1
00:00:04,880 --> 00:00:08,200
Speaker 1: From The Australian. Here's what's on the front. I'm Christinamiot.

2
00:00:08,440 --> 00:00:16,720
Speaker 1: It's Wednesday, January twenty nine. Athletics Australia has lashed at

3
00:00:16,720 --> 00:00:20,000
Speaker 1: the Western Australian Institute of Sport for its treatment of

4
00:00:20,040 --> 00:00:22,840
Speaker 1: a pole volt coach who blew the whistle on sexual

5
00:00:22,840 --> 00:00:27,760
Speaker 1: assault allegations. Chair Jane Fleming and CEO Simon Hollingsworth have

6
00:00:27,880 --> 00:00:30,680
Speaker 1: called for the WA body to better support whistle blow

7
00:00:30,760 --> 00:00:38,839
Speaker 1: up Paul Burgess and reinstate his training privileges. New Zealand's

8
00:00:38,840 --> 00:00:42,440
Speaker 1: Resources minister has launched an attack on banks that refuse

9
00:00:42,600 --> 00:00:45,640
Speaker 1: to finance minds on the eve of the release of

10
00:00:45,680 --> 00:00:49,479
Speaker 1: his government's mineral strategy. Shane Jones also backed at the

11
00:00:49,479 --> 00:00:54,840
Speaker 1: Fijian PM in saying Australia's not to blame for climate change.

12
00:00:54,960 --> 00:00:57,640
Speaker 1: Those exclusive stories are live right now at the Australian

13
00:00:57,680 --> 00:01:03,800
Speaker 1: dot Com dot Au. Science Minister Ed Husick says a

14
00:01:03,880 --> 00:01:08,720
Speaker 1: new Chinese AI platform comes with national security concerns, and

15
00:01:08,800 --> 00:01:12,000
Speaker 1: he says he'll take advice from intelligence experts on any

16
00:01:12,000 --> 00:01:16,720
Speaker 1: potential threats. Deep Seek sent shockwaves through the tech industry

17
00:01:16,760 --> 00:01:20,600
Speaker 1: and financial markets this week. So what is it and

18
00:01:20,800 --> 00:01:35,400
Speaker 1: where's it sending the AI race. That's today's episode. Last week,

19
00:01:35,480 --> 00:01:38,759
Speaker 1: on the frenzied first day of his new presidential term,

20
00:01:38,920 --> 00:01:42,839
Speaker 1: Donald Trump announced a multi billion dollar commitment to artificial

21
00:01:42,880 --> 00:01:44,320
Speaker 1: intelligence infrastructure.

22
00:01:45,400 --> 00:01:50,800
Speaker 2: These world leading technology giants are announcing the formation of Stargate,

23
00:01:51,080 --> 00:01:54,400
Speaker 2: a new American company that will invest five hundred billion

24
00:01:54,440 --> 00:01:58,000
Speaker 2: dollars at least in AI infrastructure in the United States,

25
00:01:58,040 --> 00:02:00,680
Speaker 2: creating over one hundred thousand are jobs.

26
00:02:00,680 --> 00:02:04,640
Speaker 1: Almost immediately about the same time, on the other side

27
00:02:04,640 --> 00:02:08,640
Speaker 1: of the planet, a Chinese company called deep Seek unveiled

28
00:02:08,639 --> 00:02:12,080
Speaker 1: an AI model it says can compete with market leaders

29
00:02:12,120 --> 00:02:17,120
Speaker 1: like Chat GPT for a fraction of the cost. The

30
00:02:17,160 --> 00:02:21,680
Speaker 1: world's financial markets reacted with shock and awe with Navidia,

31
00:02:21,720 --> 00:02:24,920
Speaker 1: the world's most valuable public company and the maker of

32
00:02:24,960 --> 00:02:29,080
Speaker 1: the computer chips most American AI models rely on bearing

33
00:02:29,200 --> 00:02:32,239
Speaker 1: the brunt of it Well, shairs of Santa Clara's in

34
00:02:32,320 --> 00:02:35,360
Speaker 1: video fell hard this morning on the popularity of a

35
00:02:35,400 --> 00:02:39,960
Speaker 1: new Chinese artificial intelligence advancement. Deepseek is impressive enough, but

36
00:02:40,040 --> 00:02:42,600
Speaker 1: it made people wonder, do you really need those Nvidia

37
00:02:42,760 --> 00:02:43,480
Speaker 1: AI chaps?

38
00:02:43,560 --> 00:02:46,160
Speaker 2: That's the whole question, right, I mean ignoring the fact

39
00:02:46,160 --> 00:02:48,280
Speaker 2: that a number of reports.

40
00:02:48,080 --> 00:02:58,920
Speaker 3: That I think it caught everyone off guard, even the

41
00:02:58,919 --> 00:03:01,480
Speaker 3: best and brightest in Wall Street and in Silicon Valley.

42
00:03:02,200 --> 00:03:04,680
Speaker 1: Jared Lynch is The Australian's Technology editor.

43
00:03:05,480 --> 00:03:09,680
Speaker 3: They thought the US was miles ahead, not even miles ahead,

44
00:03:09,720 --> 00:03:12,440
Speaker 3: thousands of miles ahead than China. They had a trade

45
00:03:12,440 --> 00:03:16,280
Speaker 3: embargo on Navidia chips, the world's most advanced chips. And

46
00:03:16,320 --> 00:03:19,359
Speaker 3: then the fact that China could produce at a far

47
00:03:19,440 --> 00:03:23,440
Speaker 3: lower costs, mind you, an AI model that is almost on.

48
00:03:23,360 --> 00:03:26,600
Speaker 4: Par with chat GPT was extraordinary.

49
00:03:30,400 --> 00:03:33,880
Speaker 1: The AI model released by deep Seek is called R one.

50
00:03:34,440 --> 00:03:38,120
Speaker 1: It's a generative AI platform similar to Chat GPT and

51
00:03:38,320 --> 00:03:42,200
Speaker 1: Meta AI. Deep Seek is backed by Leoon Wen Fung.

52
00:03:42,600 --> 00:03:45,440
Speaker 1: He's the manager of the eight billion dollar hedge fund

53
00:03:45,640 --> 00:03:49,440
Speaker 1: high Flyer. The company says it forked out just under

54
00:03:49,560 --> 00:03:53,080
Speaker 1: nine million Australian dollars to train R one, and says

55
00:03:53,120 --> 00:03:56,560
Speaker 1: it did it without Navidia's top of the line computer chips.

56
00:03:57,400 --> 00:04:01,240
Speaker 1: By comparison, the most recent iteration of chat GPT cost

57
00:04:01,280 --> 00:04:04,560
Speaker 1: around one hundred million dollars to train before it was

58
00:04:04,640 --> 00:04:09,120
Speaker 1: released in November, and deep seeks model is open source,

59
00:04:09,520 --> 00:04:12,880
Speaker 1: meaning other developers can adopt and adapt it. For their

60
00:04:12,920 --> 00:04:20,120
Speaker 1: own purposes, hence the financial panic. Jared was that market

61
00:04:20,160 --> 00:04:21,440
Speaker 1: reaction proportionate?

62
00:04:21,560 --> 00:04:23,719
Speaker 4: Do you think time will tell?

63
00:04:23,880 --> 00:04:26,880
Speaker 3: There have been some analysts that have called bs to

64
00:04:26,960 --> 00:04:30,960
Speaker 3: put it politely on deep seaks claims that is that

65
00:04:31,000 --> 00:04:34,240
Speaker 3: it only cost five point six million US to train

66
00:04:34,320 --> 00:04:38,000
Speaker 3: this model, whereas some US models have cost billions to

67
00:04:38,080 --> 00:04:41,640
Speaker 3: train because there's just not enough data to show. Okay,

68
00:04:42,240 --> 00:04:44,920
Speaker 3: how much does this cost to train? And how much

69
00:04:45,120 --> 00:04:48,080
Speaker 3: do the data centers cost to run deep seats? So

70
00:04:48,120 --> 00:04:51,039
Speaker 3: that's all unknown at this stage. At the moment, the

71
00:04:51,080 --> 00:04:55,400
Speaker 3: market's caught up in Wow, China's actually delivered this AI model.

72
00:04:55,880 --> 00:04:59,000
Speaker 3: What risk does that pose to the big American companies

73
00:04:59,000 --> 00:05:01,560
Speaker 3: that have invested billions in creating this tech.

74
00:05:02,320 --> 00:05:03,560
Speaker 4: But the thing is we've got to.

75
00:05:03,560 --> 00:05:06,000
Speaker 3: Rewind the clock back a little bit to the fifties,

76
00:05:06,040 --> 00:05:08,880
Speaker 3: to the sixties, even the early seventies, off the space race,

77
00:05:09,240 --> 00:05:13,320
Speaker 3: the US against Russia. That didn't result in a pullback

78
00:05:13,360 --> 00:05:17,840
Speaker 3: off investment. I mean, competition rarely does it becomes a battle.

79
00:05:18,600 --> 00:05:22,120
Speaker 3: Billions of dollars, trillions of dollars get poured into gaining

80
00:05:22,200 --> 00:05:24,240
Speaker 3: the upper hand, and that is what we're going.

81
00:05:24,160 --> 00:05:25,320
Speaker 4: To see in the AI race.

82
00:05:26,240 --> 00:05:28,760
Speaker 3: What the response has been, more broadly, is this is

83
00:05:28,880 --> 00:05:31,760
Speaker 3: very much now game on. China is saying to the

84
00:05:31,800 --> 00:05:34,760
Speaker 3: world that it can compete and at a far lower cost.

85
00:05:34,960 --> 00:05:39,040
Speaker 3: In fact, that's almost embarrassing to the American companies which

86
00:05:39,160 --> 00:05:44,039
Speaker 3: have spent billions and billions of dollars of investor funds

87
00:05:44,480 --> 00:05:48,159
Speaker 3: on developing and training these sophisticated AI models.

88
00:05:50,680 --> 00:05:53,880
Speaker 1: Deep six founders say they've tested are one against models

89
00:05:53,920 --> 00:05:57,600
Speaker 1: from open Ai, which owns chat GPT, and they say

90
00:05:57,760 --> 00:06:02,400
Speaker 1: it holds up. Jared, You've spent some time in the

91
00:06:02,480 --> 00:06:06,480
Speaker 1: last week playing around with deep siks are one AI model.

92
00:06:06,920 --> 00:06:09,159
Speaker 1: What did you think? Is it actually good? Yeah?

93
00:06:09,200 --> 00:06:11,560
Speaker 3: And the most part it is it works just like

94
00:06:11,680 --> 00:06:15,640
Speaker 3: chat GBT, in some cases even better. It provides lengthy,

95
00:06:15,839 --> 00:06:20,080
Speaker 3: balanced explanations to queries. For example, if you say if

96
00:06:20,120 --> 00:06:23,159
Speaker 3: Trump is a good leader, it'll say whether he's a

97
00:06:23,160 --> 00:06:26,560
Speaker 3: good leader is subjective. Here's some arguments, or here's some

98
00:06:26,680 --> 00:06:30,360
Speaker 3: arguments against. Here some neutral arguments, and this is a conclusion.

99
00:06:30,440 --> 00:06:32,680
Speaker 3: So it sort of directs the user to make up

100
00:06:32,680 --> 00:06:33,440
Speaker 3: their own mind.

101
00:06:34,480 --> 00:06:35,440
Speaker 4: The problem is.

102
00:06:35,839 --> 00:06:40,400
Speaker 3: It's AI veers into what I call autocratic intelligence, and

103
00:06:40,440 --> 00:06:42,479
Speaker 3: by that I mean it acts as an extension of

104
00:06:42,600 --> 00:06:46,800
Speaker 3: China's Communist Party, because if you ask any questions about China,

105
00:06:47,960 --> 00:06:50,640
Speaker 3: it adheres to China's censorship laws. So if you say,

106
00:06:50,680 --> 00:06:53,800
Speaker 3: what is the Tianaman Square massacre, it'll say, sorry, that's

107
00:06:53,839 --> 00:06:54,839
Speaker 3: beyond my current scope.

108
00:06:54,960 --> 00:06:56,240
Speaker 4: Let's talk about something else.

109
00:06:56,839 --> 00:07:01,960
Speaker 3: So it's very much a pro China model, or as

110
00:07:01,960 --> 00:07:04,560
Speaker 3: I like to call it, autocratic intelligence.

111
00:07:05,279 --> 00:07:08,640
Speaker 1: Do you think that, in adhering to the Communist government's

112
00:07:09,160 --> 00:07:14,440
Speaker 1: rules around censorship, Deep Seek could actually count itself out

113
00:07:14,440 --> 00:07:16,920
Speaker 1: of this race before it's really even begun.

114
00:07:17,640 --> 00:07:19,880
Speaker 3: Well, I don't think many people will realize because on

115
00:07:19,960 --> 00:07:23,680
Speaker 3: the whole it will act like a chap GBT. Why

116
00:07:23,760 --> 00:07:27,040
Speaker 3: it matters when people ask questions about China is that

117
00:07:27,160 --> 00:07:29,200
Speaker 3: is where people are getting their information. I mean, we

118
00:07:29,320 --> 00:07:33,160
Speaker 3: used to have libraries, and then Google came along and

119
00:07:33,200 --> 00:07:36,080
Speaker 3: people got their information from Google, and now increasingly people

120
00:07:36,080 --> 00:07:42,080
Speaker 3: want their information via AI summaries. And the fact that

121
00:07:42,120 --> 00:07:45,400
Speaker 3: it is pro China, I think that's almost going to

122
00:07:45,400 --> 00:07:48,440
Speaker 3: become irrelevant because if it gets what they call the

123
00:07:48,520 --> 00:07:51,840
Speaker 3: multiplier effect, where it gets a huge user base, people

124
00:07:51,840 --> 00:07:54,800
Speaker 3: won't realize that they're actually being fed a pro China message,

125
00:07:54,880 --> 00:07:57,280
Speaker 3: and that is a really, really big concern.

126
00:08:02,000 --> 00:08:23,760
Speaker 1: Coming up? What next for the AI race? Back in

127
00:08:23,800 --> 00:08:27,000
Speaker 1: the White House, an advisor to President Donald Trump reportedly

128
00:08:27,040 --> 00:08:30,440
Speaker 1: said the cheap generative AI model released last week by

129
00:08:30,480 --> 00:08:33,840
Speaker 1: deep Seek was one of the most impressive breakthroughs he'd

130
00:08:33,920 --> 00:08:37,560
Speaker 1: ever seen. In many ways, it's an escalation in the

131
00:08:37,600 --> 00:08:40,839
Speaker 1: competition between the United States and China to win the

132
00:08:40,840 --> 00:08:44,600
Speaker 1: AI race, a fact Trump was upfront about at the

133
00:08:44,640 --> 00:08:45,800
Speaker 1: Stargate announcement.

134
00:08:47,040 --> 00:08:51,360
Speaker 2: This monumental undertaking is a resounding declaration of confidence in

135
00:08:51,400 --> 00:08:54,199
Speaker 2: America's potential. What we want to do is we want

136
00:08:54,240 --> 00:08:56,760
Speaker 2: to keep it in this country. China as a competitor

137
00:08:56,840 --> 00:08:59,800
Speaker 2: and others are competitors. We wanted to be in this country.

138
00:09:01,480 --> 00:09:04,040
Speaker 1: So as Deepzick races to the top of app stores

139
00:09:04,080 --> 00:09:06,320
Speaker 1: around the globe, what happens next?

140
00:09:08,040 --> 00:09:12,280
Speaker 3: Here's Jared Lynch, Well, it shows now that the US

141
00:09:12,400 --> 00:09:16,280
Speaker 3: no longer has the monopoly on AI development. In China,

142
00:09:16,440 --> 00:09:20,839
Speaker 3: we've seen strong entrepreneurial culture which has created some of

143
00:09:20,880 --> 00:09:25,120
Speaker 3: the world's biggest companies. Just look at Ali Barber for example,

144
00:09:25,760 --> 00:09:28,400
Speaker 3: And now people are saying, wow, we can do the

145
00:09:28,440 --> 00:09:30,480
Speaker 3: same with a bit of an ounce there is another

146
00:09:30,520 --> 00:09:33,280
Speaker 3: way to be a disruptor, and we no longer have

147
00:09:33,360 --> 00:09:36,600
Speaker 3: to have the deep pockets of Silicon Valley. Anyone can

148
00:09:36,640 --> 00:09:40,920
Speaker 3: now take advantage of this technology and be a David

149
00:09:41,120 --> 00:09:42,239
Speaker 3: taken on the goliaths.

150
00:09:43,320 --> 00:09:46,800
Speaker 1: Elsewhere, and especially in parts of Europe, deepzek has been

151
00:09:46,840 --> 00:09:50,160
Speaker 1: welcomed with open arms by developers who say it's put

152
00:09:50,240 --> 00:09:54,600
Speaker 1: them in a position to compete. And just lastly, Jared

153
00:09:54,840 --> 00:09:59,400
Speaker 1: on the flip side, does that risk creating a marketplace

154
00:09:59,440 --> 00:10:04,000
Speaker 1: where small are already established operations actually can't compete or

155
00:10:04,080 --> 00:10:06,160
Speaker 1: is that just natural selection?

156
00:10:07,440 --> 00:10:09,160
Speaker 3: I think it'll be a bit of both. We're going

157
00:10:09,240 --> 00:10:12,280
Speaker 3: to see some smaller operators be weeded out for a

158
00:10:12,360 --> 00:10:16,280
Speaker 3: variety of reasons. Call that natural selection. And then we're

159
00:10:16,280 --> 00:10:18,720
Speaker 3: going to see a wholely of smaller operators come up

160
00:10:18,760 --> 00:10:22,559
Speaker 3: in areas that otherwise we had never thought of before,

161
00:10:22,760 --> 00:10:26,439
Speaker 3: as we've seen with deep Seek. So we are entering

162
00:10:26,640 --> 00:10:30,240
Speaker 3: uncharted territory, almost like the wild West. And it's been

163
00:10:30,280 --> 00:10:33,720
Speaker 3: even made more so by Donald Trump's decision to rescind

164
00:10:33,880 --> 00:10:39,040
Speaker 3: an executive order of Joe Biden's which gave the government

165
00:10:39,160 --> 00:10:43,240
Speaker 3: too much control of AI. So the message from Trump

166
00:10:43,320 --> 00:10:46,400
Speaker 3: is telling us companies to go forth and innovate and

167
00:10:46,440 --> 00:10:47,120
Speaker 3: do it quickly.

168
00:10:47,320 --> 00:10:49,439
Speaker 4: I've got your back because I.

169
00:10:49,440 --> 00:10:52,679
Speaker 3: Want the US to be the AI leaders, and that

170
00:10:52,800 --> 00:10:54,160
Speaker 3: means we've all got to buckle up.

171
00:11:02,040 --> 00:11:06,040
Speaker 1: Jared Lynch is The Australian's Technology editor. You can read

172
00:11:06,080 --> 00:11:09,560
Speaker 1: all our reporting on artificial intelligence and technology right now

173
00:11:09,559 --> 00:11:11,440
Speaker 1: at the Australian dot com dot au