1
00:00:03,680 --> 00:00:09,200
Speaker 1: I'm t T and I'm Zakiyah and this is Dope Labs.

2
00:00:11,360 --> 00:00:14,600
Speaker 1: Welcome to Dope Labs, a weekly podcast that mixes hardcore

3
00:00:14,680 --> 00:00:17,599
Speaker 1: science with pop culture and a healthy dose of friendship.

4
00:00:22,800 --> 00:00:25,600
Speaker 1: We don't really show our faces a lot or at

5
00:00:25,640 --> 00:00:28,840
Speaker 1: all with the podcast. We don't do because I know

6
00:00:28,880 --> 00:00:31,920
Speaker 1: a lot of podcasts now are like, oh, you can

7
00:00:31,960 --> 00:00:35,360
Speaker 1: come and watch us, Yeah, you can come and watch

8
00:00:35,479 --> 00:00:38,000
Speaker 1: us on YouTube or whatever, and we don't do that

9
00:00:38,120 --> 00:00:40,000
Speaker 1: because we don't feel like we need.

10
00:00:40,000 --> 00:00:40,320
Speaker 2: To do that.

11
00:00:40,840 --> 00:00:42,879
Speaker 3: But what I have said this is one of the

12
00:00:42,880 --> 00:00:46,160
Speaker 3: things that I did do in preparation for this lab

13
00:00:47,120 --> 00:00:51,920
Speaker 3: was I asked chat GBT to make a photo of us,

14
00:00:52,000 --> 00:00:55,480
Speaker 3: and so I uploaded some reference photos and I just

15
00:00:55,520 --> 00:01:01,440
Speaker 3: sent you what they were, and I want your initial reactions.

16
00:01:03,960 --> 00:01:08,000
Speaker 1: That's the first draft, right because this is giving AI,

17
00:01:08,240 --> 00:01:11,440
Speaker 1: you know it is it is. And I'm confused because

18
00:01:11,480 --> 00:01:13,399
Speaker 1: I see other people that are like, I generated this

19
00:01:13,560 --> 00:01:17,080
Speaker 1: using AI and it doesn't look like it looks very realistic.

20
00:01:17,319 --> 00:01:17,959
Speaker 2: Uh huh.

21
00:01:18,000 --> 00:01:20,800
Speaker 1: There's this black woman, her name is Jessica E Boyd,

22
00:01:21,800 --> 00:01:24,000
Speaker 1: So she's on Instagram at Jessica E Boyd and she

23
00:01:24,080 --> 00:01:26,800
Speaker 1: does these incredible AI images. It's just a great job

24
00:01:27,280 --> 00:01:30,520
Speaker 1: and it looks like real afro texture hair, you know,

25
00:01:31,040 --> 00:01:33,920
Speaker 1: but this one you created no shade. But how many

26
00:01:33,920 --> 00:01:34,920
Speaker 1: fingers do I have?

27
00:01:37,240 --> 00:01:40,639
Speaker 3: It looks like you've either got fourteen or six.

28
00:01:41,600 --> 00:01:42,479
Speaker 2: I can't really tell.

29
00:01:42,520 --> 00:01:46,960
Speaker 3: All the fingers are melting together, and I'm like, this

30
00:01:47,040 --> 00:01:47,360
Speaker 3: is fun.

31
00:01:47,400 --> 00:01:50,000
Speaker 1: I see people that have been making themselves, like not barbies,

32
00:01:50,040 --> 00:01:52,120
Speaker 1: but like action figures in the plastic.

33
00:01:51,960 --> 00:01:53,880
Speaker 2: I've been seeing that. I've been seeing that.

34
00:01:54,040 --> 00:01:55,880
Speaker 1: I've seen people do themselves in animation.

35
00:01:56,280 --> 00:01:59,360
Speaker 2: Yeah, the studio ghibli ones. Those are very cute.

36
00:01:59,680 --> 00:02:02,960
Speaker 1: But I think like, no, don't get me wrong. I

37
00:02:02,960 --> 00:02:05,040
Speaker 1: love all the latest technology, and you know, I've been

38
00:02:05,040 --> 00:02:08,280
Speaker 1: out here trying everything okay, but sometimes I have to

39
00:02:08,280 --> 00:02:11,520
Speaker 1: stop and wonder, like who's doing all this stuff? Whoever's

40
00:02:11,560 --> 00:02:13,760
Speaker 1: doing it, or wherever's being done. It's just so far

41
00:02:13,800 --> 00:02:16,600
Speaker 1: away from us, and it makes me think, like, what

42
00:02:16,720 --> 00:02:19,359
Speaker 1: is the cost to generate these outputs?

43
00:02:19,840 --> 00:02:20,360
Speaker 2: Honestly?

44
00:02:20,480 --> 00:02:23,200
Speaker 3: And I think that's something that has been talked about

45
00:02:23,360 --> 00:02:26,160
Speaker 3: a lot on social media. Most recently, I've been seeing

46
00:02:26,160 --> 00:02:30,480
Speaker 3: people saying it takes sixty seven trillion gallons of water

47
00:02:30,680 --> 00:02:34,240
Speaker 3: to just if you say hello, chat GPT, And I'm like,

48
00:02:35,400 --> 00:02:36,280
Speaker 3: is that how that works.

49
00:02:36,360 --> 00:02:37,800
Speaker 2: I don't know if that's how that works.

50
00:02:38,760 --> 00:02:41,560
Speaker 1: I don't think that's how that works. But it's time

51
00:02:41,600 --> 00:02:42,400
Speaker 1: to pull back the curtain.

52
00:02:42,760 --> 00:02:44,480
Speaker 2: Let's jump into the recitation.

53
00:02:46,760 --> 00:02:49,840
Speaker 1: So today we're diving into AI. We're gonna talk a

54
00:02:49,880 --> 00:02:53,720
Speaker 1: little bit about automation, large language models like chat, GPT

55
00:02:54,400 --> 00:03:00,560
Speaker 1: and their impacts, specifically their environmental impacts. So with what

56
00:03:00,680 --> 00:03:01,519
Speaker 1: we know TT.

57
00:03:02,160 --> 00:03:04,680
Speaker 2: So we know what automation is.

58
00:03:04,880 --> 00:03:09,119
Speaker 3: That's the use of technology to perform tasks without human intervention.

59
00:03:09,360 --> 00:03:14,040
Speaker 3: So it's often rule based, so like if then if

60
00:03:14,040 --> 00:03:16,680
Speaker 3: I do this, then it'll do this or do this

61
00:03:16,880 --> 00:03:20,680
Speaker 3: until and it's also repetitive, so it's able to keep

62
00:03:20,760 --> 00:03:23,760
Speaker 3: going over and over and over again without need for

63
00:03:23,800 --> 00:03:24,399
Speaker 3: a human.

64
00:03:25,320 --> 00:03:27,680
Speaker 1: But I think a step above that or beyond that

65
00:03:27,880 --> 00:03:32,079
Speaker 1: is artificial intelligence, and sometimes we see people conflating those two.

66
00:03:32,280 --> 00:03:35,480
Speaker 1: Artificial intelligence is when you have systems that are designed

67
00:03:35,480 --> 00:03:40,080
Speaker 1: to mimic human intelligence and so unlike automation, which is

68
00:03:40,200 --> 00:03:43,000
Speaker 1: just a rule and it just every time it encounters

69
00:03:43,080 --> 00:03:44,760
Speaker 1: this one thing, it does whatever you told it to,

70
00:03:45,160 --> 00:03:49,640
Speaker 1: artificial intelligence is capable of learning and problem solving, and

71
00:03:49,680 --> 00:03:52,560
Speaker 1: so it'll say, okay, if you put this input, you

72
00:03:52,600 --> 00:03:54,400
Speaker 1: tend to like this, so the next time I see

73
00:03:54,400 --> 00:03:56,040
Speaker 1: that input, I'll bring it up. Even if that's not

74
00:03:56,080 --> 00:03:59,520
Speaker 1: a rule, it's a learning. It's able to expand its

75
00:03:59,560 --> 00:04:01,720
Speaker 1: knowledge based on what the inputs you give it.

76
00:04:02,000 --> 00:04:04,680
Speaker 3: Right, So like if you had a button factory and said,

77
00:04:04,720 --> 00:04:07,560
Speaker 3: every time this piece of plastic comes here, poke three

78
00:04:07,600 --> 00:04:10,440
Speaker 3: holes in it, and it does that, and it does

79
00:04:10,440 --> 00:04:12,720
Speaker 3: it over and over again, that's automation.

80
00:04:13,440 --> 00:04:16,240
Speaker 2: Now if you had a machine that was able.

81
00:04:16,000 --> 00:04:19,039
Speaker 3: To say, based on fashion trends, I'm going to make

82
00:04:19,080 --> 00:04:21,320
Speaker 3: this button out of wood instead of plastic, or I'm

83
00:04:21,320 --> 00:04:24,320
Speaker 3: gonna make this button bigger because people started liking bigger buttons.

84
00:04:24,360 --> 00:04:26,600
Speaker 2: Now that would be aim.

85
00:04:27,760 --> 00:04:31,280
Speaker 1: Then you have large language models, and that's advanced AIS

86
00:04:31,400 --> 00:04:35,400
Speaker 1: systems that are trained on vast amounts of data, particularly

87
00:04:35,440 --> 00:04:38,960
Speaker 1: text data, to understand and generate human like language. So

88
00:04:39,040 --> 00:04:41,240
Speaker 1: that's CHAT, GPT, Gemini, all of that.

89
00:04:41,640 --> 00:04:44,200
Speaker 3: Yeah, they're calling the Internet and gathering all of this

90
00:04:44,320 --> 00:04:48,240
Speaker 3: information and creating context around all of that information so

91
00:04:48,279 --> 00:04:52,200
Speaker 3: that they're able to learn. The next thing is proprietary models,

92
00:04:52,600 --> 00:04:55,640
Speaker 3: and what they do is they highlight that companies develop

93
00:04:56,120 --> 00:04:57,600
Speaker 3: specialized models.

94
00:04:57,839 --> 00:05:00,760
Speaker 2: So open AI they have.

95
00:05:00,720 --> 00:05:04,839
Speaker 3: The GPT variant, so we know open AI, most commonly

96
00:05:04,880 --> 00:05:08,240
Speaker 3: for chat GPT and chat GBT within it, that's just

97
00:05:08,279 --> 00:05:11,760
Speaker 3: the big umbrella. There's multiple types of chat GPT, so

98
00:05:12,000 --> 00:05:15,960
Speaker 3: they have four oh they have I think it's four

99
00:05:16,040 --> 00:05:19,840
Speaker 3: and four mini. They also have deep research now, so

100
00:05:19,920 --> 00:05:22,400
Speaker 3: it's tailored specific for a different task.

101
00:05:23,400 --> 00:05:25,280
Speaker 1: So now that we all have kind of the same

102
00:05:25,360 --> 00:05:28,400
Speaker 1: understanding of what these terms mean, let's talk about what

103
00:05:28,440 --> 00:05:29,160
Speaker 1: we want to know.

104
00:05:30,320 --> 00:05:34,679
Speaker 3: I want to know how the energy consumption of training

105
00:05:34,680 --> 00:05:38,400
Speaker 3: these AI models and running these AI models compared to

106
00:05:38,440 --> 00:05:41,599
Speaker 3: other technologies, like is it one to one or is

107
00:05:41,720 --> 00:05:43,120
Speaker 3: AI the big bad wolf?

108
00:05:44,560 --> 00:05:47,000
Speaker 1: And then like what are the implications? So like water

109
00:05:47,120 --> 00:05:49,919
Speaker 1: usage is one that I've seen people talking about on

110
00:05:50,040 --> 00:05:53,640
Speaker 1: social media, so about cooling data centers and all of

111
00:05:53,640 --> 00:05:54,920
Speaker 1: this stuff. I was like, I thought we were all

112
00:05:54,920 --> 00:05:56,599
Speaker 1: in the water cycle. So all the word is here

113
00:05:56,839 --> 00:05:59,960
Speaker 1: is here, but I don't know we did an expert

114
00:06:00,040 --> 00:06:02,960
Speaker 1: for that. I also want to know the mechanics. You know,

115
00:06:03,360 --> 00:06:06,080
Speaker 1: my background isn't engineering all of those things that it's

116
00:06:06,120 --> 00:06:07,279
Speaker 1: just interesting to me.

117
00:06:07,720 --> 00:06:11,239
Speaker 2: So what processes occur when you're.

118
00:06:11,160 --> 00:06:14,760
Speaker 3: Using a large language model like chat GBT or Gemini.

119
00:06:15,880 --> 00:06:17,839
Speaker 1: We also want a future cast and the same way

120
00:06:17,960 --> 00:06:22,200
Speaker 1: that we looked at what policies are around new technology.

121
00:06:22,400 --> 00:06:25,080
Speaker 1: We thought about this in our Space episode. You know,

122
00:06:25,240 --> 00:06:27,119
Speaker 1: we say, like, all right, now that all these people

123
00:06:27,120 --> 00:06:29,279
Speaker 1: are going to space, who's making the rules, who's said

124
00:06:29,279 --> 00:06:31,919
Speaker 1: in the speed limit? The same thing for AI? You know,

125
00:06:32,640 --> 00:06:35,680
Speaker 1: who's making the rules, who's saying how much is too much?

126
00:06:35,720 --> 00:06:36,400
Speaker 2: What do you use?

127
00:06:36,440 --> 00:06:38,400
Speaker 1: What's the costs, what's you know, there's a cost and

128
00:06:38,920 --> 00:06:41,160
Speaker 1: benefit for all these things, and so what is it

129
00:06:41,160 --> 00:06:43,640
Speaker 1: for AI? It's out of site, out of mind really.

130
00:06:43,680 --> 00:06:47,240
Speaker 3: Right, Well, I think that sets the stage perfectly for

131
00:06:47,360 --> 00:06:48,840
Speaker 3: us to jump into the dissection.

132
00:06:53,200 --> 00:06:56,600
Speaker 1: For today's dissection, we are talking to doctor Challet Wren.

133
00:06:57,120 --> 00:06:58,960
Speaker 1: Doctor Wren, we wanted to start by kind of setting

134
00:06:58,960 --> 00:07:01,719
Speaker 1: the stage about them and some measurement. You know, I

135
00:07:01,760 --> 00:07:04,200
Speaker 1: think for all tech and innovation, there's a cost and

136
00:07:04,240 --> 00:07:07,719
Speaker 1: a benefit, and often those costs are immutely presented in

137
00:07:08,360 --> 00:07:13,240
Speaker 1: price and things you buy to maintain the technology. Sometimes

138
00:07:13,320 --> 00:07:16,440
Speaker 1: it's subscriptions for software, Like I remember when I had

139
00:07:16,440 --> 00:07:19,720
Speaker 1: a game Boy, it was the batteries. Okay, yeah, burn

140
00:07:19,880 --> 00:07:24,240
Speaker 1: was like burning through batteries. For cars, it's gas, it's maintenance,

141
00:07:24,240 --> 00:07:27,600
Speaker 1: it's inspections for emissions. But when we're physically removed from

142
00:07:27,640 --> 00:07:29,840
Speaker 1: the technology, it can be harder to have a good

143
00:07:29,920 --> 00:07:33,120
Speaker 1: grasp of like what it costs to run these things.

144
00:07:33,480 --> 00:07:37,720
Speaker 1: How do we understand and appreciate costs now.

145
00:07:37,520 --> 00:07:41,920
Speaker 4: So this is a huge energy consumption, and I mean,

146
00:07:41,960 --> 00:07:44,040
Speaker 4: by the way, using energy itself, I think is a

147
00:07:44,080 --> 00:07:47,440
Speaker 4: great thing because it just shows that we have more productivity.

148
00:07:48,360 --> 00:07:50,800
Speaker 4: You know, usually energy consumption is a indicator of the

149
00:07:50,840 --> 00:07:55,320
Speaker 4: economic activity, but the consequence of the using energy that's

150
00:07:55,480 --> 00:07:58,440
Speaker 4: not always good. For example, the energy produces heat and

151
00:07:58,480 --> 00:08:02,000
Speaker 4: we need water to take away the heat, and that

152
00:08:02,280 --> 00:08:06,080
Speaker 4: consumes our natural resources. And also when we generate how

153
00:08:06,120 --> 00:08:09,360
Speaker 4: we generate the electricity, how we generate the energy. Usually,

154
00:08:09,480 --> 00:08:13,640
Speaker 4: even in the United States, a large fraction of the

155
00:08:13,760 --> 00:08:16,200
Speaker 4: energy is coming from fossil fuels, and that comes with

156
00:08:16,360 --> 00:08:19,880
Speaker 4: the carbon emission problem. Right. So this, you know, it

157
00:08:19,920 --> 00:08:23,480
Speaker 4: has this long term impact on the climate change as

158
00:08:23,560 --> 00:08:25,040
Speaker 4: many of the research that they have shown.

159
00:08:25,200 --> 00:08:27,680
Speaker 1: I think that's a great point. What doctor Renna is

160
00:08:27,680 --> 00:08:31,000
Speaker 1: saying is it costs even before we get to water,

161
00:08:31,080 --> 00:08:33,680
Speaker 1: there's just energy. This is something I'm always talking about

162
00:08:33,679 --> 00:08:36,520
Speaker 1: the people who drive electric cars. I'm like, Okay, you're

163
00:08:36,559 --> 00:08:38,280
Speaker 1: not using gas, but what do you think how you

164
00:08:38,320 --> 00:08:40,679
Speaker 1: think we get the energy that's coming to you.

165
00:08:40,640 --> 00:08:45,840
Speaker 3: Right, That is one of the laws of thermodynamics. Energy

166
00:08:45,880 --> 00:08:50,600
Speaker 3: cannot be created or destroyed. So we're not creating this

167
00:08:50,840 --> 00:08:53,440
Speaker 3: energy out of nowhere and you can't get rid of it.

168
00:08:53,440 --> 00:08:54,680
Speaker 2: It all goes somewhere.

169
00:08:54,720 --> 00:08:57,840
Speaker 3: It's a closed loop like energy is flowing from one

170
00:08:57,880 --> 00:08:58,680
Speaker 3: thing to another.

171
00:08:59,320 --> 00:09:01,600
Speaker 1: And I think the other part that he mentioned that

172
00:09:01,640 --> 00:09:04,280
Speaker 1: I didn't really think about for these data centers is

173
00:09:05,640 --> 00:09:08,840
Speaker 1: the air quality. So PM two point five that means

174
00:09:08,920 --> 00:09:12,840
Speaker 1: particular matter that is two point five micrometers or smaller

175
00:09:13,320 --> 00:09:17,679
Speaker 1: in diameter. Now, biologically, when you're that small, you can

176
00:09:18,000 --> 00:09:22,880
Speaker 1: get very deep into the lungs, okay, and so whatever

177
00:09:22,920 --> 00:09:24,920
Speaker 1: you have can get into the lungs and it can

178
00:09:24,960 --> 00:09:30,000
Speaker 1: potentially enter the bloodstream. Not always, but potential is enough

179
00:09:30,000 --> 00:09:33,760
Speaker 1: for me to be like humhm, right, that's enough for me.

180
00:09:34,120 --> 00:09:35,679
Speaker 2: That sounds scary.

181
00:09:35,800 --> 00:09:39,280
Speaker 4: So the air pllutants includes particular matter two point five

182
00:09:39,440 --> 00:09:46,720
Speaker 4: PM two point five and also nitrogen dioxide soultur dioxide.

183
00:09:46,920 --> 00:09:49,840
Speaker 4: So these type of air pollutants can travel hundreds of miles.

184
00:09:50,280 --> 00:09:52,120
Speaker 4: So even though we don't leave next to a data

185
00:09:52,120 --> 00:09:54,800
Speaker 4: center or next to some power plants, we are the

186
00:09:54,960 --> 00:09:57,079
Speaker 4: air pollutants that we breathe in could be coming from

187
00:09:57,120 --> 00:10:01,240
Speaker 4: those places, and so this air just have a direct

188
00:10:01,280 --> 00:10:05,280
Speaker 4: toll on people's health. They could create asthma, heart attack,

189
00:10:06,559 --> 00:10:11,000
Speaker 4: lung cancer, and also even premature deaths. And there's been many,

190
00:10:11,000 --> 00:10:16,480
Speaker 4: many research studies showing that causal relationship between these air

191
00:10:16,480 --> 00:10:17,880
Speaker 4: pollutants and people health.

192
00:10:18,480 --> 00:10:22,360
Speaker 3: But what about the water that's associated with running these

193
00:10:22,400 --> 00:10:23,320
Speaker 3: AI models?

194
00:10:23,640 --> 00:10:25,920
Speaker 2: Is it really as bad as folks are saying.

195
00:10:26,440 --> 00:10:30,120
Speaker 4: Our study find based on the very limited data in

196
00:10:30,120 --> 00:10:33,120
Speaker 4: the public domain, we show that if you have a

197
00:10:33,160 --> 00:10:38,280
Speaker 4: conversation with a large language model, you likely consume about

198
00:10:39,400 --> 00:10:44,720
Speaker 4: five hundred million water for ten to fifty queries, depending

199
00:10:44,720 --> 00:10:48,440
Speaker 4: on where have you random model, And that's for the

200
00:10:48,520 --> 00:10:50,880
Speaker 4: inference and also when you train the model, it's also

201
00:10:50,920 --> 00:10:52,680
Speaker 4: consumed quite a bit of water.

202
00:11:07,280 --> 00:11:10,679
Speaker 1: Okay, so we're talking to doctor Charlet Wrenn about AI

203
00:11:11,040 --> 00:11:15,839
Speaker 1: and water usage, and doctor Rand said, basically, with ten

204
00:11:15,880 --> 00:11:18,640
Speaker 1: to fifty queries, you use about five hundred millileaters of water,

205
00:11:18,679 --> 00:11:21,680
Speaker 1: which is sixteen point nine ounces or a Deer Park

206
00:11:21,760 --> 00:11:24,600
Speaker 1: water bottle. Now, I think there are two things to

207
00:11:24,679 --> 00:11:28,000
Speaker 1: unpack their tt the water usage and the note that

208
00:11:28,040 --> 00:11:30,559
Speaker 1: he said about it depending on where you are when

209
00:11:30,600 --> 00:11:32,960
Speaker 1: you run the model. Let's start with the water usage.

210
00:11:33,040 --> 00:11:36,520
Speaker 3: Okay, so before we get too deep into the water consumption,

211
00:11:36,600 --> 00:11:38,560
Speaker 3: I really want to take a step back and like

212
00:11:38,840 --> 00:11:42,000
Speaker 3: set the stage for everybody about what actually happens when

213
00:11:42,040 --> 00:11:45,200
Speaker 3: you type something into chat GPT. When you put a

214
00:11:45,280 --> 00:11:48,560
Speaker 3: prompt into CHATGPT, it travels through the Internet to a

215
00:11:48,679 --> 00:11:51,800
Speaker 3: data center where that model. Remember we talked about those

216
00:11:51,800 --> 00:11:56,440
Speaker 3: different specific models associated with different organizations, So open ai

217
00:11:56,559 --> 00:11:59,520
Speaker 3: has the GPT models. So where that model is hosted,

218
00:11:59,559 --> 00:12:02,559
Speaker 3: it goes to that data center, and then the servers

219
00:12:02,640 --> 00:12:05,839
Speaker 3: process your input by running it through that model, which

220
00:12:05,880 --> 00:12:09,520
Speaker 3: consists of layers of neural networks, and then the servers

221
00:12:09,679 --> 00:12:12,360
Speaker 3: generate a response and send it back to your device,

222
00:12:12,440 --> 00:12:15,040
Speaker 3: so your phone or your laptop or however you're using

223
00:12:15,080 --> 00:12:17,199
Speaker 3: that AI model.

224
00:12:17,640 --> 00:12:20,240
Speaker 2: And so it's a very fast process.

225
00:12:20,280 --> 00:12:22,520
Speaker 3: If you've ever used any of these AI tools, you

226
00:12:22,600 --> 00:12:26,360
Speaker 3: know it happens in seconds. So when we're talking about

227
00:12:26,480 --> 00:12:31,040
Speaker 3: water consumption, I think knowing where that water comes in

228
00:12:31,120 --> 00:12:33,080
Speaker 3: because it's like, what are we talking about here? Why

229
00:12:33,200 --> 00:12:36,840
Speaker 3: is water associated with the Internet. Data centers use a

230
00:12:36,880 --> 00:12:41,320
Speaker 3: significant amount of water to cool those servers down because

231
00:12:41,320 --> 00:12:42,920
Speaker 3: they generate a lot of heat.

232
00:12:43,080 --> 00:12:43,960
Speaker 2: We all know that.

233
00:12:43,880 --> 00:12:46,280
Speaker 3: This technology gets hot when we're on our laptops for

234
00:12:46,320 --> 00:12:48,199
Speaker 3: a long time, on a computer for a long time,

235
00:12:48,480 --> 00:12:50,600
Speaker 3: when we're on our phones for a long time, they

236
00:12:50,640 --> 00:12:53,400
Speaker 3: start to feel a little bit warm, right, And that's

237
00:12:53,440 --> 00:12:57,040
Speaker 3: because you know, all of the technology inside is working

238
00:12:57,080 --> 00:12:59,839
Speaker 3: and it's generating heat to do that work. And so

239
00:13:00,120 --> 00:13:02,560
Speaker 3: what they do is they're running this cooling system which

240
00:13:02,679 --> 00:13:05,880
Speaker 3: uses water to cool down these servers.

241
00:13:06,360 --> 00:13:10,360
Speaker 4: When we cool down the data center facility, it involves

242
00:13:10,400 --> 00:13:13,199
Speaker 4: two typically involves two stages. The first stage is moving

243
00:13:13,240 --> 00:13:16,320
Speaker 4: the heat from the servers to a heat exchanger or

244
00:13:16,360 --> 00:13:20,400
Speaker 4: to the facility. This process either uses air or uses

245
00:13:20,440 --> 00:13:24,000
Speaker 4: a closed loop with some special liquids so there's no

246
00:13:24,080 --> 00:13:26,360
Speaker 4: water loss, and there shouldn't be any water loss unless

247
00:13:26,400 --> 00:13:29,720
Speaker 4: there's a leak, okay. And the second stage is moving

248
00:13:29,720 --> 00:13:32,880
Speaker 4: the heat from the heat exchanger to the outside environment,

249
00:13:33,320 --> 00:13:37,439
Speaker 4: to the atmosphere, to the sky and this process, of course,

250
00:13:37,480 --> 00:13:41,360
Speaker 4: there are different options for moving the heat, but one

251
00:13:41,600 --> 00:13:46,960
Speaker 4: Cammer approach is using water evaporation. And water evaporation just

252
00:13:47,559 --> 00:13:52,600
Speaker 4: is the water that is temporarily lost into the sky,

253
00:13:53,280 --> 00:13:57,280
Speaker 4: and that's what we call water consumption. So of course

254
00:13:57,440 --> 00:13:59,600
Speaker 4: the water is still within our planet. It doesn't go

255
00:13:59,679 --> 00:14:06,800
Speaker 4: to the sound whatever somewhere. Yeah, yeah, within our globe.

256
00:14:06,800 --> 00:14:09,320
Speaker 4: But it's just you know, when the water comes back

257
00:14:09,400 --> 00:14:13,560
Speaker 4: and where it'll come back is highly uncertain, especially in

258
00:14:13,640 --> 00:14:18,800
Speaker 4: those route regions, and also even when the water comes back,

259
00:14:18,840 --> 00:14:21,880
Speaker 4: it may not be in the in the freshwater state.

260
00:14:22,240 --> 00:14:24,040
Speaker 1: But I think in the same way that they're using

261
00:14:24,080 --> 00:14:25,760
Speaker 1: water to cool down a service. Isn't that the same

262
00:14:25,800 --> 00:14:29,960
Speaker 1: way like basically air conditioning works and cooler and stuff

263
00:14:30,000 --> 00:14:30,440
Speaker 1: in there.

264
00:14:31,280 --> 00:14:34,560
Speaker 3: And the other thing is like the Internet been using servers,

265
00:14:34,600 --> 00:14:38,720
Speaker 3: like AI using servers. That's not new. That's not new technology.

266
00:14:38,760 --> 00:14:43,080
Speaker 3: We've always had servers. AI is just advancing and using

267
00:14:43,120 --> 00:14:46,920
Speaker 3: those exame, exact servers that we're already being used for

268
00:14:47,040 --> 00:14:49,000
Speaker 3: the Internet, like the same types of servers.

269
00:14:49,000 --> 00:14:50,440
Speaker 2: They just have different models in it.

270
00:14:50,760 --> 00:14:54,760
Speaker 3: But when you type in www, dot Dope Labs, podcast

271
00:14:54,800 --> 00:14:58,000
Speaker 3: dot com, that information has to go to a server

272
00:14:58,440 --> 00:14:59,040
Speaker 3: in order.

273
00:14:58,840 --> 00:14:59,920
Speaker 1: To type it in right now.

274
00:15:00,120 --> 00:15:04,960
Speaker 2: He tied it right now and click on stuff. Oh

275
00:15:05,080 --> 00:15:09,120
Speaker 2: my goodness. Yeah, it's it's they're all doing the same things.

276
00:15:09,880 --> 00:15:12,240
Speaker 1: I think one of the things that was really interesting

277
00:15:12,280 --> 00:15:15,840
Speaker 1: to me that Charlett explained is that sometimes people are

278
00:15:15,840 --> 00:15:19,560
Speaker 1: comparing apples and oranges because that water that's those servers

279
00:15:19,600 --> 00:15:21,240
Speaker 1: are using is being recycled.

280
00:15:21,240 --> 00:15:22,359
Speaker 2: It's being used.

281
00:15:22,160 --> 00:15:24,720
Speaker 1: Over and over again. And they were comparing it to

282
00:15:24,760 --> 00:15:27,240
Speaker 1: like eating a hamburger, but they were looking at the

283
00:15:27,280 --> 00:15:31,240
Speaker 1: water to cow, the water to water the grass.

284
00:15:31,280 --> 00:15:34,760
Speaker 4: Like a few weeks ago, the CEO of a leading

285
00:15:34,880 --> 00:15:38,080
Speaker 4: AI company said, you know the study by the study

286
00:15:38,120 --> 00:15:40,200
Speaker 4: on as water US. It was something made up by

287
00:15:40,240 --> 00:15:45,200
Speaker 4: those anti AI crowd. But this, I would say, this

288
00:15:45,240 --> 00:15:50,120
Speaker 4: is really misleading comparison because the Hamburg's water days quoted

289
00:15:50,320 --> 00:15:54,880
Speaker 4: includes the rainwater in the in the soils to grow

290
00:15:54,920 --> 00:15:56,960
Speaker 4: the grasses, to feed the cattle, to make the paddy.

291
00:15:57,040 --> 00:16:00,000
Speaker 4: So that's really live cycle water using and it's mostly

292
00:16:00,160 --> 00:16:03,440
Speaker 4: water which we call green water in the technical community.

293
00:16:04,520 --> 00:16:07,720
Speaker 4: But the air water consumption is only for the operational

294
00:16:07,720 --> 00:16:10,160
Speaker 4: stage when we actually random all, we're not talking about

295
00:16:10,160 --> 00:16:14,000
Speaker 4: the water used for mining the rare earth or making

296
00:16:14,040 --> 00:16:17,800
Speaker 4: the AI chapes or recycling the summer. So it's not

297
00:16:17,960 --> 00:16:21,640
Speaker 4: life cycle for water for AI, and so it's a

298
00:16:21,640 --> 00:16:25,120
Speaker 4: different scope of water comparison, different type of water. And

299
00:16:25,160 --> 00:16:28,920
Speaker 4: the air water is mostly blue water, that's the water

300
00:16:28,960 --> 00:16:32,560
Speaker 4: in the rivers, lakes, and groundwater sources that human can

301
00:16:32,600 --> 00:16:36,440
Speaker 4: directly use. So I think those type of comparison are

302
00:16:36,560 --> 00:16:41,320
Speaker 4: just showing that they either do not have the necessary

303
00:16:41,360 --> 00:16:45,320
Speaker 4: knowledge in this space or they're having some other agendas.

304
00:16:46,520 --> 00:16:51,400
Speaker 4: So yeah, those are not really scientifically correct comparison.

305
00:16:52,600 --> 00:16:54,200
Speaker 2: Oh okay, I think I understand.

306
00:16:54,240 --> 00:16:57,920
Speaker 3: So the water consumption is the water that doesn't make

307
00:16:57,960 --> 00:17:01,880
Speaker 3: it into the drain and into our sewage. So that

308
00:17:01,880 --> 00:17:04,880
Speaker 3: would be the water that in the washing your hands example,

309
00:17:05,119 --> 00:17:08,199
Speaker 3: is being absorbed into your skin a little bit and

310
00:17:08,240 --> 00:17:11,600
Speaker 3: then also absorbed in like a paper towel or a

311
00:17:11,640 --> 00:17:14,400
Speaker 3: dish towel, or however you wipe your hands off after

312
00:17:14,440 --> 00:17:16,200
Speaker 3: you wash your hands. I hope y'alla washing our hands.

313
00:17:17,640 --> 00:17:23,119
Speaker 4: Yeah, So technically that's called water withdrawal, and the water

314
00:17:23,160 --> 00:17:25,640
Speaker 4: you put it that goes into the sewage is called

315
00:17:25,720 --> 00:17:29,240
Speaker 4: water discharge. The difference in water wizdow and water discharge

316
00:17:29,240 --> 00:17:33,359
Speaker 4: for washing our hands is very minimum, so okay, the

317
00:17:33,359 --> 00:17:34,960
Speaker 4: difference is called water consumption.

318
00:17:35,920 --> 00:17:43,520
Speaker 3: So Challet did a study about how AI and human

319
00:17:43,520 --> 00:17:48,480
Speaker 3: work compared to each other, about the efficiency and about

320
00:17:48,520 --> 00:17:52,560
Speaker 3: the energy required to perform a task, and what they

321
00:17:52,600 --> 00:17:55,520
Speaker 3: found was is that of course, you know, in a

322
00:17:55,520 --> 00:17:58,560
Speaker 3: lot of situations, AI is more efficient. It can work faster,

323
00:17:58,640 --> 00:18:01,639
Speaker 3: it can work longer. Know where humans require you know,

324
00:18:01,720 --> 00:18:06,439
Speaker 3: lunch breaks and sleep. AI doesn't it's constantly running. But

325
00:18:06,520 --> 00:18:10,560
Speaker 3: the energy consumption of AI is a lot higher, you know.

326
00:18:10,680 --> 00:18:13,680
Speaker 3: So there's like pluses and minuses to both our studies.

327
00:18:13,720 --> 00:18:17,160
Speaker 4: Not trying to say, let's AI is bad and that's

328
00:18:17,200 --> 00:18:20,560
Speaker 4: not our purpose. Its AI is great. It's why why

329
00:18:20,560 --> 00:18:23,919
Speaker 4: it's bad. It's it's using some resources. It's just like

330
00:18:24,040 --> 00:18:27,600
Speaker 4: anything else, right, So we need to have an understanding

331
00:18:27,640 --> 00:18:30,119
Speaker 4: of the cost. And turns out the water cost is

332
00:18:30,160 --> 00:18:31,919
Speaker 4: just one part of the cost, and where there are

333
00:18:31,960 --> 00:18:36,840
Speaker 4: also other costs like the climate, the energy, the public health.

334
00:18:37,200 --> 00:18:40,359
Speaker 4: So we need to look at this, evalu the cost

335
00:18:40,440 --> 00:18:43,280
Speaker 4: more holistically and think about what is the best way

336
00:18:43,320 --> 00:18:49,680
Speaker 4: to support the technology. That's our purpose. Our research goal

337
00:18:49,920 --> 00:18:53,840
Speaker 4: is not like hey, please stop using it now, No,

338
00:18:54,720 --> 00:18:59,359
Speaker 4: we shure to use AI for good purposes and you

339
00:18:59,359 --> 00:19:00,640
Speaker 4: know more response way.

340
00:19:01,119 --> 00:19:03,359
Speaker 3: Can you talk about that a little bit more the

341
00:19:04,160 --> 00:19:07,679
Speaker 3: cost to the environment, public health, so that people can

342
00:19:07,800 --> 00:19:09,080
Speaker 3: understand what you mean by that.

343
00:19:10,280 --> 00:19:14,480
Speaker 4: So because this is a huge energy consumption, and I mean,

344
00:19:14,520 --> 00:19:17,480
Speaker 4: by the way, using energy itself is not I think

345
00:19:17,600 --> 00:19:19,960
Speaker 4: is a great thing because we want I mean, it

346
00:19:20,080 --> 00:19:23,600
Speaker 4: just shows that we have more productivity. You know, usually

347
00:19:23,680 --> 00:19:27,080
Speaker 4: energy consumption is the indicator of the economic activity. So,

348
00:19:27,119 --> 00:19:31,320
Speaker 4: but the consequence of using energy that's probably not always good.

349
00:19:31,359 --> 00:19:34,240
Speaker 4: For example, the energy produces heat, and we need water

350
00:19:34,400 --> 00:19:39,800
Speaker 4: to take away the heat, and that consumes our natural resources.

351
00:19:39,960 --> 00:19:42,960
Speaker 4: And also when we generate how we generate the electricity,

352
00:19:43,080 --> 00:19:46,560
Speaker 4: how we generate energy. Usually even in the United States,

353
00:19:48,200 --> 00:19:50,879
Speaker 4: a large fraction of the energy is coming from fossil fuels,

354
00:19:50,920 --> 00:19:54,520
Speaker 4: and that comes with the carbon emission problem. Right, So this, uh,

355
00:19:55,080 --> 00:19:58,160
Speaker 4: you know, it has this long term impact on the

356
00:19:58,200 --> 00:20:15,080
Speaker 4: climate change. As many of the research I've shown, I feel.

357
00:20:14,800 --> 00:20:17,119
Speaker 1: Like so much of the conversation has been framed as

358
00:20:17,119 --> 00:20:20,280
Speaker 1: an all or nothing is don't have any costs or

359
00:20:20,440 --> 00:20:22,920
Speaker 1: use AI, And I'm like, what people don't think about

360
00:20:23,000 --> 00:20:27,479
Speaker 1: is how many people are using Zoom, Google Meet, Skype

361
00:20:27,520 --> 00:20:32,400
Speaker 1: if that's still operational. I don't know Microsoft teams and

362
00:20:32,880 --> 00:20:36,720
Speaker 1: you know, when the pandemic hit, Zoom saw elite from

363
00:20:36,840 --> 00:20:40,479
Speaker 1: ten million daily users to over three hundred million daily users.

364
00:20:41,040 --> 00:20:43,640
Speaker 1: And there was a study from Purdue University. A group

365
00:20:43,680 --> 00:20:46,640
Speaker 1: there said that an hour on zoom generates one hundred

366
00:20:46,640 --> 00:20:49,560
Speaker 1: and fifty two one thousand grams of CO two. Nobody

367
00:20:49,640 --> 00:20:51,959
Speaker 1: was thinking about that when we were trying to all

368
00:20:52,000 --> 00:20:55,119
Speaker 1: stay connected and stay saying now turn off your camera

369
00:20:55,160 --> 00:20:57,560
Speaker 1: can help with that is whether this study is in.

370
00:20:58,240 --> 00:21:02,119
Speaker 1: But I think us considering, you know, there's a cost

371
00:21:02,160 --> 00:21:04,600
Speaker 1: to all of these things that we often don't think about,

372
00:21:04,880 --> 00:21:08,960
Speaker 1: and so I think it just becomes important to consider, like, hey,

373
00:21:09,080 --> 00:21:10,880
Speaker 1: everything here has a cost.

374
00:21:11,040 --> 00:21:14,680
Speaker 3: Right, and the cost varies depending on where you are.

375
00:21:14,800 --> 00:21:18,600
Speaker 3: So the amount of water required for the AI chatbots

376
00:21:19,080 --> 00:21:22,800
Speaker 3: using chat GBT four to generate a one hundred word

377
00:21:22,840 --> 00:21:28,000
Speaker 3: email varies by location. The amount goes up depending on

378
00:21:28,040 --> 00:21:31,800
Speaker 3: where you are, Like in Washington State, it'll require almost

379
00:21:31,800 --> 00:21:35,320
Speaker 3: fifteen hundred milli liters of water, while in Texas it's

380
00:21:35,480 --> 00:21:38,760
Speaker 3: only two hundred and thirty five. So we can't just say, oh,

381
00:21:39,000 --> 00:21:43,920
Speaker 3: all of it requires all this water, and that's not

382
00:21:43,960 --> 00:21:45,119
Speaker 3: necessarily true.

383
00:21:45,600 --> 00:21:46,440
Speaker 2: Is it using water?

384
00:21:46,600 --> 00:21:49,199
Speaker 3: Yes, in a closed loop, because the water we have

385
00:21:49,320 --> 00:21:53,280
Speaker 3: is the water we've got, But also other forms of

386
00:21:53,320 --> 00:21:58,800
Speaker 3: technology are also using water as well. It takes six

387
00:21:59,240 --> 00:22:02,240
Speaker 3: nine hundred and nine liters of water to produce one

388
00:22:02,320 --> 00:22:07,240
Speaker 3: pound of beef one pound of beef six ninety two liters.

389
00:22:07,720 --> 00:22:11,480
Speaker 3: Training a chat GBT three has the same water cost

390
00:22:11,520 --> 00:22:13,560
Speaker 3: as producing one hundred pounds of beef.

391
00:22:15,040 --> 00:22:18,720
Speaker 1: So if one hundred people do meatless mondays yea, then.

392
00:22:19,280 --> 00:22:21,960
Speaker 3: It's nearly double the amount and average American eats in

393
00:22:22,000 --> 00:22:24,679
Speaker 3: a year. So the amount of water you're using to

394
00:22:24,840 --> 00:22:28,160
Speaker 3: eat beef is way more than you're using to run

395
00:22:28,200 --> 00:22:32,240
Speaker 3: a CHATGBT model. So in another comparison, is it takes

396
00:22:32,400 --> 00:22:35,280
Speaker 3: nine and fifty six liters of water to produce one

397
00:22:35,320 --> 00:22:39,120
Speaker 3: pound of rice because rice has grown completely submerged underwater.

398
00:22:39,240 --> 00:22:42,080
Speaker 2: Requires a lot of water to grow rice.

399
00:22:42,480 --> 00:22:44,760
Speaker 3: But we haven't been hearing nothing about that, or maybe

400
00:22:44,760 --> 00:22:46,920
Speaker 3: people been talking about it, and I just don't know that.

401
00:22:47,040 --> 00:22:47,640
Speaker 1: I know that I'm.

402
00:22:49,640 --> 00:22:52,520
Speaker 3: People been talking about AI and how all this water

403
00:22:52,600 --> 00:22:54,320
Speaker 3: and all these things like that, and I'm just like, man,

404
00:22:54,760 --> 00:22:58,160
Speaker 3: we really got to think about everything else that requires water.

405
00:22:58,560 --> 00:23:00,439
Speaker 3: I remember we talked about it when I said I

406
00:23:00,480 --> 00:23:03,440
Speaker 3: was drinking almond milk and they were like, you know,

407
00:23:04,359 --> 00:23:07,440
Speaker 3: text that growing all men, And I was like, damn.

408
00:23:07,280 --> 00:23:10,000
Speaker 1: You know how much water it takes up. If I

409
00:23:10,040 --> 00:23:12,639
Speaker 1: have a reaction to where I'm lactose.

410
00:23:12,280 --> 00:23:13,320
Speaker 2: Intolerant I know.

411
00:23:16,520 --> 00:23:20,119
Speaker 1: People, you know, I just think I really, when it

412
00:23:20,160 --> 00:23:23,840
Speaker 1: comes down to it, I think we are just we're

413
00:23:24,119 --> 00:23:26,639
Speaker 1: bringing in so much information, but we're not grounding it

414
00:23:26,720 --> 00:23:32,280
Speaker 1: at anything. No, really, I blame the media. Uh and

415
00:23:32,359 --> 00:23:36,200
Speaker 1: not our media, but like Hollywood, because everything you see

416
00:23:36,240 --> 00:23:39,200
Speaker 1: represented about AI, nothing is really talking about this kind

417
00:23:39,200 --> 00:23:42,000
Speaker 1: of stuff, or you would never see. I don't think

418
00:23:42,000 --> 00:23:45,159
Speaker 1: there's a lot of grounding what the cost is to

419
00:23:45,240 --> 00:23:47,760
Speaker 1: the benefit and convenience of things, and so when people

420
00:23:47,760 --> 00:23:51,800
Speaker 1: start talking costs, it's like, oh my goodness. I'm sure

421
00:23:51,800 --> 00:23:53,679
Speaker 1: people thought the same thing when the radio hit the

422
00:23:53,720 --> 00:23:57,160
Speaker 1: market or when the calculator started being used. I mean,

423
00:23:57,200 --> 00:23:59,320
Speaker 1: I feel like AI, the way people talk about AI

424
00:23:59,480 --> 00:24:02,679
Speaker 1: in the education system, they always just put it in

425
00:24:02,720 --> 00:24:05,240
Speaker 1: this thing where it's like they're cheating, They're cheating, they're cheating.

426
00:24:05,240 --> 00:24:07,240
Speaker 2: I'm like, first of all, that's the thing in the past.

427
00:24:07,520 --> 00:24:10,200
Speaker 3: That was something that was an issue with early AI,

428
00:24:10,359 --> 00:24:11,480
Speaker 3: early chat GBT.

429
00:24:11,840 --> 00:24:13,600
Speaker 2: That's no longer an issue anymore.

430
00:24:13,800 --> 00:24:15,920
Speaker 1: I think the main thing I can tell people I

431
00:24:15,960 --> 00:24:19,800
Speaker 1: would want our listeners to consider is to consider all

432
00:24:19,840 --> 00:24:24,719
Speaker 1: of this in the grand scope of the natural resources

433
00:24:24,720 --> 00:24:26,880
Speaker 1: that we have available. Think about it when you turn

434
00:24:26,880 --> 00:24:31,200
Speaker 1: it on your ring light when you are live streaming

435
00:24:31,280 --> 00:24:35,760
Speaker 1: but not even looking at your device, when you are.

436
00:24:35,760 --> 00:24:38,720
Speaker 3: Like scrolling on TikTok for hours and hours and hours

437
00:24:38,720 --> 00:24:41,120
Speaker 3: and falling asleep with the TikTok running and the TikTok

438
00:24:41,400 --> 00:24:42,760
Speaker 3: keeps looping and looping.

439
00:24:43,359 --> 00:24:45,879
Speaker 1: Yes, I think there is a cost to all of

440
00:24:45,920 --> 00:24:50,640
Speaker 1: the things that are entertaining, convenient and brought to us

441
00:24:50,720 --> 00:24:53,639
Speaker 1: through our devices, and I think we should be considering

442
00:24:53,680 --> 00:24:56,400
Speaker 1: that for everything. I think one of the key things

443
00:24:57,000 --> 00:25:00,600
Speaker 1: that I'm interested in is how public policy begins to shape,

444
00:25:00,640 --> 00:25:02,480
Speaker 1: Like are we going to have Now this may be

445
00:25:02,520 --> 00:25:04,960
Speaker 1: a little dystopian, I'm like, are we gonna have AI tokens?

446
00:25:05,000 --> 00:25:09,480
Speaker 1: You know, not only certain can you use it? But

447
00:25:10,000 --> 00:25:10,720
Speaker 1: who knows?

448
00:25:11,040 --> 00:25:13,720
Speaker 3: Yeah, I mean I think that that is what's up next.

449
00:25:13,840 --> 00:25:17,479
Speaker 3: Agentic ai is the next big thing where you know,

450
00:25:17,600 --> 00:25:21,000
Speaker 3: you can make your own personal assistant. You can create

451
00:25:21,080 --> 00:25:23,720
Speaker 3: these AI models that do a task for you, that

452
00:25:23,880 --> 00:25:27,959
Speaker 3: can you know, read emails and populate your calendar and

453
00:25:28,200 --> 00:25:30,560
Speaker 3: really work as a personal assistant for you. And that's

454
00:25:30,640 --> 00:25:33,399
Speaker 3: where where a lot of industries are moving towards. A

455
00:25:33,440 --> 00:25:36,439
Speaker 3: lot of these corporations they want agentic ai to be

456
00:25:36,520 --> 00:25:38,960
Speaker 3: a part of everyone's daily life. So as soon as

457
00:25:38,960 --> 00:25:43,040
Speaker 3: you wake up and log into your computer, your agent

458
00:25:43,840 --> 00:25:45,919
Speaker 3: can tell you, Hey, this is what you missed. This

459
00:25:46,000 --> 00:25:47,119
Speaker 3: is what you need to do, this is what you

460
00:25:47,119 --> 00:25:48,240
Speaker 3: should be prioritizing.

461
00:25:48,800 --> 00:25:49,840
Speaker 1: No, this is what you missed.

462
00:25:49,840 --> 00:25:50,399
Speaker 2: I was sleep.

463
00:25:53,359 --> 00:25:57,960
Speaker 1: I think that is definitely already in the works, So

464
00:25:58,040 --> 00:25:59,959
Speaker 1: I think it'll just be the next wave of adoption.

465
00:26:00,000 --> 00:26:01,800
Speaker 1: Shouldn't that we see next? Yeah.

466
00:26:01,840 --> 00:26:04,560
Speaker 3: And the other thing that I think people should be

467
00:26:04,600 --> 00:26:09,320
Speaker 3: thinking about and like keeping an eye on is the

468
00:26:09,359 --> 00:26:14,119
Speaker 3: fact that a lot of regulations don't exist yet, you know,

469
00:26:15,000 --> 00:26:17,879
Speaker 3: and we've already seen it with like the actors strike

470
00:26:18,040 --> 00:26:22,000
Speaker 3: because of folks being able to use their face and

471
00:26:22,359 --> 00:26:25,560
Speaker 3: make them do things, and like it's a it's a

472
00:26:25,600 --> 00:26:29,159
Speaker 3: really big issue when it comes to creativity because now,

473
00:26:29,359 --> 00:26:31,479
Speaker 3: you know, artists are feeling like, okay, art is going

474
00:26:31,560 --> 00:26:33,720
Speaker 3: to be lost if you can just take Michael B.

475
00:26:33,840 --> 00:26:37,040
Speaker 3: Jordan's face and body and get him to do whatever

476
00:26:37,119 --> 00:26:46,679
Speaker 3: you want him to do a jail, like he should

477
00:26:46,720 --> 00:26:48,680
Speaker 3: have a say in that, Like there should be loss

478
00:26:48,680 --> 00:26:51,399
Speaker 3: that say that that is illegal, and the laws have

479
00:26:51,520 --> 00:26:52,400
Speaker 3: not caught up yet.

480
00:26:52,440 --> 00:26:56,359
Speaker 2: The regulations have not caught up yet, So keep trying.

481
00:26:56,480 --> 00:27:05,200
Speaker 1: There yeah. Yeah.

482
00:27:05,640 --> 00:27:08,520
Speaker 3: You can find us on X and Instagram at Dope

483
00:27:08,600 --> 00:27:10,440
Speaker 3: Labs podcast, tt.

484
00:27:10,440 --> 00:27:14,240
Speaker 1: Is on X and Instagram at d R Underscore T Shoe.

485
00:27:13,960 --> 00:27:16,560
Speaker 2: And you can find Takiya at z said So.

486
00:27:16,960 --> 00:27:19,200
Speaker 1: Dope Labs is a production of Lamanada Media.

487
00:27:19,480 --> 00:27:24,080
Speaker 3: Our senior supervising producer is Kristin Lapour and our associate

488
00:27:24,119 --> 00:27:26,200
Speaker 3: producer is Isara Savez.

489
00:27:26,920 --> 00:27:30,639
Speaker 1: Dope Labs is sound design, edited and mixed by James Farber.

490
00:27:31,320 --> 00:27:35,520
Speaker 1: Lamanada Media's Vice President of Partnerships and Production is Jackie Danziger.

491
00:27:36,200 --> 00:27:40,320
Speaker 1: Executive producer from iHeart podcast is Katrina Norvil. Marketing lead

492
00:27:40,520 --> 00:27:44,879
Speaker 1: is Alison Kanter. Original music composed and produced by Taka

493
00:27:44,960 --> 00:27:50,000
Speaker 1: Yatsuzawa and Alex sugi Ura, with additional music by Elijah Harvey.

494
00:27:50,840 --> 00:27:54,280
Speaker 1: Dope Lab is executive produced by US T T Show

495
00:27:54,320 --> 00:27:55,880
Speaker 1: Dia and Nakia Wattley.