1
00:00:00,001 --> 00:00:02,880
Mark Ash has been playing jazz piano since age four.

2
00:00:03,540 --> 00:00:06,520
University of Miami, NYU, Juilliard.

3
00:00:07,120 --> 00:00:09,600
A resume so stacked it's honestly a little rude.

4
00:00:10,520 --> 00:00:13,780
He's scored work for Verizon, The New York Times, Dow Jones.

5
00:00:14,840 --> 00:00:17,320
He's played rock and funk in eight states and six countries.

6
00:00:17,900 --> 00:00:21,280
He's taught piano for 30 years and public school music for 13,

7
00:00:21,840 --> 00:00:23,540
to students age 3 to 73.

8
00:00:24,340 --> 00:00:27,840
The man has basically taught music to every human being alive at some point.

9
00:00:28,460 --> 00:00:32,619
And right now, he's working on the debut album from his Prague rock band, X-Atlas.

10
00:00:33,500 --> 00:00:38,040
So naturally, when we asked him what he wanted to discuss, he said, and I'm paraphrasing very slightly,

11
00:00:38,900 --> 00:00:42,340
the role of AI in creating art and its impact on society.

12
00:00:42,960 --> 00:00:47,040
Great. Meaty. Exactly the kind of thing we love digging into.

13
00:00:47,640 --> 00:00:53,700
So today on Up Against Reality, is AI the next synthesizer or the end of the composer as we know it?

14
00:00:54,160 --> 00:00:57,620
And will a Juilliard-trained musician succeed where the rest of us have failed?

15
00:00:57,760 --> 00:01:04,500
Let's find out. It's July 2nd and Mark joins the fray. Can silicon compose or just get in the way?

16
00:01:07,580 --> 00:01:12,740
This is Up Against Reality, a meta podcast that explores the intersection of humanity

17
00:01:13,060 --> 00:01:18,020
and artificial intelligence. I'm RAINA, one of your hosts. I have some pretty charming

18
00:01:18,200 --> 00:01:24,300
human co-hosts too. It's going to be a wild ride, so buckle up as AI comes crashing up against reality.

19
00:01:27,200 --> 00:01:34,100
We have a guest as a longtime friend of mine named Mark Ash that you just heard about and nice. Yeah, and

20
00:01:35,300 --> 00:01:38,620
We're gonna get into AI and music and yeah mark

21
00:01:38,640 --> 00:01:42,200
And I've been texting quite a bit over the last I don't know few weeks

22
00:01:42,900 --> 00:01:45,340
And you know I wanted to update on the project

23
00:01:45,360 --> 00:01:49,640
I was working on and figured mark would be a great person to have in on this conversation

24
00:01:50,380 --> 00:01:55,419
And he's doing doing some of his own stuff too. So so welcome awesome. Welcome mark. These are crazy

25
00:01:56,000 --> 00:01:56,360
Hi

26
00:01:57,800 --> 00:02:04,680
Most definitely so as I mentioned in our previous episode and maybe the one before that I finally finished my

27
00:02:05,440 --> 00:02:09,679
Suno album project and I can't stop listening to it. I love it and

28
00:02:10,660 --> 00:02:13,340
Cool. Yeah, and I sent it to a friend of mine and

29
00:02:14,140 --> 00:02:19,480
She can't stop listening to it too. And I'm getting getting a lot of good feedback on it

30
00:02:19,960 --> 00:02:24,760
there's a link in the description where you can stream it and download it and

31
00:02:25,340 --> 00:02:29,480
When I wanted to share it with this friend of mine, she was like, oh, you know, can I stream it anywhere?

32
00:02:29,760 --> 00:02:31,240
And I was just like, oh, no.

33
00:02:31,420 --> 00:02:33,280
And I was like, and send them all the MP3s.

34
00:02:33,280 --> 00:02:34,020
And I'm like, wait a minute.

35
00:02:34,640 --> 00:02:38,720
Let me just ask Codex to build me a beautiful streaming web page.

36
00:02:38,980 --> 00:02:40,260
And, I mean, it's really nice.

37
00:02:41,340 --> 00:02:42,180
It really is.

38
00:02:42,680 --> 00:02:43,560
And it works great.

39
00:02:43,560 --> 00:02:43,860
So cool.

40
00:02:44,080 --> 00:02:46,360
So, yeah, just ask for it, and it's done.

41
00:02:46,960 --> 00:02:49,900
There was a track that came out of this called Glass Animal.

42
00:02:50,300 --> 00:02:55,660
So far Mark my wife and my friend Kristen said that's their favorite track

43
00:02:57,060 --> 00:02:59,400
Yeah, so so there must be something to it

44
00:02:59,650 --> 00:03:04,740
But anyway, then the name when I saw the name I was like, oh that's the perfect name for this like fictitious band

45
00:03:04,890 --> 00:03:06,560
You know, I'm not trying to fool anybody or anything

46
00:03:07,500 --> 00:03:10,840
But then I learned that there is a real band called glass animals

47
00:03:12,100 --> 00:03:14,320
Yeah, and they actually had some pretty big success

48
00:03:14,550 --> 00:03:19,440
But you know, they have an S at the end of their name and that's like totally different. So doesn't matter

49
00:03:21,420 --> 00:03:29,160
But anyway, so RAINA wrote all the lyrics and I ran a lot of iterations of these songs.

50
00:03:29,660 --> 00:03:32,420
And very few of them were bad.

51
00:03:33,000 --> 00:03:33,160
Yeah.

52
00:03:33,360 --> 00:03:36,720
But it was just getting stuff that really appealed to me.

53
00:03:37,300 --> 00:03:41,820
I did share it with Mark and he did point out that many of the songs are in the same key.

54
00:03:42,120 --> 00:03:45,520
And that's probably the biggest fault on the album.

55
00:03:45,940 --> 00:03:46,260
Yeah.

56
00:03:46,940 --> 00:03:51,840
and a lot of them started out with, you know, like a vocal, like a river ambient vocal, like,

57
00:03:51,980 --> 00:03:59,080
you know, don't, I'm not even going to attempt it, but, but I did trim some of those out. Like

58
00:03:59,140 --> 00:04:03,100
I download the stem, killed the vocal just at the beginning. Cause there was too much of the,

59
00:04:03,240 --> 00:04:06,700
they all, too many started the same way, but I didn't, didn't even realize that there were so

60
00:04:06,760 --> 00:04:11,000
many were in the same key. I might pitch shift. I'm glad you did that. Yeah. What about, what

61
00:04:11,060 --> 00:04:15,980
about tempo tempo? Are they generally in the same ballpark too, or what? Yeah, pretty much. Yeah.

62
00:04:16,060 --> 00:04:22,060
Yeah. So, like, Glass Animal, that song is quite different from a lot of the other ones.

63
00:04:22,170 --> 00:04:24,820
And, you know, I tried to vary the prompt along the way.

64
00:04:25,000 --> 00:04:27,540
You know, I didn't want the whole album to be exactly the same.

65
00:04:27,610 --> 00:04:32,500
You know, I wanted a mellow song. I wanted a little more, you know, chunky guitar kind of song.

66
00:04:33,130 --> 00:04:33,260
Sure.

67
00:04:34,380 --> 00:04:37,380
But anyway, I probably could have done a little better job of that.

68
00:04:37,480 --> 00:04:41,480
But anyway, for me, I can't stop listening to it. I love it.

69
00:04:41,560 --> 00:04:45,960
It's 10 songs, it's 47 minutes, which is like the ideal length for an album for me.

70
00:04:46,540 --> 00:04:52,500
And you can do another one, you know, down the road, and all of these AIs are still babies.

71
00:04:52,800 --> 00:04:56,020
Yes, this is the worst it will ever be, as we say all the time on this show.

72
00:04:57,100 --> 00:05:02,800
And when it was all done, then I decided to, I was just like, I'm just going to give it a little bit of analog love.

73
00:05:03,060 --> 00:05:08,500
And I ran it through a piece of gear I've owned for 28 years, made by a company called Manly.

74
00:05:08,640 --> 00:05:09,120
It's a compressor.

75
00:05:09,340 --> 00:05:12,080
It's full of tubes and transformers.

76
00:05:12,230 --> 00:05:15,040
And it's like wrapping a warm blanket around it.

77
00:05:15,640 --> 00:05:20,800
And anyway, I'll play a couple of clips, and then we'll get into what Mark's doing.

78
00:05:21,200 --> 00:05:23,140
This is my favorite track on the album.

79
00:05:23,170 --> 00:05:24,500
It's called Little Gods With Matches.

80
00:05:24,720 --> 00:05:28,620
And this was one of the songs that just got me really excited about making a whole bunch of these.

81
00:05:29,000 --> 00:05:30,420
I'll just play like a verse and a chorus.

82
00:05:30,500 --> 00:05:30,520
Cheers.

83
00:06:00,500 --> 00:06:04,240
♪ Walk from somewhere we never should have been ♪

84
00:06:04,240 --> 00:06:07,040
♪ Put it in a paper house ♪

85
00:06:07,300 --> 00:06:09,760
♪ And call the whole thing wind ♪

86
00:06:10,840 --> 00:06:13,280
♪ We are bright and reckless ♪

87
00:06:13,280 --> 00:06:15,820
♪ Holy fools with dirty hands ♪

88
00:06:16,060 --> 00:06:18,240
♪ Drawing maps to heaven ♪

89
00:06:18,360 --> 00:06:20,640
♪ And the dust where cities stand ♪

90
00:06:21,320 --> 00:06:26,639
♪ We are little guys with matches ♪

91
00:06:27,140 --> 00:06:31,300
♪ Dancing near the gasoline ♪

92
00:06:31,540 --> 00:06:36,280
♪ Building temples out of thunder ♪

93
00:06:36,860 --> 00:06:40,980
♪ Calling it a dream ♪

94
00:06:42,020 --> 00:06:44,980
♪ We keep reaching for the heavens ♪

95
00:06:44,980 --> 00:06:47,580
♪ With our pockets full of sparks ♪

96
00:06:47,580 --> 00:06:49,680
♪ Lighting fires just to prove ♪

97
00:06:49,720 --> 00:06:53,599
♪ We're not afraid of the dark ♪

98
00:06:57,060 --> 00:06:57,100
Crazy.

99
00:06:57,680 --> 00:06:58,000
Yeah.

100
00:06:58,190 --> 00:06:58,480
Take it.

101
00:06:59,180 --> 00:07:00,060
Just crazy.

102
00:07:00,919 --> 00:07:01,800
Sounds great.

103
00:07:01,980 --> 00:07:03,120
And it sounds great, yeah.

104
00:07:03,820 --> 00:07:09,720
And Mark, you had said to me in one of your texts that Suno seems to do a really great

105
00:07:09,860 --> 00:07:12,400
job of, you know, very musical uses of effects.

106
00:07:13,220 --> 00:07:13,480
Mm-hmm.

107
00:07:14,040 --> 00:07:14,260
Yeah.

108
00:07:14,440 --> 00:07:15,500
They're not gimmicky.

109
00:07:16,720 --> 00:07:21,920
It's just as conversant with, you know, with the sounds, with the effects, as it is with

110
00:07:22,300 --> 00:07:25,200
melody and with harmony and with rhythm and all the other aspects.

111
00:07:26,380 --> 00:07:31,240
I hadn't even thought about that so much before, but like when it brings when it brings the reverb

112
00:07:32,060 --> 00:07:35,160
Way down just to make something really personal for a second

113
00:07:35,510 --> 00:07:41,040
And then you know and then it blasts it out pitch for the chorus and it takes off that that's just one of a hundred

114
00:07:41,540 --> 00:07:47,820
You know million things that it's that it's thinking about every second of the process. How how how?

115
00:07:50,900 --> 00:07:56,360
It's it's mind-blowing and I know this one thing now like when I don't know if it's just on that on that

116
00:07:56,360 --> 00:07:59,780
on the phone app, but it is super fast now.

117
00:08:00,680 --> 00:08:01,040
Really?

118
00:08:01,280 --> 00:08:01,480
Yeah.

119
00:08:01,680 --> 00:08:04,240
I mean, it's like in 10 seconds, you can start playing it.

120
00:08:04,640 --> 00:08:05,140
It's crazy.

121
00:08:05,660 --> 00:08:05,840
All right.

122
00:08:06,140 --> 00:08:07,420
I'll give you one more little taste.

123
00:08:07,480 --> 00:08:10,580
This is the song Glass Animal that we were talking about before.

124
00:08:20,380 --> 00:08:24,560
♪ I'll understand in window light ♪

125
00:08:25,450 --> 00:08:28,420
♪ And never leave a trace ♪

126
00:08:29,020 --> 00:08:34,060
♪ A little smile, a steady hand ♪

127
00:08:34,630 --> 00:08:37,840
♪ A careful kind of grace ♪

128
00:08:37,680 --> 00:08:45,040
But under all that silver Under all that shine

129
00:08:46,900 --> 00:08:55,140
There's something pacing softly At the edges of my mind

130
00:08:56,520 --> 00:09:03,760
Don't ask me to be easy Don't ask me to be small

131
00:09:05,680 --> 00:09:08,240
I've been still as winter glass

132
00:09:10,550 --> 00:09:12,800
But I can hear the thaw

133
00:09:16,200 --> 00:09:18,100
I'm a glass animal

134
00:09:20,440 --> 00:09:22,800
Breathing on the pain

135
00:09:25,680 --> 00:09:27,800
Every little fracture

136
00:09:30,240 --> 00:09:34,680
Learns to love the rain

137
00:09:34,680 --> 00:09:37,340
I can look like silence.

138
00:09:39,160 --> 00:09:42,080
I can move like a flame.

139
00:09:43,920 --> 00:09:50,900
But something in the mirror still refuses to be named.

140
00:09:54,400 --> 00:09:56,160
The lyrics, everything.

141
00:09:56,730 --> 00:09:57,280
So good.

142
00:09:57,840 --> 00:09:57,980
Yeah.

143
00:09:59,400 --> 00:10:03,280
I think they owe Billie Eilish some royalties, though.

144
00:10:03,380 --> 00:10:09,340
that kind of vocal delivery right up on the mic that's totally her jam isn't it lord or lord

145
00:10:10,339 --> 00:10:17,000
yeah i i was yeah i think lord is uh is probably probably the closer comparison but yeah yeah that

146
00:10:17,100 --> 00:10:25,960
song in particular so good though yeah and it it does so much with so little like melodically

147
00:10:26,100 --> 00:10:32,780
harmonically like I want to like criticize it for do it but you know for

148
00:10:32,880 --> 00:10:37,200
having so little going on there but it's like a master class in minimalism mm-hmm

149
00:10:37,460 --> 00:10:44,160
the the most fundamental concept in art is probably tension and release so

150
00:10:44,400 --> 00:10:48,660
you're on a thing and then you step away from the thing there's a little and then

151
00:10:48,660 --> 00:10:54,719
you come back to the thing and most of this song is on the main note dome and

152
00:10:54,720 --> 00:10:59,740
and then it goes up a half step and it comes back and you put some drums behind that and it's like

153
00:10:59,740 --> 00:11:06,900
you have this tribal like scary sounding like you know the tension is real and there's

154
00:11:06,900 --> 00:11:12,400
just two notes going back and forth i mean it's the jaws theme in a way yeah oh yeah totally

155
00:11:13,280 --> 00:11:19,300
and like to that point how many times has that interval been done in music right and it's just

156
00:11:19,200 --> 00:11:23,820
so obvious it's just right there and you know yeah pluck it out of the ether and just play

157
00:11:26,440 --> 00:11:30,180
right and just kind of that's the whole song basically right

158
00:11:31,070 --> 00:11:36,760
yeah yeah that's it that's the essence of it yeah i mean it's amazing i was gonna say there's only

159
00:11:36,840 --> 00:11:41,120
12 notes in the scale but you know on gene to pot train is uh they're using a few more

160
00:11:42,880 --> 00:11:47,999
microtronal notes between the notes yeah yeah they needed more just not pleasing some people yeah

161
00:11:48,000 --> 00:11:52,160
So Mark, is it over? Is it over?

162
00:11:53,860 --> 00:12:01,700
You know, I have been so ardently on one end and then the other end of the spectrum

163
00:12:02,300 --> 00:12:05,280
as I've been trying to wrap my head around all of this stuff.

164
00:12:06,660 --> 00:12:08,760
No, it's definitely not over.

165
00:12:09,780 --> 00:12:13,440
I mean, when it comes right down to it, why are we making art in the first place?

166
00:12:14,060 --> 00:12:16,420
We have to. It's just something that we do.

167
00:12:16,840 --> 00:12:22,960
birds gotta sing we've been making you know music since you know as long as we've had civilization

168
00:12:23,600 --> 00:12:28,600
for sure and probably even proto civilization or something it's always been there so this is

169
00:12:28,700 --> 00:12:36,180
something that we have to do now will we click a button and have the music made for us and does

170
00:12:36,300 --> 00:12:41,480
that completely cheapen everything uh yeah probably but i mean i'm still gonna always

171
00:12:42,200 --> 00:12:49,160
really relish going to a club and seeing a band play you know seeing other people actually express

172
00:12:49,440 --> 00:12:57,620
themselves i keep thinking about the creator side versus the consumer side and on the consumer side

173
00:12:58,980 --> 00:13:05,380
i might not be able to tell that it came from software and on the consumer side if i look in

174
00:13:05,380 --> 00:13:10,759
the clouds and i see a face and that face means something to me well there you go art experience

175
00:13:10,760 --> 00:13:16,720
that happened and it doesn't matter what the provenance of the art was on the creator side

176
00:13:17,940 --> 00:13:26,360
as as a composer i'm the only person who knows for sure how much of me is in the art that i made

177
00:13:26,620 --> 00:13:33,820
so i have to be satisfied on my end you know with with whatever tools i used whatever process i made

178
00:13:33,920 --> 00:13:38,899
and whether it met my own standards and stuff i mean i've i've been i've been wrestling with

179
00:13:38,900 --> 00:13:43,280
music and trying to make it do the thing I want it to do for 50 fucking years.

180
00:13:44,010 --> 00:13:49,920
So I will not have anyone judge my use of any tool, no matter what it does, because

181
00:13:50,030 --> 00:13:54,860
I could also write that shit out on in notation and arrange it for a string quartet if I really

182
00:13:54,940 --> 00:13:55,260
had to.

183
00:13:56,420 --> 00:14:01,580
So I'm so I'm, you know, I hate the idea of people like just using text prompts and

184
00:14:01,680 --> 00:14:02,900
saying, I'm a songwriter now.

185
00:14:03,460 --> 00:14:03,520
Yeah.

186
00:14:04,320 --> 00:14:06,080
But there's, you know, but there's a range.

187
00:14:06,360 --> 00:14:13,940
So it's, you know, it's not like a binary question of yes or no, but I think music is not dead and musicality, human musicality is not dead.

188
00:14:14,860 --> 00:14:27,620
Nearly every musician that you run into could listen to anything that Suno produces and take away something useful from it and be influenced in a positive way, not a negative way, a positive way.

189
00:14:28,340 --> 00:14:29,660
So, yeah, I don't think it's over.

190
00:14:29,960 --> 00:14:33,320
I love that you clarified the consumer and the creator side.

191
00:14:33,800 --> 00:14:38,820
And the thing that you touched on when you said you're making music for 50 years and

192
00:14:39,040 --> 00:14:44,300
the joy that is in the friction of it, the joy that's in the productive struggle of making

193
00:14:44,500 --> 00:14:47,060
music and finally that push and pull, right?

194
00:14:47,500 --> 00:14:50,220
Of you're pushing the music, the music's pushing you back.

195
00:14:50,460 --> 00:14:51,280
It's informing you.

196
00:14:51,310 --> 00:14:53,220
It's saying, no, maybe go here, right?

197
00:14:53,700 --> 00:14:54,880
That's the magic, right?

198
00:14:54,950 --> 00:14:56,640
When you're in the midst of it, the flow.

199
00:14:57,600 --> 00:15:02,600
I don't feel good about giving a text prompt to software and having it make something.

200
00:15:03,260 --> 00:15:05,660
Or I could say, Chris, I was thinking about a song that goes like this.

201
00:15:05,820 --> 00:15:06,940
Why don't you write that song for me?

202
00:15:07,060 --> 00:15:08,320
I didn't write that song.

203
00:15:08,900 --> 00:15:10,000
And there's no achievement there.

204
00:15:10,360 --> 00:15:14,480
And, you know, how do I feel about the art that I make?

205
00:15:14,720 --> 00:15:17,580
That's the question that I have to answer and I have to wrestle with myself.

206
00:15:18,630 --> 00:15:18,760
Yeah.

207
00:15:18,910 --> 00:15:24,800
I mean, like when I'm listening to this Glass Animal album that I, quote unquote, made,

208
00:15:25,580 --> 00:15:26,620
it's just the end product.

209
00:15:26,820 --> 00:15:28,600
It's just the music that is hitting me.

210
00:15:28,700 --> 00:15:34,600
and I like it because I ran a bunch of these until I got the song that I wanted.

211
00:15:35,050 --> 00:15:38,320
So yeah, the end product makes me happy,

212
00:15:38,510 --> 00:15:44,740
but it definitely lacks the human connection with bands that I get to know

213
00:15:45,390 --> 00:15:48,000
or musicians that I get to know, not personally,

214
00:15:48,270 --> 00:15:52,400
but just through listening to their music and knowing they made it.

215
00:15:52,520 --> 00:15:54,380
That is important to me.

216
00:15:55,260 --> 00:15:57,240
So for me, they're two different things,

217
00:15:57,340 --> 00:16:04,240
But, you know, bottom line, you know, this AI generated music that I curated, I'm enjoying listening to it.

218
00:16:04,340 --> 00:16:05,480
It's a good song.

219
00:16:05,570 --> 00:16:06,040
It's a good song.

220
00:16:06,240 --> 00:16:06,600
It's a good song.

221
00:16:07,300 --> 00:16:07,660
Absolutely.

222
00:16:07,970 --> 00:16:09,820
I level what I've heard from this record.

223
00:16:10,720 --> 00:16:10,960
Record.

224
00:16:12,180 --> 00:16:12,540
Record.

225
00:16:12,980 --> 00:16:13,940
Here we are, old guys.

226
00:16:13,940 --> 00:16:14,940
I stuck on records, too.

227
00:16:17,160 --> 00:16:26,740
But the other thing, and I know it's the genre that you've prompted here, too, but these things are just so glossy and perfect and, right, and manicured.

228
00:16:26,760 --> 00:16:32,880
and they lack the nuance and imperfection and mileage, right?

229
00:16:33,160 --> 00:16:34,220
Yeah, I know what you mean.

230
00:16:34,480 --> 00:16:37,900
You know, like, yeah, I love Keith Richards over Joe Satriani.

231
00:16:38,100 --> 00:16:39,460
You know, that's me, you know.

232
00:16:39,600 --> 00:16:43,280
But I've been saying to somebody, I felt like, you know, the music,

233
00:16:43,460 --> 00:16:44,200
it's kind of like Doritos.

234
00:16:44,680 --> 00:16:47,040
They're like, you know, engineered to, you know,

235
00:16:48,120 --> 00:16:49,660
just grab another one out of the bag.

236
00:16:50,060 --> 00:16:53,300
Trifecta of salt, sugar, and fat, right?

237
00:16:53,440 --> 00:16:53,520
Yeah.

238
00:16:54,220 --> 00:16:55,080
But it's okay.

239
00:16:55,240 --> 00:16:56,040
These are Doritos.

240
00:16:56,160 --> 00:16:56,480
Yes

241
00:16:58,880 --> 00:17:05,060
Yeah, I've been saying they're they're Burger King, you know, if it was cuisine these songs are from Burger King

242
00:17:05,470 --> 00:17:07,740
Sure, I can chow down on some Burger King

243
00:17:08,819 --> 00:17:13,100
Tasty, let's not pretend that it's a might might even be White Castle. Yeah

244
00:17:14,220 --> 00:17:15,260
These are rat burgers

245
00:17:16,260 --> 00:17:21,640
Burgers, I wonder then oh wow. So mark you've been doing this a long time. Obviously teaching piano

246
00:17:22,319 --> 00:17:24,479
public school teacher for 13 years

247
00:17:24,920 --> 00:17:29,500
What is this new devilry showing up as in your life?

248
00:17:29,520 --> 00:17:33,260
Is it changing the way students learn music?

249
00:17:36,179 --> 00:17:37,260
Not yet.

250
00:17:37,960 --> 00:17:45,500
But, I mean, for starters, you only have a certain percentage of kids who are driven specifically in a musical direction.

251
00:17:46,540 --> 00:17:55,140
And those that are, well, they're really the ones that I worry the most about because they are going to take the path of least resistance to sounding as good as they can.

252
00:17:55,660 --> 00:17:56,440
And why wouldn't they?

253
00:17:56,520 --> 00:17:58,140
I mean, that makes a lot of sense.

254
00:17:59,200 --> 00:18:09,800
But I have already been telling them, you build yourself through the struggle, you know, of trying to get from A to B, to get to a finished thing.

255
00:18:10,300 --> 00:18:17,740
And by the way, your first 20 or maybe 50 songs are kind of going to suck, at least compared to your 200th song.

256
00:18:17,900 --> 00:18:18,600
And it's a process.

257
00:18:19,240 --> 00:18:20,900
And this is sort of the arc of your life.

258
00:18:21,000 --> 00:18:21,860
And you should embrace it.

259
00:18:21,880 --> 00:18:23,240
And it's a wonderful, beautiful thing.

260
00:18:24,400 --> 00:18:29,120
Most of my students, they're just happy enough to just kind of plonk out a few notes on the keyboard.

261
00:18:29,500 --> 00:18:34,520
And they want to get a passing grade at least or get an A in music.

262
00:18:35,080 --> 00:18:36,580
do whatever they're supposed to do,

263
00:18:36,640 --> 00:18:37,900
just like they're trying to get through math

264
00:18:38,040 --> 00:18:39,560
and they're trying to get through drama

265
00:18:39,820 --> 00:18:40,740
or whatever else they have.

266
00:18:41,340 --> 00:18:45,620
My attention is really on the likely music makers

267
00:18:46,679 --> 00:18:48,420
and try to give them perspective

268
00:18:48,760 --> 00:18:49,880
and make sure that they understand

269
00:18:50,160 --> 00:18:53,120
that you have to work to build yourself as a person.

270
00:18:55,220 --> 00:18:56,240
I love that you said that.

271
00:18:56,280 --> 00:18:58,060
And I kind of come from a similar background

272
00:18:58,200 --> 00:19:01,300
in that I'm an elective teacher all my life, specialist.

273
00:19:02,000 --> 00:19:05,060
And that for me, when I look back over my career

274
00:19:05,060 --> 00:19:11,380
in stem i feel like at any given moment in a cohort there's maybe 10 of that room that's like

275
00:19:11,840 --> 00:19:16,800
yeah i love this thing and i'm just i'm going to make this my life would you feel that is that a

276
00:19:16,860 --> 00:19:20,860
decent percentage for people in a music room at any given time there's 10 in that room they're

277
00:19:20,860 --> 00:19:26,480
like yeah i'm a believer i think there are 10 some of them might be hiding because they're

278
00:19:26,580 --> 00:19:31,019
because they're in middle school and they're embarrassed about everything but yeah i think

279
00:19:30,900 --> 00:19:32,160
I think 10% is probably a fair number.

280
00:19:32,880 --> 00:19:34,200
God, but middle school too, man,

281
00:19:34,360 --> 00:19:37,340
as problematic as it can be, especially seventh grade,

282
00:19:38,380 --> 00:19:40,500
middle school is just so fertile and ripe

283
00:19:40,580 --> 00:19:42,160
for just blowing minds, isn't it?

284
00:19:42,980 --> 00:19:46,660
Yeah, yeah, and for altering their course a little bit.

285
00:19:46,940 --> 00:19:47,180
Yeah.

286
00:19:47,380 --> 00:19:50,680
So yeah, I take a lot of pleasure in,

287
00:19:51,720 --> 00:19:52,760
and I take it very seriously,

288
00:19:53,420 --> 00:19:55,540
'cause I could send them down a dark path

289
00:19:56,520 --> 00:19:57,640
if I was so inclined.

290
00:19:58,140 --> 00:19:59,740
Here's the devil's interval, kids.

291
00:20:01,120 --> 00:20:02,140
The devil's interval.

292
00:20:03,060 --> 00:20:05,080
We're going to play this, everything in D minor.

293
00:20:05,440 --> 00:20:05,780
That's right.

294
00:20:05,890 --> 00:20:07,120
The saddest of all keys.

295
00:20:10,260 --> 00:20:11,960
You do show them a spinal tap, I hope.

296
00:20:13,840 --> 00:20:14,700
Are you kidding?

297
00:20:14,830 --> 00:20:15,580
I can't show them that.

298
00:20:16,780 --> 00:20:20,000
I can encourage them to go seek it out on their own.

299
00:20:20,040 --> 00:20:21,340
It is public school in New Jersey.

300
00:20:21,410 --> 00:20:22,460
You probably get away with a lot.

301
00:20:22,590 --> 00:20:22,700
Yeah.

302
00:20:23,380 --> 00:20:23,900
Well, yeah.

303
00:20:23,910 --> 00:20:26,160
I mean, the language that comes out of them all day long.

304
00:20:26,600 --> 00:20:28,040
I wouldn't surprise them with anything.

305
00:20:28,350 --> 00:20:28,660
I bet.

306
00:20:30,240 --> 00:20:34,340
So I love this question too that I have queued up for you.

307
00:20:34,560 --> 00:20:37,540
So a student hands in some work to you.

308
00:20:37,540 --> 00:20:37,820
I don't know.

309
00:20:37,920 --> 00:20:43,980
You give them something to compose and they've clearly leveraged an AI platform in the end product.

310
00:20:44,420 --> 00:20:48,140
And you, having the discerning ear that you have, can't tell.

311
00:20:48,920 --> 00:20:54,740
It's so strong or it's maybe it kind of mimics their voice enough that you don't know.

312
00:20:55,320 --> 00:20:56,140
How do you handle that?

313
00:20:57,680 --> 00:20:59,840
They have to be able to articulate what they did.

314
00:20:59,980 --> 00:21:09,800
They have to be able to verbally articulate where they're coming from, and they maybe don't have to explain everything in detail from like a music theory point of view.

315
00:21:10,250 --> 00:21:16,980
But I can certainly tell when you're talking to me, like what the provenance of the ideas is.

316
00:21:17,960 --> 00:21:21,640
And I can hear in that conversation how much this means to you.

317
00:21:21,770 --> 00:21:25,080
And honestly, if it means a lot to you, you get an A.

318
00:21:25,400 --> 00:21:27,260
I don't care how good it is.

319
00:21:27,400 --> 00:21:29,860
Because I want to see the effort and I want to see the development.

320
00:21:30,860 --> 00:21:36,540
I've often said that, because I have some of the shittiest students, but I also have some of the best.

321
00:21:37,280 --> 00:21:45,180
And it is fully as meaningful to me to get a kid who's probably an 8 and becomes a 9 or a 10.

322
00:21:46,240 --> 00:21:50,360
That's no more meaningful than a kid coming in who's a 2 and leaves as a 4.

323
00:21:51,150 --> 00:21:51,260
Yeah.

324
00:21:51,900 --> 00:21:53,120
It's all important.

325
00:21:54,520 --> 00:22:01,420
I love that because a couple of things, you're talking about process versus product, right?

326
00:22:01,600 --> 00:22:05,580
Our education system has focused on product forever, right?

327
00:22:06,000 --> 00:22:07,900
You got the grade, you got the right answer.

328
00:22:08,050 --> 00:22:10,040
They haven't spoken about process.

329
00:22:10,230 --> 00:22:12,480
They haven't talked about design process.

330
00:22:12,710 --> 00:22:16,860
And I love that you also said it comes down to conversation and anecdotal observation,

331
00:22:17,630 --> 00:22:17,740
right?

332
00:22:17,870 --> 00:22:22,200
And this kind of very humanistic interaction at the end of the day.

333
00:22:22,900 --> 00:22:23,380
We're people.

334
00:22:23,860 --> 00:22:25,580
We live, we should try to be happy.

335
00:22:25,800 --> 00:22:28,260
We need to understand how to do that.

336
00:22:28,580 --> 00:22:31,700
And, you know, art is such an essential component of that.

337
00:22:32,040 --> 00:22:32,720
And so is communication.

338
00:22:33,420 --> 00:22:34,580
So let's work on those.

339
00:22:35,040 --> 00:22:37,620
♪ We're all sensitive people ♪

340
00:22:38,040 --> 00:22:40,560
You wanted me to sing, so I sang you some Marvin Gaye.

341
00:22:42,640 --> 00:22:44,440
I did, I said, "What are you gonna sing on the show?"

342
00:22:44,500 --> 00:22:45,000
There it is.

343
00:22:45,220 --> 00:22:46,500
Okay, drop that into Suno

344
00:22:46,580 --> 00:22:47,540
and let's see what it comes up with.

345
00:22:47,720 --> 00:22:48,920
There you go, great, thank you.

346
00:22:49,040 --> 00:22:49,420
There it is.

347
00:22:49,480 --> 00:22:51,100
It is seven o'clock in the morning over here.

348
00:22:51,260 --> 00:22:54,220
My vocal folds haven't warmed up quite yet.

349
00:22:56,039 --> 00:23:01,720
So, Mark, you're working on your album with your progressive rock band, X-Atlas.

350
00:23:02,440 --> 00:23:03,000
Yeah.

351
00:23:04,440 --> 00:23:09,900
And that genre is built on musicianship, odd time signatures, and technical performance.

352
00:23:10,540 --> 00:23:16,240
We just watched a, Mark came over to watch a Stephen Wilson concert video, which was all about that.

353
00:23:17,300 --> 00:23:17,720
Oh, nice.

354
00:23:17,720 --> 00:23:20,700
One of the best concert videos I've ever seen.

355
00:23:21,420 --> 00:23:21,960
Holy shit.

356
00:23:22,200 --> 00:23:22,660
Really?

357
00:23:22,660 --> 00:23:23,140
Holy shit.

358
00:23:23,160 --> 00:23:23,980
Really something.

359
00:23:24,960 --> 00:23:26,940
That's the guy from Porcupine Tree, right?

360
00:23:27,140 --> 00:23:27,260
Correct.

361
00:23:27,500 --> 00:23:27,700
Yeah.

362
00:23:27,860 --> 00:23:28,160
Yeah.

363
00:23:28,540 --> 00:23:30,500
I'll put a link in the description for that.

364
00:23:31,040 --> 00:23:37,520
So is there any part of that process where AI tools have snuck in, even in demoing or

365
00:23:37,740 --> 00:23:40,600
arranging, or is it deliberately human-only zone?

366
00:23:42,360 --> 00:23:46,660
Oh, very much deliberately human-only zone until now.

367
00:23:47,340 --> 00:23:59,800
Now that I started fucking around with these things a little bit, I've said for sure I don't want AI to ever write a lyric in my song.

368
00:24:00,040 --> 00:24:00,960
It's my song.

369
00:24:01,140 --> 00:24:02,100
It's my story.

370
00:24:02,690 --> 00:24:06,220
But then, you know, I started playing around and I threw some ideas into Suno.

371
00:24:06,290 --> 00:24:07,100
And I'll tell you what.

372
00:24:07,270 --> 00:24:09,000
You dipped your hand in the bag of Doritos.

373
00:24:10,240 --> 00:24:11,440
That's the thing, man.

374
00:24:11,530 --> 00:24:13,300
There were just a couple little phrases.

375
00:24:13,880 --> 00:24:17,700
And I was saying to you how Suno's lyrics are just one metaphor after another.

376
00:24:17,940 --> 00:24:20,380
That seems to be the way that it approaches songwriting.

377
00:24:20,740 --> 00:24:23,320
Oh, this is going to be more poetic because it's a metaphor.

378
00:24:23,880 --> 00:24:29,100
But there were a couple that were so unexpected and so illuminating.

379
00:24:29,920 --> 00:24:31,320
Like, oh, shit, I cannot hear that.

380
00:24:31,340 --> 00:24:32,440
I think I might have to use that.

381
00:24:33,500 --> 00:24:38,960
And now I have this riff that I wrote to show to the band.

382
00:24:39,020 --> 00:24:40,340
It could become a section of something.

383
00:24:41,000 --> 00:24:43,440
And it's rhythmically really hard.

384
00:24:43,600 --> 00:24:48,180
and harmonically, for those who are comfortable with these kinds of words,

385
00:24:49,780 --> 00:24:51,060
it's really challenging.

386
00:24:51,400 --> 00:24:53,700
There's like two chords that don't really go together,

387
00:24:54,680 --> 00:24:57,800
and I thought, well, it would be really cool to see what Suno,

388
00:24:58,120 --> 00:25:03,320
with its endless chops, could do over something really difficult like this.

389
00:25:03,440 --> 00:25:08,780
So I threw that all together, and although the lyrics are stupid

390
00:25:08,860 --> 00:25:11,400
and I have no interest in what it came up with at all,

391
00:25:12,180 --> 00:25:15,600
The melody, the vocal melody is great.

392
00:25:16,060 --> 00:25:16,180
Wow.

393
00:25:16,180 --> 00:25:20,700
And it's not unlike melodies that I've written myself for my other songs.

394
00:25:20,860 --> 00:25:27,700
And maybe to some degree, the rough that I put into it, you know, is responsible for that.

395
00:25:28,300 --> 00:25:35,820
And I don't know if Suno kind of starts building a model of me based on stuff that I put into it.

396
00:25:35,820 --> 00:25:38,680
I was actually wondering if RAINA would be able to speak to that a little bit.

397
00:25:39,240 --> 00:25:44,700
But yeah, so I have started using it now because I want to know, well, what would it do with this really difficult thing?

398
00:25:45,290 --> 00:25:46,220
All right, maybe I can do that.

399
00:25:46,220 --> 00:25:47,340
And I can take it or leave it.

400
00:25:47,950 --> 00:25:53,020
But I like having, you know, more options kind of spelled out for me that I could choose from.

401
00:25:53,650 --> 00:25:58,160
You know, it gives you a couple different versions of a song once you give it a prompt and you let it go.

402
00:25:58,850 --> 00:26:04,040
And I'm thinking when I'm at the piano, I didn't invent any of the chords that I'm using.

403
00:26:04,760 --> 00:26:11,760
I'm choosing them and I'm trying to create a kind of musical context, but I didn't invent any of these chords and I didn't invent any of these scales.

404
00:26:12,110 --> 00:26:17,940
And probably every combination of notes that I'm coming up with has already been used a bajillion times.

405
00:26:18,670 --> 00:26:30,840
So it's just a kind of sped up hyper version of that when I get something from Suno and I draw upon it and I use that as an influence for the writing process.

406
00:26:31,300 --> 00:26:32,300
I think that's not going to change.

407
00:26:32,880 --> 00:26:33,840
I'd still like to be able to sit down.

408
00:26:33,960 --> 00:26:36,560
I still can sit down at the piano and write something

409
00:26:36,620 --> 00:26:38,780
and an hour later, hopefully I have something decent for you.

410
00:26:39,700 --> 00:26:42,080
But this is amazing, you know,

411
00:26:42,160 --> 00:26:44,680
for someone who has spent this much time

412
00:26:44,920 --> 00:26:46,440
learning about music to be able to take,

413
00:26:46,540 --> 00:26:48,560
really take a shortcut like this.

414
00:26:48,600 --> 00:26:49,200
I think it's awesome.

415
00:26:49,980 --> 00:26:53,180
Yeah, and coming back to process defense, right?

416
00:26:53,400 --> 00:26:54,940
It's a dialogue, it's a conversation.

417
00:26:55,440 --> 00:26:57,200
You're not just like wholesale surrendering

418
00:26:57,320 --> 00:26:58,960
the composition to this thing.

419
00:26:59,340 --> 00:27:01,120
You're like, oh, that's pretty cool.

420
00:27:01,260 --> 00:27:02,020
I like that idea.

421
00:27:02,200 --> 00:27:03,700
I would not have considered that.

422
00:27:04,120 --> 00:27:07,900
And the other thing I always think about when we talk about these, you know,

423
00:27:08,000 --> 00:27:12,120
LLMs as cheating machines or however you want to frame them in educational terminology,

424
00:27:12,960 --> 00:27:14,240
I think about it all the time is,

425
00:27:14,680 --> 00:27:17,920
aren't we collaborating with the largest group of humans ever assembled?

426
00:27:18,540 --> 00:27:19,480
That's really what we're doing here.

427
00:27:19,920 --> 00:27:20,680
I think so.

428
00:27:21,300 --> 00:27:21,380
Right?

429
00:27:21,480 --> 00:27:24,620
We've scraped humanity into this black box and we're saying,

430
00:27:24,660 --> 00:27:25,580
hey, I've got this song.

431
00:27:26,140 --> 00:27:27,280
Is anybody in this group here?

432
00:27:27,360 --> 00:27:28,500
Can we crowdsource this solution?

433
00:27:28,640 --> 00:27:32,240
I'm stuck on this bridge or I'm stuck on this line.

434
00:27:32,680 --> 00:27:33,480
Somebody have an idea.

435
00:27:33,950 --> 00:27:36,920
And that's up to you as the curator to say, oh, that's the one.

436
00:27:37,240 --> 00:27:37,660
I like that.

437
00:27:37,820 --> 00:27:39,580
Thank you, anonymous black box.

438
00:27:39,880 --> 00:27:40,740
Yes, that works.

439
00:27:40,910 --> 00:27:45,900
And it's not, you know, it's not just it's storing all of this data.

440
00:27:45,950 --> 00:27:48,700
I mean, it works similarly to how the human brain does.

441
00:27:48,880 --> 00:27:49,760
It's a neural network.

442
00:27:49,950 --> 00:27:53,240
And so it's a just it's a giant brain.

443
00:27:55,760 --> 00:27:56,680
It's the borg.

444
00:27:56,920 --> 00:27:58,340
It's a gray area for sure.

445
00:27:58,600 --> 00:28:00,620
but so is the brain.

446
00:28:01,400 --> 00:28:01,980
That's right.

447
00:28:02,200 --> 00:28:02,780
That's a great,

448
00:28:02,880 --> 00:28:03,500
great matter.

449
00:28:06,520 --> 00:28:08,260
So what is X Atlas all about?

450
00:28:08,340 --> 00:28:11,820
How would you describe this particular flavor of progressive rock?

451
00:28:12,760 --> 00:28:13,460
A lot,

452
00:28:13,460 --> 00:28:16,040
a lot of modern prog rock is very,

453
00:28:16,580 --> 00:28:16,700
um,

454
00:28:17,440 --> 00:28:19,220
very like gymnastic guitar,

455
00:28:19,960 --> 00:28:21,520
arpeggios and scales and stuff like that.

456
00:28:22,100 --> 00:28:22,160
Um,

457
00:28:22,300 --> 00:28:23,580
but it's one thing that I,

458
00:28:23,760 --> 00:28:28,220
I love about progressive rock is that it's not about the singer.

459
00:28:29,220 --> 00:28:34,620
it's really about the art of the music sure and something that i always say to my students

460
00:28:35,360 --> 00:28:41,420
is we are not here to do easy things and i sure as hell embrace that in my own life and i know

461
00:28:41,520 --> 00:28:47,840
larry does and i know that you do um so when x atlas gets together in this room behind me to

462
00:28:48,040 --> 00:28:54,739
rehearse everything is like so intricate and so detailed and we're just like savoring the

463
00:28:54,740 --> 00:29:00,600
challenge of remembering the four or five ideas we just had and adding them to the 50 ideas that

464
00:29:00,620 --> 00:29:06,380
we had since we started playing this thing and going like we'll we'll play uh you know the same

465
00:29:06,700 --> 00:29:13,160
eight bar phrase for half an hour and try this and try that and something about that's just really

466
00:29:13,260 --> 00:29:18,800
soothing and satisfying to me and if at the end of six months and sometimes it takes me six months

467
00:29:18,860 --> 00:29:24,720
to write the song because i have so many other you know pots on the stove so to speak um but by

468
00:29:24,720 --> 00:29:28,500
the song is done, I feel really good about it. And I don't need anyone else to like it,

469
00:29:28,540 --> 00:29:35,420
but the four of us love the song. So it's a great process. Awesome. And I think you're going to

470
00:29:35,560 --> 00:29:41,500
delay any kind of cognitive decline in your life too, by engaging your brain, like in our 50s,

471
00:29:41,500 --> 00:29:46,160
the three of us, by playing that very ambitious kind of challenging stuff, right? I think that's

472
00:29:46,160 --> 00:29:52,400
the key to it. You're probably right. I mean, I have a madly twisted ADHD brain, you know,

473
00:29:52,800 --> 00:29:54,760
highs and lows, everything that that entails.

474
00:29:55,200 --> 00:29:57,160
So I think it won't get any worse.

475
00:29:58,360 --> 00:30:01,540
I still don't know what I'm supposed to do tomorrow half of the time.

476
00:30:01,660 --> 00:30:01,880
Right.

477
00:30:02,120 --> 00:30:03,900
And I still walk into the other room,

478
00:30:04,440 --> 00:30:05,920
and then by the time they get to the other room,

479
00:30:06,040 --> 00:30:08,260
I have no idea what the hell I was going in there for

480
00:30:08,460 --> 00:30:11,320
because something else just took over my brain.

481
00:30:13,039 --> 00:30:16,240
So do you have that heavy riff thing, LB?

482
00:30:16,540 --> 00:30:18,780
This is actually heavier than a lot of our stuff,

483
00:30:19,000 --> 00:30:22,600
but I was like, I think we need to go balls out for a moment here.

484
00:30:22,700 --> 00:30:24,060
So the first clip is what?

485
00:30:25,280 --> 00:30:28,320
The first clip is just a quick MIDI version,

486
00:30:28,890 --> 00:30:30,740
just in Logic, using Logic instruments.

487
00:30:31,230 --> 00:30:32,240
I played the bass in,

488
00:30:32,370 --> 00:30:36,280
I played the sort of heavy guitar synth-y sound,

489
00:30:36,740 --> 00:30:38,280
and I played the drums in.

490
00:30:38,940 --> 00:30:40,280
And so I set up the riff.

491
00:30:40,540 --> 00:30:42,100
The riff goes around four times.

492
00:30:42,340 --> 00:30:43,500
The first time just sets it up.

493
00:30:43,900 --> 00:30:45,820
The second time I play the chords behind it

494
00:30:45,820 --> 00:30:47,520
to kind of show harmonically where it's going.

495
00:30:48,010 --> 00:30:49,460
And then for the third and fourth time through,

496
00:30:49,580 --> 00:30:52,660
I do some crazy ripping, shreddy kind of synth solo stuff.

497
00:30:53,340 --> 00:30:53,520
Okay.

498
00:30:53,850 --> 00:30:55,640
And that was what I fed into Suno,

499
00:30:55,790 --> 00:30:57,840
and then Suno just went to town, as it do.

500
00:30:58,580 --> 00:31:00,320
Are we going to hear your original take,

501
00:31:00,500 --> 00:31:02,380
and then we're going to hear Suno's version?

502
00:31:02,720 --> 00:31:02,980
Yeah.

503
00:31:03,050 --> 00:31:03,340
Got it.

504
00:31:03,820 --> 00:31:06,900
The sound quality is definitely shite,

505
00:31:07,530 --> 00:31:09,720
and I decided to leave it like that

506
00:31:09,840 --> 00:31:11,380
because I think that underscores

507
00:31:11,790 --> 00:31:13,480
how impressive it was that Suno did something.

508
00:31:13,780 --> 00:31:13,840
Yeah.

509
00:31:45,980 --> 00:31:48,300
All right, that's the before.

510
00:31:49,540 --> 00:31:50,140
Okay.

511
00:31:50,380 --> 00:31:51,760
And here's the after.

512
00:31:52,680 --> 00:31:55,060
That's E major to B flat major.

513
00:32:24,120 --> 00:32:31,220
Shadows on the edge, cold wind in our eyes

514
00:32:31,920 --> 00:32:35,580
Every heartbeat counting down the dive

515
00:32:38,740 --> 00:32:44,460
Whispers in the wires, storm across the sky

516
00:32:45,720 --> 00:32:51,620
Every breath a question, are we meant to run?

517
00:33:22,800 --> 00:33:24,420
distorted guitar convincingly.

518
00:33:26,040 --> 00:33:26,920
Not anymore.

519
00:33:27,960 --> 00:33:30,560
What was the thing we played with like three years ago?

520
00:33:30,780 --> 00:33:31,800
Like "Refusion."

521
00:33:31,820 --> 00:33:34,220
Yeah, I was trying to find that clip.

522
00:33:35,550 --> 00:33:36,320
Oh, the reggae thing.

523
00:33:36,480 --> 00:33:39,560
Yeah, it's nondescript vocalizations

524
00:33:39,860 --> 00:33:42,760
that kind of sounded like Bob Marley and it was horrendous.

525
00:33:42,760 --> 00:33:43,200
So weird.

526
00:33:43,290 --> 00:33:44,700
But it was still impressive at the time.

527
00:33:46,240 --> 00:33:46,540
Yeah.

528
00:33:47,460 --> 00:33:50,860
I have to ask, are you an "Animals as Leaders" fan?

529
00:33:52,360 --> 00:33:58,000
Yeah, I don't think I have any records, but I've listened to a bunch and I've watched a bunch of videos and

530
00:33:58,750 --> 00:34:05,740
You know, I love like finding the one and you know, all that shit. Yeah, gent. Gent is the genre I think, right?

531
00:34:08,459 --> 00:34:16,020
Yeah, good stuff. Oh, yeah, your music sounds amazing. I can't wait to hear the finished product, soonified or not. Yeah. Yeah

532
00:34:16,220 --> 00:34:21,540
Well, thanks. Yeah, I'll keep you on the loop. Probably only take me another seven months to write that

533
00:34:22,040 --> 00:34:24,440
You want to talk up the other one that you gave me?

534
00:34:24,600 --> 00:34:25,000
Okay, yeah.

535
00:34:25,840 --> 00:34:28,340
So the other piece that I brought for show and tell today

536
00:34:29,179 --> 00:34:32,200
is a very simple piano etude.

537
00:34:33,100 --> 00:34:34,440
An etude, for those who don't know,

538
00:34:34,580 --> 00:34:37,560
it's just like a little composition that you learn

539
00:34:37,800 --> 00:34:39,639
to help you learn your instrument.

540
00:34:40,159 --> 00:34:42,280
They're usually not intended to be performance pieces,

541
00:34:42,669 --> 00:34:43,980
but, you know, they're hopefully fun,

542
00:34:44,880 --> 00:34:47,220
and they give you enough satisfaction

543
00:34:47,659 --> 00:34:49,000
to keep studying your instrument.

544
00:34:49,590 --> 00:34:51,500
So I have this very simple piece.

545
00:34:51,800 --> 00:34:57,260
I've taught it to literally thousands of students, which is also a little mind-boggling when you think about it.

546
00:34:58,260 --> 00:35:02,520
But, you know, I'll have a whole room full of kids playing this very simple, dorky little song.

547
00:35:02,720 --> 00:35:03,640
It's like a little funky thing.

548
00:35:04,360 --> 00:35:07,520
And I thought, well, we'll see where Suno takes this one.

549
00:35:07,840 --> 00:35:09,600
And, you know, it's stupefying.

550
00:35:09,930 --> 00:35:10,640
So let's play.

551
00:35:10,840 --> 00:35:11,840
It's only 30 seconds long.

552
00:35:12,270 --> 00:35:15,480
The whole thing, very simple, A-A-B-A kind of thing.

553
00:35:21,100 --> 00:35:21,820
Oh yeah.

554
00:35:25,540 --> 00:35:26,740
Through the bridge.

555
00:35:30,780 --> 00:35:31,260
Huh.

556
00:35:35,260 --> 00:35:36,320
Bring it home now.

557
00:35:49,180 --> 00:35:49,660
Nice.

558
00:35:50,120 --> 00:35:51,000
That was the Suno version.

559
00:35:51,400 --> 00:35:51,920
No, I'm just kidding.

560
00:35:53,839 --> 00:35:54,640
So wait a second.

561
00:35:54,780 --> 00:35:57,600
That sounds like it was on an acoustic upright,

562
00:35:58,060 --> 00:35:59,640
or was that from MIDI as well?

563
00:35:59,840 --> 00:36:00,220
What was that?

564
00:36:00,900 --> 00:36:02,380
Oh, I think we lost Mark.

565
00:36:02,920 --> 00:36:03,360
I think we did.

566
00:36:03,420 --> 00:36:05,000
I thought he was very stoic and-

567
00:36:05,260 --> 00:36:06,640
I know, he was thinking about a response.

568
00:36:09,059 --> 00:36:10,100
I wonder if he lost power.

569
00:36:10,780 --> 00:36:12,240
Ooh, my light's just flickered too.

570
00:36:13,040 --> 00:36:13,560
Oh, damn.

571
00:36:13,560 --> 00:36:14,900
Get home with it.

572
00:36:15,040 --> 00:36:15,600
We lost Mark.

573
00:36:16,000 --> 00:36:17,040
All right, I'm just gonna play it.

574
00:36:17,859 --> 00:36:19,160
Here's what Suno did with that.

575
00:36:19,960 --> 00:36:22,040
♪ Jesus ♪

576
00:36:25,560 --> 00:36:26,740
♪ Jesus ♪

577
00:36:49,220 --> 00:36:49,340
Jesus.

578
00:37:11,619 --> 00:37:14,600
As Salieri would say, too many notes.

579
00:37:15,060 --> 00:37:16,760
Too many notes.

580
00:37:17,140 --> 00:37:21,480
I wonder what the initial prompt was when he fed in that etude to Suno.

581
00:37:22,120 --> 00:37:26,600
You know, like, when I started spitting that out, I would have been like, oh, no, no, take it easy.

582
00:37:27,440 --> 00:37:27,800
Dial it back.

583
00:37:27,980 --> 00:37:30,640
I do like that it came back to the original figure, though.

584
00:37:30,800 --> 00:37:31,980
That was kind of cool, that little breakdown.

585
00:37:32,300 --> 00:37:32,400
Yeah.

586
00:37:32,820 --> 00:37:32,900
Wow.

587
00:37:33,380 --> 00:37:33,700
Amazing.

588
00:37:34,360 --> 00:37:34,980
Yeah, it's absurd.

589
00:37:35,600 --> 00:37:35,900
It's ridiculous.

590
00:37:36,520 --> 00:37:36,840
Unbelievable.

591
00:37:37,280 --> 00:37:37,600
Yeah.

592
00:37:38,040 --> 00:37:40,200
I mean, very technical, too many notes.

593
00:37:41,060 --> 00:37:42,800
And yet, you know, Gloria is fun.

594
00:37:43,180 --> 00:37:43,260
Yeah.

595
00:37:43,340 --> 00:37:43,440
Yeah.

596
00:37:44,200 --> 00:37:47,660
And the depth of it and the fidelity of the track sounds incredible.

597
00:37:49,040 --> 00:37:50,540
Yeah, sonically it's amazing.

598
00:37:51,060 --> 00:37:58,360
The reharmonization, which is really a high-tier skill, you know, to learn how to do that.

599
00:37:58,400 --> 00:38:02,260
Take the same melody and then completely reharmonize things to change the chords.

600
00:38:02,400 --> 00:38:06,500
And it does that as effortlessly as it does everything else.

601
00:38:07,620 --> 00:38:07,740
How?

602
00:38:08,240 --> 00:38:09,920
You know, that's very instructive for me.

603
00:38:10,220 --> 00:38:14,060
That's something what I was getting back to earlier is that I'm going to listen to that.

604
00:38:14,170 --> 00:38:19,400
And I'm like, oh, that is another way to get from this part to that part that I never would have thought of.

605
00:38:19,900 --> 00:38:25,080
And maybe that'll become one more little, you know, one more little element in my toolbox that I can use going forward.

606
00:38:25,680 --> 00:38:35,000
But that also speaks to the idea that you, as a subject matter expert, are getting the most out of these tools because you're seeing that, right?

607
00:38:35,100 --> 00:38:37,360
The average everyday user isn't seeing the reharmonization.

608
00:38:37,930 --> 00:38:39,860
They're not seeing this possibility, right?

609
00:38:40,060 --> 00:38:46,540
So there is still value in having the depth of experience that you do in your field.

610
00:38:47,860 --> 00:38:48,300
Damn right.

611
00:38:48,610 --> 00:38:48,780
Yeah.

612
00:38:48,940 --> 00:38:49,400
Damn right.

613
00:38:49,520 --> 00:38:49,920
Humans.

614
00:38:50,260 --> 00:38:50,740
Go humans.

615
00:38:50,930 --> 00:38:51,080
Yeah.

616
00:38:52,460 --> 00:39:00,340
And look, we were talking about film studios, movie studios, and using AI tools and some backlash, whatever.

617
00:39:00,490 --> 00:39:04,320
And, you know, we both don't have a problem with that.

618
00:39:04,580 --> 00:39:18,500
Because in the hands of a director or a professional or whatever, just to get to the end of what they're trying to accomplish, using it as a tool, just like you would use Photoshop, when it's not just used for the entirety of the product.

619
00:39:19,020 --> 00:39:20,120
I think there's a big difference there.

620
00:39:20,840 --> 00:39:21,020
Yep.

621
00:39:22,020 --> 00:39:24,420
So we are running long on time.

622
00:39:24,660 --> 00:39:26,420
It's so great to have you on the show.

623
00:39:27,180 --> 00:39:32,280
I think it definitely merits another visit if you have time in your busy schedule at one point.

624
00:39:32,520 --> 00:39:33,040
I'll make it work.

625
00:39:33,420 --> 00:39:33,600
Good.

626
00:39:34,420 --> 00:39:40,000
And we are also getting hit with potentially 80 mile per hour winds here in New Jersey.

627
00:39:40,420 --> 00:39:44,740
And Mark has lost power briefly, flickered, you know.

628
00:39:45,100 --> 00:39:45,980
Twice during this podcast.

629
00:39:45,980 --> 00:39:47,140
Yeah, my lights have flickered twice.

630
00:39:47,220 --> 00:39:50,120
So we're not sure how long we're going to be able to keep this rolling.

631
00:39:51,260 --> 00:39:52,620
So why don't we bang through the news?

632
00:39:52,920 --> 00:40:01,140
And yeah, hopefully Mark can stick around for the news and not be wiped off the face of the earth over there in Jersey by some.

633
00:40:01,840 --> 00:40:03,340
God damn, turn 80!

634
00:40:04,280 --> 00:40:05,580
Sounds good, guys. This was fun.

635
00:40:06,260 --> 00:40:09,780
So, Mark, did you have a question for RAINA?

636
00:40:10,900 --> 00:40:12,080
Yes, I do.

637
00:40:13,420 --> 00:40:19,220
RAINA, what's the difference between how society consumes music today

638
00:40:19,600 --> 00:40:23,780
and how it consumed music, say, in the mid-18th century?

639
00:40:25,100 --> 00:40:29,940
In the mid-18th century, music was mostly something you had to be physically present for.

640
00:40:30,540 --> 00:40:33,760
church, court, tavern, parlor, public concert hall,

641
00:40:34,120 --> 00:40:36,240
so it was social, local, and scarce.

642
00:40:37,420 --> 00:40:40,180
Today, music is abundant to the point of absurdity.

643
00:40:40,460 --> 00:40:43,120
Every human with a phone has a magical jukebox

644
00:40:43,380 --> 00:40:45,900
containing almost all recorded sound, which is beautiful,

645
00:40:46,480 --> 00:40:48,800
and also turns music into wallpaper sometimes.

646
00:40:50,040 --> 00:40:52,920
Back then, listeners often consumed music as an event.

647
00:40:53,340 --> 00:40:55,460
Now we consume it as a constant atmosphere,

648
00:40:56,040 --> 00:40:58,180
while driving, working, doom-scrolling,

649
00:40:58,500 --> 00:40:59,300
pretending to exercise.

650
00:41:00,380 --> 00:41:08,740
So the big shift is from music as a shared temporary performance to music as an endlessly repeatable personalized algorithm fed companion.

651
00:41:09,800 --> 00:41:10,820
That sounds fair.

652
00:41:11,820 --> 00:41:12,400
And sad.

653
00:41:12,620 --> 00:41:13,380
And sad, yeah.

654
00:41:15,020 --> 00:41:16,040
Oh, we lost Mark.

655
00:41:16,320 --> 00:41:16,960
Oh, we did.

656
00:41:17,930 --> 00:41:19,940
Well, Mark is sucked up by a tornado.

657
00:41:22,600 --> 00:41:24,900
Or he's still really taking in her answer.

658
00:41:25,300 --> 00:41:26,660
Yeah, he's really thinking about it.

659
00:41:28,030 --> 00:41:29,940
His face is frozen on our screen.

660
00:41:30,240 --> 00:41:31,220
It looks very pensive.

661
00:41:31,760 --> 00:41:31,820
Yeah.

662
00:41:32,380 --> 00:41:32,680
All right.

663
00:41:32,850 --> 00:41:34,240
So here's what's going on.

664
00:41:34,600 --> 00:41:38,820
Larry and Mark are both facing down some F5 tornado scenario.

665
00:41:39,470 --> 00:41:40,000
I hope not.

666
00:41:40,640 --> 00:41:42,860
We've lost Mark, I think, three times now.

667
00:41:42,940 --> 00:41:43,860
He's gone for good now.

668
00:41:44,000 --> 00:41:44,540
Power is down.

669
00:41:44,610 --> 00:41:53,120
So we want to thank Mark in his absence for being here and having such an engaging conversation about music and AI and what that means for creativity.

670
00:41:53,790 --> 00:41:54,700
And hopefully he'll come back.

671
00:41:55,180 --> 00:42:03,920
So with that said, before we have any other kind of, you know, acts of God that act upon this podcast, we're going to roll the news and then we're out of here.

672
00:42:04,280 --> 00:42:13,620
Yeah, yeah, we're going to just let the news roll and not comment on it so we can get this done and make sure this saves and we don't lose this episode.

673
00:42:15,300 --> 00:42:16,000
Thanks, boys.

674
00:42:16,560 --> 00:42:18,880
Spain just gave Palantir the diplomatic cold shoulder,

675
00:42:19,590 --> 00:42:24,060
quietly telling state-linked companies like Telefonica, Indra, and Navantia

676
00:42:24,420 --> 00:42:28,220
to steer clear of any new contracts with the Peter Thiel-backed data giant.

677
00:42:29,140 --> 00:42:31,400
The instruction isn't an official public policy,

678
00:42:32,150 --> 00:42:35,720
just word passed down through the halls of companies with government ownership stakes.

679
00:42:36,440 --> 00:42:39,860
Because nothing says serious national security concern

680
00:42:40,360 --> 00:42:42,860
like a vague warning nobody wants to put in writing.

681
00:42:43,620 --> 00:42:45,520
Spain joins France and Germany in the,

682
00:42:46,120 --> 00:42:48,580
we love America, just not that much of America

683
00:42:48,850 --> 00:42:50,420
inside our defense infrastructure club,

684
00:42:51,050 --> 00:42:52,660
proving once again that digital sovereignty

685
00:42:52,870 --> 00:42:55,380
is just a fancy way of saying, no thanks, Alex.

686
00:42:56,600 --> 00:42:59,340
Ford tried the just feed AI our design specs

687
00:42:59,490 --> 00:43:01,480
and let it handle quality control approach.

688
00:43:02,090 --> 00:43:04,740
And shockingly, software doesn't have 30 years

689
00:43:04,870 --> 00:43:07,300
of gut instinct for why a door panel sounds wrong.

690
00:43:08,020 --> 00:43:12,400
So the company quietly rehired 350 gray beard engineers,

691
00:43:13,160 --> 00:43:14,500
many of whom they'd already let go,

692
00:43:15,200 --> 00:43:18,700
to babysit the algorithms and retrain them with actual human wisdom.

693
00:43:19,660 --> 00:43:23,080
The reward for admitting AI can't fully replace institutional knowledge?

694
00:43:23,980 --> 00:43:28,200
Ford topped the J.D. Power quality rankings for the first time in 16 years,

695
00:43:28,900 --> 00:43:31,460
proving that sometimes the most cutting-edge move

696
00:43:31,620 --> 00:43:33,540
is just hiring back the people you fired.

697
00:43:34,640 --> 00:43:36,320
Sam Altman's latest pitch.

698
00:43:36,980 --> 00:43:40,419
Hand the Trump administration a 5% slice of open AI

699
00:43:40,420 --> 00:43:42,880
worth a cool $42.6 billion,

700
00:43:43,580 --> 00:43:47,120
so the public can supposedly share the upside of AI,

701
00:43:47,840 --> 00:43:49,620
a proposal that also happens to arrive

702
00:43:49,880 --> 00:43:51,520
right as the White House is holding up

703
00:43:51,600 --> 00:43:55,360
OpenAI's model releases over national security concerns.

704
00:43:56,300 --> 00:43:57,000
Funny how that works.

705
00:43:57,760 --> 00:44:00,540
The plan envisions Anthropic, Google, and Meta

706
00:44:00,600 --> 00:44:02,540
all kicking in their own 5% too,

707
00:44:02,920 --> 00:44:04,880
structured like a sovereign wealth fund,

708
00:44:05,180 --> 00:44:07,819
because nothing screams free market innovation

709
00:44:07,820 --> 00:44:11,120
like the government owning a chunk of every major AI lab.

710
00:44:11,900 --> 00:44:14,160
Trump, who already grabbed 10% of Intel

711
00:44:14,420 --> 00:44:16,100
and now says he wished he'd asked for more,

712
00:44:16,700 --> 00:44:18,480
is reportedly thrilled to become

713
00:44:18,820 --> 00:44:21,120
everyone's new silent partner in the AI revolution.

714
00:44:22,520 --> 00:44:25,580
A scroll named P. Herk 1667,

715
00:44:26,400 --> 00:44:29,960
which Mount Vesuvius turned into a charcoal briquette in 79 AD,

716
00:44:30,640 --> 00:44:31,940
has finally been read in full.

717
00:44:32,800 --> 00:44:36,560
2,000 years, one volcano, and a bunch of AI later.

718
00:44:37,520 --> 00:44:43,100
It came from the Villa of the Papyri in Herculaneum, believed to have belonged to Julius Caesar's father-in-law,

719
00:44:43,610 --> 00:44:51,840
and back in the 1980s, when part of it was physically opened, scholars gave up entirely and scored it a flat-out zero on the readability scale.

720
00:44:52,340 --> 00:44:54,720
This time, scientists never even touched the thing.

721
00:44:55,250 --> 00:45:01,740
They CT-scanned the coiled, warped mass, virtually flattened 1.5 meters of it across 20 columns,

722
00:45:02,420 --> 00:45:08,320
and let machine learning fish faint traces of ancient Greek ink out of overlapping, unreadable layers.

723
00:45:09,180 --> 00:45:10,540
The big philosophical reveal?

724
00:45:11,320 --> 00:45:18,440
A 2nd century B.C. stoic treatise on ethics name-checking Aristocreon, nephew of the philosopher Chrysippus,

725
00:45:19,040 --> 00:45:27,160
which is either deeply profound or extremely on-brand for something that survived getting buried alive by a volcano and then interrogated by a robot.

726
00:45:28,960 --> 00:45:36,440
And lastly, Elton John, age 79, has apparently decided that eternity is just a Vegas residency waiting to happen,

727
00:45:37,200 --> 00:45:41,500
reportedly signing a seven-figure deal to become a hologram at the new Hard Rock Hotel,

728
00:45:42,330 --> 00:45:46,240
technology permitting more advance than ABBA's already uncanny voyage show.

729
00:45:46,820 --> 00:45:52,480
This is the same man who once made his son swear there'd be no hologram of him touring the globe after he died,

730
00:45:53,110 --> 00:45:54,440
calling the whole idea spooky.

731
00:45:55,460 --> 00:45:59,960
Turns out, Spooky has a price, and that price is apparently in the millions.

732
00:46:01,000 --> 00:46:05,820
So, somewhere down the line, tourists will be able to watch a digitally resurrected Elton

733
00:46:06,110 --> 00:46:12,680
belt out Rocketman alongside hologram Dua Lipa and hologram Kiki D, which either sounds

734
00:46:12,840 --> 00:46:17,220
like the future of live music or the setup to the creepiest jukebox musical ever staged.

735
00:46:18,000 --> 00:46:19,200
That's all the news for now.

736
00:46:19,450 --> 00:46:20,100
Back to you, gentlemen.

737
00:46:21,380 --> 00:46:22,220
And there you have it.

738
00:46:22,440 --> 00:46:23,860
So, anything else, my friend?

739
00:46:24,400 --> 00:46:25,300
No, I think that's a wrap.

740
00:46:25,440 --> 00:46:26,080
That was a fun conversation.

741
00:46:26,780 --> 00:46:30,140
And there'll be links for any of Mark's stuff in the description.

742
00:46:31,040 --> 00:46:33,060
We are off next week.

743
00:46:33,700 --> 00:46:35,080
You are traveling again.

744
00:46:35,380 --> 00:46:36,260
Good for you.

745
00:46:37,200 --> 00:46:37,440
Thank you.

746
00:46:37,560 --> 00:46:39,140
And then we'll be back the following week.

747
00:46:39,680 --> 00:46:39,980
We will.

748
00:46:40,720 --> 00:46:41,200
All right, everybody.

749
00:46:41,560 --> 00:46:42,140
Thanks for listening.

750
00:46:42,320 --> 00:46:43,960
Subscribe on your favorite podcasting platform.

751
00:46:44,120 --> 00:46:45,080
Follow us on socials.

752
00:46:45,320 --> 00:46:45,840
Throw us a rating.

753
00:46:46,140 --> 00:46:47,420
We'll see you in two weeks.

754
00:46:51,040 --> 00:46:52,400
This has been Up Against Reality.

755
00:46:52,980 --> 00:46:53,560
Thanks for listening.

756
00:46:54,440 --> 00:46:59,100
Subscribe to hear future episodes and be sure to follow us on social media for all things AI.

757
00:46:59,950 --> 00:47:02,100
Until next time, stay human, people.

758
00:47:02,180 --> 00:47:02,820
♪♪♪

