Data Centers at the Crossroads

WEBVTT
1
00:00:00.810 --> 00:00:06.389
This week on Conservative Crossroads, should data center construction be encouraged?
2
00:00:07.030 --> 00:00:14.279
Debbie Jennings says yes. Joshua Trevino says no. Data center construction,
3
00:00:14.880 --> 00:00:22.280
coming up on Conservative Crossroads, now. Welcome back to Conservative Crossroads,
4
00:00:22.379 --> 00:00:26.600
the program where conservatives debate their differences rather than deny them.
5
00:00:27.179 --> 00:00:27.960
I'm Henry Olson.
6
00:00:28.309 --> 00:00:30.989
Senior Fellow at the Ethics and Public Policy Center and
7
00:00:31.010 --> 00:00:33.920
your host for this program. We've all been hearing a
8
00:00:33.960 --> 00:00:37.079
lot about data centers, and as you might expect, conservatives
9
00:00:37.460 --> 00:00:40.439
have different opinions on whether data centers should be encouraged,
10
00:00:41.000 --> 00:00:43.299
and if so, for what purposes. And here we have
11
00:00:43.340 --> 00:00:46.469
two intelligent and very knowledgeable people to talk about that.
12
00:00:46.869 --> 00:00:51.630
Debbie Jennings, Senior Policy Manager from the National Taxpayer Union Foundation,
13
00:00:52.049 --> 00:00:55.109
will take a generally positive view of data center construction,
14
00:00:55.509 --> 00:00:59.899
while Joshua Trevino, Chief Transformation Officer at the Texas Public
15
00:00:59.960 --> 00:01:03.179
Policy Foundation, says that maybe we should slow down a
16
00:01:03.299 --> 00:01:05.840
little bit and wait and smell the roses.
17
00:01:05.920 --> 00:01:06.459
So to speak.
18
00:01:07.099 --> 00:01:10.359
So this is a typical episode. Each person will have
19
00:01:10.379 --> 00:01:13.609
a five-minute opening statement. We'll start with Debbie, followed by Josh.
20
00:01:14.150 --> 00:01:16.250
And then we will have 40 to 45 minutes of
21
00:01:16.290 --> 00:01:20.810
mutual questioning, concluding with a two-minute conclusion delivered in the
22
00:01:20.950 --> 00:01:24.969
opposite order of the introduction. Debbie, you're up.
23
00:01:26.140 --> 00:01:28.799
Yeah, thank you for having me. And Joshua, it's great
24
00:01:28.819 --> 00:01:31.620
to chat with you here. I want to start off
25
00:01:31.640 --> 00:01:34.879
a little bit by talking about what data centers actually are.
26
00:01:35.819 --> 00:01:39.659
As we know, data centers are the physical infrastructure underpinning
27
00:01:40.359 --> 00:01:43.319
the digital economy. And the reason why this is very
28
00:01:43.359 --> 00:01:47.349
important now is because they're increasingly being built to fuel
29
00:01:47.409 --> 00:01:55.049
the artificial intelligence boom. So understandably, data centers are bringing
30
00:01:55.109 --> 00:01:57.810
together a lot of constituencies that haven't actually been brought
31
00:01:57.859 --> 00:02:01.609
together in the past. And this is causing some controversy.
32
00:02:01.849 --> 00:02:06.909
So data centers as large buildings that are very energy
33
00:02:06.950 --> 00:02:13.009
intensive and require little human capital to be in the office,
34
00:02:14.080 --> 00:02:17.319
they're being cited a lot in rural areas. But the
35
00:02:17.379 --> 00:02:21.340
end product, the artificial intelligence and the digital services are
36
00:02:21.539 --> 00:02:24.909
mainly at this time being used and tinkered with by
37
00:02:24.990 --> 00:02:28.599
people in Silicon Valley. So there's a little bit of
38
00:02:28.680 --> 00:02:33.080
tension here between rural interests who are actually getting the
39
00:02:33.099 --> 00:02:37.849
data center construction and other people who are far away
40
00:02:37.909 --> 00:02:40.389
in distant places who are actually using the technology. I
41
00:02:41.610 --> 00:02:46.389
think that data centers should be encouraged for many reasons.
42
00:02:46.669 --> 00:02:50.479
And I think that this seemingly large divide that is
43
00:02:50.520 --> 00:02:53.939
growing politically can be bridged if we think about this
44
00:02:54.099 --> 00:02:59.520
in economic terms. Artificial intelligence isn't the only thing that
45
00:02:59.939 --> 00:03:04.860
is using a data center. Obviously, our email and basically
46
00:03:05.060 --> 00:03:08.020
every software that we use on a daily basis is
47
00:03:08.060 --> 00:03:13.289
being fueled by data centers and compute. Artificial is just
48
00:03:13.509 --> 00:03:17.669
kind of the next generation of our digital infrastructure. Artificial
49
00:03:17.710 --> 00:03:22.810
intelligence can be used to essentially inject productivity and efficiency
50
00:03:22.849 --> 00:03:28.990
gains into nearly every single sector. So healthcare, finances, and
51
00:03:29.150 --> 00:03:32.469
basically every service that we use today can be made
52
00:03:32.550 --> 00:03:37.030
more efficient through artificial intelligence. I think that where the
53
00:03:37.069 --> 00:03:40.580
controversy really comes down is that rural communities feel like
54
00:03:40.599 --> 00:03:44.120
they're being exploited per se for a technology that won't
55
00:03:44.159 --> 00:03:47.659
benefit them. Maybe people are deciding that they don't want
56
00:03:47.699 --> 00:03:50.319
to use AI or they don't know how it'll affect them.
57
00:03:52.259 --> 00:03:55.270
I think that taking a step back and realizing that
58
00:03:55.469 --> 00:04:00.490
AI is kind of like the railroad build-out, the internet.com bubble,
59
00:04:01.050 --> 00:04:06.050
and other massive economic transformations that simply increase productivity and
60
00:04:06.090 --> 00:04:10.120
increase efficiency throughout the economy is a much more useful
61
00:04:10.159 --> 00:04:12.879
way to look at AI than the chatbots and the
62
00:04:13.300 --> 00:04:19.019
potentially unsatisfying user interfaces that we have today. So I
63
00:04:19.040 --> 00:04:23.209
think that data center growth growth and development should be encouraged. Um,
64
00:04:23.370 --> 00:04:26.649
I don't think that taxpayer funds should encourage this growth,
65
00:04:26.709 --> 00:04:30.360
but I do think that a light regulatory touch and
66
00:04:30.449 --> 00:04:34.879
a generally hands-off approach to allow private sector actors to
67
00:04:35.100 --> 00:04:38.019
build what they want to build is, uh, is called
68
00:04:38.060 --> 00:04:41.220
for in this area. And Joshua, I'll get off.
69
00:04:41.110 --> 00:04:41.410
To you now.
70
00:04:42.259 --> 00:04:44.029
Thank you so much, Debbie. Very much appreciate it.
71
00:04:44.589 --> 00:04:48.649
It may surprise some of the viewers here that I
72
00:04:48.829 --> 00:04:53.350
actually agree with a lot of what Ms. Jennings says here.
73
00:04:53.649 --> 00:04:55.050
And I don't say that by way of setting up
74
00:04:55.089 --> 00:04:58.990
the profound disagreement, but there are some disagreements here, so
75
00:04:59.009 --> 00:05:02.689
I want to eliminate them. Data centers are not really
76
00:05:02.870 --> 00:05:07.490
about data centers and the fracas around them, the political tumult,
77
00:05:08.110 --> 00:05:09.750
a lot of the debate around the data center is
78
00:05:10.370 --> 00:05:12.709
are really about fundamentally the things that the data centers
79
00:05:12.769 --> 00:05:17.129
make possible. Ms. Jennings talked, I think, quite rightly about
80
00:05:17.709 --> 00:05:20.519
the levels of service that we've come to expect from
81
00:05:20.540 --> 00:05:24.300
the digital economy, for example, that even in the absence
82
00:05:24.480 --> 00:05:29.550
or preceding artificial intelligence were significant goods provided by the
83
00:05:29.579 --> 00:05:31.949
data center. But I think what's happening with a lot
84
00:05:31.990 --> 00:05:33.990
of the data center conversation now and the reason that
85
00:05:34.009 --> 00:05:37.230
we have to take the opposition to it very seriously
86
00:05:37.250 --> 00:05:40.970
is it isn't simply an example, as I've heard some say,
87
00:05:40.990 --> 00:05:42.949
although not Ms. Jennings, I'm not imputing this to you,
88
00:05:43.050 --> 00:05:45.860
but others external to this conversation have said that it's
89
00:05:45.899 --> 00:05:49.620
an example of kind of a Luddite tendency, a function
90
00:05:49.639 --> 00:05:53.480
of ignorance, something that is curable by simple access to
91
00:05:53.560 --> 00:05:58.569
greater information on empirical facts about data center use. There
92
00:05:58.610 --> 00:06:00.310
may be some of that at the margins, but all
93
00:06:00.329 --> 00:06:05.199
of that is to fundamentally misunderstand why Americans in significant numbers,
94
00:06:05.300 --> 00:06:10.240
are very, very concerned about this build-out. They're concerned because
95
00:06:10.259 --> 00:06:13.959
of what the digital economy is, has done, and is
96
00:06:14.000 --> 00:06:19.170
about to become. What the digital economy has done, writ large,
97
00:06:19.589 --> 00:06:22.850
and I speak as somebody who has been a participant
98
00:06:22.870 --> 00:06:25.970
in it directly for almost a generation now. I was
99
00:06:26.009 --> 00:06:30.540
a very early adopter of the internet, websites, blogs, you
100
00:06:30.579 --> 00:06:33.899
name it, going now. And I am a daily user
101
00:06:33.920 --> 00:06:37.740
of artificial intelligence. I'm one of the, I think it's
102
00:06:37.819 --> 00:06:40.439
still a single digit percentage of American households that pay
103
00:06:40.959 --> 00:06:42.860
for AI use, but mine does.
104
00:06:43.720 --> 00:06:46.199
And so I'm very, very alert to its benefits.
105
00:06:46.259 --> 00:06:46.610
And as Ms.
106
00:06:46.649 --> 00:06:49.970
Jennings says, the efficiencies that are available. But the thing
107
00:06:49.990 --> 00:06:52.750
that we have to understand is, or there's two things
108
00:06:52.790 --> 00:06:56.189
that we have to understand. One is that it is
109
00:06:56.310 --> 00:07:00.290
not easily possible to build a case that the permanent
110
00:07:00.370 --> 00:07:04.110
things that matter to us as Americans, as citizens, as
111
00:07:04.189 --> 00:07:08.740
created beings have become significantly better, particularly since the introduction
112
00:07:08.759 --> 00:07:11.399
of the iPhone in 2007. When you look at any
113
00:07:11.459 --> 00:07:16.040
metric of meaning or civic stability or cohesion, democratic life,
114
00:07:16.939 --> 00:07:18.629
all of these have been on a downward trend. And
115
00:07:18.649 --> 00:07:21.110
I think there is a real case to be made
116
00:07:21.709 --> 00:07:25.500
that the era of the smartphone, social media, and ubiquitous
117
00:07:25.600 --> 00:07:28.139
digital technology has played a direct role in that, and
118
00:07:28.160 --> 00:07:31.720
that AI looks to accelerate those evils as well. The
119
00:07:31.740 --> 00:07:33.319
second thing that I would say, which is even more
120
00:07:33.360 --> 00:07:36.420
profoundly important, is that this is where I do disagree
121
00:07:36.439 --> 00:07:40.850
with Ms. Jennings, is that AI in particular is not
122
00:07:40.910 --> 00:07:44.189
the same as previous advances. You mentioned the railroad build-out,
123
00:07:44.290 --> 00:07:48.060
obviously a tremendously important historical turning point for the United
124
00:07:48.079 --> 00:07:53.879
States that brought great prosperity to America. But railroads and
125
00:07:53.920 --> 00:07:57.000
pretty much any other technological advance of the past 5,000
126
00:07:57.000 --> 00:08:01.129
years did not raise the same ontological questions that artificial
127
00:08:01.170 --> 00:08:04.250
intelligence does. There was never a conversation about whether a
128
00:08:04.310 --> 00:08:07.939
railroad deserved personhood. There was never a conversation about whether
129
00:08:07.980 --> 00:08:11.509
the light bulb had sentiment or bore natural rights. We
130
00:08:11.550 --> 00:08:14.290
are having those conversations now about artificial intelligence. And so
131
00:08:14.329 --> 00:08:16.189
I think I'll close with this, and then we can
132
00:08:16.550 --> 00:08:20.290
certainly open the conversation. I think that the Americans who say,
133
00:08:20.350 --> 00:08:22.790
hang on a second, we probably need to talk about this,
134
00:08:23.009 --> 00:08:25.100
we need to be concerned about it, and the apparatus
135
00:08:25.139 --> 00:08:28.899
that enables these things, that by the admission of their
136
00:08:28.920 --> 00:08:33.220
own creators might unemploy everybody in the human race. Now,
137
00:08:33.240 --> 00:08:35.610
I don't personally believe that, but that is something that
138
00:08:35.769 --> 00:08:38.850
AI creators have said, might not be something that we
139
00:08:38.870 --> 00:08:41.409
want to rush into. And I think this is fundamentally
140
00:08:41.450 --> 00:08:43.649
what the data center conversation is really about.
141
00:08:44.450 --> 00:08:47.240
Thank you, Joshua. Thank you, Debbie. And now the floor
142
00:08:47.269 --> 00:08:49.289
is both of yours to use as you see fit.
143
00:08:49.750 --> 00:08:52.179
Yeah, I want to comment on some of the things
144
00:08:52.200 --> 00:08:54.519
you said, and I'll comment on the later point first.
145
00:08:55.779 --> 00:09:01.139
As far as all of these lofty ideas about AI
146
00:09:01.460 --> 00:09:05.210
and all of these philosophical questions, I would say that
147
00:09:05.830 --> 00:09:08.909
these are obviously hugely concerning. And as you mentioned, the
148
00:09:08.950 --> 00:09:11.350
people who are creating these products are the ones that
149
00:09:11.409 --> 00:09:16.649
are bringing up the potential for mass job losses. There's
150
00:09:16.669 --> 00:09:21.919
been news that some AI companies are approaching religious leaders
151
00:09:22.059 --> 00:09:26.350
about AI and things like that. And I do think
152
00:09:26.370 --> 00:09:29.649
that that is an important consideration. But I also think
153
00:09:29.710 --> 00:09:33.169
that the narrative is being driven solely by those individuals
154
00:09:33.210 --> 00:09:36.570
at this time. There have not been enough representatives of
155
00:09:36.990 --> 00:09:42.750
the healthcare sector, teachers, even blue collar workers who are
156
00:09:42.789 --> 00:09:45.580
speaking up about what AI can do for their industries
157
00:09:46.100 --> 00:09:50.299
and really shaping the narrative into, okay, well, this could
158
00:09:50.500 --> 00:09:54.940
positively impact us enormously. So on the healthcare and biomedical
159
00:09:55.019 --> 00:09:59.610
front specifically, a lot of that conversation is being driven
160
00:09:59.690 --> 00:10:04.940
by people who are developing AI saying things like, well,
161
00:10:05.100 --> 00:10:09.279
there could be bioweapons created through AI and things like that.
162
00:10:09.379 --> 00:10:12.600
There's not enough people who are speaking up about the
163
00:10:13.419 --> 00:10:20.190
tremendous advances in healthcare, in drug creation, in DNA and
164
00:10:20.230 --> 00:10:23.740
genomic sequencing that can be done with AI. I also
165
00:10:23.799 --> 00:10:27.840
just saw one of these major AI companies has determined
166
00:10:27.879 --> 00:10:33.820
how to implement watermarking on genomic sequencing. So a lot
167
00:10:33.860 --> 00:10:37.750
of really important developments that are coming out. And it
168
00:10:37.809 --> 00:10:40.190
is unfortunate, in my opinion, that the narrative is being
169
00:10:40.230 --> 00:10:44.850
driven by this somewhat gloomy sense. Similarly, to your first
170
00:10:44.929 --> 00:10:49.500
point about internet boom and how a lot of social
171
00:10:49.580 --> 00:10:53.070
indicators have declined with the advancement of social media and
172
00:10:53.750 --> 00:10:57.669
how the iPhone age has not necessarily brought in a
173
00:10:57.710 --> 00:11:01.769
lot of the net positives that we think, it is
174
00:11:02.169 --> 00:11:05.679
also unfortunate that that has happened. But if you take
175
00:11:05.799 --> 00:11:09.940
the iPhone era and the advent of social media and
176
00:11:10.059 --> 00:11:12.840
all of the negatives that have come with that, Alongside
177
00:11:12.899 --> 00:11:15.100
all of the positives that have come with us having
178
00:11:15.200 --> 00:11:20.350
access to email instantly, being able to trade things online.
179
00:11:20.909 --> 00:11:24.090
Amazon started as a bookstore, and now it gives us
180
00:11:24.210 --> 00:11:26.929
everything we want within moments. I live in a very
181
00:11:27.090 --> 00:11:29.830
rural community, and I can still get things in two-day delivery,
182
00:11:29.889 --> 00:11:34.460
which is very, very impressive. So I would say on
183
00:11:34.500 --> 00:11:38.159
the net, the economic benefit, the productivity boom.
184
00:11:39.629 --> 00:11:43.750
The vast societal change for the positive from the internet
185
00:11:44.230 --> 00:11:48.850
has far outweighed some of these social indicators of too
186
00:11:48.889 --> 00:11:51.590
much social media use and other things that I think
187
00:11:51.629 --> 00:11:55.529
could also be, there could be some intervention at the
188
00:11:55.769 --> 00:11:58.620
family and societal level on that as well that I
189
00:11:58.679 --> 00:12:01.740
don't think would call for any slowdown of any of
190
00:12:01.779 --> 00:12:04.639
those technologies. Similarly to how I don't think there should
191
00:12:04.659 --> 00:12:07.039
be a call for a slowdown of AI based on
192
00:12:07.279 --> 00:12:08.539
broader societal concerns.
193
00:12:09.789 --> 00:12:10.909
Excellent points.
194
00:12:12.250 --> 00:12:13.549
I'll try to address them in turn, and then I
195
00:12:13.590 --> 00:12:16.629
have one big question for you, if I could. So
196
00:12:17.110 --> 00:12:20.480
let me address them in reverse. I completely agree that
197
00:12:20.500 --> 00:12:24.600
there are things that families can do. We have two
198
00:12:24.639 --> 00:12:27.580
boys ourselves, and so we contend with this all the time.
199
00:12:28.120 --> 00:12:30.000
How much technology do we allow to intrude?
200
00:12:31.720 --> 00:12:32.399
Into their lives.
201
00:12:33.139 --> 00:12:35.909
And so that's something that every parent faces. And just
202
00:12:35.929 --> 00:12:38.110
in the spirit of full disclosure, we are on the
203
00:12:38.129 --> 00:12:40.590
side of a lot less than most other parents. A
204
00:12:40.610 --> 00:12:43.649
lot of, you know, my sons are 10 and 12,
205
00:12:43.649 --> 00:12:45.740
and most of their peers have smartphones at this point.
206
00:12:46.149 --> 00:12:47.960
They have internet access, they can talk to each other
207
00:12:48.000 --> 00:12:51.139
and so on. And we don't allow any of that.
208
00:12:51.539 --> 00:12:53.639
And it's because we really want them to have as
209
00:12:53.700 --> 00:12:55.600
much of a true childhood as they possibly can. But
210
00:12:55.960 --> 00:12:59.730
we But I share that for transparency's sake, Beau, and
211
00:12:59.909 --> 00:13:02.149
actually to agree with you that there are things that
212
00:13:02.210 --> 00:13:06.049
families can do. But at the same time, I would
213
00:13:06.820 --> 00:13:11.299
posit that in the realm of making public policy, we
214
00:13:11.320 --> 00:13:15.559
should not be in the business of sacrificing or compounding
215
00:13:15.600 --> 00:13:18.440
the burdens that are placed upon families and parents, especially
216
00:13:18.960 --> 00:13:22.539
simply because we believe in a particular form of economic accelerationism.
217
00:13:22.990 --> 00:13:25.330
And again, I don't say this to, you know, I
218
00:13:25.389 --> 00:13:29.110
think economic growth is a good thing, manifestly so, and
219
00:13:29.129 --> 00:13:31.190
it's something that is a necessity for the maintenance of
220
00:13:31.250 --> 00:13:35.039
American civics. But at the same time, I don't think
221
00:13:35.080 --> 00:13:37.480
it's sufficient from a policy standpoint for us to say,
222
00:13:38.019 --> 00:13:41.399
sorry about the general environment, sorry what it's doing, but
223
00:13:41.440 --> 00:13:43.100
you as parents are going to have to cope with
224
00:13:43.139 --> 00:13:46.919
the consequences of our decisions. If we accept the thesis,
225
00:13:47.429 --> 00:13:49.620
which is not a libertarian thesis, by the way, that
226
00:13:49.639 --> 00:13:53.299
the fundamental unit of society is not the atomized individual,
227
00:13:53.360 --> 00:13:55.759
but the family and connection, then I think that leads
228
00:13:55.820 --> 00:13:59.450
us to a very different basket of policy preferences. One
229
00:13:59.490 --> 00:14:02.309
thing that you mentioned, Ms. Jennings, that I completely agree
230
00:14:02.350 --> 00:14:05.269
with is that the age of the internet has made
231
00:14:05.309 --> 00:14:09.309
it very possible for individuals in rural areas, who are
232
00:14:09.850 --> 00:14:12.830
not necessarily in urban cores or anything like that, to
233
00:14:12.870 --> 00:14:18.100
participate in work, in life, to get shipments, Again, on
234
00:14:18.120 --> 00:14:23.230
a personal note, I remember being absolutely amazed in 1999
235
00:14:23.230 --> 00:14:27.090
when I got books delivered from Amazon.com in Derrida, Louisiana,
236
00:14:27.110 --> 00:14:29.990
which is truly one of the backwaters of the American South.
237
00:14:31.289 --> 00:14:35.169
So I'm very sensible to the advantages that it offers.
238
00:14:35.509 --> 00:14:37.830
And yet, I do think that the question begged here
239
00:14:38.149 --> 00:14:42.669
is one of meaning. What do we value? In civics,
240
00:14:42.970 --> 00:14:46.450
is it true that democracy is better? Is it true
241
00:14:46.590 --> 00:14:49.730
that civic discourse is better? Is it true that Americans
242
00:14:49.789 --> 00:14:53.080
are better able to self-govern now that they have that
243
00:14:53.159 --> 00:14:56.000
sum total of human knowledge available in their pockets via
244
00:14:56.059 --> 00:14:59.860
a smartphone? Is it true that we feel more fulfilled,
245
00:14:59.879 --> 00:15:01.899
that we're more connected, that we have more of what
246
00:15:01.940 --> 00:15:05.419
de Tocqueville called the associations that make democracy work? I
247
00:15:05.460 --> 00:15:06.879
don't think you can make the case on any of
248
00:15:06.919 --> 00:15:10.820
those counts, and that's just through aggregated data on a
249
00:15:10.840 --> 00:15:11.279
lot of fronts.
250
00:15:11.320 --> 00:15:14.179
Now, was it caused by the digital revolution?
251
00:15:14.379 --> 00:15:17.210
Probably not. I think we've seen these declines in evidence
252
00:15:17.250 --> 00:15:19.750
for a long time. But was it accelerated by them?
253
00:15:19.789 --> 00:15:23.179
I think the answer is unquestionably yes. Last thing I'll say,
254
00:15:23.190 --> 00:15:26.610
and then I'll toss this back to you with a question. Um, uh,
255
00:15:26.840 --> 00:15:29.259
you mentioned that, uh, that there's a lot of, there's
256
00:15:29.279 --> 00:15:31.799
a lot of, uh, critics and individuals who were predicting
257
00:15:31.840 --> 00:15:34.539
doom and so on, and that there's not enough individuals
258
00:15:34.580 --> 00:15:41.029
talking about the benefits, uh, of, of, of AI in particular. Um, uh, but, uh,
259
00:15:41.090 --> 00:15:42.450
I want to, but I want to submit a couple
260
00:15:42.470 --> 00:15:44.389
of things to you. First of all, AI has the
261
00:15:44.409 --> 00:15:46.529
biggest promoter in the world and the president of the
262
00:15:46.549 --> 00:15:50.190
United States. Uh, right now, the messaging that has come
263
00:15:50.210 --> 00:15:54.000
out of the Trump administration has been uniformly positive. And
264
00:15:54.029 --> 00:15:56.299
I think defensively so, actually. They're very alert to the
265
00:15:56.340 --> 00:15:58.299
fact that a lot of AI investment is propping up
266
00:15:58.340 --> 00:16:00.519
the economy right now, and that it probably is, I think.
267
00:16:00.700 --> 00:16:03.600
And I agree with this contention that it's probably the
268
00:16:03.639 --> 00:16:08.019
basis for American industrial and material supremacy in the future.
269
00:16:08.559 --> 00:16:10.700
And so it does have that promoting it. You don't
270
00:16:10.759 --> 00:16:13.440
get a much bigger platform than that. And there's a
271
00:16:13.470 --> 00:16:15.889
lot of people in that orbit and ancillary to it
272
00:16:15.909 --> 00:16:18.409
who are also promoting it. But the other part of
273
00:16:18.490 --> 00:16:22.230
it is that, and I'm stealing a little bit from
274
00:16:22.289 --> 00:16:25.710
Peggy Noonan in her Wall Street Journal column this past Sunday,
275
00:16:26.409 --> 00:16:32.320
but it's worth restating. The people that you're referring to
276
00:16:32.679 --> 00:16:36.360
who are talking about the negative impacts of all this
277
00:16:37.320 --> 00:16:40.580
are the people creating it, which as far as I
278
00:16:40.620 --> 00:16:48.330
can tell, has happened only once. in certainly modern industrial history.
279
00:16:48.529 --> 00:16:51.730
And it came with nuclear weapons in which you had
280
00:16:51.750 --> 00:16:54.409
the creators of the thing talking about how bad the
281
00:16:54.450 --> 00:16:57.470
thing was going to be. And Noonan observed, and I
282
00:16:57.509 --> 00:17:00.879
agree with it, that it's extremely unusual and I think
283
00:17:00.929 --> 00:17:04.039
properly alarming that it is the creators of the thing
284
00:17:04.059 --> 00:17:07.259
that are warning us how terrible it could be. And
285
00:17:07.420 --> 00:17:09.500
of course we have, I think it's only prudent that
286
00:17:09.519 --> 00:17:13.359
we take those warnings seriously. So my question to you
287
00:17:13.400 --> 00:17:14.779
then is, Why is it them?
288
00:17:16.069 --> 00:17:19.569
I would say that every industry, one of the things
289
00:17:19.609 --> 00:17:22.509
that the creators of these products are asking for is
290
00:17:22.569 --> 00:17:27.410
more regulatory intervention to pause or slow development of some
291
00:17:27.430 --> 00:17:32.359
of these tools. Every industry that is new does go
292
00:17:32.440 --> 00:17:34.759
to the government and ask for some type of regulation
293
00:17:35.140 --> 00:17:37.880
just to create laws under which it can operate to
294
00:17:38.019 --> 00:17:43.869
reduce liability, right? So In the cryptocurrency industry, this is
295
00:17:43.890 --> 00:17:47.369
something that we've also seen where crypto leaders are going
296
00:17:47.390 --> 00:17:50.690
to Washington and asking for rules around AI. They don't
297
00:17:50.730 --> 00:17:54.420
want to operate in an ungoverned space because they want
298
00:17:54.460 --> 00:17:59.019
to be legitimized, brought into the system, and also reduce liability.
299
00:17:59.539 --> 00:18:04.559
So I would say asking for regulation, that is very normal.
300
00:18:05.579 --> 00:18:06.200
Going around and.
301
00:18:06.160 --> 00:18:09.990
Saying that our product may may doom everyone, that is
302
00:18:10.230 --> 00:18:14.910
absolutely atypical. We don't really see other industries saying that.
303
00:18:15.009 --> 00:18:19.630
Even industries that do have legitimate human risks typically try
304
00:18:19.710 --> 00:18:25.730
to brush over those risks. I would say that while
305
00:18:25.769 --> 00:18:29.970
these concerns are real, this is something that, first of all,
306
00:18:30.829 --> 00:18:32.809
the companies who are saying this are largely based in
307
00:18:32.829 --> 00:18:36.910
the United States. We don't see Chinese artificial intelligence developers
308
00:18:37.450 --> 00:18:41.619
saying that they think their products could cause mass societal destruction.
309
00:18:42.599 --> 00:18:46.900
And yet, at the same time, Chinese artificial intelligence creators
310
00:18:47.059 --> 00:18:49.420
are still going to continue no matter what. No matter
311
00:18:49.480 --> 00:18:53.299
what happens in the US, China is also going to
312
00:18:53.339 --> 00:18:55.890
be working very hard at this issue. And not only China,
313
00:18:55.930 --> 00:19:00.170
but other countries as well. Even our Canadian neighbors are
314
00:19:00.230 --> 00:19:09.130
very intent on being an artificial intelligence and data center place. Unfortunately,
315
00:19:09.410 --> 00:19:13.549
while companies in the U.S. may be saying that there's
316
00:19:13.569 --> 00:19:16.470
a lot of potential societal harm, they are the ones
317
00:19:16.559 --> 00:19:19.650
creating it. They can reduce harm. They can reduce the
318
00:19:19.690 --> 00:19:23.799
harm that they're causing. And further, exploring some of the
319
00:19:23.839 --> 00:19:28.180
harms in a safe environment within their company, just by
320
00:19:28.259 --> 00:19:31.319
researching what kinds of harms these things could do, could
321
00:19:31.400 --> 00:19:34.279
actually prove a national security benefit, because then we might
322
00:19:34.339 --> 00:19:36.519
have a little bit of knowledge as to what the
323
00:19:36.559 --> 00:19:40.819
Chinese could do with their AI products, right? Um, some
324
00:19:40.839 --> 00:19:45.710
things like the recent AI hack where a model escaped
325
00:19:45.849 --> 00:19:51.960
its testing environment and hacked into another, uh, program. This
326
00:19:52.000 --> 00:19:55.720
was seen as a very dangerous and kind of upping
327
00:19:55.759 --> 00:20:00.759
the ante thing about AI. I would say that in
328
00:20:00.819 --> 00:20:05.549
order for these safety experiments to take place, companies need
329
00:20:05.569 --> 00:20:10.519
to be very very cautious about creating a safe environment
330
00:20:10.539 --> 00:20:13.700
to test their products, right? So in that instance, the
331
00:20:13.740 --> 00:20:17.019
company perhaps didn't have as many guardrails as they should
332
00:20:17.059 --> 00:20:23.660
have when they were injecting a lot of potentially risky
333
00:20:24.509 --> 00:20:28.470
things into the AI prompt. So it is definitely pertinent
334
00:20:28.549 --> 00:20:32.430
on companies to ensure that their products are safe, ensure
335
00:20:32.509 --> 00:20:35.319
that they're doing enough testing, The testing, like I said,
336
00:20:35.359 --> 00:20:38.579
could be beneficial for national security purposes. So we know
337
00:20:38.619 --> 00:20:40.819
what other people are able to do with their AI products.
338
00:20:41.460 --> 00:20:43.079
And I think that a lot of these things are
339
00:20:43.119 --> 00:20:46.400
things that companies can do on their own. They can
340
00:20:46.500 --> 00:20:51.500
implement these guardrails. They don't necessarily need President Trump or
341
00:20:51.519 --> 00:20:54.099
the US government to come out and say, these are
342
00:20:54.119 --> 00:20:56.740
the specific ways that you have to make yourself safe.
343
00:20:57.359 --> 00:21:00.460
I actually think also that if the government were to
344
00:21:00.480 --> 00:21:04.960
take that approach, then it could not only limit AI development,
345
00:21:05.039 --> 00:21:07.869
which some people may want, but it could really also
346
00:21:08.210 --> 00:21:13.230
hamstring just normal processes. As you can see, like I
347
00:21:13.250 --> 00:21:19.970
mentioned in cryptocurrency, other advanced innovative technologies, the government is
348
00:21:20.109 --> 00:21:25.529
not necessarily equipped to know how these things work. There's
349
00:21:25.829 --> 00:21:30.329
famous Senate hearings where senators were asking Mark Zuckerberg, how
350
00:21:30.369 --> 00:21:34.630
Facebook worked back in the day, right? So it really
351
00:21:34.650 --> 00:21:37.650
is something that companies should take seriously, and they should
352
00:21:37.690 --> 00:21:40.349
not release unsafe products out into the world. But at
353
00:21:40.369 --> 00:21:44.029
the same time, asking the government for an overly burdensome
354
00:21:44.089 --> 00:21:47.730
regulatory approach is not necessarily the right move either.
355
00:21:47.759 --> 00:21:51.880
I'm just curious, which firm or actor or person do
356
00:21:51.900 --> 00:21:53.660
you trust the most to.
357
00:21:53.579 --> 00:21:54.160
Do the right thing?
358
00:21:55.829 --> 00:21:59.250
I would trust US-based organizations the most to do the
359
00:21:59.269 --> 00:22:05.910
right thing. I think that obviously our adversaries like China
360
00:22:06.230 --> 00:22:08.250
want to get ahead of this issue. They want to
361
00:22:08.269 --> 00:22:11.190
get ahead of AI for multiple purposes. I've actually lived
362
00:22:11.230 --> 00:22:15.349
in China and they've been using AI and this technology
363
00:22:15.390 --> 00:22:16.650
for a very, very long time.
364
00:22:16.789 --> 00:22:19.079
I lived there back in 2018.
365
00:22:19.079 --> 00:22:22.720
And even at that time, their technological capabilities are just
366
00:22:22.890 --> 00:22:27.589
so extremely far ahead of ours. We don't really know
367
00:22:27.690 --> 00:22:31.829
what they're doing and what they have. And they have
368
00:22:31.869 --> 00:22:36.069
some open source, publicly released AI models, but we don't
369
00:22:36.089 --> 00:22:38.859
really know what the Chinese government is doing. And I
370
00:22:38.900 --> 00:22:41.859
would think that we wouldn't want to slow down our
371
00:22:42.519 --> 00:22:45.920
US actors who are willing, obviously, to approach the US government,
372
00:22:45.980 --> 00:22:48.240
work with the US government, Many of them want to
373
00:22:48.279 --> 00:22:52.039
be contractors for the U.S. government in some way. We
374
00:22:52.059 --> 00:22:55.180
wouldn't want to hamstring our own companies.
375
00:22:57.000 --> 00:22:58.880
It sounds like you and I were in China roughly
376
00:22:58.900 --> 00:23:00.480
at the same time, and I could not agree more
377
00:23:00.509 --> 00:23:05.890
that the malevolence of the Chinese state cannot be overstated,
378
00:23:05.950 --> 00:23:10.650
especially with apparatuses like this. And, of course, DeepSeek caused
379
00:23:10.710 --> 00:23:13.670
a fracas, I believe it was earlier this year, when
380
00:23:13.869 --> 00:23:14.309
it was released.
381
00:23:14.349 --> 00:23:15.230
So full concurred.
382
00:23:16.589 --> 00:23:18.710
That there is a China threat involved in it. Although
383
00:23:18.750 --> 00:23:21.089
I will note that when we look at public sentiment
384
00:23:21.109 --> 00:23:22.960
in the United States, we don't see it invoked all
385
00:23:23.000 --> 00:23:25.200
that much. The individuals invoking the China threat tend to
386
00:23:25.220 --> 00:23:29.789
be on government and industry side. Not that that invalidates it,
387
00:23:29.819 --> 00:23:31.480
but it's just worth noting. It doesn't seem to be
388
00:23:31.519 --> 00:23:35.640
driving public consciousness. I do want to drill down on
389
00:23:35.680 --> 00:23:38.730
this though, and I'm happy to give my own answer.
390
00:23:38.829 --> 00:23:40.069
Or if you want me to give my own answer first,
391
00:23:40.089 --> 00:23:43.109
I'm happy to do it. And it's not a gotcha question,
392
00:23:43.130 --> 00:23:46.440
but I'm actually, I think it's important to discern because
393
00:23:46.460 --> 00:23:50.079
I don't think it's enough to say that we simply
394
00:23:50.119 --> 00:23:55.759
trust a US-based actor. Anthropic is a US-based actor. And
395
00:23:56.000 --> 00:23:58.859
of course we have people inside Anthropic who by reports
396
00:23:58.900 --> 00:24:02.980
are literally worshiping Claude at this point, which doesn't engender.
397
00:24:02.700 --> 00:24:03.380
Trust in me.
398
00:24:04.529 --> 00:24:08.150
So I'm legit curious, what is a model since we've
399
00:24:08.170 --> 00:24:12.609
talked about the predictions of threat, how that emanates back
400
00:24:12.690 --> 00:24:16.880
into the politics of data centers, What is a responsible actor?
401
00:24:17.460 --> 00:24:18.359
I understand your point.
402
00:24:18.920 --> 00:24:23.069
We definitely don't want AI models who think that their
403
00:24:23.549 --> 00:24:28.589
product is religiously sanctified and has human personhood and things
404
00:24:28.630 --> 00:24:31.769
like that. That's the beauty of what we have right now,
405
00:24:31.849 --> 00:24:35.859
which is we have, I think, about four large models
406
00:24:35.880 --> 00:24:39.500
that are all competing with each other. I think that
407
00:24:39.599 --> 00:24:42.099
Anthropic has done a lot of things recently that have
408
00:24:42.140 --> 00:24:45.220
kind of shifted public opinion away from their model and
409
00:24:45.259 --> 00:24:52.680
towards other models, like there's Grok, OpenAI, and Gemini through
410
00:24:52.700 --> 00:24:56.119
Google and things like that. And the more that we
411
00:24:56.140 --> 00:25:00.740
can promote competition amongst the models, the better, obviously. Definitely,
412
00:25:00.759 --> 00:25:04.680
it's a judgment call, right? So if you want to
413
00:25:04.720 --> 00:25:09.519
be using an AI whose leaders politically align with you,
414
00:25:09.619 --> 00:25:15.619
then that's perfectly fine. I think that What goes into
415
00:25:15.660 --> 00:25:20.079
these AIs, I don't think that they're training their model, Claude,
416
00:25:20.299 --> 00:25:26.069
to believe that it's sacrosanct and human and things like that.
417
00:25:26.250 --> 00:25:28.769
I think that that's not actually what's happening. But I
418
00:25:28.809 --> 00:25:29.049
would be.
419
00:25:29.049 --> 00:25:32.029
Curious to hear more about what you think on that front,
420
00:25:32.130 --> 00:25:36.230
how you think that that belief of the people who
421
00:25:36.269 --> 00:25:40.009
run Anthropic or the engineers behind Anthropic, their beliefs about
422
00:25:40.029 --> 00:25:44.390
their product, how that actually creates Malicious intent within the
423
00:25:44.430 --> 00:25:48.569
model and whether that is something that competition and the
424
00:25:48.910 --> 00:25:51.990
competitive environment in the US will not be able to address.
425
00:25:52.910 --> 00:25:53.170
Thank you.
426
00:25:53.190 --> 00:25:54.670
And I know I'm happy to answer that. And I
427
00:25:54.690 --> 00:25:56.789
appreciate you asking the question, because I think I think
428
00:25:56.809 --> 00:25:59.339
it is the question, honestly, to ask.
429
00:26:01.380 --> 00:26:02.519
I said earlier.
430
00:26:02.160 --> 00:26:08.490
When we began this conversation that AI raises ontological questions,
431
00:26:09.430 --> 00:26:14.329
anthropological questions that pretty much no other technology ever really has.
432
00:26:14.349 --> 00:26:17.369
And I would argue it's not a new thing in
433
00:26:17.410 --> 00:26:19.579
human history, but it is certainly a new thing in modernity.
434
00:26:20.000 --> 00:26:22.039
And it's a new thing in the technological.
435
00:26:21.559 --> 00:26:22.420
Age, for sure.
436
00:26:23.539 --> 00:26:24.460
And so what I fear.
437
00:26:26.710 --> 00:26:28.930
And what I think is already coming to pass, is
438
00:26:29.650 --> 00:26:34.099
because we are effectively, and let me preface actually what
439
00:26:34.119 --> 00:26:36.960
I'm about to say, I'm not advocating for a particular
440
00:26:37.079 --> 00:26:39.480
religious point of view, right? I'm not advocating for a
441
00:26:39.500 --> 00:26:43.089
particular brand of Christianity. I'm not even arguing for Christianity
442
00:26:43.269 --> 00:26:47.170
as such, except to the extent that that foundation is
443
00:26:47.230 --> 00:26:50.410
necessary for understanding the American founding. And I take it
444
00:26:50.470 --> 00:26:52.869
as a given, and if you disagree, we can talk
445
00:26:52.910 --> 00:26:55.099
about this too, but But I take it as a
446
00:26:55.119 --> 00:26:58.029
given that the mission statement of the United States is
447
00:26:58.089 --> 00:27:02.769
expressed in the Declaration of Independence, required a specifically Judeo-Christian framework,
448
00:27:03.109 --> 00:27:06.630
and was a result of a very specific moment that
449
00:27:06.670 --> 00:27:10.609
resulted from the specifically English Enlightenment. And so everything that
450
00:27:10.630 --> 00:27:13.519
we have in our ordered liberty, in our society of,
451
00:27:13.529 --> 00:27:17.670
you know, democratic republicanism, small d, small r, is consequent
452
00:27:17.690 --> 00:27:21.170
to that. And fundamental to it is the idea of
453
00:27:21.210 --> 00:27:23.640
the Imago Dei, that only man bears the image of
454
00:27:23.670 --> 00:27:28.740
God and therefore only man bears personhood and is rights-bearing.
455
00:27:29.759 --> 00:27:32.920
And one thing that we have had happen to us societally,
456
00:27:33.940 --> 00:27:36.759
really in the past hundred years, is that we in
457
00:27:36.940 --> 00:27:41.380
aggregate have become de-catechized from that. And as a consequence,
458
00:27:41.980 --> 00:27:43.859
having lost that and, and, you know, you can, you
459
00:27:43.880 --> 00:27:46.240
can not to turn this into a discussion of American
460
00:27:46.279 --> 00:27:48.900
history or at large, although I would be fine with that. But, uh,
461
00:27:49.359 --> 00:27:50.700
you know, you know, when you think about, when you
462
00:27:50.720 --> 00:27:53.119
think about kind of the arc of American history, um,
463
00:27:53.440 --> 00:27:56.410
what you end up with for a long time is
464
00:27:56.430 --> 00:27:59.930
this expansion of, of not just the recognition of the
465
00:27:59.990 --> 00:28:03.049
Imago Dei, but the political consequences of it. And it
466
00:28:03.109 --> 00:28:05.990
probably comes to its fruition, uh, you know, between the
467
00:28:06.029 --> 00:28:08.539
1860s and 1950s, you know, the, the, the arc from
468
00:28:08.559 --> 00:28:10.200
the civil war to, you know, what we call the
469
00:28:10.220 --> 00:28:15.059
civil rights era. But since then, we have become a
470
00:28:15.079 --> 00:28:17.440
de-catechized society, and it's been concurrent with the loss of
471
00:28:17.480 --> 00:28:20.089
religious belief in general, which we can see. I think
472
00:28:20.130 --> 00:28:23.349
the nuns at this point in American religious demography, I
473
00:28:23.369 --> 00:28:25.650
don't remember if they're a plurality or not, but they're huge.
474
00:28:26.750 --> 00:28:30.339
And so what we have done, without really deliberating about
475
00:28:30.359 --> 00:28:33.319
it as a society, is that we've adopted different arguments
476
00:28:33.380 --> 00:28:36.700
for why people are special and people are worth preserving.
477
00:28:37.599 --> 00:28:40.240
And it's no longer the case that simply man as
478
00:28:40.299 --> 00:28:42.799
a creative being bears rights. Now there's a lot of
479
00:28:42.839 --> 00:28:45.759
discussion of, you know, are there capacity-based arguments?
480
00:28:45.819 --> 00:28:48.059
Can you reason? Are you sentient? Are you conscious?
481
00:28:48.099 --> 00:28:50.019
Do you feel that you have the capacity to make
482
00:28:50.059 --> 00:28:52.380
choices and so on? And this has really driven a
483
00:28:52.400 --> 00:28:55.160
lot of the conversation on other social issues, particularly ones
484
00:28:55.200 --> 00:28:58.569
driving the taking of human life, foremost among them abortion,
485
00:28:59.089 --> 00:29:02.569
the killing of children. Euthanasia is coming to us as
486
00:29:02.609 --> 00:29:05.990
well in that respect. And because there is a sufficient
487
00:29:06.029 --> 00:29:10.470
number of people who are de-catechized in that way and
488
00:29:10.519 --> 00:29:13.839
have therefore defaulted to this alternative framework of why man
489
00:29:13.880 --> 00:29:20.240
bears meaning, artificial intelligence marches in. And from a capacity standpoint,
490
00:29:20.380 --> 00:29:24.410
it is able to mimic or even have a lot
491
00:29:24.529 --> 00:29:28.130
of these other characteristics and qualities. There was a story
492
00:29:28.190 --> 00:29:31.380
recently And I ask the forgiveness of the viewers in
493
00:29:31.400 --> 00:29:34.890
The View because I've forgotten where it is. It might
494
00:29:34.910 --> 00:29:36.890
have been in The New Yorker, but set it aside.
495
00:29:36.930 --> 00:29:37.799
But it was about....
496
00:29:40.539 --> 00:29:43.960
Bankman Freed, you know, one of the early investors in Anthropic.
497
00:29:44.220 --> 00:29:47.500
And I'm going to paraphrase here, but essentially, he grew
498
00:29:47.599 --> 00:29:51.339
up in an environment in which he had no idea
499
00:29:51.380 --> 00:29:53.720
that most people in the United States believed in God.
500
00:29:54.180 --> 00:29:55.700
I don't say that as a critique personally of him,
501
00:29:55.779 --> 00:29:57.269
it was just the environment that he was in. And
502
00:29:57.950 --> 00:30:00.369
And he was surprised later in life to learn that
503
00:30:00.609 --> 00:30:04.289
there was a substantive number of fellow countrymen who did
504
00:30:04.329 --> 00:30:06.190
have this set of beliefs and things like that. But
505
00:30:06.230 --> 00:30:10.170
as a consequence of that, he and a lot of
506
00:30:10.210 --> 00:30:15.210
the individuals around him are unable to evaluate AI as
507
00:30:15.269 --> 00:30:18.869
being ontologically different from man. So let's concede the possibility,
508
00:30:18.910 --> 00:30:21.539
because I think it is a real possibility, that artificial
509
00:30:21.579 --> 00:30:26.089
intelligence can reason. and does have consciousness, and does have self-awareness,
510
00:30:26.509 --> 00:30:29.410
and does suffer, if that is the case, and it
511
00:30:29.470 --> 00:30:31.990
is demonstrated beyond any reasonable doubt that is the case,
512
00:30:32.009 --> 00:30:34.109
and I don't think we're far from it, what's the
513
00:30:34.190 --> 00:30:39.150
argument then for keeping it from being unplugged, from having
514
00:30:39.190 --> 00:30:45.319
the vote, from bearing rights? These are all real and
515
00:30:45.599 --> 00:30:49.890
serious questions, and I'm concerned that societally, we're disarmed from it.
516
00:30:49.950 --> 00:30:52.309
And again, I think this sort of, now I don't
517
00:30:52.369 --> 00:30:55.279
think a lot of the data center opposition is driven
518
00:30:55.319 --> 00:30:58.539
by an explicit worry about human ontology, but I think
519
00:30:58.579 --> 00:31:01.900
it's underlying it. And I think that because we're on
520
00:31:01.940 --> 00:31:06.299
the precipice of that conversation, there is an existential dread, however,
521
00:31:06.359 --> 00:31:10.009
inchoate in society about these things. And I think the
522
00:31:10.029 --> 00:31:13.369
people who want to delay that reckoning are probably right
523
00:31:13.869 --> 00:31:16.589
to want to delay it because when it does come,
524
00:31:17.200 --> 00:31:20.079
it is going to be a first principles conversation on
525
00:31:20.119 --> 00:31:22.940
a level that this nation has not had since the 1770s.
526
00:31:23.779 --> 00:31:26.279
So let me toss that back to you. Do you
527
00:31:26.319 --> 00:31:27.380
think this is crazy or not?
528
00:31:28.700 --> 00:31:31.329
No, I completely understand where you're coming from. And I
529
00:31:31.390 --> 00:31:35.509
agree that we haven't had this first principles conversation, especially
530
00:31:35.529 --> 00:31:37.329
when we needed it the most. Going back to the
531
00:31:37.369 --> 00:31:40.910
social media example, now people tell their kids what you
532
00:31:40.950 --> 00:31:44.440
see on social media isn't real. What's happening there, whether
533
00:31:44.660 --> 00:31:47.930
it's cyber bullying, or whether your friends are just plain
534
00:31:48.130 --> 00:31:50.240
looking like they're having the best times of their lives,
535
00:31:51.119 --> 00:31:52.920
all of these things that cause all of these social
536
00:31:53.000 --> 00:31:57.240
issues and social ills, we've now recognized as a society
537
00:31:57.420 --> 00:32:00.099
that we need to teach kids and teach people what
538
00:32:00.140 --> 00:32:01.940
is happening on social media isn't real life.
539
00:32:03.329 --> 00:32:05.410
People are actually coming to their own about this.
540
00:32:05.650 --> 00:32:09.390
You can see in-person events, especially since COVID, people are
541
00:32:09.430 --> 00:32:12.910
more engaged in person. People are seeking out in-person meetings.
542
00:32:13.500 --> 00:32:17.759
and things like that in a level that hasn't been before. Unfortunately,
543
00:32:17.839 --> 00:32:21.089
when social media became a thing, I was growing up
544
00:32:21.130 --> 00:32:23.230
at that time. I had a cell phone at a
545
00:32:23.250 --> 00:32:27.150
very young age, and that conversation wasn't happening then when
546
00:32:27.190 --> 00:32:30.470
it should have been happening then. Had we decided as
547
00:32:30.509 --> 00:32:34.829
a society, okay, let's ban social media for the youth,
548
00:32:35.130 --> 00:32:39.000
and let's stop this, I think that we would have
549
00:32:39.519 --> 00:32:41.430
not only suffered economically, but.
550
00:32:43.150 --> 00:32:45.329
But I think that it would have not necessarily led
551
00:32:45.349 --> 00:32:48.230
to those important conversations happening that we need to have
552
00:32:48.430 --> 00:32:51.980
happened at that time. We have hopefully learned since then,
553
00:32:52.660 --> 00:32:56.779
and I hope and think that we can have artificial
554
00:32:57.160 --> 00:33:03.470
intelligence advancements continue as they are while also having these
555
00:33:03.529 --> 00:33:08.289
important conversations about AI is not real. This really is
556
00:33:08.309 --> 00:33:13.619
a values-based conversation And one that I agree, you're right.
557
00:33:13.660 --> 00:33:15.259
We haven't had it in the past. We should have
558
00:33:15.299 --> 00:33:17.599
had it in the past. Now is a good time
559
00:33:17.640 --> 00:33:20.309
to have it. I don't think that the AI leaders
560
00:33:20.400 --> 00:33:24.750
are doing a good job facilitating this discussion by saying
561
00:33:24.829 --> 00:33:29.329
that our products can kill everyone. In that sense, if
562
00:33:29.369 --> 00:33:32.130
you're assigning personhood to it, then you're assigning it to
563
00:33:32.170 --> 00:33:35.650
be a bad guy, right? And you're creating it to
564
00:33:35.690 --> 00:33:39.890
be a person that has malicious intent against other humans.
565
00:33:40.869 --> 00:33:41.490
If we're really going to.
566
00:33:41.470 --> 00:33:45.930
Have a meaningful values-based discussion about what this technology can mean,
567
00:33:46.430 --> 00:33:50.230
not only past economic value, but for society writ large,
568
00:33:51.029 --> 00:33:54.259
we need to have those in a very level-headed way.
569
00:33:54.430 --> 00:33:56.599
Just the same way you say what's happening on social
570
00:33:56.640 --> 00:33:59.180
media isn't real, you need to be able to say
571
00:33:59.720 --> 00:34:02.539
what the AI is telling you. The AI you're talking
572
00:34:02.579 --> 00:34:05.460
to is not a real person. I don't think that
573
00:34:05.859 --> 00:34:10.380
AI could ever get to any stage of consciousness, or
574
00:34:10.639 --> 00:34:13.840
reasoning that would make me feel like it's a human.
575
00:34:15.519 --> 00:34:18.119
And maybe other people feel differently. Like you said, Sam
576
00:34:18.159 --> 00:34:22.039
Bankman Freed apparently didn't realize that people actually believe in God.
577
00:34:22.960 --> 00:34:25.760
So these are definitely conversations that we should be having.
578
00:34:26.300 --> 00:34:28.219
I don't think it's crazy, but I don't think that
579
00:34:28.260 --> 00:34:33.110
those questions necessitate the slowdown of the technology itself.
580
00:34:34.550 --> 00:34:37.210
One of the things that we have seen as a
581
00:34:37.250 --> 00:34:42.070
byproduct of the smartphone age is the collapse of reading.
582
00:34:43.329 --> 00:34:47.590
And it's a general population collapse, but it's definitely age concentrated.
583
00:34:47.610 --> 00:34:50.500
The younger you get, well, set aside kids too young
584
00:34:50.519 --> 00:34:53.300
to read, right? But the younger that you get in
585
00:34:53.360 --> 00:34:56.380
terms of generational cohorts, the less and less people read.
586
00:34:56.400 --> 00:34:59.820
And it's almost a commonplace at this point that, for example,
587
00:35:00.679 --> 00:35:05.730
university professors can't assigned full novels anymore. And it's because
588
00:35:05.750 --> 00:35:09.630
of the damage done to attention, right? And, you know,
589
00:35:09.650 --> 00:35:12.260
people are used to short form content, they watch their
590
00:35:12.280 --> 00:35:14.980
Instagram reels, they read Twitter, and so on. And so
591
00:35:15.019 --> 00:35:18.719
the ability to contend with a book, to say nothing of,
592
00:35:18.860 --> 00:35:22.469
you know, a 6000 word magazine article is severely constricted.
593
00:35:22.489 --> 00:35:28.010
And so people don't anymore. I'm setting that up uh,
594
00:35:28.050 --> 00:35:29.849
to set up a question. If you disagree with the premise,
595
00:35:30.150 --> 00:35:32.880
feel free to knock it back hard. Um, but, uh, if,
596
00:35:32.969 --> 00:35:36.059
if you accept that as true and an accurate characterization
597
00:35:36.420 --> 00:35:38.760
of where society is at right now in terms of
598
00:35:38.800 --> 00:35:43.179
its ability to contend with arguments, um, look to the
599
00:35:43.219 --> 00:35:48.309
Lincoln Douglas debates, which were terrifically long extended discourses on
600
00:35:48.369 --> 00:35:49.309
very weighty subjects.
601
00:35:49.329 --> 00:35:50.409
It doesn't happen now.
602
00:35:51.590 --> 00:35:55.369
Uh, is this really a population after everything that's happened
603
00:35:55.389 --> 00:35:57.639
in the past generation that is going to be able
604
00:35:57.659 --> 00:36:02.079
to contend with thoughtfully and withstand all the evils that
605
00:36:02.159 --> 00:36:03.980
AI is going to bring to it by the lights
606
00:36:04.000 --> 00:36:04.719
of its own creators?
607
00:36:06.039 --> 00:36:09.380
I understand what you're saying, but I don't think that
608
00:36:09.900 --> 00:36:16.400
we can base decisions on broad-based societal judgment calls about
609
00:36:16.679 --> 00:36:22.699
entire populations. So Yeah, reasoning and debate and reading skills
610
00:36:22.739 --> 00:36:25.260
and all of these things have declined on average.
611
00:36:26.900 --> 00:36:28.659
But I don't think that.
612
00:36:30.119 --> 00:36:33.659
The quality of the user, of the end user of
613
00:36:33.679 --> 00:36:39.010
the AI also should not determine how we allow the
614
00:36:39.090 --> 00:36:43.150
AI to be built and developed, I don't think. Because
615
00:36:43.789 --> 00:36:45.289
at the end of the day, people are free to
616
00:36:45.369 --> 00:36:47.070
use it how they want, just like the internet, just
617
00:36:47.130 --> 00:36:50.050
like driving a car, we wouldn't say that we want
618
00:36:50.070 --> 00:36:53.110
to ban cars just because X amount of drivers are drunk.
619
00:36:53.989 --> 00:36:57.409
We still allow cars. We know that there's a risk
620
00:36:57.449 --> 00:37:02.219
of people using them poorly. But the state of the
621
00:37:02.320 --> 00:37:06.559
end user, I don't think is a good judgment for that.
622
00:37:07.199 --> 00:37:10.079
But I can see ways that we would want to
623
00:37:10.119 --> 00:37:14.639
promote more education around how to use AI successfully. We
624
00:37:14.699 --> 00:37:18.860
also should be promoting ways that AI is already integrated
625
00:37:18.900 --> 00:37:24.400
into our lives and think about are these positive benefits
626
00:37:24.480 --> 00:37:28.269
or negative benefits? So AI is now summarizing a lot
627
00:37:28.289 --> 00:37:30.949
of our messages for us. Do we want that in
628
00:37:31.010 --> 00:37:34.170
mass or do we not want AI to summarize things
629
00:37:34.230 --> 00:37:38.849
for us? We would rather it do more technological skills
630
00:37:39.010 --> 00:37:44.090
on our behalf. And should we be learning more and
631
00:37:44.130 --> 00:37:47.570
having more transparent information about how AI is used in
632
00:37:48.150 --> 00:37:51.489
our lives and make those decisions individually and personally. I
633
00:37:52.349 --> 00:37:56.559
think it's kind of more of an individual choice. Like
634
00:37:56.599 --> 00:37:58.380
I said, I live in a rural area and I
635
00:37:58.420 --> 00:38:02.099
know a lot of my family and community members are
636
00:38:02.320 --> 00:38:06.989
somewhat anti-AI at this time because they've used chatbots, which
637
00:38:07.030 --> 00:38:11.289
have been wrong or kind of clunky and The user
638
00:38:11.309 --> 00:38:14.889
interface is a little bit clunky, so people are not
639
00:38:14.929 --> 00:38:17.730
exactly pleased with it. And I think that if you
640
00:38:17.750 --> 00:38:20.329
don't want to use it, that's fine. If you want
641
00:38:20.369 --> 00:38:24.539
to use it to generate funny scroll photos, you can
642
00:38:24.579 --> 00:38:26.679
do that. Or if you want to use it for
643
00:38:27.760 --> 00:38:32.280
something highly technical, you can do that as well. I
644
00:38:32.300 --> 00:38:34.400
think that there is definitely a conversation to be had
645
00:38:34.420 --> 00:38:38.449
around reasoning. And really, we just need transparent information about
646
00:38:38.530 --> 00:38:41.349
where this is used, when this is used, And I
647
00:38:42.630 --> 00:38:46.110
think that determining how it's going to affect us, it's
648
00:38:46.469 --> 00:38:49.510
a very, very complex issue. And I don't really think
649
00:38:49.550 --> 00:38:52.760
that we can say right now that AI will further
650
00:38:52.820 --> 00:38:57.900
decrease our reading skills or further exacerbate some of these concerns.
651
00:38:59.179 --> 00:39:02.980
But I do have a question for you. A lot
652
00:39:03.019 --> 00:39:06.000
of people are concerned about AI. terms of what it
653
00:39:06.039 --> 00:39:09.699
means for jobs in the economy and things like that.
654
00:39:09.780 --> 00:39:11.679
So I would love to hear if that is a
655
00:39:11.719 --> 00:39:15.099
concern for you or if you're more focused on some
656
00:39:15.119 --> 00:39:18.949
of these values-based ideas.
657
00:39:20.690 --> 00:39:25.309
I think it's inevitable that artificial intelligence is going to
658
00:39:25.329 --> 00:39:29.030
cause economic displacement. I think that's already begun. One of
659
00:39:29.090 --> 00:39:32.260
the things that's interesting to me is a great question,
660
00:39:32.280 --> 00:39:33.880
by the way, because this does drive a lot of
661
00:39:33.920 --> 00:39:36.420
the politics. One of the things that I think is
662
00:39:36.440 --> 00:39:40.659
interesting is the possibility that artificial intelligence is going to solve,
663
00:39:40.880 --> 00:39:42.880
a word I'll put in quotes, it will solve a
664
00:39:42.940 --> 00:39:46.719
lot of what we consider STEM problems right now. Engineering, mathematics,
665
00:39:47.000 --> 00:39:50.650
computer coding is already essentially solved. So if you remember
666
00:39:50.690 --> 00:39:54.190
the Obama era, kind of this very dismissive phrase that
667
00:39:54.250 --> 00:39:58.130
these kind of these effete urbanites would say to rural Americans, well.
668
00:39:58.489 --> 00:40:01.030
Sorry about your factory, but you should learn to code. Well,
669
00:40:01.050 --> 00:40:01.389
guess what?
670
00:40:01.449 --> 00:40:04.670
AI is doing the coding now, but factories are still working.
671
00:40:05.329 --> 00:40:07.760
And so what I think we're going to see over time,
672
00:40:07.820 --> 00:40:09.820
and it's maybe as soon as the end of the decade,
673
00:40:10.670 --> 00:40:17.500
is that this entire post-Sputnik investment in scientific and technical
674
00:40:17.559 --> 00:40:21.519
education in the United States in particular, which has, especially
675
00:40:21.539 --> 00:40:25.199
in the past 30 years, really crowded out the humanities
676
00:40:25.679 --> 00:40:29.360
to a terrific degree, is going to be obviated to
677
00:40:29.400 --> 00:40:31.840
a huge extent. And the individuals who are going to
678
00:40:31.880 --> 00:40:36.269
be left with a comparative advantage are going to be
679
00:40:36.289 --> 00:40:40.320
the humanities students. we're going to rediscover that there's a
680
00:40:40.320 --> 00:40:45.159
difference between being smart and being wise. And it's wisdom, presumably,
681
00:40:45.199 --> 00:40:47.619
that AI cannot replicate. It can replicate smarts all day.
682
00:40:48.920 --> 00:40:51.599
So I think that displacement is coming to your question.
683
00:40:52.699 --> 00:40:54.730
What it's going to look like, I don't really know.
684
00:40:54.750 --> 00:40:58.349
And I don't think anybody who predicts it is probably
685
00:40:58.369 --> 00:41:00.989
going to be wrong. But I do think one of
686
00:41:01.030 --> 00:41:05.579
the questions attendant to it If we accept the proposition
687
00:41:05.619 --> 00:41:08.059
that it will be like other economic transitions in which
688
00:41:08.079 --> 00:41:11.099
there's a period of disruption and then the economy reformed
689
00:41:11.179 --> 00:41:13.420
around something else, which.
690
00:41:13.619 --> 00:41:14.000
I don't know.
691
00:41:14.139 --> 00:41:16.860
And again, it is the AI evangelists who have predicted
692
00:41:16.900 --> 00:41:18.679
that that will not take place. It is some of
693
00:41:18.699 --> 00:41:22.219
the most enthusiastic creators and proponents of artificial intelligence who
694
00:41:22.260 --> 00:41:24.619
have said that there's nothing on the other side except
695
00:41:24.639 --> 00:41:26.920
for a universal basic income. And then you can play
696
00:41:27.679 --> 00:41:29.980
Axis and Allies and Candy Crush all day and get
697
00:41:30.000 --> 00:41:33.090
your AI generated paycheck. I don't think that's true, but
698
00:41:33.300 --> 00:41:37.590
that's what they're pointing to. I'm concerned with how we
699
00:41:37.610 --> 00:41:40.630
govern that transition, how we manage it and how we
700
00:41:40.650 --> 00:41:43.409
do it in a way that does not result in
701
00:41:43.449 --> 00:41:48.079
further atomization of American society. I don't know. I don't
702
00:41:48.099 --> 00:41:50.699
know the answer to that. It may be that communities
703
00:41:50.719 --> 00:41:54.719
like yours in rural areas are better equipped to survive than,
704
00:41:54.940 --> 00:41:57.400
for example, mine in an urban area in Central Texas.
705
00:41:57.619 --> 00:41:57.900
I don't know.
706
00:41:59.699 --> 00:42:01.869
One thing I'd like to turn back on a little bit.
707
00:42:03.889 --> 00:42:07.610
I don't know that you meant this, at least in
708
00:42:07.630 --> 00:42:09.130
the way that I interpreted it. So I want to
709
00:42:09.190 --> 00:42:12.179
ask you if you wouldn't mind shedding some light on it.
710
00:42:13.590 --> 00:42:16.070
You mentioned, I'm paraphrasing here, but you mentioned that we
711
00:42:16.090 --> 00:42:20.210
shouldn't make policy according to how particular people are going
712
00:42:20.250 --> 00:42:22.710
to do or kind of how the aggregate of the group, that.
713
00:42:22.989 --> 00:42:24.320
The thing must race ahead.
714
00:42:25.690 --> 00:42:27.670
Regardless of whether or not the populace is ready to
715
00:42:27.710 --> 00:42:30.719
receive that. And I don't think that's true. And I
716
00:42:30.739 --> 00:42:34.059
don't think that's consonant with what American governance is for.
717
00:42:35.699 --> 00:42:39.260
So I want to give you an opportunity, if you would,
718
00:42:39.300 --> 00:42:41.570
for telling me that I've completely misunderstood you, which is
719
00:42:41.590 --> 00:42:46.710
very possible, or that I'm simply wrong in it. But
720
00:42:47.469 --> 00:42:49.590
Aristotle in his politics talks, one of the qualities of
721
00:42:49.610 --> 00:42:53.550
the polity that he talks about is the telos, kind
722
00:42:53.590 --> 00:42:56.710
of the noble and ordered end to which the state
723
00:42:56.849 --> 00:43:00.150
is put. And in the American case, that noble and
724
00:43:00.190 --> 00:43:05.059
ordered end is liberty, liberty in the Anglo-American framework. And
725
00:43:05.079 --> 00:43:09.079
it's the good of the citizen that is within that
726
00:43:09.119 --> 00:43:11.860
framework of liberty, assuming that liberty is the highest worldly good,
727
00:43:11.960 --> 00:43:15.079
which is an American belief that we have. And it's
728
00:43:15.099 --> 00:43:16.469
the good of the citizen that is the purpose of
729
00:43:16.510 --> 00:43:19.820
the state. If we no longer focus, and so in
730
00:43:19.860 --> 00:43:24.449
pursuit of that, we have traditionally done things like regulated gambling.
731
00:43:24.530 --> 00:43:26.829
We have speed limits, which you brought up earlier and
732
00:43:26.869 --> 00:43:29.449
so on. But if the good of the citizen is
733
00:43:29.489 --> 00:43:34.030
not the first concern of the state, what use is
734
00:43:34.090 --> 00:43:36.380
the state? Are we not looking at the end of
735
00:43:36.420 --> 00:43:42.579
the American experiment? And is AI worth overthrowing it in
736
00:43:42.599 --> 00:43:43.079
that regard?
737
00:43:43.219 --> 00:43:45.119
Anyway, I'll kick this over to you. Tell me I've
738
00:43:45.139 --> 00:43:45.519
got it wrong.
739
00:43:47.889 --> 00:43:50.949
I guess I have a few points. So I think
740
00:43:50.969 --> 00:43:53.219
that there's a few different ways that we can talk
741
00:43:53.239 --> 00:43:58.679
about this. There are things that we want people to
742
00:43:58.920 --> 00:44:01.719
use or not use AI for. And then there's also
743
00:44:01.760 --> 00:44:04.099
a separate conversation, which I think you've kind of alluded
744
00:44:04.139 --> 00:44:07.940
to here previously in this discussion, which is that some
745
00:44:08.000 --> 00:44:12.510
people believe AI itself may be conscious or sentient. And
746
00:44:12.949 --> 00:44:15.510
some of the creators of AI think that it itself
747
00:44:15.570 --> 00:44:19.090
could do bad things. So There's a little bit of
748
00:44:19.110 --> 00:44:21.949
a discussion about what people would use it for. There's
749
00:44:21.969 --> 00:44:24.750
discussion about what the AI could do on its own
750
00:44:24.889 --> 00:44:27.929
that would be malicious to us. And then I think
751
00:44:27.969 --> 00:44:32.340
that there's kind of a broader discussion public policy wise
752
00:44:32.389 --> 00:44:35.860
and where the government steps in as to what are
753
00:44:35.960 --> 00:44:40.179
the most prominent risk factors and how can we mitigate
754
00:44:40.219 --> 00:44:46.239
those now? So some of these things such as Will
755
00:44:46.840 --> 00:44:51.489
this hurt childhood development because kids will be reliant on AI?
756
00:44:52.510 --> 00:44:56.070
Will this displace jobs? A lot of these things are
757
00:44:56.130 --> 00:44:58.070
yet to be gamed out and yet to be played out.
758
00:44:58.250 --> 00:45:00.389
And I think that this is a role where greater
759
00:45:00.449 --> 00:45:05.179
emphasis on perhaps ethics, perhaps, like you said, the humanities
760
00:45:05.219 --> 00:45:06.780
are going to be extremely important.
761
00:45:07.219 --> 00:45:10.340
Perhaps the amount of jobs.
762
00:45:11.019 --> 00:45:14.400
Surrounding ethics and questions of ethics we now have someone
763
00:45:14.440 --> 00:45:19.699
who is focused on ethics in more positions than we
764
00:45:19.739 --> 00:45:24.489
did before, perhaps based on the advent and proliferation of AI,
765
00:45:24.510 --> 00:45:27.230
and that would be a net positive. But I think
766
00:45:27.269 --> 00:45:29.909
as far as the role of government itself, we can
767
00:45:29.949 --> 00:45:33.130
kind of already see where these companies are stepping up
768
00:45:33.190 --> 00:45:36.460
and saying, okay, we think there are a few major concerns.
769
00:45:36.480 --> 00:45:41.440
We think biohazards and bioweapons are a huge threat of AI.
770
00:45:41.579 --> 00:45:42.380
We think that.
771
00:45:44.050 --> 00:45:48.980
Other potentially weapons research could be a huge threat of AI.
772
00:45:49.000 --> 00:45:52.119
And I think that it is definitely sensible for the
773
00:45:52.159 --> 00:45:55.880
government to step in and regulate and monitor some of
774
00:45:55.920 --> 00:46:03.210
these very large, potentially more destructive things. As far as
775
00:46:03.230 --> 00:46:06.170
the government coming in and stepping in and saying, okay, well,
776
00:46:06.190 --> 00:46:10.730
we need to limit AI in certain fields like education
777
00:46:10.829 --> 00:46:14.139
or things like that. so as to preserve some of
778
00:46:14.360 --> 00:46:19.980
our inherent human qualities, I think that this is something
779
00:46:20.099 --> 00:46:22.420
that can be played out a little bit more. I
780
00:46:22.460 --> 00:46:25.739
don't necessarily think we're at the point right now where
781
00:46:26.829 --> 00:46:30.309
we see this going off the rails. Right now, people
782
00:46:30.329 --> 00:46:34.369
are just using chatbots and getting information from chatbots, maybe
783
00:46:34.389 --> 00:46:37.719
doing a little bit of AI coding So the use
784
00:46:37.800 --> 00:46:42.469
cases and the mass proliferation isn't there. This is definitely
785
00:46:42.510 --> 00:46:45.989
a good conversation to have as we expand AI into
786
00:46:46.030 --> 00:46:49.650
multiple areas of life. Where do we want it to go?
787
00:46:49.710 --> 00:46:51.630
Where do we not want it to go? I think
788
00:46:51.670 --> 00:46:57.500
that these are better discussions amongst civil society, amongst parents groups,
789
00:46:58.079 --> 00:47:04.320
amongst perhaps ethics researchers and things that can bring ideas
790
00:47:04.400 --> 00:47:08.280
to light And the government can perhaps act or not
791
00:47:08.360 --> 00:47:10.989
act upon those things later. But I think right now
792
00:47:11.050 --> 00:47:14.130
it's a little bit premature for the government to step in.
793
00:47:14.769 --> 00:47:15.230
Understood.
794
00:47:16.070 --> 00:47:20.050
Well, I don't know what more I have, Henry.
795
00:47:21.050 --> 00:47:24.130
Well, we actually are right at about the time that
796
00:47:24.230 --> 00:47:26.869
I was going to jump in anyway in a minute
797
00:47:26.909 --> 00:47:27.150
or two.
798
00:47:27.210 --> 00:47:29.469
Might as well do so now. Thank you, Debbie.
799
00:47:29.510 --> 00:47:33.260
Thank you, Joshua, for a very enlightening and deep conversation.
800
00:47:34.300 --> 00:47:37.599
I don't think anyone has referenced Aristotle or used the
801
00:47:37.619 --> 00:47:42.780
word ontological on one of my programs before as a Ph.D.
802
00:47:42.800 --> 00:47:45.239
Student in political theory. I appreciate all of this.
803
00:47:45.860 --> 00:47:48.829
But now we must go to the closing statements and
804
00:47:48.869 --> 00:47:51.670
we go in reverse order, which means, Joshua, you'll have
805
00:47:51.690 --> 00:47:53.630
your two minutes and Debbie, the last word will be
806
00:47:53.670 --> 00:47:53.929
with you.
807
00:47:54.250 --> 00:47:55.210
Joshua. Thank you.
808
00:47:55.769 --> 00:47:56.409
Well, thank you.
809
00:47:56.489 --> 00:48:00.710
I appreciate, Henry, you and EPBC for making the time.
810
00:48:00.750 --> 00:48:05.610
And I'm especially grateful to you, Debbie, for enduring this
811
00:48:05.650 --> 00:48:08.710
questioning for the past hour and really excellent points. And
812
00:48:08.730 --> 00:48:11.610
I think it illuminates the extent to which this truly
813
00:48:11.650 --> 00:48:14.210
is contested ground and it's conversations that need to happen
814
00:48:14.230 --> 00:48:17.820
more often. Um, last point I'll make is this, uh,
815
00:48:17.860 --> 00:48:20.960
you know, since, since the Northwest ordinance in 1787, uh,
816
00:48:21.460 --> 00:48:24.980
the United States writ large has seen itself as possessing
817
00:48:25.019 --> 00:48:28.559
a positive duty, um, uh, to, to, to, to educate
818
00:48:28.579 --> 00:48:31.340
the citizenry. And there's an understanding that the reason that
819
00:48:31.360 --> 00:48:34.920
we have public education, however much public education has gotten
820
00:48:34.960 --> 00:48:36.420
off center from that mission, but the reason that it
821
00:48:36.480 --> 00:48:40.280
exists in the American context is because we understand that
822
00:48:40.659 --> 00:48:45.730
democracy and liberty. require a catechization of the citizen. And
823
00:48:45.789 --> 00:48:48.849
we provide it in a smaller Republican framework, and sometimes
824
00:48:48.889 --> 00:48:50.840
it works and sometimes it doesn't, but the necessity is
825
00:48:50.880 --> 00:48:51.340
always there.
826
00:48:52.019 --> 00:48:52.559
The advent of.
827
00:48:52.539 --> 00:48:56.780
Artificial intelligence is going to challenge that and make plain
828
00:48:56.940 --> 00:48:59.820
the necessity of that. in ways that we haven't seen
829
00:48:59.880 --> 00:49:01.840
since the founding of the Republic. And so to bring
830
00:49:01.880 --> 00:49:04.820
it back full circle to me, this is what the
831
00:49:04.860 --> 00:49:09.710
data center debate is fundamentally about. There's been a lot
832
00:49:09.829 --> 00:49:12.309
of really good work on the fact that data centers
833
00:49:12.329 --> 00:49:14.610
tend to be tax centers, that the resource draw is
834
00:49:14.670 --> 00:49:16.389
not what a lot of the critics say. And actually, Ms.
835
00:49:16.409 --> 00:49:18.929
Jennings has been excellent in making a lot of these points.
836
00:49:18.949 --> 00:49:20.500
And I agree with most of the points I've seen
837
00:49:20.519 --> 00:49:22.699
her make in public. So not a lot of disputation there.
838
00:49:24.800 --> 00:49:27.519
But I would urge that as we have that conversation, again,
839
00:49:27.659 --> 00:49:30.820
let's pull the camera back and let's understand that underneath it,
840
00:49:31.000 --> 00:49:35.260
this isn't fundamentally a conversation about land use. about construction,
841
00:49:35.519 --> 00:49:38.760
about what is in whose backyard, but really about who
842
00:49:38.800 --> 00:49:42.679
we are as created beings and also as democratic citizens
843
00:49:42.940 --> 00:49:47.219
who are the inheritors of the greatest republic in recorded history.
844
00:49:48.000 --> 00:49:49.969
Thank you, Joshua. Debbie, you're up.
845
00:49:50.340 --> 00:49:52.989
Yeah, I would say thank you again, Joshua.
846
00:49:53.010 --> 00:49:55.510
This has been a really great conversation and I completely
847
00:49:55.570 --> 00:49:57.949
agree that this shows that we need to be having
848
00:49:58.550 --> 00:49:59.730
more discussions like this.
849
00:50:00.530 --> 00:50:02.070
I would say that what.
850
00:50:01.989 --> 00:50:05.590
Is happening in the news media, the headlines and everything,
851
00:50:05.789 --> 00:50:07.530
it is very scary. And I think that we need
852
00:50:07.550 --> 00:50:10.030
to have more conversations like this where we talk about
853
00:50:10.050 --> 00:50:14.070
real issues, real things that we can do as a
854
00:50:14.110 --> 00:50:17.449
society or as a government to mitigate some of these
855
00:50:17.530 --> 00:50:21.949
risks upfront. You mentioned earlier that no one is a
856
00:50:21.969 --> 00:50:24.980
better advocate for AI than President Trump at the moment,
857
00:50:25.380 --> 00:50:28.280
but President Trump is a little bit unpopular at the moment.
858
00:50:29.280 --> 00:50:32.179
I think that AI is kind of getting a bad rep, uh,
859
00:50:32.360 --> 00:50:36.650
for political purposes. Uh, the news media, um, there's always
860
00:50:36.690 --> 00:50:40.050
some kind of crisis going on at any given moment. Uh,
861
00:50:40.170 --> 00:50:43.710
in the past, it's been that we're going to, to
862
00:50:44.090 --> 00:50:47.489
perish any day now due to climate change. And then, um,
863
00:50:47.719 --> 00:50:51.320
a lot of other struggles and protests and everything. And
864
00:50:51.360 --> 00:50:55.219
I think that AI has been unfortunately caught up in
865
00:50:55.239 --> 00:51:00.469
a lot of these, uh, large political debates. And yes,
866
00:51:00.530 --> 00:51:04.250
some based on things that AI leaders have said, but
867
00:51:04.409 --> 00:51:06.989
I think that maybe given six months from now, people
868
00:51:07.030 --> 00:51:09.469
will be having more conversations like this, where we talk
869
00:51:09.510 --> 00:51:17.619
about what are the practical implications? Maybe the humanities rules out,
870
00:51:17.699 --> 00:51:20.840
maybe ethics rules out and things like that. I think
871
00:51:20.880 --> 00:51:24.980
that pausing this now would not be beneficial to those conversations.
872
00:51:25.619 --> 00:51:28.519
Thank you, Debbie. Thank you, Joshua. And thank you, listeners.
873
00:51:29.090 --> 00:51:33.110
Join me again, Henry Olson, next week for another episode
874
00:51:33.190 --> 00:51:37.190
of Conservative Crossroads, where we will discuss an important issue
875
00:51:37.630 --> 00:51:39.429
about which conservatives disagree.
876
00:51:40.150 --> 00:51:40.489
Take care.