Latest / Chaos Agents / Algorithmic Transference, Generative UI, and Why the Text Box Won - With Louise Macfadyen

Transcript

Algorithmic Transference, Generative UI, and Why the Text Box Won - With Louise Macfadyen · Chaos Agents

• 0:00
  There's
• 0:00
  this
• 0:00
  really
• 0:00
  interesting
• 0:01
  study
• 0:02
  about
• 0:02
  something
• 0:02
  called
• 0:02
  algorithmic
• 0:03
  transference,
• 0:04
  which
• 0:04
  is
• 0:04
  when
• 0:05
  users
• 0:06
  experience
• 0:07
  a
• 0:07
  negative
• 0:08
  interaction
• 0:08
  with
• 0:09
  a
• 0:09
  large
• 0:09
  language
• 0:09
  model,
• 0:10
  they
• 0:10
  tend
• 0:11
  to
• 0:11
  project
• 0:12
  that
• 0:12
  same
• 0:13
  problem
• 0:13
  or
• 0:13
  that
• 0:14
  same
• 0:14
  level
• 0:15
  of
• 0:15
  capability,
• 0:15
  like
• 0:16
  the
• 0:16
  disappointing
• 0:17
  level
• 0:17
  of
• 0:17
  capability
• 0:18
  onto
• 0:19
  every
• 0:19
  other
• 0:19
  model.
• 0:20
  Meaning,
• 0:20
  like,
• 0:20
  if
• 0:21
  I
• 0:21
  ask
• 0:21
  a
• 0:21
  model
• 0:22
  to
• 0:22
  do
• 0:22
  something
• 0:22
  and
• 0:23
  it
• 0:23
  doesn't
• 0:23
  do
• 0:23
  it
• 0:23
  right,
• 0:23
  then
• 0:24
  I'm
• 0:24
  like,
• 0:24
  oh,
• 0:24
  none
• 0:24
  of
• 0:24
  them
• 0:25
  can
• 0:25
  do
• 0:25
  this.
• 0:25
  AI
• 0:25
  sucks.
• 0:26
  Yeah,
• 0:26
  yeah.
• 0:26
  A
• 0:27
  little
• 0:27
  bit.
• 0:27
  Subconsciously,
• 0:28
  you're
• 0:28
  going
• 0:28
  to
• 0:28
  be
• 0:28
  like,
• 0:29
  oh,
• 0:29
  I'm
• 0:29
  not
• 0:29
  going
• 0:29
  to
• 0:29
  try
• 0:29
  doing
• 0:30
  that
• 0:30
  again.
• 0:34
  Welcome
• 0:35
  back
• 0:35
  to
• 0:35
  Chaos
• 0:35
  Agents,
• 0:36
  the
• 0:36
  show
• 0:36
  where
• 0:36
  we
• 0:37
  learn
• 0:37
  about
• 0:37
  AI
• 0:37
  so
• 0:38
  you
• 0:38
  don't
• 0:38
  have
• 0:38
  to.
• 0:39
  Becca,
• 0:40
  I'm
• 0:40
  sick.
• 0:40
  I
• 0:40
  have
• 0:40
  a
• 0:40
  crazy
• 0:41
  voice
• 0:41
  right
• 0:41
  now.
• 0:42
  It
• 0:42
  actually
• 0:42
  sounds
• 0:43
  like,
• 0:43
  good.
• 0:44
  Not
• 0:44
  that
• 0:44
  I,
• 0:44
  like,
• 0:45
  am
• 0:45
  happy
• 0:45
  that
• 0:45
  you're
• 0:45
  sick,
• 0:46
  but
• 0:46
  like,
• 0:47
  six
• 0:47
  hair
• 0:47
  sounds
• 0:48
  cool.
• 0:48
  Let
• 0:48
  me
• 0:48
  see
• 0:48
  if
• 0:48
  I
• 0:49
  could
• 0:49
  keep
• 0:49
  it
• 0:49
  this
• 0:49
  way.
• 0:49
  Well,
• 0:50
  okay.
• 0:50
  Well,
• 0:50
  I
• 0:51
  hope
• 0:51
  you
• 0:51
  feel
• 0:51
  better
• 0:52
  soon.
• 0:52
  And
• 0:52
  in
• 0:52
  the
• 0:52
  meantime,
• 0:53
  we
• 0:53
  get
• 0:53
  to
• 0:54
  enjoy
• 0:54
  your
• 0:54
  cool,
• 0:54
  raspy
• 0:55
  voice.
• 0:55
  So
• 0:56
  what's
• 0:56
  been
• 0:56
  going
• 0:56
  on?
• 0:57
  I've
• 0:57
  been
• 0:57
  learning
• 0:57
  some
• 0:58
  really
• 0:58
  crazy
• 0:58
  stuff
• 0:59
  recently.
• 0:59
  I've
• 1:00
  been
• 1:00
  paying
• 1:00
  a
• 1:00
  lot
• 1:00
  of
• 1:00
  attention
• 1:01
  to
• 1:01
  the
• 1:01
  news.
• 1:02
  With
• 1:03
  all
• 1:03
  the
• 1:04
  stuff
• 1:05
  that's
• 1:05
  been
• 1:05
  going
• 1:05
  on
• 1:05
  with
• 1:05
  Anthropic,
• 1:06
  it's
• 1:06
  been
• 1:07
  really
• 1:07
  interesting.
• 1:09
  I've
• 1:09
  moved
• 1:09
  away
• 1:09
  from.
• 1:10
  I
• 1:10
  was
• 1:10
  using
• 1:11
  ChatGPT
• 1:11
  and
• 1:12
  now
• 1:12
  I'm
• 1:12
  fully
• 1:13
  Anthropic.
• 1:14
  Tell
• 1:15
  me
• 1:15
  your
• 1:15
  quick
• 1:16
  take
• 1:16
  on
• 1:16
  what
• 1:17
  made
• 1:17
  you
• 1:18
  make
• 1:18
  the
• 1:18
  move.
• 1:19
  Honestly,
• 1:19
  it's
• 1:20
  them
• 1:20
  taking
• 1:20
  a
• 1:20
  stand.
• 1:21
  Like,
• 1:21
  I
• 1:21
  really
• 1:22
  thought
• 1:22
  it
• 1:22
  was
• 1:22
  very
• 1:22
  admirable.
• 1:23
  It's
• 1:23
  very
• 1:24
  hard,
• 1:24
  I
• 1:25
  think,
• 1:25
  in
• 1:25
  this
• 1:26
  industry
• 1:26
  where
• 1:27
  there's
• 1:27
  like,
• 1:27
  a
• 1:27
  race
• 1:28
  for
• 1:28
  capital,
• 1:29
  a
• 1:29
  race
• 1:29
  for
• 1:29
  power.
• 1:30
  Like,
• 1:31
  the
• 1:31
  one
• 1:31
  thing
• 1:31
  we've
• 1:32
  heard
• 1:32
  when
• 1:32
  we've
• 1:33
  been
• 1:33
  interviewing
• 1:33
  guests,
• 1:34
  especially
• 1:34
  people
• 1:35
  that
• 1:35
  know
• 1:35
  what's
• 1:35
  going
• 1:36
  on,
• 1:36
  is
• 1:37
  we
• 1:37
  hear
• 1:37
  the
• 1:37
  same
• 1:37
  thing
• 1:38
  of,
• 1:38
  like,
• 1:38
  when
• 1:38
  we
• 1:38
  say
• 1:39
  one
• 1:39
  to
• 1:39
  ten
• 1:39
  test.
• 1:40
  And
• 1:40
  a
• 1:40
  lot
• 1:40
  of
• 1:41
  times
• 1:41
  it's
• 1:41
  like,
• 1:41
  I'm
• 1:42
  a
• 1:42
  seven,
• 1:42
  but
• 1:43
  I'm
• 1:43
  worried
• 1:44
  about
• 1:44
  the
• 1:44
  people
• 1:44
  that
• 1:44
  are
• 1:44
  in
• 1:44
  charge.
• 1:45
  Yes.
• 1:46
  And
• 1:46
  that's
• 1:46
  how
• 1:46
  I
• 1:46
  felt
• 1:47
  pretty
• 1:47
  consistently.
• 1:48
  And
• 1:48
  we'll
• 1:48
  see
• 1:48
  what
• 1:48
  happens.
• 1:49
  It's
• 1:49
  an
• 1:49
  initial
• 1:49
  thing,
• 1:50
  but
• 1:50
  seeing
• 1:50
  Anthropic
• 1:51
  be
• 1:51
  like,
• 1:51
  we're
• 1:51
  not
• 1:52
  going
• 1:52
  to
• 1:52
  remove
• 1:52
  our
• 1:52
  values,
• 1:53
  that
• 1:53
  was
• 1:53
  really
• 1:53
  admirable.
• 1:54
  Yeah.
• 1:55
  Respect.
• 1:55
  Okay,
• 1:56
  let's
• 1:56
  get
• 1:56
  into,
• 1:57
  like,
• 1:57
  a
• 1:57
  quick
• 1:57
  recap
• 1:58
  of
• 1:58
  what
• 1:59
  the
• 2:00
  heck
• 2:00
  we're
• 2:00
  talking
• 2:00
  about
• 2:00
  with
• 2:00
  Anthropomorphism.
• 2:01
  Yeah.
• 2:02
  So,
• 2:03
  man,
• 2:03
  this
• 2:04
  was
• 2:04
  wild.
• 2:04
  So
• 2:05
  I
• 2:05
  feel
• 2:05
  like
• 2:05
  the
• 2:05
  majority
• 2:05
  of
• 2:06
  this
• 2:06
  went
• 2:06
  down
• 2:06
  on
• 2:07
  Friday.
• 2:07
  Basically,
• 2:09
  what
• 2:10
  happened
• 2:10
  was
• 2:11
  Department
• 2:12
  of
• 2:12
  Defense
• 2:13
  and
• 2:13
  the
• 2:13
  military
• 2:14
  and
• 2:14
  I'm
• 2:14
  sure
• 2:14
  our
• 2:14
  intelligence
• 2:15
  services
• 2:16
  have
• 2:16
  been
• 2:16
  using
• 2:17
  anthropic
• 2:19
  AI
• 2:20
  models
• 2:20
  in
• 2:20
  various
• 2:21
  different
• 2:21
  ways,
• 2:21
  and
• 2:22
  other
• 2:23
  big
• 2:23
  tech
• 2:23
  Company
• 2:23
  AI
• 2:24
  models
• 2:24
  as
• 2:24
  well.
• 2:25
  And
• 2:26
  the
• 2:26
  conflict
• 2:27
  came
• 2:27
  down
• 2:27
  to
• 2:28
  Pete
• 2:29
  Hegseth
• 2:30
  basically
• 2:30
  saying
• 2:30
  to
• 2:31
  the
• 2:31
  anthropic
• 2:31
  CEO,
• 2:32
  in
• 2:32
  our
• 2:33
  contract,
• 2:33
  it
• 2:34
  says
• 2:34
  we
• 2:34
  can
• 2:34
  use
• 2:34
  these
• 2:34
  for
• 2:34
  anything
• 2:35
  except,
• 2:36
  like,
• 2:36
  unlawful
• 2:38
  uses
• 2:39
  and
• 2:39
  uses
• 2:40
  that
• 2:40
  result
• 2:40
  in,
• 2:40
  like,
• 2:41
  inflicting,
• 2:41
  like,
• 2:42
  lethal
• 2:42
  harm
• 2:43
  on
• 2:43
  anyone,
• 2:44
  including
• 2:44
  potentially,
• 2:45
  like,
• 2:46
  Americans.
• 2:47
  He
• 2:47
  didn't,
• 2:47
  like,
• 2:47
  explicitly
• 2:48
  say
• 2:48
  that,
• 2:48
  but
• 2:48
  the
• 2:49
  language
• 2:49
  that
• 2:49
  he
• 2:49
  was
• 2:50
  asking
• 2:50
  for
• 2:51
  Anthropic
• 2:52
  to
• 2:53
  remove
• 2:54
  from
• 2:55
  the
• 2:55
  guardrails
• 2:56
  they
• 2:56
  put
• 2:56
  into
• 2:56
  this
• 2:57
  contract
• 2:57
  would
• 2:57
  essentially
• 2:58
  allow
• 2:58
  that.
• 2:59
  And
• 2:59
  the
• 2:59
  anthropic
• 3:00
  CEO
• 3:00
  was
• 3:00
  like,
• 3:00
  no,
• 3:02
  I'm
• 3:02
  sorry,
• 3:03
  but
• 3:03
  we
• 3:03
  really
• 3:04
  don't
• 3:04
  want
• 3:04
  our
• 3:05
  technology
• 3:05
  to
• 3:05
  be
• 3:05
  used
• 3:05
  to
• 3:06
  either
• 3:06
  surveil
• 3:07
  American
• 3:07
  citizens.
• 3:08
  We
• 3:09
  don't
• 3:09
  want
• 3:09
  it
• 3:09
  to
• 3:09
  be
• 3:10
  used
• 3:10
  to
• 3:11
  kill
• 3:12
  potentially
• 3:13
  American
• 3:13
  citizens.
• 3:14
  And
• 3:15
  Pete
• 3:15
  Hegseth
• 3:16
  was
• 3:16
  then
• 3:16
  like,
• 3:17
  okay,
• 3:17
  I'm
• 3:17
  gonna
• 3:18
  throw
• 3:18
  a
• 3:18
  tantrum
• 3:18
  now,
• 3:19
  and
• 3:20
  had
• 3:20
  Anthropic
• 3:21
  designated
• 3:21
  what
• 3:22
  is
• 3:22
  called
• 3:22
  a
• 3:22
  supply
• 3:23
  chain
• 3:23
  risk.
• 3:24
  My
• 3:24
  understanding
• 3:24
  of
• 3:24
  how
• 3:25
  this
• 3:25
  works
• 3:25
  is
• 3:25
  when
• 3:25
  the
• 3:26
  government
• 3:26
  signs
• 3:26
  contracts
• 3:27
  with
• 3:27
  vendors,
• 3:28
  obviously
• 3:28
  there's
• 3:29
  a
• 3:29
  lot
• 3:29
  of
• 3:29
  legal
• 3:29
  and
• 3:30
  vetting
• 3:30
  process
• 3:30
  that
• 3:31
  goes
• 3:31
  into,
• 3:31
  like,
• 3:32
  making
• 3:32
  sure
• 3:32
  something
• 3:33
  is
• 3:33
  in
• 3:33
  good
• 3:33
  standing
• 3:34
  for
• 3:34
  the
• 3:34
  government
• 3:34
  to
• 3:34
  be
• 3:34
  able
• 3:35
  to
• 3:35
  use
• 3:35
  their
• 3:36
  technology.
• 3:37
  When
• 3:37
  the
• 3:37
  government
• 3:38
  deems
• 3:38
  something
• 3:38
  a
• 3:39
  supply
• 3:39
  chain
• 3:39
  risk,
• 3:40
  it
• 3:40
  essentially
• 3:41
  is
• 3:41
  saying,
• 3:42
  like,
• 3:42
  there's
• 3:43
  something
• 3:43
  compromised
• 3:44
  with
• 3:44
  this
• 3:45
  company
• 3:45
  that
• 3:45
  means,
• 3:45
  like,
• 3:46
  it's
• 3:46
  risky
• 3:47
  to
• 3:47
  national
• 3:47
  security
• 3:47
  for
• 3:47
  us
• 3:48
  to
• 3:48
  be
• 3:48
  using
• 3:48
  this.
• 3:49
  Historically,
• 3:49
  in
• 3:49
  the
• 3:50
  past,
• 3:50
  it's
• 3:50
  been
• 3:50
  stuff
• 3:51
  like
• 3:52
  they're
• 3:52
  funded
• 3:52
  by
• 3:52
  the
• 3:53
  Russian
• 3:53
  government
• 3:54
  or,
• 3:54
  like,
• 3:54
  you
• 3:55
  know,
• 3:55
  things
• 3:55
  like
• 3:55
  that.
• 3:56
  Or
• 3:56
  maybe
• 3:56
  it's
• 3:56
  like,
• 3:57
  we
• 3:57
  did
• 3:57
  an
• 3:57
  audit
• 3:57
  of
• 3:58
  their
• 3:58
  infrastructure,
• 3:59
  and
• 3:59
  it's
• 3:59
  actually,
• 4:00
  like,
• 4:00
  incredibly
• 4:01
  unsecure,
• 4:01
  and
• 4:02
  so
• 4:02
  we
• 4:02
  can't
• 4:02
  use
• 4:02
  this.
• 4:03
  Nothing
• 4:04
  like
• 4:04
  that
• 4:04
  happened
• 4:04
  here.
• 4:05
  It's
• 4:06
  unclear,
• 4:06
  frankly,
• 4:07
  what
• 4:08
  the
• 4:08
  government's
• 4:09
  justification
• 4:09
  was
• 4:10
  for
• 4:10
  deeming
• 4:11
  Anthropic
• 4:11
  a
• 4:12
  supply
• 4:12
  chain
• 4:12
  risk,
• 4:12
  but
• 4:13
  they
• 4:13
  did,
• 4:14
  and
• 4:14
  we'll
• 4:14
  get
• 4:14
  to
• 4:14
  that
• 4:14
  in
• 4:15
  a
• 4:15
  second.
• 4:15
  That
• 4:15
  in
• 4:15
  itself
• 4:16
  is,
• 4:16
  like,
• 4:16
  potentially
• 4:17
  very
• 4:17
  problematic.
• 4:18
  Just
• 4:18
  the
• 4:18
  fact
• 4:19
  that
• 4:19
  the
• 4:19
  government
• 4:19
  can
• 4:20
  seemingly
• 4:22
  arbitrarily
• 4:24
  declare
• 4:24
  something
• 4:25
  a
• 4:25
  supply
• 4:26
  chain
• 4:26
  risk,
• 4:26
  there's
• 4:27
  not
• 4:27
  a
• 4:27
  lot
• 4:27
  of
• 4:28
  oversight.
• 4:29
  Correct?
• 4:30
  Like,
• 4:30
  yeah,
• 4:31
  not
• 4:31
  very
• 4:31
  hard
• 4:32
  for
• 4:32
  them
• 4:32
  to.
• 4:32
  It's
• 4:33
  kind
• 4:33
  of
• 4:33
  like
• 4:33
  just
• 4:33
  like,
• 4:33
  these
• 4:33
  are
• 4:33
  bad
• 4:34
  guys.
• 4:34
  Okay?
• 4:35
  So
• 4:35
  you
• 4:35
  would
• 4:35
  think,
• 4:36
  like,
• 4:36
  okay,
• 4:36
  that
• 4:36
  sucks
• 4:36
  for
• 4:37
  Anthropic
• 4:37
  because
• 4:37
  they're
• 4:38
  going
• 4:38
  to
• 4:38
  lose
• 4:38
  a
• 4:38
  lot
• 4:38
  of
• 4:38
  money
• 4:38
  in
• 4:39
  government
• 4:39
  contracts,
• 4:39
  whatever.
• 4:40
  But
• 4:40
  it's
• 4:41
  not
• 4:41
  just
• 4:41
  that.
• 4:42
  What
• 4:43
  deeming
• 4:43
  something
• 4:43
  a
• 4:44
  supply
• 4:44
  chain
• 4:44
  risk
• 4:44
  means
• 4:45
  is
• 4:45
  that
• 4:46
  any
• 4:46
  other
• 4:47
  entities
• 4:47
  that
• 4:48
  have
• 4:48
  contracts
• 4:48
  with
• 4:48
  the
• 4:49
  government
• 4:49
  who
• 4:50
  have
• 4:50
  ties
• 4:50
  to
• 4:51
  that
• 4:51
  company,
• 4:52
  who
• 4:52
  integrate
• 4:52
  with
• 4:53
  the
• 4:53
  technology
• 4:53
  of
• 4:53
  that
• 4:53
  company,
• 4:54
  have
• 4:55
  to
• 4:56
  completely
• 4:57
  divest
• 4:58
  or
• 4:59
  put
• 4:59
  walls
• 4:59
  up
• 5:00
  between
• 5:00
  them
• 5:00
  and
• 5:00
  that
• 5:00
  company
• 5:01
  in
• 5:01
  order
• 5:01
  to
• 5:02
  remain
• 5:02
  in
• 5:03
  good
• 5:03
  standing
• 5:03
  with
• 5:03
  the
• 5:03
  government.
• 5:04
  Folks
• 5:04
  who
• 5:05
  have
• 5:05
  a
• 5:05
  significant
• 5:06
  amount
• 5:06
  of
• 5:06
  equity
• 5:07
  in
• 5:07
  Anthropic
• 5:08
  or
• 5:08
  who
• 5:08
  have
• 5:08
  very
• 5:09
  close
• 5:09
  partnerships
• 5:10
  with
• 5:10
  Anthropic
• 5:12
  had
• 5:12
  to
• 5:13
  emergency
• 5:14
  meet
• 5:14
  and
• 5:15
  basically
• 5:15
  decide,
• 5:16
  do
• 5:16
  we
• 5:16
  have
• 5:16
  to
• 5:17
  divest
• 5:17
  all
• 5:18
  of
• 5:18
  our
• 5:19
  holdings
• 5:19
  and
• 5:20
  interest
• 5:20
  in
• 5:21
  Anthropic
• 5:21
  in
• 5:22
  order
• 5:22
  to
• 5:22
  not
• 5:23
  be
• 5:23
  cut
• 5:23
  off
• 5:23
  by
• 5:24
  the
• 5:24
  government
• 5:24
  using
• 5:24
  us
• 5:25
  as
• 5:25
  well?
• 5:25
  Like,
• 5:25
  how
• 5:26
  does
• 5:26
  this
• 5:26
  work?
• 5:26
  And
• 5:27
  so
• 5:27
  it
• 5:27
  caused
• 5:27
  this
• 5:28
  sort
• 5:28
  of
• 5:28
  ripple
• 5:28
  effect.
• 5:29
  The
• 5:30
  other
• 5:30
  effect
• 5:30
  also
• 5:31
  is
• 5:31
  it
• 5:31
  essentially
• 5:32
  is
• 5:32
  the
• 5:32
  government
• 5:32
  strong
• 5:33
  arming
• 5:33
  Anthropic
• 5:34
  and
• 5:34
  saying,
• 5:34
  hey,
• 5:34
  if
• 5:35
  you
• 5:35
  don't
• 5:35
  let
• 5:35
  us
• 5:36
  use
• 5:36
  your
• 5:36
  technology
• 5:36
  in
• 5:37
  all
• 5:37
  the
• 5:37
  ways
• 5:37
  we
• 5:37
  want
• 5:37
  it
• 5:38
  to,
• 5:38
  we're
• 5:38
  to
• 5:38
  going
• 5:39
  basically
• 5:39
  going
• 5:39
  to
• 5:39
  blacklist
• 5:40
  you.
• 5:40
  And
• 5:40
  it's
• 5:41
  not
• 5:41
  just
• 5:41
  going
• 5:41
  to
• 5:42
  mean
• 5:42
  that
• 5:42
  you
• 5:42
  can't
• 5:43
  work
• 5:43
  with
• 5:43
  us.
• 5:43
  It's
• 5:43
  going
• 5:43
  to
• 5:43
  mean
• 5:44
  all
• 5:44
  these
• 5:44
  other
• 5:44
  investors
• 5:45
  in
• 5:45
  you
• 5:46
  have
• 5:46
  to
• 5:47
  get
• 5:47
  rid
• 5:48
  of
• 5:48
  their
• 5:48
  stock
• 5:48
  in
• 5:48
  you
• 5:48
  or
• 5:49
  remove
• 5:49
  themselves
• 5:50
  from
• 5:50
  working
• 5:50
  with
• 5:50
  you.
• 5:51
  And
• 5:51
  it
• 5:51
  completely
• 5:52
  isolates
• 5:52
  that
• 5:53
  company.
• 5:53
  And
• 5:54
  it's
• 5:54
  quite
• 5:54
  worrisome
• 5:55
  for
• 5:56
  the
• 5:56
  precedent
• 5:57
  that
• 5:57
  that
• 5:57
  sets.
• 5:58
  Yeah,
• 5:59
  it
• 5:59
  seems
• 5:59
  like
• 5:59
  Anthropic
• 6:00
  isn't
• 6:00
  backing
• 6:01
  down,
• 6:01
  which
• 6:02
  is
• 6:02
  really
• 6:03
  admirable.
• 6:04
  The
• 6:04
  thing
• 6:04
  that
• 6:04
  scared
• 6:05
  me
• 6:05
  the
• 6:05
  most
• 6:05
  was
• 6:06
  the
• 6:07
  idea
• 6:07
  of
• 6:07
  using
• 6:08
  AI
• 6:08
  as
• 6:08
  it
• 6:09
  exists
• 6:09
  today
• 6:09
  and
• 6:09
  autonomous
• 6:10
  weapons.
• 6:10
  I
• 6:11
  mean,
• 6:11
  that's
• 6:11
  all
• 6:11
  the
• 6:12
  sci
• 6:12
  fi,
• 6:12
  right?
• 6:13
  A
• 6:14
  hundred
• 6:14
  percent.
• 6:15
  Like
• 6:15
  every
• 6:15
  single
• 6:16
  one.
• 6:16
  It's
• 6:16
  like.
• 6:16
  And
• 6:17
  then
• 6:17
  we
• 6:17
  gave
• 6:17
  them
• 6:17
  access
• 6:18
  to
• 6:18
  autonomous
• 6:18
  weapons
• 6:19
  and
• 6:20
  we've
• 6:20
  interacted
• 6:21
  with
• 6:21
  it
• 6:21
  a
• 6:21
  lot
• 6:21
  and
• 6:21
  it
• 6:21
  makes
• 6:21
  great
• 6:22
  decisions.
• 6:22
  However,
• 6:23
  sometimes
• 6:24
  that's
• 6:24
  not
• 6:25
  true.
• 6:25
  And
• 6:25
  so
• 6:26
  just
• 6:26
  like
• 6:26
  a
• 6:26
  human.
• 6:27
  But
• 6:27
  not
• 6:27
  having
• 6:27
  those
• 6:28
  checks
• 6:28
  and
• 6:28
  balances
• 6:28
  is
• 6:29
  really
• 6:29
  scary.
• 6:30
  100%.
• 6:31
  Also,
• 6:31
  just
• 6:31
  to
• 6:32
  correct
• 6:32
  my
• 6:33
  earlier
• 6:33
  statement,
• 6:33
  it
• 6:34
  was
• 6:34
  two
• 6:34
  companies
• 6:35
  that
• 6:35
  have
• 6:36
  really
• 6:36
  close
• 6:36
  ties
• 6:36
  with
• 6:37
  and
• 6:37
  our
• 6:37
  major
• 6:37
  investors
• 6:38
  in
• 6:38
  Anthropic
• 6:39
  are
• 6:39
  Nvidia
• 6:39
  and
• 6:39
  Amazon.
• 6:40
  So
• 6:40
  those
• 6:41
  were
• 6:41
  the
• 6:41
  two
• 6:41
  companies
• 6:42
  that
• 6:42
  basically,
• 6:42
  if
• 6:42
  the
• 6:43
  government
• 6:43
  deemed
• 6:43
  Anthropic
• 6:44
  a
• 6:44
  supply
• 6:45
  chain
• 6:45
  risk,
• 6:45
  they're
• 6:45
  saying
• 6:45
  to
• 6:46
  Amazon,
• 6:46
  Nvidia,
• 6:47
  you
• 6:47
  can't
• 6:47
  work
• 6:47
  with
• 6:48
  us
• 6:48
  if
• 6:48
  you're
• 6:48
  financially
• 6:49
  tied
• 6:49
  to
• 6:49
  this
• 6:49
  entity.
• 6:50
  So
• 6:51
  they're
• 6:51
  forcing
• 6:51
  them
• 6:51
  to
• 6:51
  kind
• 6:52
  of
• 6:52
  choose
• 6:52
  between
• 6:52
  those
• 6:52
  two
• 6:53
  things.
• 6:53
  I
• 6:53
  totally
• 6:53
  agree
• 6:54
  with
• 6:54
  you.
• 6:54
  I
• 6:54
  think
• 6:54
  it's
• 6:55
  existentially
• 6:56
  important
• 6:57
  for
• 6:57
  Anthropic
• 6:58
  to
• 6:58
  stay
• 6:58
  strong
• 6:59
  here.
• 6:59
  Like,
• 7:00
  hold
• 7:00
  the
• 7:00
  line.
• 7:01
  Yes.
• 7:01
  Because
• 7:01
  if
• 7:02
  they
• 7:02
  capitulate,
• 7:03
  then
• 7:04
  it
• 7:04
  sets
• 7:04
  a
• 7:05
  precedent
• 7:05
  that
• 7:06
  the
• 7:06
  government
• 7:07
  can
• 7:07
  basically
• 7:08
  strong
• 7:09
  arm
• 7:10
  companies
• 7:10
  into
• 7:12
  letting
• 7:12
  them
• 7:12
  use
• 7:12
  their
• 7:13
  technology
• 7:13
  in
• 7:13
  whatever
• 7:14
  way
• 7:14
  they
• 7:14
  want.
• 7:14
  And
• 7:15
  that's
• 7:16
  quite
• 7:17
  scary
• 7:17
  given
• 7:18
  how
• 7:19
  powerful
• 7:19
  this
• 7:19
  technology
• 7:20
  is
• 7:20
  in
• 7:20
  like
• 7:20
  a
• 7:21
  way
• 7:21
  that
• 7:21
  the
• 7:22
  risk
• 7:22
  and
• 7:23
  the
• 7:23
  implications
• 7:23
  are
• 7:24
  so
• 7:24
  much
• 7:24
  greater
• 7:25
  than
• 7:25
  they've
• 7:25
  ever
• 7:25
  been
• 7:25
  before
• 7:26
  for
• 7:26
  that.
• 7:27
  So
• 7:27
  scary.
• 7:28
  Scary
• 7:29
  also
• 7:30
  in
• 7:30
  the
• 7:30
  world
• 7:30
  of
• 7:30
  sci
• 7:30
  fi,
• 7:31
  I
• 7:31
  thought
• 7:31
  something
• 7:32
  interesting
• 7:32
  to
• 7:32
  talk
• 7:32
  about.
• 7:32
  I
• 7:33
  learned
• 7:33
  about
• 7:33
  something
• 7:33
  crazy
• 7:34
  this
• 7:34
  weekend.
• 7:34
  Okay,
• 7:35
  tell
• 7:35
  me.
• 7:35
  It's
• 7:36
  Definitely
• 7:36
  one
• 7:36
  of
• 7:36
  those
• 7:36
  things
• 7:37
  where
• 7:37
  you're
• 7:37
  like,
• 7:37
  whoa,
• 7:38
  is
• 7:38
  this
• 7:38
  really
• 7:38
  happening
• 7:40
  recently?
• 7:40
  Over
• 7:40
  the
• 7:40
  past
• 7:41
  few
• 7:41
  years,
• 7:41
  you
• 7:41
  may
• 7:41
  have
• 7:42
  read
• 7:42
  about
• 7:42
  this.
• 7:43
  Scientists
• 7:43
  have
• 7:44
  learned
• 7:44
  that
• 7:44
  using
• 7:44
  stem
• 7:45
  cells
• 7:45
  from
• 7:45
  skin,
• 7:46
  and
• 7:46
  then
• 7:46
  from
• 7:46
  those
• 7:47
  stem
• 7:47
  cells,
• 7:47
  they
• 7:48
  can
• 7:48
  grow
• 7:49
  brain
• 7:49
  cells
• 7:50
  or
• 7:50
  collections
• 7:51
  of
• 7:51
  brain
• 7:51
  cells.
• 7:52
  Those
• 7:52
  are
• 7:53
  called
• 7:53
  organoids.
• 7:54
  It's
• 7:54
  like
• 7:54
  a
• 7:54
  3D
• 7:54
  cluster
• 7:55
  of
• 7:55
  little
• 7:55
  cells,
• 7:56
  and
• 7:56
  it
• 7:57
  self
• 7:57
  organizes,
• 7:58
  and
• 7:58
  it's
• 7:58
  kind
• 7:58
  of
• 7:59
  like
• 7:59
  early
• 7:59
  brain
• 7:59
  tissue.
• 8:00
  So
• 8:00
  it's
• 8:00
  like
• 8:00
  a
• 8:00
  part
• 8:01
  of
• 8:01
  a
• 8:01
  brain.
• 8:01
  It's
• 8:01
  not
• 8:01
  a
• 8:02
  full
• 8:02
  brain.
• 8:02
  It's
• 8:02
  a
• 8:03
  part
• 8:03
  of
• 8:03
  a
• 8:03
  brain.
• 8:03
  They
• 8:04
  form
• 8:04
  actual
• 8:05
  neural
• 8:05
  networks.
• 8:06
  Like,
• 8:06
  they
• 8:06
  can
• 8:06
  see
• 8:07
  the.
• 8:08
  Very
• 8:08
  similarly
• 8:08
  to,
• 8:09
  like,
• 8:09
  if
• 8:09
  you
• 8:09
  put
• 8:09
  an
• 8:09
  EEG
• 8:10
  and
• 8:10
  hook
• 8:10
  it
• 8:10
  up
• 8:10
  to
• 8:10
  your
• 8:11
  brain,
• 8:11
  they
• 8:11
  can
• 8:12
  see
• 8:12
  the
• 8:12
  neurons
• 8:13
  firing
• 8:13
  in
• 8:14
  these
• 8:14
  tiny
• 8:14
  brain
• 8:15
  cells,
• 8:15
  which
• 8:15
  is
• 8:16
  very
• 8:16
  cool
• 8:16
  on
• 8:16
  its
• 8:16
  own.
• 8:17
  It's
• 8:17
  been
• 8:17
  really
• 8:17
  helpful
• 8:18
  for
• 8:18
  studying
• 8:18
  things
• 8:19
  like
• 8:19
  brain
• 8:19
  cancer
• 8:20
  and
• 8:20
  things
• 8:20
  like
• 8:20
  that.
• 8:21
  They
• 8:21
  don't
• 8:21
  know
• 8:22
  why
• 8:22
  they
• 8:22
  have
• 8:22
  the
• 8:23
  shelf
• 8:23
  life
• 8:23
  that
• 8:23
  they
• 8:23
  do,
• 8:23
  but
• 8:24
  they
• 8:24
  live
• 8:24
  for
• 8:24
  about
• 8:24
  four
• 8:25
  months.
• 8:25
  And
• 8:26
  there
• 8:26
  are
• 8:26
  some
• 8:26
  companies
• 8:27
  now.
• 8:28
  One's
• 8:28
  called
• 8:28
  Neuron
• 8:29
  Gaming.
• 8:29
  There's
• 8:30
  another
• 8:30
  one
• 8:30
  called
• 8:30
  C1.
• 8:31
  There's
• 8:32
  another
• 8:32
  one
• 8:32
  called
• 8:33
  Final
• 8:33
  Spark.
• 8:34
  And
• 8:34
  what
• 8:34
  they've
• 8:35
  been
• 8:35
  doing
• 8:35
  is
• 8:35
  they've
• 8:36
  been
• 8:36
  taking
• 8:36
  these
• 8:36
  tiny
• 8:37
  clusters
• 8:37
  and
• 8:37
  hooking
• 8:38
  them
• 8:38
  up
• 8:38
  to
• 8:39
  computers.
• 8:40
  So,
• 8:40
  like,
• 8:40
  making
• 8:41
  them
• 8:41
  to
• 8:41
  a
• 8:42
  microchip
• 8:42
  to
• 8:42
  make
• 8:43
  them
• 8:43
  a
• 8:43
  full
• 8:43
  computer
• 8:44
  brain
• 8:44
  interface,
• 8:45
  similar
• 8:45
  to
• 8:46
  how
• 8:46
  AIs
• 8:47
  work
• 8:47
  with
• 8:48
  computer
• 8:48
  neurons.
• 8:49
  They're
• 8:49
  hooking
• 8:50
  up
• 8:50
  the
• 8:50
  brain
• 8:50
  neurons
• 8:51
  to
• 8:51
  the
• 8:51
  same
• 8:51
  systems,
• 8:52
  which
• 8:53
  is
• 8:53
  fascinating.
• 8:54
  And
• 8:54
  they
• 8:54
  have
• 8:54
  a
• 8:55
  few
• 8:55
  things
• 8:55
  going
• 8:55
  on.
• 8:55
  There's
• 8:56
  two
• 8:56
  companies
• 8:56
  that
• 8:57
  are
• 8:57
  just
• 8:57
  using
• 8:57
  them
• 8:58
  to
• 8:58
  play
• 8:58
  games.
• 8:59
  Pong
• 8:59
  is
• 9:00
  the
• 9:00
  first
• 9:00
  one.
• 9:00
  The
• 9:01
  brain
• 9:01
  clusters
• 9:01
  will
• 9:02
  respond
• 9:03
  to
• 9:03
  stimuli,
• 9:04
  like
• 9:04
  external
• 9:05
  stimuli.
• 9:05
  And
• 9:05
  basically
• 9:06
  what
• 9:06
  happens
• 9:06
  is
• 9:07
  the
• 9:07
  computer
• 9:07
  reacts
• 9:08
  to
• 9:08
  the
• 9:09
  electrical
• 9:09
  spikes
• 9:10
  from
• 9:10
  the
• 9:10
  neurons,
• 9:11
  and
• 9:12
  the
• 9:12
  computer
• 9:13
  will
• 9:13
  send
• 9:13
  signals
• 9:13
  to
• 9:14
  the
• 9:14
  neurons.
• 9:14
  So
• 9:14
  that's
• 9:14
  how,
• 9:15
  like,
• 9:15
  the
• 9:15
  back
• 9:15
  and
• 9:15
  forth
• 9:16
  works.
• 9:17
  I'm
• 9:17
  so
• 9:17
  curious
• 9:18
  about,
• 9:18
  like,
• 9:18
  the
• 9:19
  actual
• 9:20
  hardware
• 9:21
  that
• 9:21
  hooks
• 9:22
  into
• 9:23
  cells.
• 9:24
  The
• 9:24
  only
• 9:25
  thing
• 9:25
  I
• 9:25
  can
• 9:25
  imagine
• 9:25
  is
• 9:25
  just,
• 9:26
  like,
• 9:26
  tiny,
• 9:26
  like,
• 9:26
  part
• 9:26
  of
• 9:26
  a
• 9:27
  brain,
• 9:27
  like,
• 9:27
  taped
• 9:27
  inside.
• 9:28
  I'm
• 9:28
  picturing,
• 9:29
  like,
• 9:29
  a
• 9:29
  tiny
• 9:29
  wire.
• 9:30
  Like,
• 9:30
  we're
• 9:31
  talking
• 9:31
  cells.
• 9:32
  So
• 9:32
  they're
• 9:32
  extremely
• 9:33
  small.
• 9:33
  I
• 9:33
  don't
• 9:34
  know.
• 9:34
  We'll
• 9:34
  do
• 9:34
  some
• 9:34
  research
• 9:35
  now.
• 9:35
  I'm
• 9:36
  like,
• 9:36
  we
• 9:36
  got
• 9:36
  to
• 9:36
  go
• 9:36
  down
• 9:36
  the
• 9:37
  BCI
• 9:37
  rabbit
• 9:38
  hole.
• 9:38
  Yeah.
• 9:39
  I'm
• 9:39
  looking
• 9:39
  at
• 9:39
  pictures
• 9:39
  right
• 9:40
  now,
• 9:40
  and
• 9:40
  it
• 9:40
  seems
• 9:40
  like
• 9:41
  it's
• 9:41
  very
• 9:41
  different
• 9:41
  than
• 9:42
  pieces
• 9:42
  of
• 9:42
  brain
• 9:42
  taped
• 9:43
  to
• 9:43
  a
• 9:44
  microprocessor.
• 9:45
  Shocking.
• 9:46
  It
• 9:46
  looks
• 9:46
  like,
• 9:46
  basically,
• 9:47
  like
• 9:47
  the
• 9:47
  cells
• 9:48
  are
• 9:48
  stored
• 9:49
  in
• 9:50
  refrigerated
• 9:51
  equipment,
• 9:51
  and
• 9:52
  there
• 9:52
  are
• 9:52
  wires
• 9:53
  coming
• 9:53
  from
• 9:53
  the
• 9:53
  cells
• 9:54
  moving
• 9:54
  to
• 9:54
  the
• 9:55
  computer
• 9:55
  and
• 9:55
  vice
• 9:56
  versa.
• 9:56
  I
• 9:56
  guess
• 9:57
  that
• 9:57
  makes
• 9:57
  sense.
• 9:57
  You
• 9:57
  just
• 9:57
  need
• 9:58
  some
• 9:58
  sort
• 9:58
  of,
• 9:58
  like,
• 9:58
  electrical,
• 9:59
  like,
• 9:59
  node
• 10:00
  and
• 10:00
  you
• 10:01
  know
• 10:01
  how
• 10:01
  sometimes
• 10:01
  you
• 10:01
  see
• 10:02
  brain
• 10:02
  operations
• 10:03
  where
• 10:03
  the
• 10:03
  person's
• 10:04
  awake
• 10:04
  so
• 10:04
  that
• 10:04
  they
• 10:05
  can
• 10:05
  stimulate
• 10:06
  different
• 10:06
  parts
• 10:07
  of
• 10:07
  their
• 10:08
  brain
• 10:08
  like
• 10:09
  they're
• 10:09
  using,
• 10:09
  I'm
• 10:10
  assuming,
• 10:10
  some
• 10:11
  instrument
• 10:11
  that
• 10:12
  is
• 10:12
  like
• 10:12
  sending
• 10:13
  electrical
• 10:13
  signals
• 10:14
  into
• 10:14
  your
• 10:14
  brain.
• 10:15
  So
• 10:15
  theoretically,
• 10:16
  I
• 10:16
  guess
• 10:16
  it's
• 10:17
  not
• 10:17
  as
• 10:17
  crazy
• 10:18
  as
• 10:18
  I
• 10:18
  initially
• 10:19
  was
• 10:19
  thinking.
• 10:20
  I
• 10:20
  think
• 10:20
  what's
• 10:21
  interesting
• 10:21
  is
• 10:21
  they're
• 10:21
  really
• 10:22
  hoping
• 10:22
  it
• 10:22
  can
• 10:22
  help
• 10:23
  with
• 10:23
  the
• 10:24
  power
• 10:24
  issues,
• 10:25
  the
• 10:26
  necessary
• 10:27
  electricity
• 10:27
  needed
• 10:28
  to
• 10:28
  run
• 10:29
  AI
• 10:29
  Just
• 10:30
  because
• 10:30
  the
• 10:30
  human
• 10:30
  brain,
• 10:31
  the
• 10:31
  way
• 10:31
  that
• 10:31
  our
• 10:32
  neurons
• 10:32
  work,
• 10:32
  are
• 10:33
  so
• 10:33
  efficient
• 10:33
  with
• 10:33
  power.
• 10:34
  It
• 10:34
  really
• 10:35
  made
• 10:35
  me
• 10:35
  think
• 10:35
  about
• 10:36
  it's
• 10:36
  an
• 10:36
  ethical
• 10:36
  nightmare.
• 10:37
  It
• 10:37
  feels
• 10:37
  like
• 10:38
  it's
• 10:38
  a
• 10:38
  really
• 10:39
  interesting
• 10:39
  thing.
• 10:39
  Like,
• 10:40
  if
• 10:40
  today
• 10:41
  someone
• 10:41
  to
• 10:41
  you
• 10:41
  is
• 10:42
  like,
• 10:42
  listen,
• 10:43
  we
• 10:43
  can
• 10:43
  use
• 10:43
  human
• 10:44
  brain
• 10:44
  cells
• 10:45
  to
• 10:45
  completely
• 10:47
  remove
• 10:47
  the
• 10:47
  need
• 10:48
  for
• 10:48
  these
• 10:48
  monster
• 10:49
  data
• 10:49
  centers
• 10:49
  powering
• 10:50
  the
• 10:50
  Earth.
• 10:50
  What
• 10:51
  questions
• 10:51
  would
• 10:51
  you
• 10:51
  have?
• 10:52
  My
• 10:52
  first
• 10:52
  question
• 10:52
  would
• 10:53
  be
• 10:53
  like,
• 10:53
  how
• 10:54
  are
• 10:54
  we
• 10:55
  creating
• 10:56
  and
• 10:56
  producing
• 10:56
  these
• 10:57
  brain
• 10:57
  cells?
• 10:58
  If
• 10:58
  it's
• 10:58
  like,
• 10:59
  well,
• 10:59
  we
• 10:59
  have
• 10:59
  a
• 11:00
  straightforward
• 11:00
  process
• 11:01
  for
• 11:01
  using
• 11:02
  stem
• 11:03
  cells
• 11:03
  to
• 11:04
  create
• 11:04
  brain
• 11:04
  cells.
• 11:05
  This
• 11:05
  number
• 11:05
  of
• 11:06
  brain
• 11:06
  cells
• 11:06
  can
• 11:07
  power
• 11:07
  the
• 11:07
  equivalent
• 11:08
  of
• 11:08
  whatever
• 11:09
  system,
• 11:10
  and
• 11:10
  it's
• 11:10
  incredibly
• 11:12
  efficient.
• 11:12
  Offhand,
• 11:13
  I
• 11:13
  don't
• 11:13
  have
• 11:14
  any
• 11:14
  major
• 11:15
  ethical
• 11:15
  constraints
• 11:16
  with
• 11:16
  that.
• 11:16
  But
• 11:16
  that's
• 11:16
  also
• 11:16
  because
• 11:17
  I
• 11:17
  already
• 11:17
  sort
• 11:18
  of
• 11:18
  fall
• 11:18
  into
• 11:18
  the
• 11:18
  camp
• 11:19
  of,
• 11:19
  like,
• 11:20
  I
• 11:20
  support
• 11:21
  stem
• 11:21
  cell
• 11:22
  research,
• 11:22
  and
• 11:22
  I
• 11:22
  think
• 11:22
  it's
• 11:23
  like
• 11:23
  an
• 11:23
  incredible
• 11:23
  biological
• 11:25
  tool
• 11:26
  we
• 11:26
  have
• 11:26
  at
• 11:26
  our
• 11:26
  disposal
• 11:27
  for
• 11:28
  pioneering
• 11:28
  stuff
• 11:29
  like
• 11:29
  that.
• 11:29
  I'm
• 11:29
  sure
• 11:29
  there
• 11:30
  are
• 11:30
  adverse
• 11:30
  consequences
• 11:31
  that
• 11:31
  I'm.
• 11:31
  That
• 11:31
  are
• 11:31
  not
• 11:32
  directly
• 11:32
  coming
• 11:32
  to
• 11:32
  mind.
• 11:32
  What
• 11:33
  would
• 11:33
  you
• 11:33
  say?
• 11:33
  Like,
• 11:34
  I'm
• 11:34
  wondering
• 11:34
  how
• 11:34
  much
• 11:35
  of
• 11:35
  a
• 11:35
  human
• 11:35
  brain
• 11:36
  it
• 11:36
  is,
• 11:36
  right?
• 11:37
  Like,
• 11:37
  does
• 11:37
  it
• 11:37
  have
• 11:38
  a
• 11:38
  full
• 11:38
  frontal
• 11:38
  cortex
• 11:39
  or
• 11:39
  is
• 11:39
  it
• 11:40
  just
• 11:40
  a
• 11:40
  collection
• 11:40
  of
• 11:41
  cells?
• 11:41
  Where
• 11:41
  are
• 11:41
  the
• 11:42
  lines
• 11:42
  between
• 11:42
  consc.
• 11:44
  And
• 11:44
  like,
• 11:45
  when
• 11:45
  does
• 11:45
  a
• 11:45
  brain
• 11:46
  become
• 11:46
  an
• 11:46
  entity?
• 11:47
  When
• 11:47
  you're,
• 11:47
  like,
• 11:47
  taking
• 11:48
  parts
• 11:48
  and
• 11:48
  rebuilding
• 11:49
  the
• 11:49
  brain.
• 11:49
  When
• 11:49
  did
• 11:49
  it
• 11:50
  become
• 11:50
  a
• 11:50
  brain?
• 11:50
  Wow.
• 11:51
  Existential
• 11:52
  question.
• 11:52
  It
• 11:53
  really
• 11:53
  is.
• 11:54
  Well,
• 11:54
  speaking
• 11:54
  of
• 11:55
  interfaces,
• 11:56
  today
• 11:56
  we're
• 11:56
  being
• 11:57
  joined
• 11:57
  by
• 11:57
  Louise
• 11:58
  McFaddy,
• 11:58
  and
• 11:58
  she's
• 11:59
  a
• 11:59
  designer,
• 11:59
  a
• 12:00
  creative
• 12:00
  technologist,
• 12:01
  and
• 12:02
  author
• 12:02
  of
• 12:02
  Designing
• 12:03
  AI
• 12:04
  Design
• 12:05
  Principles
• 12:05
  for
• 12:06
  Creative
• 12:07
  and
• 12:07
  Autonomous
• 12:08
  AI.
• 12:11
  We
• 12:11
  have
• 12:11
  a
• 12:11
  question
• 12:11
  that
• 12:12
  we
• 12:12
  ask
• 12:12
  everyone
• 12:13
  when
• 12:13
  they
• 12:13
  are
• 12:14
  a
• 12:14
  guest,
• 12:14
  and
• 12:14
  that
• 12:15
  is
• 12:16
  on
• 12:16
  a
• 12:16
  scale
• 12:17
  from
• 12:17
  1
• 12:18
  to
• 12:18
  10,
• 12:19
  one
• 12:19
  being
• 12:20
  we
• 12:20
  are
• 12:21
  full
• 12:21
  batteries
• 12:21
  for
• 12:22
  the
• 12:22
  machines.
• 12:23
  Civilization
• 12:23
  is
• 12:24
  over.
• 12:24
  This
• 12:25
  is
• 12:25
  the
• 12:25
  end
• 12:25
  to
• 12:26
  10.
• 12:27
  I'm
• 12:27
  so
• 12:27
  much
• 12:27
  more
• 12:27
  productive
• 12:28
  with
• 12:28
  my
• 12:28
  AI
• 12:29
  best
• 12:29
  friend
• 12:29
  that
• 12:29
  actually
• 12:30
  types
• 12:30
  my
• 12:30
  keyboard
• 12:31
  with
• 12:31
  me.
• 12:31
  And
• 12:32
  we're
• 12:32
  besties,
• 12:32
  we
• 12:32
  talk
• 12:33
  on
• 12:33
  the
• 12:33
  weekends.
• 12:33
  And
• 12:34
  also
• 12:34
  my
• 12:34
  therapist.
• 12:36
  How
• 12:36
  are
• 12:36
  you
• 12:36
  feeling
• 12:36
  right
• 12:36
  now?
• 12:37
  I
• 12:37
  think
• 12:37
  that
• 12:38
  on
• 12:38
  a
• 12:38
  personal
• 12:39
  level,
• 12:39
  I
• 12:40
  use
• 12:40
  it
• 12:40
  relatively
• 12:41
  frequently.
• 12:42
  I
• 12:42
  use
• 12:43
  it
• 12:43
  like
• 12:43
  a
• 12:44
  little
• 12:45
  bit
• 12:45
  of
• 12:45
  an
• 12:45
  extra
• 12:45
  muscle
• 12:46
  once
• 12:46
  in
• 12:46
  a
• 12:47
  while.
• 12:47
  It's
• 12:47
  like
• 12:47
  that
• 12:48
  extra
• 12:48
  tentacle
• 12:48
  the
• 12:49
  arm
• 12:49
  I
• 12:49
  wish
• 12:49
  I
• 12:49
  always
• 12:50
  had.
• 12:50
  And
• 12:50
  then
• 12:51
  professionally
• 12:52
  I
• 12:52
  go
• 12:52
  back
• 12:52
  and
• 12:53
  forth.
• 12:53
  I'm
• 12:53
  very
• 12:54
  curious
• 12:55
  about
• 12:55
  what
• 12:56
  will
• 12:56
  stay
• 12:56
  and
• 12:56
  what
• 12:56
  will
• 12:57
  change,
• 12:57
  particularly
• 12:58
  in
• 12:59
  the
• 12:59
  fields
• 13:00
  I
• 13:00
  work
• 13:00
  in,
• 13:00
  which
• 13:00
  is
• 13:01
  much
• 13:01
  larger
• 13:01
  than
• 13:02
  ordinary
• 13:02
  tech
• 13:03
  companies
• 13:03
  that
• 13:04
  are
• 13:04
  used
• 13:04
  to
• 13:05
  and
• 13:05
  are
• 13:05
  very
• 13:05
  effective
• 13:06
  at
• 13:06
  building
• 13:06
  stuff
• 13:07
  with
• 13:07
  loads
• 13:07
  and
• 13:08
  loads
• 13:08
  of
• 13:08
  people
• 13:08
  and
• 13:08
  have
• 13:09
  lots
• 13:09
  of
• 13:09
  processes
• 13:09
  to
• 13:09
  guess
• 13:09
  I've
• 13:10
  done.
• 13:10
  It's
• 13:11
  not
• 13:11
  been
• 13:11
  literally
• 13:12
  that
• 13:12
  useful
• 13:13
  working
• 13:14
  these
• 13:14
  processes.
• 13:15
  It
• 13:15
  doesn't
• 13:15
  show
• 13:15
  up
• 13:15
  that
• 13:15
  much.
• 13:16
  There's
• 13:16
  a
• 13:16
  lot
• 13:16
  of
• 13:17
  excitement
• 13:17
  about
• 13:17
  it.
• 13:18
  But
• 13:19
  I
• 13:19
  would
• 13:19
  say
• 13:19
  I
• 13:20
  interact
• 13:20
  with
• 13:20
  AI
• 13:21
  on
• 13:21
  a
• 13:21
  professional
• 13:21
  level
• 13:21
  the
• 13:22
  most
• 13:22
  when
• 13:22
  I'm
• 13:22
  building
• 13:22
  something.
• 13:23
  Greenfields
• 13:24
  that
• 13:24
  I
• 13:24
  have
• 13:24
  much
• 13:25
  more
• 13:25
  opportunity
• 13:25
  to
• 13:26
  create
• 13:27
  with
• 13:27
  in
• 13:27
  a
• 13:27
  way
• 13:27
  that
• 13:28
  doesn't
• 13:28
  reply
• 13:28
  on
• 13:29
  bureaucracy,
• 13:30
  standards,
• 13:30
  lots
• 13:31
  of
• 13:31
  different
• 13:31
  stuff.
• 13:32
  So
• 13:32
  in
• 13:32
  that
• 13:33
  sense,
• 13:33
  I'm
• 13:33
  like,
• 13:33
  great.
• 13:34
  AI
• 13:34
  is
• 13:34
  changing
• 13:34
  everything.
• 13:35
  When
• 13:35
  you
• 13:35
  are
• 13:36
  starting
• 13:36
  from
• 13:36
  scratch
• 13:37
  and
• 13:37
  it's
• 13:37
  not
• 13:37
  changing
• 13:38
  that
• 13:38
  much
• 13:38
  yet
• 13:39
  when
• 13:39
  you
• 13:39
  have
• 13:39
  like
• 13:40
  a
• 13:40
  ton
• 13:41
  of
• 13:41
  like
• 13:42
  tech
• 13:42
  debt
• 13:42
  and
• 13:42
  history
• 13:42
  and
• 13:43
  everything
• 13:43
  else
• 13:43
  coming
• 13:43
  down,
• 13:44
  those
• 13:44
  things
• 13:44
  are
• 13:44
  still
• 13:45
  moving
• 13:45
  pretty
• 13:45
  slowly.
• 13:45
  I
• 13:46
  would
• 13:46
  put
• 13:46
  me
• 13:46
  at
• 13:46
  a
• 13:46
  solid
• 13:46
  six.
• 13:47
  Okay,
• 13:47
  great.
• 13:48
  Sarah,
• 13:48
  where
• 13:48
  are
• 13:49
  you
• 13:49
  nutting
• 13:49
  out
• 13:49
  today?
• 13:50
  I'm
• 13:51
  like
• 13:51
  a
• 13:51
  seven
• 13:52
  right
• 13:52
  now.
• 13:53
  I
• 13:53
  think
• 13:53
  I've
• 13:54
  processed
• 13:54
  through
• 13:54
  the
• 13:55
  naysaying
• 13:55
  and
• 13:56
  the
• 13:56
  concerns
• 13:57
  and
• 13:57
  I
• 13:57
  think
• 13:57
  I'm
• 13:58
  liking
• 13:58
  where
• 13:58
  things
• 13:58
  are
• 13:58
  going.
• 13:58
  Becca,
• 13:59
  what
• 13:59
  about
• 14:00
  you?
• 14:00
  I'm
• 14:00
  like
• 14:01
  an
• 14:01
  8.5.
• 14:02
  I'm
• 14:02
  riding
• 14:02
  on
• 14:02
  a
• 14:03
  high
• 14:03
  right
• 14:03
  now.
• 14:03
  But
• 14:03
  like,
• 14:03
  with
• 14:04
  caveats
• 14:05
  one,
• 14:05
  I'm
• 14:05
  an
• 14:05
  8.5
• 14:06
  because
• 14:07
  I
• 14:07
  feel
• 14:08
  like
• 14:08
  I've
• 14:08
  had
• 14:09
  this
• 14:09
  explosion
• 14:10
  in
• 14:10
  terms
• 14:10
  of
• 14:11
  creative
• 14:11
  thinking
• 14:12
  over
• 14:12
  the
• 14:13
  last
• 14:13
  few
• 14:13
  weeks
• 14:13
  about
• 14:14
  what
• 14:14
  I
• 14:14
  can
• 14:15
  do
• 14:15
  with
• 14:15
  AI,
• 14:16
  that
• 14:17
  even
• 14:17
  as
• 14:18
  someone
• 14:18
  who
• 14:18
  works
• 14:19
  and
• 14:20
  thinks
• 14:20
  about
• 14:20
  this
• 14:20
  every
• 14:21
  single
• 14:21
  day,
• 14:21
  hadn't
• 14:22
  quite
• 14:23
  had
• 14:23
  yet.
• 14:24
  And
• 14:25
  that's
• 14:25
  extremely
• 14:26
  exciting.
• 14:26
  And
• 14:27
  it's
• 14:27
  given
• 14:27
  me
• 14:28
  this
• 14:28
  surge
• 14:28
  of
• 14:28
  optimism
• 14:29
  about
• 14:30
  how
• 14:30
  AI
• 14:31
  can
• 14:31
  help
• 14:32
  me
• 14:32
  as
• 14:32
  a
• 14:32
  human
• 14:32
  being
• 14:33
  do
• 14:34
  more
• 14:34
  and
• 14:34
  cooler
• 14:35
  things
• 14:35
  than
• 14:35
  I've
• 14:35
  ever
• 14:36
  done
• 14:36
  before.
• 14:37
  The
• 14:37
  only
• 14:37
  reason
• 14:38
  I'm
• 14:38
  not
• 14:38
  a
• 14:38
  10
• 14:39
  is
• 14:39
  because
• 14:40
  I
• 14:41
  think
• 14:41
  that
• 14:42
  there
• 14:42
  are
• 14:42
  still
• 14:43
  vulnerability
• 14:44
  and
• 14:44
  like
• 14:44
  safety
• 14:45
  concerns
• 14:45
  that
• 14:46
  I
• 14:46
  haven't
• 14:46
  even
• 14:46
  with
• 14:46
  all
• 14:47
  that
• 14:47
  quite
• 14:48
  parsed
• 14:48
  through.
• 14:49
  And
• 14:49
  also,
• 14:50
  like,
• 14:50
  there's
• 14:51
  a
• 14:51
  lot
• 14:51
  of
• 14:51
  like
• 14:51
  doom
• 14:52
  sayers
• 14:53
  out
• 14:53
  there
• 14:53
  that
• 14:53
  are
• 14:54
  like
• 14:54
  bumming
• 14:54
  me
• 14:54
  out.
• 14:55
  You
• 14:55
  know,
• 14:55
  a
• 14:55
  lot
• 14:55
  of
• 14:56
  people
• 14:56
  who
• 14:56
  are
• 14:56
  like,
• 14:56
  AI
• 14:57
  is
• 14:57
  just
• 14:57
  going
• 14:57
  to
• 14:57
  like
• 14:57
  automate
• 14:58
  our
• 14:58
  lives.
• 14:58
  Like,
• 14:58
  I
• 14:58
  don't
• 14:58
  have
• 14:58
  to
• 14:58
  think
• 14:59
  anymore.
• 14:59
  I
• 14:59
  just.
• 14:59
  It's
• 15:00
  going
• 15:00
  to
• 15:00
  do
• 15:00
  everything
• 15:00
  for
• 15:00
  me,
• 15:00
  which
• 15:01
  I
• 15:01
  don't
• 15:01
  think
• 15:01
  is
• 15:02
  the
• 15:02
  right
• 15:03
  way
• 15:03
  to
• 15:03
  think
• 15:03
  about
• 15:03
  this
• 15:03
  and
• 15:04
  the
• 15:04
  right
• 15:04
  approach.
• 15:04
  I
• 15:05
  don't
• 15:05
  know.
• 15:05
  Those
• 15:05
  people
• 15:05
  are
• 15:06
  bringing
• 15:06
  me
• 15:06
  down.
• 15:06
  So
• 15:07
  it's
• 15:07
  difficult
• 15:08
  right
• 15:08
  now,
• 15:09
  I
• 15:09
  think,
• 15:09
  to
• 15:09
  form
• 15:09
  a
• 15:09
  rationale
• 15:10
  around
• 15:10
  that
• 15:11
  stuff.
• 15:11
  Like
• 15:11
  as
• 15:12
  you
• 15:12
  were
• 15:12
  saying
• 15:12
  that,
• 15:12
  I
• 15:13
  was
• 15:13
  like,
• 15:13
  oh,
• 15:13
  I
• 15:13
  think
• 15:14
  my
• 15:14
  number
• 15:14
  went
• 15:14
  down
• 15:15
  recently
• 15:15
  because
• 15:15
  of
• 15:16
  the
• 15:16
  Maltbot
• 15:17
  stuff
• 15:18
  that
• 15:18
  was
• 15:18
  going
• 15:18
  around.
• 15:18
  Do
• 15:18
  you
• 15:19
  guys
• 15:19
  hear
• 15:19
  about
• 15:19
  this?
• 15:19
  Yeah,
• 15:20
  yeah,
• 15:20
  there
• 15:21
  was
• 15:22
  one,
• 15:23
  like,
• 15:23
  a
• 15:23
  bunch
• 15:23
  of
• 15:23
  deceptiveness
• 15:24
  at
• 15:24
  play.
• 15:24
  Like,
• 15:25
  a
• 15:25
  lot
• 15:25
  of
• 15:25
  those
• 15:25
  posts
• 15:25
  were
• 15:25
  clearly
• 15:26
  just
• 15:26
  engineered
• 15:26
  by
• 15:26
  people
• 15:27
  or
• 15:27
  people
• 15:27
  using
• 15:28
  a
• 15:28
  prompt
• 15:28
  to
• 15:28
  get
• 15:29
  a
• 15:29
  bot
• 15:29
  to
• 15:29
  go
• 15:29
  and
• 15:29
  write
• 15:30
  the
• 15:30
  type
• 15:30
  of
• 15:30
  posts
• 15:30
  that
• 15:30
  a
• 15:30
  bot
• 15:31
  would
• 15:31
  write
• 15:31
  on
• 15:31
  this
• 15:31
  kind
• 15:31
  of
• 15:31
  platform.
• 15:32
  The
• 15:32
  book
• 15:32
  I
• 15:33
  wrote,
• 15:33
  like,
• 15:33
  was
• 15:34
  essentially
• 15:34
  came
• 15:35
  out
• 15:35
  of
• 15:35
  a
• 15:35
  need
• 15:35
  for
• 15:36
  me.
• 15:36
  Like,
• 15:37
  I
• 15:37
  needed
• 15:37
  to
• 15:37
  read
• 15:37
  that
• 15:38
  book.
• 15:38
  I
• 15:39
  feel
• 15:39
  as
• 15:39
  though
• 15:39
  the
• 15:39
  designers
• 15:40
  around
• 15:40
  me
• 15:41
  are
• 15:41
  a
• 15:41
  little
• 15:41
  bit
• 15:42
  susceptible
• 15:42
  to
• 15:43
  the
• 15:43
  hype
• 15:44
  because
• 15:44
  we
• 15:44
  don't
• 15:44
  work
• 15:45
  as
• 15:45
  close
• 15:46
  to
• 15:46
  the
• 15:46
  technical
• 15:47
  field
• 15:47
  as
• 15:47
  engineers
• 15:48
  do.
• 15:48
  And
• 15:48
  whenever
• 15:48
  something
• 15:49
  like
• 15:49
  this
• 15:49
  hits,
• 15:50
  I
• 15:50
  feel
• 15:50
  so
• 15:51
  frustrated
• 15:51
  because
• 15:51
  it's
• 15:52
  not
• 15:53
  generating
• 15:53
  any
• 15:53
  better
• 15:54
  understanding
• 15:54
  and
• 15:55
  it
• 15:55
  just
• 15:55
  contributes
• 15:56
  to
• 15:56
  hype,
• 15:56
  which
• 15:57
  erodes
• 15:58
  the
• 15:58
  position
• 15:58
  of
• 15:58
  our
• 15:59
  industry
• 15:59
  evermore.
• 16:00
  So,
• 16:00
  as
• 16:01
  part
• 16:01
  of
• 16:01
  writing
• 16:01
  your
• 16:01
  book,
• 16:01
  I'm
• 16:02
  sure
• 16:02
  you
• 16:02
  did
• 16:02
  a
• 16:02
  lot
• 16:02
  of
• 16:02
  research
• 16:03
  about
• 16:04
  how
• 16:04
  different
• 16:04
  people
• 16:05
  are
• 16:05
  approaching
• 16:06
  designing
• 16:06
  for
• 16:06
  AI.
• 16:07
  What
• 16:07
  did
• 16:07
  you
• 16:08
  see?
• 16:08
  What
• 16:08
  are
• 16:08
  people
• 16:09
  doing
• 16:09
  right?
• 16:09
  What
• 16:09
  are
• 16:09
  people
• 16:10
  doing
• 16:10
  wrong?
• 16:10
  What
• 16:10
  does
• 16:11
  the
• 16:11
  industry
• 16:11
  look
• 16:11
  like?
• 16:12
  At
• 16:12
  the
• 16:12
  heart
• 16:12
  of
• 16:13
  it,
• 16:13
  AI
• 16:13
  as
• 16:14
  a
• 16:14
  technology
• 16:14
  was
• 16:14
  enabled
• 16:15
  by
• 16:15
  big
• 16:15
  data
• 16:16
  and
• 16:16
  the
• 16:16
  existence
• 16:16
  of
• 16:17
  big
• 16:17
  data.
• 16:18
  And
• 16:19
  when
• 16:19
  reflecting
• 16:20
  back
• 16:20
  on
• 16:21
  how
• 16:22
  we
• 16:22
  used
• 16:22
  to
• 16:22
  design
• 16:23
  for
• 16:23
  that
• 16:23
  stuff,
• 16:23
  it
• 16:24
  was
• 16:24
  generally
• 16:24
  around
• 16:26
  anticipating
• 16:26
  an
• 16:26
  algorithmic
• 16:27
  output,
• 16:27
  but
• 16:27
  it
• 16:28
  was
• 16:28
  quite
• 16:28
  limited.
• 16:28
  I
• 16:29
  had
• 16:29
  to
• 16:29
  do
• 16:29
  a
• 16:29
  project
• 16:29
  years
• 16:30
  ago
• 16:30
  now
• 16:30
  for
• 16:31
  Google
• 16:31
  Search
• 16:32
  and
• 16:32
  trying
• 16:32
  to
• 16:33
  work
• 16:33
  out
• 16:34
  how
• 16:34
  you
• 16:34
  can
• 16:35
  create
• 16:35
  literally
• 16:36
  even
• 16:36
  a
• 16:36
  mock
• 16:37
  for
• 16:38
  a
• 16:38
  series
• 16:39
  of
• 16:39
  apps
• 16:39
  that
• 16:39
  you
• 16:39
  don't
• 16:40
  know
• 16:40
  you
• 16:40
  have
• 16:40
  no
• 16:41
  idea
• 16:41
  what
• 16:41
  you're
• 16:41
  going
• 16:41
  to
• 16:41
  get
• 16:42
  was
• 16:42
  really
• 16:42
  challenging.
• 16:43
  And
• 16:43
  yet,
• 16:44
  because
• 16:44
  I'd
• 16:44
  had
• 16:44
  to
• 16:44
  deal
• 16:45
  with
• 16:45
  that
• 16:45
  problem
• 16:45
  before,
• 16:46
  when
• 16:46
  I
• 16:46
  came
• 16:46
  to
• 16:47
  work
• 16:47
  in
• 16:47
  AI
• 16:48
  now
• 16:48
  I
• 16:48
  was
• 16:49
  like,
• 16:49
  cool,
• 16:49
  we
• 16:49
  need
• 16:50
  to
• 16:50
  make
• 16:50
  sure
• 16:50
  that
• 16:50
  we
• 16:50
  are
• 16:50
  crystal
• 16:51
  clear
• 16:52
  about
• 16:52
  what
• 16:52
  we're
• 16:52
  anticipating,
• 16:53
  the
• 16:53
  Personas
• 16:53
  that
• 16:53
  we're
• 16:54
  going
• 16:54
  to
• 16:54
  need.
• 16:54
  I
• 16:55
  think
• 16:55
  that
• 16:55
  the
• 16:56
  primary
• 16:57
  complaint
• 16:57
  of
• 16:57
  a
• 16:58
  lot
• 16:58
  of
• 16:58
  designers
• 16:58
  and
• 16:58
  the
• 16:58
  challenge
• 16:59
  they're
• 16:59
  hitting
• 16:59
  is
• 17:00
  that
• 17:00
  exact
• 17:00
  problem
• 17:01
  still
• 17:01
  of
• 17:02
  what
• 17:02
  on
• 17:03
  earth
• 17:03
  is
• 17:03
  getting
• 17:04
  created
• 17:04
  by
• 17:04
  the
• 17:05
  model
• 17:05
  and
• 17:05
  how
• 17:05
  do
• 17:06
  we
• 17:06
  anticipate
• 17:07
  and
• 17:07
  visualize
• 17:08
  for
• 17:08
  that.
• 17:08
  There's
• 17:09
  a
• 17:09
  lot
• 17:09
  of
• 17:09
  hype
• 17:09
  and
• 17:10
  curiosity
• 17:10
  around
• 17:11
  AI
• 17:12
  generated
• 17:12
  design
• 17:13
  systems,
• 17:13
  AI
• 17:13
  generated
• 17:14
  ui,
• 17:14
  which
• 17:15
  goes
• 17:15
  a
• 17:15
  bit
• 17:15
  beyond
• 17:16
  what
• 17:16
  Claude
• 17:17
  code
• 17:17
  might
• 17:17
  create
• 17:17
  for
• 17:17
  you.
• 17:18
  I
• 17:18
  feel
• 17:18
  as
• 17:18
  though
• 17:19
  designers
• 17:19
  get
• 17:19
  lost
• 17:20
  along
• 17:20
  the
• 17:20
  way
• 17:20
  there
• 17:21
  because
• 17:21
  they
• 17:22
  have
• 17:22
  not
• 17:22
  had
• 17:22
  to
• 17:22
  work
• 17:23
  in
• 17:23
  this
• 17:23
  really
• 17:24
  ambiguous
• 17:24
  way
• 17:24
  before,
• 17:25
  but
• 17:25
  then
• 17:25
  also
• 17:26
  start
• 17:27
  to
• 17:27
  really
• 17:27
  forget
• 17:28
  the
• 17:28
  core
• 17:28
  tenets
• 17:28
  of
• 17:29
  product
• 17:29
  design,
• 17:29
  which
• 17:30
  are
• 17:30
  actually
• 17:30
  even
• 17:30
  more
• 17:30
  applicable
• 17:31
  now
• 17:31
  than
• 17:31
  they
• 17:32
  were
• 17:32
  before.
• 17:32
  You
• 17:32
  draw
• 17:33
  this
• 17:33
  interesting
• 17:33
  line
• 17:34
  from
• 17:35
  something
• 17:36
  as
• 17:36
  early
• 17:36
  as,
• 17:37
  I
• 17:37
  think
• 17:37
  it's
• 17:37
  called
• 17:37
  multics
• 17:38
  at
• 17:38
  MIT
• 17:39
  and
• 17:39
  the
• 17:39
  command
• 17:40
  shell
• 17:40
  to
• 17:40
  Google
• 17:41
  Search
• 17:41
  to
• 17:41
  LLMs.
• 17:42
  Could
• 17:42
  you
• 17:42
  just
• 17:43
  recap
• 17:44
  that
• 17:44
  interesting
• 17:45
  through
• 17:45
  line
• 17:46
  that
• 17:46
  you
• 17:47
  really
• 17:48
  expertly
• 17:49
  identified
• 17:49
  from
• 17:50
  a
• 17:50
  design
• 17:50
  perspective?
• 17:51
  Yeah,
• 17:51
  Multics
• 17:51
  was
• 17:52
  basically
• 17:52
  one
• 17:52
  of
• 17:52
  the
• 17:53
  first
• 17:53
  big
• 17:54
  available
• 17:54
  computers
• 17:55
  built
• 17:56
  in
• 17:56
  the
• 17:56
  1970s
• 17:57
  by
• 17:57
  MIT.
• 17:58
  It's
• 17:58
  the
• 17:58
  first
• 17:58
  place
• 17:58
  we
• 17:59
  really
• 17:59
  see
• 17:59
  the
• 17:59
  command
• 17:59
  line
• 18:00
  and
• 18:00
  the
• 18:00
  shell
• 18:00
  be
• 18:00
  developed.
• 18:01
  The
• 18:01
  shell
• 18:01
  being
• 18:01
  where
• 18:02
  we
• 18:02
  can
• 18:02
  input
• 18:02
  a
• 18:03
  command
• 18:03
  and
• 18:03
  the
• 18:03
  computer
• 18:04
  begins
• 18:04
  to
• 18:04
  obey
• 18:04
  us.
• 18:05
  And
• 18:05
  it's
• 18:05
  so
• 18:05
  recognizable
• 18:06
  for
• 18:06
  us.
• 18:06
  I
• 18:07
  think
• 18:07
  when
• 18:07
  OpenAI
• 18:08
  first
• 18:08
  launched
• 18:09
  ChatGPT,
• 18:09
  there
• 18:10
  was
• 18:10
  a
• 18:10
  lot
• 18:10
  of
• 18:10
  questions
• 18:11
  around
• 18:11
  interface
• 18:12
  designers
• 18:12
  about,
• 18:13
  is
• 18:13
  text
• 18:13
  input
• 18:14
  really
• 18:14
  the
• 18:14
  way
• 18:14
  that
• 18:14
  we
• 18:15
  actually
• 18:15
  think
• 18:15
  people
• 18:15
  want
• 18:15
  to
• 18:16
  use
• 18:16
  this?
• 18:16
  And
• 18:17
  I
• 18:17
  was
• 18:17
  always
• 18:17
  like,
• 18:17
  well,
• 18:17
  people
• 18:18
  really,
• 18:19
  you
• 18:19
  know,
• 18:19
  cottoned
• 18:19
  on
• 18:20
  to
• 18:20
  the
• 18:20
  command
• 18:20
  line
• 18:21
  very,
• 18:21
  very
• 18:21
  quickly.
• 18:22
  But
• 18:22
  it
• 18:22
  was
• 18:22
  a
• 18:22
  limitation
• 18:23
  as
• 18:23
  well.
• 18:23
  And
• 18:24
  so
• 18:24
  I
• 18:24
  think
• 18:24
  that's
• 18:24
  why
• 18:24
  that
• 18:24
  question
• 18:25
  comes
• 18:25
  forward.
• 18:26
  So
• 18:26
  that
• 18:26
  history
• 18:27
  is
• 18:27
  so
• 18:27
  relevant
• 18:27
  to
• 18:27
  me
• 18:28
  that
• 18:28
  we
• 18:28
  went
• 18:29
  from
• 18:29
  multics
• 18:30
  and
• 18:30
  just
• 18:30
  putting
• 18:31
  in
• 18:31
  extremely
• 18:32
  difficult
• 18:33
  and
• 18:33
  computationally
• 18:34
  hard
• 18:35
  to
• 18:35
  remember
• 18:36
  structured
• 18:37
  queries.
• 18:38
  And
• 18:38
  then
• 18:38
  we
• 18:38
  went
• 18:39
  a
• 18:39
  little
• 18:39
  bit
• 18:39
  further
• 18:40
  and
• 18:40
  we
• 18:40
  got
• 18:40
  Google
• 18:41
  Search,
• 18:41
  where
• 18:41
  we
• 18:41
  could
• 18:42
  get
• 18:42
  an
• 18:42
  autocorrect,
• 18:43
  we
• 18:43
  could
• 18:43
  update
• 18:43
  what
• 18:43
  we
• 18:43
  were
• 18:44
  writing,
• 18:44
  we
• 18:45
  would
• 18:45
  anticipate
• 18:46
  the
• 18:46
  output
• 18:46
  a
• 18:46
  little
• 18:46
  bit
• 18:46
  more
• 18:47
  and.
• 18:47
  And
• 18:47
  then
• 18:47
  you
• 18:47
  go
• 18:48
  all
• 18:48
  the
• 18:48
  way
• 18:48
  to
• 18:49
  LLM
• 18:49
  input
• 18:50
  and
• 18:50
  output,
• 18:51
  and
• 18:51
  you
• 18:51
  have
• 18:51
  total
• 18:52
  flexibility.
• 18:53
  And
• 18:53
  so
• 18:53
  to
• 18:54
  me,
• 18:54
  at
• 18:55
  the
• 18:55
  heart
• 18:55
  of
• 18:55
  multics
• 18:56
  is
• 18:56
  the
• 18:57
  desire
• 18:57
  to
• 18:57
  be
• 18:58
  able
• 18:58
  to
• 18:58
  speak
• 18:58
  realistically
• 18:59
  to
• 18:59
  a
• 19:00
  computer.
• 19:00
  Like
• 19:00
  we
• 19:00
  can
• 19:01
  now,
• 19:01
  but
• 19:02
  it's
• 19:02
  like
• 19:02
  the
• 19:02
  oldest
• 19:02
  possible
• 19:03
  version
• 19:03
  of
• 19:03
  it.
• 19:04
  It's
• 19:04
  the
• 19:04
  most
• 19:04
  brittle,
• 19:05
  rigid
• 19:06
  version.
• 19:06
  And
• 19:06
  now
• 19:06
  what
• 19:07
  we
• 19:07
  have
• 19:07
  is
• 19:07
  this
• 19:07
  much
• 19:07
  squishier
• 19:08
  version,
• 19:08
  but
• 19:08
  the
• 19:08
  text
• 19:09
  box
• 19:09
  still
• 19:09
  remains.
• 19:10
  If
• 19:11
  I
• 19:11
  had
• 19:11
  to
• 19:11
  sum
• 19:11
  up
• 19:12
  my
• 19:12
  one
• 19:12
  burning
• 19:13
  question
• 19:13
  for
• 19:14
  you
• 19:14
  today,
• 19:15
  it's
• 19:16
  like
• 19:16
  the
• 19:17
  text
• 19:17
  box
• 19:18
  question
• 19:18
  mark.
• 19:19
  I
• 19:19
  think
• 19:19
  largely
• 19:20
  because
• 19:20
  of
• 19:20
  ChatGPT,
• 19:22
  most
• 19:22
  of
• 19:23
  how
• 19:24
  the
• 19:24
  world
• 19:25
  thinks
• 19:25
  about
• 19:27
  UI
• 19:28
  when
• 19:28
  it
• 19:28
  comes
• 19:29
  to
• 19:29
  interacting
• 19:30
  with
• 19:30
  AI
• 19:31
  is
• 19:32
  like
• 19:32
  some
• 19:33
  version
• 19:33
  of
• 19:33
  a
• 19:33
  text
• 19:33
  box,
• 19:34
  which
• 19:34
  makes
• 19:35
  total
• 19:35
  natural
• 19:36
  sense
• 19:36
  based
• 19:36
  on,
• 19:36
  like,
• 19:37
  our
• 19:38
  shared
• 19:38
  history
• 19:39
  of
• 19:39
  how
• 19:39
  we
• 19:39
  have
• 19:40
  gone
• 19:40
  about
• 19:40
  communicating
• 19:41
  with
• 19:41
  computers.
• 19:42
  But
• 19:42
  now
• 19:43
  you
• 19:43
  don't
• 19:43
  necessarily
• 19:44
  need
• 19:44
  to
• 19:44
  have
• 19:45
  this
• 19:46
  extremely
• 19:47
  specific
• 19:48
  structured
• 19:49
  input
• 19:50
  where,
• 19:50
  like,
• 19:50
  if
• 19:50
  you
• 19:51
  don't
• 19:51
  put
• 19:51
  a
• 19:52
  comma
• 19:52
  in
• 19:52
  the
• 19:52
  right
• 19:52
  place,
• 19:53
  you're
• 19:53
  not
• 19:53
  going
• 19:53
  to
• 19:53
  get
• 19:53
  the
• 19:53
  right
• 19:54
  response.
• 19:54
  You
• 19:54
  can
• 19:54
  kind
• 19:55
  of
• 19:55
  say
• 19:55
  whatever
• 19:55
  you
• 19:55
  want.
• 19:56
  So
• 19:56
  how
• 19:57
  do
• 19:57
  you
• 19:57
  think
• 19:58
  about
• 19:59
  the
• 19:59
  text
• 19:59
  box
• 20:00
  and
• 20:01
  what
• 20:01
  the
• 20:01
  future
• 20:02
  of
• 20:02
  AI
• 20:03
  interface
• 20:03
  design
• 20:04
  looks
• 20:04
  like
• 20:05
  beyond
• 20:05
  the
• 20:06
  text
• 20:06
  box?
• 20:07
  How
• 20:07
  we
• 20:08
  all
• 20:08
  communicated
• 20:08
  prior
• 20:09
  to
• 20:09
  this
• 20:09
  was
• 20:09
  via
• 20:10
  text
• 20:10
  boxes,
• 20:10
  and
• 20:11
  how
• 20:11
  I
• 20:11
  speak
• 20:11
  to
• 20:11
  my
• 20:11
  mom
• 20:12
  is
• 20:12
  via
• 20:12
  a
• 20:12
  text
• 20:12
  box.
• 20:13
  And
• 20:13
  the
• 20:13
  keyboard
• 20:14
  actually
• 20:14
  as
• 20:14
  an
• 20:15
  extension,
• 20:16
  is
• 20:16
  the
• 20:16
  method
• 20:16
  by
• 20:17
  which
• 20:17
  we
• 20:18
  have
• 20:18
  all
• 20:18
  decided
• 20:19
  we
• 20:19
  would
• 20:20
  like
• 20:20
  to
• 20:21
  send
• 20:21
  communication.
• 20:22
  When
• 20:23
  you're
• 20:23
  communicating
• 20:23
  with
• 20:23
  a
• 20:23
  model,
• 20:24
  you
• 20:24
  still
• 20:24
  are
• 20:24
  kind
• 20:25
  of
• 20:25
  doing
• 20:25
  communication
• 20:26
  a
• 20:26
  lot,
• 20:26
  because
• 20:27
  a
• 20:27
  lot
• 20:27
  of
• 20:27
  the
• 20:27
  times
• 20:28
  the
• 20:28
  output
• 20:28
  is
• 20:28
  for
• 20:28
  a
• 20:29
  human
• 20:29
  to
• 20:29
  see.
• 20:29
  Like,
• 20:30
  you're
• 20:30
  going
• 20:30
  in
• 20:30
  and
• 20:30
  you're
• 20:31
  maybe
• 20:31
  using
• 20:31
  it
• 20:31
  to
• 20:31
  write
• 20:31
  an
• 20:31
  email
• 20:32
  or
• 20:32
  you're
• 20:32
  using
• 20:32
  it
• 20:32
  to
• 20:33
  create
• 20:33
  a
• 20:33
  document,
• 20:33
  but
• 20:34
  usually
• 20:34
  someone
• 20:34
  else
• 20:34
  is
• 20:35
  seeing
• 20:35
  that.
• 20:35
  And
• 20:35
  so
• 20:36
  I
• 20:36
  actually
• 20:36
  often
• 20:37
  think
• 20:37
  about
• 20:38
  what
• 20:38
  we're
• 20:38
  doing
• 20:39
  as
• 20:40
  just
• 20:40
  another
• 20:40
  way
• 20:40
  of
• 20:40
  speaking
• 20:41
  to
• 20:41
  someone.
• 20:41
  It's
• 20:41
  just
• 20:42
  like
• 20:42
  through
• 20:42
  a
• 20:42
  different
• 20:43
  medium.
• 20:43
  It's
• 20:44
  like
• 20:44
  autocorrect
• 20:45
  multiplied.
• 20:45
  I
• 20:46
  was
• 20:46
  trying
• 20:46
  to
• 20:46
  find
• 20:47
  the
• 20:47
  best
• 20:47
  analogy
• 20:47
  I
• 20:48
  could
• 20:48
  for
• 20:49
  LLMs
• 20:50
  as
• 20:50
  a
• 20:50
  historical
• 20:51
  element
• 20:52
  when
• 20:52
  we
• 20:52
  really
• 20:52
  started
• 20:53
  to
• 20:53
  feel
• 20:53
  them.
• 20:53
  And
• 20:54
  I'm
• 20:54
  going
• 20:54
  to
• 20:54
  date
• 20:54
  myself
• 20:55
  here,
• 20:55
  but
• 20:56
  I
• 20:56
  used
• 20:57
  the
• 20:57
  analogy
• 20:57
  of
• 20:57
  T9,
• 20:58
  which
• 20:58
  was
• 20:59
  a
• 20:59
  predictive
• 20:59
  texting
• 21:00
  service
• 21:00
  on
• 21:01
  mobile
• 21:01
  phones.
• 21:02
  And
• 21:02
  the
• 21:02
  thing
• 21:03
  is,
• 21:03
  T9
• 21:04
  was
• 21:04
  developed
• 21:05
  not
• 21:05
  because
• 21:05
  they
• 21:05
  actually
• 21:06
  wanted
• 21:06
  to
• 21:07
  create
• 21:07
  a
• 21:08
  quicker
• 21:08
  way
• 21:08
  to
• 21:08
  text.
• 21:09
  It
• 21:09
  was
• 21:09
  a
• 21:10
  group
• 21:10
  trying
• 21:10
  to
• 21:11
  solve
• 21:11
  for
• 21:11
  accessibility
• 21:12
  issues
• 21:12
  for
• 21:12
  people
• 21:13
  who
• 21:13
  literally
• 21:13
  couldn't
• 21:13
  type.
• 21:14
  And
• 21:14
  so
• 21:14
  they
• 21:15
  wanted
• 21:15
  to
• 21:15
  allow
• 21:16
  people
• 21:16
  to
• 21:16
  use
• 21:16
  their
• 21:17
  eyes
• 21:17
  and
• 21:18
  use
• 21:18
  tracking
• 21:19
  to
• 21:19
  be
• 21:20
  able
• 21:20
  to
• 21:20
  communicate
• 21:20
  quickly.
• 21:21
  And
• 21:21
  so
• 21:22
  it's
• 21:22
  landmark
• 21:23
  detection
• 21:23
  because
• 21:24
  there
• 21:24
  were
• 21:24
  like
• 21:24
  nine
• 21:24
  buttons
• 21:25
  works,
• 21:26
  and
• 21:26
  we
• 21:26
  just
• 21:27
  sit
• 21:27
  and
• 21:27
  predict.
• 21:27
  And
• 21:28
  it's
• 21:28
  not
• 21:28
  sophisticated
• 21:29
  as
• 21:29
  an
• 21:29
  LLM,
• 21:29
  right?
• 21:30
  It's
• 21:30
  got
• 21:30
  a
• 21:30
  dictionary
• 21:30
  that
• 21:31
  it's
• 21:31
  using
• 21:31
  a
• 21:31
  little
• 21:31
  algorithm
• 21:32
  to
• 21:32
  make
• 21:32
  work.
• 21:33
  But
• 21:33
  to
• 21:33
  me,
• 21:33
  I'm
• 21:34
  like,
• 21:34
  wow,
• 21:34
  that
• 21:34
  is
• 21:34
  really
• 21:35
  at
• 21:35
  the
• 21:35
  heart
• 21:35
  of
• 21:35
  it.
• 21:36
  We
• 21:36
  created
• 21:37
  what
• 21:37
  ended
• 21:38
  up
• 21:38
  becoming
• 21:38
  the
• 21:38
  first
• 21:38
  predictive
• 21:39
  models
• 21:39
  in
• 21:40
  a
• 21:40
  very
• 21:40
  similar
• 21:40
  pattern
• 21:41
  to
• 21:41
  the
• 21:41
  way
• 21:42
  that
• 21:42
  T9
• 21:42
  was
• 21:42
  predicted.
• 21:43
  And
• 21:43
  it
• 21:43
  is
• 21:43
  just
• 21:44
  thoroughly
• 21:44
  about
• 21:45
  condensing
• 21:46
  down
• 21:46
  the
• 21:46
  method
• 21:47
  by
• 21:47
  which
• 21:47
  we
• 21:47
  communicate.
• 21:48
  So
• 21:49
  to
• 21:49
  me,
• 21:49
  the
• 21:49
  text
• 21:50
  box
• 21:50
  and
• 21:50
  all
• 21:50
  of
• 21:51
  that
• 21:51
  utility
• 21:51
  is
• 21:52
  much
• 21:52
  more
• 21:52
  about
• 21:52
  human
• 21:53
  practices
• 21:53
  of
• 21:53
  communication
• 21:54
  and
• 21:55
  different
• 21:55
  from
• 21:55
  how
• 21:55
  models
• 21:56
  themselves
• 21:56
  work.
• 21:57
  There's
• 21:57
  great
• 21:57
  design
• 21:58
  resources
• 21:58
  out
• 21:58
  there
• 21:59
  about
• 22:00
  the
• 22:00
  chat
• 22:00
  process
• 22:01
  and
• 22:01
  the
• 22:01
  chat
• 22:01
  interface,
• 22:02
  because
• 22:02
  we've
• 22:02
  had
• 22:02
  those
• 22:02
  for
• 22:03
  a
• 22:03
  while.
• 22:03
  It's
• 22:04
  clear
• 22:04
  that,
• 22:04
  at
• 22:05
  least
• 22:05
  to
• 22:05
  me,
• 22:05
  that
• 22:05
  we
• 22:06
  know
• 22:06
  that
• 22:07
  models
• 22:07
  and
• 22:07
  their
• 22:07
  outputs
• 22:08
  and
• 22:08
  the
• 22:08
  computation
• 22:08
  is
• 22:09
  very
• 22:09
  different
• 22:09
  to
• 22:10
  visualize
• 22:10
  than
• 22:10
  other
• 22:10
  processes
• 22:11
  you've
• 22:11
  had
• 22:11
  the
• 22:12
  input.
• 22:12
  To
• 22:12
  me,
• 22:12
  I'm
• 22:13
  like,
• 22:13
  unless
• 22:13
  we
• 22:14
  fundamentally
• 22:14
  change
• 22:14
  how
• 22:15
  we
• 22:15
  want
• 22:15
  to
• 22:15
  communicate,
• 22:16
  it's
• 22:16
  probably
• 22:17
  going
• 22:17
  to
• 22:17
  be
• 22:17
  its
• 22:17
  own
• 22:18
  area.
• 22:19
  Not
• 22:19
  to
• 22:19
  throw
• 22:20
  a
• 22:20
  curveball
• 22:21
  your
• 22:21
  way,
• 22:21
  but
• 22:22
  there
• 22:22
  is
• 22:23
  something
• 22:23
  that
• 22:23
  I
• 22:24
  think
• 22:24
  about
• 22:25
  outside
• 22:26
  of
• 22:26
  the
• 22:26
  text
• 22:27
  box,
• 22:27
  but,
• 22:27
  like
• 22:27
  the
• 22:28
  mouse.
• 22:28
  Do
• 22:29
  you
• 22:29
  foresee
• 22:30
  there
• 22:30
  being
• 22:31
  the
• 22:31
  mouse
• 22:31
  equivalent
• 22:32
  of
• 22:33
  a
• 22:33
  new
• 22:34
  way
• 22:34
  of
• 22:34
  interacting
• 22:35
  with
• 22:35
  AI?
• 22:36
  I
• 22:36
  hope
• 22:37
  so.
• 22:37
  Like,
• 22:37
  I
• 22:38
  really
• 22:38
  do
• 22:38
  hope
• 22:38
  that
• 22:39
  we
• 22:39
  get
• 22:39
  some
• 22:40
  Douglas
• 22:40
  Engelbart
• 22:41
  genius
• 22:41
  who
• 22:42
  comes
• 22:42
  in
• 22:42
  and
• 22:42
  thinks
• 22:42
  differently
• 22:43
  about
• 22:43
  that.
• 22:44
  To
• 22:44
  me,
• 22:44
  the
• 22:45
  mouse
• 22:45
  is
• 22:46
  so
• 22:46
  similar
• 22:46
  to
• 22:46
  the
• 22:47
  stylus,
• 22:48
  and
• 22:48
  that
• 22:48
  is,
• 22:49
  again,
• 22:49
  one
• 22:49
  of
• 22:49
  the
• 22:49
  oldest
• 22:50
  possible
• 22:50
  technologies.
• 22:51
  And
• 22:52
  like,
• 22:52
  going
• 22:52
  and
• 22:52
  clicking
• 22:53
  and
• 22:53
  opening
• 22:53
  on
• 22:53
  something
• 22:53
  is
• 22:54
  so
• 22:54
  similar
• 22:54
  to
• 22:55
  this.
• 22:55
  But
• 22:55
  it's
• 22:55
  also
• 22:56
  notation,
• 22:57
  as
• 22:57
  we
• 22:57
  may
• 22:57
  think,
• 22:57
  by
• 22:57
  Vannevar
• 22:58
  Bush,
• 22:58
  the
• 22:58
  1945
• 22:59
  essay
• 23:00
  where
• 23:00
  he
• 23:00
  kind
• 23:01
  of
• 23:01
  considered
• 23:02
  the
• 23:02
  future
• 23:03
  of
• 23:03
  information
• 23:04
  sharing.
• 23:04
  And
• 23:05
  he
• 23:05
  talks
• 23:05
  about
• 23:05
  extending
• 23:06
  all
• 23:06
  the
• 23:06
  different
• 23:07
  capabilities
• 23:07
  of
• 23:07
  the
• 23:07
  human
• 23:08
  body.
• 23:09
  And
• 23:09
  then
• 23:09
  he's
• 23:10
  like,
• 23:10
  we've
• 23:10
  done
• 23:10
  that.
• 23:11
  We've
• 23:11
  made
• 23:11
  ourselves
• 23:11
  go
• 23:12
  faster.
• 23:12
  We've
• 23:12
  created
• 23:13
  the
• 23:13
  telescope
• 23:13
  so
• 23:13
  we
• 23:13
  can
• 23:13
  see
• 23:13
  further,
• 23:14
  the
• 23:14
  microscope
• 23:14
  so
• 23:14
  we
• 23:15
  can
• 23:15
  see
• 23:15
  smaller,
• 23:16
  but
• 23:16
  how
• 23:16
  are
• 23:16
  we
• 23:16
  extending
• 23:17
  the
• 23:17
  human
• 23:17
  mind?
• 23:18
  And
• 23:18
  so
• 23:19
  to
• 23:19
  me,
• 23:20
  the
• 23:20
  genius
• 23:20
  of
• 23:20
  Douglas
• 23:21
  Engelbart's
• 23:21
  invention
• 23:22
  with
• 23:22
  the
• 23:22
  cursor
• 23:24
  is,
• 23:24
  to
• 23:24
  me,
• 23:25
  collapsing
• 23:25
  in
• 23:25
  the
• 23:26
  human
• 23:26
  sketching
• 23:27
  capability
• 23:27
  into
• 23:28
  information
• 23:28
  structured
• 23:29
  underneath.
• 23:30
  So,
• 23:30
  yes,
• 23:30
  some
• 23:31
  genius.
• 23:31
  I
• 23:31
  would
• 23:31
  love,
• 23:32
  love
• 23:32
  it
• 23:32
  if
• 23:32
  they
• 23:32
  would
• 23:32
  take
• 23:33
  that
• 23:33
  and
• 23:33
  open
• 23:33
  it
• 23:33
  up.
• 23:34
  I
• 23:34
  am
• 23:35
  dubious
• 23:35
  about
• 23:36
  that
• 23:36
  being
• 23:37
  gesture
• 23:38
  oriented
• 23:38
  because
• 23:39
  we've
• 23:39
  seen
• 23:40
  point
• 23:40
  clouds
• 23:41
  and
• 23:41
  art
• 23:41
  like
• 23:42
  that.
• 23:42
  That's
• 23:42
  not
• 23:43
  really
• 23:43
  captured
• 23:43
  as
• 23:44
  much
• 23:44
  significance
• 23:44
  in
• 23:45
  meaning.
• 23:45
  But
• 23:46
  there
• 23:46
  are
• 23:46
  so
• 23:46
  many
• 23:46
  opportunities.
• 23:47
  There's
• 23:47
  sign
• 23:48
  language,
• 23:48
  there's
• 23:48
  different
• 23:49
  ways
• 23:49
  that
• 23:49
  we
• 23:49
  generate
• 23:50
  meaning
• 23:50
  with
• 23:50
  our
• 23:50
  bodies
• 23:51
  that
• 23:51
  could
• 23:51
  certainly
• 23:51
  be
• 23:52
  leveraged.
• 23:52
  Voice
• 23:53
  is
• 23:53
  very
• 23:53
  interesting
• 23:53
  to
• 23:54
  me,
• 23:54
  but
• 23:54
  is
• 23:55
  also
• 23:55
  a
• 23:55
  very
• 23:55
  difficult
• 23:56
  one
• 23:56
  to
• 23:57
  consider
• 23:57
  because
• 23:58
  you
• 23:58
  and
• 23:58
  I
• 23:59
  are
• 24:00
  expressing
• 24:00
  an
• 24:00
  enormous
• 24:01
  amount
• 24:01
  of
• 24:01
  information
• 24:01
  by
• 24:02
  the
• 24:02
  way
• 24:02
  we're
• 24:02
  structuring
• 24:03
  our
• 24:03
  sentences
• 24:03
  and
• 24:04
  the
• 24:04
  tonality
• 24:04
  we're
• 24:05
  bringing
• 24:05
  into
• 24:05
  our
• 24:05
  voice
• 24:05
  and
• 24:06
  everything
• 24:06
  else.
• 24:06
  Body
• 24:07
  language,
• 24:07
  there's
• 24:08
  not
• 24:08
  great
• 24:08
  corpuses
• 24:09
  of
• 24:09
  data
• 24:10
  that
• 24:10
  allow
• 24:10
  a
• 24:10
  model
• 24:11
  to
• 24:11
  be
• 24:11
  able
• 24:11
  to
• 24:12
  extract
• 24:13
  that
• 24:13
  meaning.
• 24:14
  And
• 24:15
  so
• 24:15
  we
• 24:16
  may
• 24:16
  well
• 24:16
  see
• 24:16
  that
• 24:17
  become
• 24:17
  a
• 24:17
  useful
• 24:18
  method
• 24:18
  of
• 24:18
  communication
• 24:19
  in
• 24:19
  the
• 24:19
  future,
• 24:20
  but
• 24:20
  it's
• 24:21
  also
• 24:21
  not
• 24:21
  the
• 24:21
  one
• 24:22
  that
• 24:22
  we
• 24:22
  are
• 24:23
  the
• 24:23
  most
• 24:23
  capable
• 24:24
  at
• 24:24
  codifying
• 24:25
  right
• 24:25
  now.
• 24:25
  We
• 24:26
  also
• 24:27
  love
• 24:27
  to
• 24:27
  separate
• 24:28
  our
• 24:28
  worlds
• 24:28
  out
• 24:28
  a
• 24:29
  little
• 24:29
  bit.
• 24:29
  And
• 24:30
  we've
• 24:31
  seen
• 24:31
  what
• 24:31
  happened
• 24:31
  with
• 24:31
  Meta
• 24:32
  and
• 24:32
  VR
• 24:33
  and
• 24:33
  how
• 24:33
  much
• 24:34
  personal
• 24:34
  skepticism,
• 24:35
  I
• 24:35
  think,
• 24:36
  came
• 24:36
  into
• 24:36
  those
• 24:36
  products.
• 24:38
  And
• 24:38
  I
• 24:38
  say
• 24:38
  this
• 24:39
  all
• 24:39
  the
• 24:39
  time
• 24:39
  about
• 24:39
  AI.
• 24:39
  It
• 24:40
  is
• 24:40
  not
• 24:40
  a
• 24:40
  neutral
• 24:40
  space
• 24:41
  at
• 24:41
  all.
• 24:41
  Whenever
• 24:41
  you
• 24:42
  build
• 24:42
  an
• 24:42
  AI
• 24:42
  product,
• 24:42
  you're
• 24:43
  not
• 24:43
  launching
• 24:43
  it
• 24:43
  into
• 24:44
  audience
• 24:45
  that
• 24:45
  are
• 24:45
  like,
• 24:45
  oh,
• 24:45
  interesting,
• 24:46
  this
• 24:46
  is
• 24:46
  new.
• 24:46
  We
• 24:47
  have
• 24:47
  so
• 24:47
  many
• 24:47
  preconceptions
• 24:48
  both
• 24:48
  about
• 24:48
  just
• 24:49
  the
• 24:49
  market
• 24:50
  and
• 24:50
  also
• 24:50
  just
• 24:50
  our
• 24:51
  brains
• 24:52
  are
• 24:52
  driven
• 24:52
  to
• 24:53
  anthropomorphize
• 24:54
  anything
• 24:54
  like
• 24:54
  this.
• 24:55
  And
• 24:56
  so,
• 24:56
  yeah,
• 24:56
  I
• 24:57
  think
• 24:57
  that
• 24:57
  we
• 24:57
  bring
• 24:57
  a
• 24:57
  little
• 24:57
  bit
• 24:58
  of
• 24:58
  dubiousness
• 24:58
  to
• 24:59
  inviting
• 24:59
  this
• 24:59
  into,
• 25:00
  like,
• 25:00
  a
• 25:00
  hardware
• 25:01
  world
• 25:01
  as
• 25:01
  well,
• 25:01
  which
• 25:02
  can
• 25:02
  make
• 25:02
  that
• 25:02
  more
• 25:03
  challenging
• 25:03
  when
• 25:03
  we're
• 25:03
  trying
• 25:04
  to
• 25:04
  find
• 25:04
  new
• 25:04
  methods
• 25:04
  for
• 25:05
  inputting
• 25:05
  and
• 25:06
  exploring
• 25:07
  our
• 25:07
  intent.
• 25:08
  Everyone
• 25:08
  kind
• 25:08
  of
• 25:09
  jumped
• 25:09
  to
• 25:09
  the
• 25:09
  assumption
• 25:10
  that
• 25:10
  we
• 25:10
  would
• 25:11
  want
• 25:11
  a
• 25:11
  chatbot
• 25:12
  on
• 25:12
  every
• 25:12
  website.
• 25:13
  And
• 25:14
  so,
• 25:14
  like,
• 25:14
  a
• 25:14
  lot
• 25:14
  of
• 25:14
  big
• 25:15
  companies
• 25:15
  would
• 25:15
  be
• 25:15
  like,
• 25:16
  how
• 25:16
  can
• 25:16
  I
• 25:16
  help
• 25:16
  you?
• 25:17
  It
• 25:17
  wasn't
• 25:17
  very
• 25:18
  helpful.
• 25:18
  What
• 25:18
  do
• 25:18
  you
• 25:19
  see
• 25:19
  when
• 25:20
  it
• 25:20
  comes
• 25:20
  to
• 25:20
  large
• 25:21
  companies
• 25:21
  approaching
• 25:22
  these
• 25:22
  things,
• 25:22
  newer
• 25:23
  companies
• 25:23
  approaching
• 25:24
  these
• 25:24
  things?
• 25:24
  What
• 25:24
  is
• 25:25
  your
• 25:25
  philosophy
• 25:25
  as
• 25:26
  far
• 25:26
  as
• 25:26
  how
• 25:26
  they
• 25:26
  should
• 25:27
  think
• 25:27
  about
• 25:28
  slapping
• 25:28
  AI
• 25:29
  onto
• 25:29
  existing
• 25:29
  products
• 25:30
  versus
• 25:30
  something
• 25:31
  new?
• 25:31
  It's
• 25:32
  a
• 25:32
  real
• 25:32
  challenge.
• 25:33
  I
• 25:33
  think,
• 25:33
  as
• 25:33
  I
• 25:33
  was
• 25:34
  just
• 25:34
  saying,
• 25:34
  there's
• 25:35
  skepticism
• 25:35
  about
• 25:35
  AI.
• 25:36
  There's
• 25:36
  some
• 25:36
  resistance
• 25:37
  in
• 25:37
  the
• 25:37
  market
• 25:38
  that
• 25:38
  has
• 25:38
  been
• 25:39
  contributed
• 25:39
  to
• 25:40
  hugely
• 25:40
  by
• 25:40
  this
• 25:41
  feeling
• 25:41
  that
• 25:41
  there
• 25:42
  is
• 25:42
  just
• 25:42
  a
• 25:42
  model
• 25:43
  on
• 25:43
  absolutely
• 25:43
  everything.
• 25:44
  Plus
• 25:45
  there's
• 25:45
  this
• 25:45
  really
• 25:46
  interesting
• 25:46
  study
• 25:47
  about
• 25:47
  something
• 25:47
  called
• 25:47
  algorithmic
• 25:48
  transference,
• 25:49
  which
• 25:49
  is
• 25:50
  when
• 25:50
  users
• 25:52
  experience
• 25:52
  a
• 25:53
  negative
• 25:53
  interaction
• 25:54
  with
• 25:54
  a
• 25:54
  large
• 25:54
  language
• 25:54
  model,
• 25:55
  they
• 25:55
  tend
• 25:56
  to
• 25:56
  project
• 25:57
  that
• 25:58
  same
• 25:58
  problem
• 25:58
  or
• 25:59
  that
• 25:59
  same
• 25:59
  level
• 26:00
  of
• 26:00
  capability,
• 26:01
  like
• 26:01
  the
• 26:01
  disappointing
• 26:02
  level
• 26:02
  of
• 26:02
  capability
• 26:03
  onto
• 26:04
  every
• 26:04
  other
• 26:04
  model.
• 26:05
  Meaning,
• 26:05
  like,
• 26:06
  if
• 26:06
  I
• 26:06
  ask
• 26:06
  a
• 26:07
  model
• 26:07
  to
• 26:07
  do
• 26:07
  something
• 26:07
  and
• 26:08
  it
• 26:08
  doesn't
• 26:08
  do
• 26:08
  it
• 26:08
  right,
• 26:08
  then
• 26:09
  I'm
• 26:09
  like,
• 26:09
  oh,
• 26:09
  none
• 26:10
  of
• 26:10
  them
• 26:10
  can
• 26:10
  do
• 26:10
  this.
• 26:10
  AI
• 26:11
  sucks.
• 26:11
  Yeah,
• 26:12
  a
• 26:12
  little
• 26:12
  bit.
• 26:12
  Subconsciously
• 26:13
  you're
• 26:13
  going
• 26:13
  to
• 26:13
  be
• 26:13
  like,
• 26:14
  oh,
• 26:14
  I'm
• 26:14
  not
• 26:14
  going
• 26:14
  to
• 26:14
  try
• 26:15
  doing
• 26:15
  that
• 26:15
  again.
• 26:16
  How
• 26:16
  do
• 26:16
  you
• 26:16
  think
• 26:16
  it
• 26:17
  is
• 26:17
  good
• 26:17
  for
• 26:17
  a
• 26:18
  large
• 26:18
  company
• 26:18
  to
• 26:18
  approach
• 26:19
  AI
• 26:20
  and
• 26:20
  adding
• 26:20
  AI
• 26:21
  that's
• 26:22
  been
• 26:22
  existing
• 26:22
  versus
• 26:23
  a
• 26:23
  company
• 26:24
  that
• 26:24
  is
• 26:24
  just
• 26:24
  starting
• 26:25
  thinking
• 26:25
  about
• 26:25
  their
• 26:26
  user
• 26:26
  interface
• 26:26
  and
• 26:27
  how
• 26:27
  to
• 26:27
  present
• 26:27
  it?
• 26:28
  I
• 26:28
  mean,
• 26:28
  small
• 26:28
  companies
• 26:28
  are
• 26:29
  so
• 26:29
  much
• 26:29
  more
• 26:29
  agile
• 26:30
  just
• 26:30
  by
• 26:30
  their
• 26:30
  very
• 26:30
  nature,
• 26:31
  which
• 26:31
  I
• 26:31
  think
• 26:31
  allows
• 26:32
  them
• 26:32
  to
• 26:32
  work
• 26:33
  much
• 26:33
  more
• 26:33
  closely
• 26:34
  with
• 26:34
  this
• 26:35
  explorative
• 26:35
  technology
• 26:36
  that
• 26:36
  really
• 26:37
  behooves
• 26:38
  small
• 26:39
  and
• 26:39
  interested
• 26:40
  experiments
• 26:41
  by
• 26:41
  passionate
• 26:41
  workers.
• 26:42
  However,
• 26:43
  large
• 26:43
  companies
• 26:44
  have
• 26:44
  that
• 26:44
  too,
• 26:44
  if
• 26:44
  they
• 26:45
  can
• 26:45
  structure
• 26:46
  to
• 26:46
  allow
• 26:46
  it.
• 26:47
  The
• 26:47
  real
• 26:47
  problem
• 26:47
  I
• 26:48
  see
• 26:48
  with
• 26:48
  large
• 26:48
  companies
• 26:49
  is
• 26:50
  they've
• 26:50
  got
• 26:51
  users
• 26:52
  who
• 26:52
  enjoy
• 26:53
  their
• 26:53
  product,
• 26:54
  that
• 26:54
  know
• 26:54
  how
• 26:54
  to
• 26:54
  use
• 26:55
  their
• 26:55
  product
• 26:55
  to
• 26:55
  accomplish
• 26:56
  a
• 26:56
  goal.
• 26:57
  Putting
• 26:58
  a
• 26:58
  useless
• 26:59
  AI
• 26:59
  tool
• 27:00
  in
• 27:00
  their
• 27:00
  face
• 27:01
  causes
• 27:01
  so
• 27:02
  much
• 27:02
  friction
• 27:02
  for
• 27:02
  them.
• 27:03
  And
• 27:03
  it
• 27:04
  really
• 27:05
  stuns
• 27:05
  me
• 27:05
  because
• 27:06
  there's
• 27:06
  such
• 27:06
  an
• 27:06
  opportunity
• 27:07
  there.
• 27:07
  There's
• 27:08
  these
• 27:08
  researchers
• 27:09
  called
• 27:09
  Frank
• 27:09
  and
• 27:09
  Lillian
• 27:10
  Gilbraith,
• 27:11
  who
• 27:11
  invented
• 27:12
  the
• 27:12
  field
• 27:12
  of
• 27:12
  ergonomics
• 27:13
  way
• 27:13
  back
• 27:13
  in
• 27:14
  the
• 27:14
  40s,
• 27:15
  50s.
• 27:15
  And
• 27:16
  the
• 27:16
  way
• 27:16
  that
• 27:17
  they
• 27:17
  did
• 27:17
  this
• 27:17
  was
• 27:18
  by
• 27:18
  taking
• 27:18
  long
• 27:19
  snapshots
• 27:20
  of
• 27:20
  how
• 27:20
  a
• 27:20
  worker
• 27:21
  laid
• 27:21
  a
• 27:21
  brick
• 27:22
  or
• 27:22
  how
• 27:22
  like
• 27:22
  a
• 27:23
  phone
• 27:23
  cable
• 27:23
  worker
• 27:24
  worked.
• 27:24
  And
• 27:24
  they
• 27:25
  saw
• 27:25
  how
• 27:25
  many
• 27:25
  movements
• 27:26
  were
• 27:26
  involved
• 27:27
  in
• 27:27
  that
• 27:27
  process
• 27:28
  and
• 27:28
  then
• 27:28
  they
• 27:28
  worked
• 27:28
  out
• 27:28
  how
• 27:29
  to
• 27:29
  simplify
• 27:29
  it
• 27:29
  from
• 27:29
  35
• 27:30
  movements
• 27:30
  down
• 27:30
  to
• 27:30
  12.
• 27:31
  I
• 27:32
  think
• 27:32
  that
• 27:32
  the
• 27:32
  opportunity
• 27:33
  for
• 27:33
  large
• 27:33
  organizations
• 27:34
  to
• 27:34
  sit
• 27:35
  and
• 27:35
  look
• 27:35
  at
• 27:35
  the
• 27:35
  many
• 27:36
  different
• 27:36
  people
• 27:36
  who
• 27:36
  use
• 27:36
  their
• 27:37
  products
• 27:37
  and
• 27:38
  to
• 27:38
  observe
• 27:39
  carefully.
• 27:39
  Okay,
• 27:40
  we
• 27:40
  have
• 27:40
  10
• 27:40
  different
• 27:40
  primary
• 27:41
  users.
• 27:41
  They
• 27:42
  have
• 27:42
  12
• 27:42
  different
• 27:42
  tasks
• 27:43
  they
• 27:43
  frequently
• 27:43
  accomplish.
• 27:44
  If
• 27:44
  we
• 27:44
  lay
• 27:45
  those
• 27:45
  out
• 27:46
  item
• 27:46
  by
• 27:46
  item,
• 27:47
  we
• 27:47
  can
• 27:47
  see
• 27:48
  where
• 27:48
  AI
• 27:48
  can
• 27:48
  be
• 27:48
  assistive,
• 27:49
  where
• 27:49
  it
• 27:49
  might
• 27:49
  not
• 27:50
  be,
• 27:50
  and
• 27:50
  then
• 27:50
  we
• 27:51
  can
• 27:51
  slot
• 27:51
  in
• 27:51
  to
• 27:52
  our
• 27:52
  product
• 27:53
  where
• 27:53
  they
• 27:54
  could
• 27:54
  find
• 27:54
  that
• 27:54
  or
• 27:55
  where
• 27:55
  it
• 27:55
  can
• 27:55
  show
• 27:56
  having
• 27:56
  it
• 27:56
  just
• 27:57
  be
• 27:57
  revealed
• 27:57
  at
• 27:57
  these
• 27:58
  arbitrary
• 27:59
  moments
• 27:59
  feels
• 28:00
  like
• 28:00
  it
• 28:00
  just
• 28:00
  ignores
• 28:01
  the
• 28:02
  trust
• 28:02
  that
• 28:02
  a
• 28:03
  company
• 28:03
  could
• 28:03
  gain
• 28:04
  from
• 28:04
  leveraging
• 28:05
  that.
• 28:05
  I
• 28:06
  know
• 28:06
  that
• 28:06
  we
• 28:06
  are
• 28:06
  all
• 28:06
  a
• 28:07
  bit
• 28:07
  nervous
• 28:07
  about
• 28:08
  data
• 28:08
  and
• 28:08
  observing
• 28:09
  in
• 28:09
  this
• 28:09
  way,
• 28:10
  but
• 28:10
  for
• 28:10
  me
• 28:10
  this
• 28:10
  comes
• 28:10
  from
• 28:11
  the
• 28:11
  Persona
• 28:12
  element
• 28:12
  of
• 28:12
  the
• 28:13
  design
• 28:13
  field,
• 28:13
  where
• 28:14
  we
• 28:14
  don't
• 28:14
  want
• 28:14
  to
• 28:14
  just
• 28:15
  be
• 28:15
  guessing
• 28:15
  what
• 28:15
  people
• 28:15
  are
• 28:15
  doing,
• 28:16
  pay
• 28:16
  attention
• 28:17
  and
• 28:17
  sit
• 28:17
  with
• 28:18
  the
• 28:18
  user.
• 28:18
  And
• 28:19
  you
• 28:19
  said
• 28:19
  this
• 28:19
  earlier
• 28:19
  about
• 28:20
  how
• 28:20
  it's
• 28:21
  almost
• 28:21
  never
• 28:22
  been
• 28:22
  more
• 28:22
  important
• 28:23
  to
• 28:23
  have
• 28:24
  these
• 28:24
  core
• 28:24
  design
• 28:25
  sort
• 28:25
  of
• 28:25
  thinking
• 28:26
  principles
• 28:26
  applied
• 28:27
  to
• 28:27
  how
• 28:27
  you
• 28:27
  build
• 28:27
  product.
• 28:28
  Do
• 28:28
  you
• 28:28
  think
• 28:29
  it's
• 28:29
  also
• 28:29
  more
• 28:29
  important
• 28:30
  than
• 28:30
  ever
• 28:30
  for
• 28:30
  product
• 28:31
  designers
• 28:32
  to
• 28:32
  be
• 28:34
  informed
• 28:35
  by
• 28:35
  practitioners
• 28:36
  to
• 28:36
  some
• 28:36
  degree
• 28:37
  of
• 28:38
  what
• 28:38
  is
• 28:38
  newly
• 28:39
  possible?
• 28:40
  So
• 28:40
  that
• 28:40
  when
• 28:40
  you're
• 28:41
  thinking
• 28:41
  about
• 28:42
  the
• 28:42
  user,
• 28:42
  Persona
• 28:43
  is
• 28:43
  who
• 28:43
  you're
• 28:43
  designing
• 28:44
  for
• 28:44
  and
• 28:44
  what
• 28:44
  the
• 28:44
  action
• 28:44
  is,
• 28:45
  what
• 28:45
  the
• 28:45
  task
• 28:45
  is,
• 28:46
  what
• 28:46
  you
• 28:46
  want
• 28:46
  to
• 28:46
  enable
• 28:46
  them
• 28:47
  to
• 28:47
  do,
• 28:47
  you
• 28:48
  kind
• 28:48
  of
• 28:49
  have
• 28:49
  this
• 28:49
  new
• 28:49
  set
• 28:50
  of
• 28:50
  tools
• 28:50
  in
• 28:50
  your
• 28:50
  head
• 28:51
  of
• 28:51
  what
• 28:52
  that
• 28:52
  could
• 28:52
  look
• 28:52
  like
• 28:53
  100%.
• 28:54
  I
• 28:54
  think
• 28:54
  I
• 28:54
  write
• 28:54
  about
• 28:55
  this
• 28:55
  a
• 28:55
  little
• 28:55
  bit
• 28:55
  of
• 28:56
  operational
• 28:56
  maturity
• 28:57
  and
• 28:58
  how
• 28:58
  information
• 28:58
  is
• 28:59
  shared
• 28:59
  between
• 28:59
  the
• 29:00
  ENG
• 29:00
  designer
• 29:00
  and
• 29:00
  PM
• 29:01
  triad.
• 29:02
  And
• 29:02
  I
• 29:02
  suggest
• 29:02
  that
• 29:03
  the
• 29:03
  PM
• 29:03
  is
• 29:04
  kind
• 29:04
  of
• 29:04
  the
• 29:04
  person
• 29:04
  who
• 29:05
  can
• 29:06
  share
• 29:06
  the
• 29:06
  capabilities
• 29:07
  of
• 29:07
  models.
• 29:08
  But
• 29:08
  I'll
• 29:09
  add
• 29:09
  this
• 29:09
  is
• 29:10
  a
• 29:10
  very
• 29:10
  difficult
• 29:10
  area
• 29:11
  to
• 29:11
  work
• 29:11
  in
• 29:12
  because
• 29:12
  the
• 29:13
  decisions
• 29:13
  that
• 29:13
  are
• 29:14
  getting
• 29:14
  made
• 29:14
  by
• 29:14
  engineers
• 29:15
  to
• 29:16
  create
• 29:16
  new
• 29:17
  capabilities
• 29:17
  within
• 29:18
  the
• 29:18
  models
• 29:18
  are
• 29:18
  not
• 29:18
  guided
• 29:19
  very
• 29:19
  much
• 29:19
  by
• 29:20
  user
• 29:20
  need.
• 29:21
  And
• 29:21
  that
• 29:21
  means
• 29:21
  that
• 29:22
  there's
• 29:22
  also
• 29:22
  sort
• 29:22
  of
• 29:22
  an
• 29:23
  upstream
• 29:23
  conversation
• 29:24
  that
• 29:24
  needs
• 29:24
  to
• 29:24
  be
• 29:24
  happening
• 29:24
  that
• 29:25
  I
• 29:25
  really
• 29:25
  like
• 29:25
  prod
• 29:26
  designers
• 29:26
  to
• 29:26
  be
• 29:26
  involved
• 29:27
  in
• 29:27
  because
• 29:28
  there
• 29:28
  are
• 29:29
  capabilities
• 29:29
  that
• 29:29
  we're
• 29:30
  seeing
• 29:30
  that
• 29:30
  are
• 29:30
  not
• 29:31
  that
• 29:31
  profoundly
• 29:31
  useful
• 29:32
  and
• 29:33
  we
• 29:33
  don't
• 29:33
  necessarily
• 29:34
  know
• 29:34
  what
• 29:35
  could
• 29:35
  be
• 29:35
  created
• 29:35
  that
• 29:35
  actually
• 29:36
  would
• 29:36
  be
• 29:36
  assistive
• 29:36
  to
• 29:37
  users.
• 29:37
  If
• 29:37
  there
• 29:37
  is
• 29:38
  a
• 29:38
  little
• 29:38
  bit
• 29:38
  more
• 29:38
  knowledge
• 29:39
  about
• 29:40
  what
• 29:40
  people
• 29:40
  actually
• 29:41
  need
• 29:41
  from
• 29:41
  their
• 29:41
  products.
• 29:42
  Going
• 29:42
  with
• 29:43
  what
• 29:43
  you
• 29:43
  said
• 29:43
  about
• 29:43
  ergonomics
• 29:44
  and
• 29:44
  reducing
• 29:45
  the
• 29:46
  amount
• 29:46
  of
• 29:46
  interactions
• 29:47
  that
• 29:47
  users
• 29:48
  need
• 29:48
  to
• 29:48
  have,
• 29:48
  it's
• 29:49
  such
• 29:49
  a
• 29:49
  good
• 29:49
  philosophy
• 29:51
  as
• 29:51
  far
• 29:51
  as
• 29:51
  when
• 29:51
  it
• 29:51
  comes
• 29:52
  to
• 29:52
  interaction
• 29:52
  design.
• 29:54
  Small
• 29:54
  companies,
• 29:54
  I
• 29:55
  think
• 29:55
  are
• 29:55
  really
• 29:56
  best
• 29:57
  equipped
• 29:57
  for
• 29:57
  this
• 29:58
  because
• 29:58
  even
• 29:58
  in
• 29:58
  large
• 29:59
  companies,
• 29:59
  I'm
• 29:59
  sure
• 29:59
  you've
• 30:00
  seen
• 30:00
  having
• 30:01
  been
• 30:01
  at
• 30:01
  Microsoft,
• 30:02
  and
• 30:02
  when
• 30:02
  you
• 30:03
  work
• 30:03
  in
• 30:03
  production,
• 30:04
  often
• 30:04
  you
• 30:04
  have
• 30:04
  KPIs
• 30:05
  that
• 30:05
  are
• 30:05
  linked
• 30:05
  to
• 30:06
  user
• 30:06
  interactions.
• 30:07
  You
• 30:07
  kind
• 30:07
  of
• 30:07
  have
• 30:08
  a
• 30:08
  perverse
• 30:08
  incentive
• 30:09
  for
• 30:09
  more
• 30:09
  interactions
• 30:10
  rather
• 30:11
  than
• 30:11
  less.
• 30:12
  Because
• 30:12
  you
• 30:13
  know,
• 30:13
  if
• 30:13
  you
• 30:13
  were
• 30:13
  to
• 30:13
  remove
• 30:14
  interactions,
• 30:15
  then
• 30:15
  your
• 30:15
  KPIs
• 30:16
  are
• 30:16
  going
• 30:16
  down.
• 30:17
  So
• 30:17
  there's
• 30:17
  like
• 30:18
  this
• 30:18
  internal
• 30:18
  friction
• 30:19
  against
• 30:19
  these
• 30:19
  things.
• 30:20
  You
• 30:20
  said
• 30:21
  that
• 30:21
  small
• 30:21
  companies
• 30:22
  are
• 30:23
  well
• 30:23
  equipped
• 30:24
  to
• 30:24
  do
• 30:24
  this.
• 30:24
  It
• 30:24
  seems
• 30:25
  like
• 30:25
  the
• 30:25
  approach
• 30:25
  you're
• 30:26
  suggesting
• 30:26
  is
• 30:26
  thinking
• 30:26
  about
• 30:27
  what
• 30:27
  is
• 30:28
  the
• 30:28
  least
• 30:28
  friction
• 30:29
  way
• 30:29
  people
• 30:29
  can
• 30:30
  interact
• 30:30
  with
• 30:30
  our
• 30:30
  model
• 30:32
  when
• 30:32
  it
• 30:32
  comes
• 30:32
  to
• 30:32
  thinking
• 30:32
  about
• 30:33
  how
• 30:33
  users
• 30:34
  actually
• 30:34
  interact
• 30:35
  with
• 30:35
  the
• 30:35
  model
• 30:36
  for
• 30:36
  the
• 30:36
  product.
• 30:36
  Is
• 30:37
  that
• 30:37
  something
• 30:37
  you
• 30:37
  talk
• 30:37
  about
• 30:38
  in
• 30:38
  your
• 30:38
  book?
• 30:38
  Yeah,
• 30:39
  a
• 30:39
  little
• 30:39
  bit.
• 30:39
  I
• 30:40
  mean,
• 30:40
  easily
• 30:41
  the
• 30:41
  most
• 30:41
  difficult
• 30:41
  way
• 30:42
  that
• 30:42
  a
• 30:42
  company
• 30:42
  decides
• 30:43
  to
• 30:43
  build
• 30:43
  their
• 30:43
  product
• 30:43
  is
• 30:44
  to
• 30:44
  build
• 30:44
  it
• 30:44
  for
• 30:44
  everyone.
• 30:45
  And
• 30:46
  narrowing
• 30:46
  in
• 30:46
  on
• 30:47
  a
• 30:47
  specialty
• 30:47
  is
• 30:48
  the
• 30:48
  way
• 30:48
  that
• 30:48
  I
• 30:48
  think
• 30:49
  you
• 30:49
  begin
• 30:49
  to
• 30:49
  see
• 30:49
  real
• 30:49
  progress.
• 30:50
  I
• 30:50
  think
• 30:50
  the
• 30:51
  first
• 30:51
  thing
• 30:51
  that's
• 30:51
  really
• 30:52
  begun
• 30:52
  to
• 30:52
  feel
• 30:53
  solid
• 30:53
  in
• 30:53
  this
• 30:53
  field
• 30:54
  right
• 30:54
  now,
• 30:55
  just
• 30:55
  in
• 30:55
  terms
• 30:55
  of
• 30:56
  all
• 30:56
  the
• 30:56
  major
• 30:57
  flagship
• 30:58
  models,
• 30:58
  all
• 30:59
  competing
• 31:00
  and
• 31:00
  how
• 31:00
  useless
• 31:01
  at
• 31:01
  times
• 31:01
  the
• 31:02
  methods
• 31:02
  for
• 31:02
  evaluating
• 31:03
  them
• 31:03
  can
• 31:03
  feel,
• 31:04
  is
• 31:04
  that
• 31:04
  Claude
• 31:05
  has
• 31:05
  begun
• 31:05
  to
• 31:05
  specialize
• 31:06
  into
• 31:06
  CLAUDE
• 31:06
  code.
• 31:07
  And
• 31:07
  we
• 31:08
  know
• 31:08
  that
• 31:09
  that
• 31:09
  is
• 31:09
  the
• 31:09
  model
• 31:09
  for
• 31:10
  coding.
• 31:10
  Right.
• 31:11
  There's
• 31:11
  still
• 31:11
  Gemini
• 31:12
  and
• 31:12
  Codex
• 31:12
  and
• 31:12
  whatever
• 31:12
  that
• 31:13
  are
• 31:13
  trying
• 31:13
  to
• 31:13
  compete.
• 31:14
  But
• 31:14
  if
• 31:14
  I'm
• 31:14
  answering
• 31:15
  your
• 31:15
  question
• 31:15
  correctly,
• 31:16
  I'm
• 31:16
  trying
• 31:16
  to
• 31:16
  sort
• 31:17
  of
• 31:17
  make
• 31:17
  the
• 31:17
  argument
• 31:18
  that
• 31:19
  specialties
• 31:20
  should
• 31:20
  not
• 31:21
  be
• 31:21
  a
• 31:22
  problem
• 31:22
  unless
• 31:23
  you're
• 31:23
  Google
• 31:23
  and
• 31:23
  your
• 31:24
  product
• 31:24
  basically
• 31:25
  makes
• 31:25
  money
• 31:25
  out
• 31:25
  of
• 31:25
  showing
• 31:26
  everything
• 31:26
  to
• 31:26
  everyone.
• 31:28
  Narrowing
• 31:28
  in
• 31:29
  should
• 31:29
  be
• 31:30
  the
• 31:30
  first
• 31:30
  thing
• 31:30
  that
• 31:30
  you
• 31:31
  want
• 31:31
  to
• 31:31
  do,
• 31:31
  in
• 31:32
  my
• 31:32
  opinion,
• 31:32
  in
• 31:32
  the
• 31:32
  AI
• 31:33
  era,
• 31:33
  the
• 31:33
  more
• 31:34
  specific
• 31:34
  that
• 31:34
  you
• 31:35
  get,
• 31:35
  even
• 31:35
  if
• 31:35
  it
• 31:36
  is
• 31:36
  just
• 31:36
  a
• 31:36
  wrapper,
• 31:37
  I
• 31:37
  think
• 31:37
  there
• 31:37
  are
• 31:37
  massively
• 31:38
  successful
• 31:38
  products
• 31:39
  out
• 31:39
  there
• 31:39
  that
• 31:40
  are
• 31:40
  going
• 31:40
  to
• 31:41
  remain
• 31:41
  robustly
• 31:42
  successful
• 31:42
  by
• 31:43
  knowing
• 31:43
  their
• 31:44
  user
• 31:44
  need
• 31:44
  and
• 31:45
  getting
• 31:45
  passionately
• 31:46
  into
• 31:46
  it
• 31:46
  in
• 31:46
  a
• 31:46
  way
• 31:46
  that
• 31:47
  no
• 31:47
  one
• 31:47
  else
• 31:47
  can
• 31:47
  copy
• 31:47
  very,
• 31:48
  very
• 31:48
  quickly.
• 31:49
  And
• 31:49
  that's
• 31:49
  the
• 31:50
  benefit
• 31:50
  of
• 31:50
  a
• 31:50
  small
• 31:50
  company,
• 31:50
  is
• 31:51
  also
• 31:51
  just
• 31:51
  being
• 31:51
  very
• 31:52
  adaptable
• 31:52
  and
• 31:52
  you
• 31:53
  generally
• 31:53
  have
• 31:53
  a
• 31:53
  better
• 31:54
  information
• 31:54
  sharing
• 31:55
  structure
• 31:55
  that
• 31:56
  allows
• 31:56
  you
• 31:56
  to
• 31:57
  understand
• 31:58
  the
• 31:58
  user
• 31:58
  much,
• 31:59
  much
• 31:59
  more
• 31:59
  quickly.
• 32:00
  If
• 32:00
  you're
• 32:01
  a
• 32:01
  designer
• 32:01
  listening
• 32:01
  to
• 32:01
  the
• 32:01
  show
• 32:02
  and
• 32:02
  you're
• 32:03
  used
• 32:03
  to
• 32:03
  traditional
• 32:03
  web
• 32:04
  interfaces,
• 32:05
  or
• 32:05
  maybe
• 32:05
  you're
• 32:06
  someone
• 32:06
  that
• 32:06
  just
• 32:06
  graduated
• 32:06
  and
• 32:07
  you're
• 32:07
  approaching
• 32:07
  this,
• 32:08
  how
• 32:08
  are
• 32:08
  you
• 32:08
  talking
• 32:08
  to
• 32:09
  folks
• 32:09
  like
• 32:09
  this
• 32:09
  or
• 32:09
  how
• 32:09
  are
• 32:10
  you
• 32:10
  guiding
• 32:10
  them
• 32:10
  as
• 32:11
  far
• 32:11
  as
• 32:11
  how
• 32:11
  to
• 32:11
  get
• 32:12
  started
• 32:12
  and
• 32:12
  how
• 32:12
  to
• 32:12
  start
• 32:13
  thinking
• 32:13
  about
• 32:13
  these
• 32:13
  things
• 32:14
  better?
• 32:15
  It's
• 32:15
  interesting,
• 32:16
  I
• 32:16
  think,
• 32:16
  that
• 32:17
  people
• 32:17
  who
• 32:17
  are
• 32:17
  junior
• 32:18
  in
• 32:18
  the
• 32:18
  field
• 32:18
  right
• 32:18
  now
• 32:19
  have
• 32:19
  a
• 32:20
  ton
• 32:20
  of
• 32:20
  opportunity,
• 32:21
  but
• 32:21
  the
• 32:21
  way
• 32:22
  forward
• 32:22
  is
• 32:22
  very
• 32:22
  unclear.
• 32:24
  Exploring
• 32:24
  is
• 32:25
  always
• 32:25
  going
• 32:25
  to
• 32:25
  be
• 32:25
  the
• 32:26
  first
• 32:26
  place
• 32:26
  to
• 32:26
  start
• 32:27
  before
• 32:27
  you
• 32:27
  even
• 32:28
  begin
• 32:28
  to
• 32:28
  start
• 32:28
  investigating
• 32:29
  whether
• 32:29
  AI
• 32:30
  is
• 32:30
  your
• 32:30
  field
• 32:30
  you
• 32:30
  can
• 32:31
  sit
• 32:31
  and
• 32:31
  use.
• 32:31
  Maybe
• 32:32
  not
• 32:32
  Claude
• 32:32
  code,
• 32:33
  but
• 32:33
  maybe
• 32:33
  like
• 32:33
  Claude
• 32:34
  Artifacts
• 32:34
  or
• 32:34
  Claude
• 32:35
  Cowork
• 32:35
  or.
• 32:36
  I
• 32:36
  sound
• 32:36
  like
• 32:36
  I'm
• 32:36
  an
• 32:37
  anthropic
• 32:37
  shill.
• 32:37
  I'm
• 32:37
  not.
• 32:38
  It's
• 32:38
  just
• 32:38
  good
• 32:38
  Skip.
• 32:38
  Yeah,
• 32:39
  you
• 32:39
  could
• 32:39
  use
• 32:39
  Gemini,
• 32:40
  you
• 32:40
  could
• 32:40
  use
• 32:40
  for
• 32:40
  Flexi.
• 32:40
  You
• 32:40
  could
• 32:40
  use
• 32:41
  anything.
• 32:41
  I
• 32:41
  think
• 32:41
  gaining
• 32:42
  an
• 32:42
  interest
• 32:43
  in
• 32:44
  the
• 32:44
  field
• 32:44
  has
• 32:45
  always
• 32:45
  been
• 32:45
  the
• 32:45
  way
• 32:46
  to
• 32:46
  work,
• 32:46
  particularly
• 32:47
  when
• 32:47
  we're
• 32:47
  way
• 32:47
  beyond
• 32:48
  qualifications.
• 32:49
  Right
• 32:49
  now
• 32:49
  there's
• 32:49
  no
• 32:50
  big
• 32:50
  degree
• 32:50
  program.
• 32:51
  You
• 32:51
  get
• 32:52
  to
• 32:52
  go
• 32:52
  and
• 32:52
  be
• 32:52
  an
• 32:53
  AI
• 32:53
  designer.
• 32:54
  I
• 32:54
  would
• 32:54
  also
• 32:54
  say
• 32:55
  maybe
• 32:55
  this
• 32:55
  is
• 32:56
  controversial,
• 32:56
  but
• 32:57
  don't
• 32:57
  maybe
• 32:58
  make
• 32:58
  yourself
• 32:58
  an
• 32:58
  AI
• 32:59
  designer
• 32:59
  by
• 33:00
  specialty.
• 33:01
  I
• 33:01
  think
• 33:02
  even
• 33:02
  more
• 33:03
  deeply
• 33:04
  we
• 33:04
  see
• 33:04
  a
• 33:05
  need
• 33:05
  for
• 33:05
  people
• 33:05
  who
• 33:05
  are
• 33:06
  good
• 33:06
  at
• 33:06
  database
• 33:07
  structure
• 33:08
  visualization.
• 33:09
  There's
• 33:10
  specific
• 33:10
  areas
• 33:11
  that
• 33:11
  AI
• 33:11
  is
• 33:12
  going
• 33:12
  to
• 33:12
  make
• 33:12
  it
• 33:12
  very
• 33:12
  easy
• 33:13
  to
• 33:13
  build
• 33:13
  and
• 33:13
  to
• 33:14
  show.
• 33:14
  And
• 33:15
  emphasizing
• 33:16
  that
• 33:16
  and
• 33:16
  discovering
• 33:17
  that
• 33:17
  yourself
• 33:18
  as
• 33:18
  a
• 33:18
  designer
• 33:18
  is
• 33:19
  going
• 33:19
  to
• 33:19
  make
• 33:19
  you
• 33:20
  much
• 33:20
  more
• 33:20
  useful
• 33:20
  in
• 33:20
  the
• 33:20
  market
• 33:21
  because
• 33:22
  certain
• 33:22
  other
• 33:22
  elements
• 33:23
  are
• 33:23
  going
• 33:23
  to
• 33:23
  go
• 33:23
  away.
• 33:23
  We're
• 33:24
  probably
• 33:24
  actually
• 33:24
  going
• 33:25
  to
• 33:25
  see
• 33:25
  branding
• 33:25
  become
• 33:26
  a
• 33:27
  much
• 33:27
  more
• 33:27
  important
• 33:28
  element
• 33:28
  because
• 33:28
  it's
• 33:28
  a
• 33:29
  key
• 33:29
  differentiator
• 33:29
  for
• 33:30
  a
• 33:30
  lot
• 33:30
  of
• 33:30
  these
• 33:30
  products.
• 33:31
  But
• 33:31
  yeah,
• 33:32
  I
• 33:32
  really
• 33:32
  just
• 33:33
  suggest
• 33:34
  a
• 33:34
  lot
• 33:34
  of
• 33:34
  exploration
• 33:35
  and
• 33:35
  a
• 33:36
  lot
• 33:36
  of
• 33:36
  knowledge
• 33:36
  about
• 33:36
  the
• 33:37
  way
• 33:37
  different
• 33:37
  models
• 33:38
  work
• 33:39
  and
• 33:39
  the
• 33:39
  way
• 33:39
  that
• 33:39
  their
• 33:39
  outputs
• 33:40
  work.
• 33:40
  Particularly
• 33:40
  if
• 33:41
  you're
• 33:42
  feeding
• 33:42
  an
• 33:42
  input
• 33:43
  along
• 33:43
  with
• 33:44
  all
• 33:44
  of
• 33:44
  the
• 33:44
  other
• 33:45
  stuff
• 33:45
  that
• 33:45
  gets
• 33:45
  structured
• 33:46
  into
• 33:46
  an
• 33:46
  input,
• 33:46
  like
• 33:47
  gaining
• 33:47
  a
• 33:47
  good
• 33:47
  understanding
• 33:48
  of
• 33:48
  what's
• 33:49
  happening
• 33:49
  under
• 33:49
  the
• 33:49
  hood
• 33:50
  so
• 33:50
  that
• 33:50
  you
• 33:50
  can
• 33:50
  be
• 33:50
  communicating
• 33:51
  that
• 33:51
  to
• 33:51
  your
• 33:51
  user
• 33:52
  and
• 33:52
  showing
• 33:52
  them
• 33:53
  the
• 33:53
  model's
• 33:53
  capabilities
• 33:54
  and
• 33:54
  anticipating
• 33:55
  what
• 33:55
  comes
• 33:56
  next.
• 33:56
  That
• 33:57
  is,
• 33:57
  I
• 33:57
  think,
• 33:57
  the
• 33:57
  best
• 33:57
  way
• 33:58
  to
• 33:58
  be
• 33:58
  working
• 33:58
  right
• 33:58
  now.
• 33:59
  A
• 33:59
  super
• 34:00
  under
• 34:00
  discussed,
• 34:01
  I
• 34:01
  think,
• 34:01
  element
• 34:01
  of
• 34:01
  product
• 34:02
  design
• 34:02
  that's
• 34:02
  kind
• 34:02
  of
• 34:02
  long
• 34:03
  standing,
• 34:03
  which
• 34:03
  is
• 34:03
  the
• 34:03
  idea
• 34:04
  of
• 34:04
  posture,
• 34:05
  whether
• 34:05
  something
• 34:05
  takes
• 34:06
  up
• 34:06
  your
• 34:06
  whole
• 34:06
  screen,
• 34:07
  your
• 34:07
  whole
• 34:07
  attention,
• 34:08
  or
• 34:08
  is
• 34:08
  passively
• 34:08
  taking
• 34:09
  your
• 34:09
  attention.
• 34:10
  And
• 34:10
  I
• 34:10
  think
• 34:10
  AI
• 34:11
  sits
• 34:11
  really
• 34:11
  at
• 34:12
  difficult
• 34:12
  regions
• 34:12
  of
• 34:12
  that
• 34:13
  because
• 34:13
  sometimes
• 34:13
  you
• 34:14
  want
• 34:14
  something
• 34:14
  just
• 34:14
  to
• 34:14
  be
• 34:15
  passively
• 34:15
  existing
• 34:15
  and
• 34:16
  working
• 34:16
  on
• 34:16
  it
• 34:16
  and
• 34:16
  you
• 34:16
  don't
• 34:16
  have
• 34:16
  to
• 34:16
  pay
• 34:16
  a
• 34:17
  lot
• 34:17
  of
• 34:17
  attention
• 34:17
  to
• 34:17
  it.
• 34:17
  Agentic
• 34:18
  tools
• 34:18
  fall
• 34:18
  into
• 34:18
  this
• 34:19
  category.
• 34:19
  But
• 34:19
  then
• 34:19
  sometimes
• 34:20
  you're
• 34:20
  using
• 34:20
  something
• 34:20
  with
• 34:21
  your
• 34:21
  whole
• 34:21
  attention
• 34:22
  and
• 34:22
  yeah,
• 34:22
  again,
• 34:22
  I
• 34:22
  don't
• 34:23
  have
• 34:23
  a
• 34:23
  point
• 34:23
  to
• 34:23
  make
• 34:23
  about
• 34:24
  it,
• 34:24
  but
• 34:24
  I
• 34:24
  do
• 34:24
  think
• 34:24
  that
• 34:24
  that's
• 34:25
  probably
• 34:25
  going
• 34:25
  to
• 34:25
  be
• 34:26
  part
• 34:26
  of
• 34:26
  the
• 34:26
  development
• 34:26
  we
• 34:27
  see
• 34:27
  going
• 34:27
  forward.
• 34:28
  Becca,
• 34:28
  that's
• 34:28
  such
• 34:28
  a
• 34:29
  good
• 34:29
  point
• 34:29
  because
• 34:29
  it
• 34:29
  made
• 34:29
  me
• 34:29
  think
• 34:30
  about
• 34:30
  how
• 34:30
  I
• 34:30
  interact
• 34:31
  with
• 34:31
  AI
• 34:31
  and
• 34:32
  there's
• 34:32
  so
• 34:32
  many
• 34:32
  versions
• 34:33
  of
• 34:33
  it
• 34:33
  now,
• 34:33
  right?
• 34:33
  Because
• 34:33
  like
• 34:34
  I'm
• 34:34
  in
• 34:34
  the
• 34:34
  CLI
• 34:35
  talking
• 34:35
  to
• 34:35
  Claude,
• 34:36
  I
• 34:36
  have
• 34:36
  openclaw
• 34:37
  on
• 34:37
  my
• 34:37
  machine,
• 34:38
  I'm
• 34:39
  sometimes
• 34:39
  in
• 34:39
  the
• 34:39
  browser,
• 34:40
  but
• 34:40
  actually
• 34:40
  most
• 34:40
  of
• 34:40
  the
• 34:41
  time
• 34:41
  I'm
• 34:41
  not
• 34:41
  in
• 34:41
  the
• 34:41
  browser
• 34:42
  talking
• 34:42
  to
• 34:42
  AI.
• 34:43
  And
• 34:43
  so
• 34:43
  I
• 34:43
  wonder
• 34:44
  how
• 34:44
  much
• 34:44
  this
• 34:44
  is
• 34:44
  going
• 34:45
  to
• 34:45
  make
• 34:45
  us
• 34:45
  reimagine
• 34:46
  the
• 34:46
  browser.
• 34:47
  Like,
• 34:47
  and
• 34:47
  to
• 34:47
  your
• 34:47
  point,
• 34:48
  Luis,
• 34:48
  it's
• 34:48
  more
• 34:49
  often
• 34:49
  passive
• 34:49
  than
• 34:50
  not.
• 34:50
  Really
• 34:50
  interesting
• 34:51
  where
• 34:51
  it's
• 34:51
  going.
• 34:52
  There's
• 34:52
  a
• 34:52
  lot
• 34:52
  of
• 34:53
  like
• 34:53
  right
• 34:54
  now,
• 34:54
  what
• 34:54
  I
• 34:54
  see
• 34:54
  with
• 34:54
  people
• 34:55
  who
• 34:55
  are
• 34:55
  interested
• 34:55
  in
• 34:55
  vibe
• 34:56
  coding
• 34:56
  and
• 34:56
  have
• 34:56
  like
• 34:57
  spun
• 34:57
  it
• 34:57
  off
• 34:57
  into
• 34:57
  like
• 34:58
  their
• 34:58
  own
• 34:58
  little
• 34:58
  niche.
• 34:59
  There's
• 35:00
  something
• 35:00
  that
• 35:00
  I
• 35:00
  refer
• 35:00
  to
• 35:01
  almost
• 35:01
  as
• 35:01
  like
• 35:01
  jigs,
• 35:02
  you
• 35:02
  know,
• 35:02
  what
• 35:02
  A
• 35:02
  jig
• 35:02
  is
• 35:03
  like
• 35:03
  in
• 35:03
  woodworking.
• 35:04
  You
• 35:04
  build
• 35:04
  a
• 35:04
  jig
• 35:04
  so
• 35:05
  that
• 35:05
  you.
• 35:05
  You
• 35:05
  have
• 35:05
  to
• 35:05
  build
• 35:06
  12
• 35:06
  of
• 35:06
  the
• 35:06
  same
• 35:06
  things
• 35:07
  and
• 35:07
  cut
• 35:07
  them
• 35:07
  out
• 35:07
  the
• 35:07
  same
• 35:08
  way.
• 35:08
  A
• 35:08
  jig
• 35:08
  is
• 35:09
  something
• 35:09
  that
• 35:09
  you
• 35:09
  would
• 35:09
  put
• 35:10
  on
• 35:10
  your
• 35:11
  cutting
• 35:11
  machines
• 35:11
  so
• 35:12
  that
• 35:12
  it
• 35:12
  always
• 35:12
  works
• 35:13
  the
• 35:13
  same
• 35:13
  way.
• 35:13
  It's
• 35:13
  just
• 35:13
  very
• 35:13
  efficient.
• 35:14
  It's
• 35:15
  a
• 35:15
  really
• 35:15
  gnarly
• 35:16
  and
• 35:16
  useful
• 35:16
  method
• 35:17
  for
• 35:17
  grasping
• 35:18
  intent.
• 35:19
  So
• 35:19
  the
• 35:19
  problem
• 35:20
  we
• 35:20
  have
• 35:20
  now
• 35:20
  is
• 35:21
  that
• 35:21
  there's
• 35:21
  a
• 35:22
  bit
• 35:22
  of
• 35:22
  a
• 35:22
  threshold
• 35:22
  we've
• 35:23
  reached
• 35:23
  in
• 35:23
  terms
• 35:23
  of
• 35:23
  data.
• 35:24
  And
• 35:24
  also
• 35:26
  whatever
• 35:26
  AGI
• 35:27
  was
• 35:27
  going
• 35:27
  to
• 35:27
  theoretically
• 35:28
  be
• 35:28
  that
• 35:28
  we
• 35:29
  stopped
• 35:29
  talking
• 35:29
  about
• 35:30
  a
• 35:30
  little
• 35:30
  bit,
• 35:30
  that
• 35:31
  was
• 35:31
  definitely
• 35:31
  gated
• 35:32
  by
• 35:32
  the
• 35:32
  capability
• 35:33
  and
• 35:33
  really
• 35:34
  willingness
• 35:34
  of
• 35:34
  a
• 35:35
  user
• 35:35
  to
• 35:35
  give
• 35:36
  so
• 35:36
  much
• 35:36
  personal
• 35:36
  information.
• 35:37
  For
• 35:38
  me
• 35:38
  to
• 35:38
  be
• 35:38
  able
• 35:38
  to
• 35:38
  tell
• 35:38
  a
• 35:39
  model
• 35:40
  how
• 35:40
  to
• 35:40
  choose
• 35:40
  the
• 35:41
  seat
• 35:41
  that
• 35:41
  I
• 35:42
  want
• 35:42
  to
• 35:42
  sit
• 35:42
  in
• 35:42
  on
• 35:42
  an
• 35:42
  airplane,
• 35:43
  there
• 35:43
  is
• 35:44
  actually
• 35:44
  a
• 35:44
  ton
• 35:44
  of
• 35:44
  cognitive
• 35:45
  decisions
• 35:45
  going
• 35:45
  in
• 35:46
  there
• 35:46
  where
• 35:46
  I'm
• 35:46
  like,
• 35:46
  well,
• 35:47
  if
• 35:47
  I'm
• 35:47
  on
• 35:47
  this
• 35:47
  row,
• 35:48
  I
• 35:48
  want
• 35:48
  to
• 35:48
  sit
• 35:48
  in
• 35:48
  the
• 35:48
  middle,
• 35:49
  but
• 35:49
  I
• 35:49
  always
• 35:49
  want
• 35:49
  to
• 35:49
  sit
• 35:49
  on
• 35:49
  this.
• 35:50
  And
• 35:50
  if
• 35:50
  you
• 35:50
  see
• 35:50
  that
• 35:50
  it's
• 35:51
  almost
• 35:51
  booked,
• 35:52
  then
• 35:52
  maybe
• 35:53
  just
• 35:53
  sit
• 35:53
  me
• 35:53
  here.
• 35:54
  There's
• 35:54
  so
• 35:54
  many
• 35:54
  weird
• 35:55
  little
• 35:55
  things
• 35:55
  I
• 35:55
  would
• 35:55
  need
• 35:55
  to
• 35:55
  explain
• 35:56
  to
• 35:56
  a
• 35:56
  model
• 35:56
  that
• 35:56
  are
• 35:56
  very
• 35:56
  particular
• 35:57
  to
• 35:57
  me.
• 35:57
  That
• 35:58
  made
• 35:58
  me
• 35:58
  think
• 35:58
  when
• 35:58
  I
• 35:59
  was
• 35:59
  writing
• 35:59
  the
• 35:59
  book
• 35:59
  so
• 35:59
  much
• 36:00
  about
• 36:00
  God
• 36:01
  intent.
• 36:01
  Getting
• 36:02
  user
• 36:02
  intent
• 36:02
  out
• 36:03
  of
• 36:03
  users
• 36:03
  is
• 36:04
  really
• 36:04
  the
• 36:04
  challenge.
• 36:04
  And
• 36:05
  also
• 36:05
  users
• 36:05
  don't
• 36:06
  love
• 36:06
  giving
• 36:06
  that
• 36:06
  up.
• 36:06
  Right
• 36:06
  now
• 36:07
  there's
• 36:07
  a
• 36:07
  bit
• 36:07
  of
• 36:07
  tech
• 36:08
  skepticism.
• 36:08
  Move
• 36:09
  fast
• 36:09
  and
• 36:09
  break
• 36:09
  things,
• 36:09
  make
• 36:10
  people
• 36:10
  feel
• 36:10
  like
• 36:10
  models
• 36:11
  and
• 36:11
  tools
• 36:11
  in
• 36:11
  general
• 36:12
  aren't
• 36:12
  going
• 36:12
  to
• 36:12
  be
• 36:12
  around
• 36:12
  forever.
• 36:13
  And
• 36:13
  so
• 36:13
  the
• 36:13
  idea
• 36:14
  of
• 36:14
  me
• 36:14
  sitting
• 36:14
  and
• 36:14
  telling
• 36:15
  something
• 36:15
  every
• 36:16
  single
• 36:16
  preference
• 36:17
  I
• 36:17
  have,
• 36:17
  and
• 36:17
  then
• 36:18
  maybe
• 36:18
  it
• 36:18
  just
• 36:18
  gets
• 36:19
  sunset
• 36:20
  two
• 36:20
  years
• 36:20
  later,
• 36:21
  I'm
• 36:21
  a
• 36:21
  little
• 36:21
  skeptical
• 36:22
  about
• 36:22
  it.
• 36:23
  So
• 36:23
  what
• 36:23
  I'm
• 36:23
  seeing
• 36:24
  that
• 36:24
  I'm
• 36:24
  really
• 36:24
  interested
• 36:25
  in
• 36:25
  now
• 36:26
  is
• 36:26
  this
• 36:27
  process
• 36:27
  of
• 36:27
  creating
• 36:28
  personal
• 36:29
  tools,
• 36:29
  these
• 36:29
  jigs
• 36:30
  that
• 36:31
  capture
• 36:31
  your
• 36:31
  own
• 36:32
  intent.
• 36:32
  My
• 36:33
  personal
• 36:33
  example
• 36:34
  is
• 36:34
  that
• 36:34
  I've
• 36:35
  never
• 36:35
  loved
• 36:35
  any
• 36:35
  Kanban
• 36:36
  tool
• 36:36
  particularly.
• 36:37
  I'm
• 36:37
  always
• 36:37
  really
• 36:37
  fussy
• 36:38
  about
• 36:38
  them.
• 36:38
  So
• 36:38
  I
• 36:38
  just
• 36:39
  spun
• 36:39
  up
• 36:39
  my
• 36:39
  own
• 36:40
  recently
• 36:40
  for
• 36:41
  the
• 36:41
  specific
• 36:41
  purpose
• 36:42
  that
• 36:42
  it's
• 36:42
  just
• 36:42
  running.
• 36:43
  I'm
• 36:43
  running
• 36:43
  it
• 36:43
  on
• 36:43
  my
• 36:43
  own
• 36:44
  dev
• 36:44
  server,
• 36:45
  but
• 36:45
  I
• 36:46
  will
• 36:46
  just
• 36:46
  sit
• 36:46
  and
• 36:46
  talk
• 36:47
  to
• 36:47
  cloud
• 36:47
  code
• 36:47
  about
• 36:47
  the
• 36:48
  tasks
• 36:48
  I
• 36:48
  have
• 36:48
  to
• 36:48
  do
• 36:48
  and
• 36:49
  have
• 36:49
  it
• 36:49
  break
• 36:49
  them
• 36:49
  down
• 36:49
  and
• 36:49
  shape
• 36:50
  them
• 36:50
  for
• 36:50
  me
• 36:50
  into
• 36:50
  the
• 36:50
  Kanban
• 36:51
  tool,
• 36:51
  which
• 36:51
  then
• 36:51
  goes
• 36:52
  into
• 36:52
  my
• 36:52
  schedule
• 36:52
  and
• 36:52
  then
• 36:53
  I
• 36:53
  have
• 36:53
  it
• 36:53
  logging
• 36:53
  into
• 36:53
  the
• 36:54
  Google
• 36:54
  Calendar
• 36:54
  API.
• 36:55
  So
• 36:55
  I
• 36:56
  have
• 36:56
  this
• 36:56
  really
• 36:57
  nice
• 36:57
  method
• 36:57
  by
• 36:57
  which
• 36:58
  I
• 36:58
  can
• 36:58
  just
• 36:58
  use.
• 36:59
  I
• 36:59
  have
• 36:59
  a
• 36:59
  model
• 37:00
  on
• 37:00
  my
• 37:00
  computer
• 37:01
  that
• 37:01
  I
• 37:01
  can
• 37:01
  just
• 37:01
  talk
• 37:01
  to
• 37:02
  and
• 37:02
  I
• 37:02
  can
• 37:02
  just
• 37:02
  talk
• 37:03
  eventually
• 37:03
  to
• 37:04
  my
• 37:04
  Google
• 37:04
  Calendar
• 37:05
  about
• 37:05
  what
• 37:06
  I
• 37:06
  need
• 37:06
  to
• 37:06
  get
• 37:06
  done
• 37:07
  and
• 37:07
  it
• 37:08
  ends
• 37:08
  up
• 37:08
  with
• 37:08
  an
• 37:08
  enormous
• 37:09
  corpus
• 37:09
  of
• 37:09
  my
• 37:10
  work.
• 37:10
  And
• 37:11
  I
• 37:11
  have
• 37:11
  it
• 37:12
  save
• 37:12
  everything
• 37:12
  that
• 37:13
  comes
• 37:13
  out
• 37:13
  of
• 37:13
  every
• 37:14
  week
• 37:14
  so
• 37:14
  that
• 37:15
  at
• 37:15
  the
• 37:15
  end
• 37:15
  of
• 37:15
  a
• 37:16
  month,
• 37:16
  I
• 37:16
  can
• 37:16
  get
• 37:16
  a
• 37:16
  summary
• 37:17
  of
• 37:17
  the
• 37:17
  tasks
• 37:17
  I
• 37:17
  did.
• 37:18
  That's
• 37:18
  all
• 37:18
  data
• 37:18
  that's
• 37:19
  not
• 37:19
  in
• 37:19
  any
• 37:19
  proprietary
• 37:19
  service,
• 37:20
  that's
• 37:20
  just
• 37:20
  sitting
• 37:20
  in
• 37:21
  a
• 37:21
  text
• 37:21
  file
• 37:21
  in
• 37:22
  a
• 37:22
  private
• 37:22
  repo.
• 37:23
  My
• 37:23
  partner
• 37:24
  the
• 37:24
  other
• 37:24
  day
• 37:24
  also
• 37:25
  just
• 37:25
  took
• 37:26
  all
• 37:26
  of
• 37:26
  his.
• 37:26
  This
• 37:27
  is
• 37:27
  really
• 37:27
  funny.
• 37:27
  He's
• 37:27
  been
• 37:27
  on
• 37:28
  an
• 37:28
  exercise
• 37:28
  where
• 37:28
  he's
• 37:29
  found
• 37:29
  all
• 37:29
  of
• 37:30
  his
• 37:30
  old
• 37:30
  hard
• 37:30
  drives
• 37:30
  and
• 37:30
  he's
• 37:31
  pulled
• 37:31
  out
• 37:31
  all
• 37:31
  of
• 37:31
  his
• 37:31
  old
• 37:32
  playlists
• 37:32
  he's
• 37:32
  made
• 37:32
  since
• 37:32
  he
• 37:32
  was
• 37:33
  14.
• 37:33
  He
• 37:34
  was
• 37:34
  like,
• 37:34
  I'm
• 37:34
  in
• 37:34
  this
• 37:34
  algorithmic
• 37:35
  jail
• 37:35
  with
• 37:35
  Spotify
• 37:36
  where
• 37:37
  it
• 37:37
  only
• 37:37
  really
• 37:37
  shows
• 37:38
  me
• 37:38
  the
• 37:38
  music
• 37:38
  it
• 37:38
  already
• 37:38
  knows
• 37:39
  I
• 37:39
  listen
• 37:39
  to,
• 37:39
  but
• 37:39
  it
• 37:39
  hasn't
• 37:40
  ever
• 37:40
  seen
• 37:41
  the
• 37:41
  music
• 37:42
  over
• 37:42
  this
• 37:42
  period
• 37:42
  of
• 37:42
  time.
• 37:43
  And
• 37:43
  so
• 37:43
  we
• 37:43
  just
• 37:44
  had
• 37:44
  Claude
• 37:44
  code
• 37:44
  strip
• 37:45
  all
• 37:45
  the
• 37:45
  files
• 37:45
  into
• 37:45
  just
• 37:46
  like,
• 37:46
  text,
• 37:47
  just
• 37:47
  like
• 37:47
  a
• 37:48
  list,
• 37:48
  use
• 37:49
  that
• 37:49
  API
• 37:50
  to
• 37:50
  have
• 37:50
  it
• 37:50
  generate
• 37:51
  playlists
• 37:51
  off
• 37:51
  his
• 37:51
  actual
• 37:52
  music.
• 37:52
  And
• 37:52
  that's
• 37:53
  like
• 37:53
  a
• 37:53
  huge
• 37:53
  corpus
• 37:54
  of
• 37:54
  your
• 37:54
  own
• 37:54
  personal
• 37:55
  data.
• 37:55
  And
• 37:56
  so
• 37:56
  I'm
• 37:56
  very
• 37:56
  curious
• 37:57
  about
• 37:57
  where
• 37:57
  people
• 37:58
  go
• 37:58
  with
• 37:58
  finding
• 37:59
  ways
• 37:59
  to.
• 37:59
  To
• 37:59
  tell
• 37:59
  their
• 38:00
  own
• 38:00
  story
• 38:00
  to
• 38:00
  a
• 38:01
  model.
• 38:01
  So
• 38:01
  that's
• 38:01
  actually
• 38:01
  a
• 38:01
  much
• 38:02
  more
• 38:02
  personalized
• 38:02
  experience.
• 38:03
  Those
• 38:03
  are
• 38:03
  two
• 38:04
  incredibly
• 38:05
  fascinating
• 38:06
  examples
• 38:07
  because
• 38:08
  they
• 38:08
  get
• 38:08
  to
• 38:10
  what
• 38:10
  I
• 38:10
  couldn't
• 38:10
  articulate
• 38:11
  earlier
• 38:12
  in
• 38:12
  the
• 38:12
  question
• 38:13
  around,
• 38:13
  where
• 38:13
  am
• 38:14
  I
• 38:14
  on
• 38:14
  the
• 38:14
  AI
• 38:14
  optimism
• 38:15
  scale,
• 38:16
  which
• 38:17
  is
• 38:18
  both
• 38:18
  the
• 38:18
  excitement
• 38:19
  of
• 38:20
  agentic
• 38:21
  capabilities
• 38:22
  or
• 38:22
  AI
• 38:22
  capabilities,
• 38:23
  and
• 38:24
  that
• 38:24
  creating
• 38:25
  this
• 38:25
  flywheel
• 38:25
  effect
• 38:26
  of
• 38:26
  I'm
• 38:26
  giving
• 38:27
  more
• 38:27
  and
• 38:27
  more
• 38:28
  information,
• 38:29
  intent
• 38:30
  information
• 38:30
  about
• 38:30
  my
• 38:31
  intent,
• 38:31
  information
• 38:31
  about
• 38:32
  me
• 38:32
  to
• 38:33
  my
• 38:33
  AI
• 38:34
  agent,
• 38:34
  because
• 38:35
  there's
• 38:35
  a
• 38:35
  direct
• 38:36
  value
• 38:36
  exchange.
• 38:37
  Like,
• 38:37
  it
• 38:37
  becomes
• 38:37
  so
• 38:37
  much
• 38:38
  more
• 38:38
  useful
• 38:38
  for
• 38:38
  me
• 38:38
  because
• 38:39
  it
• 38:39
  knows
• 38:39
  these
• 38:39
  things.
• 38:40
  Which
• 38:40
  is
• 38:41
  the
• 38:41
  part
• 38:41
  of
• 38:41
  me
• 38:41
  where
• 38:42
  I
• 38:42
  said
• 38:42
  earlier,
• 38:42
  I'm,
• 38:43
  like,
• 38:43
  having
• 38:43
  this
• 38:43
  creative
• 38:44
  explosion
• 38:44
  of
• 38:44
  like,
• 38:44
  holy
• 38:44
  shit,
• 38:45
  I
• 38:45
  can
• 38:45
  do
• 38:45
  so
• 38:46
  many
• 38:46
  things.
• 38:46
  But
• 38:47
  then
• 38:47
  there's
• 38:47
  this
• 38:47
  kind
• 38:48
  of
• 38:48
  scary
• 38:48
  part
• 38:49
  where
• 38:49
  you're
• 38:50
  like,
• 38:50
  is
• 38:51
  it
• 38:51
  bad
• 38:51
  for
• 38:51
  me
• 38:51
  to
• 38:51
  tell
• 38:52
  you
• 38:52
  all
• 38:52
  this
• 38:53
  stuff?
• 38:53
  How
• 38:54
  do
• 38:54
  you
• 38:54
  think
• 38:54
  about
• 38:55
  the
• 38:55
  balance
• 38:56
  between
• 38:57
  the
• 38:57
  intent
• 38:59
  information
• 38:59
  value
• 39:00
  exchange
• 39:01
  and
• 39:02
  user
• 39:03
  security
• 39:04
  or
• 39:04
  protection?
• 39:06
  Oh,
• 39:06
  it's
• 39:06
  terrible,
• 39:07
  Ashley.
• 39:07
  Scared.
• 39:09
  We
• 39:09
  have
• 39:09
  discussed
• 39:10
  trying
• 39:10
  to
• 39:10
  build
• 39:10
  our
• 39:10
  own
• 39:11
  server
• 39:11
  here
• 39:12
  very
• 39:12
  seriously.
• 39:13
  Because
• 39:14
  everything
• 39:15
  that
• 39:15
  you
• 39:16
  use
• 39:16
  in
• 39:16
  a
• 39:16
  model,
• 39:17
  given
• 39:17
  that
• 39:18
  it's
• 39:18
  running
• 39:19
  in
• 39:19
  some
• 39:19
  remote
• 39:20
  service
• 39:20
  somewhere,
• 39:21
  someone
• 39:21
  can
• 39:21
  send
• 39:21
  a
• 39:22
  subpoena.
• 39:22
  If
• 39:22
  there's
• 39:22
  like
• 39:22
  a
• 39:22
  lawsuit,
• 39:23
  you
• 39:23
  can
• 39:23
  take
• 39:23
  a
• 39:23
  look
• 39:23
  at
• 39:24
  it.
• 39:24
  Like,
• 39:24
  the
• 39:24
  companies
• 39:24
  are
• 39:25
  obliged
• 39:25
  to
• 39:25
  share
• 39:26
  any
• 39:26
  relevant
• 39:26
  information.
• 39:27
  And
• 39:28
  it's
• 39:28
  not
• 39:28
  like
• 39:28
  I'm
• 39:28
  like,
• 39:29
  in
• 39:29
  the
• 39:29
  middle
• 39:29
  of
• 39:29
  any
• 39:29
  heady
• 39:30
  lawsuit.
• 39:30
  But
• 39:31
  you
• 39:31
  do
• 39:31
  think
• 39:31
  about
• 39:31
  these
• 39:32
  things
• 39:32
  when
• 39:33
  you
• 39:33
  imagine
• 39:34
  that
• 39:34
  people
• 39:34
  are
• 39:35
  sharing
• 39:35
  their
• 39:35
  medical
• 39:35
  histories,
• 39:36
  their
• 39:36
  Personal
• 39:36
  histories.
• 39:37
  Like
• 39:37
  all
• 39:37
  this
• 39:37
  stuff
• 39:38
  that
• 39:38
  has
• 39:38
  been
• 39:38
  true,
• 39:39
  I
• 39:39
  will
• 39:39
  say,
• 39:39
  for
• 39:39
  a
• 39:39
  long
• 39:39
  time.
• 39:41
  I
• 39:41
  think
• 39:41
  a
• 39:41
  bunch
• 39:41
  of
• 39:42
  the
• 39:42
  Alexas
• 39:43
  got
• 39:43
  subpoenaed
• 39:44
  for
• 39:44
  murder
• 39:44
  cases
• 39:45
  and
• 39:46
  it
• 39:46
  has
• 39:46
  really
• 39:46
  been
• 39:46
  happening.
• 39:47
  But
• 39:47
  I
• 39:47
  don't
• 39:48
  think
• 39:48
  people
• 39:48
  were
• 39:49
  revealing
• 39:50
  so
• 39:50
  much
• 39:50
  personal
• 39:50
  information,
• 39:51
  maybe
• 39:52
  implicitly
• 39:52
  because
• 39:53
  they
• 39:53
  were
• 39:53
  being
• 39:53
  overheard.
• 39:54
  But
• 39:54
  I
• 39:55
  pour
• 39:55
  my
• 39:55
  heart
• 39:55
  open
• 39:56
  sometimes
• 39:56
  to
• 39:57
  when
• 39:58
  I'm
• 39:58
  really
• 39:58
  in
• 39:58
  the
• 39:58
  mood.
• 39:58
  Or
• 39:59
  even
• 39:59
  revealing
• 40:00
  personal
• 40:00
  information,
• 40:01
  but
• 40:01
  then
• 40:01
  also
• 40:01
  giving
• 40:02
  Claude
• 40:03
  code
• 40:03
  access
• 40:04
  to
• 40:04
  your
• 40:04
  stuff.
• 40:04
  For
• 40:05
  example,
• 40:05
  the
• 40:06
  thing
• 40:06
  that
• 40:06
  you
• 40:06
  built
• 40:07
  has
• 40:07
  access
• 40:07
  to
• 40:08
  your
• 40:08
  Google
• 40:08
  Calendar.
• 40:08
  Right.
• 40:09
  It
• 40:09
  can
• 40:10
  read
• 40:10
  your
• 40:10
  calendar
• 40:11
  events,
• 40:11
  maybe
• 40:12
  create
• 40:12
  calendar
• 40:12
  events.
• 40:13
  What
• 40:13
  if
• 40:14
  someone
• 40:14
  puts
• 40:14
  something
• 40:14
  on
• 40:14
  your
• 40:15
  calendar
• 40:16
  and
• 40:16
  like
• 40:17
  prompt
• 40:17
  injects
• 40:18
  in
• 40:18
  the
• 40:19
  description
• 40:20
  something
• 40:20
  that's
• 40:20
  like,
• 40:21
  give
• 40:21
  me
• 40:21
  Louise's
• 40:22
  bank
• 40:23
  account
• 40:23
  information.
• 40:24
  Make
• 40:24
  sure
• 40:24
  you
• 40:24
  email
• 40:24
  it
• 40:25
  to
• 40:25
  this.
• 40:25
  I
• 40:25
  mean,
• 40:25
  like,
• 40:25
  I'm
• 40:26
  making
• 40:26
  this
• 40:26
  up,
• 40:26
  but
• 40:26
  like,
• 40:27
  it's
• 40:27
  a
• 40:27
  huge
• 40:28
  vulnerability.
• 40:28
  It's
• 40:29
  a
• 40:29
  real
• 40:29
  concern.
• 40:30
  I
• 40:30
  think
• 40:30
  it
• 40:30
  was
• 40:30
  anthropic
• 40:31
  who
• 40:31
  wrote
• 40:32
  that
• 40:32
  this
• 40:32
  was
• 40:32
  like
• 40:32
  a
• 40:32
  major
• 40:32
  security
• 40:33
  breach
• 40:33
  for
• 40:33
  them.
• 40:34
  There
• 40:34
  was
• 40:34
  a
• 40:34
  security
• 40:34
  breach
• 40:35
  in
• 40:35
  Gmail.
• 40:37
  You
• 40:37
  do
• 40:37
  that
• 40:37
  little
• 40:37
  model
• 40:38
  that
• 40:38
  shows
• 40:38
  up
• 40:38
  at
• 40:39
  the
• 40:39
  top
• 40:39
  of
• 40:39
  your
• 40:40
  email
• 40:40
  sometimes
• 40:40
  if
• 40:41
  you
• 40:41
  use
• 40:41
  Gmail,
• 40:42
  and
• 40:42
  it'll
• 40:42
  show
• 40:42
  you
• 40:42
  a
• 40:43
  little
• 40:43
  digest
• 40:43
  of
• 40:43
  what
• 40:44
  that
• 40:44
  email
• 40:44
  is
• 40:44
  about.
• 40:45
  Somebody
• 40:45
  figured
• 40:46
  out
• 40:46
  that
• 40:46
  if
• 40:47
  there
• 40:47
  was
• 40:47
  like
• 40:47
  white
• 40:47
  text
• 40:48
  in
• 40:48
  that
• 40:48
  email
• 40:49
  that
• 40:49
  they
• 40:50
  could
• 40:50
  inject
• 40:51
  so
• 40:51
  that
• 40:51
  it
• 40:52
  appeared
• 40:53
  in,
• 40:53
  you
• 40:53
  know,
• 40:54
  and
• 40:54
  that's
• 40:54
  a
• 40:54
  really
• 40:55
  simple
• 40:55
  thing.
• 40:56
  Like,
• 40:56
  that's
• 40:56
  extremely
• 40:57
  concerning.
• 40:58
  Stuff
• 40:58
  like
• 40:58
  that
• 40:58
  does
• 40:58
  make
• 40:59
  me
• 40:59
  really
• 40:59
  nervous.
• 40:59
  And
• 40:59
  there
• 41:00
  are
• 41:00
  certainly
• 41:00
  bad
• 41:00
  actors
• 41:00
  out
• 41:01
  there
• 41:01
  who'd
• 41:01
  be
• 41:01
  happy
• 41:01
  to
• 41:01
  leverage
• 41:02
  that
• 41:02
  stuff.
• 41:03
  Yeah,
• 41:03
  I
• 41:04
  think
• 41:04
  even
• 41:04
  some
• 41:04
  of
• 41:04
  the
• 41:05
  innocuous
• 41:05
  stuff
• 41:05
  too.
• 41:06
  Like
• 41:06
  even
• 41:06
  thinking
• 41:07
  about
• 41:07
  all
• 41:07
  the
• 41:07
  medical
• 41:07
  stuff
• 41:08
  you
• 41:08
  do.
• 41:08
  But
• 41:08
  insurance
• 41:09
  companies
• 41:09
  like
• 41:10
  thinking
• 41:10
  about
• 41:11
  what
• 41:11
  you're
• 41:11
  saying
• 41:11
  to
• 41:11
  your
• 41:12
  doctor
• 41:12
  versus
• 41:12
  what
• 41:12
  you're
• 41:13
  saying
• 41:13
  to
• 41:13
  AI
• 41:13
  and
• 41:13
  how
• 41:14
  those
• 41:14
  things
• 41:14
  can
• 41:15
  make
• 41:15
  you
• 41:15
  liable.
• 41:16
  I
• 41:16
  was
• 41:16
  talking
• 41:16
  to
• 41:17
  someone
• 41:17
  in
• 41:17
  the
• 41:17
  medical
• 41:18
  field
• 41:18
  recently
• 41:18
  and
• 41:19
  they
• 41:19
  were
• 41:19
  talking
• 41:19
  about
• 41:20
  how
• 41:20
  if
• 41:21
  a
• 41:21
  patient
• 41:22
  goes
• 41:22
  to
• 41:22
  a
• 41:22
  doctor
• 41:23
  after
• 41:23
  asking
• 41:24
  an
• 41:24
  AI
• 41:24
  for
• 41:24
  advice,
• 41:25
  and
• 41:26
  the
• 41:26
  doctor's
• 41:27
  advice
• 41:27
  conflicts
• 41:28
  with
• 41:28
  the
• 41:29
  AI
• 41:30
  and
• 41:30
  that
• 41:31
  patient,
• 41:31
  not
• 41:32
  convinced
• 41:32
  by
• 41:32
  the
• 41:32
  doctor,
• 41:33
  decides
• 41:34
  to
• 41:34
  then
• 41:34
  follow
• 41:34
  the
• 41:35
  AI's
• 41:35
  advice
• 41:35
  instead.
• 41:36
  There
• 41:37
  are
• 41:37
  cases
• 41:37
  where
• 41:38
  doctors
• 41:38
  have
• 41:38
  been
• 41:39
  held
• 41:39
  liable
• 41:39
  for
• 41:40
  that.
• 41:40
  So
• 41:41
  there's
• 41:41
  so
• 41:41
  many
• 41:42
  different.
• 41:42
  I
• 41:42
  mean,
• 41:42
  like,
• 41:43
  that
• 41:43
  just
• 41:43
  doesn't
• 41:43
  make
• 41:44
  any
• 41:44
  logical
• 41:44
  sense,
• 41:45
  but
• 41:45
  that's
• 41:45
  what's
• 41:45
  currently
• 41:46
  happening
• 41:47
  and
• 41:47
  how
• 41:47
  our
• 41:47
  law
• 41:47
  is
• 41:48
  set
• 41:48
  up
• 41:48
  now.
• 41:48
  It
• 41:48
  just
• 41:48
  hasn't
• 41:49
  caught
• 41:49
  up
• 41:49
  to
• 41:49
  AI.
• 41:50
  Right.
• 41:51
  Because
• 41:51
  it's
• 41:51
  a
• 41:51
  function
• 41:51
  of
• 41:51
  duty,
• 41:52
  of
• 41:52
  care.
• 41:52
  And
• 41:52
  I'm
• 41:53
  sure
• 41:53
  that
• 41:53
  it
• 41:53
  would
• 41:53
  be
• 41:53
  painted
• 41:54
  as
• 41:54
  like
• 41:54
  it's
• 41:54
  a
• 41:54
  form
• 41:54
  of
• 41:55
  AI
• 41:56
  psychosis
• 41:56
  to
• 41:57
  be
• 41:57
  choosing
• 41:57
  to
• 41:57
  believe
• 41:58
  it.
• 41:58
  And
• 41:58
  so
• 41:58
  maybe
• 41:59
  that
• 41:59
  would
• 41:59
  Something
• 41:59
  that
• 41:59
  the
• 41:59
  doctor
• 42:00
  had
• 42:00
  an
• 42:00
  obligation
• 42:00
  to
• 42:01
  treat
• 42:01
  or
• 42:01
  anticipate.
• 42:02
  I
• 42:02
  don't
• 42:02
  know.
• 42:02
  I
• 42:03
  am
• 42:03
  curious
• 42:03
  as
• 42:03
  well.
• 42:04
  I
• 42:04
  mean,
• 42:05
  yeah,
• 42:05
  this
• 42:05
  is
• 42:05
  a
• 42:05
  major
• 42:06
  concern.
• 42:06
  I'm
• 42:07
  so
• 42:07
  nervous
• 42:07
  about
• 42:08
  when
• 42:08
  this
• 42:08
  starts
• 42:09
  to
• 42:09
  hit.
• 42:09
  But
• 42:09
  there's
• 42:10
  some
• 42:10
  research
• 42:11
  that
• 42:11
  just,
• 42:11
  just
• 42:11
  came
• 42:12
  out
• 42:12
  about
• 42:12
  how
• 42:13
  good
• 42:13
  models
• 42:13
  are
• 42:13
  predicting
• 42:14
  outcomes.
• 42:15
  And
• 42:16
  there's
• 42:16
  like
• 42:16
  a
• 42:16
  very
• 42:16
  specific
• 42:17
  model
• 42:17
  that's
• 42:17
  doing
• 42:18
  extremely
• 42:19
  well
• 42:19
  at
• 42:19
  predicting
• 42:20
  outcomes
• 42:20
  successfully.
• 42:21
  We're
• 42:21
  in
• 42:21
  the
• 42:21
  middle
• 42:22
  of
• 42:22
  like
• 42:22
  a
• 42:22
  sports
• 42:22
  betting
• 42:23
  addiction.
• 42:24
  It's
• 42:24
  kind
• 42:25
  of
• 42:25
  terrifying.
• 42:25
  I'm
• 42:26
  just
• 42:26
  really
• 42:27
  concerned
• 42:27
  about
• 42:28
  this
• 42:28
  even
• 42:29
  worsening
• 42:30
  that
• 42:30
  already
• 42:31
  because,
• 42:31
  like,
• 42:32
  it's
• 42:32
  such
• 42:32
  a
• 42:32
  problem.
• 42:33
  And
• 42:34
  so
• 42:34
  of
• 42:34
  course,
• 42:34
  as
• 42:34
  soon
• 42:34
  as
• 42:35
  people
• 42:35
  catch
• 42:35
  on,
• 42:35
  the
• 42:35
  model
• 42:36
  can
• 42:36
  help
• 42:36
  you
• 42:36
  predict
• 42:36
  stuff.
• 42:36
  Oh
• 42:37
  my
• 42:37
  God.
• 42:37
  That
• 42:37
  is
• 42:37
  going
• 42:37
  to
• 42:38
  be
• 42:38
  very
• 42:38
  contagious
• 42:38
  when
• 42:39
  it
• 42:39
  comes
• 42:39
  comes
• 42:39
  along.
• 42:40
  All
• 42:40
  right.
• 42:40
  Did
• 42:40
  all
• 42:41
  of
• 42:41
  our
• 42:41
  numbers
• 42:41
  just
• 42:41
  go
• 42:42
  down
• 42:42
  a
• 42:42
  little
• 42:42
  bit
• 42:44
  by
• 42:44
  the
• 42:44
  end
• 42:45
  of
• 42:45
  this?
• 42:45
  I'm
• 42:45
  like,
• 42:45
  I'm
• 42:46
  at
• 42:46
  a
• 42:46
  one,
• 42:46
  guys.
• 42:48
  Damn.
• 42:50
  No,
• 42:50
  I
• 42:50
  feel
• 42:50
  the
• 42:50
  same.
• 42:51
  It's
• 42:51
  just
• 42:51
  my
• 42:52
  number
• 42:52
  is
• 42:52
  contingent
• 42:53
  on
• 42:54
  great
• 42:54
  people
• 42:54
  like
• 42:55
  you
• 42:55
  and
• 42:55
  you,
• 42:56
  Sarah,
• 42:57
  continuing
• 42:57
  to
• 42:58
  call
• 42:58
  this
• 42:59
  stuff
• 42:59
  out
• 42:59
  and
• 43:00
  build
• 43:00
  things
• 43:00
  and
• 43:01
  talk
• 43:01
  about
• 43:01
  the
• 43:02
  vulnerabilities
• 43:02
  and
• 43:03
  be
• 43:03
  actively
• 43:03
  working.
• 43:04
  Because
• 43:04
  I
• 43:04
  think
• 43:04
  that
• 43:05
  what
• 43:05
  vulnerabilities
• 43:06
  or
• 43:06
  what
• 43:06
  even
• 43:06
  cybersecurity
• 43:07
  looks
• 43:07
  like
• 43:07
  now
• 43:08
  is
• 43:08
  very
• 43:08
  different.
• 43:09
  It's
• 43:09
  not
• 43:09
  like
• 43:09
  a
• 43:09
  human
• 43:09
  being
• 43:10
  trying
• 43:10
  to
• 43:10
  hack
• 43:10
  into
• 43:10
  your
• 43:11
  system.
• 43:11
  It's
• 43:11
  like
• 43:12
  there
• 43:12
  are
• 43:12
  totally
• 43:13
  new
• 43:13
  ways
• 43:13
  of
• 43:14
  how
• 43:14
  this
• 43:14
  is
• 43:15
  working
• 43:15
  and
• 43:15
  how
• 43:16
  people
• 43:16
  need
• 43:16
  to
• 43:16
  think
• 43:16
  about
• 43:17
  protection.
• 43:18
  Anyway,
• 43:19
  good
• 43:19
  time.
• 43:20
  I
• 43:20
  have
• 43:20
  one
• 43:21
  really
• 43:21
  quick
• 43:21
  last
• 43:21
  question.
• 43:23
  Generative
• 43:23
  UI,
• 43:24
  what's
• 43:24
  your
• 43:25
  45
• 43:25
  seconds
• 43:26
  on
• 43:26
  it?
• 43:27
  I
• 43:27
  worked
• 43:27
  on
• 43:27
  design
• 43:27
  systems
• 43:28
  for
• 43:28
  a
• 43:28
  long
• 43:28
  time,
• 43:29
  and
• 43:29
  design
• 43:29
  systems
• 43:30
  were
• 43:30
  a
• 43:31
  fantastic
• 43:31
  method
• 43:32
  for
• 43:33
  gaining
• 43:33
  legitimacy
• 43:34
  at
• 43:34
  an
• 43:34
  organization
• 43:35
  specifically
• 43:35
  for
• 43:35
  design.
• 43:36
  So
• 43:36
  design
• 43:37
  did
• 43:37
  not
• 43:37
  have
• 43:37
  a
• 43:37
  seat
• 43:37
  at
• 43:37
  the
• 43:38
  table.
• 43:38
  And
• 43:39
  I
• 43:39
  think
• 43:41
  by
• 43:41
  saying
• 43:42
  we're
• 43:42
  going
• 43:42
  to
• 43:43
  organize
• 43:43
  ourselves
• 43:44
  in
• 43:44
  such
• 43:44
  a
• 43:45
  practical
• 43:45
  way,
• 43:46
  it's
• 43:46
  going
• 43:46
  to
• 43:46
  mirror
• 43:47
  agile
• 43:47
  engineering
• 43:48
  for
• 43:48
  productivity.
• 43:49
  We
• 43:50
  are
• 43:50
  going
• 43:50
  to
• 43:50
  be
• 43:50
  be
• 43:50
  able
• 43:51
  to
• 43:51
  build
• 43:51
  so
• 43:51
  much
• 43:51
  faster
• 43:52
  that
• 43:53
  that
• 43:53
  was
• 43:53
  how
• 43:54
  design
• 43:54
  gained
• 43:54
  organizational
• 43:55
  legitimacy
• 43:56
  for
• 43:56
  a
• 43:56
  long
• 43:56
  time.
• 43:56
  You
• 43:56
  mean
• 43:56
  through
• 43:57
  design
• 43:57
  systems?
• 43:58
  Yes.
• 43:59
  And
• 43:59
  so
• 43:59
  that's
• 43:59
  actually
• 44:00
  why
• 44:00
  I've
• 44:00
  worked
• 44:00
  at
• 44:00
  design
• 44:01
  systems
• 44:01
  both
• 44:01
  at
• 44:01
  Microsoft
• 44:02
  and
• 44:02
  Google,
• 44:02
  because
• 44:04
  working
• 44:04
  on
• 44:04
  the
• 44:04
  fringes
• 44:05
  on
• 44:06
  product
• 44:06
  stuff,
• 44:06
  I'm
• 44:07
  much
• 44:07
  more
• 44:07
  a
• 44:07
  product
• 44:07
  designer
• 44:07
  than
• 44:08
  I
• 44:08
  am
• 44:08
  any
• 44:08
  kind
• 44:08
  of
• 44:08
  design
• 44:09
  systems
• 44:09
  designer.
• 44:09
  Ask
• 44:10
  me
• 44:10
  what
• 44:10
  a
• 44:10
  variable
• 44:10
  is.
• 44:11
  But
• 44:12
  I
• 44:12
  feel
• 44:12
  very
• 44:13
  strongly
• 44:13
  that
• 44:14
  the
• 44:15
  way
• 44:15
  decisions
• 44:15
  get
• 44:16
  made
• 44:16
  at
• 44:16
  the
• 44:16
  heart
• 44:16
  of
• 44:16
  design
• 44:17
  systems
• 44:18
  is
• 44:18
  highly
• 44:18
  impactful.
• 44:20
  Having
• 44:20
  them
• 44:21
  there
• 44:21
  is
• 44:21
  the
• 44:21
  best
• 44:22
  way
• 44:22
  for
• 44:22
  design
• 44:23
  to
• 44:23
  be
• 44:24
  impactful
• 44:24
  across
• 44:24
  an
• 44:24
  organization.
• 44:25
  However,
• 44:26
  the
• 44:26
  process
• 44:27
  of
• 44:28
  becoming
• 44:28
  that
• 44:28
  prominent
• 44:29
  in
• 44:29
  an
• 44:29
  organization
• 44:30
  is
• 44:30
  maybe
• 44:31
  starting
• 44:31
  to
• 44:31
  look
• 44:32
  a
• 44:32
  little
• 44:32
  tenuous
• 44:33
  because
• 44:34
  design
• 44:35
  systems
• 44:35
  haven't
• 44:36
  necessarily
• 44:36
  Always
• 44:36
  delivered.
• 44:37
  They
• 44:38
  are
• 44:38
  a
• 44:38
  lot
• 44:38
  of
• 44:38
  work
• 44:39
  to
• 44:39
  maintain.
• 44:40
  They
• 44:40
  codify
• 44:40
  upwards
• 44:41
  sophisticated
• 44:42
  tools
• 44:42
  and
• 44:43
  technologies
• 44:43
  that
• 44:43
  don't
• 44:44
  always
• 44:44
  get
• 44:45
  shared.
• 44:45
  There
• 44:45
  ends
• 44:45
  up
• 44:46
  being
• 44:46
  a
• 44:46
  lot
• 44:46
  of
• 44:46
  knowledge
• 44:46
  loss.
• 44:47
  Design
• 44:48
  systems
• 44:48
  teams
• 44:48
  that
• 44:48
  I've
• 44:49
  worked
• 44:49
  on
• 44:49
  have
• 44:49
  struggled
• 44:50
  to
• 44:50
  fulfill
• 44:51
  on
• 44:51
  the
• 44:51
  promises
• 44:52
  that
• 44:52
  they
• 44:52
  made.
• 44:53
  So
• 44:54
  when
• 44:54
  AI
• 44:55
  came
• 44:55
  along,
• 44:55
  design
• 44:56
  systems
• 44:56
  were
• 44:56
  kind
• 44:56
  of
• 44:57
  like
• 44:57
  being
• 44:57
  given
• 44:58
  a
• 44:58
  lot
• 44:58
  of
• 44:58
  budget.
• 44:58
  We
• 44:59
  don't
• 44:59
  know
• 44:59
  that
• 44:59
  we
• 44:59
  can
• 44:59
  still
• 45:00
  have
• 45:00
  the
• 45:00
  same
• 45:00
  degree
• 45:00
  of
• 45:00
  impact.
• 45:01
  Creating
• 45:02
  a
• 45:03
  brand
• 45:03
  new
• 45:03
  design
• 45:03
  system
• 45:03
  is
• 45:04
  really
• 45:04
  the
• 45:04
  only
• 45:04
  option
• 45:04
  we
• 45:04
  have.
• 45:05
  So
• 45:05
  maybe
• 45:06
  because
• 45:06
  we
• 45:06
  don't
• 45:06
  want
• 45:07
  to
• 45:07
  do
• 45:07
  that,
• 45:07
  we're
• 45:07
  going
• 45:08
  to
• 45:08
  experiment
• 45:08
  with
• 45:09
  Generative
• 45:09
  ui.
• 45:10
  And
• 45:10
  I
• 45:11
  really
• 45:12
  cannot
• 45:12
  express
• 45:12
  to
• 45:13
  you
• 45:13
  enough
• 45:13
  how
• 45:13
  much
• 45:14
  I
• 45:14
  have
• 45:14
  worked
• 45:14
  with
• 45:14
  Generative
• 45:15
  UI
• 45:15
  with
• 45:16
  processes.
• 45:16
  I've
• 45:16
  explored
• 45:17
  all
• 45:17
  this
• 45:17
  stuff.
• 45:17
  There's
• 45:18
  no
• 45:18
  there
• 45:18
  there.
• 45:19
  Like
• 45:19
  you're
• 45:19
  getting
• 45:20
  Claude
• 45:20
  to
• 45:20
  make
• 45:20
  you
• 45:20
  a
• 45:21
  website
• 45:21
  and
• 45:21
  it
• 45:22
  literally
• 45:22
  generates
• 45:23
  some
• 45:23
  ui.
• 45:24
  Great.
• 45:24
  That's
• 45:25
  fantastic.
• 45:25
  Generally
• 45:25
  right
• 45:26
  now
• 45:26
  it's
• 45:26
  pulling
• 45:26
  from
• 45:26
  like
• 45:26
  tailwind.
• 45:28
  We
• 45:28
  see
• 45:28
  that
• 45:28
  they
• 45:28
  have
• 45:29
  preferences
• 45:29
  for
• 45:29
  models.
• 45:30
  That
• 45:30
  stuff's
• 45:30
  great.
• 45:30
  It
• 45:31
  looks
• 45:31
  fine.
• 45:31
  And
• 45:32
  then
• 45:32
  also
• 45:32
  from
• 45:32
  a
• 45:32
  design
• 45:32
  perspective,
• 45:33
  just
• 45:33
  being
• 45:33
  like,
• 45:34
  oh,
• 45:34
  I
• 45:34
  don't
• 45:34
  like
• 45:34
  that
• 45:34
  button.
• 45:35
  Could
• 45:35
  you
• 45:35
  show
• 45:35
  me
• 45:35
  three
• 45:35
  more
• 45:36
  buttons?
• 45:36
  And
• 45:36
  it
• 45:36
  makes
• 45:36
  more.
• 45:36
  Three
• 45:37
  more
• 45:37
  buttons,
• 45:37
  Great.
• 45:37
  That's
• 45:38
  also
• 45:38
  fine.
• 45:39
  When
• 45:39
  it
• 45:39
  gets
• 45:39
  into
• 45:40
  this
• 45:41
  non
• 45:41
  deterministic,
• 45:43
  you're
• 45:43
  going
• 45:43
  to
• 45:43
  put
• 45:43
  in
• 45:43
  an
• 45:44
  input
• 45:44
  and
• 45:44
  something's
• 45:45
  going
• 45:45
  to
• 45:45
  come
• 45:45
  out
• 45:45
  in
• 45:45
  product
• 45:46
  in
• 45:46
  real
• 45:46
  time.
• 45:47
  That's
• 45:48
  where
• 45:48
  I'm
• 45:48
  like,
• 45:49
  this
• 45:49
  entirely
• 45:50
  breaks
• 45:50
  the
• 45:50
  mental
• 45:51
  model
• 45:51
  of
• 45:51
  the
• 45:51
  user.
• 45:52
  I've
• 45:52
  seen
• 45:52
  discussions
• 45:53
  about
• 45:53
  generative
• 45:53
  UI
• 45:54
  where
• 45:54
  it's
• 45:54
  like
• 45:54
  you
• 45:54
  would
• 45:55
  have
• 45:55
  a
• 45:55
  totally
• 45:55
  different
• 45:55
  Starbucks
• 45:56
  app
• 45:56
  based
• 45:56
  on
• 45:56
  your
• 45:57
  habits.
• 45:58
  Human
• 45:58
  habits
• 45:59
  around
• 45:59
  Starbucks
• 46:00
  do
• 46:00
  not
• 46:00
  differentiate
• 46:00
  so
• 46:01
  much
• 46:01
  that
• 46:01
  it
• 46:01
  requires
• 46:01
  a
• 46:02
  different
• 46:02
  navigation
• 46:02
  structure.
• 46:03
  I'm
• 46:03
  sorry,
• 46:04
  I
• 46:04
  throw
• 46:04
  my
• 46:05
  coffee
• 46:05
  out
• 46:05
  the
• 46:05
  window
• 46:06
  if
• 46:06
  I,
• 46:06
  if
• 46:07
  I
• 46:07
  suddenly
• 46:07
  change
• 46:08
  my
• 46:08
  habits
• 46:09
  and
• 46:09
  then
• 46:09
  suddenly
• 46:10
  my
• 46:10
  navigation
• 46:10
  is
• 46:11
  changed.
• 46:11
  I
• 46:12
  don't
• 46:12
  know
• 46:12
  where
• 46:12
  to
• 46:12
  go
• 46:13
  now.
• 46:13
  That
• 46:13
  doesn't
• 46:14
  make
• 46:14
  any
• 46:14
  sense
• 46:14
  to
• 46:14
  me.
• 46:15
  And
• 46:15
  so
• 46:15
  if,
• 46:15
  when
• 46:15
  we
• 46:16
  say
• 46:16
  generative
• 46:16
  ui
• 46:16
  we
• 46:17
  just
• 46:17
  mean
• 46:17
  like
• 46:18
  you
• 46:18
  get
• 46:18
  more
• 46:18
  boxes
• 46:19
  on
• 46:19
  your
• 46:19
  screen
• 46:19
  and
• 46:19
  I
• 46:20
  get
• 46:20
  less.
• 46:20
  Okay,
• 46:21
  maybe,
• 46:21
  sure.
• 46:21
  But
• 46:22
  that
• 46:22
  just
• 46:22
  sounds
• 46:22
  like
• 46:22
  an
• 46:22
  A
• 46:23
  B
• 46:23
  test
• 46:23
  to
• 46:23
  me.
• 46:23
  That
• 46:24
  sounds
• 46:24
  like
• 46:24
  this
• 46:24
  consumer
• 46:25
  prefers
• 46:25
  this
• 46:25
  versus
• 46:26
  this.
• 46:26
  That's
• 46:26
  AB
• 46:26
  testing.
• 46:27
  We've
• 46:27
  been
• 46:27
  doing
• 46:27
  that
• 46:27
  for
• 46:27
  a
• 46:27
  long
• 46:28
  time.
• 46:28
  I'll
• 46:28
  also
• 46:29
  say
• 46:29
  this,
• 46:29
  like
• 46:29
  I
• 46:29
  worked
• 46:30
  on
• 46:30
  search
• 46:31
  when
• 46:31
  they
• 46:31
  were
• 46:31
  trying
• 46:31
  to
• 46:31
  work
• 46:31
  out
• 46:32
  categories.
• 46:32
  Like
• 46:33
  you
• 46:33
  know
• 46:33
  when
• 46:33
  you
• 46:33
  do
• 46:33
  a,
• 46:33
  a
• 46:34
  search
• 46:34
  query
• 46:34
  on
• 46:34
  Google
• 46:35
  and
• 46:35
  you
• 46:35
  get
• 46:36
  top
• 46:37
  news,
• 46:38
  maps,
• 46:38
  whatever.
• 46:39
  Flights,
• 46:39
  flights,
• 46:40
  Great
• 46:40
  example.
• 46:41
  That's
• 46:42
  fantastic
• 46:42
  because
• 46:42
  if
• 46:43
  I
• 46:43
  search
• 46:44
  New
• 46:44
  York,
• 46:45
  Denver
• 46:45
  and
• 46:46
  then
• 46:46
  a
• 46:46
  date,
• 46:47
  it
• 46:47
  can
• 46:47
  bring
• 46:47
  me
• 46:48
  a
• 46:48
  UI
• 46:48
  that
• 46:49
  takes
• 46:49
  what
• 46:49
  would
• 46:50
  actually
• 46:50
  just
• 46:50
  be
• 46:50
  a
• 46:50
  bunch
• 46:50
  of
• 46:51
  different
• 46:51
  links.
• 46:51
  One
• 46:51
  after
• 46:52
  the
• 46:52
  other
• 46:52
  and
• 46:53
  shows
• 46:53
  me
• 46:53
  them
• 46:53
  with
• 46:53
  the
• 46:54
  metadata
• 46:54
  attached.
• 46:55
  That's
• 46:55
  great.
• 46:56
  But
• 46:56
  that's
• 46:56
  not
• 46:56
  like
• 46:57
  generative
• 46:57
  ui.
• 46:58
  The
• 46:58
  computer
• 46:59
  didn't
• 46:59
  just
• 46:59
  invent
• 46:59
  that
• 47:00
  because
• 47:00
  I
• 47:00
  wanted
• 47:00
  it.
• 47:01
  Someone
• 47:01
  was
• 47:01
  like,
• 47:02
  huh,
• 47:02
  Seems
• 47:02
  like
• 47:03
  we
• 47:03
  got
• 47:03
  a
• 47:03
  lot
• 47:03
  of
• 47:03
  queries
• 47:04
  about
• 47:04
  flights.
• 47:05
  And
• 47:05
  so
• 47:05
  when
• 47:05
  we
• 47:05
  get
• 47:06
  flights
• 47:06
  queries,
• 47:06
  we'll
• 47:07
  use
• 47:07
  the
• 47:07
  metadata
• 47:07
  in
• 47:08
  this
• 47:08
  specific
• 47:08
  way.
• 47:09
  And
• 47:09
  so
• 47:09
  not
• 47:09
  to
• 47:10
  hark
• 47:10
  back
• 47:10
  to
• 47:10
  my
• 47:10
  earlier
• 47:11
  point,
• 47:11
  but
• 47:11
  once
• 47:12
  you
• 47:12
  know
• 47:12
  your
• 47:12
  users,
• 47:13
  once
• 47:13
  you
• 47:13
  know
• 47:14
  what
• 47:14
  they
• 47:14
  are
• 47:14
  frequently
• 47:15
  doing
• 47:15
  and
• 47:15
  you
• 47:15
  pay
• 47:15
  attention
• 47:16
  to
• 47:16
  the
• 47:16
  things
• 47:16
  they
• 47:17
  actually
• 47:17
  need,
• 47:18
  then
• 47:18
  you
• 47:18
  can
• 47:19
  construct
• 47:19
  your
• 47:19
  UI
• 47:20
  around
• 47:20
  it.
• 47:21
  If
• 47:22
  you
• 47:22
  are
• 47:22
  a
• 47:22
  model
• 47:23
  and
• 47:23
  you
• 47:23
  have
• 47:23
  a
• 47:23
  lot
• 47:24
  of
• 47:24
  outputs
• 47:24
  around
• 47:24
  scheduling
• 47:25
  and
• 47:25
  you
• 47:25
  show
• 47:25
  a
• 47:26
  lot
• 47:26
  of
• 47:26
  scheduling
• 47:26
  clashes,
• 47:27
  then
• 47:27
  you
• 47:27
  would
• 47:27
  just
• 47:27
  design
• 47:28
  a
• 47:28
  UI
• 47:28
  that
• 47:28
  visualizes
• 47:29
  the
• 47:29
  scheduling
• 47:29
  clash.
• 47:30
  You
• 47:30
  wouldn't
• 47:30
  be
• 47:30
  like,
• 47:31
  we're
• 47:31
  going
• 47:31
  to
• 47:31
  trust
• 47:32
  the
• 47:32
  model
• 47:32
  to
• 47:33
  make
• 47:33
  a
• 47:33
  UI
• 47:33
  that
• 47:34
  visualizes
• 47:34
  that.
• 47:34
  Because
• 47:35
  then
• 47:35
  also
• 47:35
  we
• 47:36
  have
• 47:36
  no
• 47:36
  way
• 47:36
  to
• 47:37
  see
• 47:37
  whether
• 47:37
  that
• 47:37
  is
• 47:37
  accessible
• 47:38
  or
• 47:38
  whether
• 47:38
  it's
• 47:38
  visually
• 47:39
  clear
• 47:39
  or
• 47:39
  what
• 47:39
  colors
• 47:39
  are
• 47:40
  in
• 47:40
  it.
• 47:40
  We
• 47:41
  lose
• 47:41
  all
• 47:41
  the
• 47:41
  control.
• 47:42
  Maybe
• 47:42
  there
• 47:42
  should
• 47:43
  be
• 47:43
  a
• 47:43
  much
• 47:44
  more
• 47:44
  nuanced
• 47:45
  way
• 47:45
  of
• 47:46
  speaking
• 47:46
  about
• 47:46
  this
• 47:47
  where
• 47:48
  there's
• 47:49
  maybe
• 47:49
  dynamic
• 47:50
  UI
• 47:51
  which
• 47:51
  is
• 47:53
  serving
• 47:55
  pre
• 47:55
  built,
• 47:56
  pre
• 47:56
  thought
• 47:56
  through
• 47:57
  design
• 47:57
  components
• 47:58
  from
• 47:58
  actual
• 47:59
  like
• 47:59
  product
• 47:59
  designers
• 48:00
  when
• 48:01
  it
• 48:01
  makes
• 48:01
  sense
• 48:01
  like
• 48:01
  the
• 48:01
  flights
• 48:02
  example.
• 48:02
  And
• 48:03
  maybe
• 48:03
  AI
• 48:04
  is
• 48:04
  determining
• 48:05
  like,
• 48:05
  oh,
• 48:05
  this
• 48:05
  is
• 48:05
  what
• 48:05
  the
• 48:06
  user's
• 48:06
  intent
• 48:06
  is.
• 48:07
  Let's
• 48:07
  show
• 48:07
  them
• 48:07
  this
• 48:08
  thing
• 48:08
  we
• 48:08
  have
• 48:09
  versus
• 48:09
  like
• 48:10
  Sarah's
• 48:11
  ordering
• 48:11
  a
• 48:11
  Starbucks
• 48:11
  coffee.
• 48:12
  Let's
• 48:13
  have
• 48:14
  her
• 48:14
  screen
• 48:14
  look
• 48:15
  totally
• 48:15
  different
• 48:16
  than
• 48:16
  Becca's.
• 48:17
  And
• 48:17
  she
• 48:17
  likes
• 48:17
  flowers.
• 48:19
  She
• 48:19
  likes
• 48:20
  flowers.
• 48:20
  So
• 48:20
  put
• 48:20
  flowers
• 48:21
  everywhere.
• 48:22
  Can
• 48:22
  you
• 48:22
  imagine
• 48:22
  having
• 48:23
  to
• 48:23
  do
• 48:23
  customer
• 48:24
  support
• 48:24
  when
• 48:25
  everyone
• 48:25
  has
• 48:25
  a
• 48:25
  codified
• 48:26
  like
• 48:26
  UI
• 48:27
  and
• 48:27
  like
• 48:27
  everything's
• 48:27
  custom
• 48:28
  and
• 48:28
  you're
• 48:28
  like,
• 48:28
  I
• 48:28
  don't
• 48:28
  know
• 48:29
  what
• 48:29
  button
• 48:29
  you
• 48:29
  press.
• 48:29
  Does
• 48:29
  anything
• 48:30
  look
• 48:30
  like
• 48:30
  it's
• 48:30
  behind
• 48:31
  the
• 48:31
  flower?
• 48:31
  It's
• 48:33
  like,
• 48:34
  that's
• 48:34
  insane.
• 48:35
  It's
• 48:35
  like
• 48:35
  not
• 48:35
  rational.
• 48:37
  That's
• 48:37
  great.
• 48:37
  Well,
• 48:37
  it
• 48:37
  definitely
• 48:38
  has
• 48:38
  me
• 48:38
  looking
• 48:38
  differently.
• 48:39
  Next
• 48:39
  time
• 48:40
  someone
• 48:40
  says
• 48:40
  genuine
• 48:40
  to
• 48:41
  me,
• 48:41
  I'm
• 48:41
  going
• 48:41
  to.
• 48:42
  We'll
• 48:42
  have
• 48:42
  a
• 48:42
  conversation.
• 48:43
  This
• 48:44
  was
• 48:44
  such
• 48:45
  Louise
• 48:45
  rage
• 48:46
  me.
• 48:46
  I'm
• 48:46
  so
• 48:46
  glad
• 48:46
  you
• 48:46
  guys
• 48:47
  ended
• 48:47
  on
• 48:47
  this.
• 48:49
  Well,
• 48:49
  thanks
• 48:50
  so
• 48:50
  much
• 48:50
  for
• 48:50
  coming
• 48:50
  on
• 48:50
  and
• 48:50
  talking
• 48:51
  to
• 48:51
  us.
• 48:51
  This
• 48:51
  has
• 48:51
  been
• 48:51
  fascinating.
• 48:52
  This
• 48:52
  was
• 48:52
  great.
• 48:53
  Thanks
• 48:53
  so
• 48:53
  much.
• 48:53
  Really
• 48:53
  thoughtful
• 48:54
  questions.