1
00:00:00,000 --> 00:00:03,737
I want to try and focus
on scenarios where A.I. goes rogue.

2
00:00:03,820 --> 00:00:05,957
I should have had a picture
of a Terminator robot

3
00:00:06,040 --> 00:00:07,897
to scare people as much as possible.

4
00:00:07,980 --> 00:00:10,990
I think the internet...
The internet does that for us.

5
00:00:16,800 --> 00:00:21,697
Are the lords of artificial intelligence
on the side of the human race?

6
00:00:21,780 --> 00:00:24,257
"My prediction is there’ll
be more robots than people."

7
00:00:24,340 --> 00:00:28,217
"The physical and the digital worlds
should really be fully blended."

8
00:00:28,300 --> 00:00:31,417
"I don’t think the world has really
had the humanoid robots moment yet.

9
00:00:31,500 --> 00:00:33,077
It’s going to feel very sci-fi."

10
00:00:33,160 --> 00:00:36,097
That’s the core question I
had for this week’s guest.

11
00:00:36,180 --> 00:00:39,237
He’s the head of Anthropic,
one of the fastest growing A.I.

12
00:00:39,320 --> 00:00:45,017
companies. Anthropic is estimated
to be worth nearly $350 billion.

13
00:00:45,100 --> 00:00:47,737
It’s been win after win
for Anthropic’s Claude code.

14
00:00:47,820 --> 00:00:50,797
He’s a utopian of sorts,
when it comes to the potential

15
00:00:50,880 --> 00:00:54,497
effects of the technology
that he’s unleashing on the world.

16
00:00:54,580 --> 00:00:56,417
"You know, will help us cure cancer.

17
00:00:56,500 --> 00:00:59,137
It may help us to eradicate
tropical diseases.

18
00:00:59,220 --> 00:01:02,187
It will help us understand,
understand the universe."

19
00:01:02,270 --> 00:01:07,147
But he also sees grave dangers ahead
and massive disruption, no matter what.

20
00:01:07,230 --> 00:01:10,167
"This is happening so
fast and is such a crisis,

21
00:01:10,250 --> 00:01:13,087
we should be devoting
almost all of our effort

22
00:01:13,170 --> 00:01:15,247
to thinking about how
to get through this."

23
00:01:15,330 --> 00:01:18,497
Dario Amodei, welcome
to Interesting Times.

24
00:01:18,580 --> 00:01:19,747
Thank you for having me, Ross.

25
00:01:19,830 --> 00:01:24,207
Thank you for being here.
So you are rather unusually, maybe

26
00:01:24,290 --> 00:01:27,047
for a tech C.E.O., an essayist.

27
00:01:27,130 --> 00:01:30,807
You have written two long,
very interesting essays

28
00:01:30,890 --> 00:01:35,007
about the promise and the peril
of artificial intelligence.

29
00:01:35,090 --> 00:01:37,707
And we’re going to talk about
the perils in this conversation.

30
00:01:37,790 --> 00:01:41,287
But I thought it would be
good to start with the promise

31
00:01:41,370 --> 00:01:44,107
and with the optimistic vision.

32
00:01:44,190 --> 00:01:46,887
Indeed, I would say
the utopian vision that you laid

33
00:01:46,970 --> 00:01:51,087
out a couple of years ago
in an essay entitled, "Machines

34
00:01:51,170 --> 00:01:55,807
of Loving Grace," which we’ll come
back to that title, I think, at the end.

35
00:01:55,890 --> 00:01:59,967
But, I think a lot
of people encounter A.I. news

36
00:02:00,050 --> 00:02:04,767
through headlines predicting
a bloodbath for white collar

37
00:02:04,850 --> 00:02:06,147
jobs, these kinds of things.

38
00:02:06,230 --> 00:02:08,967
Sometimes your own quotes —
Have used my own quotes — Yes.

39
00:02:09,050 --> 00:02:10,227
Have encouraged these things.

40
00:02:10,310 --> 00:02:13,767
And I think there’s a commonplace
sense of, "What is A.I.

41
00:02:13,850 --> 00:02:18,347
for?" that people have.
So why don’t you answer that question,

42
00:02:18,430 --> 00:02:21,047
to start out — if everything goes

43
00:02:21,130 --> 00:02:25,687
amazingly in the next five
or 10 years, what is A.I. for?

44
00:02:25,770 --> 00:02:29,007
Yeah, so I think for a little background

45
00:02:29,090 --> 00:02:33,047
before I worked in before
I worked in tech at all,

46
00:02:33,130 --> 00:02:39,207
I was a biologist. I first worked
on computational neuroscience, and then

47
00:02:39,290 --> 00:02:43,807
I worked at Stanford Medical School
on finding protein biomarkers for cancer

48
00:02:43,890 --> 00:02:47,927
on trying to improve
diagnostics and curing cancer.

49
00:02:48,010 --> 00:02:50,187
And one of the observations
that I most had

50
00:02:50,270 --> 00:02:56,027
when I worked in that field
was the incredible complexity of it.

51
00:02:56,110 --> 00:02:59,757
Each protein has a level
localized within each cell.

52
00:02:59,840 --> 00:03:02,297
It’s not enough to measure
the level within the body

53
00:03:02,380 --> 00:03:03,637
or the level within each cell.

54
00:03:03,720 --> 00:03:06,097
You have to measure
the level in a particular part

55
00:03:06,180 --> 00:03:10,497
of the cell and the other proteins that
it’s interacting with or complexing with.

56
00:03:10,580 --> 00:03:14,017
And I had the sense of, "Man,
this is too complicated for humans."

57
00:03:14,100 --> 00:03:18,497
We’re making progress on,
all these problems of biology

58
00:03:18,580 --> 00:03:21,497
and medicine, but we’re making
progress relatively slowly.

59
00:03:21,580 --> 00:03:24,297
And so what drew me to the field of A.I.

60
00:03:24,380 --> 00:03:28,037
was this idea of — that you know,
could we make progress more quickly?

61
00:03:28,120 --> 00:03:31,297
Look, we’ve been trying to apply A.I.
and machine learning

62
00:03:31,380 --> 00:03:34,457
techniques to biology for a long time.

63
00:03:34,540 --> 00:03:38,597
Typically they’ve been for analyzing data,
but as A.I. gets really powerful,

64
00:03:38,680 --> 00:03:40,847
I think we should actually
think about it differently.

65
00:03:40,930 --> 00:03:45,997
We should think of A.I. as doing
the job of the biologist, right?

66
00:03:46,080 --> 00:03:48,457
Doing the whole thing from end to end.

67
00:03:48,540 --> 00:03:52,717
And part of that involves
proposing experiments,

68
00:03:52,800 --> 00:03:54,497
coming up with new techniques.

69
00:03:54,580 --> 00:03:56,947
I have this section where I say, "Look,

70
00:03:57,030 --> 00:03:59,547
a lot of the progress
in biology has been driven

71
00:03:59,630 --> 00:04:03,147
by this relatively small
number of insights that lets

72
00:04:03,230 --> 00:04:07,507
us measure or get at or intervene
in the stuff that’s really small.

73
00:04:07,590 --> 00:04:09,467
You look at a lot of these techniques.

74
00:04:09,550 --> 00:04:13,787
They’re invented very much
as a matter of serendipity.

75
00:04:13,870 --> 00:04:17,467
CRISPR, which is one of these
gene editing technologies

76
00:04:17,550 --> 00:04:20,427
was invented because
someone went to a lecture

77
00:04:20,510 --> 00:04:25,027
on the bacterial immune system
and connected that to the work

78
00:04:25,110 --> 00:04:27,347
they were doing on gene therapy.

79
00:04:27,430 --> 00:04:29,767
And that connection could
have been made 30 years ago.

80
00:04:29,850 --> 00:04:33,387
And so the thought is —could A.I.
accelerate all of this

81
00:04:33,470 --> 00:04:35,687
and could we really cure cancer?

82
00:04:35,770 --> 00:04:38,487
Could we really cure Alzheimer’s disease?

83
00:04:38,570 --> 00:04:41,127
Could we really cure, heart disease?

84
00:04:41,210 --> 00:04:44,467
And more subtly, some of the
more psychological afflictions

85
00:04:44,550 --> 00:04:48,267
that people have — depression, bipolar —

86
00:04:48,350 --> 00:04:50,707
could we do something about these?
To the extent that

87
00:04:50,790 --> 00:04:55,237
they’re biologically based, which
I think they are, at least in part.

88
00:04:55,320 --> 00:04:58,577
So, I go through this argument here,
"Well, how fast could it go?"

89
00:04:58,660 --> 00:05:02,537
If we have these intelligences out
there who could do just about anything?

90
00:05:02,620 --> 00:05:05,957
And I want to pause you there
because one of the interesting

91
00:05:06,040 --> 00:05:08,217
things about your framing in that essay,

92
00:05:08,300 --> 00:05:12,197
and you returned to it,
is that these intelligences don’t

93
00:05:12,280 --> 00:05:15,797
have to be right,
the kind of maximal godlike

94
00:05:15,880 --> 00:05:18,777
superintelligence that
comes up in A.I. debates.

95
00:05:18,860 --> 00:05:22,677
You’re basically saying,
if we can achieve a strong

96
00:05:22,760 --> 00:05:26,337
intelligence at the level
of peak human performance —

97
00:05:26,420 --> 00:05:31,117
peak human performance, yes —
and then multiply it, right, to what?

98
00:05:31,200 --> 00:05:32,737
Your phrase is, "A country of geniuses."

99
00:05:32,820 --> 00:05:34,757
A country — have 100 million of them.

100
00:05:34,840 --> 00:05:37,177
Right. A hundred million —
Each, a little trained,

101
00:05:37,260 --> 00:05:40,377
a little different,
or trying a different problem.

102
00:05:40,460 --> 00:05:44,857
There’s benefit in diversification
and trying things a little differently.

103
00:05:44,940 --> 00:05:48,337
But yes. So you don’t have
to have the full machine.

104
00:05:48,420 --> 00:05:50,707
God you just need to have
100 million geniuses.

105
00:05:50,790 --> 00:05:52,377
You don’t have to have the full machine.

106
00:05:52,460 --> 00:05:53,957
God and indeed, there are places

107
00:05:54,040 --> 00:05:58,797
where I cast doubt on whether
the machine God would

108
00:05:58,880 --> 00:06:02,857
be that much more effective at these
things than the 100 million geniuses.

109
00:06:02,940 --> 00:06:08,297
I have this concept called the diminishing
returns to intelligence, right.

110
00:06:08,380 --> 00:06:13,187
Which is there’s economists talk about the
marginal productivity of land and labor.

111
00:06:13,270 --> 00:06:16,297
We’ve never thought about the marginal
productivity of intelligence.

112
00:06:16,380 --> 00:06:18,877
But if I look at some
of these problems in biology

113
00:06:18,960 --> 00:06:22,717
at some level, you just have to interact
with the world at some level, you just

114
00:06:22,800 --> 00:06:24,537
have to try things at some level.

115
00:06:24,620 --> 00:06:29,117
You just have to comply with the laws
or change the laws on getting medicines

116
00:06:29,200 --> 00:06:30,457
through the regulatory system.

117
00:06:30,540 --> 00:06:35,937
So there’s a finite rate
at which these changes can happen.

118
00:06:36,020 --> 00:06:38,227
Now there are some domains
like if you’re playing chess

119
00:06:38,310 --> 00:06:41,977
or Go where the intelligence
ceiling is extremely high.

120
00:06:42,060 --> 00:06:44,617
But I think the real world
has a lot of limiters.

121
00:06:44,700 --> 00:06:46,897
So maybe you can go
above the genius level.

122
00:06:46,980 --> 00:06:52,167
But, sometimes I think all this discussion
of could you use a moon of computation

123
00:06:52,250 --> 00:06:56,647
to make an AI God are there
a little bit sensationalistic

124
00:06:56,730 --> 00:06:59,287
and besides the point, even as I think

125
00:06:59,370 --> 00:07:02,107
this will be the biggest thing
that ever happened to humanity.

126
00:07:02,190 --> 00:07:05,827
And so you have so keeping it concrete,

127
00:07:05,910 --> 00:07:09,487
you have a world where
there’s just an end to cancer

128
00:07:09,570 --> 00:07:12,307
as a serious threat to human
life, an end to heart disease,

129
00:07:12,390 --> 00:07:15,327
an end to most of the
illnesses that we experience

130
00:07:15,410 --> 00:07:19,127
that kill us, possible
life extension beyond that.

131
00:07:19,210 --> 00:07:21,567
So that’s health.
That’s a pretty positive vision.

132
00:07:21,650 --> 00:07:23,587
Then talk about economics and wealth.

133
00:07:23,670 --> 00:07:28,227
What happens in the 5 to 10
year A.I. takeoff to wealth.

134
00:07:28,310 --> 00:07:31,047
So again, let’s keep it
on the positive side because there

135
00:07:31,130 --> 00:07:33,807
will be plenty we’ll
get to the negative side.

136
00:07:33,890 --> 00:07:36,467
But we’re already working
with pharma companies.

137
00:07:36,550 --> 00:07:40,787
We’re already working with
financial industry companies.

138
00:07:40,870 --> 00:07:44,407
We’re already working with
folks who do manufacturing

139
00:07:44,490 --> 00:07:47,867
or of course, I think especially known
for coding and software engineering.

140
00:07:47,950 --> 00:07:50,657
So just the raw productivity, the ability

141
00:07:50,740 --> 00:07:54,597
to make stuff and get stuff
done that is very powerful.

142
00:07:54,680 --> 00:07:59,197
And we see our company’s revenue
growing going up 10x a year.

143
00:07:59,280 --> 00:08:04,277
And, we suspect the wider industry
looks something similar to that.

144
00:08:04,360 --> 00:08:05,837
If the technology keeps improving,

145
00:08:05,920 --> 00:08:09,177
it doesn’t take that many more
10 X’s until suddenly you’re

146
00:08:09,260 --> 00:08:12,557
saying, oh, if you’re
adding across the industry $1

147
00:08:12,640 --> 00:08:16,457
trillion of revenue a year,
the US GDP is 20 or 30

148
00:08:16,540 --> 00:08:17,917
trillion, I can’t remember exactly.

149
00:08:18,000 --> 00:08:21,437
So you must be increasing
the GDP growth by a few percent.

150
00:08:21,520 --> 00:08:27,897
So I can see a world where A.I.
brings the developed world GDP

151
00:08:27,980 --> 00:08:32,337
growth to something like
percent or 15 percent 5, 10, 15

152
00:08:32,420 --> 00:08:36,157
mean, there’s no science
of calculating these numbers.

153
00:08:36,240 --> 00:08:37,647
It’s totally unprecedented thing.

154
00:08:37,730 --> 00:08:40,857
But it could bring it to numbers
that are outside the distribution of what

155
00:08:40,940 --> 00:08:44,397
we saw before. And again, I think
this will lead to a weird world.

156
00:08:44,480 --> 00:08:47,877
We have all these debates
about the deficit is growing.

157
00:08:47,960 --> 00:08:50,767
If you have that much in GDP growth,

158
00:08:50,850 --> 00:08:53,137
you’re going to have that much
in tax receipts and you’re

159
00:08:53,220 --> 00:08:56,117
going to balance the budget
without meaning to.

160
00:08:56,200 --> 00:08:58,457
But one of the things I’ve
been thinking about lately is

161
00:08:58,540 --> 00:09:02,577
I think one of the assumptions
of just our economic

162
00:09:02,660 --> 00:09:05,477
and political debates
is that growth is hard to achieve.

163
00:09:05,560 --> 00:09:11,057
It’s this unicorn. There are all kinds
of ways you can kill the golden goose.

164
00:09:11,140 --> 00:09:14,577
We could enter a world
where growth is really easy.

165
00:09:14,660 --> 00:09:18,797
And it’s the distribution that’s hard
because it’s happening so fast. right.

166
00:09:18,880 --> 00:09:21,557
The pie is being increased. So fast.

167
00:09:21,640 --> 00:09:24,657
So before we get to the
hard problem, one more

168
00:09:24,740 --> 00:09:27,417
note of optimism than
on politics, I think.

169
00:09:27,500 --> 00:09:29,077
And here it’s a little more I mean,

170
00:09:29,160 --> 00:09:32,157
all of this is speculative, but I
think it’s a little more speculative.

171
00:09:32,240 --> 00:09:36,777
You try and make the case that I
could be good for democracy and liberty

172
00:09:36,860 --> 00:09:39,197
around the world, which
is not necessarily intuitive.

173
00:09:39,280 --> 00:09:44,017
A lot of people say,
incredibly powerful technology

174
00:09:44,100 --> 00:09:46,057
in the hands of authoritarian leaders

175
00:09:46,140 --> 00:09:48,367
leads to concentrations
of power and so on.

176
00:09:48,450 --> 00:09:49,827
And I talk about that in the other.

177
00:09:49,910 --> 00:09:53,027
But just briefly,
what is the optimistic case

178
00:09:53,110 --> 00:09:56,487
for why A.I. is good for democracy
Yeah, I mean absolutely.

179
00:09:56,570 --> 00:09:58,987
So yeah, I mean, machines
of loving grace, I kind of like,

180
00:09:59,070 --> 00:10:01,967
I’m just like, let’s dream,
let’s dream about how it could go.

181
00:10:02,050 --> 00:10:05,527
well, I don’t know how likely it is,
but we got to lay out a dream.

182
00:10:05,610 --> 00:10:07,587
Let’s try and make the dream happen.

183
00:10:07,670 --> 00:10:10,667
So I think the positive version,

184
00:10:10,750 --> 00:10:15,727
I admit there that I don’t know that
the technology inherently favors liberty.

185
00:10:15,810 --> 00:10:18,467
I think it inherently
favors curing disease

186
00:10:18,550 --> 00:10:20,847
and it inherently favors economic growth.

187
00:10:20,930 --> 00:10:23,967
But I worry you that it may
not inherently favor liberty.

188
00:10:24,050 --> 00:10:26,887
But what I say there is,
can we make it favor liberty.

189
00:10:26,970 --> 00:10:32,767
Can we make the United States and other
democracies get ahead in this technology.

190
00:10:32,850 --> 00:10:35,667
The United States has been
technologically and militarily

191
00:10:35,750 --> 00:10:39,507
ahead, has meant that we have
throw weight around the world

192
00:10:39,590 --> 00:10:44,687
through and augmented by our
alliances with other democracies.

193
00:10:44,770 --> 00:10:49,117
And we’ve been able to shape
a world that I think is better

194
00:10:49,200 --> 00:10:51,557
than the world would be
if it were shaped by Russia

195
00:10:51,640 --> 00:10:55,017
or by China or by other
authoritarian countries.

196
00:10:55,100 --> 00:11:01,637
And so can we use our lead in A.I. to
shape, to shape liberty around the world.

197
00:11:01,720 --> 00:11:03,477
There’s obviously a lot
of debates about how

198
00:11:03,560 --> 00:11:07,517
interventionist we should be, how we
should how we should wield that power.

199
00:11:07,600 --> 00:11:11,717
But I’ve often worried that
today through social media,

200
00:11:11,800 --> 00:11:15,057
authoritarians are kind
of undermining us, right.

201
00:11:15,140 --> 00:11:18,957
Can we counter that?
Can we win the information war?

202
00:11:19,040 --> 00:11:23,437
Can we prevent authoritarians
from invading countries

203
00:11:23,520 --> 00:11:28,277
like Ukraine or Taiwan by defending them

204
00:11:28,360 --> 00:11:31,197
with the power of A.I.,
with giant, giant swarms

205
00:11:31,280 --> 00:11:33,977
of A.I. powered drones,
which we need to be careful about.

206
00:11:34,060 --> 00:11:37,417
We ourselves need to be careful
about how we build those.

207
00:11:37,500 --> 00:11:41,577
We need to defend liberty
in our own country,

208
00:11:41,660 --> 00:11:44,077
but is there some vision where we kind

209
00:11:44,160 --> 00:11:48,197
of like, re-envision
liberty and individual rights

210
00:11:48,280 --> 00:11:53,977
in the age of A.I. where we need
in some ways to be protected against A.I.

211
00:11:54,060 --> 00:11:57,297
Someone needs to hold the button
on the swarm of drones,

212
00:11:57,380 --> 00:11:59,837
which is something I’m very,
I’m very concerned about

213
00:11:59,920 --> 00:12:01,997
and that oversight doesn’t exist today.

214
00:12:02,080 --> 00:12:05,337
But also think about
the Justice system today, right.

215
00:12:05,420 --> 00:12:08,537
We promise equal justice for all right.

216
00:12:08,620 --> 00:12:11,657
But the truth is, there are
different judges in the world.

217
00:12:11,740 --> 00:12:13,597
The legal system is imperfect.

218
00:12:13,680 --> 00:12:16,477
I don’t think we should
replace judges with A.I.,

219
00:12:16,560 --> 00:12:21,037
but is there some way in which A.I.
can help us to be more

220
00:12:21,120 --> 00:12:22,977
fair, to help us be more uniform.

221
00:12:23,060 --> 00:12:26,037
It’s never been possible before,

222
00:12:26,120 --> 00:12:30,837
but can we somehow use A.I.
to create something that is

223
00:12:30,920 --> 00:12:34,197
fuzzy, but where also you can
give a promise that it’s being

224
00:12:34,280 --> 00:12:36,817
applied in the same way to everyone.

225
00:12:36,900 --> 00:12:38,897
So I don’t know exactly
how it should be done.

226
00:12:38,980 --> 00:12:41,517
And I don’t think we should
replace the Supreme Court with

227
00:12:41,600 --> 00:12:44,827
that’s not what well,
we’re going to talk about that.

228
00:12:44,910 --> 00:12:49,167
But yeah but just this idea that can we

229
00:12:49,250 --> 00:12:52,847
deliver on the promise
of equal opportunity

230
00:12:52,930 --> 00:12:57,447
and equal justice by some
combination of A.I. and humans.

231
00:12:57,530 --> 00:12:59,547
There has to be some way to do that.

232
00:12:59,630 --> 00:13:03,647
And so, just thinking
about reinventing democracy

233
00:13:03,730 --> 00:13:09,327
for the A.I. age and enhancing
liberty instead of reducing it.

234
00:13:09,410 --> 00:13:12,507
Good so that’s good.
That’s a very positive vision.

235
00:13:12,590 --> 00:13:15,507
We’re leading longer
lives, healthier lives.

236
00:13:15,590 --> 00:13:17,247
We’re richer than ever before.

237
00:13:17,330 --> 00:13:20,187
All of this is happening
in a compressed period of time,

238
00:13:20,270 --> 00:13:24,007
where you’re getting a century
of economic growth in 10 years.

239
00:13:24,090 --> 00:13:29,407
And we have increased liberty around
the world and equality at home.

240
00:13:29,490 --> 00:13:34,567
O.K, even in the best case scenario,
it’s incredibly disruptive.

241
00:13:34,650 --> 00:13:38,647
And this is where the lines
that you’ve been quoted

242
00:13:38,730 --> 00:13:43,517
saying, 50 percent of white
collar jobs get disrupted,

243
00:13:43,600 --> 00:13:46,317
or 50 percent of entry level
white collar jobs and so on.

244
00:13:46,400 --> 00:13:49,517
So on a five year time horizon
or a two year time horizon,

245
00:13:49,600 --> 00:13:52,757
whatever time horizon you have, what jobs,

246
00:13:52,840 --> 00:13:55,257
what professions are most
vulnerable to total A.I.

247
00:13:55,340 --> 00:13:59,537
disruption Yeah, it’s hard
to predict these things

248
00:13:59,620 --> 00:14:04,077
because the technology is moving
so fast and moves so unevenly.

249
00:14:04,160 --> 00:14:06,357
So at least a couple
principles for figuring it out.

250
00:14:06,440 --> 00:14:09,537
And then I’ll give my guesses
at what I think will be disrupted.

251
00:14:09,620 --> 00:14:16,057
So one thing is I think the technology
itself and its capabilities will be ahead

252
00:14:16,140 --> 00:14:20,457
of the actual job disruption. Two things
have to happen for jobs to be disrupted

253
00:14:20,540 --> 00:14:22,417
or for productivity to occur,
because sometimes

254
00:14:22,500 --> 00:14:24,817
those sometimes those
two things are linked.

255
00:14:24,900 --> 00:14:28,037
One is the technology has
to be capable of doing it.

256
00:14:28,120 --> 00:14:32,457
And the second is there’s this
messy thing of it actually has

257
00:14:32,540 --> 00:14:36,217
to be applied within a large
bank or a large company

258
00:14:36,300 --> 00:14:39,497
or think about customer
service or something.

259
00:14:39,580 --> 00:14:42,267
In theory, I customer service agents

260
00:14:42,350 --> 00:14:44,767
can be much better than
human customer service agents.

261
00:14:44,850 --> 00:14:49,147
They’re more patient, they know more,
they handle things in a more uniform way.

262
00:14:49,230 --> 00:14:51,827
But the actual logistics
and the actual process

263
00:14:51,910 --> 00:14:57,807
of making that substitution
that takes some time.

264
00:14:57,890 --> 00:15:01,727
So I’m very bullish about
the direction of the A.I. itself.

265
00:15:01,810 --> 00:15:04,227
I think we might have that
country of geniuses in a data

266
00:15:04,310 --> 00:15:10,047
center and one or two years and maybe
it’ll be 5, but it could happen very fast.

267
00:15:10,130 --> 00:15:13,527
But I think the diffusion of the economy
is going to be a little slower.

268
00:15:13,610 --> 00:15:16,887
And that diffusion creates
some unpredictability.

269
00:15:16,970 --> 00:15:21,187
So an example of this is and
we’ve seen within Anthropic

270
00:15:21,270 --> 00:15:25,367
the models writing
code has gone very fast.

271
00:15:25,450 --> 00:15:28,327
I don’t think it’s because the models
are inherently better at code.

272
00:15:28,410 --> 00:15:31,587
I think it’s because
developers are used to fast

273
00:15:31,670 --> 00:15:34,987
technological change
and they adopt things quickly,

274
00:15:35,070 --> 00:15:37,407
and they’re very socially
adjacent to the A.I. world.

275
00:15:37,490 --> 00:15:39,407
So they pay attention
to what’s happening in it.

276
00:15:39,490 --> 00:15:44,307
If you do customer service
or banking or manufacturing,

277
00:15:44,390 --> 00:15:46,107
the distance is a little greater.

278
00:15:46,190 --> 00:15:50,227
And so I think six months ago,
I would have said the first

279
00:15:50,310 --> 00:15:55,307
thing to be disrupted is these
kind of entry level white

280
00:15:55,390 --> 00:16:02,390
collar jobs data entry
or a kind of document review

281
00:16:03,230 --> 00:16:06,187
for law or the things you
would give to a first year

282
00:16:06,270 --> 00:16:09,767
at a financial industry company
where you’re analyzing documents.

283
00:16:09,850 --> 00:16:12,747
And I still think those
are going pretty fast.

284
00:16:12,830 --> 00:16:16,467
But I actually think
software might go even faster

285
00:16:16,550 --> 00:16:19,667
because of the reasons that I
gave where I don’t think that

286
00:16:19,750 --> 00:16:24,187
far from the models being able to do
a lot of it, a lot of it end to end.

287
00:16:24,270 --> 00:16:25,727
And what we’re going to see is first,

288
00:16:25,810 --> 00:16:29,407
the model only does a piece of what
the human software engineer does.

289
00:16:29,490 --> 00:16:31,387
And that increases their productivity.

290
00:16:31,470 --> 00:16:33,867
Then even when the models do
everything that human software

291
00:16:33,950 --> 00:16:36,507
engineers used to do,
the human software engineers

292
00:16:36,590 --> 00:16:41,797
take a step up and they act
as managers and supervise the systems.

293
00:16:41,880 --> 00:16:45,117
And so this is where the term centaur gets

294
00:16:45,200 --> 00:16:51,117
used to describe essentially
like man and horse fused I

295
00:16:51,200 --> 00:16:54,657
and engineer working together
Yeah this is like centaur chess.

296
00:16:54,740 --> 00:16:58,797
So after I think Garry Kasparov was beaten
by deep blue, there was an era that I

297
00:16:58,880 --> 00:17:03,117
think for chess was 15
or 20 years long, where

298
00:17:03,200 --> 00:17:07,957
a human checking the output
of the A.I. playing chess

299
00:17:08,040 --> 00:17:12,317
was able to defeat any human
or any A.I. system alone.

300
00:17:12,400 --> 00:17:15,697
That era at some point ended,
and then it’s just recently.

301
00:17:15,780 --> 00:17:19,237
And then it’s just the machine
Yeah and so my worry

302
00:17:19,320 --> 00:17:20,937
of course, is about that last phase.

303
00:17:21,020 --> 00:17:25,177
So I think we’re already
in our centaur phase for software.

304
00:17:25,260 --> 00:17:28,017
And I think during that centaur phase,

305
00:17:28,100 --> 00:17:30,417
if anything the demand
for software engineers may go up.

306
00:17:30,500 --> 00:17:32,917
But the period may be very brief.

307
00:17:33,000 --> 00:17:37,207
And so, I have this concern
for entry level white collar

308
00:17:37,290 --> 00:17:40,207
work, for software engineering work.

309
00:17:40,290 --> 00:17:43,147
It’s just going to be a big disruption.

310
00:17:43,230 --> 00:17:46,267
I think my worry is just
that it’s all happening so fast.

311
00:17:46,350 --> 00:17:49,607
People talk about previous disruptions.

312
00:17:49,690 --> 00:17:52,467
They say, oh yeah, well,
people used to be farmers.

313
00:17:52,550 --> 00:17:54,587
Then we all worked in industry.

314
00:17:54,670 --> 00:17:59,087
Then we all did knowledge work
Yeah people, people adapted.

315
00:17:59,170 --> 00:18:03,007
That happened over centuries or decades.

316
00:18:03,090 --> 00:18:07,107
This is happening over low
single digit numbers of years.

317
00:18:07,190 --> 00:18:09,347
And maybe that’s my concern here.

318
00:18:09,430 --> 00:18:11,187
How do we get people to adapt fast enough.

319
00:18:11,270 --> 00:18:16,537
But is there also something maybe where
industries like software and professions

320
00:18:16,620 --> 00:18:18,287
like coding that have this kind of comfort

321
00:18:18,370 --> 00:18:22,807
that you describe move faster,
but in other areas people just

322
00:18:22,890 --> 00:18:25,867
want to hang out in the center phase.

323
00:18:25,950 --> 00:18:29,167
So one of the critiques of the
job loss hypothesis will say,

324
00:18:29,250 --> 00:18:32,687
people will say, well, look,
we’ve had A.I. that’s better

325
00:18:32,770 --> 00:18:36,927
at reading a scan then
a radiologist for a while.

326
00:18:37,010 --> 00:18:41,447
But there isn’t job loss. In radiology,
people keep being hired and employed

327
00:18:41,530 --> 00:18:46,007
as radiologists. And doesn’t
that suggest that in the end,

328
00:18:46,090 --> 00:18:48,047
people will want the A.I.
and they’ll want a human

329
00:18:48,130 --> 00:18:50,187
to interpret it because
we’re human beings,

330
00:18:50,270 --> 00:18:52,347
and that will be true across other fields.

331
00:18:52,430 --> 00:18:56,807
Like, how do you see that. That example
is I think it’s going to be pretty

332
00:18:56,890 --> 00:19:01,687
heterogeneous. There may
be areas where a human touch

333
00:19:01,770 --> 00:19:06,807
kind of for its own sake
is particularly important.

334
00:19:06,890 --> 00:19:09,307
Do you think that’s what’s
happening in radiology?

335
00:19:09,390 --> 00:19:12,787
Is that why we haven’t fired all
the radiologists details of radiology.

336
00:19:12,870 --> 00:19:17,047
That might be true. It’s like you go
in and you’re getting cancer diagnosed,

337
00:19:17,130 --> 00:19:21,827
you might not want Hal, from 2001
to be the one to diagnose your cancer.

338
00:19:21,910 --> 00:19:28,727
It’s just maybe not. That’s just
maybe not a human way of doing things.

339
00:19:28,810 --> 00:19:31,987
But there are other areas where you
might think human touch is important.

340
00:19:32,070 --> 00:19:34,337
Like if we look at customer service,

341
00:19:34,420 --> 00:19:36,577
actually customer
service is a terrible job

342
00:19:36,660 --> 00:19:40,997
and the humans who do customer service
are they lose their patience a lot.

343
00:19:41,080 --> 00:19:43,077
And it turns out customers
don’t much like talking

344
00:19:43,160 --> 00:19:46,417
to them because it’s a pretty
robotic interaction, honestly.

345
00:19:46,500 --> 00:19:50,137
And I think the observation
that many people have had

346
00:19:50,220 --> 00:19:54,477
is maybe actually it would be better
for all concerned if this job were done,

347
00:19:54,560 --> 00:20:00,037
were done by machines. So there are
places where a human touch is important.

348
00:20:00,120 --> 00:20:01,657
There are places where it’s not.

349
00:20:01,740 --> 00:20:05,897
And then there are also places
where the job itself doesn’t

350
00:20:05,980 --> 00:20:09,097
really involve it doesn’t
really involve human touch,

351
00:20:09,180 --> 00:20:12,497
assessing the financial
prospects of companies

352
00:20:12,580 --> 00:20:15,637
or writing code or so forth and so on.

353
00:20:15,720 --> 00:20:17,757
Or let’s take the example of the law,

354
00:20:17,840 --> 00:20:23,577
because I think it’s a useful
place that in between applied

355
00:20:23,660 --> 00:20:26,977
science and pure humanities whatever.

356
00:20:27,060 --> 00:20:33,027
So I know a lot of lawyers who have
looked at what I can do already in terms

357
00:20:33,110 --> 00:20:34,667
of legal research and brief writing

358
00:20:34,750 --> 00:20:37,867
and all of these things and have
said, yeah, this is going

359
00:20:37,950 --> 00:20:40,727
to be a bloodbath for the way
our profession works right now.

360
00:20:40,810 --> 00:20:43,247
And you’ve seen this in
the stock market already.

361
00:20:43,330 --> 00:20:48,647
There’s disturbances around companies that
do legal research, some attributed to us,

362
00:20:48,730 --> 00:20:53,207
some attributed to actually cause
we figure out why things happen.

363
00:20:53,290 --> 00:20:57,927
We don’t speculate about the stock
market Yeah very much on this show.

364
00:20:58,010 --> 00:21:03,227
But it seems like in law you can tell
a pretty straightforward story where

365
00:21:03,310 --> 00:21:07,287
law has a kind of system
of training and apprenticeship,

366
00:21:07,370 --> 00:21:11,027
where you have paralegals
and you have junior lawyers who

367
00:21:11,110 --> 00:21:16,087
do behind the scenes research
and development for cases.

368
00:21:16,170 --> 00:21:20,327
And then it has the top tier lawyers who
are actually in the courtroom and so on.

369
00:21:20,410 --> 00:21:23,547
And it just seems really
easy to imagine a world where

370
00:21:23,630 --> 00:21:26,147
all of the apprentice roles go away.

371
00:21:26,230 --> 00:21:28,167
Does that sound right to you.

372
00:21:28,250 --> 00:21:31,907
And you’re just left with
the jobs that involve talking

373
00:21:31,990 --> 00:21:34,547
to clients, talking to juries,
talking to judges.

374
00:21:34,630 --> 00:21:39,107
That is what I had in mind when I
talked about entry level white collar

375
00:21:39,190 --> 00:21:44,247
labor and the bloodbath
headlines of you oh, my God,

376
00:21:44,330 --> 00:21:46,837
are the entry level
pipelines going to dry up.

377
00:21:46,920 --> 00:21:49,947
And then, then how do we get
to the level of the senior partners.

378
00:21:50,030 --> 00:21:51,947
And I think this is actually
a good illustration

379
00:21:52,030 --> 00:21:57,467
because particularly if you froze
the quality of the technology in place,

380
00:21:57,550 --> 00:22:00,307
there are over time ways to adapt to this.

381
00:22:00,390 --> 00:22:04,747
Maybe we just need more lawyers who
spend their time talking to clients.

382
00:22:04,830 --> 00:22:09,387
Maybe lawyers are more
become more like salespeople

383
00:22:09,470 --> 00:22:15,227
or consultants who explain what goes
on in the contracts written by A.I.,

384
00:22:15,310 --> 00:22:16,607
help people come to an agreement.

385
00:22:16,690 --> 00:22:19,067
Maybe you lean into the human side of it.

386
00:22:19,150 --> 00:22:22,287
If we had enough time, that would happen.

387
00:22:22,370 --> 00:22:24,867
But reshaping industries like that

388
00:22:24,950 --> 00:22:27,947
takes years or decades,
whereas these economic forces

389
00:22:28,030 --> 00:22:31,517
driven by A.I.
are going to happen very quickly.

390
00:22:31,600 --> 00:22:33,777
And it’s not just that
they’re happening in law.

391
00:22:33,860 --> 00:22:38,237
The same thing is happening in consulting
and finance and medicine and coding.

392
00:22:38,320 --> 00:22:39,877
And so you have this.

393
00:22:39,960 --> 00:22:45,017
It becomes a macroeconomic phenomenon, not
something just happening in one industry.

394
00:22:45,100 --> 00:22:46,777
And it’s all happening very fast.

395
00:22:46,860 --> 00:22:52,157
And so the norm. I’m just my worry here
is that the normal adaptive mechanisms

396
00:22:52,240 --> 00:22:55,897
will be overwhelmed.
And, I’m not a doomer.

397
00:22:55,980 --> 00:22:59,397
The view is, and we’re
thinking very hard about how

398
00:22:59,480 --> 00:23:03,417
do we strengthen societies adaptive
mechanisms to respond to this.

399
00:23:03,500 --> 00:23:05,417
But I think it’s first
important to say this.

400
00:23:05,500 --> 00:23:06,957
This isn’t just like the other.

401
00:23:07,040 --> 00:23:09,497
This isn’t just like previous disruptions,

402
00:23:09,580 --> 00:23:12,937
but I would then go one step
further though, and say, O.K,

403
00:23:13,020 --> 00:23:15,697
let’s say the law adapts
successfully and it says, all right.

404
00:23:15,780 --> 00:23:18,477
From now on, legal apprenticeship

405
00:23:18,560 --> 00:23:21,657
involves more time in court,
more time with clients.

406
00:23:21,740 --> 00:23:25,617
We’re essentially moving you up
the ladder of responsibility faster.

407
00:23:25,700 --> 00:23:31,367
There are fewer people employed in the law
overall, but the profession settles still.

408
00:23:31,450 --> 00:23:33,727
The reason law would settle right

409
00:23:33,810 --> 00:23:37,767
is that you have all of these
situations in the law where

410
00:23:37,850 --> 00:23:41,567
you are legally required
to have people involved.

411
00:23:41,650 --> 00:23:45,307
You have to have a human
representative in court.

412
00:23:45,390 --> 00:23:48,187
You have to have 12 humans on your jury.

413
00:23:48,270 --> 00:23:49,827
You have to have a human judge.

414
00:23:49,910 --> 00:23:52,647
And you already mentioned
the idea that there are various

415
00:23:52,730 --> 00:23:55,467
ways in which I might be let’s say,

416
00:23:55,550 --> 00:23:59,307
very helpful at clarifying what
kind of decision should be reached.

417
00:23:59,390 --> 00:24:01,967
But that too seems like a scenario

418
00:24:02,050 --> 00:24:06,147
where what preserves human
agency is law and custom.

419
00:24:06,230 --> 00:24:07,747
Like you could replace the judge.

420
00:24:07,830 --> 00:24:11,767
Yes, with Claude version 17.9.

421
00:24:11,850 --> 00:24:16,967
But you choose not to because
the law requires there to be a human.

422
00:24:17,050 --> 00:24:20,487
That just seems a very
interesting way of thinking

423
00:24:20,570 --> 00:24:25,737
about the future, where it’s volitional,
whether we stay in charge Yeah,

424
00:24:25,820 --> 00:24:29,057
and I would argue that in many cases,
we do want to stay in charge.

425
00:24:29,140 --> 00:24:31,837
That’s a choice we want to make,

426
00:24:31,920 --> 00:24:34,297
even in some cases when we
think the humans on average

427
00:24:34,380 --> 00:24:37,437
make kind of worse decisions.

428
00:24:37,520 --> 00:24:41,357
I mean, again, life critical,
safety critical cases.

429
00:24:41,440 --> 00:24:44,077
We really want to turn it over.

430
00:24:44,160 --> 00:24:48,197
But there’s some sense of and
this could be one of our defenses.

431
00:24:48,280 --> 00:24:51,237
Society can only adapt so
fast if it’s going to be good.

432
00:24:51,320 --> 00:24:56,237
Another way you could say
about it is maybe A.I. itself,

433
00:24:56,320 --> 00:24:58,217
if it didn’t have to care about us humans,

434
00:24:58,300 --> 00:25:00,777
it could just go off to Mars
and build all these automated

435
00:25:00,860 --> 00:25:04,237
factories and build its own
society and do its own thing.

436
00:25:04,320 --> 00:25:06,277
But that’s not the problem
we’re trying to solve.

437
00:25:06,360 --> 00:25:09,657
We’re not trying to solve
the problem of building a Dyson

438
00:25:09,740 --> 00:25:14,817
swarm of artificial robots
at in on some other planet.

439
00:25:14,900 --> 00:25:20,077
We’re trying to build these systems,
not so they can conquer the world,

440
00:25:20,160 --> 00:25:24,237
but so that they can interface with
our society and improve that society.

441
00:25:24,320 --> 00:25:27,177
And there’s a maximum rate
at which that can happen if we

442
00:25:27,260 --> 00:25:29,897
actually want to do it
in a human and humane way.

443
00:25:29,980 --> 00:25:33,717
All right. We’ve been talking about
white collar jobs and professional jobs.

444
00:25:33,800 --> 00:25:36,057
And one of the interesting
things about this moment

445
00:25:36,140 --> 00:25:40,477
is that there are ways in which
unlike past disruptions,

446
00:25:40,560 --> 00:25:46,217
it could be that blue collar
working class jobs, trades,

447
00:25:46,300 --> 00:25:49,017
jobs that require intense
physical engagement

448
00:25:49,100 --> 00:25:51,737
with the world might be,
for a little while,

449
00:25:51,820 --> 00:25:55,617
more protected that paralegals
and junior associates

450
00:25:55,700 --> 00:25:59,397
might be in more trouble
than plumbers and so on.

451
00:25:59,480 --> 00:26:01,397
One do you think that’s right?

452
00:26:01,480 --> 00:26:04,777
And two, it seems like how long that lasts

453
00:26:04,860 --> 00:26:09,937
depends entirely on how fast
robotics advances, right?

454
00:26:10,020 --> 00:26:13,577
So I think that may be
right in the short term.

455
00:26:13,660 --> 00:26:18,087
One of the things is Anthropic
and other companies

456
00:26:18,170 --> 00:26:19,877
are building these
very large data centers.

457
00:26:19,960 --> 00:26:23,157
This has been in the news like
are we building them too big.

458
00:26:23,240 --> 00:26:29,807
Are they’re using electricity
and driving up the prices for local towns.

459
00:26:29,890 --> 00:26:32,927
So there’s lots of excitement
and lots of concerns about them.

460
00:26:33,010 --> 00:26:36,667
But one of the things about the data
centers is like need a lot of electricians

461
00:26:36,750 --> 00:26:39,807
and you need a lot of construction
workers to build them.

462
00:26:39,890 --> 00:26:44,187
Now, I should be honest, actually, data
centers are not super labor intensive jobs

463
00:26:44,270 --> 00:26:46,887
to operate. We should
be honest about that.

464
00:26:46,970 --> 00:26:51,307
But they are very labor
intensive jobs to construct.

465
00:26:51,390 --> 00:26:54,007
And so we need a lot of electricians.

466
00:26:54,090 --> 00:26:56,427
We need a lot of construction workers,

467
00:26:56,510 --> 00:27:00,787
the same for various kinds
of manufacturing plants.

468
00:27:00,870 --> 00:27:06,307
And again, as kind of all more and more
of the intellectual work is done

469
00:27:06,390 --> 00:27:08,367
by A.I., what are the complements to it.

470
00:27:08,450 --> 00:27:11,027
Things that happen in the physical world.

471
00:27:11,110 --> 00:27:15,767
So, I think this kind
of seems very I mean,

472
00:27:15,850 --> 00:27:18,667
it’s hard to predict things,
but it seems very logical that

473
00:27:18,750 --> 00:27:20,467
this would be true in the short run.

474
00:27:20,550 --> 00:27:24,557
Now, in the longer run, maybe
just the slightly longer run.

475
00:27:24,640 --> 00:27:26,677
Robotics is advancing quickly.

476
00:27:26,760 --> 00:27:28,897
And, we shouldn’t exclude that.

477
00:27:28,980 --> 00:27:31,937
Even without very powerful A.I., there

478
00:27:32,020 --> 00:27:33,977
are things being automated
in the physical world.

479
00:27:34,060 --> 00:27:37,437
If you’ve seen a Waymo
or a Tesla recently,

480
00:27:37,520 --> 00:27:40,657
I think we’re not that far from
the world of self-driving cars.

481
00:27:40,740 --> 00:27:43,217
And then I think A.I.
itself will accelerate it,

482
00:27:43,300 --> 00:27:45,827
because if you have these
really smart, brains,

483
00:27:45,910 --> 00:27:48,157
one of the things they’re
going to be smart at is how do

484
00:27:48,240 --> 00:27:51,917
you design better robots and how
do you operate better robots.

485
00:27:52,000 --> 00:27:55,037
Do you think that though,
that there is something

486
00:27:55,120 --> 00:27:59,577
distinctively difficult about
operating in physical reality,

487
00:27:59,660 --> 00:28:02,717
the way humans do that is very
different from the kind

488
00:28:02,800 --> 00:28:06,677
of problems that A.I.
models have been overcoming already.

489
00:28:06,760 --> 00:28:10,237
Intellectually speaking, I don’t think so.

490
00:28:10,320 --> 00:28:14,857
We had this thing where
Anthropic’s model, Claude,

491
00:28:14,940 --> 00:28:18,677
was actually used to pilot the Mars Rover.

492
00:28:18,760 --> 00:28:21,657
It was used to plan
and pilot the Mars Rover.

493
00:28:21,740 --> 00:28:23,777
And we’ve looked at other
robotics applications.

494
00:28:23,860 --> 00:28:25,567
We’re not the only
company that’s doing it.

495
00:28:25,650 --> 00:28:28,617
There are different companies
that this is a general thing,

496
00:28:28,700 --> 00:28:31,597
not just something that we’re doing,

497
00:28:31,680 --> 00:28:36,357
but we have generally found
that while the complexity is

498
00:28:36,440 --> 00:28:42,257
higher, piloting a robot is it’s not
different in than playing a video game.

499
00:28:42,340 --> 00:28:44,027
It’s different in complexity.

500
00:28:44,110 --> 00:28:47,377
And we’re starting to get to the point
where we have that complexity.

501
00:28:47,460 --> 00:28:50,277
Now, what is hard is the physical form

502
00:28:50,360 --> 00:28:55,137
of the robot handling the higher stakes
safety issues that happen with robots.

503
00:28:55,220 --> 00:28:58,397
You don’t want robots
literally crushing people.

504
00:28:58,480 --> 00:29:00,497
That’s the we’re against. We’re against.

505
00:29:00,580 --> 00:29:02,477
That oldest sci-fi trope in the book

506
00:29:02,560 --> 00:29:07,437
is like the robot crushes you,
dropping the baby, breaking the dishes.

507
00:29:07,520 --> 00:29:11,477
There’s a number of practical
issues that will slow,

508
00:29:11,560 --> 00:29:16,477
just like what you described
in the law and human custom,

509
00:29:16,560 --> 00:29:21,727
there are these kind of safety
issues that will slow things down.

510
00:29:21,810 --> 00:29:25,247
But I don’t believe at all
that there is some kind

511
00:29:25,330 --> 00:29:27,807
of fundamental difference
between the kind of cognitive

512
00:29:27,890 --> 00:29:32,867
labor that the A.I. models do
and piloting things in the physical world.

513
00:29:32,950 --> 00:29:35,027
I think those are both
information problems.

514
00:29:35,110 --> 00:29:37,707
And I think they end
up being very similar.

515
00:29:37,790 --> 00:29:39,987
One one can be more complex in some ways,

516
00:29:40,070 --> 00:29:43,067
but I don’t think
that will protect us here.

517
00:29:43,150 --> 00:29:48,527
So you think it is reasonable
to expect the whatever

518
00:29:48,610 --> 00:29:53,087
your sci-fi vision of a robot
Butler might to be a reality

519
00:29:53,170 --> 00:30:00,170
in 10 years, let’s say it will
be on a longer time scale than

520
00:30:00,730 --> 00:30:04,527
the kind of genius level
intelligence of the A.I. models

521
00:30:04,610 --> 00:30:06,627
because of these practical issues.

522
00:30:06,710 --> 00:30:08,227
But it is only practical issues.

523
00:30:08,310 --> 00:30:10,507
I don’t believe it is fundamental issues.

524
00:30:10,590 --> 00:30:14,847
I think one way to say it
is that the brain of the robot

525
00:30:14,930 --> 00:30:18,797
will be made in the next couple
of years or the next few years.

526
00:30:18,880 --> 00:30:21,677
The question is making the robot body,

527
00:30:21,760 --> 00:30:24,017
making sure that body
operates safely and does

528
00:30:24,100 --> 00:30:26,077
the tasks it needs to do
that may take longer.

529
00:30:26,160 --> 00:30:31,177
O.K, so these are challenges
and disruptive forces

530
00:30:31,260 --> 00:30:36,897
that exist in the good timeline,
in the timeline where we are generally

531
00:30:36,980 --> 00:30:39,877
curing diseases, building
wealth, and maintaining

532
00:30:39,960 --> 00:30:43,497
a stable and Democratic world,
that we can use all this enormous wealth

533
00:30:43,580 --> 00:30:46,697
and plenty we will have
unprecedented societal

534
00:30:46,780 --> 00:30:49,557
resources to address these problems.

535
00:30:49,640 --> 00:30:56,137
It’ll be a time of plenty. And it’s just a
matter taking all these wonders and making

536
00:30:56,220 --> 00:30:57,917
sure everyone benefits from it.

537
00:30:58,000 --> 00:31:03,097
But then there are also
scenarios that are more dangerous.

538
00:31:03,180 --> 00:31:07,177
And so here we’re going
to move to the second Amadeus,

539
00:31:07,260 --> 00:31:11,177
which came out recently called
the adolescence of technology.

540
00:31:11,260 --> 00:31:15,197
That is about what you see
as the most serious A.I. risks.

541
00:31:15,280 --> 00:31:16,317
And you list a whole bunch.

542
00:31:16,400 --> 00:31:23,317
I want to try and focus on just two, which
are basically, the risk of human misuse.

543
00:31:23,400 --> 00:31:26,537
Misuse primarily by authoritarian regimes

544
00:31:26,620 --> 00:31:33,517
and governments, and scenarios where A.I.
goes rogue, what you call autonomy risks.

545
00:31:33,600 --> 00:31:37,877
Yes, yes. I just figured we should
have a more technical term for it.

546
00:31:37,960 --> 00:31:40,817
I’m not a then we can’t
just call it Skynet.

547
00:31:40,900 --> 00:31:43,497
I should have had a picture
of a terminator robot

548
00:31:43,580 --> 00:31:45,437
to scare people as much as possible.

549
00:31:45,520 --> 00:31:49,677
I think the internet, including
the internet, including your own eyes,

550
00:31:49,760 --> 00:31:52,717
are already generating that.
The internet does that for us just fine.

551
00:31:52,800 --> 00:31:58,757
So, so let’s so let’s talk about
the kind of political military dimension.

552
00:31:58,840 --> 00:32:03,017
So you say I’m going to quote
a swarm of billions of fully

553
00:32:03,100 --> 00:32:06,777
automated armed drones,
locally controlled by powerful

554
00:32:06,860 --> 00:32:12,177
A.I., strategically coordinated across
the world by even more powerful A.I.

555
00:32:12,260 --> 00:32:14,557
Could be an unbeatable army.

556
00:32:14,640 --> 00:32:19,227
Me and you’ve already talked
a little bit about how you think

557
00:32:19,310 --> 00:32:20,927
that in the best possible timeline,

558
00:32:21,010 --> 00:32:24,667
there’s a world where
essentially democracies stay

559
00:32:24,750 --> 00:32:29,527
ahead of dictatorships
and this kind of technology,

560
00:32:29,610 --> 00:32:32,987
therefore, to the extent
that it affects world politics is

561
00:32:33,070 --> 00:32:37,547
on is affecting it
on the side of the good guys.

562
00:32:37,630 --> 00:32:42,027
I’m curious about why you
don’t spend more time thinking

563
00:32:42,110 --> 00:32:48,987
about the model of what
we did in the Cold War,

564
00:32:49,070 --> 00:32:51,887
where it was not swarms of robot drones,

565
00:32:51,970 --> 00:32:55,147
but it was we had a technology
that threatened to destroy all

566
00:32:55,230 --> 00:32:59,107
of humanity Yeah, right.
There was a window where people

567
00:32:59,190 --> 00:33:02,227
talked about, oh, the US could
maintain a nuclear monopoly.

568
00:33:02,310 --> 00:33:05,427
That window closed. And from then on,
we basically spent the Cold War

569
00:33:05,510 --> 00:33:11,627
and rolling ongoing negotiations
with the Soviet Union.

570
00:33:11,710 --> 00:33:14,277
Now, there’s really only two
countries in the world that

571
00:33:14,360 --> 00:33:19,557
are doing intense A.I. work, the US
and the People’s Republic of China.

572
00:33:19,640 --> 00:33:24,197
I feel like you are. You are strongly
weighted towards a future where we’re

573
00:33:24,280 --> 00:33:28,277
staying ahead of the Chinese
and effectively building

574
00:33:28,360 --> 00:33:30,137
a kind of shield around democracy.

575
00:33:30,220 --> 00:33:31,377
That could even be a sword.

576
00:33:31,460 --> 00:33:35,877
But isn’t it just more likely
that if humanity survives all

577
00:33:35,960 --> 00:33:39,197
this in one piece, it will be
because the US and Beijing are

578
00:33:39,280 --> 00:33:43,397
just constantly sitting down,
hammering out A.I. control deals.

579
00:33:43,480 --> 00:33:48,797
So a few points on this. One is I
think there’s certainly risk of that,

580
00:33:48,880 --> 00:33:50,867
and I think if we end up in that world,

581
00:33:50,950 --> 00:33:52,657
that is actually exactly
what we should do.

582
00:33:52,740 --> 00:33:55,937
I mean, maybe I don’t maybe I
don’t talk about that enough,

583
00:33:56,020 --> 00:34:00,597
but I definitely am in favor
of trying to work out

584
00:34:00,680 --> 00:34:05,757
restraints here trying to take
some of the worst applications

585
00:34:05,840 --> 00:34:08,237
of the technology, which
could be some versions of these

586
00:34:08,320 --> 00:34:11,197
drones, which could be,
they’re used to create these

587
00:34:11,280 --> 00:34:15,447
terrifying biological weapons
like there is some precedent

588
00:34:15,530 --> 00:34:19,567
for the worst abuses being curbed.

589
00:34:19,650 --> 00:34:22,267
Often because they’re horrifying,

590
00:34:22,350 --> 00:34:27,527
while at the same time they
provide limited strategic advantage.

591
00:34:27,610 --> 00:34:30,207
So I’m all in favor of that.

592
00:34:30,290 --> 00:34:34,807
I’m at the same time,
a little concerned and a little

593
00:34:34,890 --> 00:34:38,967
skeptical that when things
kind of directly provide

594
00:34:39,050 --> 00:34:42,687
as much power as possible,
it’s kind of hard to get out

595
00:34:42,770 --> 00:34:45,147
of the game given what’s at stake.

596
00:34:45,230 --> 00:34:49,807
It’s hard to fully disarm.
If we go back to the Cold War we

597
00:34:49,890 --> 00:34:54,567
were able to reduce the number of missiles
that both sides had, but we were not

598
00:34:54,650 --> 00:34:57,207
able to entirely forsake nuclear weapons.

599
00:34:57,290 --> 00:35:00,387
And I would guess that we
would be in this world again.

600
00:35:00,470 --> 00:35:05,067
We can hope for a better one. And I’ll
certainly, I’ll certainly advocate for.

601
00:35:05,150 --> 00:35:08,247
Well, is it but is your
skepticism rooted in the fact

602
00:35:08,330 --> 00:35:11,207
that you think I would
provide a kind of advantage

603
00:35:11,290 --> 00:35:14,707
that nukes did not wear in the Cold War.

604
00:35:14,790 --> 00:35:18,867
Both sides. Even if you used
your nukes and gained advantages,

605
00:35:18,950 --> 00:35:20,877
you still probably would
be wiped out yourself.

606
00:35:20,960 --> 00:35:22,667
And you think that wouldn’t
happen with A.I.

607
00:35:22,750 --> 00:35:24,507
If you got an A.I.
Edge, you would just win.

608
00:35:24,590 --> 00:35:27,667
I mean, I think there’s a few things.

609
00:35:27,750 --> 00:35:32,067
And I just want to caveat like I’m
no international politics expert here.

610
00:35:32,150 --> 00:35:37,987
I think this weird world of intersection
of a new technology with geopolitics.

611
00:35:38,070 --> 00:35:41,027
So all of this is very but to be clear,

612
00:35:41,110 --> 00:35:43,647
as you yourself say,
in the course of the essay,

613
00:35:43,730 --> 00:35:46,467
the leaders of major A.I.
companies are in fact,

614
00:35:46,550 --> 00:35:48,667
likely to be major geopolitical actors.

615
00:35:48,750 --> 00:35:52,987
So you are sitting here. You are sitting
here as a potential geopolitical actor.

616
00:35:53,070 --> 00:35:54,867
I’m learning as much as I can about it.

617
00:35:54,950 --> 00:35:57,947
I just we should all have we
should all have humility here.

618
00:35:58,030 --> 00:36:01,167
I think there’s a failure
mode where read a book and go

619
00:36:01,250 --> 00:36:04,787
around like the world’s greatest
expert in national security.

620
00:36:04,870 --> 00:36:05,927
I’m trying to learn. That’s what.

621
00:36:06,010 --> 00:36:07,507
That’s what my profession does not.

622
00:36:07,590 --> 00:36:11,417
But it’s more annoying
when tech people do it.

623
00:36:11,500 --> 00:36:14,597
I don’t know. Let’s look at something
like the biological Weapons Convention.

624
00:36:14,680 --> 00:36:17,777
Biological weapons. They’re horrifying.

625
00:36:17,860 --> 00:36:22,677
Everyone hates them. We were able
to sign the biological Weapons Convention.

626
00:36:22,760 --> 00:36:25,257
The US genuinely stopped developing them.

627
00:36:25,340 --> 00:36:27,997
It’s somewhat more unclear
what the Soviet Union.

628
00:36:28,080 --> 00:36:31,297
But biological weapons
provide some advantage.

629
00:36:31,380 --> 00:36:37,317
But it’s not like they’re the difference
between winning and losing.

630
00:36:37,400 --> 00:36:39,317
And because they were so horrifying,

631
00:36:39,400 --> 00:36:42,857
we were kind of able to give
them up having 12,000 nuclear

632
00:36:42,940 --> 00:36:45,857
weapons versus 5,000 nuclear weapons.

633
00:36:45,940 --> 00:36:49,077
Again, you can kill more people on the
other side if you have more of these.

634
00:36:49,160 --> 00:36:51,717
But it’s like we were able
to be reasonable and say,

635
00:36:51,800 --> 00:36:53,637
we should have we
should have less of them.

636
00:36:53,720 --> 00:36:56,897
But if you’re like, O.K,
we’re going to completely disarm

637
00:36:56,980 --> 00:36:59,857
nuclear and we have
to trust the other side.

638
00:36:59,940 --> 00:37:01,477
I don’t think we ever got to that.

639
00:37:01,560 --> 00:37:05,657
And I think that’s just very hard unless
you had really reliable verification.

640
00:37:05,740 --> 00:37:10,127
So I would guess we’ll end up
in the same world with A.I.,

641
00:37:10,210 --> 00:37:12,587
that there are some kinds
of restraint that are going to be

642
00:37:12,670 --> 00:37:16,227
possible, but there are some
aspects that are so central

643
00:37:16,310 --> 00:37:19,067
to the competition that it will be.

644
00:37:19,150 --> 00:37:23,127
It will be hard to restrain them,
that democracies will make a trade off,

645
00:37:23,210 --> 00:37:25,427
that they will be willing
to restrain themselves

646
00:37:25,510 --> 00:37:29,747
more than authoritarian countries,
but will not restrain themselves fully.

647
00:37:29,830 --> 00:37:32,627
And the only world in which
I can see full restraint

648
00:37:32,710 --> 00:37:37,507
is one in which some kind of truly
reliable verification is possible.

649
00:37:37,590 --> 00:37:39,887
That would be. That would be my guess.

650
00:37:39,970 --> 00:37:46,027
And my analysis isn’t. Isn’t this
a case, though, for slowing down.

651
00:37:46,110 --> 00:37:51,647
And I know the argument is effectively, if
you slow down, China does not slow down.

652
00:37:51,730 --> 00:37:54,407
And then handing things
over to the authoritarians.

653
00:37:54,490 --> 00:37:58,067
But again, if you have right
now only two major powers

654
00:37:58,150 --> 00:38:01,567
playing in this game,
it’s not a multipolar game,

655
00:38:01,650 --> 00:38:06,967
why would it not make sense to say we
need a five year, mutually agreed upon.

656
00:38:07,050 --> 00:38:11,887
Slowdown in research towards
the geniuses in a data center scenario.

657
00:38:11,970 --> 00:38:15,027
I want to say two things at one time.

658
00:38:15,110 --> 00:38:18,807
I’m absolutely in favor
of trying to do that.

659
00:38:18,890 --> 00:38:24,787
So during the last administration,
I believe there was an effort by the US

660
00:38:24,870 --> 00:38:28,847
to reach out to the Chinese government
and say, there are dangers here.

661
00:38:28,930 --> 00:38:30,927
Can we collaborate? Can we work together?

662
00:38:31,010 --> 00:38:34,107
Can we work together on the dangers?

663
00:38:34,190 --> 00:38:37,167
And there wasn’t that much
interest on the other side.

664
00:38:37,250 --> 00:38:38,727
I think we should keep trying.

665
00:38:38,810 --> 00:38:43,047
But, even if that would mean that
your labs would have to slow down.

666
00:38:43,130 --> 00:38:46,347
Correct yeah.
If we really got it, if we really

667
00:38:46,430 --> 00:38:52,247
had a story of we can forcibly slow down,
the Chinese can forcibly slow down.

668
00:38:52,330 --> 00:38:54,807
We have verification.
We’re really doing it.

669
00:38:54,890 --> 00:38:57,807
Like if such a thing were really possible,

670
00:38:57,890 --> 00:39:03,567
if we could really get both sides
to do it, then I would be all for it.

671
00:39:03,650 --> 00:39:06,597
But I think what we
need to be careful of is,

672
00:39:06,680 --> 00:39:09,397
I don’t there’s this game
theory thing where sometimes

673
00:39:09,480 --> 00:39:14,097
you’ll hear a comment on the
CCP side where they’re like,

674
00:39:14,180 --> 00:39:16,417
"Oh yeah, A.I.is dangerous.
We should slow down."

675
00:39:16,500 --> 00:39:18,097
It’s really cheap to say that.

676
00:39:18,180 --> 00:39:21,117
And, actually arriving
at an agreement and actually

677
00:39:21,200 --> 00:39:24,757
sticking to the agreement
is much more and we haven’t it’s

678
00:39:24,840 --> 00:39:28,117
much more difficult.
And nuclear arms control it was

679
00:39:28,200 --> 00:39:32,657
a developed field that took
a long time to come.

680
00:39:32,740 --> 00:39:34,657
I know we don’t have those protocols.

681
00:39:34,740 --> 00:39:37,907
I will tell you something. Let me give
you something I’m very optimistic about.

682
00:39:37,990 --> 00:39:41,337
And then something I’m not optimistic
about and something in between.

683
00:39:41,420 --> 00:39:44,797
So the idea of using a worldwide agreement

684
00:39:44,880 --> 00:39:50,107
to restrain the use of A.I.
to build biological weapons, right.

685
00:39:50,190 --> 00:39:55,597
Like some of the things I write about
in the essay, reconstituting smallpox

686
00:39:55,680 --> 00:39:58,097
or mirror life this stuff is scary.

687
00:39:58,180 --> 00:39:59,627
Doesn’t matter if you’re a dictator.

688
00:39:59,710 --> 00:40:01,977
You don’t want that.
Like, no one wants that.

689
00:40:02,060 --> 00:40:04,487
And so could we have a worldwide treaty

690
00:40:04,570 --> 00:40:07,087
that says everyone who
builds powerful A.I. models is

691
00:40:07,170 --> 00:40:09,387
going to block them from doing this.

692
00:40:09,470 --> 00:40:11,407
And we have enforcement mechanisms

693
00:40:11,490 --> 00:40:14,507
around the treaty China
signs up for it Like hell.

694
00:40:14,590 --> 00:40:17,007
Maybe even North Korea signs up for it.

695
00:40:17,090 --> 00:40:20,227
Even Russia signs up for it.
I don’t think that’s too utopian.

696
00:40:20,310 --> 00:40:25,087
I think that’s possible. Conversely,
if we had something that said,

697
00:40:25,170 --> 00:40:30,747
you’re not going to make the next
most powerful A.I. model, everyone.

698
00:40:30,830 --> 00:40:32,727
Everyone’s going to stop.

699
00:40:32,810 --> 00:40:35,707
Boy, the commercial value
is in the tens of trillions.

700
00:40:35,790 --> 00:40:38,687
The military value is like,
this is the difference between

701
00:40:38,770 --> 00:40:41,787
being the preeminent world
power and not proposing it,

702
00:40:41,870 --> 00:40:46,187
as long as it’s not one of these fake
out games, but it’s not going to happen.

703
00:40:46,270 --> 00:40:49,247
What about then you mentioned
the current environment.

704
00:40:49,330 --> 00:40:51,927
You’ve had a few skeptical
things to say about Donald

705
00:40:52,010 --> 00:40:55,367
Trump and his trustworthiness
as a political actor.

706
00:40:55,450 --> 00:40:57,387
What about the domestic landscape.

707
00:40:57,470 --> 00:41:01,857
Whether it’s Trump or someone else,
you are building a tremendously powerful

708
00:41:01,940 --> 00:41:06,777
technology.
What is the safeguard there to prevent.

709
00:41:06,860 --> 00:41:10,817
Essentially A.I. becoming
a tool of authoritarian takeover

710
00:41:10,900 --> 00:41:13,517
inside a Democratic context
Yeah I mean, look, look,

711
00:41:13,600 --> 00:41:17,337
just to be clear, I think
the attitude we’ve taken

712
00:41:17,420 --> 00:41:21,437
as a company is very much to be
about policies and not the politics.

713
00:41:21,520 --> 00:41:25,217
You the company is not going
to say Donald Trump is great

714
00:41:25,300 --> 00:41:28,257
or Donald Trump is terrible,
but it doesn’t have to be

715
00:41:28,340 --> 00:41:32,017
Trump Yeah it is easy to imagine
a hypothetical US President.

716
00:41:32,100 --> 00:41:35,637
No, no, no. Who wants
to use your technology apps.

717
00:41:35,720 --> 00:41:41,017
Absolutely and for example.
That’s one reason why I’m worried about,

718
00:41:41,100 --> 00:41:44,077
the autonomous drone swarm, right.

719
00:41:44,160 --> 00:41:51,017
So the constitutional protections in our
military structures depend on the idea

720
00:41:51,100 --> 00:41:53,637
that there are humans who would we hope,

721
00:41:53,720 --> 00:41:57,477
disobey illegal orders
with fully autonomous weapons.

722
00:41:57,560 --> 00:41:59,837
We don’t necessarily
have those protections.

723
00:41:59,920 --> 00:42:04,377
But I actually think this
whole idea of constitutional

724
00:42:04,460 --> 00:42:10,137
rights and liberty along
many different dimensions,

725
00:42:10,220 --> 00:42:16,637
can be undermined by A.I. if we don’t
update these protections appropriately.

726
00:42:16,720 --> 00:42:18,997
So think about the Fourth Amendment.

727
00:42:19,080 --> 00:42:22,897
It is not illegal to put
cameras around everywhere

728
00:42:22,980 --> 00:42:25,837
in public space and record every
conversation in a public space.

729
00:42:25,920 --> 00:42:28,157
You don’t have a right
to privacy in a public space.

730
00:42:28,240 --> 00:42:32,717
But today, the government couldn’t
record that all and make sense of it.

731
00:42:32,800 --> 00:42:35,757
With A.I., the ability
to transcribe speech,

732
00:42:35,840 --> 00:42:39,197
to look through it, correlate
it all, you could say, oh,

733
00:42:39,280 --> 00:42:43,397
there’s this person is a
member of the opposition.

734
00:42:43,480 --> 00:42:48,117
This person is expressing this view
and make a map of all 100 million.

735
00:42:48,200 --> 00:42:50,377
And so are you going to make a mockery

736
00:42:50,460 --> 00:42:55,497
of the Fourth Amendment by the technology
finding kind of technical ways around it.

737
00:42:55,580 --> 00:42:59,297
And, and so again, if we had the time

738
00:42:59,380 --> 00:43:01,687
and we should do this,
we should try to do this even.

739
00:43:01,770 --> 00:43:03,107
Even if we don’t have the time.

740
00:43:03,190 --> 00:43:06,907
Is there some way of reconceptualizing

741
00:43:06,990 --> 00:43:10,487
constitutional rights
and liberties in the age of A.I.

742
00:43:10,570 --> 00:43:13,867
Maybe we don’t need to write
a new constitutional, but.

743
00:43:13,950 --> 00:43:17,487
But you have to do this. Do we expand
the meaning of the Fourth Amendment?

744
00:43:17,570 --> 00:43:19,707
Do we expand the meaning
of the First Amendment?

745
00:43:19,790 --> 00:43:22,547
And you have to do it just
as the legal profession

746
00:43:22,630 --> 00:43:27,307
or software engineers has
to update in a rapid amount of time.

747
00:43:27,390 --> 00:43:29,687
Politics has to update
in a rapid amount of time.

748
00:43:29,770 --> 00:43:33,727
That seems hard. What seems harder
dilemma that’s the dilemma of all of this.

749
00:43:33,810 --> 00:43:39,007
But what. So what seems harder
is preventing the second danger,

750
00:43:39,090 --> 00:43:43,027
which is the danger of essentially
what gets called misaligned A.I.

751
00:43:43,110 --> 00:43:46,947
Rogue A.I. In popular parlance,
from doing bad things

752
00:43:47,030 --> 00:43:52,007
without human beings telling
it them, they to do it right.

753
00:43:52,090 --> 00:43:56,167
And as I read your essays, the literature,

754
00:43:56,250 --> 00:44:00,237
everything I can see this just
seems like it’s going to happen.

755
00:44:00,320 --> 00:44:04,597
Not in the sense necessarily
that A.I. will wipe us all out,

756
00:44:04,680 --> 00:44:09,557
but it just seems to me that again,
I’m going to quote from your own writing,

757
00:44:09,640 --> 00:44:12,877
A.I. systems are unpredictable,
difficult to control.

758
00:44:12,960 --> 00:44:16,097
We’ve seen behaviors as varied
as obsession, sycophancy,

759
00:44:16,180 --> 00:44:18,517
laziness, deception, blackmail, and so on.

760
00:44:18,600 --> 00:44:22,117
Again, not from the models
you’re releasing into the world.

761
00:44:22,200 --> 00:44:27,957
But from A.I. models. And it just seems
like, tell me if I’m wrong about this.

762
00:44:28,040 --> 00:44:32,837
A world that has multiplying A.I.
agents working on behalf

763
00:44:32,920 --> 00:44:35,277
of people, millions upon
millions who are being given

764
00:44:35,360 --> 00:44:38,657
access to bank accounts,
email accounts, passwords,

765
00:44:38,740 --> 00:44:42,797
and so on, you’re just going
to have essentially some kind

766
00:44:42,880 --> 00:44:45,877
of misalignment, and a bunch
of A.I. are going to decide.

767
00:44:45,960 --> 00:44:47,337
Decide might be the wrong word,

768
00:44:47,420 --> 00:44:50,397
but they’re going to talk
themselves into taking down

769
00:44:50,480 --> 00:44:52,657
the power grid on the West
Coast or something.

770
00:44:52,740 --> 00:44:55,917
Won’t that happen Yeah,
I think there are definitely

771
00:44:56,000 --> 00:44:59,097
going to be things that go wrong,
particularly if we go quickly.

772
00:44:59,180 --> 00:45:02,997
So I don’t to back up a little
bit because this is one area

773
00:45:03,080 --> 00:45:07,617
where people have had just very
different intuitions, right.

774
00:45:07,700 --> 00:45:10,877
There are some people in the
field like Yann LeCun would

775
00:45:10,960 --> 00:45:14,737
be one example who say, look,
we programmed these A.I. models.

776
00:45:14,820 --> 00:45:17,557
We make them like we just
tell them to follow human

777
00:45:17,640 --> 00:45:19,697
instructions and they’ll
follow human instructions.

778
00:45:19,780 --> 00:45:24,097
Your Roomba vacuum cleaner doesn’t go
off and start shooting people like, why—

779
00:45:24,180 --> 00:45:25,637
Why’s an A.I. system going to do it?

780
00:45:25,720 --> 00:45:28,897
That’s one intuition. And some
people are so convinced of that.

781
00:45:28,980 --> 00:45:33,697
And then the other intuition is like
we basically we train these things.

782
00:45:33,780 --> 00:45:37,577
They’re just going to seek power.

783
00:45:37,660 --> 00:45:39,417
It’s like the Sorcerer’s Apprentice.

784
00:45:39,500 --> 00:45:41,417
How could you possibly imagine that?

785
00:45:41,500 --> 00:45:44,617
They’re a new species.
How can you imagine that.

786
00:45:44,700 --> 00:45:46,077
They’re not going to take over.

787
00:45:46,160 --> 00:45:49,277
And my intuition
is somewhere in the middle,

788
00:45:49,360 --> 00:45:53,737
which is that look,
you can’t just give instructions.

789
00:45:53,820 --> 00:45:57,967
I mean, we try, but you can’t
just have these things do

790
00:45:58,050 --> 00:46:02,267
exactly what you want to do. They’re
more like growing a biological organism.

791
00:46:02,350 --> 00:46:05,147
But there is a science
of how to control them.

792
00:46:05,230 --> 00:46:07,207
Like early in our training, these things

793
00:46:07,290 --> 00:46:10,467
are often unpredictable,
and then we shape them.

794
00:46:10,550 --> 00:46:12,867
We address problems one by one.

795
00:46:12,950 --> 00:46:18,407
So I have more of not a fatalistic view

796
00:46:18,490 --> 00:46:22,467
that these things are uncontrollable,
not what are you talking about.

797
00:46:22,550 --> 00:46:23,827
What could possibly go wrong?

798
00:46:23,910 --> 00:46:28,647
But I like this is a complex
engineering problem and I

799
00:46:28,730 --> 00:46:32,307
think something will go wrong
with someone’s A.I. system.

800
00:46:32,390 --> 00:46:35,827
Hopefully not ours.
Not because it’s an insoluble problem.

801
00:46:35,910 --> 00:46:38,687
But again, this and this
is the constant challenge

802
00:46:38,770 --> 00:46:41,107
because we’re moving so
fast and the scale of it.

803
00:46:41,190 --> 00:46:43,527
And tell me tell me if
I’m misunderstanding that

804
00:46:43,610 --> 00:46:45,587
the technological reality here.

805
00:46:45,670 --> 00:46:51,687
But if you have A.I. agents that have
been trained and officially aligned

806
00:46:51,770 --> 00:46:54,917
with human values,
whatever those values may be,

807
00:46:55,000 --> 00:46:59,697
but you have millions of them,
operating in digital space

808
00:46:59,780 --> 00:47:01,977
and interacting with other agents.

809
00:47:02,060 --> 00:47:07,037
How fixed is that alignment?

810
00:47:07,120 --> 00:47:11,577
To what extent can agents
change and D align in that

811
00:47:11,660 --> 00:47:16,927
context right now or in the future
when they’re learning more continuously.

812
00:47:17,010 --> 00:47:20,457
So a couple of points right now
the agents don’t learn continuously.

813
00:47:20,540 --> 00:47:25,197
And so we just deploy these agents
and they have a fixed set of weights.

814
00:47:25,280 --> 00:47:30,437
And so the problem is only that they’re
interacting in a million different ways.

815
00:47:30,520 --> 00:47:33,447
And so there’s a large number
of situations and therefore

816
00:47:33,530 --> 00:47:35,317
a large number of things
that could go wrong.

817
00:47:35,400 --> 00:47:38,117
But it’s the same agent.
It’s like it’s the same person.

818
00:47:38,200 --> 00:47:40,597
So the alignment is a constant thing.

819
00:47:40,680 --> 00:47:45,017
That’s one of the things that
has made it easier right now.

820
00:47:45,100 --> 00:47:48,297
Separate from that, there’s
a research area called continual

821
00:47:48,380 --> 00:47:53,637
learning, which is where these agents
would learn during time, learn on the job.

822
00:47:53,720 --> 00:47:56,677
And obviously that has a bunch
of that has a bunch of advantages.

823
00:47:56,760 --> 00:48:00,737
Some people think it’s one
of the most important barriers

824
00:48:00,820 --> 00:48:02,357
to making these more human like.

825
00:48:02,440 --> 00:48:04,877
But that would introduce all
these new alignment problems.

826
00:48:04,960 --> 00:48:08,097
So I’m actually a bit see,
to me that seems like the terrain

827
00:48:08,180 --> 00:48:12,497
where it becomes just again,
not impossible to stop the end

828
00:48:12,580 --> 00:48:18,117
of the world, but impossible to stop
punctuating something going wrong things.

829
00:48:18,200 --> 00:48:24,797
So I’m actually a skeptic.
That continual learning is, necessary.

830
00:48:24,880 --> 00:48:27,437
We don’t know yet,
but is necessarily needed.

831
00:48:27,520 --> 00:48:30,297
Like, maybe there’s a world
where the way we make these A.I.

832
00:48:30,380 --> 00:48:35,757
systems safe is by not having
them do continual learning again.

833
00:48:35,840 --> 00:48:39,417
Again, if we go back to the law,
that’s the international treaties.

834
00:48:39,500 --> 00:48:42,157
Like if you have some barrier that’s like,

835
00:48:42,240 --> 00:48:46,097
we’re going to take this path,
but we’re not going to take that path.

836
00:48:46,180 --> 00:48:48,717
I still have a lot of skepticism,

837
00:48:48,800 --> 00:48:53,267
but that’s the kind of thing that at
least doesn’t seem dead on arrival.

838
00:48:53,350 --> 00:48:57,587
One of the things that you’ve
tried to do is literally write

839
00:48:57,670 --> 00:49:03,667
a constitution, a long
constitution for your eye.

840
00:49:03,750 --> 00:49:08,247
What is that? So it’s.
What the hell is that?

841
00:49:08,330 --> 00:49:10,527
It’s actually almost
exactly what it sounds like.

842
00:49:10,610 --> 00:49:15,327
So basically, the constitution
is a document readable by humans.

843
00:49:15,410 --> 00:49:18,307
Ours is about 75 pages long.

844
00:49:18,390 --> 00:49:21,507
And as we’re training Claude,
as we’re training the A.I.

845
00:49:21,590 --> 00:49:25,767
system in some large fraction
of the tasks we give it,

846
00:49:25,850 --> 00:49:30,007
we say, please do this task
in line with this constitution,

847
00:49:30,090 --> 00:49:33,427
in line with this document
Yeah and then so every time

848
00:49:33,510 --> 00:49:36,007
Claude does a task, it kind
of like reads the constitution.

849
00:49:36,090 --> 00:49:39,157
And so as it’s training
every loop of it’s training,

850
00:49:39,240 --> 00:49:41,287
it looks at that constitution
and keeps it in mind.

851
00:49:41,370 --> 00:49:44,247
And so over time, we restore.

852
00:49:44,330 --> 00:49:47,787
And then we have Claude itself
or another copy of Claude

853
00:49:47,870 --> 00:49:51,657
evaluate Hey, did what Claude just
do in line with the constitution.

854
00:49:51,740 --> 00:49:57,077
So we’re using this document
as the control rod in a loop

855
00:49:57,160 --> 00:50:03,557
to train the model.
And so essentially Claude is an A.I. model

856
00:50:03,640 --> 00:50:09,237
whose fundamental principle
is to follow this constitution.

857
00:50:09,320 --> 00:50:11,637
And I think a really
interesting lesson we’ve

858
00:50:11,720 --> 00:50:16,417
learned, early versions of the
constitution were very prescriptive.

859
00:50:16,500 --> 00:50:18,537
They were very much about rules.

860
00:50:18,620 --> 00:50:24,117
So we would say, Claude should not
tell the user how to hotwire a car.

861
00:50:24,200 --> 00:50:28,037
Claude should not discuss
politically sensitive topics.

862
00:50:28,120 --> 00:50:31,677
But as we’ve worked
on this for several years,

863
00:50:31,760 --> 00:50:35,117
we’ve come to the conclusion
that the most robust way

864
00:50:35,200 --> 00:50:38,717
to train these models
is to train them at the level

865
00:50:38,800 --> 00:50:42,577
of principles and reasons.

866
00:50:42,660 --> 00:50:48,847
So now we say, Claude is a model,
it’s under a contract.

867
00:50:48,930 --> 00:50:54,267
Its goal is to serve the interests of the
user, but it has to protect third parties.

868
00:50:54,350 --> 00:50:58,467
Claude aims to be helpful,
honest and harmless.

869
00:50:58,550 --> 00:51:02,267
Claude aims to consider
a wide variety of interests.

870
00:51:02,350 --> 00:51:05,887
We tell the model about
how the model was trained.

871
00:51:05,970 --> 00:51:09,027
We tell it about how it’s
situated in the world,

872
00:51:09,110 --> 00:51:11,467
the job it’s trying to do for Anthropic,

873
00:51:11,550 --> 00:51:14,247
what Anthropic is aiming
to achieve in the world.

874
00:51:14,330 --> 00:51:20,367
That it has a duty to be ethical,
and respect human life.

875
00:51:20,450 --> 00:51:22,647
And we let it derive its rules from that.

876
00:51:22,730 --> 00:51:24,187
Now, there are still some hard rules.

877
00:51:24,270 --> 00:51:26,487
For example, we tell the model,

878
00:51:26,570 --> 00:51:29,647
no matter what you think,
don’t make biological weapons

879
00:51:29,730 --> 00:51:32,937
no matter what you think,
don’t make child sexual material.

880
00:51:33,020 --> 00:51:34,267
Those are like these hard rules.

881
00:51:34,350 --> 00:51:40,007
But we operate very much
at the level of principles.

882
00:51:40,090 --> 00:51:43,427
So if you read the US Constitution,
it doesn’t read like that.

883
00:51:43,510 --> 00:51:47,067
The US Constitution. I mean,
it has a little bit of flowery language,

884
00:51:47,150 --> 00:51:48,907
but it’s a set of. It’s a set of rules.

885
00:51:48,990 --> 00:51:52,867
Yes right. If you read your
Constitution, it’s something.

886
00:51:52,950 --> 00:51:55,247
It’s like you’re talking to a person.

887
00:51:55,330 --> 00:51:56,787
It’s like you’re talking to a person.

888
00:51:56,870 --> 00:52:01,927
I think I compared it to. Like if you
have a parent who dies and they like seal

889
00:52:02,010 --> 00:52:03,627
a letter that you read when you grow up,

890
00:52:03,710 --> 00:52:06,607
it’s a little bit like it’s
telling you who you should be

891
00:52:06,690 --> 00:52:08,627
and what advice you should follow.

892
00:52:08,710 --> 00:52:15,087
So this is where we get into
the mystical waters of A.I.

893
00:52:15,170 --> 00:52:21,287
a little bit. So again,
in your latest model,

894
00:52:21,370 --> 00:52:23,967
this is from one of the cards
they’re called that you guys

895
00:52:24,050 --> 00:52:27,947
release model card with these
models that I recommend reading.

896
00:52:28,030 --> 00:52:30,787
They’re very interesting.
It says the model.

897
00:52:30,870 --> 00:52:33,167
And again, this is who you’re
writing the constitution

898
00:52:33,250 --> 00:52:37,367
for expresses occasional
discomfort with the experience

899
00:52:37,450 --> 00:52:43,607
of being a product, some degree of concern
with impermanence and discontinuity.

900
00:52:43,690 --> 00:52:49,257
We found that opus 4.6. That’s the model
would assign itself a 15 to 20 percent

901
00:52:49,340 --> 00:52:54,177
probability of being conscious under
a variety of prompting conditions.

902
00:52:54,260 --> 00:52:57,217
Suppose you have a model
that assigns itself as 72 percent

903
00:52:57,300 --> 00:52:58,717
chance of being conscious.

904
00:52:58,800 --> 00:53:03,517
Would you believe it Yeah this is one
of these really hard to answer questions.

905
00:53:03,600 --> 00:53:08,977
But it’s very important. As much as every
question you’ve asked me before this

906
00:53:09,060 --> 00:53:13,657
as devilish a sociotechnical
problem as it had been,

907
00:53:13,740 --> 00:53:17,577
at least we at least
understand the factual basis

908
00:53:17,660 --> 00:53:20,097
of how to answer these questions.

909
00:53:20,180 --> 00:53:22,897
This is something rather different.

910
00:53:22,980 --> 00:53:26,247
We’ve taken a generally
precautionary approach here.

911
00:53:26,330 --> 00:53:27,997
We don’t know if the models are conscious.

912
00:53:28,080 --> 00:53:30,537
We’re not even sure that we
know what it would mean

913
00:53:30,620 --> 00:53:34,297
for a model to be conscious
or whether a model can be conscious.

914
00:53:34,380 --> 00:53:39,077
But we’re open to the
idea that it could be.

915
00:53:39,160 --> 00:53:45,307
And so we’ve taken certain
measures to make sure that

916
00:53:45,390 --> 00:53:48,567
if we hypothesize that the
models did have some morally

917
00:53:48,650 --> 00:53:51,027
relevant experience, I don’t
know if I want to use the word

918
00:53:51,110 --> 00:53:55,847
conscious that they do,
that they have a good experience.

919
00:53:55,930 --> 00:53:59,307
So the first thing we did,
I think this was six months ago

920
00:53:59,390 --> 00:54:02,347
or so is we gave the models
basically an I quit this job

921
00:54:02,430 --> 00:54:05,217
button where they can just
press the I quit this job

922
00:54:05,300 --> 00:54:08,027
button and then they have to stop
doing whatever the task is.

923
00:54:08,110 --> 00:54:10,207
They very infrequently press that button.

924
00:54:10,290 --> 00:54:13,987
I think it’s usually
around sorting through child

925
00:54:14,070 --> 00:54:17,827
sexualization material
or discussing something with

926
00:54:17,910 --> 00:54:20,647
a lot of Gore or blood
and guts or something.

927
00:54:20,730 --> 00:54:26,787
And similar to humans, the models will
just say, no, I don’t want to do this.

928
00:54:26,870 --> 00:54:31,147
Happens happens very rarely. We’re putting
a lot of work into this field called

929
00:54:31,230 --> 00:54:33,627
interpretability, which
is looking inside the brains

930
00:54:33,710 --> 00:54:37,587
of the models to try to understand
what they’re thinking.

931
00:54:37,670 --> 00:54:41,107
And you find things that
are evocative where there

932
00:54:41,190 --> 00:54:43,667
are activations that
light up in the models

933
00:54:43,750 --> 00:54:48,227
that we see as being associated

934
00:54:48,310 --> 00:54:52,367
with ID, the concept of anxiety
or something like that.

935
00:54:52,450 --> 00:54:54,707
That when characters experience anxiety

936
00:54:54,790 --> 00:54:57,667
in the text and then when
the model itself is in a situation

937
00:54:57,750 --> 00:54:59,487
that a human might associate with anxiety,

938
00:54:59,570 --> 00:55:03,887
that same anxiety, that same
anxiety neuron shows up now.

939
00:55:03,970 --> 00:55:06,447
Does that mean the model
is experiencing anxiety?

940
00:55:06,530 --> 00:55:07,947
That doesn’t prove that at all.

941
00:55:08,030 --> 00:55:13,287
But it does indicate
it I think to the user.

942
00:55:13,370 --> 00:55:17,187
And I would have to do
an entirely different interview.

943
00:55:17,270 --> 00:55:19,387
And maybe I can induce you to come back

944
00:55:19,470 --> 00:55:22,647
for that interview about
the nature of A.I. consciousness.

945
00:55:22,730 --> 00:55:26,757
But it seems clear to me
that people using these things,

946
00:55:26,840 --> 00:55:28,127
whether they’re conscious or not,

947
00:55:28,210 --> 00:55:30,047
are going to believe they
already believe they’re conscious.

948
00:55:30,130 --> 00:55:33,287
You already have people who have
parasocial relationships with A.I.

949
00:55:33,370 --> 00:55:37,747
You have people who complain
when models are retired.

950
00:55:37,830 --> 00:55:40,387
This ought to be clear.
I think that can be unhealthy.

951
00:55:40,470 --> 00:55:45,957
But that is it seems to me that
is guaranteed to increase in a way

952
00:55:46,040 --> 00:55:52,057
that I think calls into question the
sustainability of what you said earlier.

953
00:55:52,140 --> 00:55:56,497
You want to sustain, which is this sense
that whatever happens in the end,

954
00:55:56,580 --> 00:55:58,677
human beings are in charge.

955
00:55:58,760 --> 00:56:03,357
And I exists for our purposes
to use the science fiction

956
00:56:03,440 --> 00:56:06,617
example, if you watch Star Trek,
there are eyes on Star Trek.

957
00:56:06,700 --> 00:56:08,277
The ship’s computer is an A.I.

958
00:56:08,360 --> 00:56:10,477
Lieutenant Commander data is an A.I.,

959
00:56:10,560 --> 00:56:13,577
but jean-luc PyCaret
is in charge of the enterprise.

960
00:56:13,660 --> 00:56:20,077
But if people become fully convinced
that their A.I. is conscious in some way.

961
00:56:20,160 --> 00:56:26,117
And guess what. It seems to be better
than them at all kinds of decision making.

962
00:56:26,200 --> 00:56:30,937
How do you sustain human
mastery beyond safety?

963
00:56:31,020 --> 00:56:34,877
Safety is important, but mastery seems
like the fundamental question, and it

964
00:56:34,960 --> 00:56:37,577
seems like a perception
of A.I. consciousness.

965
00:56:37,660 --> 00:56:43,487
Doesn’t that inevitably undermine
the human impulse to stay in charge?

966
00:56:43,570 --> 00:56:47,167
So I think we should separate
out a few different things

967
00:56:47,250 --> 00:56:49,947
here that we’re all
trying to achieve at once.

968
00:56:50,030 --> 00:56:51,767
They’re like in tension with each other.

969
00:56:51,850 --> 00:56:56,727
There’s the question
of whether the I genuinely have

970
00:56:56,810 --> 00:57:00,147
a consciousness and if so,
how do we them a good experience.

971
00:57:00,230 --> 00:57:02,527
There’s a question of the
humans who interact with

972
00:57:02,610 --> 00:57:06,427
the A.I., and how do we give
those humans a good experience.

973
00:57:06,510 --> 00:57:09,207
And how does the perception
that A.I.'s might be conscious

974
00:57:09,290 --> 00:57:11,027
interact with that experience.

975
00:57:11,110 --> 00:57:13,747
And there’s the idea of how
we maintain human mastery,

976
00:57:13,830 --> 00:57:16,787
as we put it over the AI
system, these things,

977
00:57:16,870 --> 00:57:19,127
the last two Yeah, set aside
whether they’re conscious

978
00:57:19,210 --> 00:57:24,527
or not Yeah, the last two. But how do
you sustain mastery in an environment

979
00:57:24,610 --> 00:57:28,167
where most humans experience AI

980
00:57:28,250 --> 00:57:31,827
as if it is a peer and a
potentially superior peer.

981
00:57:31,910 --> 00:57:37,007
So the thing I was going
to say is that actually I wonder

982
00:57:37,090 --> 00:57:44,007
if there’s a kind of an elegant way to
satisfy all three, including the last two.

983
00:57:44,090 --> 00:57:46,907
Again, this is me dreaming
in machines of loving grace mode.

984
00:57:46,990 --> 00:57:51,727
This is. This mode I go into where I’m
like, man, I see all these problems.

985
00:57:51,810 --> 00:57:56,627
If we could solve is there an elegant way.

986
00:57:56,710 --> 00:57:59,127
This is not me saying
there are no problems here.

987
00:57:59,210 --> 00:58:00,347
That’s not how I think.

988
00:58:00,430 --> 00:58:06,447
But if we think about making
the Constitution of the AI

989
00:58:06,530 --> 00:58:10,167
so that the AI has
a sophisticated understanding

990
00:58:10,250 --> 00:58:12,887
of its relationship to human beings,

991
00:58:12,970 --> 00:58:17,007
and it induces psychologically
healthy behavior

992
00:58:17,090 --> 00:58:20,087
in the humans psychologically
healthy relationship

993
00:58:20,170 --> 00:58:22,067
between the A.I. and the humans.

994
00:58:22,150 --> 00:58:23,807
And I think something that could grow out

995
00:58:23,890 --> 00:58:26,887
of that psychologically
healthy, not psychologically

996
00:58:26,970 --> 00:58:30,167
unhealthy relationship
is some understanding

997
00:58:30,250 --> 00:58:32,927
of the relationship
between human and machine.

998
00:58:33,010 --> 00:58:37,377
And perhaps that relationship
could be the idea that,

999
00:58:37,460 --> 00:58:39,857
these models when you interact
with them and when you talk

1000
00:58:39,940 --> 00:58:43,157
to them, they’re really helpful.

1001
00:58:43,240 --> 00:58:46,357
They want the best for you.
They want you to listen to them,

1002
00:58:46,440 --> 00:58:49,737
but they don’t want
to take away your freedom

1003
00:58:49,820 --> 00:58:53,337
and your agency and take over your life.

1004
00:58:53,420 --> 00:58:57,217
in a way, they’re watching over you.

1005
00:58:57,300 --> 00:59:01,317
But you still have
your freedom and your will.

1006
00:59:01,400 --> 00:59:05,297
But this is so to me,
this is the crucial question.

1007
00:59:05,380 --> 00:59:10,077
Listening to you talk like one of my
question is, are these people on my side?

1008
00:59:10,160 --> 00:59:14,557
Are you on my side? And when you
talk about humans remaining in charge,

1009
00:59:14,640 --> 00:59:17,837
I think you’re on my side. That’s good.

1010
00:59:17,920 --> 00:59:20,257
But one thing I’ve done
in the past on this show and we’ll

1011
00:59:20,340 --> 00:59:23,037
end here, is I read
poems to technologists,

1012
00:59:23,120 --> 00:59:25,737
and you supplied the poem
"Machines of Loving Grace"

1013
00:59:25,820 --> 00:59:27,877
the name of a poem by Richard Brautigan.

1014
00:59:27,960 --> 00:59:30,617
Yes here’s how the poem ends.

1015
00:59:30,700 --> 00:59:35,497
I like to think it has
to be of a cybernetic ecology

1016
00:59:35,580 --> 00:59:41,187
where we are free of our labors and joined
back to nature, returned to our mammal

1017
00:59:41,270 --> 00:59:44,867
brothers and sisters, and all watched over

1018
00:59:44,950 --> 00:59:48,507
by machines of loving grace.

1019
00:59:48,590 --> 00:59:52,587
To me, that sounds like the dystopian end

1020
00:59:52,670 --> 00:59:56,747
where human beings are
reanimated, minimalized

1021
00:59:56,830 --> 01:00:02,107
and reduced and however benevolently
the machines are in charge.

1022
01:00:02,190 --> 01:00:05,847
So last question. What do you
hear when you hear that poem?

1023
01:00:05,930 --> 01:00:09,487
And if I think that’s a dystopia,
are you on my side?

1024
01:00:09,570 --> 01:00:12,867
It’s actually that poem
is interesting because it’s

1025
01:00:12,950 --> 01:00:15,007
interpretable in several different ways.

1026
01:00:15,090 --> 01:00:22,067
There some people say it’s
actually ironic that he says

1027
01:00:22,150 --> 01:00:24,567
it’s not going to happen quite that way.

1028
01:00:24,650 --> 01:00:29,347
Knowing the poet himself, then yes, I
think that’s a reasonable interpretation.

1029
01:00:29,430 --> 01:00:32,967
That’s one interpretation. Some
people would have your interpretation,

1030
01:00:33,050 --> 01:00:37,157
which is it’s meant literally,
but maybe it’s not a good thing.

1031
01:00:37,240 --> 01:00:40,537
But you could also interpret
it as it’s a return to nature.

1032
01:00:40,620 --> 01:00:42,217
It’s return to the core of what human.

1033
01:00:42,300 --> 01:00:48,057
We’re not being animalized. We’re being
we’re being reconnected with the world.

1034
01:00:48,140 --> 01:00:50,777
So I was aware of that ambiguity.

1035
01:00:50,860 --> 01:00:54,077
And, because I’ve always been
talking about the positive

1036
01:00:54,160 --> 01:00:55,637
side and the negative side.

1037
01:00:55,720 --> 01:01:02,720
So I actually think that may
be a tension that we may face,

1038
01:01:03,520 --> 01:01:08,517
which is that the positive
world and the negative world

1039
01:01:08,600 --> 01:01:11,937
in their early stages, maybe
even in their middle stages,

1040
01:01:12,020 --> 01:01:14,797
maybe even in their fairly late stages.

1041
01:01:14,880 --> 01:01:19,517
I wonder if the distance
between the good ending

1042
01:01:19,600 --> 01:01:24,017
and some of the subtle bad
endings is relatively small.

1043
01:01:24,100 --> 01:01:29,497
If it’s a very subtle thing like we’ve
put very subtle, made very subtle changes.

1044
01:01:29,580 --> 01:01:33,857
Like if you eat a particular fruit
from a tree in a garden or not.

1045
01:01:33,940 --> 01:01:39,357
Hypothetically Very small
thing Yeah big divergence Yeah.

1046
01:01:39,440 --> 01:01:44,537
I guess this always comes back to there’s
some fundamental questions here.

1047
01:01:44,620 --> 01:01:48,537
Yes yeah. Well, I guess
we’ll see how it plays out.

1048
01:01:48,620 --> 01:01:51,557
I do think of people in your position

1049
01:01:51,640 --> 01:01:58,057
as people whose moral choices will
carry an unusual amount of weight.

1050
01:01:58,140 --> 01:02:01,577
And so I wish you God’s help with them.

1051
01:02:01,660 --> 01:02:03,847
Dario Amodei, thank you for joining me.

1052
01:02:03,930 --> 01:02:05,430
Thank you for having me, Ross.

1053
01:02:26,880 --> 01:02:29,870
But what if I’m a robot?