1
00:00:02,360 --> 00:00:05,200
Speaker 1: Welcome back to a Numbers Game with Ryan Gardowski. Thank

2
00:00:05,200 --> 00:00:07,320
Speaker 1: you all for being here on this Thursday episode. A

3
00:00:07,360 --> 00:00:09,480
Speaker 1: lot of news broke since Monday, so let me give

4
00:00:09,480 --> 00:00:12,240
Speaker 1: you guys some quick hits before talking with the main topic.

5
00:00:12,320 --> 00:00:14,520
Speaker 1: You should know a little bit of news to keep

6
00:00:14,560 --> 00:00:17,520
Speaker 1: you more informed than the average person. In Poland, the

7
00:00:17,600 --> 00:00:21,079
Speaker 1: Nationalist Party candidate, the Law and Justice Party nominee for president,

8
00:00:21,640 --> 00:00:25,520
Speaker 1: Carol Norwiki, won the presidency in an absolute come from

9
00:00:25,520 --> 00:00:29,720
Speaker 1: behind victory. Law and Justice is a nationalist party in Europe.

10
00:00:29,760 --> 00:00:31,520
Speaker 1: Not my favorite because they do a lot of like

11
00:00:31,960 --> 00:00:35,320
Speaker 1: public outreach that's not really They come out as a

12
00:00:35,400 --> 00:00:39,840
Speaker 1: very anti immigration policy immigrant immigrant politicians and political party,

13
00:00:39,880 --> 00:00:42,239
Speaker 1: but they're very pro immigrant. Really, it's a lot of

14
00:00:42,240 --> 00:00:47,200
Speaker 1: pr that's bigger than the actual policy anyway. Norwicky's win, though,

15
00:00:47,440 --> 00:00:51,200
Speaker 1: is notable because he was double digits behind the polls

16
00:00:51,240 --> 00:00:55,640
Speaker 1: as recently in April and had an absolute monster comeback,

17
00:00:55,640 --> 00:00:59,000
Speaker 1: and it dispels the myth that Trump is so toxic

18
00:00:59,040 --> 00:01:01,760
Speaker 1: to nationalist and pop those candidates around the globe. This

19
00:01:01,840 --> 00:01:04,200
Speaker 1: is the third straight presidential election with the Law and

20
00:01:04,360 --> 00:01:07,360
Speaker 1: Justice Party one in Poland and on the other side

21
00:01:07,360 --> 00:01:11,800
Speaker 1: of Europe, National's populace firebrand Gets Wilders removed his party

22
00:01:11,800 --> 00:01:14,960
Speaker 1: from the coalition government and it collapsed the coalition. It

23
00:01:14,959 --> 00:01:18,440
Speaker 1: will force early elections later on this year. Here's how

24
00:01:18,480 --> 00:01:22,240
Speaker 1: that went down. So the Netherlands is a multi party system.

25
00:01:22,319 --> 00:01:25,119
Speaker 1: They have lots of parties. I'm talking they have more

26
00:01:25,160 --> 00:01:27,560
Speaker 1: parties than gen Z has Genders like. It is a

27
00:01:27,600 --> 00:01:30,760
Speaker 1: lot of parties. In twenty twenty three, Gets Wilders, who

28
00:01:30,800 --> 00:01:34,440
Speaker 1: is the longest serving Dutch politician and always a political

29
00:01:34,480 --> 00:01:37,680
Speaker 1: outsider for being a hardliner against mass immigration and the

30
00:01:37,720 --> 00:01:41,400
Speaker 1: Islamification of the Netherlands, and had his party, the Freedom Party,

31
00:01:41,720 --> 00:01:45,080
Speaker 1: surprise everyone in come in first place. But in order

32
00:01:45,080 --> 00:01:48,200
Speaker 1: to form a government you need seventy six seats in parliament.

33
00:01:48,320 --> 00:01:51,040
Speaker 1: His party only had thirty seven. So we entered a

34
00:01:51,040 --> 00:01:54,600
Speaker 1: coalition government with three other center right parties and populist

35
00:01:54,640 --> 00:01:57,920
Speaker 1: parties on the condition that he could not be prime minister,

36
00:01:58,240 --> 00:02:02,160
Speaker 1: but they would concede his demands and immigration. Eleven months

37
00:02:02,240 --> 00:02:05,960
Speaker 1: later the coalition. After the coalition government was formed, Fielders

38
00:02:06,320 --> 00:02:09,239
Speaker 1: has decided to leave the coalition and trigger early elections

39
00:02:09,280 --> 00:02:12,359
Speaker 1: because the other three parties refused to move forward on

40
00:02:12,400 --> 00:02:15,880
Speaker 1: what he called the strictest asylum policies in Europe. This

41
00:02:15,960 --> 00:02:18,720
Speaker 1: is not the first time Builders has lost left the

42
00:02:18,720 --> 00:02:22,959
Speaker 1: coalition government created a snap election. In twenty ten, he

43
00:02:23,480 --> 00:02:26,799
Speaker 1: pulled a similar move because the coalition government moved forward

44
00:02:26,800 --> 00:02:31,440
Speaker 1: on austerity measures. The result was a political instability and

45
00:02:31,480 --> 00:02:34,800
Speaker 1: Wilder's party was punished in the next election, even though

46
00:02:34,880 --> 00:02:39,359
Speaker 1: austerity measures were unpopular. Because if there's anything you should

47
00:02:39,400 --> 00:02:42,040
Speaker 1: know about voters in all of the West, regardless of

48
00:02:42,080 --> 00:02:46,160
Speaker 1: what country it is, they punish in political instability more

49
00:02:46,160 --> 00:02:49,520
Speaker 1: than anything else. Now, whether it be Builders leaving the

50
00:02:49,520 --> 00:02:53,080
Speaker 1: coalition government or Republicans shutting down the government or Democrats

51
00:02:53,120 --> 00:02:57,400
Speaker 1: passing Obamacare, people don't like to make it the feeling

52
00:02:57,560 --> 00:03:01,160
Speaker 1: of being unstable in their politics. So we'll see what

53
00:03:01,200 --> 00:03:04,360
Speaker 1: happens in the next election. If Wilders's party can survive

54
00:03:04,480 --> 00:03:07,640
Speaker 1: and grow, I don't know. I'm pessimistic on that. Stranger

55
00:03:07,680 --> 00:03:10,200
Speaker 1: things have happened. Maybe immigration is a big enough issue

56
00:03:10,240 --> 00:03:13,720
Speaker 1: that voters won't mind, but it's something worth keeping an

57
00:03:13,720 --> 00:03:15,720
Speaker 1: eye on. It's all the politics I have for Europe.

58
00:03:15,760 --> 00:03:18,560
Speaker 1: I love talking about European politics, I love foreign politics.

59
00:03:18,600 --> 00:03:20,840
Speaker 1: I know it's not for everybody. I try to make

60
00:03:20,880 --> 00:03:23,680
Speaker 1: sure that you understand how it's connects to American politics.

61
00:03:23,760 --> 00:03:26,600
Speaker 1: But if you're okay with me doing more episodes on

62
00:03:26,919 --> 00:03:30,280
Speaker 1: other countries, particularly in Europe's politics, let me know. I'm

63
00:03:30,320 --> 00:03:33,040
Speaker 1: building the show with you for you, so your feedback

64
00:03:33,120 --> 00:03:35,840
Speaker 1: is very important. I read every email. Emailm me Ryan

65
00:03:35,880 --> 00:03:38,960
Speaker 1: at numbers gamepodcast dot com and shoot me either an

66
00:03:39,000 --> 00:03:40,680
Speaker 1: idea for an episode or if you want to hear

67
00:03:40,720 --> 00:03:42,920
Speaker 1: more about European politics, I'd love to do an episode

68
00:03:42,920 --> 00:03:46,880
Speaker 1: about it. Okay, Now to the States. Elon Musk has

69
00:03:47,000 --> 00:03:49,360
Speaker 1: officially left the White House and DOGE is all but

70
00:03:49,480 --> 00:03:53,480
Speaker 1: dead after just a few months of being inactive. And

71
00:03:53,560 --> 00:03:56,480
Speaker 1: I told people that I didn't think what Elon was

72
00:03:56,480 --> 00:03:58,480
Speaker 1: trying to do was what he said he was going

73
00:03:58,560 --> 00:04:00,040
Speaker 1: to do. I don't think it was about trying to

74
00:04:00,080 --> 00:04:02,840
Speaker 1: balance the budget, but that's what it was sold at,

75
00:04:02,880 --> 00:04:06,040
Speaker 1: and that is what he's leaving on. On Tuesday, Elon

76
00:04:06,080 --> 00:04:10,360
Speaker 1: took a swipe at President Trump's big beautiful bill, tweeting quote,

77
00:04:10,400 --> 00:04:13,200
Speaker 1: I am sorry, but I can't stand it anymore. The massive,

78
00:04:13,280 --> 00:04:18,599
Speaker 1: outrageous Porkfield Congressional spending bill is a disgusting abomination. Shame

79
00:04:18,680 --> 00:04:20,919
Speaker 1: on those who voted for it. You know you did wrong.

80
00:04:21,080 --> 00:04:23,200
Speaker 1: You know it, and then he followed up by saying

81
00:04:23,240 --> 00:04:26,760
Speaker 1: it will massively increase the already gigantic budget deficit to

82
00:04:26,839 --> 00:04:30,680
Speaker 1: two point five trillion and burden American citizens with crushingly

83
00:04:30,839 --> 00:04:36,039
Speaker 1: unstable debt. Elon has a point, right, not my favorite

84
00:04:36,040 --> 00:04:38,840
Speaker 1: personal world, but Elon has a point, and he has

85
00:04:38,880 --> 00:04:43,240
Speaker 1: been warring with members of this administration for months now,

86
00:04:43,800 --> 00:04:46,800
Speaker 1: according to Axeos Treasury Secretary Scout Besset, and he got

87
00:04:46,839 --> 00:04:50,240
Speaker 1: into his screaming match outside the Oval office, where Besset

88
00:04:50,320 --> 00:04:53,839
Speaker 1: called him a fraud for not finding two trillion dollars

89
00:04:53,920 --> 00:04:56,279
Speaker 1: dollars in wasteful spending. It he claimed he would find,

90
00:04:56,600 --> 00:05:00,720
Speaker 1: and to be fair, Beset, he's right. Elon promised brillions

91
00:05:00,800 --> 00:05:03,680
Speaker 1: in cuts without having any pain because he was going

92
00:05:03,720 --> 00:05:06,080
Speaker 1: to find it in a waste fraud abuse. He gave

93
00:05:06,240 --> 00:05:11,159
Speaker 1: Republicans cover to increase spending and then didn't deliver. It's

94
00:05:11,240 --> 00:05:15,560
Speaker 1: just calling balls and strikes. But here's the thing. Among Republicans,

95
00:05:15,839 --> 00:05:21,839
Speaker 1: especially grassroots donors, Elon is very popular. So I don't

96
00:05:21,880 --> 00:05:24,159
Speaker 1: think this is all about spending. I think there's a

97
00:05:24,200 --> 00:05:26,960
Speaker 1: lot of sour grapes between Elon and Trump that isn't

98
00:05:26,960 --> 00:05:30,080
Speaker 1: going away. But this part of this bill is going

99
00:05:30,120 --> 00:05:33,320
Speaker 1: to Elon's words about this bill are going to impact

100
00:05:33,360 --> 00:05:36,359
Speaker 1: the base of the Republican party. Parts of this bill

101
00:05:36,440 --> 00:05:40,400
Speaker 1: are very unpopular. I think it's still likely to pass,

102
00:05:40,560 --> 00:05:43,800
Speaker 1: given that almost all of Trump's legislative agenda is wrapped

103
00:05:43,880 --> 00:05:46,400
Speaker 1: up in this single bill, and Republicans can't afford not

104
00:05:46,480 --> 00:05:49,520
Speaker 1: to pass anything. But the one provision I want to

105
00:05:49,560 --> 00:05:52,359
Speaker 1: talk about, the one persition I'm thinking about, is on

106
00:05:52,400 --> 00:05:54,400
Speaker 1: page two hundred and seventy eight and to two hundred

107
00:05:54,440 --> 00:05:55,920
Speaker 1: and seventy nine of the bill. If you want to

108
00:05:55,920 --> 00:05:58,600
Speaker 1: go in the congressional website and read it, which up

109
00:05:58,680 --> 00:06:02,560
Speaker 1: till this week had almost nobody noticed. On page two

110
00:06:02,600 --> 00:06:05,320
Speaker 1: hundred and seventy eight, Congressional Republicans snuck in a ten

111
00:06:05,440 --> 00:06:10,479
Speaker 1: year moudatorium on states regulating artificial intelligence. The section reads,

112
00:06:10,480 --> 00:06:12,320
Speaker 1: and this is a bit long, but I want I

113
00:06:12,360 --> 00:06:16,080
Speaker 1: want to read the whole thing. It's important. Quote no

114
00:06:16,279 --> 00:06:19,960
Speaker 1: state or political subdivision thereof may enforce during the ten

115
00:06:20,040 --> 00:06:23,599
Speaker 1: year period beginning on the date of the enactment of

116
00:06:23,640 --> 00:06:28,120
Speaker 1: this Act any law or regulation of that state or

117
00:06:28,400 --> 00:06:34,799
Speaker 1: political subdivision thereof limiting, restricting, or otherwise regulating artificial intelligence models.

118
00:06:35,080 --> 00:06:40,680
Speaker 1: Artificial intelligence systems or automated decision systems entered into interstate commerce.

119
00:06:41,520 --> 00:06:45,760
Speaker 1: Paragraph one may not be construed as prohibited enforcement of

120
00:06:45,839 --> 00:06:49,880
Speaker 1: any law or regulation that the primary purpose and effect

121
00:06:50,000 --> 00:06:53,680
Speaker 1: of which is to remove legal impediments or facilitate the

122
00:06:53,760 --> 00:06:58,480
Speaker 1: development or operation of artificial intelligence models, artificial intelligence systems,

123
00:06:58,600 --> 00:07:05,560
Speaker 1: or automated decision system or streamline licensing, permitting, routing, zoning, procurement,

124
00:07:05,960 --> 00:07:09,760
Speaker 1: or reporting procedures in a manner that facilitates the adoption

125
00:07:09,840 --> 00:07:15,040
Speaker 1: of artificial intelligence models, systems or automated decision systems. Does

126
00:07:15,200 --> 00:07:24,200
Speaker 1: not impose any substantial design, performance, data handling, documentation, civil liability, taxation, fee,

127
00:07:24,640 --> 00:07:29,360
Speaker 1: or other requirement of artificial intelligence models, artificial in intelligence systems,

128
00:07:29,560 --> 00:07:35,120
Speaker 1: or automated decision systems unless required, unless re requirement is

129
00:07:35,160 --> 00:07:37,880
Speaker 1: imposed under the federal law. Sorry, that was long. I

130
00:07:37,880 --> 00:07:39,760
Speaker 1: know I started at the end, but that is important.

131
00:07:39,800 --> 00:07:42,120
Speaker 1: I want you to hear the actual law, not just

132
00:07:42,160 --> 00:07:45,760
Speaker 1: what somebody is saying about it. So, states which already

133
00:07:45,840 --> 00:07:48,440
Speaker 1: have begun to regulate AI. More than twenty states have

134
00:07:48,480 --> 00:07:51,800
Speaker 1: a law on the books to regulate AI in some fashion,

135
00:07:52,320 --> 00:07:56,360
Speaker 1: will not only not be allowed to pass new regulations,

136
00:07:56,400 --> 00:08:00,800
Speaker 1: but they cannot even enforce the ones currently on the books.

137
00:08:01,560 --> 00:08:04,520
Speaker 1: They are also not allowed to make any AI civil

138
00:08:04,720 --> 00:08:08,640
Speaker 1: civilly liable and cannot create a special tax for the

139
00:08:08,680 --> 00:08:11,640
Speaker 1: AI system. Okay, I'm going to give you both sides

140
00:08:11,640 --> 00:08:14,960
Speaker 1: the argument, and then my take. First, the most ideal

141
00:08:15,160 --> 00:08:18,440
Speaker 1: case would be for a national standard in all fifty states,

142
00:08:18,480 --> 00:08:21,080
Speaker 1: not to have a patchwork across the country. It's what's

143
00:08:21,080 --> 00:08:24,680
Speaker 1: best for business. It helps companies know the regulations and

144
00:08:24,720 --> 00:08:27,120
Speaker 1: the laws and how to work properly. I think that's

145
00:08:27,160 --> 00:08:29,840
Speaker 1: best business practices. A lot of economist would agree with me.

146
00:08:30,120 --> 00:08:32,480
Speaker 1: It's what we've done for our whole host of industries

147
00:08:32,520 --> 00:08:36,960
Speaker 1: like cars, telecom, food, drugs, and the AI innovation has

148
00:08:37,000 --> 00:08:39,280
Speaker 1: a lot of positives to it. A lot of doctors

149
00:08:39,280 --> 00:08:42,240
Speaker 1: told me they use AI to double check research. Scientists

150
00:08:42,280 --> 00:08:44,880
Speaker 1: are using AI to help discover new curs for diseases.

151
00:08:45,120 --> 00:08:47,720
Speaker 1: I use AI when doing research for podcasts and articles.

152
00:08:47,760 --> 00:08:51,199
Speaker 1: I find it to superior to Google. So that's I mean,

153
00:08:51,240 --> 00:08:54,280
Speaker 1: there are positives this new technology, and there is a

154
00:08:54,320 --> 00:08:57,200
Speaker 1: giant elephant in the room, being China. We want to

155
00:08:57,200 --> 00:09:01,120
Speaker 1: be more advanced than our main global adversary. Now here's

156
00:09:01,160 --> 00:09:05,200
Speaker 1: the negative side. Congress is broken and does very little,

157
00:09:05,400 --> 00:09:09,080
Speaker 1: very quickly, and AI is moving at light speed. Remember

158
00:09:09,160 --> 00:09:12,680
Speaker 1: Chat GPT is less than three years old, and colleges

159
00:09:12,679 --> 00:09:15,840
Speaker 1: and high schools are struggling to adapt. We don't know

160
00:09:15,840 --> 00:09:18,720
Speaker 1: where this technology is going. And with an aging population

161
00:09:18,840 --> 00:09:22,200
Speaker 1: in Congress that knows very little about technology and is

162
00:09:22,240 --> 00:09:26,080
Speaker 1: dependent on donation from the text industry, how can we

163
00:09:26,120 --> 00:09:28,400
Speaker 1: hope that they properly regulate it and it moves with

164
00:09:28,480 --> 00:09:33,359
Speaker 1: the times. There are legitimate concerns what happens to intellectual

165
00:09:33,440 --> 00:09:37,000
Speaker 1: property with AI? What happens to deep fakes? Especially regarding

166
00:09:37,000 --> 00:09:41,319
Speaker 1: to health science politics. Boomers can barely already tell the

167
00:09:41,360 --> 00:09:43,240
Speaker 1: difference between a real video and a fake video and

168
00:09:43,240 --> 00:09:46,320
Speaker 1: pictures on Facebook. What about if an AI system is

169
00:09:46,360 --> 00:09:49,760
Speaker 1: incorporated in healthcare database and screws up and gets people

170
00:09:49,800 --> 00:09:52,680
Speaker 1: injured or god forbid, dies and they can't be held

171
00:09:52,720 --> 00:09:56,199
Speaker 1: civilly liable. What happens if an AI company makes a

172
00:09:56,280 --> 00:09:59,960
Speaker 1: ship and some it's designed to put into kids' brains?

173
00:10:00,040 --> 00:10:02,280
Speaker 1: I know that sounds insane, but that's what tech people

174
00:10:02,360 --> 00:10:04,880
Speaker 1: are saying that they want to do. That's what the

175
00:10:04,920 --> 00:10:06,520
Speaker 1: future they want to have. I don't know if they

176
00:10:06,520 --> 00:10:09,120
Speaker 1: can get there, but what if they can. Are we

177
00:10:09,120 --> 00:10:12,200
Speaker 1: supposed to wait for Congress to just do something? How

178
00:10:12,280 --> 00:10:16,000
Speaker 1: is that going? We are thirty years, thirty plus years

179
00:10:16,000 --> 00:10:19,040
Speaker 1: into the Internet revolution, and Congress has only passed a

180
00:10:19,200 --> 00:10:22,360
Speaker 1: handful of regulatory bills over the Internet, many of which

181
00:10:22,480 --> 00:10:26,840
Speaker 1: protect Internet companies and not consumers. We're twenty one years

182
00:10:27,040 --> 00:10:30,280
Speaker 1: since the creation of Facebook and the rise of social media,

183
00:10:30,360 --> 00:10:33,720
Speaker 1: which we know has massive effects on children's mental health.

184
00:10:33,840 --> 00:10:36,679
Speaker 1: We know that drug dealers use Snapchat to pedal illegal

185
00:10:36,720 --> 00:10:40,160
Speaker 1: substance substances to miners and other users. We know how

186
00:10:40,160 --> 00:10:42,920
Speaker 1: social media companies censor news and affect our elections. And

187
00:10:43,000 --> 00:10:48,359
Speaker 1: until last year, Congress didn't do anything to protect Americans

188
00:10:48,520 --> 00:10:52,560
Speaker 1: on social media. And what they've passed is very limited

189
00:10:52,880 --> 00:10:56,640
Speaker 1: and so in scope except when they ban TikTok, and

190
00:10:56,679 --> 00:10:59,800
Speaker 1: that isn't even being enforced by either the Democrat president

191
00:11:00,280 --> 00:11:03,880
Speaker 1: left office or the current Republican president. So don't tell

192
00:11:03,920 --> 00:11:06,679
Speaker 1: me that you're so concerned over the future of the

193
00:11:06,760 --> 00:11:10,040
Speaker 1: Chinese Communist Party when you won't even stop using their

194
00:11:10,080 --> 00:11:12,719
Speaker 1: spy app. You know, because nurses have to do their

195
00:11:12,720 --> 00:11:16,400
Speaker 1: little TikTok dances in hospitals. We can't sit there and

196
00:11:16,400 --> 00:11:18,880
Speaker 1: stop Chinese spyware, but we have to beat them in

197
00:11:18,880 --> 00:11:23,920
Speaker 1: the AI race. It is not a coherent message. There's

198
00:11:23,960 --> 00:11:26,400
Speaker 1: also the worry about what's going to happen to the

199
00:11:26,520 --> 00:11:30,000
Speaker 1: job market in the future. AI Dario Emodi, and I'm

200
00:11:30,000 --> 00:11:33,160
Speaker 1: probably mispronouncing his name. I'm sorry. He is the CEO,

201
00:11:33,280 --> 00:11:36,200
Speaker 1: anthropic and a leader in the AI industry. He even

202
00:11:36,600 --> 00:11:39,960
Speaker 1: an interview to Anderson Cooper on CNN. Sign note, I

203
00:11:40,000 --> 00:11:42,720
Speaker 1: watched the interview, and Dario looks exactly what you would

204
00:11:42,760 --> 00:11:46,240
Speaker 1: think tech AI CEO looks like like a virgin who

205
00:11:46,280 --> 00:11:50,480
Speaker 1: just fell up a building. Anyway. Anyway, he believes that

206
00:11:50,559 --> 00:11:54,200
Speaker 1: we are just five years away from twenty percent of

207
00:11:54,400 --> 00:11:58,560
Speaker 1: all entry level white collar jobs being erased. He says

208
00:11:58,600 --> 00:12:00,880
Speaker 1: he has an idea of the future where GDP is

209
00:12:00,920 --> 00:12:04,200
Speaker 1: at ten percent, the national debt is erased because of

210
00:12:04,240 --> 00:12:10,280
Speaker 1: massive GDP growth, and unemployment is at twenty percent. He said,

211
00:12:10,320 --> 00:12:12,640
Speaker 1: that's not out of the question. Now, remember during the

212
00:12:12,640 --> 00:12:16,080
Speaker 1: Great Recession of the two thousand and eight, which led

213
00:12:16,120 --> 00:12:19,400
Speaker 1: to rise of a lot of socialists thinking in our country,

214
00:12:19,520 --> 00:12:21,320
Speaker 1: led to a rise of the Bernie Sanders movement in

215
00:12:21,320 --> 00:12:24,040
Speaker 1: a certain way, led to rise of Barack Obama in

216
00:12:24,120 --> 00:12:27,400
Speaker 1: a certain way, that was when unemployment was at nine

217
00:12:27,480 --> 00:12:34,079
Speaker 1: percent nine point five nationwide. Twenty percent is double more

218
00:12:34,160 --> 00:12:37,880
Speaker 1: than double that, and he's not the only one. David Sue,

219
00:12:38,320 --> 00:12:41,400
Speaker 1: the founder of retools, says his goal is to automate

220
00:12:41,440 --> 00:12:45,160
Speaker 1: ten percent of the labor force in the next five years.

221
00:12:45,360 --> 00:12:47,440
Speaker 1: I want a trip to the border with some tech CEOs,

222
00:12:47,480 --> 00:12:50,160
Speaker 1: like maybe two years ago, and they were all talking

223
00:12:50,160 --> 00:12:52,240
Speaker 1: about this, that millions of jobs were going to be

224
00:12:52,240 --> 00:12:54,880
Speaker 1: wiped out and people would not be able to find work.

225
00:12:55,160 --> 00:12:56,880
Speaker 1: They all said, we're going to have to either have

226
00:12:56,960 --> 00:12:59,560
Speaker 1: some kind of government work projects or some people have

227
00:12:59,559 --> 00:13:03,319
Speaker 1: something to do, or a universal basic income to subsidize

228
00:13:03,320 --> 00:13:06,480
Speaker 1: people who won't be able to find work. We already

229
00:13:06,480 --> 00:13:09,319
Speaker 1: could be seeing the signs. Derek Thompson from The Atlantic

230
00:13:09,400 --> 00:13:13,199
Speaker 1: Track the recent college graduates have a higher unemployment rate

231
00:13:13,240 --> 00:13:16,560
Speaker 1: than the national average for the first time ever. This

232
00:13:16,600 --> 00:13:20,640
Speaker 1: is especially true for kids going into STEM fields. Computer

233
00:13:20,720 --> 00:13:25,679
Speaker 1: engineers have the third highest unemployment rate among recent college graduates.

234
00:13:25,760 --> 00:13:29,600
Speaker 1: That's right now, and Americans are increasingly worried about the

235
00:13:29,640 --> 00:13:33,079
Speaker 1: fear that AI will cost their jobs and their industries.

236
00:13:33,360 --> 00:13:36,240
Speaker 1: In March twenty twenty three, a Yuga poll found that

237
00:13:36,320 --> 00:13:39,240
Speaker 1: twenty nine percent of Americans thought that the advances in

238
00:13:39,320 --> 00:13:41,839
Speaker 1: AI would lead to a decrease in the amount of

239
00:13:41,920 --> 00:13:45,520
Speaker 1: jobs available in their industry. Fast forward to August twenty

240
00:13:45,559 --> 00:13:49,640
Speaker 1: twenty four, that number increased to forty eight percent. So

241
00:13:49,720 --> 00:13:53,839
Speaker 1: what's the reaction then to the big beautiful bills AI moratorium. Well,

242
00:13:53,920 --> 00:13:56,559
Speaker 1: Ted Cruz says that he's all for it. Josh Howley

243
00:13:56,600 --> 00:13:58,440
Speaker 1: says he's got some anxiety about it. I think he'll

244
00:13:58,440 --> 00:14:01,600
Speaker 1: fold like a house of cards. And MTG, who voted

245
00:14:01,600 --> 00:14:03,079
Speaker 1: for the bill in the House, the first version of

246
00:14:03,080 --> 00:14:04,720
Speaker 1: the bill in the House, she says she's going to

247
00:14:04,760 --> 00:14:07,880
Speaker 1: vote against it, the compromise bill if the Senate does

248
00:14:07,920 --> 00:14:10,800
Speaker 1: not strip that language. Now, remember the bill passed by

249
00:14:10,800 --> 00:14:13,400
Speaker 1: a single vote in the House Representative. So if MGG

250
00:14:14,160 --> 00:14:18,120
Speaker 1: sticks to her guns, then we might get the ten

251
00:14:18,200 --> 00:14:22,560
Speaker 1: year moratorium out of the legislation. We'll see Speaker Mike

252
00:14:22,640 --> 00:14:25,960
Speaker 1: Johnson said that he feels very compassionate. We need to

253
00:14:26,040 --> 00:14:28,760
Speaker 1: keep that in there. More than three hundred and sixty

254
00:14:28,840 --> 00:14:32,280
Speaker 1: state legislators, both Republicans and Democrats, I need a letter

255
00:14:32,680 --> 00:14:35,040
Speaker 1: asking Congress to strip the language from the bill. This

256
00:14:35,080 --> 00:14:39,040
Speaker 1: includes very progressive Democrats and Super Mago Republicans who signed

257
00:14:39,080 --> 00:14:41,720
Speaker 1: with this letter. It wasn't like, you know, just a

258
00:14:41,800 --> 00:14:44,600
Speaker 1: rhino and a bunch of Democrats. No, there were hardcore

259
00:14:45,280 --> 00:14:48,880
Speaker 1: state legislators signed on to this letter. Senate Republicans will

260
00:14:48,920 --> 00:14:51,680
Speaker 1: likely have to make some reform simply because something called

261
00:14:51,680 --> 00:14:55,040
Speaker 1: the Bird Rule, which allows Senators to block provisions and

262
00:14:55,080 --> 00:14:59,480
Speaker 1: a reconciliation bill that are deemed extoraneous. The federal budget

263
00:15:00,040 --> 00:15:02,720
Speaker 1: can't put anything in the reconciliation bill. It has to

264
00:15:02,760 --> 00:15:05,800
Speaker 1: be kind of stuck to spending. I spoke to a

265
00:15:05,840 --> 00:15:08,400
Speaker 1: Republican Senate staff or a top Republican Senate center for

266
00:15:08,520 --> 00:15:11,440
Speaker 1: some cenator you would all know, and he told me

267
00:15:11,520 --> 00:15:15,040
Speaker 1: that they plan on attaching the regulation to federal money

268
00:15:15,360 --> 00:15:18,360
Speaker 1: so that states could go forward regulation on AI, but

269
00:15:18,400 --> 00:15:20,800
Speaker 1: they'll lose access to federal money, something that some states

270
00:15:20,840 --> 00:15:23,800
Speaker 1: can't afford to do, but many can't. And when I

271
00:15:23,920 --> 00:15:27,520
Speaker 1: brought up the concerns with AI, the staff are simply

272
00:15:27,560 --> 00:15:30,320
Speaker 1: said there is no AI crisis, and that this is

273
00:15:31,080 --> 00:15:34,520
Speaker 1: to prevent future gridlock, to create a national standard something

274
00:15:34,560 --> 00:15:36,840
Speaker 1: that we've seen the kind of gridlock we've seen over

275
00:15:37,040 --> 00:15:39,920
Speaker 1: national data privacy laws. We can't get it because states

276
00:15:39,960 --> 00:15:43,080
Speaker 1: already started making their own policy. He also assured me

277
00:15:43,120 --> 00:15:46,080
Speaker 1: that this is only a moratorium undeveloping AI models, and

278
00:15:46,160 --> 00:15:49,280
Speaker 1: states can do whatever they want about specific harms or

279
00:15:49,320 --> 00:15:51,800
Speaker 1: offering it in their states. I cast a lot of

280
00:15:51,880 --> 00:15:55,200
Speaker 1: doubt on the second part, because why would they ban

281
00:15:55,360 --> 00:15:59,760
Speaker 1: legislation on civil liabilities? Republicans in favor of the ten

282
00:15:59,840 --> 00:16:02,720
Speaker 1: year more term have fallen into the scarecrow argument that

283
00:16:02,720 --> 00:16:06,200
Speaker 1: if Congress doesn't forbid it, then California will create the

284
00:16:06,280 --> 00:16:11,000
Speaker 1: regulations and we can't trust gavenusom Okay, if California becomes

285
00:16:11,040 --> 00:16:14,920
Speaker 1: too onerous, they'll move to Texas or Florida, or Utah

286
00:16:15,160 --> 00:16:19,120
Speaker 1: or Tennessee. Elon Musk moved to Tennessee. They'll make their

287
00:16:19,120 --> 00:16:22,920
Speaker 1: own set of standards, the ones that Republican legislators would

288
00:16:22,920 --> 00:16:25,880
Speaker 1: work better with. The Truth is that twenty states, including

289
00:16:25,920 --> 00:16:30,000
Speaker 1: many Red states, already put AI regulation in the law,

290
00:16:30,400 --> 00:16:34,800
Speaker 1: while Congress has done nothing. But by waiting ten years

291
00:16:34,920 --> 00:16:40,040
Speaker 1: with the hope that Congress will make regulation that protects consumers, children, workers,

292
00:16:40,160 --> 00:16:44,280
Speaker 1: and somehow manages to avoid mass unemployment and wealth consolidation,

293
00:16:44,760 --> 00:16:49,160
Speaker 1: leaves someone like me feeling very skeptical and as far

294
00:16:49,200 --> 00:16:52,000
Speaker 1: as winning the AI race, against China, goes when does

295
00:16:52,040 --> 00:16:55,360
Speaker 1: this race end? Where's the finish line? What are we

296
00:16:55,560 --> 00:16:58,560
Speaker 1: racing tours? And can we for one second have an

297
00:16:58,560 --> 00:17:03,720
Speaker 1: adult conversation without hyperbole before doing something that probably can't

298
00:17:03,760 --> 00:17:07,119
Speaker 1: be undone? And my last thought on this monologue to

299
00:17:07,200 --> 00:17:09,800
Speaker 1: any Republican who is running for office or considering to

300
00:17:10,000 --> 00:17:12,919
Speaker 1: or in office and hearing this, if you want to

301
00:17:13,000 --> 00:17:18,560
Speaker 1: turn twenty million Trump voters into hardcore AOC or Bernie voters,

302
00:17:18,720 --> 00:17:21,399
Speaker 1: in the blink of an eye, replace their jobs with

303
00:17:21,440 --> 00:17:25,679
Speaker 1: AI without ensuring they can find another one. Now, I

304
00:17:25,720 --> 00:17:28,240
Speaker 1: have two guests coming on No Far More but AI

305
00:17:28,280 --> 00:17:30,159
Speaker 1: and the policies around it and how it affects the

306
00:17:30,160 --> 00:17:36,440
Speaker 1: economy than I do. They're coming up next. Stay tune.

307
00:17:36,560 --> 00:17:39,200
Speaker 1: My guest for today's episode is Brad Carson. He's the

308
00:17:39,240 --> 00:17:41,800
Speaker 1: former president of the University of Tulsa and the current

309
00:17:41,840 --> 00:17:45,440
Speaker 1: co founder of Americans for Responsible Innovation, and Mark Beam,

310
00:17:45,480 --> 00:17:48,960
Speaker 1: who's the president of Government Affairs for the AI Policy Network.

311
00:17:49,000 --> 00:17:51,760
Speaker 1: Thank you both for being here. I want to start

312
00:17:51,760 --> 00:17:54,280
Speaker 1: by talking about the Big Beautiful Bill and the ten

313
00:17:54,359 --> 00:17:59,600
Speaker 1: year moratorium on states regulating AI systems. Brad, your organization,

314
00:17:59,680 --> 00:18:02,359
Speaker 1: America for Responsible Innovation took part in an effort to

315
00:18:02,359 --> 00:18:05,360
Speaker 1: get three hundred and sixty state legislators to sign onto

316
00:18:05,359 --> 00:18:08,080
Speaker 1: a letter to oppose that section of the bill. Why

317
00:18:08,119 --> 00:18:10,840
Speaker 1: do you feel it's better to have states regulate the

318
00:18:10,880 --> 00:18:13,880
Speaker 1: industry instead of like a federal proposition.

319
00:18:14,880 --> 00:18:16,960
Speaker 2: Well, I think you have to view the choices as

320
00:18:17,000 --> 00:18:20,639
Speaker 2: not simply a binary between states regulating it and the

321
00:18:20,640 --> 00:18:25,199
Speaker 2: federal government. We would prefer a uniform federal regulatory scheme.

322
00:18:25,480 --> 00:18:28,840
Speaker 2: That's the best option. The second best option is states

323
00:18:28,920 --> 00:18:32,399
Speaker 2: experimenting with something. The worst option is to have a

324
00:18:32,480 --> 00:18:36,520
Speaker 2: moratorium on state action and have no federal approach. And

325
00:18:36,640 --> 00:18:39,560
Speaker 2: my worry is going forward that third option is actually

326
00:18:39,560 --> 00:18:42,120
Speaker 2: what Congress is going to take. We know how hard

327
00:18:42,160 --> 00:18:44,600
Speaker 2: it is it get things through Congress. So a moratoria

328
00:18:44,720 --> 00:18:47,600
Speaker 2: without a regulatory scheme in place is the worst of

329
00:18:47,680 --> 00:18:48,720
Speaker 2: all possible worlds.

330
00:18:48,800 --> 00:18:52,720
Speaker 1: What about the argument from supporters of a more robust

331
00:18:52,800 --> 00:18:56,840
Speaker 1: AI policy that has very little regulation, that we don't

332
00:18:56,880 --> 00:19:00,800
Speaker 1: want California setting the standards for the and that's just

333
00:19:00,880 --> 00:19:02,800
Speaker 1: too onerous for companies to fallow.

334
00:19:03,760 --> 00:19:06,000
Speaker 2: You know, I'm open to the idea that California or

335
00:19:06,080 --> 00:19:09,040
Speaker 2: New York shouldn't set a national standard, but states seem

336
00:19:09,080 --> 00:19:12,120
Speaker 2: to be free to experiment with regulation in the absence

337
00:19:12,119 --> 00:19:15,640
Speaker 2: of a federal regulatory scheme, and the preemption as it's

338
00:19:15,640 --> 00:19:19,040
Speaker 2: written today, would ensure that a lot of very important

339
00:19:19,040 --> 00:19:21,760
Speaker 2: consumer protection laws they really have nothing to do with

340
00:19:21,840 --> 00:19:25,399
Speaker 2: frontier AI regulation, which is the cutting edge, would be

341
00:19:25,400 --> 00:19:28,240
Speaker 2: stopped at their tracks too. So what we have going

342
00:19:28,280 --> 00:19:30,640
Speaker 2: forward with the Congress that can't pass laws, and I'm

343
00:19:30,680 --> 00:19:33,280
Speaker 2: a former congressman, so these are my colleagues. They can't

344
00:19:33,320 --> 00:19:36,879
Speaker 2: pass laws. A moratorium would basically leave a vacuum in

345
00:19:36,880 --> 00:19:39,560
Speaker 2: which there'd be no oversight at all of what's probably

346
00:19:39,560 --> 00:19:41,840
Speaker 2: going to be the world's most transformative technology.

347
00:19:42,680 --> 00:19:46,800
Speaker 1: Mark You had a tweet in response to Dario Anodi's prediction,

348
00:19:46,920 --> 00:19:48,440
Speaker 1: I don't know if I'm saying his last name correctly.

349
00:19:48,480 --> 00:19:50,480
Speaker 1: I think I am prediction that AI will lead to

350
00:19:50,520 --> 00:19:53,399
Speaker 1: wiping out twenty percent of white collar jobs for recent

351
00:19:53,440 --> 00:19:56,480
Speaker 1: college graduates. You said, quote nine to eleven may meet

352
00:19:56,520 --> 00:20:00,240
Speaker 1: study intelligence failures. How we had warnings but couldn't believe them.

353
00:20:00,280 --> 00:20:03,560
Speaker 1: AI execs are telling us what's coming millions of jobs

354
00:20:03,600 --> 00:20:06,879
Speaker 1: gone or worse. Will we act on this intelligence overly

355
00:20:06,920 --> 00:20:09,080
Speaker 1: wait for an impact. History doesn't repeat it something it

356
00:20:09,119 --> 00:20:13,679
Speaker 1: does rhyme what does responsible AI regulation look like? Because

357
00:20:13,720 --> 00:20:17,399
Speaker 1: tech execs basically say, if you either have almost no

358
00:20:17,480 --> 00:20:20,440
Speaker 1: regulation or you allow us to lose to China.

359
00:20:21,720 --> 00:20:25,000
Speaker 3: Yeah, I think we love to sometimes frame a bit

360
00:20:25,000 --> 00:20:28,280
Speaker 3: of a false choice and a false dichotomy here. And

361
00:20:28,359 --> 00:20:30,840
Speaker 3: I think, certainly, you know, as as folks in the

362
00:20:30,840 --> 00:20:34,159
Speaker 3: Senate Commerce Committee, like to say, it's either going to

363
00:20:34,200 --> 00:20:38,240
Speaker 3: be some European Union style heavy handed regulation or America

364
00:20:38,320 --> 00:20:40,320
Speaker 3: is going to win and accelerate. And I think we

365
00:20:40,400 --> 00:20:43,720
Speaker 3: have to find a middle path, you know. And I

366
00:20:43,760 --> 00:20:46,960
Speaker 3: think AI, given how disruptive and transformative it might be,

367
00:20:47,640 --> 00:20:50,760
Speaker 3: and also given some of the uncertainty about the timelines

368
00:20:50,800 --> 00:20:53,520
Speaker 3: associated with how fast it's moving, I think what would

369
00:20:53,560 --> 00:20:56,479
Speaker 3: be obvious to do sort of time now would include

370
00:20:56,520 --> 00:21:00,720
Speaker 3: basically increasing the US government's capacity to test and evaluate

371
00:21:00,840 --> 00:21:04,040
Speaker 3: these systems for things like loss of control risks, for

372
00:21:04,119 --> 00:21:07,520
Speaker 3: things like weaponization, and even things like starting to track

373
00:21:07,920 --> 00:21:11,960
Speaker 3: the types of tasks and jobs that are being automated

374
00:21:11,960 --> 00:21:14,520
Speaker 3: at least in the Fortune five hundred companies. And this

375
00:21:14,560 --> 00:21:17,240
Speaker 3: is an important set of data that can help drive

376
00:21:18,119 --> 00:21:21,640
Speaker 3: a more refined regulatory approach. And the fact that we're

377
00:21:21,680 --> 00:21:24,800
Speaker 3: not even seeming really interested in getting our arms around

378
00:21:24,800 --> 00:21:27,120
Speaker 3: where this has head. It seems to be a significant

379
00:21:27,320 --> 00:21:28,639
Speaker 3: potential intelligence failure.

380
00:21:28,680 --> 00:21:32,000
Speaker 1: Oh question, I forgive me if I sound naive, because

381
00:21:32,000 --> 00:21:35,800
Speaker 1: I don't know this answer. If China invades Taiwan, is

382
00:21:35,880 --> 00:21:39,000
Speaker 1: the AI race essentially over? You know, I obviously it's

383
00:21:39,119 --> 00:21:41,600
Speaker 1: it's sort of the chips, That's why I'm thinking that.

384
00:21:43,280 --> 00:21:46,480
Speaker 3: So obviously, you know, the Taiwan some In Conductor Manufacturing

385
00:21:46,520 --> 00:21:51,040
Speaker 3: Center is perhaps the world's most strategic foundry for producing

386
00:21:51,119 --> 00:21:54,800
Speaker 3: the chips that go into the training runs for advanced

387
00:21:54,840 --> 00:21:58,560
Speaker 3: AI systems. And you know TSMC is looking to diverse

388
00:21:58,640 --> 00:22:03,480
Speaker 3: its manufacturing capacity, including in places like Arizona. But I think, yeah,

389
00:22:03,640 --> 00:22:07,800
Speaker 3: one of the biggest for strategic challenges if the China

390
00:22:07,800 --> 00:22:10,400
Speaker 3: were to invade Taiwan would be associated with that fabric.

391
00:22:10,520 --> 00:22:13,040
Speaker 3: That fab and whether we even survive the war, our

392
00:22:13,080 --> 00:22:15,720
Speaker 3: survivor conflict would be sort of a top of mind question.

393
00:22:15,760 --> 00:22:17,639
Speaker 3: But if it's in fact true that China were to

394
00:22:17,880 --> 00:22:20,119
Speaker 3: seize that capability and have it for its own, then

395
00:22:20,119 --> 00:22:22,719
Speaker 3: I think it would certainly put us in a significant

396
00:22:22,720 --> 00:22:24,040
Speaker 3: strategic disadvantage.

397
00:22:24,160 --> 00:22:27,040
Speaker 1: Yeah, and I so, speaking of China, we always hear

398
00:22:27,080 --> 00:22:29,000
Speaker 1: about the race against China. I want to ask you

399
00:22:29,000 --> 00:22:30,760
Speaker 1: both this question. I'll start with Brad and go to Mark.

400
00:22:30,960 --> 00:22:34,680
Speaker 1: But there's the race against China. What is the finish

401
00:22:34,720 --> 00:22:38,520
Speaker 1: line of that race look like? Because that's the thing

402
00:22:38,560 --> 00:22:42,560
Speaker 1: that we all have. Every job and military thing is

403
00:22:42,560 --> 00:22:45,880
Speaker 1: completely replaced by computers. Like what are we racing towards?

404
00:22:45,960 --> 00:22:47,240
Speaker 1: And I don't know that answer.

405
00:22:47,880 --> 00:22:50,040
Speaker 2: It's a great question, they don't think people who use

406
00:22:50,080 --> 00:22:53,160
Speaker 2: that metaphor have always thoughted through. Usually in the past

407
00:22:53,240 --> 00:22:57,000
Speaker 2: we talk about arms races. We use that term pejoratively, right,

408
00:22:57,040 --> 00:23:01,879
Speaker 2: It's about excessive expenditures of an escalation of threats that

409
00:23:01,960 --> 00:23:04,200
Speaker 2: often leaves a lot of people dead in the back.

410
00:23:04,560 --> 00:23:06,920
Speaker 2: And so it's a good question. Is the arms race

411
00:23:06,960 --> 00:23:09,400
Speaker 2: the right metaphor for AI? I think what we're trying

412
00:23:09,400 --> 00:23:12,640
Speaker 2: to get to is some sense of artificial superintelligence. Right,

413
00:23:12,760 --> 00:23:16,640
Speaker 2: the first person to get there, in some speculative scenarios,

414
00:23:17,000 --> 00:23:20,720
Speaker 2: I could have a decisive military advantage. The issue is like,

415
00:23:20,840 --> 00:23:23,399
Speaker 2: is that actually possible? What does that mean? How quickly

416
00:23:23,440 --> 00:23:25,960
Speaker 2: would it diffuse to China? And I think we'd actually

417
00:23:25,960 --> 00:23:28,240
Speaker 2: probably be wise to get away from the arms race

418
00:23:28,320 --> 00:23:30,600
Speaker 2: framing of it and instead think about what we can

419
00:23:30,640 --> 00:23:34,760
Speaker 2: actually do in this country sometimes cooperatively recognize that China

420
00:23:34,800 --> 00:23:38,239
Speaker 2: will likely have AI whatever we do shortly thereafter, So

421
00:23:38,280 --> 00:23:40,800
Speaker 2: it's going to be difficult to keep to have us

422
00:23:40,880 --> 00:23:43,040
Speaker 2: possession of it alone and try to think of a

423
00:23:43,080 --> 00:23:44,920
Speaker 2: way to use AI for the good and get away

424
00:23:44,920 --> 00:23:47,280
Speaker 2: from this kind of militaristic framing of my.

425
00:23:48,600 --> 00:23:53,520
Speaker 1: Yeah, I mean self taught AI is and I push

426
00:23:53,600 --> 00:23:56,840
Speaker 1: this over to Mark. Self taught AI isn't that everyone

427
00:23:56,920 --> 00:24:03,320
Speaker 1: says could be possible or super intelligence AI, but it's

428
00:24:03,600 --> 00:24:06,439
Speaker 1: years in the future. If then, but there was a

429
00:24:07,119 --> 00:24:09,280
Speaker 1: thing instance with a I might be saying this wrong.

430
00:24:09,520 --> 00:24:14,240
Speaker 1: Palisade AI open AI three model rewriting its own code

431
00:24:14,280 --> 00:24:18,199
Speaker 1: to avoid being shut down, and even blackmailing one of

432
00:24:18,200 --> 00:24:22,520
Speaker 1: the developers saying they'll reveal their spouses indiscretion or their

433
00:24:22,560 --> 00:24:25,879
Speaker 1: indiscretion with their spouse for the fact, are we close

434
00:24:25,920 --> 00:24:30,040
Speaker 1: to that? And like is I mean, do we want

435
00:24:30,160 --> 00:24:34,040
Speaker 1: a situation where we're like doctor Frankenstein making the self

436
00:24:34,080 --> 00:24:38,040
Speaker 1: taught monster and saying, okay, this is a superior for humanity.

437
00:24:38,960 --> 00:24:43,400
Speaker 3: This is a very concerning development, and unfortunately it's one

438
00:24:43,640 --> 00:24:46,320
Speaker 3: we've seen little warning signs along the way that these

439
00:24:46,359 --> 00:24:49,520
Speaker 3: AI models behave in strange ways. And to be clear, Ryan,

440
00:24:50,119 --> 00:24:52,720
Speaker 3: when developers who are making these systems are not really

441
00:24:52,800 --> 00:24:56,240
Speaker 3: building them themselves, they're almost like growing them. And so

442
00:24:56,280 --> 00:24:58,200
Speaker 3: they take a whole bunch of compute resources and a

443
00:24:58,200 --> 00:25:00,440
Speaker 3: whole bunch of data and run these algorithms ruin these

444
00:25:00,480 --> 00:25:04,679
Speaker 3: AI write their own ways, and as a result, it's

445
00:25:04,680 --> 00:25:06,480
Speaker 3: it's like there's no it's like a very much a

446
00:25:06,480 --> 00:25:09,000
Speaker 3: black box. We can't like crack these things open and

447
00:25:09,160 --> 00:25:11,199
Speaker 3: understand how they work and reason about them, and they

448
00:25:11,240 --> 00:25:14,520
Speaker 3: make these really weird decisions sometimes, like the example that

449
00:25:14,560 --> 00:25:17,720
Speaker 3: you made with the team at Palisade Research in three

450
00:25:17,880 --> 00:25:24,480
Speaker 3: avoiding shutdown anthropics model Opus four attempting to blackmail. It's uh,

451
00:25:24,520 --> 00:25:27,640
Speaker 3: it's it's engineer, you know. I think if you extrapolate

452
00:25:28,080 --> 00:25:30,920
Speaker 3: this further, I think super intelligence might actually be a

453
00:25:30,960 --> 00:25:33,760
Speaker 3: little bit closer than than folks may realize. Although it's

454
00:25:33,760 --> 00:25:36,199
Speaker 3: some some some disagreement in the field there. But if

455
00:25:36,240 --> 00:25:39,159
Speaker 3: you have a super intelligent system that is capable of

456
00:25:39,200 --> 00:25:41,760
Speaker 3: rewriting its own code and avoiding shutdown, this is the

457
00:25:41,760 --> 00:25:43,879
Speaker 3: scenario on which a lot of the experts are starting

458
00:25:43,880 --> 00:25:45,480
Speaker 3: to sound the warm bells about right now.

459
00:25:45,760 --> 00:25:49,240
Speaker 1: Yeah, there was there was that a conversation Dario had

460
00:25:49,280 --> 00:25:52,520
Speaker 1: with Dara Modi had with r Anderson Cooper talking about

461
00:25:52,560 --> 00:25:55,560
Speaker 1: this entire thing. And I have a young recent college

462
00:25:55,600 --> 00:25:57,840
Speaker 1: grad who's a researcher for me, helps kill like some

463
00:25:57,920 --> 00:26:00,680
Speaker 1: data for this podcast. And it's a part time job,

464
00:26:00,800 --> 00:26:03,440
Speaker 1: you know, it's just data for my twice a week podcast.

465
00:26:04,359 --> 00:26:06,119
Speaker 1: But he can't find a full time job as a

466
00:26:06,160 --> 00:26:09,639
Speaker 1: recent college graduate. And he blames in part automation and

467
00:26:09,720 --> 00:26:13,400
Speaker 1: they just don't hire these types of entry level jobs anymore.

468
00:26:13,720 --> 00:26:17,199
Speaker 1: We've seen that college graduates, recent college graduates have a

469
00:26:17,280 --> 00:26:19,680
Speaker 1: higher unemployment rate than the national average for the first

470
00:26:19,720 --> 00:26:23,320
Speaker 1: time in forty years. You're both dads. You said this

471
00:26:23,359 --> 00:26:25,600
Speaker 1: on Twitter, so I'm not exploding some new information. But

472
00:26:25,640 --> 00:26:27,240
Speaker 1: your both dads. You both have kids. I don't know

473
00:26:27,240 --> 00:26:31,360
Speaker 1: how old they are. Are you both worried about their

474
00:26:31,440 --> 00:26:35,360
Speaker 1: opportunity for the workforce and what are college kids supposed

475
00:26:35,359 --> 00:26:39,240
Speaker 1: to do to ensure they can get work in this future.

476
00:26:41,040 --> 00:26:41,919
Speaker 2: It's a great question.

477
00:26:42,240 --> 00:26:42,399
Speaker 1: You know.

478
00:26:42,480 --> 00:26:45,720
Speaker 2: We do see companies like do Alinko and Shopify say

479
00:26:45,760 --> 00:26:48,160
Speaker 2: that before you post a job, you have to assert

480
00:26:48,200 --> 00:26:50,959
Speaker 2: that AI cannot do it. As you mentioned, the un

481
00:26:50,960 --> 00:26:54,800
Speaker 2: deployment rate for new college graduates is larger higher than

482
00:26:54,760 --> 00:26:58,520
Speaker 2: the national average. The Washington Post recently said that it's

483
00:26:58,520 --> 00:27:02,160
Speaker 2: the worst market for software years since nineteen seventy nine,

484
00:27:02,440 --> 00:27:06,399
Speaker 2: we hardly even had computers in wide dissemination. So I

485
00:27:06,440 --> 00:27:08,960
Speaker 2: do think the jobs are going to go away. There's

486
00:27:09,000 --> 00:27:11,600
Speaker 2: a debate in the AI community about how rapidly that

487
00:27:11,640 --> 00:27:14,720
Speaker 2: will be, but most people think that white college jobs,

488
00:27:14,760 --> 00:27:17,760
Speaker 2: especially will be increasingly automated. And you might see in

489
00:27:17,800 --> 00:27:20,240
Speaker 2: the next five or seven years fifty percent of white

490
00:27:20,240 --> 00:27:23,960
Speaker 2: college jobs could be done by AI. And so it's

491
00:27:24,000 --> 00:27:26,600
Speaker 2: a very good question, and there's no obviously answer for

492
00:27:26,640 --> 00:27:28,800
Speaker 2: what you should study. You know. One hand, one could

493
00:27:28,800 --> 00:27:30,840
Speaker 2: try to study machine learning itself, where you're one of

494
00:27:30,880 --> 00:27:33,840
Speaker 2: the engineers making these products, you know. On the other hand,

495
00:27:34,680 --> 00:27:36,639
Speaker 2: it's not obvious what you could do. And I think

496
00:27:36,920 --> 00:27:39,800
Speaker 2: this calls into question, like the very social compact, what

497
00:27:39,880 --> 00:27:42,480
Speaker 2: is democracy in a world where lots of people don't

498
00:27:42,520 --> 00:27:45,919
Speaker 2: have jobs, where we have incredible economic growth, perhaps, but

499
00:27:45,960 --> 00:27:49,280
Speaker 2: it's concentrated in a very very few number of people,

500
00:27:49,320 --> 00:27:51,560
Speaker 2: the people who are running these labs, and the rest

501
00:27:51,560 --> 00:27:54,520
Speaker 2: of us find ourselves and penury. I think it's actually

502
00:27:54,600 --> 00:27:57,280
Speaker 2: going to be a devastating problem for us, and it's coming.

503
00:27:57,520 --> 00:27:59,359
Speaker 2: As Mark said, we very much agree on most of

504
00:27:59,359 --> 00:28:02,240
Speaker 2: these issues. It's coming a lot faster to the average

505
00:28:02,280 --> 00:28:06,040
Speaker 2: American things, whether it's two years, five years, seven years,

506
00:28:06,200 --> 00:28:08,879
Speaker 2: it's coming very rapidly for all of our jobs. And

507
00:28:08,960 --> 00:28:10,400
Speaker 2: there's no easy answer to it.

508
00:28:10,720 --> 00:28:12,680
Speaker 1: Yeah, it might be the number one issue going into

509
00:28:12,680 --> 00:28:16,880
Speaker 1: the twenty twenty eight presidential election. Mark, what I mean,

510
00:28:17,119 --> 00:28:19,800
Speaker 1: what would you say about fifty percent wipe out of

511
00:28:19,840 --> 00:28:26,160
Speaker 1: why college educated jobs is like absolutely devastating. And when

512
00:28:26,200 --> 00:28:29,359
Speaker 1: you hear these techeos talk and I've been on trips

513
00:28:29,359 --> 00:28:31,960
Speaker 1: with tech CEOs and they've talked about this going back

514
00:28:32,000 --> 00:28:35,760
Speaker 1: to three years. In my case, to me, they're like, oh,

515
00:28:35,760 --> 00:28:38,360
Speaker 1: we have to do ubi universal basic income. There is

516
00:28:38,400 --> 00:28:41,880
Speaker 1: no other They will just be permanently unemployed people.

517
00:28:43,520 --> 00:28:45,880
Speaker 3: Yeah, you know, I agree with a lot of what

518
00:28:45,880 --> 00:28:48,640
Speaker 3: with Brad said. I have a sixteen year old son.

519
00:28:48,840 --> 00:28:50,560
Speaker 3: We were talking about, like what do you study in

520
00:28:50,600 --> 00:28:55,240
Speaker 3: college these days? I mentioned maybe physics and philosophy could

521
00:28:55,240 --> 00:28:57,800
Speaker 3: be things that could be useful, But I mean, I

522
00:28:57,840 --> 00:29:00,960
Speaker 3: think even the Bureau of Labor Statistics earlier this year

523
00:29:01,080 --> 00:29:04,080
Speaker 3: reported something like a twenty percent year over year drop

524
00:29:04,320 --> 00:29:07,920
Speaker 3: of entry leveled positions. And I think to your point,

525
00:29:08,040 --> 00:29:10,800
Speaker 3: it's not that people will become unemployed, it's like they'll

526
00:29:10,840 --> 00:29:15,720
Speaker 3: become unemployable. And I think this is a significant disruption.

527
00:29:15,840 --> 00:29:17,959
Speaker 3: I think on the good news is that looking at

528
00:29:18,000 --> 00:29:20,840
Speaker 3: where this administration is and some of the remarks that

529
00:29:20,880 --> 00:29:23,320
Speaker 3: the Vice President made in Paris, it seems like we're

530
00:29:23,320 --> 00:29:25,320
Speaker 3: going to try to give workers. I see it at

531
00:29:25,360 --> 00:29:28,240
Speaker 3: the table. I know folks like Senator Cruz are very

532
00:29:28,240 --> 00:29:32,080
Speaker 3: focused on jobs, jobs, jobs. I think this is one

533
00:29:32,120 --> 00:29:33,920
Speaker 3: that we're going to have to grapple with and candidly

534
00:29:34,040 --> 00:29:36,280
Speaker 3: data Like Ryan, I, you know, when people say UBI,

535
00:29:36,480 --> 00:29:39,440
Speaker 3: it's probably the most under specified term I've ever heard.

536
00:29:40,080 --> 00:29:43,160
Speaker 3: I think by default it's the wealth and power will

537
00:29:43,200 --> 00:29:47,000
Speaker 3: amass with a few folks, and I think the rest

538
00:29:47,040 --> 00:29:48,920
Speaker 3: of us are going to be left holding the bag.

539
00:29:48,960 --> 00:29:52,680
Speaker 3: And I'll say that whenever anyone sort of talks about

540
00:29:52,880 --> 00:29:56,720
Speaker 3: this utopian vision, I like to think that. You know,

541
00:29:56,800 --> 00:29:59,360
Speaker 3: if you look at history, every time someone promises utopia

542
00:29:59,440 --> 00:30:02,120
Speaker 3: ends in one place, and that place is the Gulag.

543
00:30:03,040 --> 00:30:06,960
Speaker 1: And so I think we need to outbeat That's not

544
00:30:07,000 --> 00:30:07,600
Speaker 1: a lot I.

545
00:30:08,440 --> 00:30:10,600
Speaker 3: See some message coming out of the White House saying like, oh,

546
00:30:10,600 --> 00:30:12,719
Speaker 3: it's just a left wing agenda, is not. I think

547
00:30:12,760 --> 00:30:15,320
Speaker 3: it's going to affect everybody, and we have to make

548
00:30:15,360 --> 00:30:18,160
Speaker 3: this as much as we can, not partisan, and together.

549
00:30:18,200 --> 00:30:20,040
Speaker 3: We have to have a national dialogue. And it's going

550
00:30:20,080 --> 00:30:24,560
Speaker 3: to provoke some very serious reconsiderations of the fundamental assumptions

551
00:30:24,560 --> 00:30:26,720
Speaker 3: of it that strames our constitutional order right now.

552
00:30:27,160 --> 00:30:30,880
Speaker 1: Yeah, you know, I hear. I hear the things that

553
00:30:31,000 --> 00:30:35,080
Speaker 1: come out of Republican and I work in republic and politics,

554
00:30:35,120 --> 00:30:38,320
Speaker 1: so this is what I'm most familiar with. Is uh,

555
00:30:38,600 --> 00:30:42,320
Speaker 1: you can't let Gavenuwsom run the entire AI economy. You

556
00:30:42,360 --> 00:30:46,960
Speaker 1: can't let China win. And even Ted Cruz, who you mentioned,

557
00:30:48,800 --> 00:30:52,360
Speaker 1: he is absolutely opposed to any state regulation. Even though

558
00:30:52,400 --> 00:30:55,520
Speaker 1: Texas regulates AI. It was one of the earlier states

559
00:30:55,560 --> 00:30:58,000
Speaker 1: to have a regulation, and I a I put into

560
00:30:58,560 --> 00:31:00,640
Speaker 1: law and signed by the governor. I don't it's not

561
00:31:00,680 --> 00:31:05,360
Speaker 1: a serious, serious piece of legislation, but it is a regulation.

562
00:31:07,560 --> 00:31:09,520
Speaker 1: So it's just curious of what see they are. I

563
00:31:09,600 --> 00:31:14,800
Speaker 1: spoke to someone, a very senior politician in office right now,

564
00:31:15,440 --> 00:31:19,480
Speaker 1: and they had a very optimistic view. They have been saying, yes,

565
00:31:19,640 --> 00:31:22,320
Speaker 1: jobs will be lost, but jobs will be created. A

566
00:31:22,320 --> 00:31:24,360
Speaker 1: bil like the Internet, We're going to see millions of

567
00:31:24,360 --> 00:31:26,600
Speaker 1: new jobs that we don't even know about be created.

568
00:31:27,320 --> 00:31:29,000
Speaker 1: Is that a possibility.

569
00:31:29,120 --> 00:31:31,960
Speaker 2: It's always a possibility, right. We talked about the lump

570
00:31:31,960 --> 00:31:35,680
Speaker 2: of labor fallacy and economics, where even when the automobile

571
00:31:35,760 --> 00:31:37,920
Speaker 2: comes and many people were put out of business, new

572
00:31:37,960 --> 00:31:41,960
Speaker 2: jobs were created as mechanics or manufacturing. I do think

573
00:31:42,080 --> 00:31:45,120
Speaker 2: there's a very real chance at this time. Is different.

574
00:31:45,800 --> 00:31:48,160
Speaker 2: I mean, we should be paying attention to what these

575
00:31:48,160 --> 00:31:51,160
Speaker 2: tech executives are saying. They have stayed it openly that

576
00:31:51,160 --> 00:31:54,120
Speaker 2: it's their job to try to create an artificial intelligence

577
00:31:54,240 --> 00:31:56,040
Speaker 2: that can do ninety five percent of the work that

578
00:31:56,120 --> 00:31:59,080
Speaker 2: humans can do today. And we're giving them hundreds of

579
00:31:59,120 --> 00:32:01,720
Speaker 2: billions of dollars and the smartest people on the planet

580
00:32:01,840 --> 00:32:03,880
Speaker 2: to make that happen. And they seem to be making

581
00:32:04,040 --> 00:32:07,840
Speaker 2: very real steps toward the realization of that goal. And

582
00:32:07,920 --> 00:32:09,640
Speaker 2: so it's a bit of hope to always say, well,

583
00:32:09,680 --> 00:32:12,160
Speaker 2: we'll have some kind of new job, you know, this time,

584
00:32:12,200 --> 00:32:14,680
Speaker 2: it seems like they're actually going to come and take

585
00:32:14,720 --> 00:32:18,080
Speaker 2: a lot of our jobs away, interestingly white collar first,

586
00:32:18,480 --> 00:32:21,160
Speaker 2: but as AI gets embedded in robotics, it will come

587
00:32:21,200 --> 00:32:23,880
Speaker 2: for the blue collar jobs too, and the goal is

588
00:32:23,920 --> 00:32:25,840
Speaker 2: to displace all of us, and they seem to be

589
00:32:25,880 --> 00:32:28,400
Speaker 2: making steps toward it. And so that's a very glib

590
00:32:28,480 --> 00:32:32,120
Speaker 2: thing to say when said you already see the software

591
00:32:32,120 --> 00:32:36,640
Speaker 2: engineer market being crushed, unemployment rising, and again you're probably

592
00:32:36,640 --> 00:32:38,360
Speaker 2: going to see more of this going into the future. So,

593
00:32:38,800 --> 00:32:41,280
Speaker 2: you know, I hope that person's right. One doesn't really

594
00:32:41,280 --> 00:32:43,200
Speaker 2: know how this will develop, but you have to take

595
00:32:43,240 --> 00:32:45,800
Speaker 2: it seriously that this time it's very different.

596
00:32:46,040 --> 00:32:47,280
Speaker 1: Yeah, Mark, what do you saying?

597
00:32:48,480 --> 00:32:52,440
Speaker 3: You know, I asked Claude Opus for who would be

598
00:32:52,480 --> 00:32:55,760
Speaker 3: the most likely the United States senator to oppose federal

599
00:32:55,760 --> 00:32:58,920
Speaker 3: preemption of state regulation, and they answer back with Senator Cruise.

600
00:33:00,480 --> 00:33:03,600
Speaker 3: And you know, I think I think we, of course are.

601
00:33:03,920 --> 00:33:06,000
Speaker 3: We want to be optimistic, we also want to be

602
00:33:06,040 --> 00:33:08,280
Speaker 3: clear eyed about the issue. We can't just pretend and

603
00:33:08,400 --> 00:33:11,440
Speaker 3: handwave that this is all just going to be fine.

604
00:33:11,560 --> 00:33:13,240
Speaker 3: And you know, if you look at what happened during

605
00:33:13,240 --> 00:33:15,600
Speaker 3: the Industrial Revolution, which is what a lot of folks

606
00:33:15,640 --> 00:33:20,800
Speaker 3: compare this to. You know, human physical labor was automated

607
00:33:21,400 --> 00:33:25,719
Speaker 3: and it became next to zero related to GDP output today.

608
00:33:26,120 --> 00:33:28,920
Speaker 3: So physical strength became automated, and that's now not a

609
00:33:29,000 --> 00:33:31,000
Speaker 3: function a factor of the economy. Well, now what we're

610
00:33:31,120 --> 00:33:34,479
Speaker 3: doing is automating our intelligence, and so the question is like,

611
00:33:34,520 --> 00:33:36,360
Speaker 3: what's what's left at that point? Are are you know,

612
00:33:36,360 --> 00:33:39,160
Speaker 3: maybe people can have access to their own, you know,

613
00:33:39,360 --> 00:33:41,520
Speaker 3: very powerful AI systems, and those AIS can go out

614
00:33:41,560 --> 00:33:44,520
Speaker 3: on the Internet and do work for them. But it's again,

615
00:33:44,520 --> 00:33:46,880
Speaker 3: it's it's quite under specified. And I admit the fact

616
00:33:46,920 --> 00:33:49,360
Speaker 3: that I might not be intellectually intelligent enough to predict

617
00:33:49,640 --> 00:33:52,440
Speaker 3: what's going to happen. But I think this seems to

618
00:33:52,440 --> 00:33:54,360
Speaker 3: be a bit of a platitude in a handwave to

619
00:33:54,440 --> 00:33:57,160
Speaker 3: keep people from panicking, and we might be at the

620
00:33:57,200 --> 00:33:59,280
Speaker 3: ponment where it's time to have some of that panic.

621
00:34:00,080 --> 00:34:02,640
Speaker 1: Yeah, someone said that there's the possibility that within the

622
00:34:02,680 --> 00:34:05,480
Speaker 1: next year, I forget who was as an AI person

623
00:34:05,480 --> 00:34:07,600
Speaker 1: on Twitter, So take it with a very small grainssalt.

624
00:34:07,880 --> 00:34:10,320
Speaker 1: But they said that it is possible for a single

625
00:34:10,360 --> 00:34:13,080
Speaker 1: person to have no employees and make a company that's

626
00:34:13,320 --> 00:34:15,600
Speaker 1: valued in the tens of millions of dollars, and then

627
00:34:15,640 --> 00:34:18,440
Speaker 1: the very near future will work the billions of dollars.

628
00:34:18,880 --> 00:34:22,840
Speaker 1: Weird though, that it feels like people are having conversations

629
00:34:22,960 --> 00:34:25,600
Speaker 1: past each other about something that will affect all of us.

630
00:34:25,760 --> 00:34:27,000
Speaker 1: Do you guys get that feeling?

631
00:34:28,440 --> 00:34:30,520
Speaker 2: I definitely think people are talking a bit past one.

632
00:34:30,520 --> 00:34:33,040
Speaker 2: Another part of it is it's still a very technical

633
00:34:33,080 --> 00:34:35,920
Speaker 2: and fairly new technology for most people, and so they

634
00:34:35,920 --> 00:34:39,120
Speaker 2: don't really understand it. They're not aware of what's really happening,

635
00:34:39,440 --> 00:34:41,759
Speaker 2: and even if they're aware of what's happening today, they're

636
00:34:41,800 --> 00:34:44,800
Speaker 2: not paying attention to where it's going. It's rapidly improving,

637
00:34:45,120 --> 00:34:47,160
Speaker 2: so it's not we don't worry so much about what's

638
00:34:47,160 --> 00:34:49,799
Speaker 2: happening at this moment, it's what's happening in a year,

639
00:34:50,000 --> 00:34:53,919
Speaker 2: two years, five years. Where they've actually radically improved even

640
00:34:53,920 --> 00:34:57,120
Speaker 2: today is quite remarkable capabilities. So I think people to

641
00:34:57,239 --> 00:35:00,000
Speaker 2: understand it, they want to believe things will just be okay.

642
00:35:00,600 --> 00:35:03,600
Speaker 2: There's a lot of reflexive opposition to any kind of regulation,

643
00:35:03,800 --> 00:35:06,920
Speaker 2: especially if you're working in Republican politics right regulating business.

644
00:35:07,000 --> 00:35:09,920
Speaker 2: It's just the per se bad which I understand that,

645
00:35:09,960 --> 00:35:12,560
Speaker 2: and they're not often wrong to believe that. But here

646
00:35:12,640 --> 00:35:15,640
Speaker 2: is a technology that's openly stating it's going to try

647
00:35:15,640 --> 00:35:19,759
Speaker 2: to transform our lives up in our world, and it's

648
00:35:19,760 --> 00:35:23,319
Speaker 2: probably worth paying attention to and putting in like reasonable safeguards,

649
00:35:23,360 --> 00:35:26,000
Speaker 2: where like Mark suggests, at the beginning, we're at least

650
00:35:26,280 --> 00:35:28,600
Speaker 2: having the capabilities to deal with it. We're getting more

651
00:35:28,640 --> 00:35:31,560
Speaker 2: information about where it's going, We're watching what jobs are

652
00:35:31,600 --> 00:35:35,120
Speaker 2: being displaced, and we begin the conversation very difficult one

653
00:35:35,360 --> 00:35:38,799
Speaker 2: about how this kind of social compact to undergirds our

654
00:35:38,840 --> 00:35:40,760
Speaker 2: life might radically change.

655
00:35:41,160 --> 00:35:42,879
Speaker 1: Okay, so I want to ask you both of our

656
00:35:42,960 --> 00:35:46,960
Speaker 1: last question. What is a state and what is a

657
00:35:47,000 --> 00:35:51,040
Speaker 1: federal regulation on AI that is not in place currently

658
00:35:51,200 --> 00:35:53,239
Speaker 1: that should be in place if you were speaking to

659
00:35:53,280 --> 00:35:56,160
Speaker 1: a state legislator or governor or someone either in Congress

660
00:35:56,200 --> 00:35:59,360
Speaker 1: or the president. So you can go first, Mark.

661
00:36:00,280 --> 00:36:03,360
Speaker 3: I mean, I think the most important thing the federal

662
00:36:03,400 --> 00:36:06,400
Speaker 3: government ought to be doing time now is taking the

663
00:36:06,440 --> 00:36:10,880
Speaker 3: testing and evaluation regime at the classified level very seriously.

664
00:36:11,680 --> 00:36:15,560
Speaker 3: I think having the data to make inform regulatory choices

665
00:36:15,680 --> 00:36:18,439
Speaker 3: is a foundational first step. And given everything Brad said

666
00:36:18,440 --> 00:36:22,200
Speaker 3: about the technical nature, the relative opacity, the relative fact

667
00:36:22,239 --> 00:36:25,040
Speaker 3: that you know, Washington's quite behind on understanding whether I

668
00:36:25,040 --> 00:36:27,520
Speaker 3: mean the Vatican seems to be further ahead than Washington,

669
00:36:27,600 --> 00:36:31,480
Speaker 3: d C. I'm appreciating the significance of this moment and

670
00:36:31,480 --> 00:36:34,840
Speaker 3: then I would urge my friends and the Republicans Republican Party,

671
00:36:35,760 --> 00:36:38,120
Speaker 3: you know, we can't take a country club attitude towards

672
00:36:38,120 --> 00:36:41,120
Speaker 3: this one. And I agree with that. The fact that,

673
00:36:41,160 --> 00:36:44,319
Speaker 3: you know, the government's oftentimes very ham fisted when it

674
00:36:44,360 --> 00:36:47,040
Speaker 3: engages in the economy. It's not efficient, but it does

675
00:36:47,080 --> 00:36:49,319
Speaker 3: have an important role to play. And we can't, as

676
00:36:49,360 --> 00:36:52,920
Speaker 3: Mike Davis use the term, let the tech brodogarchs you know,

677
00:36:53,120 --> 00:36:55,240
Speaker 3: run the show. We have to think about the broader

678
00:36:55,560 --> 00:36:59,239
Speaker 3: social compact and issues associated with that. And so, but

679
00:36:59,320 --> 00:37:00,839
Speaker 3: the first thing we need to do to get through

680
00:37:00,840 --> 00:37:03,640
Speaker 3: some of this talking past each other is generating the

681
00:37:03,719 --> 00:37:06,840
Speaker 3: right data and having that informed conversation being grounded in

682
00:37:06,880 --> 00:37:08,840
Speaker 3: the facts, and then from there we can have actually

683
00:37:08,840 --> 00:37:09,840
Speaker 3: a productive discussion.

684
00:37:10,440 --> 00:37:11,000
Speaker 2: That's great.

685
00:37:11,160 --> 00:37:13,239
Speaker 1: I love the idea of the country club attitude because

686
00:37:13,280 --> 00:37:20,000
Speaker 1: that is pervasive still despite the whole working class overtures

687
00:37:20,120 --> 00:37:23,759
Speaker 1: being made at least verbally towards voters. Right now, it

688
00:37:23,840 --> 00:37:26,480
Speaker 1: is still very much a country club attitude. What on't you, Brad,

689
00:37:26,520 --> 00:37:29,320
Speaker 1: what would you say is a state or federal policy

690
00:37:29,360 --> 00:37:31,279
Speaker 1: that it's not an ACTI currently that should be an actor,

691
00:37:31,320 --> 00:37:32,680
Speaker 1: or it's an acted in one state and all the

692
00:37:32,719 --> 00:37:33,719
Speaker 1: other states should adopt it.

693
00:37:34,680 --> 00:37:37,239
Speaker 2: I think what marks is something I would associate myself with,

694
00:37:37,320 --> 00:37:39,680
Speaker 2: but I would add just this. We have the AI

695
00:37:39,760 --> 00:37:44,280
Speaker 2: Safety Institute, something that the Treasury Department yesterday renamed as CASEY,

696
00:37:44,440 --> 00:37:48,920
Speaker 2: the Center for AI Standards. So having an institution in

697
00:37:48,960 --> 00:37:52,440
Speaker 2: the federal government that can collect data that brings together expertise.

698
00:37:52,760 --> 00:37:55,040
Speaker 2: It can be the repository of this kind of data

699
00:37:55,120 --> 00:37:58,280
Speaker 2: and to help work with the frontier labs to get

700
00:37:58,360 --> 00:38:01,279
Speaker 2: information and to fund them adequately. But they have the

701
00:38:01,360 --> 00:38:04,399
Speaker 2: right people who are often quite high paid and very

702
00:38:04,520 --> 00:38:08,400
Speaker 2: high levels of skill. That's an important kind of bedrock

703
00:38:08,480 --> 00:38:10,560
Speaker 2: policy we have to have in place. We've had something

704
00:38:10,640 --> 00:38:12,880
Speaker 2: like that, but it's never been codified. Right, it's an

705
00:38:12,880 --> 00:38:15,800
Speaker 2: executive order Biden. Now they've changed a little bit of

706
00:38:15,880 --> 00:38:18,200
Speaker 2: the President Trump. We need to put that into law.

707
00:38:18,480 --> 00:38:21,080
Speaker 2: We have an institution, it's like, dedicated to looking at

708
00:38:21,120 --> 00:38:22,840
Speaker 2: what's happening AI in our economy.

709
00:38:23,120 --> 00:38:27,160
Speaker 1: It's politicians is one class of jobs that probably will

710
00:38:27,160 --> 00:38:30,080
Speaker 1: not be automated, and that's probably why we're not seeing

711
00:38:30,200 --> 00:38:32,600
Speaker 1: much moving on it. Uh Bra, Where can people go

712
00:38:32,640 --> 00:38:33,960
Speaker 1: to find more of your work.

713
00:38:34,800 --> 00:38:37,880
Speaker 2: In the American's responsible innovation can be found at AARI

714
00:38:38,000 --> 00:38:40,600
Speaker 2: dot us. You know, read about all our policies, what

715
00:38:40,640 --> 00:38:43,520
Speaker 2: we support, blog posts, links to many of the kind

716
00:38:43,560 --> 00:38:46,919
Speaker 2: of other thinkers that we find important the work we're doing.

717
00:38:47,000 --> 00:38:48,880
Speaker 2: So ar dot us.

718
00:38:49,080 --> 00:38:52,080
Speaker 1: And Mark where can people find more about them? Government

719
00:38:52,080 --> 00:38:54,160
Speaker 1: affairs for the AI Policy Network.

720
00:38:54,320 --> 00:38:59,640
Speaker 3: Yeah, THEAIPN dot org. We're focused on accelerating federal preparedness

721
00:38:59,680 --> 00:39:04,080
Speaker 3: for tras formative AI, looking at national security, at economic security,

722
00:39:04,440 --> 00:39:07,600
Speaker 3: and human flourishing, and would be excited to partner with

723
00:39:07,640 --> 00:39:10,320
Speaker 3: anyone out there who wants to help promote education around

724
00:39:10,360 --> 00:39:14,080
Speaker 3: this content and help Congress make informed decisions that's going

725
00:39:14,160 --> 00:39:16,280
Speaker 3: to be driving towards the betterment of the American people.

726
00:39:17,120 --> 00:39:18,880
Speaker 1: Well, you guys do great work, so I hope that

727
00:39:19,760 --> 00:39:22,120
Speaker 1: hope they listen, and hope that we are not as

728
00:39:22,120 --> 00:39:24,600
Speaker 1: slow moving on AI as we were on social media

729
00:39:24,640 --> 00:39:26,719
Speaker 1: and everything else. So thank you both for being here.

730
00:39:27,040 --> 00:39:27,839
Speaker 3: Thank you so much.

731
00:39:28,160 --> 00:39:30,480
Speaker 1: You're listening to It's the Numbers Game with Ryan Grodowski.

732
00:39:30,560 --> 00:39:36,239
Speaker 1: We'll be right back after this message. Okay, I know

733
00:39:36,320 --> 00:39:38,080
Speaker 1: the show's gone a little long today and I hope

734
00:39:38,080 --> 00:39:40,600
Speaker 1: you've learned something I did. But before we go, I

735
00:39:40,640 --> 00:39:43,040
Speaker 1: need to do the listener question they Ask Me Anything

736
00:39:43,080 --> 00:39:46,359
Speaker 1: segment you This might be actually a great episode one

737
00:39:46,400 --> 00:39:48,759
Speaker 1: day to do just ask me anythings. If I get

738
00:39:48,840 --> 00:39:51,000
Speaker 1: enough emails, I would love to do that. If you

739
00:39:51,040 --> 00:39:52,720
Speaker 1: want to be part of the ask Me Anything segment,

740
00:39:52,760 --> 00:39:56,000
Speaker 1: please m them email me Ryan at Numbers Game podcast

741
00:39:56,080 --> 00:40:00,359
Speaker 1: dot com. That's ryanat Numbers gamepodcast dot com. Today's question

742
00:40:00,400 --> 00:40:03,440
Speaker 1: comes from David Inda. He writes, my name is Dave.

743
00:40:03,480 --> 00:40:06,040
Speaker 1: I'm a huge fan of your podcast. My question is

744
00:40:06,080 --> 00:40:09,759
Speaker 1: regarding the changing voting patterns of different demographics in the US,

745
00:40:09,800 --> 00:40:13,640
Speaker 1: specifically black and Latino, and any possible conversations you may

746
00:40:13,640 --> 00:40:15,400
Speaker 1: have had with your friend Ann Kolzer, who I'm a

747
00:40:15,400 --> 00:40:18,120
Speaker 1: massive fan of AND's great. That's not the thing, but

748
00:40:18,280 --> 00:40:20,960
Speaker 1: and is great anyway, and talks extensively how the GOP

749
00:40:21,000 --> 00:40:23,480
Speaker 1: should only ever focus on winning over white voters. But

750
00:40:23,520 --> 00:40:25,960
Speaker 1: in light of the recent data that has been released

751
00:40:25,960 --> 00:40:28,200
Speaker 1: post election, I am wondering have you tried to convince

752
00:40:28,239 --> 00:40:31,200
Speaker 1: her that the GOP tried to win black and Latino voters?

753
00:40:31,280 --> 00:40:33,200
Speaker 1: Is that a good idea? Or do you think the

754
00:40:33,239 --> 00:40:35,719
Speaker 1: only thing Republicans should talk about is crime and immigration?

755
00:40:35,880 --> 00:40:38,919
Speaker 1: This is what I believe, and these two policies seems

756
00:40:38,960 --> 00:40:41,959
Speaker 1: to be universally popular regarding race and ethnicity. I would

757
00:40:41,960 --> 00:40:43,920
Speaker 1: love to hear your thoughts on this. Also, I currently

758
00:40:43,960 --> 00:40:46,560
Speaker 1: live in Michigan planning to move to Tennessee, and I'm

759
00:40:46,600 --> 00:40:49,280
Speaker 1: wondering do Republicans have any chance of winning the open

760
00:40:49,360 --> 00:40:51,680
Speaker 1: Senate or governor seat in Michigan in twenty twenty six.

761
00:40:51,760 --> 00:40:55,000
Speaker 1: Along the same lines, why are some states voting one

762
00:40:55,000 --> 00:40:57,520
Speaker 1: way at state levels at either red or blue and

763
00:40:57,520 --> 00:41:00,160
Speaker 1: the opposite of presidential elections? Namely places like George in

764
00:41:00,200 --> 00:41:02,840
Speaker 1: New Hampshire and Michigan come to mind? And this question

765
00:41:03,160 --> 00:41:06,279
Speaker 1: is party affiliation as salient as people typically believe. Where

766
00:41:06,280 --> 00:41:08,880
Speaker 1: does candidate quality in a particular year have a bigger

767
00:41:08,920 --> 00:41:11,200
Speaker 1: impact on the people's votes. I know there's a lot

768
00:41:11,239 --> 00:41:13,480
Speaker 1: of rambling, so I apologize and massive fan of the

769
00:41:13,480 --> 00:41:17,440
Speaker 1: podcast and all the important work you do. One final question, Sorry,

770
00:41:17,480 --> 00:41:19,759
Speaker 1: the ADHD is really strong today. I am originally from

771
00:41:19,800 --> 00:41:22,439
Speaker 1: Illinois and wondering if there's any hope for a place

772
00:41:22,480 --> 00:41:24,560
Speaker 1: like that or is it goondla with California? Wow? Dave, Okay,

773
00:41:24,960 --> 00:41:27,759
Speaker 1: lots of questions. Highly recommend switching a decaf, but I

774
00:41:27,840 --> 00:41:30,400
Speaker 1: love the passion I have ADHD two so I know

775
00:41:30,640 --> 00:41:33,200
Speaker 1: how difficult the struggle com based going to go through

776
00:41:33,200 --> 00:41:35,440
Speaker 1: as quickly. Okay, so first and a culture. I try

777
00:41:35,480 --> 00:41:38,480
Speaker 1: to keep our conversations private. But what I believe Annas

778
00:41:38,480 --> 00:41:41,040
Speaker 1: said is that they shouldn't try to win over non

779
00:41:41,120 --> 00:41:45,840
Speaker 1: white voters by remaking policies like we should try to

780
00:41:45,880 --> 00:41:50,920
Speaker 1: win over non white voters, but focus on policies that

781
00:41:51,000 --> 00:41:53,880
Speaker 1: actually help our voter base, which is namely whites without

782
00:41:53,920 --> 00:41:57,400
Speaker 1: a college degree. We haven't talked about the demographic breakdown

783
00:41:57,400 --> 00:41:59,360
Speaker 1: the election yet. It's something we're trying to get a

784
00:41:59,440 --> 00:42:01,720
Speaker 1: dinner around to sit there and go through it and

785
00:42:01,760 --> 00:42:04,560
Speaker 1: we talk about it. I'll send her your woman's regards.

786
00:42:04,560 --> 00:42:06,680
Speaker 1: But I think that's what ann really says. It's not

787
00:42:06,719 --> 00:42:08,920
Speaker 1: that we shouldn't try, It's that we shouldn't sell out

788
00:42:08,960 --> 00:42:12,040
Speaker 1: on policies like on crime, immigration, which are two very

789
00:42:12,040 --> 00:42:15,040
Speaker 1: popular policies. Secondly, do Republicans have a chance at winning

790
00:42:15,080 --> 00:42:18,520
Speaker 1: the Michigan governor or Senate election next year? Both seats

791
00:42:18,600 --> 00:42:21,760
Speaker 1: are going to be vacant, that is correct. My bet

792
00:42:21,840 --> 00:42:24,440
Speaker 1: is on the governor's race. And here's why. Mike Dugan

793
00:42:24,800 --> 00:42:28,160
Speaker 1: Dugan whatever's last name is the Democrat mayor of Detroit

794
00:42:28,239 --> 00:42:31,120
Speaker 1: is running as an independent and will likely take more

795
00:42:31,200 --> 00:42:34,880
Speaker 1: votes from Democrats than Republicans. Duggan has been on the

796
00:42:35,000 --> 00:42:37,040
Speaker 1: rise in the polls, but it's essentially a toss up

797
00:42:37,040 --> 00:42:41,680
Speaker 1: between the likely Democrat Joyce Leen Benson and Republican John James.

798
00:42:42,280 --> 00:42:44,279
Speaker 1: As for the Senate seat, it just depends on who

799
00:42:44,280 --> 00:42:47,600
Speaker 1: the Democrats nominate, because Republicans will likely pick Mike Rogers again.

800
00:42:48,080 --> 00:42:51,400
Speaker 1: Early polls show it's close, but the only canet Rogers

801
00:42:51,400 --> 00:42:54,680
Speaker 1: has a decent lead against is Wayne County Health director

802
00:42:54,719 --> 00:42:58,960
Speaker 1: and Bernie Sanders supporter. Abdul l said, I think that's

803
00:42:58,960 --> 00:43:02,279
Speaker 1: how you pronounce that name. So we'll see if that

804
00:43:02,840 --> 00:43:05,040
Speaker 1: you know who they pick. That's a big question. And

805
00:43:05,239 --> 00:43:08,160
Speaker 1: how many third party nominees get on the ballot is

806
00:43:08,200 --> 00:43:10,360
Speaker 1: a big question. There were a lot of third party, conservative,

807
00:43:10,400 --> 00:43:14,160
Speaker 1: third party, libertarian, third party. Last time, they took enough

808
00:43:14,200 --> 00:43:15,759
Speaker 1: of the vote to matter. I think he lost by

809
00:43:15,840 --> 00:43:18,360
Speaker 1: nineteen thousand votes and they took well over two hundred

810
00:43:18,360 --> 00:43:21,400
Speaker 1: thousand or something like that. Anyway, why do you states

811
00:43:21,440 --> 00:43:24,360
Speaker 1: vote differently in federal elections state elections? This depends on

812
00:43:24,360 --> 00:43:27,000
Speaker 1: one canon. Quality does matter. If you run some link

813
00:43:27,040 --> 00:43:29,440
Speaker 1: Kerry Lake, you're probably going to lose irregardless of the

814
00:43:29,440 --> 00:43:33,880
Speaker 1: fact that there's more Republicans than Democrats. Party affiliation matters

815
00:43:34,080 --> 00:43:37,080
Speaker 1: to a point. It is the most likely outcome. If

816
00:43:37,080 --> 00:43:39,600
Speaker 1: you're readister Democrat, you will most likely vote Democrat. That's

817
00:43:39,719 --> 00:43:43,080
Speaker 1: the biggest consideration than any almost anything else. Right, I

818
00:43:43,080 --> 00:43:46,279
Speaker 1: think gun ownership is the only other larger indicator of

819
00:43:46,320 --> 00:43:50,080
Speaker 1: how you will vote than party affiliation. But party affiliation

820
00:43:50,360 --> 00:43:53,919
Speaker 1: is very, very important. But in places like Kentucky which

821
00:43:54,040 --> 00:43:57,480
Speaker 1: vote for Republicans federally but Democrats locally, or Vermont, which

822
00:43:57,560 --> 00:44:02,520
Speaker 1: vote for Republicans for govern but Democrats in the legislature

823
00:44:02,560 --> 00:44:06,200
Speaker 1: and federally, a lot of it matters in the minds

824
00:44:06,200 --> 00:44:11,520
Speaker 1: of voters is who can put a check on the legislature. So,

825
00:44:11,719 --> 00:44:15,640
Speaker 1: because the working class people of Kentucky oftentimes have a

826
00:44:15,680 --> 00:44:19,520
Speaker 1: Republican party that has a lot of country club attitudes,

827
00:44:19,520 --> 00:44:22,480
Speaker 1: as I said earlier on this podcast, they will support

828
00:44:22,560 --> 00:44:25,480
Speaker 1: a Democrat who's more pro worker, pro union. I mean

829
00:44:25,480 --> 00:44:29,040
Speaker 1: it was a coal mining state forever Likewise, Vermont, they

830
00:44:29,080 --> 00:44:33,280
Speaker 1: will support a socially liberal Republican who's more pro economic

831
00:44:33,320 --> 00:44:36,480
Speaker 1: growth because the legislature is so anti economic growth. A

832
00:44:36,520 --> 00:44:38,920
Speaker 1: lot of it depends on what the state's going through

833
00:44:39,360 --> 00:44:43,360
Speaker 1: and the character and how the character of the candidate

834
00:44:43,360 --> 00:44:45,480
Speaker 1: and how much they reflect the interest of that state.

835
00:44:45,840 --> 00:44:47,920
Speaker 1: The last question was on Illinois, has it gone the

836
00:44:47,920 --> 00:44:50,239
Speaker 1: way of California. I actually think California is in a

837
00:44:50,280 --> 00:44:54,560
Speaker 1: better place in respect of how Republicans are winning over

838
00:44:54,640 --> 00:44:58,760
Speaker 1: voters than Illinois Black voters, which make a bigger population

839
00:44:58,840 --> 00:45:02,080
Speaker 1: Illinois than in CALIFORNI and you are moving slower at

840
00:45:02,120 --> 00:45:04,520
Speaker 1: a slower speed towards the GP than Asians and Hispanics,

841
00:45:04,880 --> 00:45:09,000
Speaker 1: and southern California whites move towards the Republicans faster than

842
00:45:09,120 --> 00:45:13,160
Speaker 1: Chicago area whites. So I think we're going to see

843
00:45:13,160 --> 00:45:16,360
Speaker 1: more gains in California than in Illinois, both in Congress

844
00:45:16,360 --> 00:45:18,120
Speaker 1: and the state legislature. I know it's probably not the

845
00:45:18,120 --> 00:45:20,400
Speaker 1: answer you want to hear, but that's how I see it.

846
00:45:20,440 --> 00:45:24,479
Speaker 1: Thank you again, though for listening. I really appreciate your email, Dave.

847
00:45:24,640 --> 00:45:27,200
Speaker 1: If you anyone else on any other emails, please email me,

848
00:45:27,239 --> 00:45:30,200
Speaker 1: and once again Ryan at numbers gamepodcast dot com. Thank

849
00:45:30,239 --> 00:45:33,080
Speaker 1: you for listening. Please like and subscribe to this podcast

850
00:45:33,120 --> 00:45:37,160
Speaker 1: on the iHeartRadio app, Apple podcast, wherever you get your podcasts.

851
00:45:37,160 --> 00:45:39,000
Speaker 1: Give me a five star review if you're feeling generous,

852
00:45:39,040 --> 00:45:42,279
Speaker 1: that really helps people find the show. And I will

853
00:45:42,280 --> 00:45:44,160
Speaker 1: see you guys on Monday. Thank you again.