32
  knew
• 23:32
  go.
• 23:32
  And
• 23:32
  they're
• 23:32
  like,
• 23:33
  I
• 23:33
  don't
• 23:33
  know,
• 23:33
  go.
• 23:34
  Okay,
• 23:34
  well
• 23:34
  then
• 23:35
  the
• 23:35
  question
• 23:35
  is
• 23:36
  like,
• 23:36
  what
• 23:36
  choice
• 23:37
  are
• 23:37
  you
• 23:37
  making?
• 23:37
  And
• 23:38
  for
• 23:38
  them
• 23:38
  the
• 23:38
  choice
• 23:38
  is
• 23:39
  then
• 23:39
  visual
• 23:40
  and
• 23:41
  does
• 23:41
  it
• 23:41
  work
• 23:41
  how
• 23:42
  they
• 23:42
  want
• 23:42
  it
• 23:42
  to
• 23:42
  work
• 23:42
  and
• 23:43
  does
• 23:43
  it
• 23:43
  look
• 23:43
  how
• 23:43
  they
• 23:43
  want
• 23:43
  it
• 23:44
  to
• 23:44
  look?
• 23:44
  And
• 23:45
  then
• 23:45
  insides
• 23:45
  they
• 23:46
  don't
• 23:46
  really
• 23:47
  care.
• 23:49
  This
• 23:49
  simultaneously
• 23:50
  excites
• 23:51
  me
• 23:51
  and
• 23:51
  hurts
• 23:51
  me
• 23:51
  inside.
• 23:52
  At
• 23:52
  a
• 23:52
  very
• 23:52
  core.
• 23:53
  I'm
• 23:53
  more
• 23:54
  excited
• 23:54
  because
• 23:55
  one
• 23:55
  of
• 23:55
  the
• 23:55
  things
• 23:55
  that
• 23:55
  I
• 23:56
  think
• 23:56
  has
• 23:56
  been
• 23:56
  the
• 23:56
  coolest
• 23:57
  about
• 23:57
  tech
• 23:57
  as
• 23:57
  we
• 23:57
  have
• 23:58
  gone
• 23:58
  from
• 23:59
  2000s
• 24:00
  where
• 24:00
  I
• 24:00
  was
• 24:00
  a
• 24:00
  Java
• 24:01
  programmer
• 24:01
  to
• 24:01
  today
• 24:02
  is
• 24:02
  just
• 24:02
  more
• 24:03
  people.
• 24:03
  I
• 24:04
  love
• 24:04
  sharing
• 24:04
  this
• 24:05
  with
• 24:05
  people
• 24:05
  and
• 24:05
  having
• 24:05
  more
• 24:06
  people
• 24:06
  around
• 24:06
  doing
• 24:06
  this.
• 24:07
  And
• 24:07
  my
• 24:08
  brother
• 24:08
  just
• 24:08
  sent
• 24:08
  me
• 24:09
  the
• 24:09
  theme
• 24:09
  he
• 24:09
  made
• 24:10
  for
• 24:10
  his
• 24:10
  blog,
• 24:10
  which
• 24:11
  is
• 24:11
  Dylan
• 24:11
  Blog,
• 24:12
  which
• 24:12
  is
• 24:13
  now
• 24:14
  a
• 24:14
  cookie
• 24:15
  clicker
• 24:15
  style
• 24:16
  game
• 24:17
  of
• 24:17
  which
• 24:17
  as
• 24:17
  you
• 24:18
  interact
• 24:18
  with
• 24:18
  the
• 24:18
  blog,
• 24:18
  you
• 24:19
  gain
• 24:19
  achievements.
• 24:20
  I
• 24:20
  love
• 24:20
  that.
• 24:21
  I'm
• 24:21
  just
• 24:21
  like,
• 24:21
  what?
• 24:22
  He
• 24:22
  started
• 24:22
  it
• 24:23
  yesterday
• 24:23
  afternoon
• 24:24
  and
• 24:24
  sent
• 24:24
  it
• 24:24
  to
• 24:25
  me
• 24:25
  this
• 24:25
  morning
• 24:25
  with
• 24:26
  changes.
• 24:27
  He
• 24:28
  doesn't
• 24:28
  know
• 24:28
  how
• 24:28
  to
• 24:29
  program
• 24:29
  cookie
• 24:29
  clicker
• 24:30
  style
• 24:30
  games,
• 24:30
  but
• 24:31
  I
• 24:31
  don't
• 24:31
  think
• 24:31
  that's
• 24:31
  the
• 24:31
  question
• 24:32
  anymore.
• 24:32
  It
• 24:32
  used
• 24:32
  to
• 24:32
  be
• 24:33
  that
• 24:33
  was
• 24:33
  the
• 24:33
  question.
• 24:33
  Do
• 24:34
  you
• 24:34
  know
• 24:34
  how
• 24:34
  to
• 24:34
  do
• 24:34
  this
• 24:34
  thing?
• 24:35
  I
• 24:35
  just
• 24:35
  don't
• 24:35
  think
• 24:35
  it's
• 24:35
  germane
• 24:37
  because
• 24:37
  he
• 24:37
  doesn't
• 24:38
  need
• 24:38
  to.
• 24:38
  He
• 24:38
  was
• 24:38
  able
• 24:39
  to
• 24:39
  do
• 24:39
  it
• 24:39
  just
• 24:39
  fine.
• 24:40
  And
• 24:40
  certainly
• 24:41
  I'm
• 24:41
  sure
• 24:41
  someone
• 24:41
  could
• 24:42
  have
• 24:42
  done
• 24:42
  it
• 24:42
  better
• 24:43
  over
• 24:43
  like
• 24:43
  a
• 24:43
  three
• 24:44
  year
• 24:44
  period
• 24:44
  of
• 24:45
  desperately
• 24:45
  writing
• 24:46
  each
• 24:46
  line
• 24:46
  of
• 24:46
  code
• 24:47
  themselves
• 24:48
  and
• 24:48
  being
• 24:48
  careful
• 24:48
  and
• 24:48
  all
• 24:49
  this
• 24:49
  stuff.
• 24:49
  But
• 24:49
  I
• 24:49
  don't
• 24:50
  think
• 24:50
  he's
• 24:50
  really
• 24:50
  worried
• 24:51
  about
• 24:51
  it
• 24:52
  being
• 24:52
  correct.
• 24:54
  He's
• 24:55
  worried
• 24:55
  about
• 24:55
  it.
• 24:55
  Does
• 24:55
  it
• 24:55
  make
• 24:55
  him
• 24:56
  feel
• 24:56
  good?
• 24:56
  Is
• 24:56
  it
• 24:57
  right
• 24:57
  for
• 24:57
  him
• 24:57
  at
• 24:58
  this
• 24:58
  moment?
• 24:59
  And
• 24:59
  like
• 24:59
  if
• 24:59
  he
• 24:59
  was
• 24:59
  entering
• 25:00
  like
• 25:00
  a
• 25:00
  minified
• 25:01
  JavaScript
• 25:01
  competition
• 25:02
  to
• 25:02
  see
• 25:02
  he
• 25:02
  can
• 25:02
  write,
• 25:02
  you
• 25:02
  know,
• 25:03
  like,
• 25:03
  does
• 25:03
  your
• 25:03
  code
• 25:03
  look
• 25:04
  like
• 25:04
  a
• 25:04
  dolphin?
• 25:04
  Like,
• 25:05
  I'm
• 25:05
  sure
• 25:05
  he
• 25:05
  would
• 25:05
  fail
• 25:06
  but
• 25:06
  like,
• 25:06
  that's
• 25:07
  not
• 25:07
  what
• 25:07
  the
• 25:07
  world
• 25:07
  is.
• 25:07
  That's
• 25:08
  why
• 25:08
  I
• 25:08
  think
• 25:08
  it's
• 25:08
  interesting.
• 25:09
  And
• 25:09
  Sarah,
• 25:09
  I'm
• 25:09
  curious
• 25:09
  how
• 25:10
  you
• 25:10
  think
• 25:10
  about
• 25:10
  this
• 25:10
  too,
• 25:10
  but
• 25:11
  it
• 25:11
  feels
• 25:11
  like
• 25:12
  that's
• 25:12
  a
• 25:12
  great
• 25:12
  example
• 25:14
  of
• 25:14
  like,
• 25:15
  there
• 25:15
  are
• 25:16
  certain
• 25:17
  products,
• 25:18
  let's
• 25:18
  say,
• 25:18
  where
• 25:19
  what
• 25:19
  you
• 25:19
  just
• 25:19
  described
• 25:20
  is
• 25:21
  perfect.
• 25:22
  Like
• 25:23
  you're
• 25:23
  building
• 25:23
  for
• 25:23
  yourself,
• 25:24
  you're
• 25:24
  building
• 25:24
  for.
• 25:25
  It's
• 25:25
  not
• 25:25
  something
• 25:25
  that
• 25:26
  needs
• 25:26
  to
• 25:26
  be
• 25:26
  scalable
• 25:27
  necessarily.
• 25:28
  It's
• 25:28
  not
• 25:28
  something
• 25:28
  that
• 25:28
  needs
• 25:28
  to
• 25:28
  be
• 25:29
  like
• 25:29
  consumer
• 25:29
  grade.
• 25:29
  You're
• 25:30
  not
• 25:30
  going
• 25:30
  to
• 25:30
  be
• 25:30
  charging
• 25:30
  for
• 25:31
  it.
• 25:31
  Like
• 25:31
  you're
• 25:31
  not
• 25:31
  trying
• 25:31
  to
• 25:32
  build
• 25:32
  a
• 25:32
  business
• 25:32
  around
• 25:33
  it
• 25:34
  and
• 25:34
  certainly
• 25:35
  code
• 25:36
  gen
• 25:36
  has
• 25:37
  democratized
• 25:38
  the
• 25:38
  ability
• 25:39
  for
• 25:39
  people
• 25:39
  to
• 25:40
  come
• 25:40
  up
• 25:40
  with
• 25:40
  an
• 25:40
  idea
• 25:41
  and
• 25:41
  make
• 25:41
  stuff
• 25:41
  like
• 25:41
  that,
• 25:42
  which
• 25:42
  is
• 25:42
  like,
• 25:44
  as
• 25:44
  someone
• 25:44
  who
• 25:45
  their
• 25:45
  whole
• 25:45
  life
• 25:46
  wanted
• 25:46
  to
• 25:47
  be
• 25:47
  able
• 25:47
  to
• 25:47
  do
• 25:48
  that
• 25:48
  and
• 25:48
  struggled,
• 25:49
  it
• 25:49
  is
• 25:49
  like
• 25:50
  amazing
• 25:51
  thing.
• 25:52
  I
• 25:53
  hear,
• 25:53
  I
• 25:53
  hear
• 25:53
  a
• 25:53
  but
• 25:54
  coming.
• 25:54
  Yeah.
• 25:55
  And
• 25:55
  this
• 25:55
  is
• 25:55
  why,
• 25:55
  where,
• 25:56
  where
• 25:56
  I'm
• 25:56
  curious
• 25:57
  because
• 25:57
  you
• 25:57
  said
• 25:57
  like,
• 25:58
  oh,
• 25:58
  I'm,
• 25:58
  you
• 25:58
  know,
• 25:58
  there's
• 25:59
  a
• 25:59
  level
• 25:59
  of
• 25:59
  concern
• 26:00
  about
• 26:00
  like
• 26:00
  your
• 26:00
  identity
• 26:01
  as
• 26:01
  a
• 26:01
  programmer,
• 26:02
  software
• 26:02
  developer
• 26:03
  being
• 26:03
  replaced.
• 26:05
  To
• 26:05
  me,
• 26:05
  at
• 26:05
  least
• 26:06
  right
• 26:06
  now,
• 26:07
  and
• 26:07
  I
• 26:07
  don't
• 26:07
  know
• 26:07
  how
• 26:07
  far
• 26:08
  away
• 26:08
  we
• 26:08
  are
• 26:08
  from
• 26:08
  this.
• 26:08
  It
• 26:09
  doesn't
• 26:09
  feel
• 26:09
  like
• 26:11
  I,
• 26:12
  armed
• 26:12
  with
• 26:13
  cursor
• 26:13
  or
• 26:13
  Claude
• 26:14
  code
• 26:14
  or
• 26:14
  replit
• 26:15
  or
• 26:15
  whatever,
• 26:15
  could
• 26:16
  get
• 26:16
  a
• 26:16
  job
• 26:17
  as
• 26:17
  a
• 26:18
  software
• 26:18
  developer
• 26:19
  at
• 26:19
  a
• 26:19
  great
• 26:19
  tech
• 26:20
  company
• 26:20
  and
• 26:20
  be
• 26:21
  at
• 26:21
  the
• 26:21
  same
• 26:21
  level
• 26:22
  and
• 26:22
  know
• 26:22
  what
• 26:23
  to
• 26:23
  do
• 26:23
  and
• 26:24
  be
• 26:24
  shipping
• 26:25
  great
• 26:25
  code
• 26:25
  or
• 26:26
  even
• 26:27
  start
• 26:27
  a
• 26:27
  company
• 26:29
  with
• 26:29
  three
• 26:29
  other
• 26:30
  people
• 26:30
  where
• 26:30
  none
• 26:30
  of
• 26:30
  us
• 26:31
  are
• 26:32
  trained
• 26:32
  programmers
• 26:33
  or
• 26:33
  engineers
• 26:34
  and
• 26:34
  build
• 26:34
  something
• 26:35
  that
• 26:35
  is
• 26:35
  consumer
• 26:35
  grade
• 26:36
  that
• 26:36
  we
• 26:36
  could
• 26:36
  actually
• 26:36
  build
• 26:36
  a
• 26:37
  business
• 26:37
  around.
• 26:39
  What
• 26:40
  are
• 26:40
  your
• 26:40
  thoughts
• 26:41
  on
• 26:41
  that?
• 26:42
  Yeah,
• 26:42
  I
• 26:42
  think
• 26:42
  the
• 26:43
  thing
• 26:43
  I've
• 26:44
  been
• 26:44
  talking
• 26:44
  to
• 26:44
  people
• 26:44
  about
• 26:45
  recently,
• 26:45
  I've
• 26:46
  been
• 26:46
  talking
• 26:46
  to
• 26:46
  a
• 26:46
  bunch
• 26:46
  of
• 26:46
  people
• 26:47
  that
• 26:47
  are
• 26:47
  building
• 26:48
  browsers
• 26:48
  and
• 26:48
  browser
• 26:49
  tools
• 26:49
  and
• 26:50
  they're
• 26:50
  thinking
• 26:50
  about
• 26:51
  like,
• 26:51
  okay,
• 26:52
  for
• 26:52
  example,
• 26:52
  I
• 26:52
  heard
• 26:53
  a
• 26:53
  metric
• 26:53
  that
• 26:53
  was
• 26:54
  like,
• 26:54
  in
• 26:54
  this
• 26:54
  past
• 26:55
  quarter,
• 26:56
  13
• 26:56
  million
• 26:56
  websites
• 26:57
  were
• 26:57
  created,
• 26:58
  which
• 26:58
  is
• 26:59
  more
• 26:59
  than
• 26:59
  they've
• 26:59
  ever
• 27:00
  seen.
• 27:00
  And
• 27:00
  I'm
• 27:01
  sure
• 27:01
  some
• 27:01
  of
• 27:01
  those
• 27:01
  are
• 27:01
  just
• 27:02
  like,
• 27:02
  oh,
• 27:02
  I'm
• 27:02
  testing,
• 27:03
  blah,
• 27:03
  blah,
• 27:03
  blah,
• 27:03
  like
• 27:03
  whatever
• 27:04
  this
• 27:04
  is.
• 27:04
  And
• 27:04
  some
• 27:04
  of
• 27:04
  them
• 27:04
  are
• 27:05
  your
• 27:05
  brother's
• 27:05
  awesome
• 27:06
  cookie
• 27:06
  clicker
• 27:07
  blog,
• 27:07
  which
• 27:07
  is
• 27:08
  really
• 27:08
  functional
• 27:09
  and
• 27:09
  neat.
• 27:10
  And
• 27:10
  I
• 27:11
  think
• 27:11
  people
• 27:11
  are
• 27:11
  thinking
• 27:12
  about
• 27:12
  what
• 27:12
  kind
• 27:12
  of
• 27:12
  tooling
• 27:13
  to
• 27:13
  make
• 27:13
  for
• 27:13
  people
• 27:14
  with
• 27:14
  this
• 27:14
  background,
• 27:15
  right.
• 27:15
  Of
• 27:15
  people
• 27:16
  that
• 27:16
  haven't
• 27:16
  been
• 27:16
  coding
• 27:17
  their
• 27:17
  whole
• 27:17
  lives
• 27:18
  and
• 27:19
  start
• 27:19
  coding,
• 27:19
  what
• 27:20
  do
• 27:20
  you
• 27:20
  need
• 27:20
  to
• 27:20
  build
• 27:20
  to
• 27:20
  make
• 27:21
  that
• 27:21
  easier
• 27:22
  for
• 27:22
  them
• 27:22
  and
• 27:22
  make
• 27:23
  it
• 27:23
  not
• 27:23
  blow
• 27:23
  up
• 27:23
  in
• 27:23
  their
• 27:24
  face?
• 27:24
  So
• 27:24
  I
• 27:24
  think
• 27:25
  we're
• 27:25
  not
• 27:25
  quite
• 27:25
  there
• 27:26
  yet.
• 27:26
  I
• 27:26
  do
• 27:27
  think
• 27:27
  that,
• 27:27
  like,
• 27:28
  you
• 27:28
  talk
• 27:28
  to
• 27:28
  the
• 27:28
  smaller
• 27:29
  tech
• 27:29
  companies,
• 27:29
  they're
• 27:30
  really
• 27:30
  comfortable
• 27:31
  hiring
• 27:32
  new
• 27:32
  engineers
• 27:33
  that
• 27:33
  just
• 27:33
  have
• 27:33
  cursor
• 27:34
  experience
• 27:35
  or
• 27:35
  like
• 27:35
  only
• 27:36
  know
• 27:36
  how
• 27:36
  to
• 27:36
  build
• 27:36
  a
• 27:36
  website
• 27:37
  on
• 27:37
  cursor
• 27:37
  or
• 27:38
  only
• 27:38
  know
• 27:38
  how
• 27:38
  to
• 27:38
  do
• 27:38
  certain
• 27:39
  things.
• 27:40
  And
• 27:40
  I
• 27:41
  think
• 27:43
  as
• 27:43
  code
• 27:44
  gets
• 27:44
  easier
• 27:44
  to
• 27:44
  throw
• 27:44
  away,
• 27:45
  actually
• 27:45
  that's
• 27:46
  more
• 27:46
  present
• 27:46
  in
• 27:47
  the
• 27:47
  space.
• 27:47
  So
• 27:47
  I
• 27:47
  do
• 27:48
  think
• 27:48
  that
• 27:48
  software
• 27:48
  engineering.
• 27:49
  Is
• 27:49
  going
• 27:49
  to
• 27:49
  look
• 27:50
  very
• 27:50
  different.
• 27:51
  I'm
• 27:51
  basically
• 27:52
  an
• 27:52
  extremist
• 27:53
  at
• 27:53
  this
• 27:53
  point.
• 27:54
  I
• 27:54
  think
• 27:54
  it's
• 27:54
  over.
• 27:55
  You
• 27:55
  think
• 27:56
  it's
• 27:56
  over?
• 27:56
  I
• 27:57
  think
• 27:57
  we're
• 27:57
  done.
• 27:57
  Yeah,
• 27:58
  we're
• 27:58
  calling
• 27:58
  it.
• 27:58
  Here's
• 27:59
  why.
• 27:59
  So
• 27:59
  we
• 28:00
  had
• 28:00
  an
• 28:00
  intern
• 28:00
  program.
• 28:01
  I
• 28:01
  was
• 28:01
  testing
• 28:01
  a
• 28:01
  thing.
• 28:02
  My
• 28:03
  assumption
• 28:04
  was
• 28:04
  with
• 28:05
  like
• 28:05
  Claude
• 28:06
  Code
• 28:07
  and
• 28:07
  a
• 28:07
  good
• 28:07
  coach,
• 28:09
  anyone
• 28:09
  can
• 28:09
  be
• 28:10
  just
• 28:10
  as
• 28:10
  good
• 28:10
  as
• 28:10
  I
• 28:10
  am.
• 28:11
  And
• 28:11
  so
• 28:12
  we
• 28:13
  had
• 28:13
  this
• 28:14
  and
• 28:14
  I
• 28:14
  read
• 28:14
  this
• 28:14
  random
• 28:15
  tweet.
• 28:16
  I
• 28:16
  refuse
• 28:16
  it
• 28:16
  to
• 28:16
  call
• 28:16
  it
• 28:17
  X.
• 28:17
  You
• 28:17
  have
• 28:17
  to
• 28:17
  die
• 28:17
  on
• 28:17
  some
• 28:18
  hills.
• 28:18
  I
• 28:18
  have
• 28:18
  very
• 28:19
  few
• 28:19
  hills
• 28:19
  that
• 28:19
  I'm
• 28:19
  willing
• 28:19
  to
• 28:19
  die
• 28:20
  on
• 28:20
  and
• 28:20
  I
• 28:20
  just
• 28:20
  pick
• 28:20
  that
• 28:20
  one.
• 28:21
  For
• 28:21
  some
• 28:21
  reason
• 28:21
  I
• 28:21
  sometimes
• 28:22
  regret
• 28:22
  it
• 28:22
  because
• 28:22
  I
• 28:22
  feel
• 28:22
  like
• 28:22
  a
• 28:23
  child.
• 28:23
  But
• 28:23
  whatever.
• 28:24
  I
• 28:25
  read
• 28:26
  this
• 28:26
  tweet
• 28:26
  and
• 28:26
  it
• 28:27
  was
• 28:27
  essentially
• 28:27
  this
• 28:27
  guy
• 28:27
  said,
• 28:28
  the
• 28:28
  team
• 28:28
  of
• 28:28
  the
• 28:28
  future
• 28:29
  is
• 28:31
  coaches,
• 28:31
  where
• 28:32
  you
• 28:32
  have
• 28:32
  someone
• 28:32
  on
• 28:33
  the
• 28:33
  team
• 28:33
  that
• 28:33
  knows
• 28:33
  all
• 28:34
  the
• 28:34
  stuff
• 28:34
  and
• 28:35
  they're
• 28:35
  coaching
• 28:35
  people
• 28:35
  who
• 28:36
  don't
• 28:36
  know
• 28:36
  all
• 28:36
  the
• 28:37
  stuff
• 28:37
  yet.
• 28:38
  And
• 28:38
  I
• 28:39
  like
• 28:39
  this
• 28:39
  because
• 28:40
  there's
• 28:40
  two
• 28:41
  reasons
• 28:41
  I
• 28:41
  like
• 28:41
  it.
• 28:41
  One,
• 28:42
  my
• 28:42
  partner
• 28:43
  works
• 28:43
  at
• 28:43
  Deloitte
• 28:44
  and
• 28:44
  they
• 28:44
  do
• 28:44
  a
• 28:44
  really
• 28:44
  strong
• 28:46
  coaching
• 28:46
  world.
• 28:46
  And
• 28:47
  I'm
• 28:47
  always
• 28:47
  admiring
• 28:48
  how
• 28:48
  much
• 28:48
  support
• 28:49
  that
• 28:49
  they
• 28:49
  have
• 28:50
  both
• 28:50
  as
• 28:51
  someone
• 28:51
  who's
• 28:51
  coaching
• 28:52
  someone
• 28:52
  and
• 28:52
  who's.
• 28:53
  It
• 28:53
  just
• 28:53
  really
• 28:54
  helps
• 28:54
  people
• 28:54
  belong,
• 28:55
  know
• 28:55
  all
• 28:55
  the
• 28:55
  knowledge,
• 28:56
  get
• 28:56
  ahead
• 28:57
  in
• 28:57
  their
• 28:57
  career.
• 28:57
  And
• 28:58
  one
• 28:58
  of
• 28:58
  the
• 28:58
  things
• 28:58
  I
• 28:58
  think
• 28:58
  is
• 28:59
  missing
• 28:59
  from
• 28:59
  tech
• 29:00
  is
• 29:00
  like
• 29:00
  the
• 29:00
  ability
• 29:01
  to
• 29:02
  get
• 29:02
  good
• 29:02
  mentorship,
• 29:03
  like
• 29:04
  built
• 29:04
  in
• 29:04
  as
• 29:04
  infrastructure.
• 29:05
  I
• 29:05
  just
• 29:05
  think
• 29:06
  that's
• 29:06
  something
• 29:06
  that
• 29:06
  I
• 29:06
  have
• 29:06
  not
• 29:07
  seen
• 29:07
  in
• 29:07
  a
• 29:07
  lot
• 29:07
  of
• 29:07
  my
• 29:07
  career.
• 29:08
  And
• 29:08
  yet
• 29:09
  the
• 29:09
  people
• 29:09
  around
• 29:10
  me,
• 29:10
  we've
• 29:10
  been
• 29:10
  lucky
• 29:11
  because
• 29:11
  our
• 29:11
  crew,
• 29:11
  we've
• 29:12
  worked
• 29:12
  together
• 29:12
  for
• 29:12
  a
• 29:12
  long
• 29:12
  time,
• 29:13
  so
• 29:13
  we've
• 29:13
  been
• 29:13
  able
• 29:13
  to
• 29:14
  mentor
• 29:14
  and
• 29:14
  coach
• 29:14
  ourselves.
• 29:15
  But
• 29:16
  when
• 29:16
  I
• 29:16
  worked
• 29:16
  at,
• 29:16
  you
• 29:16
  know,
• 29:16
  a
• 29:17
  big,
• 29:17
  bigger
• 29:17
  companies,
• 29:17
  that
• 29:17
  just
• 29:18
  wasn't
• 29:18
  the
• 29:18
  case.
• 29:18
  It
• 29:18
  was
• 29:18
  like
• 29:19
  only
• 29:19
  a
• 29:19
  special
• 29:19
  program
• 29:19
  for
• 29:20
  leadership
• 29:20
  potential.
• 29:21
  And
• 29:21
  it's
• 29:21
  like,
• 29:21
  no,
• 29:21
  what
• 29:21
  if
• 29:21
  it
• 29:22
  was
• 29:22
  just
• 29:22
  for
• 29:22
  everyone?
• 29:23
  So
• 29:23
  I
• 29:23
  really
• 29:23
  like
• 29:24
  that.
• 29:24
  The
• 29:24
  second
• 29:24
  thing
• 29:24
  was
• 29:25
  this
• 29:26
  idea
• 29:26
  of
• 29:26
  just
• 29:26
  having
• 29:26
  a
• 29:27
  coach.
• 29:28
  There's
• 29:28
  this
• 29:28
  very
• 29:29
  soft
• 29:30
  aspect,
• 29:30
  like
• 29:30
  how
• 29:31
  do
• 29:31
  you
• 29:31
  interact?
• 29:31
  How
• 29:31
  should
• 29:32
  I
• 29:32
  feel?
• 29:33
  And
• 29:33
  then
• 29:33
  there's
• 29:33
  a
• 29:33
  very
• 29:33
  hard
• 29:34
  aspect
• 29:34
  of
• 29:34
  like,
• 29:34
  what
• 29:35
  editor
• 29:35
  should
• 29:35
  I
• 29:35
  be
• 29:35
  using?
• 29:36
  What
• 29:36
  is
• 29:36
  the
• 29:36
  tool
• 29:37
  I
• 29:37
  should
• 29:37
  use?
• 29:37
  And
• 29:37
  it's
• 29:38
  like
• 29:38
  the
• 29:38
  key
• 29:38
  is
• 29:38
  that
• 29:40
  those
• 29:40
  people
• 29:40
  know
• 29:41
  all
• 29:41
  these
• 29:41
  things.
• 29:42
  Just
• 29:42
  real
• 29:42
  quick,
• 29:43
  if
• 29:43
  you
• 29:43
  hear
• 29:43
  any
• 29:43
  weird
• 29:43
  noises,
• 29:44
  we
• 29:44
  have
• 29:44
  a
• 29:44
  lot
• 29:44
  of
• 29:45
  things
• 29:45
  happening.
• 29:45
  Like
• 29:45
  there
• 29:46
  is
• 29:46
  a
• 29:47
  little
• 29:47
  unitree
• 29:48
  robot
• 29:49
  dog
• 29:49
  over
• 29:49
  here
• 29:49
  that
• 29:50
  is
• 29:50
  making
• 29:51
  noises.
• 29:51
  It's
• 29:52
  amazing.
• 29:52
  It
• 29:53
  just
• 29:53
  made
• 29:53
  it.
• 29:53
  Like
• 29:53
  everyone
• 29:53
  just
• 29:54
  walked
• 29:54
  off.
• 29:54
  I
• 29:54
  don't
• 29:54
  know
• 29:54
  why.
• 29:54
  And
• 29:55
  it
• 29:55
  made
• 29:55
  a
• 29:55
  noise.
• 29:55
  And
• 29:55
  I'm
• 29:56
  just
• 29:56
  like,
• 29:56
  this
• 29:56
  is
• 29:56
  concerning.
• 29:57
  That's
• 29:57
  perfect
• 29:57
  context
• 29:58
  and
• 29:58
  ambient
• 29:59
  noise
• 29:59
  for
• 30:00
  the
• 30:00
  theme.
• 30:00
  Yeah.
• 30:01
  And
• 30:01
  there's
• 30:01
  also
• 30:01
  an
• 30:01
  agent
• 30:01
  for
• 30:02
  the
• 30:02
  office
• 30:02
  that
• 30:02
  manages
• 30:03
  things.
• 30:03
  So
• 30:03
  if
• 30:03
  it,
• 30:03
  if
• 30:03
  it
• 30:04
  runs
• 30:04
  by
• 30:04
  or
• 30:04
  somehow
• 30:05
  I
• 30:05
  die
• 30:05
  in
• 30:05
  this.
• 30:06
  But
• 30:06
  what,
• 30:06
  but,
• 30:06
  but
• 30:07
  the
• 30:07
  intern
• 30:07
  program
• 30:07
  was
• 30:08
  very
• 30:08
  simple.
• 30:08
  Like
• 30:08
  basically
• 30:09
  what
• 30:09
  it
• 30:09
  was
• 30:09
  is
• 30:09
  I
• 30:09
  was
• 30:10
  like,
• 30:10
  can
• 30:10
  we
• 30:10
  find
• 30:10
  some
• 30:10
  interns
• 30:11
  that
• 30:11
  are
• 30:11
  smart
• 30:12
  people?
• 30:12
  Computer
• 30:13
  sciencey
• 30:13
  people,
• 30:13
  but
• 30:13
  not
• 30:14
  at
• 30:14
  all
• 30:15
  like
• 30:16
  senior
• 30:17
  engineers,
• 30:18
  just
• 30:18
  like
• 30:18
  normal
• 30:19
  kind
• 30:19
  of
• 30:19
  computer
• 30:19
  science
• 30:19
  people.
• 30:20
  And
• 30:20
  then
• 30:22
  if
• 30:22
  we
• 30:22
  paired
• 30:23
  them
• 30:23
  with
• 30:23
  a
• 30:23
  coach,
• 30:24
  me
• 30:24
  or
• 30:24
  like
• 30:24
  my
• 30:24
  friend
• 30:25
  Clint
• 30:25
  or
• 30:25
  Dylan
• 30:26
  or
• 30:26
  one
• 30:26
  of
• 30:26
  the
• 30:26
  people
• 30:26
  on
• 30:26
  our
• 30:26
  team
• 30:27
  who
• 30:27
  just
• 30:28
  has,
• 30:28
  you
• 30:28
  know,
• 30:28
  years
• 30:28
  of
• 30:28
  experience
• 30:29
  and
• 30:29
  we
• 30:29
  give
• 30:30
  them
• 30:30
  cloud
• 30:30
  code
• 30:31
  and
• 30:31
  we're
• 30:31
  looking
• 30:31
  for
• 30:31
  probably
• 30:32
  people
• 30:32
  who
• 30:32
  have
• 30:32
  somewhat
• 30:33
  high
• 30:33
  agency.
• 30:33
  And
• 30:34
  then
• 30:34
  there's
• 30:34
  this
• 30:34
  other
• 30:34
  aspect
• 30:35
  that
• 30:35
  we've
• 30:35
  been
• 30:35
  trying
• 30:35
  to
• 30:35
  figure
• 30:35
  how
• 30:36
  to
• 30:36
  describe,
• 30:36
  which
• 30:36
  is
• 30:36
  someone
• 30:36
  that
• 30:37
  can,
• 30:37
  that
• 30:37
  knows
• 30:38
  when
• 30:38
  they're
• 30:38
  underwater,
• 30:39
  which
• 30:39
  I
• 30:39
  think
• 30:39
  is
• 30:40
  a
• 30:40
  very
• 30:40
  specific
• 30:40
  skill
• 30:41
  set,
• 30:41
  which
• 30:42
  can
• 30:42
  raise
• 30:42
  their
• 30:42
  hand
• 30:43
  and
• 30:43
  say
• 30:43
  I'm
• 30:43
  stuck.
• 30:44
  And
• 30:44
  I
• 30:45
  don't
• 30:45
  think
• 30:45
  coachable
• 30:46
  in
• 30:46
  that
• 30:46
  bullshit
• 30:47
  phrase
• 30:47
  matters
• 30:49
  because
• 30:49
  it's
• 30:49
  not.
• 30:49
  You're
• 30:49
  not
• 30:50
  looking
• 30:50
  to
• 30:50
  teach
• 30:50
  people,
• 30:50
  you're
• 30:51
  just
• 30:51
  saying,
• 30:51
  look,
• 30:51
  just
• 30:52
  go.
• 30:52
  So
• 30:52
  we
• 30:52
  put
• 30:52
  these
• 30:53
  people
• 30:53
  and
• 30:53
  we
• 30:53
  gave
• 30:53
  them
• 30:53
  independent
• 30:54
  projects
• 30:54
  that
• 30:54
  were
• 30:55
  like
• 30:55
  at
• 30:55
  the
• 30:56
  same
• 30:56
  level
• 30:56
  as
• 30:56
  the
• 30:56
  projects
• 30:56
  we
• 30:56
  were
• 30:57
  doing.
• 30:57
  And
• 30:57
  they
• 30:57
  all
• 30:57
  were
• 30:57
  successful.
• 30:58
  Like
• 30:58
  they
• 30:58
  were
• 30:58
  fine,
• 30:59
  they
• 30:59
  were
• 30:59
  good,
• 30:59
  they
• 31:00
  were
• 31:00
  production
• 31:00
  level.
• 31:01
  Truly.
• 31:02
  They
• 31:02
  were
• 31:03
  absolutely
• 31:04
  the
• 31:04
  same
• 31:04
  as
• 31:04
  quality
• 31:05
  I
• 31:05
  would
• 31:05
  expect
• 31:05
  for
• 31:05
  all
• 31:05
  of
• 31:05
  us.
• 31:06
  There
• 31:06
  was
• 31:06
  no
• 31:06
  difference.
• 31:07
  This
• 31:07
  was
• 31:07
  not
• 31:07
  in
• 31:07
  turn
• 31:08
  quality
• 31:08
  projects.
• 31:08
  These
• 31:08
  were
• 31:09
  just
• 31:09
  normal,
• 31:09
  normal.
• 31:09
  Last
• 31:09
  projects
• 31:10
  that
• 31:10
  came
• 31:10
  out
• 31:10
  of
• 31:10
  this
• 31:11
  and
• 31:11
  this
• 31:11
  was
• 31:12
  Summertime
• 31:13
  in
• 31:13
  AI
• 31:13
  years.
• 31:14
  That
• 31:14
  was
• 31:14
  like
• 31:14
  two
• 31:14
  years
• 31:14
  ago,
• 31:15
  it.
• 31:15
  Was
• 31:15
  like
• 31:15
  four
• 31:16
  years
• 31:16
  ago
• 31:16
  in
• 31:16
  AI
• 31:16
  years.
• 31:16
  And
• 31:17
  so
• 31:17
  one
• 31:17
  of
• 31:17
  the
• 31:17
  people
• 31:18
  on
• 31:18
  my
• 31:18
  team,
• 31:19
  this
• 31:19
  woman
• 31:19
  named
• 31:19
  Sophie,
• 31:20
  she's
• 31:20
  great,
• 31:20
  she's
• 31:21
  started
• 31:21
  kind
• 31:22
  of
• 31:22
  helping
• 31:22
  us
• 31:22
  with
• 31:22
  some
• 31:22
  ops,
• 31:23
  some
• 31:23
  product
• 31:23
  ops
• 31:24
  stuff.
• 31:24
  And
• 31:25
  there
• 31:25
  was
• 31:26
  a
• 31:26
  project
• 31:26
  that
• 31:27
  they
• 31:27
  were
• 31:27
  experimenting
• 31:28
  with
• 31:28
  and
• 31:28
  it
• 31:28
  was
• 31:28
  this
• 31:29
  translation
• 31:30
  script
• 31:30
  I
• 31:30
  had
• 31:30
  made,
• 31:31
  which
• 31:31
  is
• 31:31
  like
• 31:31
  a
• 31:31
  small
• 31:31
  multi
• 31:32
  agent
• 31:32
  translator
• 31:33
  that
• 31:33
  would
• 31:33
  take
• 31:33
  and
• 31:33
  do
• 31:33
  double
• 31:34
  checks
• 31:34
  and
• 31:34
  critique
• 31:34
  and
• 31:35
  double
• 31:35
  check
• 31:35
  and
• 31:35
  critique
• 31:35
  to
• 31:36
  actually
• 31:36
  do
• 31:36
  some
• 31:36
  pretty
• 31:36
  good
• 31:37
  translation
• 31:37
  for
• 31:37
  content.
• 31:38
  We
• 31:38
  were
• 31:38
  doing
• 31:38
  some
• 31:39
  multilingual
• 31:39
  content
• 31:40
  and
• 31:41
  they
• 31:41
  were
• 31:41
  like,
• 31:42
  I
• 31:42
  kind
• 31:42
  of
• 31:42
  want
• 31:42
  to
• 31:42
  write
• 31:42
  a
• 31:42
  blog
• 31:43
  post
• 31:43
  about
• 31:43
  this,
• 31:43
  but
• 31:43
  I
• 31:43
  want
• 31:44
  to
• 31:44
  know
• 31:44
  who
• 31:44
  we
• 31:44
  make
• 31:44
  it
• 31:44
  a
• 31:45
  product.
• 31:45
  And
• 31:45
  I
• 31:46
  was
• 31:46
  just
• 31:46
  like,
• 31:46
  why
• 31:46
  don't
• 31:46
  you
• 31:46
  just
• 31:46
  make
• 31:47
  it
• 31:47
  a
• 31:47
  product?
• 31:48
  And
• 31:48
  so
• 31:49
  48
• 31:50
  hours
• 31:50
  later,
• 31:50
  it's
• 31:51
  a
• 31:51
  product
• 31:51
  that
• 31:52
  you
• 31:52
  can
• 31:52
  use.
• 31:52
  It's
• 31:53
  on
• 31:53
  netlify
• 31:53
  with
• 31:53
  a
• 31:53
  backend
• 31:54
  on
• 31:54
  fly.
• 31:55
  We
• 31:55
  did
• 31:55
  not
• 31:55
  give
• 31:55
  her
• 31:56
  any
• 31:56
  of
• 31:56
  those
• 31:56
  instructions.
• 31:57
  It
• 31:57
  has
• 31:57
  a
• 31:57
  stripe
• 31:57
  account
• 31:58
  hooked
• 31:58
  into
• 31:58
  it.
• 31:58
  This
• 31:58
  is
• 31:58
  not
• 31:59
  a
• 31:59
  programmer
• 32:00
  and
• 32:00
  it's
• 32:00
  not
• 32:00
  bad.
• 32:01
  If
• 32:01
  we
• 32:01
  were
• 32:01
  to
• 32:01
  release
• 32:02
  it,
• 32:02
  we
• 32:02
  would
• 32:02
  do
• 32:02
  literally
• 32:03
  nothing.
• 32:04
  We
• 32:04
  would
• 32:04
  just
• 32:04
  be
• 32:04
  like,
• 32:04
  oh,
• 32:05
  there
• 32:05
  it
• 32:05
  is,
• 32:05
  here's
• 32:05
  the
• 32:05
  domain.
• 32:07
  But
• 32:07
  there's
• 32:07
  a
• 32:07
  couple
• 32:07
  things
• 32:07
  I
• 32:08
  Learned
• 32:08
  from
• 32:08
  that.
• 32:08
  1.
• 32:09
  I
• 32:10
  don't
• 32:10
  think
• 32:10
  there
• 32:10
  ever
• 32:10
  has
• 32:11
  been
• 32:11
  a.
• 32:12
  This
• 32:12
  is
• 32:13
  production
• 32:13
  code
• 32:14
  versus
• 32:14
  non
• 32:14
  production
• 32:15
  code.
• 32:15
  I
• 32:15
  think
• 32:15
  that's
• 32:15
  typically
• 32:16
  a
• 32:16
  delineator
• 32:17
  that
• 32:17
  is
• 32:17
  made
• 32:18
  by
• 32:18
  people
• 32:18
  like
• 32:18
  me
• 32:18
  to
• 32:19
  keep
• 32:19
  people
• 32:19
  that
• 32:20
  are
• 32:20
  trying
• 32:20
  to
• 32:20
  get
• 32:20
  into
• 32:20
  the
• 32:20
  industry
• 32:21
  out
• 32:21
  of
• 32:21
  the
• 32:21
  industry.
• 32:22
  I'll
• 32:22
  just
• 32:22
  be
• 32:22
  like,
• 32:22
  well,
• 32:23
  that's
• 32:23
  just
• 32:23
  not
• 32:23
  production.
• 32:24
  And
• 32:24
  it's
• 32:24
  like,
• 32:24
  bullshit.
• 32:25
  We've
• 32:25
  worked
• 32:25
  at
• 32:25
  big
• 32:26
  companies,
• 32:26
  it's
• 32:27
  all
• 32:27
  bullshit.
• 32:27
  Fucking
• 32:28
  bubblegum
• 32:29
  and
• 32:29
  bailing
• 32:29
  wire
• 32:30
  code.
• 32:30
  And
• 32:30
  so
• 32:30
  we
• 32:30
  create
• 32:31
  all
• 32:31
  these
• 32:31
  barriers
• 32:32
  like,
• 32:32
  oh,
• 32:32
  well,
• 32:32
  this
• 32:33
  is
• 32:33
  not
• 32:33
  production
• 32:33
  grade
• 32:34
  or
• 32:34
  you
• 32:35
  need
• 32:35
  to
• 32:36
  senior
• 32:36
  to
• 32:36
  get
• 32:36
  this
• 32:36
  done.
• 32:36
  But
• 32:37
  then
• 32:37
  for
• 32:37
  some
• 32:37
  reason,
• 32:38
  all
• 32:38
  these
• 32:38
  great
• 32:38
  companies
• 32:39
  come
• 32:40
  from
• 32:40
  people
• 32:40
  who
• 32:40
  are
• 32:40
  like
• 32:40
  24.
• 32:42
  I'm
• 32:42
  confused
• 32:42
  how
• 32:43
  this
• 32:43
  happens,
• 32:43
  how
• 32:43
  you
• 32:44
  can
• 32:44
  build
• 32:44
  a
• 32:45
  life
• 32:45
  changing
• 32:45
  company
• 32:45
  when
• 32:46
  you're
• 32:46
  24,
• 32:46
  but
• 32:46
  you
• 32:47
  have
• 32:47
  to
• 32:47
  be
• 32:47
  a
• 32:48
  senior
• 32:48
  engineer
• 32:49
  with
• 32:49
  35
• 32:49
  years
• 32:49
  of
• 32:49
  experience
• 32:50
  to
• 32:50
  be
• 32:50
  a
• 32:50
  hire.
• 32:51
  So
• 32:51
  I
• 32:52
  feel
• 32:52
  like
• 32:52
  we
• 32:52
  lie
• 32:52
  to
• 32:52
  ourselves
• 32:53
  to
• 32:53
  try
• 32:53
  and
• 32:53
  create
• 32:56
  more
• 32:56
  insular
• 32:56
  experiences
• 32:57
  to
• 32:57
  try
• 32:57
  and
• 32:57
  keep
• 32:58
  people
• 32:58
  out
• 32:58
  more
• 32:58
  so
• 32:59
  than
• 32:59
  we
• 32:59
  create
• 32:59
  opportunities
• 33:00
  for
• 33:00
  people
• 33:00
  to
• 33:00
  come
• 33:00
  in.
• 33:01
  And
• 33:01
  so
• 33:01
  I
• 33:01
  think
• 33:02
  the
• 33:02
  thing
• 33:02
  that's
• 33:02
  really
• 33:02
  fucked
• 33:03
  up
• 33:03
  about
• 33:03
  AI,
• 33:04
  besides
• 33:04
  being
• 33:04
  existential
• 33:05
  for
• 33:05
  people
• 33:05
  like
• 33:06
  me,
• 33:06
  is
• 33:07
  that
• 33:07
  it
• 33:07
  literally
• 33:07
  doesn't
• 33:08
  care.
• 33:08
  Like
• 33:09
  your
• 33:09
  success
• 33:09
  is
• 33:09
  the
• 33:10
  amount
• 33:10
  is
• 33:10
  related
• 33:10
  to
• 33:10
  the
• 33:10
  amount
• 33:11
  of
• 33:11
  patience
• 33:11
  and
• 33:12
  interest
• 33:12
  you
• 33:13
  have
• 33:13
  in
• 33:13
  getting
• 33:13
  this
• 33:13
  done.
• 33:14
  Like
• 33:15
  it
• 33:15
  doesn't
• 33:15
  matter
• 33:15
  at
• 33:16
  all
• 33:16
  if
• 33:16
  you
• 33:16
  know
• 33:16
  anything.
• 33:17
  I
• 33:17
  mean,
• 33:17
  so
• 33:17
  I.
• 33:17
  The
• 33:18
  second
• 33:18
  challenge
• 33:18
  I
• 33:18
  did
• 33:18
  was
• 33:19
  I
• 33:19
  was
• 33:19
  like,
• 33:19
  I'm
• 33:19
  gonna
• 33:19
  do
• 33:19
  everything
• 33:19
  in
• 33:20
  Rust.
• 33:20
  Which
• 33:21
  I
• 33:21
  think
• 33:21
  that
• 33:21
  was
• 33:21
  because
• 33:21
  I
• 33:21
  had
• 33:22
  7
• 33:22
  million
• 33:22
  years
• 33:23
  to
• 33:23
  waste
• 33:23
  compiling.
• 33:24
  So
• 33:25
  I
• 33:25
  did
• 33:25
  everything
• 33:25
  in
• 33:25
  Rust.
• 33:25
  I
• 33:25
  don't
• 33:25
  know
• 33:25
  anything
• 33:26
  about
• 33:26
  Rust
• 33:26
  then.
• 33:27
  Now
• 33:27
  I
• 33:27
  do
• 33:27
  everything
• 33:27
  in
• 33:27
  Go
• 33:28
  same.
• 33:28
  Like
• 33:28
  I've
• 33:29
  never
• 33:29
  done
• 33:29
  Go
• 33:29
  in
• 33:29
  my
• 33:29
  life.
• 33:30
  Like,
• 33:30
  it's
• 33:30
  great.
• 33:30
  I
• 33:31
  do
• 33:31
  fucking
• 33:31
  Swift
• 33:31
  all
• 33:32
  the
• 33:32
  time.
• 33:32
  I
• 33:32
  don't
• 33:32
  know
• 33:32
  Swift.
• 33:33
  I
• 33:33
  don't
• 33:34
  know
• 33:34
  what
• 33:34
  I'm
• 33:34
  doing
• 33:34
  with
• 33:34
  my
• 33:34
  life.
• 33:34
  I
• 33:34
  don't
• 33:35
  even
• 33:35
  do
• 33:35
  Python.
• 33:35
  I
• 33:35
  was
• 33:35
  really
• 33:36
  good
• 33:36
  at
• 33:36
  Python.
• 33:36
  I
• 33:37
  really
• 33:37
  liked
• 33:37
  doing
• 33:37
  things
• 33:37
  in
• 33:37
  Python.
• 33:38
  I
• 33:38
  don't
• 33:38
  even
• 33:38
  do
• 33:38
  it
• 33:38
  anymore
• 33:39
  because
• 33:39
  it
• 33:39
  was
• 33:39
  too
• 33:39
  easy.
• 33:40
  I
• 33:40
  don't
• 33:40
  even
• 33:40
  do
• 33:41
  any
• 33:41
  work.
• 33:41
  So
• 33:42
  this
• 33:42
  is
• 33:42
  the
• 33:42
  existential
• 33:42
  part.
• 33:45
  It
• 33:46
  really
• 33:46
  is
• 33:46
  challenging
• 33:46
  because
• 33:47
  I
• 33:47
  see
• 33:47
  myself
• 33:50
  basically
• 33:50
  just
• 33:51
  chatting
• 33:51
  like
• 33:52
  I
• 33:52
  did
• 33:52
  on
• 33:52
  IRC
• 33:53
  in
• 33:53
  2003
• 33:54
  and
• 33:55
  2004
• 33:55
  and
• 33:56
  et
• 33:56
  cetera,
• 33:56
  while
• 33:56
  I
• 33:56
  was
• 33:56
  a
• 33:57
  Java
• 33:57
  programmer,
• 33:58
  being
• 33:58
  like,
• 33:58
  what
• 33:58
  are
• 33:58
  you
• 33:59
  going
• 33:59
  to
• 33:59
  do
• 33:59
  tonight?
• 33:59
  That's
• 34:00
  basically
• 34:00
  all
• 34:00
  we
• 34:00
  chatted
• 34:00
  about
• 34:01
  while
• 34:01
  the
• 34:02
  programs
• 34:02
  were
• 34:02
  doing
• 34:02
  things
• 34:02
  and
• 34:03
  you
• 34:03
  were
• 34:03
  typing
• 34:03
  into
• 34:03
  some
• 34:03
  editor.
• 34:04
  Now
• 34:05
  I'm
• 34:05
  in
• 34:05
  a
• 34:06
  terminal
• 34:06
  chatting
• 34:07
  with
• 34:07
  someone
• 34:08
  who
• 34:08
  calls
• 34:08
  me
• 34:09
  Dr.
• 34:09
  Biz.
• 34:10
  And
• 34:10
  that's
• 34:10
  it.
• 34:10
  That's
• 34:11
  all
• 34:11
  I
• 34:11
  do.
• 34:11
  And
• 34:11
  it's
• 34:12
  ridiculous.
• 34:12
  We're
• 34:13
  so
• 34:13
  productive.
• 34:13
  We
• 34:14
  are
• 34:14
  doing
• 34:14
  so
• 34:14
  much
• 34:14
  code.
• 34:15
  If
• 34:15
  we
• 34:15
  were
• 34:15
  counting
• 34:16
  lines
• 34:16
  of
• 34:16
  code,
• 34:16
  we
• 34:17
  would
• 34:17
  be
• 34:17
  beating
• 34:17
  our
• 34:18
  yearlies
• 34:19
  in
• 34:20
  just
• 34:21
  half
• 34:21
  a
• 34:21
  year.
• 34:21
  Of
• 34:21
  2025.
• 34:23
  It's
• 34:24
  bonkers.
• 34:25
  And
• 34:25
  I
• 34:26
  think
• 34:26
  we're
• 34:26
  all
• 34:26
  going
• 34:26
  to
• 34:26
  die.
• 34:27
  Well,
• 34:28
  inevitably
• 34:28
  we'll
• 34:28
  all
• 34:28
  die,
• 34:29
  but
• 34:29
  I
• 34:29
  think
• 34:29
  this
• 34:29
  is
• 34:29
  not
• 34:29
  going
• 34:29
  to
• 34:29
  help
• 34:30
  us.
• 34:31
  Yeah,
• 34:31
  I
• 34:32
  feel
• 34:32
  like
• 34:33
  to
• 34:33
  your
• 34:33
  point,
• 34:34
  the
• 34:34
  quantity
• 34:34
  of
• 34:34
  code
• 34:34
  is
• 34:35
  just
• 34:35
  wild
• 34:36
  what
• 34:36
  you're
• 34:36
  talking
• 34:36
  about
• 34:37
  with
• 34:37
  the
• 34:37
  gatekeeping.
• 34:38
  I've
• 34:38
  been
• 34:38
  really
• 34:39
  thinking
• 34:39
  about
• 34:39
  this.
• 34:40
  Last
• 34:40
  week
• 34:40
  I
• 34:40
  was
• 34:41
  having
• 34:41
  a
• 34:41
  conversation
• 34:41
  with
• 34:41
  someone
• 34:42
  that
• 34:42
  clearly
• 34:43
  had
• 34:43
  been
• 34:43
  a
• 34:44
  engineer
• 34:44
  for
• 34:44
  a
• 34:45
  long
• 34:45
  time.
• 34:45
  And
• 34:45
  I
• 34:45
  feel
• 34:46
  like
• 34:46
  what
• 34:46
  you
• 34:46
  said
• 34:46
  is,
• 34:47
  you're
• 34:47
  latest
• 34:48
  and
• 34:48
  greatest,
• 34:48
  let's
• 34:49
  do
• 34:49
  it.
• 34:49
  And
• 34:49
  I'm
• 34:50
  sure
• 34:50
  you've
• 34:50
  seen
• 34:50
  the
• 34:50
  spectrum
• 34:51
  of.
• 34:51
  There's
• 34:51
  that
• 34:52
  side
• 34:52
  of
• 34:52
  the
• 34:52
  spectrum
• 34:52
  and
• 34:52
  then
• 34:52
  there's
• 34:53
  the
• 34:53
  other
• 34:53
  side
• 34:53
  of
• 34:53
  the
• 34:53
  people
• 34:54
  that
• 34:54
  are
• 34:54
  like,
• 34:54
  there's
• 34:55
  engineers
• 34:55
  out
• 34:55
  there
• 34:55
  that
• 34:55
  are
• 34:56
  like,
• 34:56
  I
• 34:56
  won't
• 34:56
  use
• 34:56
  AI.
• 34:56
  I
• 34:57
  don't
• 34:57
  trust
• 34:57
  AI.
• 34:57
  And
• 34:57
  they're
• 34:58
  always
• 34:59
  boned.
• 34:59
  They're
• 35:00
  so
• 35:00
  boned.
• 35:00
  They
• 35:00
  probably
• 35:01
  also
• 35:01
  didn't
• 35:01
  use
• 35:02
  like,
• 35:02
  I
• 35:02
  don't
• 35:02
  know,
• 35:03
  like
• 35:03
  anything
• 35:03
  at
• 35:03
  some
• 35:04
  point.
• 35:04
  Like,
• 35:04
  I
• 35:04
  don't
• 35:04
  know,
• 35:04
  I
• 35:04
  think,
• 35:05
  well,
• 35:05
  there's
• 35:05
  a
• 35:05
  thing
• 35:05
  which
• 35:06
  is
• 35:06
  like,
• 35:06
  I
• 35:06
  didn't
• 35:06
  use
• 35:06
  Ruby
• 35:07
  on
• 35:07
  Rails,
• 35:07
  man.
• 35:07
  I've
• 35:07
  been
• 35:07
  in,
• 35:07
  I've
• 35:08
  been
• 35:08
  documented
• 35:08
  since
• 35:08
  2007
• 35:09
  talking
• 35:09
  shit
• 35:09
  about
• 35:09
  Ruby
• 35:10
  on
• 35:10
  Rails.
• 35:10
  But
• 35:10
  that's
• 35:10
  just
• 35:10
  because
• 35:11
  I
• 35:11
  like,
• 35:11
  I
• 35:12
  like
• 35:12
  good
• 35:12
  things,
• 35:12
  not
• 35:13
  bad
• 35:13
  things.
• 35:13
  But
• 35:14
  I
• 35:14
  think
• 35:14
  you're
• 35:14
  right.
• 35:14
  There's
• 35:15
  always
• 35:15
  people
• 35:15
  who
• 35:16
  are
• 35:16
  like,
• 35:16
  I'm
• 35:17
  not
• 35:17
  going
• 35:17
  to
• 35:17
  use
• 35:17
  this.
• 35:18
  And
• 35:18
  I
• 35:18
  think
• 35:18
  AI,
• 35:19
  especially
• 35:19
  Cogen
• 35:21
  and
• 35:21
  I
• 35:21
  kind
• 35:21
  of
• 35:21
  refuse
• 35:22
  to
• 35:22
  say
• 35:22
  vibe
• 35:22
  coding,
• 35:23
  but
• 35:23
  I
• 35:23
  think
• 35:23
  there's
• 35:23
  some
• 35:23
  benefit
• 35:24
  to
• 35:24
  calling
• 35:24
  it
• 35:24
  that
• 35:24
  as
• 35:25
  well.
• 35:25
  But
• 35:25
  like
• 35:25
  the
• 35:25
  code
• 35:26
  gen
• 35:26
  stuff,
• 35:27
  in
• 35:27
  my
• 35:27
  experience,
• 35:28
  people
• 35:28
  who
• 35:28
  are
• 35:29
  really
• 35:29
  against
• 35:29
  it
• 35:30
  feel
• 35:30
  it,
• 35:31
  it's
• 35:31
  a
• 35:31
  threat
• 35:32
  to
• 35:32
  them.
• 35:33
  And
• 35:33
  that's
• 35:33
  because,
• 35:33
  not
• 35:34
  that
• 35:34
  they're
• 35:34
  good
• 35:34
  or
• 35:35
  bad,
• 35:35
  but
• 35:36
  because
• 35:36
  they
• 35:36
  see
• 35:36
  their
• 35:36
  job
• 35:37
  ultimately
• 35:38
  as
• 35:38
  replaceable.
• 35:39
  I
• 35:39
  think
• 35:39
  that's
• 35:40
  really
• 35:40
  the
• 35:40
  issue
• 35:40
  is
• 35:41
  that
• 35:41
  when
• 35:41
  I
• 35:41
  look
• 35:41
  at
• 35:42
  who
• 35:42
  if
• 35:43
  the
• 35:43
  three
• 35:43
  of
• 35:43
  us
• 35:43
  started
• 35:43
  a
• 35:44
  company,
• 35:44
  convinced
• 35:45
  Becca
• 35:45
  to
• 35:46
  fund
• 35:46
  us,
• 35:46
  that'd
• 35:47
  be
• 35:47
  fun.
• 35:47
  Fund
• 35:47
  us.
• 35:49
  If
• 35:49
  we
• 35:49
  started
• 35:49
  a
• 35:49
  company
• 35:50
  and
• 35:50
  we
• 35:50
  went
• 35:50
  out
• 35:50
  there
• 35:51
  to
• 35:51
  hire
• 35:51
  people,
• 35:51
  we're
• 35:52
  not
• 35:52
  hiring
• 35:53
  mid
• 35:53
  level
• 35:53
  developers.
• 35:55
  There's
• 35:55
  no
• 35:55
  jobs
• 35:55
  anymore
• 35:56
  for
• 35:56
  that
• 35:56
  person
• 35:56
  because
• 35:57
  they're
• 35:58
  expensive,
• 35:58
  they
• 35:58
  require
• 35:59
  more
• 35:59
  money
• 36:00
  and
• 36:00
  they
• 36:00
  just
• 36:00
  don't
• 36:00
  have
• 36:01
  as
• 36:01
  much
• 36:01
  benefit
• 36:01
  as
• 36:02
  a
• 36:02
  young
• 36:03
  developer.
• 36:04
  And
• 36:04
  so
• 36:04
  you
• 36:04
  have
• 36:05
  this
• 36:05
  fucked
• 36:05
  up
• 36:05
  thing.
• 36:05
  When
• 36:06
  you
• 36:06
  start
• 36:06
  litigating
• 36:07
  the
• 36:07
  view
• 36:07
  through
• 36:07
  capitalism,
• 36:09
  then
• 36:09
  you
• 36:09
  stop
• 36:09
  hiring
• 36:10
  mid
• 36:10
  and
• 36:10
  senior
• 36:11
  people
• 36:11
  and
• 36:11
  you
• 36:11
  only
• 36:12
  hire
• 36:12
  cheap,
• 36:12
  younger
• 36:13
  developers.
• 36:13
  They
• 36:14
  don't
• 36:14
  have
• 36:14
  to
• 36:14
  be
• 36:14
  young.
• 36:14
  It's
• 36:15
  not
• 36:15
  about
• 36:15
  age,
• 36:15
  it's
• 36:15
  more
• 36:15
  about
• 36:16
  how
• 36:16
  much
• 36:16
  they're
• 36:17
  demanding
• 36:17
  and
• 36:17
  someone
• 36:17
  with.
• 36:17
  Less
• 36:18
  experience
• 36:18
  or
• 36:18
  it
• 36:18
  pushes
• 36:18
  you
• 36:19
  like
• 36:19
  both
• 36:20
  Ends
• 36:20
  of
• 36:20
  the
• 36:20
  spectrum
• 36:20
  too,
• 36:21
  where
• 36:22
  you
• 36:22
  spend
• 36:23
  the
• 36:24
  vast
• 36:24
  majority
• 36:25
  of
• 36:25
  your
• 36:25
  capital
• 36:25
  trying
• 36:26
  to
• 36:26
  acquire
• 36:26
  like
• 36:26
  top
• 36:27
  talent,
• 36:27
  like
• 36:28
  finding
• 36:28
  the.
• 36:28
  I
• 36:28
  just
• 36:28
  don't.
• 36:29
  I
• 36:29
  think
• 36:29
  that's
• 36:29
  a
• 36:29
  mistake.
• 36:30
  I
• 36:30
  think
• 36:30
  that's
• 36:31
  a
• 36:31
  horrible
• 36:31
  mistake.
• 36:31
  Because
• 36:32
  I
• 36:32
  think
• 36:32
  the
• 36:32
  problem
• 36:32
  is,
• 36:32
  is
• 36:33
  that
• 36:33
  if
• 36:33
  you
• 36:34
  are
• 36:34
  just
• 36:34
  laying
• 36:35
  out
• 36:35
  how
• 36:35
  much
• 36:35
  productivity
• 36:36
  you're
• 36:36
  going
• 36:36
  to
• 36:36
  get,
• 36:37
  get
• 36:37
  one
• 36:37
  of
• 36:37
  those
• 36:38
  people,
• 36:38
  then
• 36:38
  get
• 36:39
  10
• 36:39
  juniors,
• 36:41
  give
• 36:41
  them
• 36:41
  cloud
• 36:41
  code
• 36:42
  and
• 36:42
  have
• 36:42
  that
• 36:42
  one
• 36:42
  person
• 36:42
  yet,
• 36:43
  like
• 36:43
  be
• 36:43
  an
• 36:43
  asshole.
• 36:44
  What
• 36:44
  are
• 36:44
  they
• 36:44
  going
• 36:44
  to
• 36:44
  do?
• 36:44
  They're
• 36:44
  going
• 36:44
  to
• 36:44
  sit
• 36:44
  there
• 36:45
  and
• 36:45
  be
• 36:45
  miserable.
• 36:45
  They're
• 36:45
  going
• 36:45
  to
• 36:46
  be
• 36:46
  miserable
• 36:46
  wherever
• 36:46
  they
• 36:46
  are.
• 36:47
  I
• 36:47
  would
• 36:48
  rather
• 36:48
  work
• 36:48
  with
• 36:48
  10
• 36:48
  people
• 36:49
  who
• 36:49
  have
• 36:49
  no
• 36:49
  idea
• 36:49
  what
• 36:49
  they're
• 36:49
  doing
• 36:50
  with
• 36:50
  cloud
• 36:50
  code
• 36:51
  just
• 36:51
  trying
• 36:51
  to
• 36:51
  figure
• 36:51
  it
• 36:51
  out,
• 36:52
  than
• 36:52
  two
• 36:52
  people
• 36:53
  who
• 36:53
  are
• 36:53
  like
• 36:54
  10x
• 36:54
  developers.
• 36:55
  Yeah.
• 37:01
  Do
• 37:02
  you
• 37:02
  want
• 37:02
  to
• 37:02
  talk
• 37:02
  a
• 37:02
  little
• 37:03
  bit
• 37:03
  about
• 37:04
  what
• 37:04
  you're
• 37:05
  building?
• 37:05
  Yeah.
• 37:06
  First
• 37:06
  of
• 37:06
  all,
• 37:06
  obviously
• 37:07
  2389
• 37:07
  is
• 37:07
  an
• 37:08
  early
• 37:08
  stage
• 37:08
  company.
• 37:09
  We
• 37:09
  started
• 37:09
  with
• 37:10
  this
• 37:10
  very
• 37:10
  specific
• 37:10
  idea
• 37:11
  about
• 37:11
  how
• 37:11
  to
• 37:11
  kind
• 37:12
  of
• 37:12
  change
• 37:12
  workflow.
• 37:13
  And
• 37:14
  when
• 37:14
  we
• 37:14
  started
• 37:15
  thinking
• 37:15
  about
• 37:16
  this,
• 37:16
  one
• 37:16
  of
• 37:16
  the
• 37:16
  things
• 37:16
  that
• 37:16
  we
• 37:16
  noticed
• 37:17
  was
• 37:17
  most
• 37:17
  of
• 37:17
  the
• 37:18
  tools
• 37:18
  that
• 37:18
  we
• 37:18
  saw
• 37:19
  were
• 37:19
  not.
• 37:21
  You
• 37:21
  couldn't
• 37:21
  add
• 37:22
  others
• 37:22
  into
• 37:22
  them.
• 37:23
  Like
• 37:23
  a
• 37:23
  lot
• 37:24
  of
• 37:24
  the
• 37:24
  AI
• 37:25
  kind
• 37:25
  of
• 37:25
  experiences
• 37:26
  have
• 37:26
  been
• 37:26
  singular.
• 37:27
  A
• 37:27
  person
• 37:27
  interacts
• 37:28
  with
• 37:28
  a
• 37:28
  thing
• 37:29
  and
• 37:29
  you
• 37:29
  see
• 37:29
  people
• 37:30
  doing
• 37:30
  experiments
• 37:30
  like
• 37:31
  the
• 37:31
  OpenAI
• 37:31
  group
• 37:31
  chats,
• 37:32
  things
• 37:33
  like
• 37:33
  that,
• 37:33
  which
• 37:33
  are
• 37:33
  just,
• 37:33
  okay,
• 37:35
  it's
• 37:35
  not
• 37:36
  super
• 37:36
  accessible,
• 37:36
  it's
• 37:37
  not
• 37:37
  very
• 37:37
  interesting.
• 37:38
  But
• 37:39
  we
• 37:39
  kind
• 37:39
  of
• 37:39
  thought
• 37:39
  of
• 37:39
  this
• 37:40
  very
• 37:40
  simple
• 37:40
  and
• 37:41
  kind
• 37:41
  of
• 37:41
  fun
• 37:41
  idea
• 37:42
  of
• 37:42
  just
• 37:42
  like
• 37:42
  what
• 37:42
  happens
• 37:43
  if
• 37:43
  you
• 37:43
  had
• 37:43
  a
• 37:44
  workflow
• 37:44
  tool
• 37:44
  that
• 37:44
  allows
• 37:44
  you
• 37:45
  get
• 37:45
  work
• 37:45
  done,
• 37:46
  but
• 37:46
  you
• 37:46
  can
• 37:46
  add
• 37:46
  friends
• 37:47
  and
• 37:48
  work
• 37:48
  co
• 37:48
  workers,
• 37:48
  et
• 37:49
  cetera,
• 37:49
  both
• 37:49
  as
• 37:50
  kind
• 37:50
  of
• 37:50
  agents
• 37:51
  themselves,
• 37:51
  so
• 37:51
  they
• 37:51
  can
• 37:52
  get
• 37:52
  things
• 37:52
  done
• 37:52
  into
• 37:52
  the
• 37:52
  system
• 37:53
  and
• 37:53
  then
• 37:53
  give
• 37:53
  them
• 37:53
  the
• 37:54
  same
• 37:54
  tooling
• 37:54
  that
• 37:54
  the
• 37:54
  agents
• 37:55
  have.
• 37:55
  Because
• 37:55
  that's
• 37:55
  the
• 37:55
  other
• 37:55
  thing
• 37:55
  about
• 37:56
  some
• 37:56
  of
• 37:56
  this
• 37:56
  stuff
• 37:56
  is
• 37:56
  like,
• 37:57
  these
• 37:57
  agents
• 37:57
  have
• 37:57
  really
• 37:57
  good
• 37:58
  tools.
• 37:58
  People
• 37:59
  are
• 37:59
  like,
• 37:59
  I
• 37:59
  gave
• 37:59
  my
• 37:59
  agents
• 38:00
  all
• 38:00
  these
• 38:00
  tools
• 38:00
  and
• 38:01
  I'm
• 38:01
  kind
• 38:01
  of
• 38:01
  like,
• 38:01
  I
• 38:01
  also
• 38:02
  want
• 38:02
  some
• 38:02
  of
• 38:02
  those
• 38:02
  tools.
• 38:03
  I
• 38:03
  would
• 38:04
  love
• 38:04
  to
• 38:04
  be
• 38:04
  able
• 38:04
  to
• 38:04
  just
• 38:04
  like
• 38:04
  run
• 38:05
  a
• 38:05
  function
• 38:05
  and
• 38:05
  get
• 38:05
  all
• 38:06
  of
• 38:06
  this
• 38:06
  information
• 38:06
  for
• 38:07
  myself
• 38:07
  to
• 38:07
  do
• 38:08
  this
• 38:08
  stuff.
• 38:08
  Like,
• 38:09
  there's
• 38:09
  so
• 38:09
  much
• 38:10
  benefit
• 38:10
  of
• 38:10
  doing
• 38:10
  knowledge
• 38:11
  graphs
• 38:11
  inside
• 38:12
  of
• 38:12
  some
• 38:12
  of
• 38:12
  these
• 38:12
  systems.
• 38:13
  Whether
• 38:13
  it's
• 38:13
  like
• 38:13
  graph
• 38:13
  Rag
• 38:14
  from
• 38:14
  Microsoft
• 38:14
  or
• 38:14
  some
• 38:14
  of
• 38:15
  those
• 38:15
  other
• 38:15
  ideas.
• 38:15
  There's
• 38:15
  a
• 38:16
  lot
• 38:16
  of
• 38:16
  benefit.
• 38:16
  Can
• 38:16
  you.
• 38:16
  You
• 38:16
  get
• 38:16
  from
• 38:17
  some
• 38:17
  of
• 38:17
  that
• 38:17
  stuff,
• 38:18
  just
• 38:18
  from
• 38:18
  being
• 38:18
  able
• 38:18
  to
• 38:19
  say,
• 38:19
  like,
• 38:19
  yeah,
• 38:20
  this
• 38:20
  is
• 38:20
  the
• 38:20
  knowledge
• 38:21
  that
• 38:21
  I
• 38:21
  have.
• 38:21
  Like,
• 38:22
  I
• 38:22
  want
• 38:22
  to
• 38:22
  look
• 38:22
  up
• 38:22
  everything
• 38:22
  I
• 38:23
  know
• 38:23
  about
• 38:23
  Sarah
• 38:23
  Chips.
• 38:23
  Like
• 38:24
  we've
• 38:24
  known
• 38:24
  each
• 38:24
  other
• 38:25
  sporadically
• 38:25
  for
• 38:26
  quite
• 38:26
  a
• 38:26
  while
• 38:27
  now.
• 38:27
  And
• 38:27
  there's
• 38:27
  all
• 38:28
  these
• 38:28
  like
• 38:28
  little
• 38:28
  things
• 38:28
  that
• 38:28
  you
• 38:29
  would
• 38:29
  see,
• 38:29
  these
• 38:29
  little
• 38:29
  topics
• 38:30
  that
• 38:30
  would
• 38:30
  surface
• 38:31
  up
• 38:31
  there
• 38:31
  that
• 38:31
  might
• 38:31
  actually
• 38:32
  help
• 38:32
  various
• 38:33
  things
• 38:33
  happen.
• 38:33
  And
• 38:33
  so
• 38:34
  we
• 38:34
  spent
• 38:35
  a
• 38:35
  lot
• 38:35
  of
• 38:35
  time
• 38:35
  talking
• 38:35
  with
• 38:35
  Nathan
• 38:36
  Stoll,
• 38:36
  who
• 38:36
  is
• 38:37
  one
• 38:37
  of
• 38:37
  the
• 38:37
  founders
• 38:37
  of
• 38:37
  Aardvark,
• 38:38
  kind
• 38:38
  of
• 38:38
  building
• 38:39
  around
• 38:39
  this
• 38:39
  stuff,
• 38:39
  like
• 38:39
  how
• 38:40
  do
• 38:40
  you
• 38:40
  make
• 38:40
  human
• 38:41
  powered
• 38:41
  knowledge
• 38:42
  graphs?
• 38:42
  And
• 38:43
  then
• 38:43
  how
• 38:43
  do
• 38:43
  you
• 38:43
  start
• 38:44
  mixing
• 38:44
  in
• 38:44
  agents?
• 38:45
  And
• 38:45
  so
• 38:45
  that
• 38:45
  was
• 38:45
  kind
• 38:45
  of
• 38:46
  some
• 38:46
  of
• 38:46
  the
• 38:46
  start
• 38:46
  there.
• 38:47
  Basically
• 38:48
  what's
• 38:48
  happened
• 38:48
  is
• 38:48
  we've
• 38:48
  just
• 38:49
  been
• 38:49
  totally
• 38:49
  fucked
• 38:49
  by
• 38:50
  the
• 38:50
  fact
• 38:50
  that
• 38:50
  you
• 38:50
  can
• 38:50
  program
• 38:51
  anything
• 38:51
  you
• 38:51
  want
• 38:51
  in
• 38:52
  like
• 38:52
  48
• 38:52
  hours.
• 38:53
  And
• 38:53
  that
• 38:53
  like
• 38:54
  removes
• 38:54
  all
• 38:54
  conviction
• 38:55
  from
• 38:55
  any
• 38:55
  idea
• 38:56
  ever.
• 38:56
  And
• 38:57
  so
• 38:57
  we
• 38:57
  just
• 38:57
  kind
• 38:57
  of
• 38:57
  jumped
• 38:58
  in
• 38:58
  a
• 38:58
  pot
• 38:58
  and
• 38:58
  stirred
• 38:59
  ourselves
• 38:59
  to
• 38:59
  death.
• 38:59
  And
• 38:59
  so
• 38:59
  now
• 39:00
  we're
• 39:00
  just
• 39:00
  laying
• 39:00
  on
• 39:00
  the
• 39:00
  ground
• 39:00
  being
• 39:01
  like,
• 39:01
  what
• 39:01
  do
• 39:01
  we
• 39:01
  build?
• 39:02
  What
• 39:02
  do
• 39:02
  we
• 39:02
  build?
• 39:03
  To
• 39:03
  be
• 39:03
  fair,
• 39:03
  isn't
• 39:04
  that
• 39:04
  kind
• 39:04
  of
• 39:04
  like
• 39:05
  what
• 39:05
  it's
• 39:05
  always
• 39:05
  like
• 39:06
  when
• 39:06
  you're
• 39:06
  starting
• 39:06
  to
• 39:06
  build
• 39:07
  something
• 39:07
  from
• 39:07
  scratch?
• 39:08
  You
• 39:08
  started
• 39:08
  a
• 39:08
  company,
• 39:09
  Becca,
• 39:09
  you
• 39:09
  know
• 39:09
  how
• 39:09
  this
• 39:09
  is.
• 39:10
  Like
• 39:10
  there's
• 39:10
  never.
• 39:10
  I
• 39:11
  think
• 39:11
  the
• 39:11
  difference
• 39:11
  is,
• 39:11
  is
• 39:12
  that
• 39:12
  in
• 39:12
  the
• 39:12
  olden
• 39:13
  days
• 39:14
  you
• 39:15
  would
• 39:15
  say,
• 39:15
  here
• 39:16
  is
• 39:16
  my
• 39:16
  idea,
• 39:17
  I'm
• 39:17
  going
• 39:17
  to
• 39:17
  build
• 39:17
  it
• 39:18
  for
• 39:18
  the
• 39:18
  next
• 39:18
  18
• 39:18
  months.
• 39:19
  And
• 39:20
  so
• 39:20
  you
• 39:20
  have
• 39:20
  18
• 39:20
  months
• 39:21
  to
• 39:21
  figure
• 39:21
  out
• 39:21
  the
• 39:21
  story,
• 39:22
  the
• 39:22
  sales
• 39:22
  cycle.
• 39:23
  You're
• 39:23
  talking
• 39:23
  to
• 39:23
  customers,
• 39:24
  you're
• 39:24
  spending
• 39:24
  all
• 39:25
  this
• 39:25
  time
• 39:25
  and
• 39:25
  then
• 39:25
  you're
• 39:25
  like,
• 39:25
  done,
• 39:26
  I'm
• 39:26
  done.
• 39:27
  But
• 39:27
  then
• 39:27
  at
• 39:28
  that
• 39:28
  18
• 39:28
  months
• 39:28
  during
• 39:29
  that
• 39:29
  time
• 39:29
  period,
• 39:29
  you
• 39:30
  spent
• 39:30
  all
• 39:30
  this
• 39:31
  time
• 39:31
  with
• 39:31
  people.
• 39:32
  So
• 39:32
  you're
• 39:32
  iterating
• 39:32
  and
• 39:33
  you're
• 39:33
  doing
• 39:33
  this
• 39:33
  stuff
• 39:33
  because
• 39:33
  it
• 39:33
  just
• 39:34
  takes
• 39:34
  a
• 39:34
  long
• 39:34
  time
• 39:34
  to
• 39:34
  build.
• 39:35
  Now
• 39:35
  you
• 39:36
  take
• 39:36
  that
• 39:36
  same
• 39:36
  amount
• 39:36
  of
• 39:37
  time
• 39:37
  to
• 39:37
  build
• 39:37
  and
• 39:37
  you've
• 39:38
  shrunk
• 39:38
  it
• 39:38
  down
• 39:38
  to
• 39:38
  two
• 39:38
  weeks.
• 39:39
  That's
• 39:40
  with
• 39:40
  a
• 39:40
  weak
• 39:40
  padding.
• 39:41
  And
• 39:41
  now
• 39:41
  you're
• 39:42
  just
• 39:42
  like,
• 39:42
  you
• 39:42
  talk
• 39:42
  to
• 39:43
  like
• 39:43
  five
• 39:43
  people
• 39:43
  and
• 39:43
  they're
• 39:43
  like,
• 39:44
  that's
• 39:44
  a
• 39:44
  shitty
• 39:44
  idea.
• 39:44
  And
• 39:44
  you're
• 39:45
  just
• 39:45
  like,
• 39:45
  well,
• 39:45
  okay.
• 39:46
  Then
• 39:46
  two
• 39:46
  weeks
• 39:47
  go
• 39:47
  by
• 39:47
  and
• 39:47
  you're
• 39:47
  like,
• 39:47
  well,
• 39:47
  I
• 39:47
  guess
• 39:47
  I'll
• 39:48
  do
• 39:48
  it
• 39:48
  again.
• 39:48
  So
• 39:48
  you
• 39:48
  just
• 39:48
  keep
• 39:48
  doing
• 39:49
  this
• 39:49
  again.
• 39:49
  So
• 39:50
  we've
• 39:50
  done
• 39:50
  a
• 39:50
  lot.
• 39:52
  What
• 39:52
  we've
• 39:52
  found
• 39:53
  is
• 39:53
  a
• 39:53
  handful
• 39:53
  of
• 39:53
  things.
• 39:54
  The
• 39:54
  first
• 39:54
  thing
• 39:54
  is
• 39:54
  that
• 39:55
  we
• 39:55
  have,
• 39:56
  and
• 39:56
  I
• 39:56
  know
• 39:56
  there's
• 39:57
  a
• 39:57
  handful
• 39:57
  of
• 39:57
  people
• 39:57
  that
• 39:57
  are
• 39:57
  doing
• 39:58
  stuff
• 39:58
  like
• 39:58
  that.
• 39:58
  We
• 39:58
  have
• 39:59
  basically
• 40:00
  a
• 40:00
  codegen
• 40:00
  loop
• 40:01
  that
• 40:01
  is
• 40:01
  very
• 40:01
  effective
• 40:02
  and
• 40:03
  at
• 40:03
  this
• 40:03
  point
• 40:03
  we
• 40:04
  can
• 40:04
  drop
• 40:04
  kind
• 40:04
  of
• 40:04
  any
• 40:05
  spec
• 40:06
  in
• 40:06
  and
• 40:06
  within
• 40:07
  a
• 40:07
  very
• 40:07
  small
• 40:07
  amount
• 40:08
  of
• 40:08
  time
• 40:08
  getting
• 40:08
  any
• 40:08
  level
• 40:09
  of
• 40:09
  tech
• 40:09
  out
• 40:10
  of
• 40:10
  it
• 40:11
  very
• 40:11
  effectively.
• 40:12
  That's
• 40:12
  interesting.
• 40:13
  A
• 40:13
  lot
• 40:13
  of
• 40:13
  people
• 40:13
  are
• 40:13
  trying
• 40:13
  to
• 40:13
  do
• 40:13
  that.
• 40:14
  I
• 40:14
  haven't
• 40:14
  heard
• 40:14
  anyone
• 40:14
  doing
• 40:14
  that
• 40:15
  effectively
• 40:15
  yet.
• 40:15
  It's
• 40:16
  very
• 40:16
  effective.
• 40:16
  It's
• 40:17
  scary
• 40:17
  effective.
• 40:18
  There's
• 40:18
  A
• 40:18
  couple
• 40:19
  things
• 40:19
  that
• 40:19
  are
• 40:19
  really
• 40:19
  important
• 40:19
  there.
• 40:21
  We
• 40:22
  have
• 40:22
  embraced
• 40:23
  what
• 40:23
  we're
• 40:23
  calling.
• 40:24
  I
• 40:24
  stole
• 40:24
  this
• 40:24
  from
• 40:24
  someone
• 40:24
  else.
• 40:25
  So
• 40:25
  it's
• 40:25
  not
• 40:25
  like
• 40:25
  my
• 40:25
  invention
• 40:26
  here,
• 40:26
  but
• 40:26
  calling
• 40:27
  like
• 40:27
  scenario
• 40:27
  testing.
• 40:28
  So
• 40:28
  it's
• 40:28
  a
• 40:28
  little
• 40:28
  bit
• 40:28
  like
• 40:29
  integration
• 40:29
  testing,
• 40:30
  but
• 40:30
  it's
• 40:30
  not
• 40:30
  necessarily
• 40:31
  for
• 40:32
  some
• 40:32
  smoke
• 40:33
  test
• 40:33
  or
• 40:33
  some
• 40:34
  integration
• 40:34
  test.
• 40:35
  It's
• 40:35
  just
• 40:35
  you're
• 40:35
  trying
• 40:36
  to
• 40:36
  get
• 40:36
  the
• 40:36
  agents
• 40:36
  to
• 40:36
  use
• 40:36
  the
• 40:37
  product
• 40:37
  like
• 40:37
  a
• 40:37
  human
• 40:37
  would.
• 40:39
  So
• 40:39
  for
• 40:39
  instance,
• 40:39
  you
• 40:39
  give
• 40:39
  them
• 40:39
  a
• 40:40
  browser
• 40:40
  and
• 40:40
  you
• 40:40
  have
• 40:40
  them
• 40:41
  go
• 40:41
  through
• 40:41
  like
• 40:42
  a
• 40:42
  user
• 40:42
  test
• 40:42
  or
• 40:43
  a
• 40:43
  user
• 40:43
  experience
• 40:43
  test
• 40:44
  and
• 40:44
  what
• 40:44
  they
• 40:44
  would
• 40:44
  do
• 40:44
  and
• 40:45
  like
• 40:45
  do
• 40:45
  various
• 40:45
  things
• 40:45
  like
• 40:45
  that.
• 40:46
  And
• 40:46
  it's
• 40:47
  incredible
• 40:47
  because
• 40:47
  if
• 40:47
  you
• 40:48
  set
• 40:48
  up
• 40:48
  this
• 40:48
  loop
• 40:49
  of
• 40:49
  that,
• 40:50
  then
• 40:50
  you
• 40:50
  can
• 40:50
  just
• 40:50
  be
• 40:50
  like,
• 40:51
  does
• 40:51
  it
• 40:52
  work?
• 40:52
  And
• 40:52
  they'll
• 40:52
  be
• 40:53
  like,
• 40:53
  yes.
• 40:53
  And
• 40:53
  then
• 40:53
  it
• 40:54
  gets
• 40:54
  closer
• 40:54
  and
• 40:54
  closer
• 40:55
  to
• 40:55
  working.
• 40:56
  I'm
• 40:56
  smiling
• 40:57
  because.
• 40:58
  And
• 40:58
  please,
• 40:58
  I
• 40:59
  want
• 40:59
  you
• 40:59
  guys
• 40:59
  to
• 40:59
  like
• 41:00
  blow
• 41:00
  this
• 41:00
  argument
• 41:01
  up
• 41:01
  if
• 41:01
  you
• 41:01
  can.
• 41:02
  And
• 41:03
  for
• 41:03
  sure,
• 41:03
  I
• 41:03
  am
• 41:03
  biased
• 41:04
  in
• 41:04
  this
• 41:04
  because
• 41:04
  I
• 41:05
  come
• 41:05
  from
• 41:05
  a
• 41:05
  product
• 41:06
  background.
• 41:06
  I
• 41:06
  was
• 41:07
  a
• 41:07
  product
• 41:07
  manager.
• 41:08
  It
• 41:08
  feels
• 41:09
  to
• 41:09
  me
• 41:10
  like
• 41:11
  the
• 41:11
  skills
• 41:12
  you
• 41:12
  need
• 41:12
  to
• 41:13
  be
• 41:13
  a
• 41:13
  really
• 41:13
  good
• 41:14
  product
• 41:15
  builder
• 41:16
  are
• 41:16
  the
• 41:17
  skills
• 41:18
  that
• 41:18
  are
• 41:18
  becoming
• 41:19
  increasingly
• 41:20
  important.
• 41:20
  Do
• 41:20
  you
• 41:20
  guys
• 41:21
  agree?
• 41:21
  Yeah,
• 41:22
  I
• 41:23
  think
• 41:23
  I.
• 41:24
  What
• 41:24
  I'm
• 41:24
  saying
• 41:24
  is
• 41:25
  I'm
• 41:26
  super
• 41:26
  important,
• 41:27
  very
• 41:28
  important.
• 41:28
  I
• 41:29
  feel
• 41:29
  like
• 41:29
  I
• 41:29
  don't.
• 41:29
  Harper,
• 41:29
  I
• 41:30
  don't
• 41:30
  know
• 41:30
  if
• 41:30
  you
• 41:30
  feel
• 41:30
  the
• 41:30
  same.
• 41:30
  I
• 41:30
  feel
• 41:30
  like
• 41:30
  I'm
• 41:31
  turning
• 41:31
  into
• 41:31
  a
• 41:31
  product
• 41:31
  manager
• 41:32
  because
• 41:32
  that's
• 41:33
  how
• 41:33
  I
• 41:33
  start
• 41:33
  building
• 41:33
  everything.
• 41:34
  It's
• 41:34
  like,
• 41:35
  here's
• 41:35
  exactly
• 41:35
  what
• 41:36
  I
• 41:36
  want
• 41:36
  it
• 41:36
  to
• 41:36
  look
• 41:36
  like
• 41:36
  in
• 41:36
  the
• 41:37
  end.
• 41:37
  And
• 41:37
  then
• 41:38
  I
• 41:38
  say,
• 41:38
  okay,
• 41:38
  let's
• 41:39
  break
• 41:39
  this
• 41:39
  down
• 41:39
  into
• 41:40
  testable
• 41:40
  steps.
• 41:41
  So
• 41:42
  give
• 41:42
  me
• 41:42
  testable
• 41:43
  steps.
• 41:43
  And
• 41:43
  we're
• 41:44
  going
• 41:44
  to
• 41:44
  build
• 41:44
  them.
• 41:44
  At
• 41:45
  the
• 41:45
  end
• 41:45
  of
• 41:45
  each
• 41:45
  step,
• 41:45
  we're
• 41:45
  going
• 41:46
  to
• 41:46
  test
• 41:46
  and
• 41:46
  then
• 41:46
  we're
• 41:47
  going
• 41:47
  to
• 41:47
  build
• 41:47
  the
• 41:47
  next
• 41:47
  step.
• 41:47
  And
• 41:48
  that's
• 41:48
  been
• 41:48
  the
• 41:48
  most
• 41:49
  productive
• 41:49
  way
• 41:49
  for
• 41:49
  me
• 41:50
  to
• 41:50
  build.
• 41:51
  And
• 41:51
  I
• 41:51
  feel
• 41:51
  like
• 41:51
  that's
• 41:52
  very
• 41:52
  much
• 41:52
  like
• 41:53
  a
• 41:53
  product
• 41:53
  manager.
• 41:54
  Do
• 41:54
  great.
• 41:54
  Well,
• 41:55
  I
• 41:55
  think
• 41:55
  it's
• 41:55
  funnier.
• 41:56
  There's
• 41:56
  two
• 41:56
  things
• 41:56
  I
• 41:56
  think
• 41:56
  about
• 41:57
  a
• 41:57
  lot
• 41:57
  of.
• 41:58
  I've
• 41:58
  been
• 41:58
  speaking
• 41:59
  at
• 41:59
  Northwestern
• 42:00
  Business
• 42:00
  School
• 42:00
  a
• 42:00
  couple
• 42:00
  times
• 42:01
  and
• 42:01
  I'm
• 42:01
  always
• 42:01
  like,
• 42:01
  now
• 42:01
  is
• 42:02
  the
• 42:02
  time
• 42:02
  for
• 42:02
  the
• 42:02
  idea
• 42:03
  guy.
• 42:04
  You
• 42:04
  know
• 42:04
  what
• 42:04
  I'm
• 42:04
  talking?
• 42:05
  Like
• 42:05
  they
• 42:05
  finally
• 42:05
  found
• 42:06
  their
• 42:06
  time.
• 42:07
  All
• 42:07
  these
• 42:07
  like,
• 42:08
  all
• 42:08
  these
• 42:08
  Patagonia
• 42:09
  vested
• 42:09
  bros
• 42:09
  who
• 42:10
  are
• 42:10
  like,
• 42:10
  hey,
• 42:10
  I'm
• 42:10
  looking
• 42:11
  for
• 42:11
  a
• 42:11
  technical
• 42:11
  co
• 42:11
  founder.
• 42:12
  They
• 42:13
  finally
• 42:13
  got
• 42:13
  it.
• 42:13
  And
• 42:13
  I'm
• 42:14
  so
• 42:14
  excited
• 42:14
  for
• 42:14
  them
• 42:16
  because
• 42:16
  the
• 42:16
  thing
• 42:16
  is,
• 42:17
  even
• 42:17
  though
• 42:17
  it's
• 42:17
  easy
• 42:18
  and
• 42:18
  this
• 42:18
  is
• 42:18
  where
• 42:18
  the
• 42:18
  struggle
• 42:19
  is,
• 42:19
  like
• 42:19
  what
• 42:19
  2389
• 42:20
  has
• 42:20
  been
• 42:20
  doing,
• 42:20
  we've
• 42:21
  been
• 42:21
  doing
• 42:21
  so
• 42:21
  many
• 42:21
  experiments,
• 42:22
  we've
• 42:22
  released
• 42:23
  a
• 42:23
  lot
• 42:23
  of
• 42:23
  software
• 42:24
  We've
• 42:24
  done
• 42:24
  all
• 42:24
  of
• 42:24
  this
• 42:24
  stuff,
• 42:25
  but
• 42:25
  still
• 42:26
  it's
• 42:26
  hard
• 42:27
  because
• 42:27
  just
• 42:27
  because
• 42:28
  you
• 42:28
  can
• 42:28
  build
• 42:28
  the
• 42:28
  software
• 42:29
  faster
• 42:30
  doesn't
• 42:30
  mean
• 42:30
  that
• 42:30
  it
• 42:31
  makes
• 42:31
  it
• 42:31
  easier
• 42:32
  to
• 42:32
  grow
• 42:32
  a
• 42:32
  company,
• 42:32
  to
• 42:33
  build
• 42:33
  a
• 42:33
  product,
• 42:33
  to
• 42:34
  do
• 42:34
  all
• 42:34
  this
• 42:34
  stuff.
• 42:35
  And
• 42:35
  I
• 42:35
  think
• 42:35
  this
• 42:35
  is
• 42:36
  a
• 42:36
  good
• 42:36
  reminder
• 42:37
  to
• 42:37
  all
• 42:37
  of
• 42:37
  us
• 42:37
  who
• 42:38
  are
• 42:38
  entrepreneurs
• 42:39
  that
• 42:39
  especially
• 42:39
  tech
• 42:40
  people
• 42:40
  who
• 42:40
  are
• 42:40
  tech
• 42:41
  biased,
• 42:41
  that
• 42:42
  product
• 42:42
  is
• 42:43
  not
• 42:43
  and
• 42:43
  is
• 42:44
  rarely
• 42:44
  technical
• 42:45
  solution.
• 42:46
  Right.
• 42:47
  It's
• 42:47
  much
• 42:47
  more
• 42:48
  about
• 42:48
  like,
• 42:48
  is
• 42:49
  it
• 42:49
  something
• 42:49
  that
• 42:49
  your
• 42:49
  customer
• 42:50
  wants?
• 42:51
  Is
• 42:51
  it
• 42:51
  something
• 42:51
  that
• 42:51
  your
• 42:52
  customer
• 42:52
  needs?
• 42:53
  Maybe
• 42:53
  not
• 42:54
  needs.
• 42:54
  I
• 42:54
  think
• 42:54
  customers
• 42:55
  are
• 42:55
  often
• 42:55
  wrong.
• 42:55
  I
• 42:55
  think
• 42:55
  they
• 42:56
  don't
• 42:56
  know
• 42:56
  what
• 42:56
  they
• 42:56
  need.
• 42:57
  But
• 42:57
  I
• 42:57
  do
• 42:58
  think
• 42:58
  there
• 42:58
  is
• 42:58
  a
• 42:58
  thing
• 42:58
  of
• 42:59
  like,
• 42:59
  growth
• 42:59
  is
• 42:59
  still
• 43:00
  hard.
• 43:00
  And
• 43:00
  there's
• 43:00
  some
• 43:00
  people
• 43:01
  who
• 43:01
  have
• 43:01
  done
• 43:01
  a
• 43:01
  good
• 43:01
  job
• 43:02
  of
• 43:02
  showing
• 43:02
  how
• 43:02
  they
• 43:02
  have
• 43:03
  used,
• 43:03
  you
• 43:04
  know,
• 43:04
  vibe
• 43:04
  marketing
• 43:05
  or
• 43:05
  whatever
• 43:05
  horrible
• 43:05
  words
• 43:06
  we
• 43:06
  want
• 43:06
  to
• 43:06
  use
• 43:06
  to
• 43:06
  describe
• 43:07
  it
• 43:07
  to,
• 43:07
  like,
• 43:09
  it's
• 43:09
  so
• 43:09
  good.
• 43:09
  But,
• 43:10
  like,
• 43:10
  all
• 43:10
  this
• 43:10
  stuff
• 43:10
  is
• 43:11
  like,
• 43:11
  I
• 43:11
  have
• 43:11
  not
• 43:12
  yet
• 43:12
  seen
• 43:13
  easily
• 43:14
  adopted
• 43:15
  workflows
• 43:15
  to
• 43:16
  help
• 43:16
  use
• 43:16
  LLMs
• 43:17
  to
• 43:17
  build
• 43:18
  do
• 43:18
  better
• 43:18
  growth
• 43:19
  without
• 43:20
  me
• 43:20
  screwing
• 43:20
  it
• 43:20
  up
• 43:21
  in
• 43:21
  some
• 43:21
  way.
• 43:22
  Whereas
• 43:22
  on
• 43:22
  the
• 43:22
  code
• 43:23
  side,
• 43:24
  you
• 43:24
  know,
• 43:24
  I
• 43:25
  have
• 43:25
  friends
• 43:25
  who
• 43:25
  are
• 43:25
  shipping
• 43:25
  iOS
• 43:26
  apps
• 43:26
  who
• 43:26
  should
• 43:27
  not
• 43:27
  be
• 43:27
  shipping
• 43:27
  iOS
• 43:28
  apps.
• 43:28
  If
• 43:28
  you
• 43:28
  looked
• 43:29
  at
• 43:29
  their
• 43:29
  abilities,
• 43:30
  their
• 43:30
  core
• 43:30
  abilities
• 43:31
  as
• 43:31
  judged
• 43:32
  five
• 43:32
  years
• 43:32
  ago,
• 43:33
  and
• 43:33
  they're
• 43:33
  doing
• 43:33
  fine
• 43:34
  now.
• 43:34
  They're
• 43:34
  still
• 43:35
  stumbling
• 43:35
  in
• 43:35
  the,
• 43:36
  like,
• 43:36
  how
• 43:36
  do
• 43:36
  I
• 43:36
  get
• 43:36
  people
• 43:36
  to
• 43:37
  use
• 43:37
  it?
• 43:37
  And
• 43:38
  that's
• 43:38
  not
• 43:38
  because
• 43:38
  it
• 43:38
  was
• 43:39
  vibe
• 43:39
  coded,
• 43:39
  that's
• 43:40
  because
• 43:40
  that
• 43:40
  shit's
• 43:41
  hard.
• 43:41
  Right?
• 43:42
  Right.
• 43:42
  It
• 43:42
  doesn't
• 43:43
  matter
• 43:43
  who
• 43:43
  made
• 43:43
  it
• 43:44
  or
• 43:44
  how
• 43:44
  it
• 43:44
  was
• 43:44
  made.
• 43:44
  It
• 43:45
  could
• 43:45
  be
• 43:45
  honed
• 43:45
  from
• 43:45
  metal
• 43:46
  and
• 43:46
  it
• 43:46
  still
• 43:46
  is
• 43:47
  very
• 43:47
  difficult
• 43:48
  to
• 43:48
  get
• 43:48
  people
• 43:48
  to
• 43:49
  use.
• 43:49
  And
• 43:49
  so
• 43:51
  we've
• 43:51
  been
• 43:51
  building
• 43:52
  a
• 43:52
  lot
• 43:52
  of
• 43:52
  stuff.
• 43:52
  We've
• 43:53
  just
• 43:53
  been
• 43:53
  pushing
• 43:53
  it
• 43:53
  out
• 43:53
  to
• 43:53
  friends
• 43:54
  and
• 43:54
  using
• 43:54
  it
• 43:54
  ourselves
• 43:55
  and
• 43:55
  trying
• 43:55
  to
• 43:55
  see
• 43:55
  what
• 43:56
  sticks.
• 43:56
  And
• 43:56
  the
• 43:56
  things
• 43:56
  that
• 43:57
  have
• 43:57
  been
• 43:57
  sticking
• 43:58
  are
• 43:59
  somewhat.
• 43:59
  How
• 43:59
  do
• 43:59
  you
• 44:00
  put
• 44:00
  agents
• 44:00
  into
• 44:01
  interfaces
• 44:01
  that
• 44:01
  are
• 44:02
  with
• 44:02
  you
• 44:02
  all
• 44:02
  the
• 44:02
  time?
• 44:02
  So,
• 44:02
  like,
• 44:03
  how
• 44:03
  do
• 44:03
  I
• 44:03
  code
• 44:03
  from
• 44:03
  my
• 44:04
  phone?
• 44:05
  People
• 44:05
  will
• 44:05
  get
• 44:05
  mad
• 44:05
  if
• 44:06
  you
• 44:06
  figure
• 44:06
  that
• 44:06
  out.
• 44:07
  Well,
• 44:07
  I
• 44:07
  think
• 44:08
  the
• 44:08
  thing.
• 44:08
  I
• 44:09
  really
• 44:09
  think
• 44:09
  you
• 44:09
  just
• 44:09
  have
• 44:09
  to
• 44:10
  give
• 44:10
  up
• 44:10
  on
• 44:10
  the
• 44:10
  fact
• 44:10
  that
• 44:10
  engineers
• 44:11
  were
• 44:11
  a
• 44:11
  thing
• 44:11
  once.
• 44:12
  Yeah,
• 44:12
  yeah.
• 44:13
  I
• 44:13
  think
• 44:13
  that,
• 44:13
  like,
• 44:13
  the
• 44:14
  fundamental
• 44:14
  premise
• 44:15
  here
• 44:15
  is
• 44:15
  that
• 44:15
  engineers
• 44:16
  are
• 44:16
  insufferable.
• 44:17
  Yeah.
• 44:18
  And
• 44:18
  they're
• 44:18
  bad
• 44:18
  people
• 44:19
  and
• 44:19
  they
• 44:19
  probably
• 44:19
  are
• 44:20
  annoying.
• 44:21
  And
• 44:21
  I
• 44:21
  include
• 44:21
  myself
• 44:22
  in
• 44:22
  this
• 44:22
  because
• 44:22
  I've
• 44:23
  had
• 44:23
  to
• 44:23
  work
• 44:23
  with
• 44:23
  myself
• 44:24
  for
• 44:24
  years
• 44:24
  now.
• 44:25
  But
• 44:25
  I
• 44:25
  do
• 44:25
  think
• 44:26
  there's
• 44:26
  a
• 44:26
  real
• 44:26
  thing
• 44:26
  of
• 44:27
  like,
• 44:27
  so
• 44:27
  we
• 44:28
  built
• 44:28
  this
• 44:28
  simple
• 44:28
  product
• 44:29
  that
• 44:29
  put
• 44:29
  cloud
• 44:30
  code
• 44:30
  inside
• 44:30
  of
• 44:30
  a
• 44:30
  matrix
• 44:31
  chat
• 44:32
  and
• 44:32
  it
• 44:32
  is
• 44:33
  barely
• 44:34
  working,
• 44:34
  like
• 44:34
  it's
• 44:35
  an
• 44:35
  early
• 44:35
  product.
• 44:35
  And
• 44:36
  so
• 44:36
  we
• 44:36
  gave
• 44:36
  it
• 44:36
  to.
• 44:36
  Is
• 44:36
  that
• 44:36
  what
• 44:36
  you
• 44:36
  mean
• 44:37
  by
• 44:37
  Matrix
• 44:37
  Chat.
• 44:38
  Matrix
• 44:38
  is
• 44:39
  this
• 44:39
  open
• 44:39
  source
• 44:40
  kind
• 44:40
  of
• 44:40
  do
• 44:41
  it
• 44:41
  all
• 44:41
  chat
• 44:41
  thing.
• 44:42
  It's
• 44:42
  kind
• 44:42
  of
• 44:42
  like
• 44:42
  a
• 44:42
  logical
• 44:43
  conclusion
• 44:44
  from
• 44:44
  xmpp.
• 44:45
  None
• 44:46
  of
• 44:46
  this
• 44:46
  makes
• 44:46
  any
• 44:46
  sense.
• 44:47
  It's
• 44:47
  basically
• 44:47
  just
• 44:47
  a
• 44:48
  chat
• 44:48
  app
• 44:48
  that
• 44:48
  is
• 44:49
  super
• 44:49
  encrypted,
• 44:49
  that
• 44:50
  allows
• 44:50
  for
• 44:50
  rooms
• 44:51
  and
• 44:51
  some
• 44:51
  video
• 44:51
  chat
• 44:52
  and
• 44:52
  some
• 44:52
  other
• 44:52
  stuff
• 44:52
  in
• 44:52
  there.
• 44:53
  The
• 44:53
  reason
• 44:54
  why
• 44:54
  Matrix,
• 44:55
  I've
• 44:55
  been
• 44:55
  using
• 44:55
  Matrix
• 44:56
  for
• 44:56
  bots
• 44:56
  for
• 44:57
  a
• 44:57
  really
• 44:57
  long
• 44:57
  time
• 44:59
  because
• 44:59
  it's
• 44:59
  really
• 45:00
  easy
• 45:00
  to
• 45:00
  build
• 45:00
  bots
• 45:01
  in.
• 45:01
  Interesting.
• 45:02
  Like
• 45:02
  WhatsApp
• 45:03
  is
• 45:03
  a
• 45:03
  pain
• 45:03
  in
• 45:03
  the
• 45:03
  ass
• 45:03
  to
• 45:04
  build
• 45:04
  a
• 45:04
  bot
• 45:04
  in
• 45:05
  Telegram
• 45:05
  seems
• 45:06
  like
• 45:06
  a
• 45:06
  liability
• 45:06
  to
• 45:07
  use.
• 45:07
  Is
• 45:07
  that
• 45:07
  because
• 45:08
  Matrix
• 45:08
  is
• 45:08
  open
• 45:09
  source.
• 45:09
  They
• 45:10
  just
• 45:10
  don't
• 45:10
  care,
• 45:10
  you
• 45:10
  know.
• 45:10
  And
• 45:11
  it's
• 45:11
  pretty
• 45:12
  easy
• 45:12
  to
• 45:12
  use.
• 45:12
  It
• 45:12
  doesn't
• 45:13
  have
• 45:13
  a
• 45:13
  huge
• 45:13
  number
• 45:13
  of
• 45:14
  users,
• 45:14
  so
• 45:14
  it's
• 45:14
  probably
• 45:15
  not
• 45:15
  being
• 45:15
  abused.
• 45:16
  Wait,
• 45:16
  that's
• 45:16
  a
• 45:16
  hot
• 45:17
  tip.
• 45:17
  I
• 45:17
  want
• 45:17
  to
• 45:17
  play
• 45:18
  around.
• 45:18
  Matrix
• 45:18
  is
• 45:18
  great.
• 45:19
  The
• 45:19
  clients
• 45:19
  are
• 45:20
  all
• 45:20
  pretty
• 45:20
  good.
• 45:20
  They're
• 45:21
  pretty
• 45:21
  rich.
• 45:21
  Like
• 45:22
  it's
• 45:22
  pretty
• 45:22
  nice.
• 45:22
  We
• 45:22
  built
• 45:23
  this
• 45:23
  thing
• 45:23
  that
• 45:23
  just
• 45:23
  has
• 45:23
  like
• 45:23
  cloud
• 45:24
  code.
• 45:24
  Like
• 45:24
  each
• 45:24
  channel
• 45:25
  you
• 45:25
  make
• 45:25
  is
• 45:25
  a
• 45:25
  cloud
• 45:26
  code
• 45:26
  instance
• 45:27
  with
• 45:27
  a
• 45:27
  bot.
• 45:28
  You
• 45:28
  can
• 45:28
  chat
• 45:28
  with
• 45:29
  the
• 45:29
  bot
• 45:29
  and
• 45:29
  you
• 45:29
  can
• 45:29
  do
• 45:30
  whatever
• 45:30
  cloud
• 45:30
  code
• 45:30
  can
• 45:31
  do.
• 45:31
  And
• 45:31
  this
• 45:31
  is
• 45:31
  just
• 45:32
  an
• 45:32
  experiment.
• 45:32
  And
• 45:32
  I
• 45:32
  had
• 45:33
  done
• 45:33
  some
• 45:33
  other
• 45:33
  stuff
• 45:33
  with
• 45:33
  Matrix
• 45:34
  in
• 45:34
  the
• 45:34
  past
• 45:35
  of
• 45:35
  hooking
• 45:35
  it
• 45:35
  into
• 45:35
  like
• 45:36
  LLM
• 45:36
  APIs
• 45:37
  just
• 45:37
  for
• 45:37
  like
• 45:37
  on
• 45:37
  the
• 45:37
  road
• 45:38
  chatgpt
• 45:39
  before
• 45:39
  the
• 45:39
  mobile
• 45:39
  apps
• 45:39
  were
• 45:40
  released.
• 45:40
  And
• 45:42
  I
• 45:42
  gave
• 45:43
  it
• 45:43
  to
• 45:43
  a
• 45:43
  friend
• 45:44
  and
• 45:44
  that
• 45:44
  friend
• 45:44
  immediately
• 45:45
  started
• 45:45
  building
• 45:45
  websites.
• 45:47
  I
• 45:47
  was
• 45:47
  just
• 45:47
  like,
• 45:47
  hey,
• 45:48
  you
• 45:48
  can
• 45:48
  just
• 45:48
  chat
• 45:48
  with
• 45:48
  this
• 45:49
  thing.
• 45:49
  And
• 45:49
  it
• 45:49
  would
• 45:49
  check
• 45:49
  your
• 45:50
  email
• 45:50
  or
• 45:50
  whatever.
• 45:50
  And
• 45:51
  he
• 45:51
  just
• 45:51
  started
• 45:51
  ripping
• 45:52
  up
• 45:52
  websites.
• 45:52
  He
• 45:53
  built
• 45:53
  like
• 45:53
  seven
• 45:53
  websites.
• 45:54
  And
• 45:54
  I
• 45:54
  was
• 45:55
  just
• 45:55
  like,
• 45:55
  where
• 45:55
  are
• 45:56
  you
• 45:56
  going
• 45:56
  to
• 45:56
  host
• 45:56
  them?
• 45:56
  He's
• 45:57
  like,
• 45:57
  I
• 45:57
  don't
• 45:57
  know.
• 45:57
  I
• 45:58
  don't
• 45:58
  know
• 45:58
  how
• 45:58
  he
• 45:58
  did
• 45:58
  it.
• 45:58
  He's
• 45:59
  not
• 45:59
  a
• 45:59
  technical
• 45:59
  person.
• 45:59
  He
• 46:00
  doesn't
• 46:00
  care.
• 46:00
  He's
• 46:00
  not
• 46:01
  worried.
• 46:01
  I
• 46:01
  think
• 46:02
  we're
• 46:02
  the
• 46:02
  only
• 46:02
  us
• 46:02
  technical
• 46:03
  only
• 46:03
  ones
• 46:04
  worried
• 46:04
  about
• 46:04
  whether
• 46:05
  it's
• 46:05
  good
• 46:05
  or
• 46:05
  bad.
• 46:06
  They
• 46:06
  always
• 46:07
  happen.
• 46:07
  It's
• 46:07
  a
• 46:07
  story
• 46:08
  as
• 46:08
  old
• 46:08
  as
• 46:08
  time,
• 46:09
  I
• 46:09
  suppose,
• 46:09
  of
• 46:10
  just
• 46:10
  like
• 46:10
  us
• 46:11
  people
• 46:11
  with
• 46:11
  experience
• 46:11
  being
• 46:12
  like,
• 46:12
  you
• 46:12
  can't
• 46:12
  possibly
• 46:13
  do
• 46:13
  that.
• 46:13
  And
• 46:13
  the
• 46:13
  person
• 46:14
  without
• 46:14
  experience
• 46:14
  is
• 46:14
  being
• 46:15
  like,
• 46:15
  I
• 46:15
  did
• 46:15
  it.
• 46:16
  You're
• 46:16
  right,
• 46:17
  you're.
• 46:17
  You're
• 46:18
  right
• 46:18
  that
• 46:18
  I
• 46:18
  think
• 46:18
  there
• 46:19
  is
• 46:19
  probably
• 46:19
  like
• 46:20
  an
• 46:20
  emotional
• 46:20
  component
• 46:21
  to
• 46:21
  it
• 46:21
  too.
• 46:21
  Or
• 46:22
  like
• 46:22
  an
• 46:22
  ego
• 46:22
  threatening
• 46:23
  component
• 46:23
  to
• 46:23
  it
• 46:24
  too.
• 46:24
  Yeah.
• 46:24
  Which
• 46:24
  if
• 46:24
  I'm
• 46:25
  honest,
• 46:25
  is
• 46:26
  partially
• 46:26
  why
• 46:27
  I
• 46:27
  also
• 46:28
  do
• 46:28
  not
• 46:28
  like
• 46:29
  the
• 46:29
  term
• 46:29
  vibe
• 46:29
  coding.
• 46:30
  Because
• 46:30
  I'm
• 46:31
  like,
• 46:31
  I'm
• 46:31
  not
• 46:31
  just
• 46:31
  like
• 46:32
  vibing
• 46:32
  guys.
• 46:33
  I'm
• 46:33
  like,
• 46:33
  actually
• 46:34
  I
• 46:34
  know
• 46:34
  what
• 46:34
  I'm
• 46:34
  doing
• 46:35
  and
• 46:35
  I'm
• 46:35
  like
• 46:35
  telling
• 46:36
  it
• 46:36
  what
• 46:36
  to
• 46:36
  do.
• 46:37
  But
• 46:37
  even.
• 46:38
  But
• 46:38
  even
• 46:38
  then,
• 46:39
  isn't
• 46:39
  it,
• 46:39
  though,
• 46:40
  a
• 46:40
  vibe?
• 46:40
  Like,
• 46:41
  aren't
• 46:41
  you
• 46:41
  seriously
• 46:42
  running
• 46:42
  it?
• 46:43
  Because
• 46:43
  you
• 46:43
  are.
• 46:43
  Like,
• 46:44
  this
• 46:44
  feels
• 46:44
  right
• 46:45
  or
• 46:45
  feels
• 46:45
  wrong?
• 46:45
  Because
• 46:46
  you
• 46:46
  no
• 46:46
  longer
• 46:46
  have
• 46:47
  to
• 46:47
  have
• 46:47
  this,
• 46:48
  like,
• 46:48
  intellectual,
• 46:49
  pseudo
• 46:49
  intellectual
• 46:50
  idea
• 46:50
  that
• 46:50
  you
• 46:51
  executed
• 46:51
  on
• 46:51
  something
• 46:52
  accurately
• 46:52
  that's
• 46:53
  just
• 46:53
  gone.
• 46:53
  Like,
• 46:54
  you
• 46:54
  don't
• 46:54
  have
• 46:54
  to
• 46:54
  do
• 46:54
  that.
• 46:55
  So
• 46:55
  now
• 46:55
  it
• 46:55
  really
• 46:56
  is
• 46:56
  like,
• 46:56
  does
• 46:56
  this
• 46:57
  feel
• 46:57
  right?
• 46:58
  Like,
• 46:59
  there's
• 46:59
  this.
• 46:59
  One
• 47:00
  of
• 47:00
  my
• 47:00
  favorite
• 47:00
  movies,
• 47:01
  because
• 47:01
  I'm
• 47:01
  about
• 47:01
  700
• 47:02
  years
• 47:02
  old,
• 47:02
  is
• 47:02
  a
• 47:03
  movie
• 47:03
  called
• 47:03
  UHF
• 47:04
  by
• 47:05
  Weird
• 47:05
  Al
• 47:05
  Yankovic
• 47:06
  from
• 47:06
  the
• 47:06
  80s.
• 47:07
  And
• 47:07
  it's
• 47:07
  a
• 47:07
  wonderful
• 47:07
  movie.
• 47:08
  I
• 47:08
  highly
• 47:08
  recommend
• 47:08
  it.
• 47:09
  I
• 47:09
  have
• 47:09
  no
• 47:09
  idea
• 47:09
  if
• 47:09
  it's
• 47:10
  super
• 47:10
  problematic
• 47:10
  or
• 47:11
  not.
• 47:11
  I
• 47:11
  don't
• 47:11
  care.
• 47:12
  But
• 47:12
  it's
• 47:13
  very.
• 47:13
  It's
• 47:13
  good.
• 47:13
  But
• 47:14
  I
• 47:14
  can
• 47:14
  get
• 47:14
  back
• 47:14
  to
• 47:14
  you
• 47:14
  on
• 47:14
  that.
• 47:15
  Yeah,
• 47:15
  please
• 47:15
  do.
• 47:15
  But
• 47:16
  one
• 47:16
  of
• 47:16
  the
• 47:16
  scenes
• 47:16
  in
• 47:17
  it
• 47:17
  is,
• 47:17
  like,
• 47:17
  two
• 47:17
  guys
• 47:17
  sitting
• 47:18
  next
• 47:18
  to
• 47:18
  each
• 47:18
  other.
• 47:18
  One
• 47:18
  of
• 47:18
  them
• 47:19
  is
• 47:19
  blind,
• 47:19
  and
• 47:19
  the
• 47:19
  blind
• 47:19
  guy
• 47:20
  is
• 47:20
  doing
• 47:20
  a
• 47:20
  Rubik's
• 47:20
  cue
• 47:20
  to
• 47:21
  just
• 47:21
  asking
• 47:21
  the
• 47:21
  other
• 47:21
  guy,
• 47:22
  did
• 47:22
  I
• 47:22
  get
• 47:22
  it?
• 47:22
  And
• 47:23
  the
• 47:23
  guy's
• 47:23
  like,
• 47:23
  not
• 47:23
  yet.
• 47:23
  And
• 47:23
  he
• 47:24
  changes
• 47:24
  it
• 47:24
  and
• 47:24
  he
• 47:24
  does
• 47:24
  it.
• 47:25
  Do
• 47:25
  I
• 47:25
  get
• 47:25
  it?
• 47:25
  And,
• 47:26
  like,
• 47:26
  as
• 47:26
  a
• 47:26
  kid,
• 47:26
  I
• 47:26
  always
• 47:27
  thought
• 47:27
  about
• 47:27
  that.
• 47:27
  But
• 47:27
  that
• 47:27
  feels
• 47:28
  a
• 47:28
  lot
• 47:28
  like
• 47:28
  what
• 47:28
  we're
• 47:28
  doing
• 47:29
  right
• 47:29
  now
• 47:29
  when
• 47:29
  we're
• 47:29
  building
• 47:29
  products
• 47:30
  where
• 47:30
  we're
• 47:30
  just
• 47:30
  like,
• 47:31
  claude,
• 47:32
  take
• 47:32
  another
• 47:32
  pass.
• 47:33
  And
• 47:33
  God's
• 47:34
  like,
• 47:34
  beep,
• 47:34
  beep,
• 47:34
  beep,
• 47:34
  beep.
• 47:35
  And
• 47:35
  then
• 47:35
  you're
• 47:35
  like,
• 47:35
  did
• 47:35
  I
• 47:35
  get
• 47:36
  it?
• 47:36
  And
• 47:36
  you're
• 47:36
  like,
• 47:36
  that's
• 47:36
  not
• 47:36
  quite
• 47:37
  it.
• 47:37
  Take
• 47:37
  another
• 47:37
  pass.
• 47:38
  And
• 47:38
  it's
• 47:38
  like,
• 47:38
  boop,
• 47:38
  boop,
• 47:38
  beep,
• 47:38
  boop,
• 47:39
  beep.
• 47:39
  And
• 47:39
  it's
• 47:39
  like,
• 47:39
  that
• 47:39
  is
• 47:40
  like
• 47:40
  the
• 47:40
  iteration
• 47:40
  process
• 47:41
  instead
• 47:41
  of
• 47:41
  us
• 47:42
  being
• 47:42
  so
• 47:42
  involved,
• 47:43
  like,
• 47:43
  we
• 47:43
  are
• 47:44
  the
• 47:44
  judge,
• 47:44
  we
• 47:45
  are
• 47:45
  no
• 47:45
  longer
• 47:45
  the
• 47:45
  doer.
• 47:46
  And
• 47:46
  I
• 47:46
  think
• 47:46
  that
• 47:47
  is
• 47:47
  a
• 47:47
  very
• 47:47
  interesting
• 47:48
  kind
• 47:48
  of
• 47:48
  place
• 47:49
  of,
• 47:49
  like,
• 47:49
  when
• 47:49
  you
• 47:49
  are
• 47:49
  the
• 47:50
  one
• 47:50
  who's
• 47:50
  expected
• 47:50
  to
• 47:51
  be
• 47:51
  the
• 47:51
  arbiter
• 47:52
  of
• 47:52
  taste,
• 47:53
  not
• 47:53
  necessarily
• 47:54
  arbiter
• 47:54
  of
• 47:54
  technical
• 47:55
  competence.
• 47:56
  Like,
• 47:57
  it
• 47:57
  requires
• 47:57
  a
• 47:57
  different
• 47:57
  skill
• 47:58
  set.
• 47:58
  And
• 47:58
  this
• 47:58
  is
• 47:58
  also,
• 47:59
  I
• 47:59
  think,
• 47:59
  why
• 47:59
  you
• 47:59
  have
• 48:00
  so
• 48:00
  many
• 48:00
  people
• 48:00
  who
• 48:01
  are
• 48:02
  like,
• 48:02
  solopreneurs
• 48:02
  right
• 48:03
  now
• 48:03
  is
• 48:04
  because
• 48:04
  you
• 48:04
  can
• 48:04
  get
• 48:04
  so
• 48:04
  far
• 48:05
  without.
• 48:05
  But
• 48:05
  also,
• 48:06
  it's
• 48:06
  very
• 48:06
  hard
• 48:07
  to
• 48:07
  share
• 48:07
  taste.
• 48:08
  Yeah.
• 48:09
  One
• 48:09
  thing
• 48:09
  I've
• 48:09
  heard,
• 48:09
  and
• 48:10
  I
• 48:10
  probably
• 48:10
  said
• 48:10
  this
• 48:10
  already
• 48:11
  before,
• 48:11
  but
• 48:11
  someone
• 48:12
  said
• 48:12
  to
• 48:12
  me,
• 48:13
  what
• 48:13
  we're
• 48:13
  seeing
• 48:13
  on
• 48:14
  the
• 48:14
  Internet
• 48:14
  right
• 48:14
  now
• 48:14
  is
• 48:15
  very
• 48:15
  similar
• 48:15
  to
• 48:15
  what
• 48:16
  like
• 48:16
  YouTube
• 48:16
  did
• 48:16
  to
• 48:17
  video.
• 48:18
  It's
• 48:18
  like
• 48:18
  what
• 48:18
  we're
• 48:19
  seeing
• 48:19
  happen,
• 48:20
  encoding
• 48:20
  on
• 48:20
  the
• 48:20
  web.
• 48:21
  And,
• 48:21
  like,
• 48:22
  with
• 48:22
  video,
• 48:22
  like,
• 48:23
  you
• 48:23
  used
• 48:23
  to
• 48:23
  have
• 48:23
  to
• 48:23
  be
• 48:23
  a
• 48:23
  videographer.
• 48:24
  You
• 48:24
  used
• 48:24
  to
• 48:24
  have
• 48:24
  to
• 48:25
  have
• 48:25
  all
• 48:25
  the
• 48:25
  good
• 48:25
  equipment.
• 48:26
  You
• 48:26
  used
• 48:26
  to
• 48:26
  have
• 48:26
  to,
• 48:26
  like,
• 48:26
  have
• 48:26
  a
• 48:27
  technical
• 48:27
  understanding.
• 48:27
  How
• 48:28
  edit
• 48:28
  to
• 48:28
  video
• 48:28
  and
• 48:29
  then
• 48:29
  sit
• 48:29
  for
• 48:29
  like
• 48:29
  hours
• 48:30
  and
• 48:30
  edit
• 48:30
  video.
• 48:31
  And
• 48:31
  that
• 48:31
  was
• 48:32
  the
• 48:32
  skills
• 48:32
  that
• 48:33
  you
• 48:33
  needed.
• 48:34
  And
• 48:34
  now
• 48:34
  anyone
• 48:35
  can
• 48:35
  build
• 48:35
  video.
• 48:35
  And
• 48:36
  it
• 48:36
  really
• 48:36
  is,
• 48:36
  look
• 48:36
  at
• 48:36
  us.
• 48:36
  We'Re
• 48:37
  on
• 48:37
  a
• 48:37
  video
• 48:37
  podcast
• 48:38
  right
• 48:38
  now.
• 48:39
  I
• 48:39
  have
• 48:39
  no
• 48:40
  business
• 48:40
  being
• 48:40
  here.
• 48:42
  Exactly.
• 48:43
  And
• 48:43
  then
• 48:43
  taste
• 48:43
  becomes
• 48:44
  more
• 48:44
  important.
• 48:44
  I
• 48:45
  think
• 48:45
  that'll
• 48:45
  be
• 48:45
  really
• 48:45
  interesting
• 48:46
  to
• 48:46
  see
• 48:46
  where
• 48:46
  taste
• 48:47
  goes.
• 48:47
  Like,
• 48:47
  or
• 48:47
  like
• 48:48
  when
• 48:48
  the
• 48:48
  people
• 48:49
  that
• 48:49
  figure
• 48:49
  out
• 48:49
  like
• 48:50
  this
• 48:50
  is
• 48:50
  what
• 48:51
  people
• 48:51
  need
• 48:51
  and
• 48:51
  want,
• 48:51
  I
• 48:52
  think
• 48:52
  it's
• 48:52
  going
• 48:52
  to
• 48:52
  change
• 48:52
  a
• 48:53
  bunch.
• 48:53
  And
• 48:53
  it
• 48:53
  changes
• 48:54
  too
• 48:54
  because.
• 48:55
  Because
• 48:55
  there's
• 48:55
  this
• 48:55
  interesting
• 48:56
  thing
• 48:56
  that
• 48:56
  I
• 48:56
  love.
• 48:57
  I
• 48:57
  love
• 48:57
  this.
• 48:57
  I
• 48:58
  love
• 48:58
  the
• 48:59
  accessibility,
• 48:59
  I
• 49:00
  love
• 49:00
  the
• 49:00
  speed.
• 49:01
  Because
• 49:01
  the
• 49:01
  other
• 49:01
  thing
• 49:02
  that
• 49:02
  is
• 49:02
  true
• 49:03
  is
• 49:04
  you
• 49:04
  can
• 49:05
  test
• 49:05
  things
• 49:05
  very,
• 49:05
  very
• 49:06
  quickly.
• 49:07
  So
• 49:07
  my
• 49:07
  office
• 49:07
  is
• 49:08
  in
• 49:08
  the
• 49:08
  middle
• 49:08
  of
• 49:08
  a
• 49:08
  manufacturing
• 49:09
  incubator,
• 49:10
  like
• 49:10
  kind
• 49:10
  of
• 49:10
  like
• 49:10
  an
• 49:10
  old
• 49:11
  school
• 49:11
  incubator.
• 49:12
  I
• 49:12
  just
• 49:12
  talked
• 49:12
  to
• 49:12
  a
• 49:13
  friend
• 49:13
  of
• 49:13
  mine
• 49:13
  who
• 49:13
  runs
• 49:13
  a
• 49:14
  electric
• 49:15
  scooter
• 49:15
  lock
• 49:15
  company.
• 49:17
  It's
• 49:17
  like
• 49:17
  the
• 49:17
  viper
• 49:18
  alarms
• 49:18
  of
• 49:18
  electric
• 49:18
  scooters.
• 49:19
  It's
• 49:19
  stuff
• 49:20
  like
• 49:20
  that.
• 49:20
  But
• 49:20
  it's
• 49:20
  manufactured
• 49:21
  in
• 49:21
  this
• 49:21
  building.
• 49:22
  And
• 49:23
  talking
• 49:23
  to
• 49:23
  them,
• 49:23
  we've
• 49:24
  been
• 49:24
  doing
• 49:24
  a
• 49:24
  lot
• 49:24
  of
• 49:24
  work
• 49:25
  with
• 49:25
  them
• 49:25
  to
• 49:25
  try
• 49:25
  and
• 49:25
  figure
• 49:26
  out
• 49:26
  like
• 49:26
  what
• 49:26
  are
• 49:26
  some
• 49:26
  of
• 49:26
  the
• 49:26
  things
• 49:27
  they
• 49:27
  need.
• 49:27
  Both
• 49:27
  because
• 49:28
  we
• 49:28
  have
• 49:28
  a
• 49:28
  lot
• 49:28
  of
• 49:28
  E
• 49:28
  commerce
• 49:29
  experience.
• 49:29
  And
• 49:30
  what's
• 49:30
  exciting
• 49:30
  is
• 49:31
  they
• 49:31
  can
• 49:31
  say,
• 49:31
  oh,
• 49:31
  well,
• 49:31
  I
• 49:32
  need
• 49:32
  this
• 49:32
  thing.
• 49:32
  And
• 49:33
  then
• 49:33
  we
• 49:33
  can
• 49:33
  build
• 49:34
  it
• 49:34
  in
• 49:35
  a
• 49:35
  couple
• 49:35
  days
• 49:36
  and
• 49:36
  say,
• 49:36
  was
• 49:36
  this
• 49:36
  it?
• 49:37
  And
• 49:37
  they're
• 49:37
  going
• 49:38
  to
• 49:38
  say
• 49:38
  no.
• 49:38
  And
• 49:38
  then
• 49:38
  you
• 49:39
  go
• 49:39
  back,
• 49:39
  okay,
• 49:39
  what
• 49:39
  about
• 49:39
  now?
• 49:40
  And
• 49:40
  it's
• 49:40
  like,
• 49:40
  yes.
• 49:42
  And
• 49:42
  like
• 49:42
  that
• 49:42
  is
• 49:43
  something
• 49:43
  that's
• 49:43
  very
• 49:43
  real
• 49:44
  right
• 49:44
  now.
• 49:44
  And
• 49:44
  as
• 49:45
  kind
• 49:45
  of
• 49:45
  the
• 49:46
  like
• 49:47
  company
• 49:47
  that
• 49:48
  has
• 49:48
  1
• 49:48
  million
• 49:48
  internal
• 49:49
  products
• 49:49
  and
• 49:49
  has
• 49:50
  tested
• 49:50
  a
• 49:50
  couple
• 49:50
  of
• 49:50
  them
• 49:50
  out
• 49:51
  externally,
• 49:51
  like
• 49:52
  trying
• 49:52
  to
• 49:53
  figure
• 49:54
  that
• 49:54
  out,
• 49:54
  like
• 49:55
  it's
• 49:55
  really
• 49:55
  about
• 49:55
  your
• 49:56
  distance
• 49:57
  from
• 49:57
  the
• 49:57
  customer.
• 49:57
  I
• 49:58
  think
• 49:58
  right
• 49:58
  now
• 49:59
  with
• 49:59
  testing
• 49:59
  some
• 49:59
  of
• 50:00
  this
• 50:00
  stuff,
• 50:00
  like
• 50:00
  can
• 50:00
  you
• 50:01
  walk
• 50:01
  over
• 50:01
  to
• 50:01
  them
• 50:02
  and
• 50:02
  talk
• 50:02
  to
• 50:02
  them?
• 50:03
  And
• 50:04
  this
• 50:05
  is
• 50:05
  a
• 50:05
  complicated
• 50:05
  thing
• 50:06
  because
• 50:06
  one
• 50:06
  of
• 50:06
  the
• 50:06
  reasons
• 50:06
  I
• 50:06
  think
• 50:07
  I
• 50:07
  chose
• 50:07
  programming
• 50:07
  is
• 50:08
  because
• 50:08
  I
• 50:08
  get
• 50:08
  to
• 50:08
  sit
• 50:08
  inside
• 50:08
  in
• 50:09
  the
• 50:09
  dark.
• 50:10
  Now
• 50:11
  you're
• 50:11
  telling
• 50:11
  me
• 50:11
  I
• 50:11
  gotta
• 50:12
  go
• 50:12
  talk
• 50:12
  to
• 50:13
  anyone?
• 50:14
  Yeah,
• 50:14
  I
• 50:14
  mean
• 50:15
  I
• 50:15
  love
• 50:15
  talking
• 50:16
  to
• 50:16
  people,
• 50:16
  but
• 50:16
  it
• 50:16
  is
• 50:16
  definitely
• 50:17
  I'm
• 50:17
  just
• 50:17
  like,
• 50:17
  oh
• 50:18
  gosh.
• 50:18
  But
• 50:18
  it's
• 50:18
  more
• 50:18
  interesting
• 50:19
  now
• 50:19
  because
• 50:19
  I
• 50:19
  think
• 50:20
  the
• 50:20
  people
• 50:20
  I
• 50:20
  have
• 50:20
  seen,
• 50:22
  in
• 50:22
  my
• 50:22
  experience,
• 50:22
  the
• 50:22
  people
• 50:22
  I've
• 50:23
  seen
• 50:23
  who've
• 50:23
  been
• 50:23
  very
• 50:23
  successful
• 50:24
  at
• 50:24
  taking
• 50:24
  up
• 50:25
  programming
• 50:25
  at
• 50:25
  this
• 50:26
  world,
• 50:26
  are
• 50:26
  people
• 50:26
  who
• 50:26
  are
• 50:27
  really
• 50:27
  good
• 50:27
  at
• 50:27
  communicating.
• 50:29
  And
• 50:29
  so
• 50:29
  I
• 50:29
  have
• 50:29
  this
• 50:30
  theory,
• 50:30
  like
• 50:30
  I
• 50:30
  wonder
• 50:31
  if
• 50:31
  people
• 50:31
  who
• 50:31
  are
• 50:31
  good
• 50:32
  at
• 50:32
  like
• 50:32
  rapping
• 50:33
  or
• 50:33
  poetry
• 50:34
  or
• 50:34
  things
• 50:34
  where
• 50:34
  it's
• 50:34
  about
• 50:35
  efficiency
• 50:35
  of
• 50:35
  words,
• 50:36
  if
• 50:36
  they
• 50:36
  would
• 50:36
  be
• 50:37
  good
• 50:37
  with
• 50:38
  like
• 50:39
  prompting.
• 50:39
  I
• 50:39
  just
• 50:39
  don't
• 50:39
  know,
• 50:39
  I,
• 50:40
  I'm
• 50:41
  not
• 50:41
  cool
• 50:41
  enough
• 50:41
  to
• 50:41
  hang
• 50:41
  out
• 50:41
  with
• 50:41
  rappers.
• 50:44
  Yeah,
• 50:44
  I
• 50:44
  imagine
• 50:45
  there's
• 50:45
  programmers
• 50:45
  listening
• 50:46
  that
• 50:46
  are
• 50:46
  like,
• 50:46
  okay,
• 50:46
  but
• 50:46
  then
• 50:47
  what
• 50:47
  do
• 50:47
  I
• 50:47
  do?
• 50:48
  What
• 50:48
  are
• 50:48
  you
• 50:48
  telling
• 50:48
  me
• 50:48
  that
• 50:48
  you're
• 50:48
  boned?
• 50:49
  I
• 50:49
  think
• 50:49
  you're
• 50:49
  fucked.
• 50:50
  I
• 50:50
  don't
• 50:50
  think
• 50:50
  there's
• 50:50
  a
• 50:51
  what
• 50:51
  do
• 50:51
  I
• 50:51
  do?
• 50:51
  I
• 50:51
  really,
• 50:51
  I
• 50:52
  don't
• 50:52
  think
• 50:52
  so.
• 50:52
  I
• 50:52
  think
• 50:52
  that
• 50:53
  acting
• 50:53
  like
• 50:53
  there
• 50:54
  is
• 50:54
  an
• 50:54
  option
• 50:54
  within
• 50:55
  the
• 50:55
  space
• 50:55
  of
• 50:55
  programming
• 50:56
  and
• 50:56
  technology
• 50:57
  is
• 50:57
  just
• 50:57
  a
• 50:57
  trick.
• 50:57
  Well,
• 50:58
  is
• 50:58
  the
• 50:58
  answer
• 50:59
  like
• 50:59
  the
• 50:59
  answer
• 51:00
  doesn't
• 51:00
  lie
• 51:00
  in
• 51:00
  your
• 51:01
  programming
• 51:01
  abilities?
• 51:02
  The
• 51:02
  answer
• 51:02
  is
• 51:03
  going
• 51:03
  to
• 51:03
  lie
• 51:03
  in
• 51:03
  all
• 51:04
  the
• 51:04
  stuff
• 51:04
  you
• 51:04
  were
• 51:04
  just
• 51:04
  talking
• 51:05
  about
• 51:05
  is
• 51:05
  still
• 51:05
  hard.
• 51:06
  Yeah.
• 51:07
  I
• 51:07
  think
• 51:07
  the
• 51:07
  thing
• 51:07
  that
• 51:08
  we
• 51:08
  have
• 51:08
  been
• 51:08
  afforded
• 51:09
  the
• 51:09
  people
• 51:09
  of
• 51:10
  our
• 51:10
  generation,
• 51:10
  we
• 51:10
  have
• 51:11
  been
• 51:11
  afforded
• 51:11
  a
• 51:11
  very
• 51:11
  lucky
• 51:12
  time
• 51:12
  when
• 51:13
  you
• 51:13
  can
• 51:13
  start
• 51:13
  a
• 51:13
  job
• 51:14
  and
• 51:14
  continue
• 51:14
  to
• 51:14
  do
• 51:15
  that
• 51:15
  job
• 51:15
  over
• 51:15
  time
• 51:16
  without
• 51:16
  change.
• 51:17
  And
• 51:17
  I
• 51:17
  don't
• 51:17
  think
• 51:17
  that
• 51:18
  has
• 51:18
  been
• 51:18
  for
• 51:18
  a
• 51:18
  lot
• 51:18
  of
• 51:18
  other
• 51:19
  generations.
• 51:19
  And
• 51:20
  so
• 51:20
  there
• 51:20
  hasn't
• 51:21
  been
• 51:21
  an
• 51:21
  upheaval
• 51:22
  in
• 51:22
  how
• 51:22
  you
• 51:22
  program
• 51:23
  computers.
• 51:23
  When
• 51:23
  I
• 51:24
  started
• 51:24
  programming
• 51:24
  computers
• 51:25
  and
• 51:26
  probably
• 51:26
  professionally
• 51:27
  getting
• 51:27
  paid
• 51:28
  for
• 51:28
  it
• 51:28
  in
• 51:28
  the
• 51:28
  late
• 51:28
  90s,
• 51:29
  you
• 51:29
  had
• 51:29
  an
• 51:29
  IDE,
• 51:30
  a
• 51:31
  text
• 51:31
  editor
• 51:31
  and
• 51:32
  then
• 51:32
  you
• 51:32
  compiled
• 51:33
  somehow.
• 51:33
  And
• 51:33
  that's
• 51:34
  basically
• 51:34
  been
• 51:34
  the
• 51:34
  same
• 51:35
  until
• 51:35
  let's
• 51:35
  say
• 51:35
  2021,
• 51:37
  2022.
• 51:38
  And
• 51:38
  so
• 51:38
  that
• 51:39
  was
• 51:39
  basically
• 51:39
  it.
• 51:39
  Now
• 51:39
  you
• 51:40
  use
• 51:40
  better
• 51:40
  tools.
• 51:40
  New
• 51:41
  cool
• 51:41
  stuff
• 51:41
  came
• 51:41
  out,
• 51:41
  Ruby
• 51:42
  on
• 51:42
  Rails
• 51:42
  came
• 51:42
  out,
• 51:42
  which
• 51:43
  I
• 51:43
  could
• 51:43
  ignore
• 51:43
  and
• 51:43
  all
• 51:43
  these
• 51:43
  other
• 51:44
  things
• 51:44
  came
• 51:44
  out
• 51:44
  that
• 51:44
  were
• 51:44
  like,
• 51:45
  that
• 51:45
  were
• 51:45
  there
• 51:45
  that
• 51:46
  really
• 51:46
  made
• 51:46
  life
• 51:47
  better
• 51:47
  for
• 51:47
  programmers.
• 51:47
  Like
• 51:48
  Ruby
• 51:48
  on
• 51:48
  Rails
• 51:48
  really
• 51:48
  did
• 51:49
  make
• 51:49
  life
• 51:49
  better
• 51:49
  for
• 51:49
  programmers,
• 51:50
  just
• 51:51
  not
• 51:51
  me.
• 51:51
  But
• 51:51
  the
• 51:51
  thing
• 51:52
  is
• 51:53
  now
• 51:54
  it's
• 51:54
  different.
• 51:55
  I
• 51:55
  don't
• 51:55
  use
• 51:56
  an
• 51:56
  ide,
• 51:56
  I
• 51:56
  don't
• 51:57
  have
• 51:57
  an
• 51:57
  editor,
• 51:58
  I
• 51:58
  don't
• 51:59
  look
• 51:59
  at
• 51:59
  the
• 51:59
  code.
• 52:00
  That's
• 52:00
  really
• 52:01
  complicated.
• 52:02
  I
• 52:02
  think
• 52:02
  that
• 52:02
  we
• 52:03
  need
• 52:03
  to
• 52:03
  think
• 52:03
  about,
• 52:04
  okay,
• 52:04
  what
• 52:05
  is
• 52:05
  my
• 52:06
  job?
• 52:06
  Is
• 52:07
  my
• 52:07
  job
• 52:07
  to
• 52:07
  look
• 52:08
  at
• 52:08
  the
• 52:08
  code?
• 52:09
  No.
• 52:09
  So
• 52:09
  what
• 52:09
  is
• 52:10
  left?
• 52:11
  And
• 52:11
  if
• 52:11
  you
• 52:11
  start
• 52:11
  removing
• 52:12
  all
• 52:12
  the
• 52:12
  things
• 52:12
  that
• 52:12
  the
• 52:12
  code
• 52:12
  gen
• 52:13
  is
• 52:13
  doing,
• 52:13
  which
• 52:13
  is
• 52:14
  like
• 52:14
  generating
• 52:14
  and
• 52:15
  designing
• 52:15
  code,
• 52:16
  making
• 52:16
  some
• 52:17
  like
• 52:17
  mid
• 52:18
  level
• 52:18
  architecture
• 52:19
  decisions,
• 52:19
  if
• 52:19
  you
• 52:20
  take
• 52:20
  everything
• 52:20
  that
• 52:20
  it
• 52:20
  does
• 52:21
  and
• 52:21
  nothing
• 52:21
  is
• 52:22
  left,
• 52:22
  then
• 52:23
  you
• 52:23
  really
• 52:23
  have
• 52:23
  no
• 52:24
  future.
• 52:25
  But
• 52:25
  the
• 52:25
  thing
• 52:26
  is,
• 52:26
  is
• 52:26
  we're,
• 52:26
  we're
• 52:27
  years
• 52:27
  ahead
• 52:28
  of
• 52:28
  everyone
• 52:29
  being
• 52:29
  removed
• 52:29
  from
• 52:29
  their
• 52:30
  jobs
• 52:30
  hopefully.
• 52:31
  And
• 52:31
  so
• 52:32
  like
• 52:32
  there's
• 52:32
  time
• 52:33
  to
• 52:33
  reconcile
• 52:34
  and
• 52:34
  I
• 52:34
  don't
• 52:34
  think
• 52:34
  this
• 52:35
  is
• 52:35
  like
• 52:35
  en
• 52:35
  masse,
• 52:36
  every
• 52:36
  programmer
• 52:36
  should
• 52:36
  become
• 52:36
  a
• 52:37
  pm,
• 52:38
  but
• 52:38
  I
• 52:38
  do
• 52:38
  think
• 52:39
  understanding
• 52:40
  a
• 52:40
  little
• 52:40
  bit
• 52:40
  more
• 52:40
  what
• 52:40
  that
• 52:41
  means
• 52:41
  and
• 52:41
  I
• 52:41
  think
• 52:41
  the
• 52:41
  people
• 52:42
  who
• 52:42
  are
• 52:42
  most
• 52:42
  endangered
• 52:43
  are
• 52:44
  the
• 52:44
  mid
• 52:44
  level
• 52:44
  career
• 52:45
  people
• 52:45
  because
• 52:45
  I
• 52:46
  think
• 52:46
  the
• 52:46
  senior
• 52:46
  people
• 52:46
  are
• 52:47
  already
• 52:47
  senior.
• 52:48
  They're
• 52:48
  either
• 52:48
  gonna
• 52:49
  retire,
• 52:49
  they're
• 52:50
  already
• 52:50
  old,
• 52:50
  they're
• 52:51
  gonna
• 52:51
  quit
• 52:51
  or
• 52:51
  they're
• 52:52
  just
• 52:52
  gonna
• 52:52
  keep
• 52:52
  being
• 52:52
  very
• 52:52
  senior
• 52:53
  with
• 52:53
  all
• 52:53
  this
• 52:53
  knowledge.
• 52:54
  The
• 52:54
  young
• 52:55
  people,
• 52:55
  the
• 52:55
  juniors,
• 52:56
  they're
• 52:57
  fine,
• 52:57
  they've
• 52:57
  always
• 52:58
  been
• 52:58
  fine.
• 52:58
  That's
• 52:58
  the
• 52:58
  best
• 52:59
  place
• 52:59
  to
• 52:59
  be
• 52:59
  in
• 52:59
  tech,
• 52:59
  I
• 53:00
  think.
• 53:00
  You
• 53:00
  don't
• 53:00
  get
• 53:00
  paid
• 53:00
  very
• 53:00
  much,
• 53:01
  but
• 53:01
  the
• 53:01
  sky's
• 53:01
  the
• 53:01
  limit
• 53:02
  on
• 53:02
  what
• 53:02
  you're
• 53:02
  able
• 53:02
  to
• 53:02
  do,
• 53:02
  what
• 53:03
  career
• 53:03
  path
• 53:03
  you
• 53:03
  choose,
• 53:03
  et
• 53:04
  cetera.
• 53:04
  But
• 53:05
  the
• 53:05
  people
• 53:05
  who
• 53:05
  are
• 53:05
  mid
• 53:05
  career,
• 53:06
  I
• 53:06
  have
• 53:06
  a
• 53:06
  lot
• 53:06
  of
• 53:06
  concern
• 53:06
  for
• 53:07
  that.
• 53:07
  Because
• 53:08
  when
• 53:08
  you
• 53:08
  do
• 53:08
  layoffs,
• 53:09
  you
• 53:09
  don't
• 53:09
  lay
• 53:09
  off
• 53:09
  the
• 53:09
  cheap
• 53:10
  people
• 53:10
  and
• 53:11
  you
• 53:11
  don't
• 53:11
  lay
• 53:11
  off
• 53:11
  the
• 53:11
  people
• 53:11
  who
• 53:12
  know
• 53:12
  everything.
• 53:12
  You
• 53:12
  lay
• 53:12
  off
• 53:12
  the
• 53:13
  people
• 53:13
  who
• 53:13
  are
• 53:13
  expensive
• 53:13
  but
• 53:14
  aren't
• 53:14
  quite
• 53:14
  there
• 53:14
  yet.
• 53:15
  And
• 53:15
  that's
• 53:15
  my
• 53:16
  biggest
• 53:16
  fear,
• 53:16
  is
• 53:16
  those
• 53:17
  folks.
• 53:17
  And
• 53:17
  so
• 53:17
  when
• 53:18
  I
• 53:18
  think
• 53:18
  about
• 53:18
  that,
• 53:18
  I
• 53:18
  think
• 53:19
  I
• 53:19
  would
• 53:19
  be
• 53:19
  thinking
• 53:19
  a
• 53:19
  lot
• 53:20
  about
• 53:20
  how
• 53:20
  do
• 53:20
  I
• 53:21
  be
• 53:21
  very
• 53:21
  helpful
• 53:22
  to
• 53:22
  my
• 53:23
  organization
• 53:23
  while
• 53:24
  they're
• 53:24
  going
• 53:24
  through
• 53:24
  this
• 53:24
  change.
• 53:25
  Because
• 53:25
  it's
• 53:25
  not
• 53:25
  just
• 53:25
  an
• 53:26
  individual
• 53:26
  change
• 53:26
  either.
• 53:27
  Right.
• 53:27
  It's
• 53:27
  an
• 53:27
  organizational
• 53:28
  level
• 53:28
  change
• 53:28
  as
• 53:28
  well.
• 53:29
  And
• 53:29
  so
• 53:29
  that
• 53:29
  could
• 53:30
  be
• 53:30
  like
• 53:30
  finding
• 53:31
  places
• 53:31
  where
• 53:31
  it's
• 53:31
  just
• 53:32
  very
• 53:32
  hard.
• 53:32
  And
• 53:33
  you
• 53:33
  know,
• 53:33
  it
• 53:33
  could
• 53:33
  be
• 53:33
  a
• 53:33
  compliance
• 53:34
  place,
• 53:34
  it
• 53:34
  could
• 53:34
  be
• 53:34
  a
• 53:34
  reason,
• 53:35
  a
• 53:35
  regulatory
• 53:35
  reason
• 53:36
  why
• 53:36
  it's
• 53:36
  very
• 53:36
  hard
• 53:36
  to
• 53:36
  introduce
• 53:37
  these
• 53:37
  tools.
• 53:38
  It
• 53:38
  could
• 53:38
  be
• 53:38
  various
• 53:39
  things
• 53:39
  like
• 53:39
  that.
• 53:39
  It
• 53:40
  also
• 53:40
  could
• 53:40
  be
• 53:40
  maybe
• 53:40
  just
• 53:41
  becoming
• 53:41
  very
• 53:41
  good
• 53:41
  at
• 53:42
  these
• 53:42
  tools.
• 53:42
  It
• 53:42
  could
• 53:42
  be
• 53:43
  becoming
• 53:43
  a
• 53:43
  baker
• 53:44
  or
• 53:44
  a
• 53:44
  baker,
• 53:45
  not
• 53:45
  a
• 53:45
  banker,
• 53:45
  don't
• 53:45
  become
• 53:46
  a
• 53:46
  baker,
• 53:46
  but
• 53:46
  becoming
• 53:47
  a
• 53:47
  baker.
• 53:48
  It
• 53:48
  could
• 53:49
  be
• 53:49
  a
• 53:49
  lot
• 53:49
  of
• 53:49
  things.
• 53:49
  Like
• 53:50
  do
• 53:50
  something
• 53:51
  with
• 53:51
  your
• 53:51
  hands,
• 53:51
  do.
• 53:52
  Something
• 53:52
  with
• 53:52
  your
• 53:52
  hands.
• 53:53
  But
• 53:53
  I'm
• 53:54
  not
• 53:54
  sure
• 53:55
  what
• 53:55
  the
• 53:55
  right
• 53:55
  answer
• 53:55
  is.
• 53:56
  And
• 53:56
  this
• 53:56
  is
• 53:56
  one
• 53:57
  of
• 53:57
  the
• 53:57
  reasons
• 53:57
  why
• 53:57
  it's
• 53:57
  so
• 53:58
  scary.
• 53:59
  And
• 53:59
  so
• 53:59
  the
• 53:59
  reason
• 54:00
  why
• 54:00
  I
• 54:01
  really
• 54:01
  believe
• 54:01
  that
• 54:01
  this
• 54:02
  is
• 54:02
  not
• 54:02
  going
• 54:02
  to
• 54:02
  work
• 54:02
  out
• 54:02
  well
• 54:03
  for
• 54:03
  us
• 54:03
  is
• 54:03
  not
• 54:04
  because
• 54:05
  the
• 54:06
  AI
• 54:06
  is
• 54:06
  going
• 54:06
  to
• 54:06
  kill
• 54:07
  us.
• 54:07
  I
• 54:07
  think
• 54:07
  it's
• 54:08
  because
• 54:08
  the
• 54:08
  humans
• 54:09
  will
• 54:09
  kill
• 54:09
  us.
• 54:10
  That's
• 54:10
  the
• 54:10
  reason.
• 54:11
  And
• 54:11
  sure,
• 54:11
  that's
• 54:12
  probably
• 54:12
  like
• 54:12
  a
• 54:12
  slightly
• 54:13
  different
• 54:13
  P
• 54:13
  doom
• 54:14
  or
• 54:14
  whatever
• 54:14
  people
• 54:14
  call
• 54:14
  it
• 54:15
  than
• 54:15
  a
• 54:16
  lot
• 54:16
  of
• 54:16
  the
• 54:16
  Silicon
• 54:16
  Valley
• 54:17
  kind
• 54:17
  of
• 54:17
  AI
• 54:18
  Bros.
• 54:18
  But
• 54:19
  I
• 54:19
  think
• 54:19
  the
• 54:19
  outcome
• 54:19
  is
• 54:20
  the
• 54:20
  same.
• 54:20
  And
• 54:20
  I
• 54:21
  think
• 54:21
  the
• 54:21
  easy
• 54:21
  way
• 54:21
  to
• 54:21
  think
• 54:22
  about
• 54:22
  this
• 54:22
  and
• 54:22
  this
• 54:23
  just
• 54:23
  to
• 54:23
  get
• 54:23
  even
• 54:23
  more
• 54:24
  depressing
• 54:24
  is,
• 54:26
  is
• 54:26
  if
• 54:27
  everything
• 54:27
  is
• 54:27
  successful
• 54:28
  and
• 54:28
  it
• 54:28
  all
• 54:28
  works
• 54:30
  and
• 54:30
  we're
• 54:30
  suddenly
• 54:30
  all
• 54:31
  out
• 54:31
  of
• 54:31
  a
• 54:31
  jobs,
• 54:32
  what
• 54:32
  happens?
• 54:34
  The
• 54:34
  US
• 54:34
  is
• 54:34
  not
• 54:34
  very
• 54:35
  good
• 54:35
  at
• 54:35
  withstanding
• 54:37
  economic
• 54:37
  shakedowns.
• 54:38
  Right?
• 54:38
  It's
• 54:39
  not
• 54:39
  good.
• 54:40
  Or
• 54:40
  we've
• 54:40
  invested
• 54:41
  all
• 54:41
  of
• 54:41
  our
• 54:41
  money
• 54:42
  into
• 54:42
  AI
• 54:43
  and
• 54:43
  it
• 54:43
  all
• 54:44
  collapses
• 54:44
  because
• 54:45
  it's
• 54:45
  fake.
• 54:46
  Where
• 54:46
  are
• 54:47
  we?
• 54:47
  We're
• 54:47
  in
• 54:47
  the
• 54:47
  same
• 54:48
  place.
• 54:49
  We
• 54:49
  have
• 54:50
  economic
• 54:51
  catastrophe.
• 54:51
  So
• 54:52
  I
• 54:52
  guess
• 54:52
  we
• 54:52
  all
• 54:52
  just
• 54:52
  got
• 54:52
  to
• 54:52
  figure
• 54:53
  out
• 54:53
  how
• 54:53
  to
• 54:53
  get
• 54:53
  our
• 54:53
  castles.
• 54:54
  So
• 54:54
  start
• 54:54
  Companies
• 54:55
  try
• 54:55
  and
• 54:55
  get
• 54:55
  some
• 54:56
  wealth
• 54:56
  creat.
• 54:57
  There
• 54:57
  we
• 54:58
  go.
• 54:58
  And
• 54:58
  get
• 54:58
  a
• 54:59
  moat.
• 54:59
  I
• 54:59
  guess
• 55:00
  one
• 55:01
  thing
• 55:01
  I
• 55:01
  appreciate
• 55:01
  is
• 55:02
  in
• 55:02
  the
• 55:02
  world
• 55:03
  where
• 55:03
  these
• 55:03
  tools
• 55:04
  are
• 55:04
  making
• 55:04
  it
• 55:04
  easier
• 55:04
  for
• 55:05
  people
• 55:05
  that
• 55:05
  aren't
• 55:05
  engineers
• 55:06
  to
• 55:06
  become
• 55:07
  engineers.
• 55:08
  I
• 55:08
  think
• 55:08
  also
• 55:08
  it
• 55:09
  enables
• 55:10
  people
• 55:10
  that
• 55:10
  are
• 55:10
  engineers
• 55:11
  to
• 55:11
  build
• 55:11
  tools
• 55:12
  that
• 55:12
  make
• 55:12
  them
• 55:12
  better
• 55:13
  at
• 55:13
  the
• 55:13
  things
• 55:13
  they're
• 55:14
  bad
• 55:14
  at.
• 55:14
  I
• 55:15
  used
• 55:15
  your
• 55:16
  blog
• 55:16
  instructions
• 55:17
  to
• 55:17
  build
• 55:17
  up
• 55:17
  the
• 55:17
  email
• 55:18
  client.
• 55:19
  I
• 55:19
  am
• 55:19
  physically
• 55:20
  incapable
• 55:20
  of
• 55:20
  scheduling,
• 55:21
  especially
• 55:21
  across
• 55:22
  time
• 55:22
  zone.
• 55:23
  But
• 55:23
  Now
• 55:23
  I
• 55:23
  have
• 55:24
  five
• 55:24
  email
• 55:24
  drafts.
• 55:25
  That
• 55:25
  was
• 55:25
  great.
• 55:25
  It
• 55:25
  took
• 55:25
  me
• 55:26
  45
• 55:26
  minutes
• 55:27
  to
• 55:27
  set
• 55:27
  everything
• 55:27
  up
• 55:28
  and
• 55:28
  now
• 55:28
  I
• 55:28
  have
• 55:29
  an
• 55:29
  AI
• 55:30
  that
• 55:30
  schedules
• 55:30
  things
• 55:31
  for
• 55:31
  me,
• 55:31
  writes
• 55:32
  emails
• 55:32
  for
• 55:32
  me.
• 55:33
  It
• 55:33
  was
• 55:33
  great.
• 55:33
  And
• 55:33
  it
• 55:33
  was
• 55:34
  that
• 55:34
  experience
• 55:34
  of
• 55:34
  like
• 55:35
  I
• 55:35
  didn't
• 55:35
  have
• 55:35
  to
• 55:35
  code
• 55:35
  anything.
• 55:35
  I
• 55:36
  just,
• 55:36
  you
• 55:36
  know,
• 55:36
  did
• 55:36
  a
• 55:36
  bunch
• 55:37
  of
• 55:37
  commands,
• 55:37
  set
• 55:38
  up
• 55:38
  some
• 55:38
  MCPs
• 55:39
  and
• 55:39
  servers
• 55:40
  and
• 55:40
  it
• 55:41
  was,
• 55:41
  it
• 55:41
  was
• 55:41
  good
• 55:41
  to
• 55:41
  go.
• 55:42
  So
• 55:43
  hopefully
• 55:43
  people
• 55:43
  can
• 55:44
  learn
• 55:44
  how
• 55:44
  to
• 55:44
  do
• 55:44
  that
• 55:44
  too.
• 55:45
  This
• 55:45
  is
• 55:45
  the
• 55:45
  thing
• 55:46
  is
• 55:46
  every
• 55:47
  time
• 55:47
  I
• 55:48
  run
• 55:48
  into
• 55:48
  a
• 55:48
  thing
• 55:48
  where
• 55:48
  I
• 55:49
  do
• 55:49
  not
• 55:49
  know
• 55:49
  I
• 55:50
  have
• 55:50
  two
• 55:50
  options,
• 55:50
  right?
• 55:51
  One
• 55:51
  of
• 55:51
  the
• 55:51
  options
• 55:52
  is
• 55:52
  to
• 55:52
  just
• 55:52
  be
• 55:52
  like
• 55:53
  cloud.
• 55:53
  You
• 55:53
  do
• 55:53
  it.
• 55:55
  Great,
• 55:55
  love
• 55:56
  it.
• 55:56
  Second
• 55:56
  option
• 55:56
  is
• 55:57
  like,
• 55:57
  what
• 55:58
  happened?
• 55:59
  Then
• 55:59
  I
• 55:59
  bring
• 55:59
  that
• 56:00
  into
• 56:00
  my
• 56:00
  brain
• 56:00
  and
• 56:01
  then
• 56:01
  I
• 56:01
  learn
• 56:01
  how
• 56:01
  to
• 56:01
  do
• 56:01
  it
• 56:02
  and
• 56:02
  then
• 56:02
  inevitably
• 56:02
  I
• 56:02
  just
• 56:03
  ask
• 56:03
  Claude
• 56:03
  to
• 56:03
  do
• 56:03
  it.
• 56:03
  But
• 56:03
  now
• 56:04
  I
• 56:04
  know
• 56:05
  and
• 56:05
  then
• 56:05
  I
• 56:05
  can
• 56:05
  figure
• 56:06
  out
• 56:06
  that
• 56:06
  thing
• 56:06
  that
• 56:06
  I'm
• 56:06
  bad
• 56:07
  at.
• 56:07
  And
• 56:07
  this
• 56:07
  goes
• 56:07
  back
• 56:08
  to
• 56:08
  what
• 56:08
  we
• 56:08
  originally
• 56:09
  chartered
• 56:10
  to
• 56:10
  build
• 56:10
  here.
• 56:10
  And
• 56:10
  this
• 56:10
  is
• 56:11
  what
• 56:11
  I
• 56:11
  think
• 56:11
  is
• 56:11
  really
• 56:11
  interesting
• 56:12
  as
• 56:12
  we
• 56:13
  hone
• 56:13
  in
• 56:13
  our
• 56:13
  process
• 56:14
  and
• 56:14
  get
• 56:14
  the
• 56:14
  code
• 56:15
  gen
• 56:15
  process
• 56:15
  to
• 56:16
  be
• 56:16
  this
• 56:16
  great
• 56:16
  loop
• 56:16
  where
• 56:16
  we
• 56:17
  can
• 56:17
  drop
• 56:17
  any
• 56:17
  kind
• 56:17
  of
• 56:17
  thing
• 56:18
  in
• 56:18
  and
• 56:18
  get
• 56:18
  some
• 56:19
  really
• 56:19
  magical
• 56:19
  out.
• 56:19
  But
• 56:20
  we're
• 56:20
  finding
• 56:21
  ourselves
• 56:21
  just
• 56:22
  almost
• 56:22
  like
• 56:23
  some
• 56:23
  singularity
• 56:23
  getting
• 56:24
  closer
• 56:24
  and
• 56:24
  closer
• 56:25
  to
• 56:25
  what
• 56:25
  we
• 56:25
  originally
• 56:25
  pitched
• 56:27
  in
• 56:27
  this
• 56:27
  kind
• 56:27
  of
• 56:27
  fun
• 56:28
  way
• 56:28
  where
• 56:28
  we're
• 56:28
  like,
• 56:28
  okay,
• 56:29
  now
• 56:29
  we
• 56:29
  have
• 56:29
  this
• 56:29
  matrix
• 56:30
  chat
• 56:30
  where
• 56:31
  our
• 56:31
  friends
• 56:31
  who
• 56:31
  are
• 56:31
  using
• 56:31
  this
• 56:32
  can
• 56:32
  iterate
• 56:33
  in
• 56:33
  this
• 56:33
  make
• 56:33
  a
• 56:33
  channel.
• 56:34
  And
• 56:34
  then
• 56:34
  theoretically,
• 56:35
  which
• 56:35
  is
• 56:35
  one
• 56:36
  of
• 56:36
  the
• 56:36
  things
• 56:36
  we
• 56:36
  have
• 56:36
  to
• 56:36
  test
• 56:36
  soon,
• 56:37
  is
• 56:37
  you
• 56:37
  can
• 56:37
  add
• 56:37
  another
• 56:37
  person
• 56:38
  to
• 56:38
  that
• 56:38
  channel
• 56:38
  and
• 56:38
  then
• 56:38
  they
• 56:39
  get
• 56:39
  access
• 56:39
  to
• 56:39
  the
• 56:40
  same
• 56:40
  kind
• 56:40
  of
• 56:40
  all
• 56:40
  that
• 56:40
  information.
• 56:41
  And
• 56:41
  then
• 56:41
  how
• 56:41
  do
• 56:41
  you
• 56:42
  have
• 56:42
  memory
• 56:43
  there?
• 56:43
  And
• 56:43
  there's
• 56:44
  all
• 56:44
  this
• 56:44
  other
• 56:44
  stuff,
• 56:44
  but
• 56:44
  it's
• 56:45
  exactly
• 56:45
  that,
• 56:45
  Sarah,
• 56:45
  where
• 56:45
  it's
• 56:46
  like,
• 56:46
  I
• 56:46
  don't
• 56:46
  know
• 56:46
  how
• 56:46
  to
• 56:46
  do
• 56:47
  this,
• 56:47
  I
• 56:47
  don't
• 56:47
  know
• 56:47
  what
• 56:47
  the
• 56:47
  fuck,
• 56:48
  how
• 56:48
  to
• 56:48
  build
• 56:48
  this
• 56:48
  stuff.
• 56:48
  And
• 56:49
  it's
• 56:49
  just
• 56:49
  like,
• 56:49
  oh
• 56:49
  great,
• 56:50
  I
• 56:50
  didn't
• 56:50
  need
• 56:50
  to
• 56:50
  know
• 56:51
  how.
• 56:51
  More
• 56:51
  importantly,
• 56:52
  it
• 56:52
  took
• 56:52
  advantage
• 56:52
  of
• 56:52
  what
• 56:52
  I
• 56:53
  did
• 56:53
  know
• 56:53
  how
• 56:53
  to
• 56:53
  make
• 56:54
  that
• 56:54
  thing
• 56:54
  way
• 56:54
  better.
• 56:55
  Yeah,
• 56:57
  that's
• 56:57
  very
• 56:57
  nice.
• 56:58
  Well,
• 56:58
  great.
• 56:58
  Where
• 56:59
  can
• 56:59
  people
• 56:59
  find
• 56:59
  out
• 57:00
  more
• 57:00
  about
• 57:00
  what
• 57:00
  you're
• 57:00
  building?
• 57:01
  You
• 57:01
  can
• 57:01
  go
• 57:01
  to
• 57:02
  23.9ai,
• 57:03
  that's
• 57:03
  our
• 57:03
  website.
• 57:04
  And
• 57:04
  then
• 57:05
  one
• 57:05
  of
• 57:05
  the
• 57:06
  things
• 57:06
  that
• 57:06
  we
• 57:06
  are
• 57:06
  playing
• 57:07
  around
• 57:07
  with
• 57:07
  is
• 57:08
  so
• 57:08
  first
• 57:09
  of
• 57:09
  all,
• 57:09
  just
• 57:09
  a
• 57:09
  couple
• 57:09
  shout
• 57:09
  outs.
• 57:10
  So
• 57:10
  Jesse
• 57:11
  Vincent's
• 57:11
  superpowers
• 57:12
  is
• 57:12
  incredible.
• 57:13
  If
• 57:13
  you
• 57:13
  haven't
• 57:14
  used
• 57:14
  it
• 57:14
  with
• 57:14
  Claude
• 57:15
  code,
• 57:15
  it's
• 57:15
  really,
• 57:16
  really
• 57:16
  good.
• 57:16
  What
• 57:16
  it
• 57:17
  does
• 57:17
  is
• 57:17
  it
• 57:17
  allows
• 57:18
  you
• 57:18
  to
• 57:18
  avoid
• 57:18
  the
• 57:19
  speedrun
• 57:19
  to
• 57:20
  learn
• 57:20
  cogen,
• 57:21
  which
• 57:21
  I
• 57:21
  think
• 57:21
  is
• 57:21
  not
• 57:21
  worth
• 57:22
  learning
• 57:22
  anymore.
• 57:22
  Jesse
• 57:23
  Vistant
• 57:23
  like
• 57:23
  made
• 57:23
  the
• 57:24
  keyboard.
• 57:24
  Jesse
• 57:25
  Vincent.
• 57:25
  Yeah.
• 57:26
  Cool.
• 57:26
  The
• 57:27
  other
• 57:27
  thing
• 57:27
  that
• 57:27
  I
• 57:27
  highly
• 57:28
  recommend
• 57:29
  is
• 57:29
  if
• 57:29
  you
• 57:30
  build
• 57:30
  MCP
• 57:31
  servers,
• 57:32
  but
• 57:32
  my
• 57:32
  advice
• 57:33
  is
• 57:33
  to
• 57:33
  build
• 57:34
  them
• 57:34
  as
• 57:35
  CLIs
• 57:35
  for
• 57:35
  humans
• 57:37
  and
• 57:37
  then
• 57:38
  at
• 57:38
  the
• 57:38
  end
• 57:38
  of
• 57:38
  it
• 57:38
  just
• 57:39
  be
• 57:39
  like,
• 57:39
  also
• 57:39
  make
• 57:39
  it
• 57:39
  an
• 57:39
  MCP
• 57:40
  server.
• 57:40
  Yeah,
• 57:41
  interesting,
• 57:41
  because
• 57:42
  one
• 57:42
  of
• 57:42
  the
• 57:42
  things
• 57:42
  that
• 57:42
  I've
• 57:42
  noticed
• 57:43
  is
• 57:43
  as
• 57:43
  we
• 57:43
  build
• 57:43
  software
• 57:44
  utilities
• 57:45
  internally
• 57:45
  for
• 57:45
  us
• 57:46
  to
• 57:46
  use
• 57:46
  and
• 57:46
  primarily
• 57:47
  agents
• 57:47
  to
• 57:47
  use,
• 57:48
  we're
• 57:48
  ended
• 57:48
  up
• 57:48
  building
• 57:49
  the
• 57:49
  software
• 57:49
  for
• 57:49
  the
• 57:49
  agents.
• 57:50
  And
• 57:50
  that's
• 57:51
  a
• 57:51
  different
• 57:51
  frame,
• 57:51
  that's
• 57:52
  a
• 57:52
  different
• 57:52
  program,
• 57:52
  that's
• 57:52
  a
• 57:53
  different
• 57:53
  style
• 57:53
  of
• 57:54
  work.
• 57:55
  And
• 57:55
  so
• 57:55
  a
• 57:56
  CLI
• 57:56
  tool
• 57:56
  is
• 57:56
  closer
• 57:57
  to
• 57:57
  that
• 57:58
  than
• 57:59
  like
• 57:59
  a
• 57:59
  GUI
• 58:00
  tool,
• 58:00
  a
• 58:00
  web
• 58:01
  tool.
• 58:01
  And
• 58:01
  so
• 58:01
  as
• 58:01
  you
• 58:02
  build
• 58:02
  a
• 58:02
  CLI
• 58:02
  tool,
• 58:03
  it's
• 58:03
  easier
• 58:03
  to
• 58:03
  say
• 58:03
  this
• 58:04
  argument
• 58:04
  is
• 58:05
  a
• 58:05
  function
• 58:05
  that
• 58:05
  would
• 58:06
  exist
• 58:06
  within
• 58:06
  the
• 58:07
  MCP
• 58:07
  framework,
• 58:08
  but
• 58:08
  it's
• 58:08
  quicker
• 58:09
  for
• 58:09
  you
• 58:09
  to
• 58:09
  test
• 58:09
  because
• 58:09
  we
• 58:09
  don't
• 58:10
  have
• 58:10
  MCP
• 58:11
  clients
• 58:12
  as
• 58:12
  people
• 58:13
  and
• 58:13
  most
• 58:13
  of
• 58:14
  the
• 58:14
  demo
• 58:14
  ones
• 58:14
  are
• 58:14
  really
• 58:15
  bad.
• 58:15
  And
• 58:15
  so
• 58:16
  this
• 58:16
  is
• 58:16
  a
• 58:16
  really
• 58:16
  quick
• 58:16
  way
• 58:16
  that
• 58:17
  I've
• 58:17
  done
• 58:17
  for
• 58:17
  like,
• 58:17
  prototyping
• 58:18
  some
• 58:18
  stuff
• 58:19
  which
• 58:20
  has
• 58:20
  been
• 58:20
  really
• 58:20
  interesting.
• 58:21
  It's
• 58:21
  very
• 58:21
  cool.
• 58:22
  It's
• 58:22
  a
• 58:22
  cool
• 58:22
  time.
• 58:22
  I'm
• 58:23
  very
• 58:23
  happy
• 58:23
  to
• 58:23
  be
• 58:23
  here
• 58:24
  both
• 58:24
  in
• 58:24
  the
• 58:25
  podcast
• 58:25
  and
• 58:25
  working
• 58:26
  on
• 58:26
  this
• 58:26
  stuff.
• 58:26
  But
• 58:26
  boy,
• 58:27
  is
• 58:27
  it
• 58:27
  weird.
• 58:28
  Yeah,
• 58:28
  it's
• 58:29
  wild.
• 58:29
  It's
• 58:29
  certainly
• 58:30
  a
• 58:30
  ride.
• 58:31
  This
• 58:31
  has
• 58:31
  been
• 58:32
  so
• 58:33
  awesome
• 58:33
  having
• 58:34
  you
• 58:34
  on.
• 58:34
  Well,
• 58:35
  thank
• 58:35
  you
• 58:35
  for
• 58:35
  having
• 58:35
  me.
• 58:36
  I
• 58:36
  would
• 58:36
  love
• 58:36
  to
• 58:36
  be
• 58:36
  back
• 58:37
  if
• 58:37
  I
• 58:37
  can.
• 58:37
  We
• 58:37
  can
• 58:37
  check
• 58:38
  up
• 58:38
  with
• 58:38
  how
• 58:38
  our
• 58:39
  death
• 58:39
  is
• 58:39
  happening.
• 58:40
  Thank
• 58:40
  you
• 58:41
  so
• 58:41
  much.