1
00:00:00,760 --> 00:00:04,600
Speaker 1: Earlier this month, the Prague Philharmonic Orchestra got on stage

2
00:00:04,640 --> 00:00:07,360
Speaker 1: to play a symphony written by the famous check composer

3
00:00:07,400 --> 00:00:11,240
Speaker 1: Antony Davor Jack. The orchestra was premiering a piece of

4
00:00:11,320 --> 00:00:27,040
Speaker 1: music the world had never heard before. Divorjack died more

5
00:00:27,080 --> 00:00:29,760
Speaker 1: than a hundred years ago, and he left behind just

6
00:00:29,960 --> 00:00:33,479
Speaker 1: the beginnings of a composition, just two sheets of music.

7
00:00:34,640 --> 00:00:38,200
Speaker 1: So a computer program powered by AI studied the rest

8
00:00:38,240 --> 00:00:42,640
Speaker 1: of divor Jack's music and completed the composer's unfinished work.

9
00:00:43,960 --> 00:00:46,920
Speaker 1: This is what the software spit out, something in the

10
00:00:46,960 --> 00:00:51,240
Speaker 1: style of Davor Jack that was still an entirely new symphony.

11
00:00:53,120 --> 00:00:56,520
Speaker 1: For years, we've worried about robots replacing the jobs of

12
00:00:56,560 --> 00:01:01,120
Speaker 1: truck drivers and call center asians and accountants, but experts

13
00:01:01,120 --> 00:01:04,600
Speaker 1: said creative jobs were going to be safe, that computers

14
00:01:04,640 --> 00:01:07,319
Speaker 1: were still so far from making things that are new

15
00:01:07,360 --> 00:01:11,720
Speaker 1: and subjective that move us. So consider the fact that

16
00:01:11,800 --> 00:01:32,200
Speaker 1: AI wrote most of this. Today, in the show Reporter,

17
00:01:32,280 --> 00:01:36,480
Speaker 1: in the tallya Drosiac visits three musicians using artificial intelligence

18
00:01:36,520 --> 00:01:40,000
Speaker 1: to make their music, including the guy behind this robo

19
00:01:40,080 --> 00:01:45,000
Speaker 1: Divorgac symphony. If computers can now compose something this beautiful,

20
00:01:45,520 --> 00:01:49,280
Speaker 1: what's left for us humans to do. Am Ito, you're

21
00:01:49,320 --> 00:02:19,440
Speaker 1: listening to Decrypted stay with us. M Hey Nat, welcome

22
00:02:19,440 --> 00:02:22,680
Speaker 1: to the show. Thanks for having me. So you're our

23
00:02:22,760 --> 00:02:25,679
Speaker 1: European tech reporter out of Brussels, and you normally write

24
00:02:25,720 --> 00:02:30,200
Speaker 1: about Europe's regulatory crackdown on the tech industry, but today

25
00:02:30,240 --> 00:02:33,840
Speaker 1: we're talking about something completely different. Yep. So we've been

26
00:02:33,840 --> 00:02:36,720
Speaker 1: writing about the use of AI and self driving cars

27
00:02:36,760 --> 00:02:40,680
Speaker 1: and chatbots and voice assistants, but in the world of music,

28
00:02:40,720 --> 00:02:43,840
Speaker 1: this is all still kind of new. I mean, musicians

29
00:02:43,960 --> 00:02:46,480
Speaker 1: just started using AI as a tool a few years ago.

30
00:02:49,040 --> 00:02:51,880
Speaker 1: So I talked to three different composers who are using

31
00:02:51,919 --> 00:02:55,040
Speaker 1: AI in really different ways. The first guy I want

32
00:02:55,040 --> 00:02:59,079
Speaker 1: to introduce you to is Ben Waki. He's a French musician.

33
00:02:59,680 --> 00:03:02,600
Speaker 1: He's known in France for these French pop songs, but

34
00:03:02,639 --> 00:03:06,080
Speaker 1: he's also composed music for artists like the famous French

35
00:03:06,160 --> 00:03:09,480
Speaker 1: rocker Johnny Halliday. So this is a pretty famous guy

36
00:03:09,520 --> 00:03:12,720
Speaker 1: out of France. Yeah, he's well known, but he's gotten

37
00:03:12,720 --> 00:03:16,480
Speaker 1: even more attention recently because of his experimentations with AI

38
00:03:16,520 --> 00:03:20,040
Speaker 1: generated music. And how did he get into that. So

39
00:03:20,160 --> 00:03:23,320
Speaker 1: there's this guy called Francois Pachet who for a long

40
00:03:23,360 --> 00:03:26,680
Speaker 1: time was ahead of the research lab run by Sony

41
00:03:26,760 --> 00:03:30,560
Speaker 1: in Paris, and Paschet has kind of become known as

42
00:03:30,680 --> 00:03:34,520
Speaker 1: the godfather of AI music because he's really done a

43
00:03:34,520 --> 00:03:38,040
Speaker 1: lot of research in this area. France called me a

44
00:03:38,040 --> 00:03:42,320
Speaker 1: long time ago. He discovered my my songs in the

45
00:03:42,400 --> 00:03:46,200
Speaker 1: late nineties, and he was interested by my way of

46
00:03:46,400 --> 00:03:52,840
Speaker 1: composing songs, always searching for unexpected coat changes, and and

47
00:03:52,920 --> 00:03:55,360
Speaker 1: he invited ben Wata's lab to try out some of

48
00:03:55,400 --> 00:03:57,920
Speaker 1: the tools that he's been developing over the past few years.

49
00:03:58,720 --> 00:04:01,720
Speaker 1: And one of the tools that Pesh's team developed is

50
00:04:01,760 --> 00:04:05,000
Speaker 1: this machine learning tool that generates music. So how does

51
00:04:05,040 --> 00:04:09,920
Speaker 1: it work? So first, the program basically ingests a bunch

52
00:04:09,960 --> 00:04:12,560
Speaker 1: of sheets of music that the musician wants the program

53
00:04:12,600 --> 00:04:16,240
Speaker 1: to train on, and then the computer cuts it up

54
00:04:16,279 --> 00:04:19,240
Speaker 1: into really tiny bits of music and rearranges it into

55
00:04:19,279 --> 00:04:22,520
Speaker 1: a whole new composition. So let me give you an example.

56
00:04:22,680 --> 00:04:25,880
Speaker 1: So a few years ago, ben wa fed this program

57
00:04:25,920 --> 00:04:29,440
Speaker 1: four seventy lead sheets of jazz standards from the thirties

58
00:04:29,600 --> 00:04:33,120
Speaker 1: up through the fifties and sixties, and out came these

59
00:04:33,120 --> 00:04:42,679
Speaker 1: two bars of music that he really liked, very jazzy.

60
00:04:43,000 --> 00:04:46,920
Speaker 1: I can definitely sense the influence. Yeah, so Ben Wah

61
00:04:47,000 --> 00:04:50,279
Speaker 1: really liked the sound of what the computer came up with.

62
00:04:50,760 --> 00:04:56,480
Speaker 1: It's just two bars at the beginning, UM with ascendant

63
00:04:57,120 --> 00:05:03,120
Speaker 1: melodique movement. That's it is really unexpected and that I loved.

64
00:05:03,440 --> 00:05:07,360
Speaker 1: I was at the first first time. So he asked

65
00:05:07,360 --> 00:05:10,800
Speaker 1: the system to generate new ones based on those new parameters,

66
00:05:11,080 --> 00:05:15,480
Speaker 1: and then I reiterate with the machine like you know,

67
00:05:15,560 --> 00:05:20,080
Speaker 1: if if it were a workmate something like that. It's

68
00:05:20,120 --> 00:05:24,039
Speaker 1: kind of like, um, two artists collaborating on a song,

69
00:05:24,160 --> 00:05:27,320
Speaker 1: but it's just that one of them is a machine exactly.

70
00:05:27,720 --> 00:05:31,240
Speaker 1: And so that ultimately led to this full song that

71
00:05:31,320 --> 00:05:39,640
Speaker 1: he released under his artist name Skeige. It's it's ding

72
00:05:40,080 --> 00:06:02,640
Speaker 1: those doings are those use Yeah, I definitely recognized the

73
00:06:02,680 --> 00:06:06,479
Speaker 1: melody from earlier, but it doesn't sound as jazzy. It

74
00:06:06,520 --> 00:06:10,600
Speaker 1: sounds some a little old timey and familiar. At the end,

75
00:06:10,680 --> 00:06:16,200
Speaker 1: I had really, uh, very interesting melody. I was able

76
00:06:16,240 --> 00:06:20,680
Speaker 1: to say, Okay, I'm proud of this melody and I

77
00:06:20,680 --> 00:06:25,240
Speaker 1: think it's new. I think it's interesting because I couldn't

78
00:06:25,279 --> 00:06:31,800
Speaker 1: have made it by myself. So the collaboration with AI

79
00:06:32,160 --> 00:06:37,360
Speaker 1: is really interesting because it gives me something new. So

80
00:06:37,880 --> 00:06:41,280
Speaker 1: for this song, the AI was really just the initial inspiration,

81
00:06:42,040 --> 00:06:45,760
Speaker 1: but Benuah made a ton of creative decisions to bring

82
00:06:45,760 --> 00:06:48,000
Speaker 1: this song to the finish line. So in that sense,

83
00:06:48,040 --> 00:06:51,240
Speaker 1: maybe you can call this AI light right. And so

84
00:06:51,320 --> 00:06:56,520
Speaker 1: Bena released this song back in and since then, Francois

85
00:06:56,560 --> 00:07:00,280
Speaker 1: Pachet from Sony moved over to Spotify, and Ben has

86
00:07:00,320 --> 00:07:04,680
Speaker 1: been helping francoise new work at Spotify and they're they're

87
00:07:04,680 --> 00:07:09,120
Speaker 1: building these tools from scratch with researchers for various tasks

88
00:07:09,200 --> 00:07:12,440
Speaker 1: that are involved in creating a song. And one of

89
00:07:12,480 --> 00:07:16,520
Speaker 1: those tools that they've developed, they've incorporated into Ben's new

90
00:07:16,560 --> 00:07:20,280
Speaker 1: album called American Folk Songs. So been one and I

91
00:07:20,360 --> 00:07:23,120
Speaker 1: talked about one song called black is a Color. The

92
00:07:23,160 --> 00:07:26,160
Speaker 1: melody is based on this famous folk song called black

93
00:07:26,240 --> 00:07:28,560
Speaker 1: is the Color of My True Loves hair. Um. It's

94
00:07:28,600 --> 00:07:32,240
Speaker 1: been covered by people like Pete Seeger Nina Simone. I'm

95
00:07:32,240 --> 00:07:36,160
Speaker 1: sure you've heard it before. I actually don't think I have.

96
00:07:36,640 --> 00:07:39,920
Speaker 1: UM is that bad? I grew up into dead well,

97
00:07:40,040 --> 00:07:43,600
Speaker 1: I think most of our listeners have. So Ben Wass

98
00:07:43,600 --> 00:07:47,360
Speaker 1: song is based on this acapella version sung by Pete Seeger.

99
00:07:47,720 --> 00:07:55,080
Speaker 1: Black Black, Black, Color of My True loves hand and

100
00:07:55,120 --> 00:07:58,200
Speaker 1: he had Spotify's AI program. I'll add all these harmonies

101
00:07:58,200 --> 00:08:00,360
Speaker 1: in the background to make it sound so much a chair.

102
00:08:00,720 --> 00:08:07,640
Speaker 1: Black black, black is the color of my true love's head.

103
00:08:09,640 --> 00:08:17,280
Speaker 1: Her face is something wondrous fair. But yourlest eyes and

104
00:08:17,520 --> 00:08:29,720
Speaker 1: the daintiest hands. I love the ground she STIs. I love.

105
00:08:30,720 --> 00:08:32,800
Speaker 1: I thought that it could be very black in Sparring,

106
00:08:32,880 --> 00:08:36,440
Speaker 1: to have something very sophisticated with a simple melody like

107
00:08:36,720 --> 00:08:39,679
Speaker 1: black is the color, and I got a really unnam

108
00:08:39,760 --> 00:08:43,080
Speaker 1: is amazing result. I was blown away by this result.

109
00:08:44,559 --> 00:08:51,840
Speaker 1: Black black, black is the color of my true loves

110
00:08:51,960 --> 00:09:04,400
Speaker 1: her her face is something wondrous fair. I think, no, no,

111
00:09:04,400 --> 00:09:08,600
Speaker 1: no human could have composed this string arrangement because it's

112
00:09:09,040 --> 00:09:12,520
Speaker 1: too weird. There are two strange things in it too,

113
00:09:12,600 --> 00:09:20,440
Speaker 1: strange chord changes and internal movements. But it's still really,

114
00:09:20,480 --> 00:09:25,920
Speaker 1: really beautiful. And I think that this kind of result

115
00:09:26,679 --> 00:09:32,040
Speaker 1: is very encouraging for for musicians like me and after me,

116
00:09:32,720 --> 00:09:38,280
Speaker 1: musicians did something new, something that they find in Sparring

117
00:09:38,400 --> 00:09:54,920
Speaker 1: for their compositions. You know, it's really different from the

118
00:09:55,040 --> 00:09:59,240
Speaker 1: music i'man're still listening to, uh and definitely a little weird,

119
00:09:59,360 --> 00:10:02,680
Speaker 1: but I think I like it. It's beautiful. Yeah, it

120
00:10:02,760 --> 00:10:05,080
Speaker 1: sounds I don't know, both old and knew at the

121
00:10:05,120 --> 00:10:07,680
Speaker 1: same time. But what's interesting to me is how he's

122
00:10:07,720 --> 00:10:11,760
Speaker 1: really revamping this traditional folk music, and I can imagine

123
00:10:12,160 --> 00:10:14,920
Speaker 1: it might sound strange for some American audiences who grew

124
00:10:15,000 --> 00:10:17,880
Speaker 1: up listening to this type of music and then have

125
00:10:18,040 --> 00:10:22,480
Speaker 1: it repackaged in a totally different way. So the original

126
00:10:22,520 --> 00:10:26,520
Speaker 1: melody and the lyrics were obviously written by a human

127
00:10:27,120 --> 00:10:30,680
Speaker 1: a long time ago, but the arrangement we hear in

128
00:10:30,720 --> 00:10:33,439
Speaker 1: the background was all composed by a machine. Is that right?

129
00:10:34,000 --> 00:10:37,040
Speaker 1: Mostly so, Ben Watt did make a few choices and

130
00:10:37,240 --> 00:10:40,400
Speaker 1: editions along the way. He used Pete Seeger's voice to

131
00:10:40,440 --> 00:10:43,719
Speaker 1: create an au ac choir, for instance. He also composed

132
00:10:43,720 --> 00:10:47,959
Speaker 1: and recorded some chord sequences. One was in the style

133
00:10:48,160 --> 00:10:52,560
Speaker 1: of like an epic Game of Thrones type sound. Another

134
00:10:52,600 --> 00:10:55,560
Speaker 1: one was in a Bossa Nova style, and he fed

135
00:10:55,600 --> 00:10:58,320
Speaker 1: that into the machine, which in turn generated a whole

136
00:10:58,360 --> 00:11:02,200
Speaker 1: new string arrangement for the background on sound and for

137
00:11:02,240 --> 00:11:05,320
Speaker 1: the final recording. He also chose the string quartet and

138
00:11:05,360 --> 00:11:08,800
Speaker 1: the director to play that new composition. Okay, so still

139
00:11:08,840 --> 00:11:12,440
Speaker 1: a lot of human intervention there, Yeah, exactly. Uh, And

140
00:11:12,480 --> 00:11:15,240
Speaker 1: this is a tool that Spotify is making available to everyone,

141
00:11:15,840 --> 00:11:18,400
Speaker 1: so not quite yet. They eventually want to release the

142
00:11:18,480 --> 00:11:21,320
Speaker 1: tool on open source on the Internet, possibly by the

143
00:11:21,400 --> 00:11:23,800
Speaker 1: end of next year, but it could take longer than that.

144
00:11:27,200 --> 00:11:29,560
Speaker 1: And so the next composer and entrepreneur I want to

145
00:11:29,600 --> 00:11:33,040
Speaker 1: introduce you to is this guy called Pierre Bahl and

146
00:11:33,120 --> 00:11:37,360
Speaker 1: he's based in Luxembourg. He studied computer science. He's a musician,

147
00:11:37,640 --> 00:11:40,160
Speaker 1: his father is a film and music producer, his mother

148
00:11:40,240 --> 00:11:42,680
Speaker 1: is a singer, and he and his brother started this

149
00:11:42,720 --> 00:11:46,040
Speaker 1: company called Ava. This is the startup that made the

150
00:11:46,120 --> 00:11:48,760
Speaker 1: Divor Jack inspired symphony that we heard at the top

151
00:11:48,760 --> 00:11:51,199
Speaker 1: of the show right, and he and his brother were

152
00:11:51,240 --> 00:11:54,880
Speaker 1: inspired by how important music is in film, but how

153
00:11:54,920 --> 00:11:58,160
Speaker 1: long that process can take in terms of creating a soundtrack.

154
00:11:58,679 --> 00:12:00,680
Speaker 1: So they wanted to see if they could train AI

155
00:12:00,960 --> 00:12:04,000
Speaker 1: and see if it could basically help a composer create

156
00:12:04,040 --> 00:12:07,200
Speaker 1: that type of music, not just for films, but also

157
00:12:07,280 --> 00:12:12,120
Speaker 1: for commercials, promotional videos, video games, and also just generally

158
00:12:12,320 --> 00:12:16,120
Speaker 1: assisting composers in their work. So purefed the machine with

159
00:12:16,360 --> 00:12:20,920
Speaker 1: thirty thou scores of history's greatest composers from Bach, Bitovin,

160
00:12:21,040 --> 00:12:25,719
Speaker 1: and Mozart. So basically it all starts by teaching and

161
00:12:25,840 --> 00:12:30,680
Speaker 1: algorithm to to learn the patterns and music. You know

162
00:12:30,760 --> 00:12:34,160
Speaker 1: that there's this common knowledge that music is sort of

163
00:12:34,200 --> 00:12:37,520
Speaker 1: emotional and it's it's the opposite of math. But actually,

164
00:12:37,559 --> 00:12:40,679
Speaker 1: if you look at music very carefully, there's a lot

165
00:12:40,720 --> 00:12:44,600
Speaker 1: of patterns in it. Um So AVA understands all these

166
00:12:44,640 --> 00:12:47,680
Speaker 1: patterns which are very very mathematical at the core, and

167
00:12:47,760 --> 00:12:50,880
Speaker 1: EVA uses that to generate different kinds of music depending

168
00:12:50,880 --> 00:12:54,000
Speaker 1: on what you're looking for. So let me play you

169
00:12:54,160 --> 00:12:57,440
Speaker 1: two songs in a completely different style. The first one

170
00:12:57,600 --> 00:13:21,320
Speaker 1: is this classical song called I Am a h This

171
00:13:21,440 --> 00:13:26,760
Speaker 1: makes me think of a winter ballet. Does that make sense? Absolutely? Yeah,

172
00:13:27,040 --> 00:13:32,160
Speaker 1: I totally think of forest blanketed and snow and little

173
00:13:32,200 --> 00:13:35,520
Speaker 1: bunny rabbits popping by. I see that too. Let me

174
00:13:35,559 --> 00:13:39,480
Speaker 1: play a totally different song. This is a pop song

175
00:13:39,559 --> 00:13:53,720
Speaker 1: called Guiding Light. So background music is a big industry,

176
00:13:54,080 --> 00:13:58,200
Speaker 1: and this just underscores the strength of Ava's business model.

177
00:13:58,440 --> 00:14:00,319
Speaker 1: I mean, for the most part, if you're come penny

178
00:14:00,520 --> 00:14:03,440
Speaker 1: or like a podcast, you can pay a composer to

179
00:14:03,480 --> 00:14:06,000
Speaker 1: make custom music, or you can use these catalogs with

180
00:14:06,120 --> 00:14:08,600
Speaker 1: songs that you're licensed to use. Right, That's what we

181
00:14:08,679 --> 00:14:13,240
Speaker 1: do for this show. We use these big catalogs of

182
00:14:13,360 --> 00:14:16,920
Speaker 1: songs UM that we're allowed to use, but it's really

183
00:14:16,960 --> 00:14:20,320
Speaker 1: hard because you know, these are already existing songs. A

184
00:14:20,320 --> 00:14:22,760
Speaker 1: lot of other shows use them, and I don't know,

185
00:14:22,800 --> 00:14:25,240
Speaker 1: I know, I drive our producers crazy because I'm really

186
00:14:25,280 --> 00:14:28,360
Speaker 1: picky about which tracks they use for different sections of

187
00:14:28,360 --> 00:14:31,440
Speaker 1: our show. Yeah, yeah, I mean, so that's the big

188
00:14:31,480 --> 00:14:35,080
Speaker 1: opportunity for EVA. You could have a custom made piece

189
00:14:35,120 --> 00:14:37,520
Speaker 1: of music, but because you're having a computer do it,

190
00:14:37,520 --> 00:14:41,960
Speaker 1: it's probably cheaper and faster than having a human composer

191
00:14:42,000 --> 00:14:45,280
Speaker 1: do it. How long does it take to create like

192
00:14:45,320 --> 00:14:47,800
Speaker 1: a new song, you know, like three minutes, three and

193
00:14:47,840 --> 00:14:50,560
Speaker 1: a half minutes, So it depends. It depends with the

194
00:14:50,560 --> 00:14:52,960
Speaker 1: algorithms that we use, because we have different algorithms that

195
00:14:53,040 --> 00:14:56,320
Speaker 1: we've used of the basket full of years. The very

196
00:14:56,360 --> 00:14:59,720
Speaker 1: first ones that we had it took usually forty eight

197
00:14:59,720 --> 00:15:02,320
Speaker 1: to say many two hours because we had to retrain

198
00:15:03,160 --> 00:15:09,480
Speaker 1: av the influences that we wanted her to specifically emanate UM.

199
00:15:09,560 --> 00:15:12,800
Speaker 1: But more recently it takes about like forty five seconds

200
00:15:12,840 --> 00:15:15,440
Speaker 1: to a minute to create a piece of music from

201
00:15:15,640 --> 00:15:20,160
Speaker 1: three minutes long. Now, so you know how earlier the

202
00:15:20,240 --> 00:15:24,320
Speaker 1: first example from ben Wa was just this starting point

203
00:15:24,760 --> 00:15:27,880
Speaker 1: and there was still like a ton of human intervention

204
00:15:28,000 --> 00:15:32,440
Speaker 1: that was involved. How complete is the music that Ava's

205
00:15:32,480 --> 00:15:36,720
Speaker 1: software is generating, well, Pierce, is it kind of depends

206
00:15:36,760 --> 00:15:39,680
Speaker 1: on who's playing around with the tool, Like someone who's

207
00:15:39,880 --> 00:15:43,200
Speaker 1: highly musically trained might make more tweaks and changes to it,

208
00:15:43,280 --> 00:15:46,920
Speaker 1: whereas someone who isn't might leave it as is, though

209
00:15:46,920 --> 00:15:48,960
Speaker 1: the quality might not be as good as a result.

210
00:15:49,560 --> 00:15:53,160
Speaker 1: So in the songs that we just listened to earlier,

211
00:15:54,360 --> 00:15:58,520
Speaker 1: did human composers tweak those songs? So nope, that was

212
00:15:58,680 --> 00:16:04,160
Speaker 1: composed entirely by Eva, but humans did assign the different

213
00:16:04,200 --> 00:16:26,600
Speaker 1: parts of music two different instruments. We'll be right back, Okay,

214
00:16:26,680 --> 00:16:29,640
Speaker 1: So before the break, we met two musicians who are

215
00:16:29,720 --> 00:16:33,520
Speaker 1: using AI to make music in really different ways. Yeah,

216
00:16:33,560 --> 00:16:37,680
Speaker 1: so Bena uses AI kind of as a collaborator, Pierre

217
00:16:37,680 --> 00:16:42,440
Speaker 1: Berrow uses AI specifically for background music. But the last

218
00:16:42,440 --> 00:16:44,880
Speaker 1: guy I want to introduce you to takes it even further.

219
00:16:45,240 --> 00:16:49,480
Speaker 1: His name's ash Kusha. Yeah, thanks so much for for

220
00:16:49,680 --> 00:16:54,000
Speaker 1: having us sober. We're really and he's an electronic musician.

221
00:16:54,080 --> 00:16:57,960
Speaker 1: He's Iranian born and based in London. The question for

222
00:16:58,080 --> 00:17:02,680
Speaker 1: me always was how I can I can make very intriguing,

223
00:17:03,160 --> 00:17:07,760
Speaker 1: complete and complex piece of music only using a computer.

224
00:17:08,359 --> 00:17:11,919
Speaker 1: And also what if the part of the composition that

225
00:17:11,960 --> 00:17:15,800
Speaker 1: I'm thinking of can be partly made by the computer.

226
00:17:16,480 --> 00:17:20,160
Speaker 1: So that question took me twelve years to go through

227
00:17:20,200 --> 00:17:23,520
Speaker 1: all the different parts, from instruments, replicating sound of violence,

228
00:17:24,080 --> 00:17:29,320
Speaker 1: field recording, synthesizing voice, and finally to finding a way

229
00:17:29,359 --> 00:17:35,120
Speaker 1: to generate lyrics. Um, that is what initiated oxyen. It

230
00:17:35,240 --> 00:17:41,919
Speaker 1: was trying to tackle what is a non human creative engine.

231
00:17:42,200 --> 00:17:45,800
Speaker 1: And a few years ago Ash started building virtual entertainers

232
00:17:45,840 --> 00:17:47,960
Speaker 1: for video games. So you know how they have these

233
00:17:47,960 --> 00:17:52,399
Speaker 1: avatars and video games, is that like Mario and Mario

234
00:17:52,520 --> 00:17:57,680
Speaker 1: kart am I adding itself is a non gamer. Well,

235
00:17:57,680 --> 00:18:01,480
Speaker 1: so Ash made a virtual avatar that makes music and

236
00:18:01,720 --> 00:18:04,639
Speaker 1: he called her Yonah. So this is like a whole

237
00:18:05,359 --> 00:18:09,359
Speaker 1: virtual character that writes her own music. Yeah, so you

238
00:18:09,400 --> 00:18:13,320
Speaker 1: can kind of think of it as infusing an AI

239
00:18:13,400 --> 00:18:17,080
Speaker 1: with a personality too. So in this case, Ash trained

240
00:18:17,160 --> 00:18:21,120
Speaker 1: Yonah's AI on Margaret Atwood's novels and also on articles

241
00:18:21,160 --> 00:18:25,840
Speaker 1: about teenage life. So Ash's company ox Human created Yonah,

242
00:18:26,359 --> 00:18:30,840
Speaker 1: and she's an angsty teen in a dystopian world. That's

243
00:18:30,880 --> 00:18:33,480
Speaker 1: a really good way of putting it. So, yeah, let

244
00:18:33,480 --> 00:18:36,359
Speaker 1: me play you a song that Ash just released in September,

245
00:18:36,480 --> 00:18:41,720
Speaker 1: written and sung by Yona. I never felt alone. You

246
00:18:41,880 --> 00:18:47,359
Speaker 1: never said a word. I fell from my prone. You

247
00:18:47,440 --> 00:18:53,280
Speaker 1: didn't want me there, you didn't want me near, You

248
00:18:53,320 --> 00:19:04,440
Speaker 1: didn't want me there. I never heard you say it's

249
00:19:04,480 --> 00:19:18,960
Speaker 1: every life I live. I it's definitely very weird, very dystopian.

250
00:19:20,440 --> 00:19:21,919
Speaker 1: I feel like she's going to come and kill me

251
00:19:21,960 --> 00:19:25,080
Speaker 1: in my sleep. Yeah, I mean it's definitely a little eerie.

252
00:19:25,240 --> 00:19:27,159
Speaker 1: And some of the comments on the video are kind

253
00:19:27,200 --> 00:19:30,639
Speaker 1: of funny, like one user says this is creepy, and

254
00:19:30,720 --> 00:19:32,800
Speaker 1: other one says, don't listen to this when you're high,

255
00:19:33,680 --> 00:19:37,360
Speaker 1: and another guy says, don't mess with organic music. Um,

256
00:19:37,480 --> 00:19:40,080
Speaker 1: And I think part of what makes it extra creepy

257
00:19:40,200 --> 00:19:42,560
Speaker 1: is that you can watch Yona sing this song on

258
00:19:42,600 --> 00:19:46,600
Speaker 1: YouTube and she looks half human half robot. If you cry,

259
00:19:46,680 --> 00:19:52,400
Speaker 1: would every smile, I'll make you stare? I get where

260
00:19:52,400 --> 00:19:56,280
Speaker 1: do you were? Were they all happened? You know? So

261
00:19:56,640 --> 00:19:59,960
Speaker 1: this definitely really creeps me out. But now that I'm

262
00:20:00,080 --> 00:20:04,520
Speaker 1: thinking about it, maybe that's the point that it's disturbing yeah.

263
00:20:04,520 --> 00:20:08,040
Speaker 1: So Ash's goal with Jonah and his other characters he

264
00:20:08,080 --> 00:20:12,280
Speaker 1: has other AI based characters is to really have um

265
00:20:12,560 --> 00:20:15,280
Speaker 1: have them in vogue emotion and the people listening to it.

266
00:20:15,880 --> 00:20:19,800
Speaker 1: So in her music you get this sense that she's sad,

267
00:20:20,520 --> 00:20:24,120
Speaker 1: and Ash says it's easier to get human reaction through

268
00:20:24,160 --> 00:20:28,040
Speaker 1: sad and romantic music. Um, we want to make something

269
00:20:28,080 --> 00:20:30,320
Speaker 1: that when you listen to Yon know you you believe

270
00:20:30,400 --> 00:20:35,840
Speaker 1: that there's something being said. So the nature of of

271
00:20:36,480 --> 00:20:38,439
Speaker 1: the scales that we use, and and the and the

272
00:20:38,480 --> 00:20:40,560
Speaker 1: type of music that we use, it's very personal, it's

273
00:20:40,640 --> 00:20:43,960
Speaker 1: very deep, it's very um yeah, it's in the sense

274
00:20:44,000 --> 00:20:47,000
Speaker 1: is romantic. I think the other thing he says is

275
00:20:47,040 --> 00:20:50,160
Speaker 1: that he's not trying to replace traditional music in any way,

276
00:20:50,200 --> 00:20:52,800
Speaker 1: but to create a whole new form of sound and

277
00:20:52,800 --> 00:20:55,440
Speaker 1: a whole new genre. We we could try and make

278
00:20:55,480 --> 00:20:59,720
Speaker 1: the best rock track, but I think there are many

279
00:20:59,760 --> 00:21:03,240
Speaker 1: many good musicians that do that. Creating genres takes time,

280
00:21:03,320 --> 00:21:06,760
Speaker 1: and it's just a subculture and it's seen pretty much,

281
00:21:07,000 --> 00:21:08,800
Speaker 1: so it takes time, so we want to give it time.

282
00:21:09,080 --> 00:21:12,320
Speaker 1: At the beginning of soundings sometimes wonky and it's a

283
00:21:12,320 --> 00:21:16,000
Speaker 1: bit weird, but that's what we're looking for. And so

284
00:21:16,040 --> 00:21:20,120
Speaker 1: with that also goes a different form of consumption. So

285
00:21:22,000 --> 00:21:24,800
Speaker 1: this type of music might be consumed through games or

286
00:21:24,880 --> 00:21:28,159
Speaker 1: virtual reality or different types of concerts, and there's the

287
00:21:28,320 --> 00:21:32,000
Speaker 1: market for that. Yeah. I mean he's already ashes already

288
00:21:32,000 --> 00:21:37,200
Speaker 1: monetizing his virtual entertainers. I mean Jonah performed at events

289
00:21:37,240 --> 00:21:41,439
Speaker 1: and shows and at a poetry festival recently. Wow. So

290
00:21:41,480 --> 00:21:45,000
Speaker 1: it's it's kind of like um hatsen Amiku and Japan.

291
00:21:45,320 --> 00:21:47,840
Speaker 1: Do you know about her? I have no idea who

292
00:21:47,920 --> 00:21:50,679
Speaker 1: that is. So she's just like they call her a

293
00:21:50,800 --> 00:21:54,919
Speaker 1: virtual pop idol um but it's it's not like they

294
00:21:55,000 --> 00:21:57,919
Speaker 1: have like tons of people who go to her concerts

295
00:21:58,040 --> 00:22:00,560
Speaker 1: or her or it or whatever you would call it.

296
00:22:09,760 --> 00:22:22,560
Speaker 1: M Yeah, and she gets projected onto a screen and

297
00:22:22,920 --> 00:22:26,359
Speaker 1: you know, they use the software for her um to

298
00:22:26,960 --> 00:22:30,000
Speaker 1: to sing, if you can call it that. That is

299
00:22:30,080 --> 00:22:34,320
Speaker 1: the idea of the modern, the contemporary of pop culture,

300
00:22:34,359 --> 00:22:38,040
Speaker 1: which is alter egos and creating almost cartoon characters. So

301
00:22:38,080 --> 00:22:40,960
Speaker 1: we we are kind of adopting all of these things

302
00:22:41,000 --> 00:22:44,720
Speaker 1: into a new form of expression. This is why we're

303
00:22:44,760 --> 00:22:48,200
Speaker 1: not apologetic about, you know, not being digital or being

304
00:22:48,240 --> 00:22:51,200
Speaker 1: a bit weird and not complete or perfect. That's how

305
00:22:51,240 --> 00:22:55,280
Speaker 1: she is. And I think, um, the gaming generation, the

306
00:22:55,400 --> 00:22:59,040
Speaker 1: video game generation and Juris and Alpha is going to

307
00:22:59,080 --> 00:23:10,560
Speaker 1: be more except thing. So now we started this episode

308
00:23:10,600 --> 00:23:15,480
Speaker 1: with this idea that creativity is the last frontier of AI,

309
00:23:16,000 --> 00:23:18,360
Speaker 1: and of course that begs the question, if AI can

310
00:23:18,359 --> 00:23:21,919
Speaker 1: now make music, is there anything left for us humans

311
00:23:21,920 --> 00:23:25,879
Speaker 1: to do? And you know, having listened to your conversations

312
00:23:26,040 --> 00:23:29,199
Speaker 1: with Benoir and Pierre and Ash and having listened to

313
00:23:29,240 --> 00:23:33,400
Speaker 1: their AI generated music, I think it's all really interesting

314
00:23:33,440 --> 00:23:36,600
Speaker 1: and there were definitely sections of their songs that I

315
00:23:36,640 --> 00:23:39,359
Speaker 1: really liked. But the question we posed at the start

316
00:23:39,520 --> 00:23:42,560
Speaker 1: feels still very premature to me. It feels like we're

317
00:23:42,600 --> 00:23:47,760
Speaker 1: still very very far from AI replacing human composers. And

318
00:23:47,840 --> 00:23:51,480
Speaker 1: I personally still prefer all of the artists I listened

319
00:23:51,520 --> 00:23:54,480
Speaker 1: to who are writing their own stuff instead of having

320
00:23:54,480 --> 00:23:57,680
Speaker 1: computers write their stuff. You know what. It reminds me

321
00:23:58,080 --> 00:24:01,359
Speaker 1: of the conversation that I had with a Barrow from Ava,

322
00:24:01,480 --> 00:24:05,800
Speaker 1: and he said that people were really upset when synthesizers

323
00:24:05,800 --> 00:24:08,840
Speaker 1: were first introduced thirty years ago, and now it's totally normal.

324
00:24:09,359 --> 00:24:11,600
Speaker 1: So I guess I can imagine a day when I

325
00:24:11,640 --> 00:24:14,040
Speaker 1: was just going to be another tool for musicians and

326
00:24:14,160 --> 00:24:26,160
Speaker 1: people won't even bat an eye. Natalia drags Jack. Thanks

327
00:24:26,160 --> 00:24:28,879
Speaker 1: for coming on the show today. Thanks for having me Aki.

328
00:24:33,840 --> 00:24:35,560
Speaker 1: So before I let you guys go, I want to

329
00:24:35,640 --> 00:24:37,840
Speaker 1: let you know that this is the last episode of

330
00:24:37,880 --> 00:24:40,960
Speaker 1: this season of Decrypted, and we're going to be publishing

331
00:24:40,960 --> 00:24:44,679
Speaker 1: a special bonus episode in two weeks where this show's

332
00:24:44,720 --> 00:24:47,240
Speaker 1: original co host, Brad Stone and I are going to

333
00:24:47,320 --> 00:24:50,240
Speaker 1: be talking about our favorite episodes that we've ever done

334
00:24:50,600 --> 00:24:52,680
Speaker 1: and update you on the people and the stories we've

335
00:24:52,680 --> 00:24:55,439
Speaker 1: covered over the years. So if there's an episode you

336
00:24:55,440 --> 00:24:58,679
Speaker 1: want to talk about, tweet at Meat at seven or

337
00:24:58,760 --> 00:25:01,760
Speaker 1: email me a I t O one six at Bloomberg

338
00:25:01,800 --> 00:25:06,639
Speaker 1: dot net. Decrypted is hosted by me. Shawn ween is

339
00:25:06,640 --> 00:25:10,760
Speaker 1: our executive producer. Ethan Brooks mixes show today. Nate Linkson

340
00:25:10,840 --> 00:25:14,439
Speaker 1: and Neville Gillette helped with recordings. Francesca Levy is the

341
00:25:14,480 --> 00:25:17,760
Speaker 1: head of Bloomberg Podcasts. We'll see you in two weeks