9/25/2025 - Exploring the Power of Suno V5 by Creating Endless Covers of a Single AI Song

Description

Chapters

Transcript

0:01 Let's go. Let's go. Let's go. Let's go.
0:04 Let's go. Let's go.
0:08 [Music]
0:19 It's not simp.
0:22 [Music]
0:29 [Music]
0:38 Good evening. Good people. What's
0:40 happening on the internet? Woohoo!
0:44 [Music]
1:02 Yeah.
1:03 [Music]
1:25 [Music]
1:30 Yeah.
1:31 [Music]
1:38 It's not simple to say
1:41 that most days
1:44 I don't recognize me that these shoes
1:48 and this apron that place and its
1:51 patrons
1:53 have taken more than I gave them.
1:57 It's not easy to know
2:01 I'm not anything
2:03 like I used to be. Although it's true I
2:07 was never attention sweet Santa. I still
2:12 remember that girl.
2:15 She's a perfect but she tries.
2:19 She is good but she lies.
2:24 She is hard on herselves.
2:28 She is broken and won't ask for help.
2:33 She is messy, but she's kind.
2:37 She's lonely
2:40 most of the time. She's all of this
2:43 mixed up and baked a beautiful pie.
2:48 She's gone, but she used to be mine.
2:53 [Music]
3:03 [Music]
3:12 All right, good people. What's
3:14 happening? What's shaking? What's going
3:15 down?
3:17 [Music]
3:33 blue.
3:41 [Music]
4:20 All right. All right. All right. All
4:23 right. Love it when your dog sings
4:25 along. Isn't he the best? He's the best
4:28 at champ. He is a good singing dog. As
4:31 far as singing dogs go, I got a good
4:33 one.
4:36 That said, I don't think there's a bad
4:37 singing dog. Singing dogs are the best.
4:39 And they don't all sing. I've had a
4:41 bunch of dogs. They didn't couldn't give
4:44 two shits. Champ, on the other hand, I
4:48 can't tune my guitar without some some
4:50 vocalization.
4:51 [Laughter]
4:54 [Music]
5:07 Hey Danny, what's happening?
5:12 Welcome everybody. If you're on YouTube,
5:14 welcome. Or LinkedIn or X, welcome,
5:17 welcome, welcome. I'm Kyle Shannon. This
5:20 is the AI learning lab brought to you by
5:21 the AI Salon.
5:23 the AI salon. If you go to the salon.ai
5:29 and say join our community.
5:34 [Music]
5:42 He's totally not into it. He sang for a
5:44 bit.
5:51 [Music]
6:07 Um, I don't know if you remember this,
6:09 but about a month or so ago, maybe it
6:11 was two months ago, Meta announced that
6:14 they did a deal with Midjourney.
6:16 Well, today
6:19 they updated the Meta AI app. It doesn't
6:22 work on my iPad or I'd show it to you,
6:26 but um
6:29 they basically added Who did this
6:32 before?
6:34 Brandon, do you remember who did the
6:36 feed? It was one of the video companies.
6:38 They did a feed that was kind of like
6:40 Tik Tok
6:41 and then now Sunno just added one. I
6:44 can't remember who it was. So now Meta,
6:47 they've got a feed and what it is is
6:50 it's 5-second midjourney videos. So it's
6:55 it's kind of a slightly wonky interface.
6:58 So you go to this feed, it's called
7:01 Vibes. So basically, if you've got the
7:03 Meta AI app, you update your Meta AI
7:06 app, and then when you launch it, you
7:08 see like videos, like music videos, but
7:11 they're like 5 seconds long.
7:13 They can be longer, but they don't make
7:15 it obvious. And then if you click on any
7:18 of the bottom icons, those are not the
7:20 ones. There's a little teeny tiny plus
7:22 button in the upper right hand corner.
7:23 And if you click that, it just basically
7:25 takes you to a prompt window. And then
7:28 you say, "Make me a video about
7:30 whatever." And it makes you images.
7:33 And then you click one of those images
7:35 and you say, "Aanate."
7:37 And then you can add music to it. And
7:40 but you can also extend it. So I I've
7:43 done two of them now and it's and it's a
7:44 feed just like a Tik Tok feed. So it
7:46 looks like Tik Tok Kyle. It works on the
7:49 Meta AI website supposedly. Oh,
7:51 interesting. Let's go see that.
7:54 It's meta.ai, right? That's the meta
7:57 website.
7:59 Meta Meta.ai.
8:08 [Music]
8:11 That's a little loud. Why is he plucking
8:13 the guitar strings so loud? I thought
8:15 this was the AI learning lab. That
8:17 sounds a little analog to me, if you
8:19 understand what I'm saying.
8:21 It's not very AI like. I thought he
8:23 would be producing songs that were
8:25 stolen from other artists or something
8:27 like that.
8:29 Okay. Um,
8:35 discover.
8:38 Yeah, here it is.
8:42 Can I turn on audio? Yes. So, I guess I
8:46 need to share this, don't I? Share this
8:48 tab instead. Okay. Tabs. Yes. Browser.
8:53 Yes. Everybody calm down. Everybody calm
8:57 down. I don't need the attitude.
9:01 I'm doing my best here. I'm an old man.
9:04 I'm an old man on the internet.
9:13 Look at those jumpy little beans. All
9:15 right, so there's the jumpy little
9:17 beans. Let's turn on audio.
9:20 Did I turn on?
9:25 [Music]
9:32 [Music]
9:35 >> All right.
9:39 So that's it. So I've got
9:42 Is this it? Yeah, this is the one I did
9:45 earlier. Here's mine.
9:48 Where's my audio?
9:52 How do I get to my stuff? New chat.
9:56 Kyle Shannon.
10:00 I already added view profile.
10:04 Do you need midjourney to make it? No,
10:05 you don't.
10:07 [Music]
10:10 Oh yeah.
10:14 Uhhuh.
10:16 [Music]
10:19 >> Oh, the Pika. That's what it was.
10:22 Danielle,
10:23 >> the Pika app. So, okay. So, there's
10:26 Listen, I said this last night. Remember
10:28 when I said last night that there's
10:30 going to be a huge opportunity in
10:32 becoming an AI creator because what's
10:35 going to happen is
10:37 the traditional entertainment
10:39 establishment is going to [ __ ] all over
10:41 AI with the exception of Lionsgate.
10:43 They're trying to embrace it, but they
10:46 did it wrong and they're having trouble
10:48 now. But but because they're trying to
10:50 basically turn AI into their old style
10:54 of production, that's not going to work.
10:55 But whatever.
10:58 Pika has a something very similar to
11:01 this. The Pika app. P I K A. This is
11:04 Meta. Um,
11:07 Sunno, the music generation thing, just
11:09 created a feed called Hooks, which is
11:12 let you put Sunseo music behind videos.
11:16 Now, they don't have the ability to
11:17 generate the videos in Sunno, but I bet
11:19 they're going to. And then and then all
11:22 of the con the video generation
11:24 platforms are also doing like TV
11:26 channels and like you know Netflix style
11:29 interfaces for content. Um so it's just
11:32 further it's it's further evidence to me
11:36 that content creation platforms are
11:38 going to become the new studios. Studios
11:41 are going to fight it. They're going to
11:42 try to sue these people into oblivion
11:44 and then that's going to shake out
11:46 however it shakes out. Some of them will
11:48 get sued into oblivion, but none of this
11:50 is stopping.
11:51 And if you're a creator that that makes
11:55 a name for yourself on one of these
11:57 platforms, like now would be the time
11:59 when nobody knows they exist, right?
12:01 Important Tik Tok pip. How can someone
12:04 with no experience get into AI? Hey
12:06 Martin Wood, great great great question.
12:10 One, step one, show up to [ __ ] like
12:13 this. So let me let me take off my
12:16 guitar. This is serious, Martin. We got
12:18 to have a talk. You and I, we got to
12:19 talk. Um,
12:23 so in the AI salon, we we the the the
12:26 sort of mission of the AI salon is is to
12:28 promote what we call AI readiness. And
12:30 AI readiness is three simple stages.
12:33 Play first,
12:36 um, create excellence or mindfully
12:38 create, create. So play, create, and
12:41 then the third step is generously lead.
12:43 And what that looks like is learn out
12:45 loud and share with other people. So if
12:48 you go to the salon.ai and say join our
12:50 community, you'll you'll enter the AI
12:52 salon community. And it's a community of
12:55 about 3,000 AI optimists. So step one is
12:58 get in community with people that are
13:00 playing with AI. What's amazing about
13:03 the AI salon is one, the people are very
13:05 very generous and they're they're
13:07 willing to share what they learn. And
13:09 the other thing is this channel, the AI
13:12 learning lab, I go live five nights a
13:15 week for about 2 hours starting at 8:00
13:17 p.m. and just talk about AI, play with
13:20 AI, show you stuff, do all that sort of
13:22 stuff. Um,
13:25 show up here night after night. Not
13:28 because I'm going to teach you anything
13:30 necessarily, but put yourself in
13:33 communication with people that are
13:35 playing with this stuff, too. And while
13:37 I'm waxing poetic or going on a rant
13:40 about AI or demoing some piece of
13:42 software poorly, you can be playing
13:44 along and either doing what I'm doing,
13:47 what you're seeing me do on screen or
13:48 just playing on chat GBT or playing in
13:50 MidJourney, playing in some other tool.
13:54 You just need to be in the conversation.
13:56 There there was a you know Malcolm
13:58 Gladwell's book I think it was outliers
14:00 talked about you know the 10,000 hour
14:02 rule that if you wanted mastery of
14:04 something you had to put in 10,000 hours
14:06 um I think AI is accelerating that but
14:10 but 10,000 prompts is the new 10,000
14:13 hours right so just start playing with
14:16 AI
14:18 there there's another thing about AI
14:21 that even in how you asked your question
14:23 I feel like you've put pressure on
14:25 yourself. I have to learn AI,
14:28 right? Like I think there's this I think
14:31 people sort of people go one of two
14:34 ways. Well, they go one of three ways.
14:36 One is they just panic. They get they
14:37 have panic and fear and they're like,
14:38 "Fuck it. I'm not going to deal with
14:39 it." One is they get excited about it,
14:41 and that's a little more rare. And then
14:44 the other one is, "Oh my god, I have to
14:46 learn this, right? It's going to change
14:47 everything. Someone just got laid off
14:49 because of AI. I got to learn this. I
14:50 got to learn this." That puts a lot of
14:53 pressure on you to get really good at a
14:55 specific task. And and unfortunately
14:59 what that does is it will lock you into
15:03 just using AI to make what you currently
15:05 do more efficient,
15:08 which it can do. You can say, "Listen, I
15:10 write emails all day and I want AI to
15:12 make my email writing more efficient."
15:14 You can do that. But what you miss out
15:17 if you if you sort of narrow your focus
15:20 too much early on is you miss out on
15:22 what it makes possible that's way beyond
15:25 what you already know. So the play
15:28 thing, one of the one of the nicknames
15:30 for this channel is chatd.
15:34 And so if if you're if you're coming
15:36 here expecting a well- constructed artic
15:41 curricula, this is not this is not for
15:43 you. Um,
15:47 but there is so much going on in AI that
15:50 you literally you could just throw a
15:52 dart at an AI tool these days. They're
15:54 all so good that you'll learn you will
15:58 probably have your mind blown in the
16:00 early days of learning this stuff two or
16:02 three times a week, if not two or three
16:04 times a day, because what what you can
16:06 do now is like it it's just shattering.
16:11 In fact, we'll go play with um we'll go
16:13 play with some tools tonight. We were we
16:15 were playing last night with Sunno,
16:17 which is a music making tool, and it's
16:19 absolutely mind-blowing. Like, it's it's
16:21 just the the quality of the music that
16:23 it creates is insane. Um so, just hang
16:26 out. Just hang out. Start playing. Um if
16:29 you haven't, go to chatgpt.com and just
16:31 start typing in prompts. Just start
16:33 playing with it. See what it can do. Um
16:35 you can watch YouTube videos, but really
16:37 just hang out, ask people questions,
16:39 things like that. Okay,
16:42 fantastic. Let's get going here, Dian.
16:45 All right, Bob. Hey, Bob. Bob. Hey, Bob.
16:49 Yeah. Okay. Could you tell them what
16:50 they've won? Oh, yeah. We uh they
16:54 haven't won anything. This is not a game
16:56 show. Okay. Um so, here we are on Meta
17:00 Vibes.
17:02 Oh yeah.
17:05 [Music]
17:08 >> So this was a single prompt. Pimp
17:11 standing next to a 70s resto mod on
17:13 Coney Island. Okay. 70s pimp. Hip7s pimp
17:19 standing next to a restood on Coney
17:21 Island.
17:23 And then I hit animate. And then I hit
17:26 extend. And then I hit add music. And
17:28 that's what it did. It's just one of
17:30 those days where you don't want to wake
17:31 up. Everything is everybody's.
17:34 >> So, those are the two I've made. So,
17:36 let's see if we can make one of these
17:37 things here on the site.
17:40 Um, go to discover.
17:43 Okay. So, yeah. So, down here below
17:47 it says ask anything. So, I'm going to
17:49 say, um, make me a picture of a,
17:56 um,
17:59 of the great Gatsby
18:04 in Clay Mason
18:09 at sunset.
18:13 Uh, we'll just see what that generates.
18:17 All right. That's not giving me what I
18:19 was hoping
18:20 because it's making square images, not
18:23 tall ones.
18:34 Animate and custom animate.
18:40 So I think I think this is midjourney
18:43 animation.
18:47 They're doing some nice little animation
18:49 effects, pixel effects while it's
18:51 generating.
18:55 This is not really how the mobile
18:56 interface looks. So, I don't I don't
19:00 think this is the same thing, but I I
19:02 don't know. I I honestly don't know. But
19:05 anyway, go if you don't have the Meta AI
19:08 app, go get it. It's Midjourney.
19:12 And then I guess we could add music to
19:14 this.
19:16 So we'll do um let's do circus music.
19:26 [Music]
19:40 >> I like that.
19:41 That's freaky.
19:47 Wait, why do I not have my music? Did I
19:50 not add it?
19:52 Circus
19:56 apply.
19:58 Yeah, I did. Oh, I see. It's a separate
20:01 video.
20:03 My cous.
20:12 >> So now it's extending.
20:15 Oh, and then you can restyle it. It
20:16 looks like they've got styles here.
20:18 That's fascinating.
20:23 Oh, the other thing about this, it's
20:25 free. So, um,
20:32 Midjourney is not free, but Meta AI
20:35 image and animate is free. Now, can you
20:39 actually download it and use it like you
20:40 would use Midjourney? I don't know.
20:43 Surprised they're not using AI music. I
20:45 know. Or Instagram. This might be on
20:47 Instagram. a freak on purpose under the
20:50 sheet. You know I'm
20:54 >> post
20:56 caption
20:58 the great
21:00 let's see the clay Gatsby
21:07 gzb
21:08 right
21:10 post to feed
21:15 but why did it make it square if it's
21:17 wanting vertical ones
21:20 I don't know people. I don't know how
21:23 any of this [ __ ] works. I'm an old man.
21:27 Instagram avatars are also free. Shut
21:29 up. Shut up.
21:35 They don't make pretty pictures though
21:37 and pretty videos.
21:40 Um,
21:42 yeah. I don't know, man.
21:48 You didn't run, you didn't lie. You knew
21:50 I wanted
21:52 [Music]
21:59 >> It's It's not a Z, it's an S. Damn it. I
22:03 got to go change my Can I edit my thing?
22:13 Who let the dogs out? View
22:17 profile.
22:19 Where's my thing? Where's the thing I
22:22 just made? There it is.
22:25 Welcome to the late night circus. Raise
22:27 to be a freak on purpose. Under the
22:30 sheets, you know I'm lurking.
22:34 Welcome to the late night circus. Raise
22:36 to be a freak on purpose. Under the
22:39 sheets, you know I'm lurking.
22:46 remove from feed. Let's go back in.
22:49 Maybe I can
22:56 circus raised to be a freak on purpose
22:59 under the sheets. You know I'm lurking.
23:04 >> Welcome to Lay Night Circus. Raised to
23:06 be a freak on purpose under the sheets.
23:09 You know I'm lurking.
23:14 post caption. Here we go. The Clay
23:20 Gatsby.
23:24 That's not very clever.
23:27 I'll admit that's not very clever. But
23:30 posted to feed. See post.
23:33 There I am. Look, I'm in the feed
23:37 and I can turn on my audio.
23:46 Yeah, this is the Meta AI app.
23:49 Uh, you have to update it. Go to the go
23:51 to the app store and update it and then
23:54 uh and then it'll have this feed. And
23:56 again, making these this this whole
23:59 episode is a copyright violation.
24:03 Probably is. Yeah. Although I am looping
24:06 the same seven seconds on each of these
24:08 songs.
24:11 So maybe it's under fair use. Uh but
24:14 whatever. I don't give a [ __ ] Okay. Um
24:17 so that's that. Let's I want to go back
24:18 to Sunno
24:21 and um
24:24 Okay. So let's go.
24:29 Here we go. Um
24:33 so if you weren't here last night,
24:36 um
24:37 let's see. surprised.
24:42 Wait, surprised on the dogs out
24:47 IP. Well, so Rick, what they're doing is
24:52 very few people are going to understand
24:54 that they can extend a video. So they're
24:57 they're 5-second midjourney videos and
25:00 then they're just you're clipping a
25:01 5-second clip to the video. So most of
25:05 the videos are 5second long. 5 seconds
25:07 long. And I think what is it? 7 seconds
25:09 or 9 seconds is considered fair use
25:12 without having to do anything. So I
25:15 think that's what they're banking on.
25:16 But yeah, it's everything is an absolute
25:18 flipping nightmare. Um, let's see.
25:22 Wendy Wii of Meta, I'm trying AI after
25:26 you said you have to do it. Lol. I did.
25:30 Awesome. Listen,
25:34 first of all, I appreciate that you
25:36 tried it. And I don't know, I I think I
25:38 said that last night that even if you
25:39 hate AI, try it. Um, I don't know where
25:42 you are on the lovehate relationship
25:44 with it. And listen, where I am with AI
25:48 on in terms of like my opinion of it is
25:50 is both, right? Like I hate a lot of how
25:54 it was created, but I love how
25:56 empowering it is. Independent of your
26:00 opinion or my opinion or anyone's
26:03 opinion about AI,
26:05 it's here and it's getting better and
26:08 better and better and the courts are
26:12 going to they're going to work out the
26:13 copyright stuff. They're going to work
26:14 out the attribution stuff. That stuff's
26:16 going to be handled. It will not make
26:19 everyone happy. I promise you that. It's
26:21 gonna favor the big companies. Like all
26:24 the same [ __ ] that happened with
26:26 music sampling in the 80s and 90s. It's
26:29 just it's gonna be that but but like on
26:31 steroids.
26:33 That's all going to play itself out. And
26:36 AI is still going to be here and it's
26:37 still going to be progressing. So
26:40 trying it, play with it. Um, just play
26:44 play play
26:47 until you've had here. Here's the thing
26:48 that I would I would encourage you to
26:50 do. Wait, what's your name again? Wendy
26:53 We of Meta. Um, if you haven't had your
26:56 first Kevin Mallister moment, so Kevin
26:59 Mallister was that kid in Home Alone
27:00 that that went like this.
27:04 When you play with AI, there will be a
27:06 point, if you haven't had it yet, just
27:08 keep playing. There will be a point
27:10 where it will do something that is so
27:12 insane that it will literally make your
27:15 jaw drop open.
27:17 And just keep playing. Just keep pushing
27:19 it. Just keep pushing it. And what you
27:21 will discover is there are things that
27:23 it can do that you just
27:26 it doesn't make any sense that computers
27:29 can do what they're doing today. They've
27:32 never behaved like this in history. um
27:34 starting with chat GBT and you know a
27:37 few of the image generation tools before
27:38 that um this this invention that Google
27:43 made called the transformer is what's
27:44 making all of this possible u and it's
27:47 absolutely mindbending so anyway
27:50 the slop bomb bomb is coming for AI
27:54 corporate use not if I have my way with
27:57 it RJ
27:58 um
28:00 you know there there are there are two
28:03 ways that you can use AI. You can use it
28:05 as a lazy
28:07 intellectual vending machine, right? You
28:10 put in a prompt, you get out a thing and
28:12 you share that squit, squit.
28:16 That's AI slop. I agree with you that
28:19 that's [ __ ]
28:21 And it's what what's gonna happen is
28:27 right now because not everyone uses AI,
28:30 people who don't use AI regularly, when
28:33 AI slop is put in the world, it looks
28:36 like decent content to them.
28:40 If you use AI even for 10 [ __ ]
28:42 minutes, when someone puts something
28:44 lazy and shitty in the world, for the
28:45 most part, you're like, "Yeah, that was
28:47 lazy." Right? Like all the LinkedIn
28:49 posts where people just put up a, you
28:51 know, the first thing that came out of
28:53 chat GPT when they said, "Make me an
28:55 image for this article." And they just
28:57 post the first thing that came out. You
28:59 can tell. I've done it. I've gotten
29:01 yelled at for it, right? And and
29:05 rightfully so.
29:07 But when you shift your relationship
29:11 with AI, not to be like a vending
29:13 machine where it's just squirting out
29:14 crap, but more like you're a creative
29:17 director or you're a producer and you're
29:20 in there with your ideas and your taste
29:24 and you're not just taking any crap that
29:28 AI gives you and sharing it with the
29:29 world. You're actually in a
29:31 collaborative conversation.
29:33 When you shift your mindset from vending
29:35 machine to producer, the quality of the
29:38 work goes up exponentially
29:41 and it actually represents who you are
29:43 and what you want to say and how you
29:44 want to impact people and that's not AI
29:46 slop. So, so don't collapse those two
29:49 because they're two very different
29:50 things. You sound like slop. Well, I you
29:53 know I then you shouldn't you shouldn't
29:56 be here. like like if if if all you want
30:00 to do is sit on the sidelines and and
30:02 and carp about how awful it is,
30:07 my fear for you is that the AI is not
30:10 going away and that AI is going to
30:12 happen to you as opposed to you saying,
30:16 "Hey, there's some things I hate about
30:17 AI
30:19 and I'm going to figure out what it is,
30:22 what it's good for, what it's not good
30:24 for, and I'm going to figure out my own
30:26 ethical boundaries for how I use it. I
30:29 totally respect that. I also totally
30:31 respect I'm going to try it and if I end
30:35 up liking it or not liking it, whatever,
30:37 and if I decide I can't deal with the
30:39 ethical stuff and I'm going to step away
30:40 from it, I totally respect that, too.
30:44 The thing I'm seeing too much of is
30:46 people being pissed off and angry about
30:48 it, making assumptions about what it is
30:50 and what it can do, and they've never
30:52 [ __ ] tried it. So, you know, that's
30:55 my two cents. But yeah, I am AI slot.
30:58 It's mathematical.
31:01 Um, okay. So,
31:04 so last night we were on Sunno. Let me
31:06 Am I sharing this tab? No. Let me share
31:08 this tab.
31:11 And we took a song that we made on this
31:14 channel, I don't know, a year and a half
31:16 ago, which is this.
31:21 [Music]
31:54 Cedar Hill.
31:55 >> All right. So, that's a town. That's a
31:57 song we did like a year and a half ago
31:58 on UIO, which is another one of these
32:00 music platforms. And we ended up, this
32:03 ended up turning into a video and it's
32:04 out there and it's a little sloppy and
32:06 muddy because it was an it was one of
32:08 the early tools. Yesterday V5 of Sunno
32:12 came out and so we played with making
32:14 some remixes of it, some covers of it.
32:17 Um, let me see. I'll find one here that
32:19 I give a thumbs up to. Here's one of the
32:21 remixes.
32:23 [Music]
32:36 So, a little cleaner.
32:42 Oh, Dar Dar the Dar the AI star. AI made
32:46 this. Yeah, AI made that. Um, and then
32:48 so last night we we so these were these
32:51 were remixes of it which you can do a
32:53 remix where it basically just takes it
32:54 and or a remaster and it remasters it.
32:57 But then we did some covers and the
32:59 covers of this actually got really
33:01 interesting last night. So this was
33:05 uh these were two covers that I did
33:07 where it analyzed the music and wrote
33:10 its own prompt.
33:12 [Music]
33:26 That's pretty good.
33:30 [Music]
33:46 Nice.
33:50 >> Right. That one's cool. That one's nice.
33:52 Here's another one.
33:56 [Music]
34:25 Cedar Hill.
34:28 >> All right. And then we started to get
34:29 into like mucking with the genre. So,
34:34 let's see. Here was some 70s funk ones.
34:38 Wait, that was Americana rock.
34:42 Let's see.
34:44 Did Were the 70s funk ones any good?
34:46 Maybe they were.
34:49 That's kind of cool.
34:55 [Music]
35:00 Oh, yeah. These didn't really do uh kind
35:02 of they didn't do funk. They they were
35:04 kind of weird.
35:08 Let's see.
35:10 The ones we got at the end of the night,
35:12 the the prompt was rap, jazz, studio
35:15 recorded, R&B, Appalachian influence.
35:23 [Music]
35:27 And we changed the vocal to a female.
35:30 [Music]
35:52 Hill.
35:54 [Music]
35:57 >> Isn't that [ __ ] insane? Okay, so
35:59 let's So, so there's a couple of things
36:01 I wanted to play with tonight. There
36:02 were there were two versions last night
36:05 that were really good, but they had they
36:08 had either weird timing in them or or
36:11 like weird chords at at one point.
36:14 So, what I thought we could do tonight
36:16 is just I want to keep playing with
36:18 different genres. I want to take this
36:20 one song and just kind of blow the doors
36:21 off it and just like push push the
36:23 envelope of what it could be. Um, and
36:27 then
36:28 if we get a song that's like super good,
36:30 but it's got a [ __ ] up part of it, I
36:33 want to go like I want to go into the
36:34 editor and see if we can fix that. And I
36:36 think we can. Um, but but we'll see. So
36:40 anyway, so let's play around with the uh
36:42 with the styles here. Let's let's make
36:43 some new ones. We'll just make a pile of
36:45 them. Waw wa guitar. We're just going to
36:47 do one version that's just Waw Wa
36:50 guitar.
36:52 Let's see. Low gritty female deep
36:55 baseline punk. That's good. Okay.
37:00 Create. We'll get rid of that.
37:04 Let's see. West Coast hip hop.
37:08 Um, let's just do West Coast hiphop.
37:13 Let's have it write its own prompt.
37:17 Oh, wait. I guess I need to
37:20 Let's see.
37:22 electronics minimalist soft piano chords
37:27 festival ready. I like that. And then
37:30 we'll have it AI improve that.
37:36 Let's see what it says. A a minimalist
37:39 electronics track with crisp punchy
37:41 drums. Pulsing subbase lays the
37:44 foundation. Soft spacious piano
37:48 chords gently float above the mix,
37:50 evolving with subtle synth textures. All
37:52 right, great. We'll do that.
37:56 [Music]
38:00 Thank you, Harshweed, for sharing the
38:01 live. Appreciate that. Get some more
38:03 people in here. Maybe we can scare Dar
38:06 back in here. Dar the Tik Tok star.
38:10 All right, let's get rid of that. Oh,
38:11 pirate shanty. Let's do a pirate shanty
38:14 version of this.
38:20 Okay. What's this? Orchestrate.
38:24 Wait. Orchestra
38:26 beagel tongue. I don't know what the
38:29 [ __ ] that is.
38:32 Booming bass. And then we'll do one more
38:34 in there. We'll throw in
38:39 pop. No, pop rock.
38:44 [Music]
38:48 All right. You want to go hear some
38:50 weird ass music? So, you heard the
38:53 original, you heard all the covers. Here
38:55 we go. Anyone else listening?
39:00 [Music]
39:04 >> Prompt here is blow on guitar.
39:06 [Music]
39:16 >> Nice.
39:18 Oo.
39:22 [Music]
39:35 >> All right. Nice but boring.
39:42 La.
39:47 [Music]
39:56 Yeah, this is this has
40:01 [Music]
40:04 [Applause]
40:11 a quiet town.
40:17 >> Nice. All right, next one is
40:22 this is low grritty female deep baseline
40:26 punk.
40:27 [Music]
40:36 Guess the uncle.
40:38 [Music]
40:51 [Applause]
40:55 [Music]
41:04 Ooh. Oh,
41:07 [Music]
41:19 >> wow.
41:24 [Music]
41:27 [Laughter]
41:31 >> Sounds like she's about to pass out.
41:39 That's I
41:41 [Music]
42:05 Wait, what? Golden golden throat
42:08 talkbox. All right, we'll try that.
42:12 That's a good style.
42:14 Golden
42:16 throat talkbox.
42:19 Oh, and let's do um um rock flute. We'll
42:23 get a little Jethro Tall going in there.
42:27 Some Peter Frampton. We're going to mix
42:29 some Peter Frampton and Jethro Tol with
42:31 this one.
42:35 Okay. Um, this is West Coast Hip Hop.
42:38 Well, you know what's funny? There
42:40 there's something about this original
42:42 track that it's it's got a it's got such
42:44 a specific opening that the covers are
42:48 not really mucking with it too much,
42:49 which I find fascinating.
42:52 That like that's not how starts.
42:55 [Music]
43:18 Oh yeah.
43:20 [Music]
43:29 Yeah, that's awful.
43:36 >> All right. Hip-hop and this original
43:38 track don't mix. But but okay. Wait. So
43:41 So wait, let's
43:44 audio influence 30, style influence 60,
43:48 weirdness 62.
43:53 Maybe lower weirdness
43:56 increase style influence.
44:01 And let's do West Coast hiphop.
44:05 West
44:08 Coast hiphop. By the way, when I was
44:10 talking earlier about um you've just got
44:14 to play with this stuff, a big part of
44:17 what I try to do here on a nightly basis
44:20 is just do weird [ __ ] Like just try to
44:24 break it. Try to get it to do something
44:26 or to not do something. And if it's like
44:29 if it's sticking too close to that
44:31 original song, there's something we can
44:33 probably do to muck with that. Um, you
44:36 know, we'll see when we get to some of
44:38 the more Maybe these minimal style
44:41 descriptions aren't good. Maybe you need
44:43 more than that for it to break the
44:45 style. I don't know. I don't know. I I I
44:49 I am unqualified.
44:53 What's this? This is electronic dance
44:56 house.
45:00 [Music]
45:08 [Applause]
45:11 [Music]
45:11 [Applause]
45:14 [Music]
45:19 Oh,
45:23 Cedar Hill is a quiet town
45:26 where the oak trees sway.
45:31 The whistle of the train echoes through
45:34 the day.
45:37 [Music]
45:45 >> I actually love this one. That one's
45:47 really good.
45:49 It's weird.
45:54 >> Let's wait till
45:56 >> I wonder if there's going to be a drop
45:57 here.
45:58 [Music]
45:59 She grew up tough
46:02 in the town we called home.
46:07 Folks never
46:09 [Music]
46:11 tone.
46:12 >> Yeah, this is Joe.
46:15 >> Longing to be on the inside,
46:18 longing for love.
46:22 She found herself as a guest in the
46:25 nervous hospital.
46:30 Weird
46:32 hills
46:36 [Music]
46:38 of sadness and shame.
46:43 >> Let's see if there's a drop here.
46:46 But when
46:48 [Music]
46:58 >> Yeah, that one's kind of Johnny one
47:00 note, but it's really pretty. I like
47:01 that one a lot.
47:05 Great lyrics. Yeah, the lyrics the
47:07 lyrics were originally AI generated, but
47:10 then over three hours we modified them
47:12 and changed them and tightened them up.
47:14 And I think we actually did
47:18 four four different sets of lyrics that
47:22 we were then mixing and matching and
47:23 then people were kind of shouting out
47:25 phrases and things like that. That's how
47:27 we got to the lyrics.
47:31 [Music]
47:44 Oh, this is pirate shanty. I was like,
47:45 that sounds like a pirate shanty cuz
47:48 it's the pirate shanty.
47:52 [Music]
48:04 Oh, that's funny. You could do an album
48:06 of all your favorite covers of Weird
48:08 Mary. It's actually not a bad idea. You
48:11 know, you know what's funny about that,
48:13 Paul?
48:16 Like
48:19 we're entering this era of
48:20 hyperpersonalization,
48:22 right? and and how I've been thinking
48:24 about that is is like as a consumer,
48:31 but imagine
48:33 imagine it as a as a creator that that
48:38 why not do an album of weird Mary
48:42 covers,
48:43 right? Why not? If you create a song you
48:46 really love and you want to explore all
48:48 these different layers of it and some of
48:50 them are really cool, like why not put
48:52 that out there as a as a Spotify
48:54 playlist or as as an album on iTunes?
48:58 That's that's that's a really
49:00 interesting idea because like
49:03 hyperpersonalization from the creator
49:05 economy is going to start to create
49:09 forms,
49:11 outputs, collections, projects
49:15 that no one would ever have considered
49:20 if you were doing production
49:22 traditionally. So that's that's a really
49:24 interesting idea. I I dig that a lot.
49:27 Um, okay.
49:30 So, that was pirate shanty. That was
49:31 okay. I think we have another pirate
49:34 shanty here. Do don't don't we? Yeah.
49:39 [Music]
49:46 >> There you go. Source camp. Just like I
49:48 have seven versions of Witchita Lineman.
49:50 Yeah. And Source Camp, you should not
49:52 turn down your weirdness, by the way.
49:57 [Music]
50:05 Yeah, KKE. This is uh this is Suna.
50:11 [Music]
50:29 Don't like those vocals. Not crazy about
50:31 them. All right. What do we got here?
50:33 Orchestrate.
50:35 This is big lung. Bagel lung. Orchestra.
50:40 Bagel lung. I assume that's some other
50:42 language for something. Booming bass.
50:45 Pop rock.
50:46 [Music]
51:03 Wait, Achilles. I want to bring out an
51:06 album of the bagpipe vacuum cleaner
51:09 disco style. Okay, wait. We're going to
51:11 make that. We're gonna go Okay. bagpipe
51:16 vacuum cleaner
51:18 [Laughter]
51:20 disco style. We're going to do We're
51:22 going to do the bagpipe vacuum cleaner
51:24 disco style of Weird Mary. And then
51:27 let's do with with uh with uh raging
51:32 vocals.
51:38 Yes. Create. Okay. So So we'll get to
51:40 that. Okay. Um, whatever that was was
51:44 boring. Oh, that was the orchestrate
51:45 one. We didn't listen to it enough, did
51:47 we? Okay, hang on.
51:50 [Music]
52:06 >> Boring.
52:08 [Music]
52:13 Cedar Hill is a quiet town
52:17 where the oak trees.
52:19 >> That one's not bad. How did this start?
52:23 [Applause]
52:24 [Music]
52:36 Yeah,
52:38 [Music]
52:47 that one's actually pretty good. We'll
52:48 give that a thumbs up. All right, next
52:51 style in the jukebox. This is Golden
52:54 Throat Talkbox
52:56 flute.
53:01 [Music]
53:08 Oh, pan flute. Sam fear. The magic of
53:12 the pan flute.
53:19 Wait, Achilles, you're on AGT with what?
53:21 With bag pipes.
53:23 I might remember that. I watch AGT all
53:25 the time. That's awesome.
53:31 Sudo needs to generate the first five
53:34 seconds and have you give it a thumbs
53:36 up. Yeah. Although you don't know like
53:38 sometimes it falls apart. The bagpipe
53:40 vacuum. Oh, that's hilarious.
53:42 [Applause]
53:48 Actually,
53:50 [Music]
53:54 you know what would be a good AGT
53:57 gimmick
53:59 is go on there and perform some banger
54:03 song and then when when Simon says, you
54:06 know, did you write that? Go, no, AI
54:08 wrote that. Like go on there and and
54:10 prove prove the haters wrong.
54:14 But like go play it with a band because
54:17 I actually I actually think that there's
54:19 a there's an interesting model. Imagine
54:21 being a cover band of AI songs that are
54:24 bangers,
54:27 which I know seems like cheating and all
54:29 [ __ ] up, but they're they're going to
54:31 be songs that no one's heard before. Um,
54:34 and if they're bangers, they're bangers.
54:38 [Music]
54:41 Oh, this is bad.
54:43 [Music]
54:58 >> Yeah, actually you call the band AI
55:00 Slop.
55:01 You absolutely call the band We're AI
55:04 Slop.
55:11 [Music]
55:15 Yeah, right. Silver Fox. That music box
55:17 is kind of creepy. It It fits the song.
55:23 [Music]
55:30 In the town we call homes
55:33 [Music]
55:36 took her noatter.
55:40 The vague childhood memories of pan
55:42 flute are from the TV commercial source
55:46 camp. You don't remember Zamfir and the
55:48 Magic of the Pan flute on on TV like
55:53 buy buying it.
56:00 Oh my god. All right. Uh here's another
56:02 one from that same genre. Golden throat
56:05 talkbox rock flute.
56:11 [Music]
56:28 [Music]
56:35 [Applause]
56:38 As cool as this one is, that would be
56:39 annoying after about the third listen.
56:41 All right, fine. Next. Uh, West Coast
56:44 hip hop again. Oh, this is one where we
56:46 changed the uh settings.
56:49 That's not hip-hop.
56:52 [Music]
56:55 >> That's cool.
57:07 [Music]
57:25 Seed Hill.
57:28 >> I don't There ain't nothing rappy about
57:30 that at all.
57:32 [Music]
57:36 >> Yeah. Weird. I don't know. It doesn't
57:37 like West Coast hip hop. Strange. Okay,
57:40 this is Okay, here's bagpipe vacuum
57:42 cleaner disco style raging vocals.
57:49 [Music]
58:00 >> Got that Scottish thing going.
58:02 [Music]
58:29 I don't want to
58:37 [Music]
58:55 Cedar Hill is a quiet town.
58:58 [Music]
59:09 [Music]
59:18 Yeah, Achilles, I agree. It's really
59:20 dumb that you have to declare AI. First
59:22 of all, you don't have to declare AI.
59:26 There's So I'm a year and a half into
59:29 writing a musical and we've used AI
59:31 throughout that, but it is not written
59:33 by AI. Like it's written by my writing
59:36 partner and I and everything that's
59:38 [ __ ] created in it has been gone over
59:40 40 [ __ ] times. And whether or not the
59:44 output was from an AI tool or not, it
59:46 was it it's it's an intentional creative
59:50 act, right? You know, we did it with
59:52 intention, with taste, with Right. So,
59:57 we're we're in a very particular time in
1:00:00 history where the the same thing
1:00:02 happened when the digital camera came
1:00:04 out. You know, photographers were like,
1:00:07 "No, that's those are not real
1:00:09 photographs. If it was not shot on film,
1:00:11 it's not a real photograph." And they
1:00:13 were like, "Look, look at the
1:00:14 resolution." You know, they it can't
1:00:16 come close to film. And until it did,
1:00:19 right? You know, it's like it it took
1:00:21 some amount of time. and and there was,
1:00:23 you know, you've got to declare it a
1:00:24 digital image because that's not real
1:00:26 photography. Same shit's going to happen
1:00:28 with AI. Yeah, I think it's absolutely
1:00:30 insane. But whatever the where where I
1:00:33 think the requirement for the
1:00:35 declaration comes from
1:00:38 is if you think about AI avatars, what
1:00:41 do we call them? We call them deep
1:00:43 fakes. In the name deep fake is the
1:00:46 nefarious intent, right? I'm gonna make
1:00:50 a deep fake of someone and have them say
1:00:52 something they didn't say, and that's
1:00:54 going to be bad.
1:00:57 Unless I make an AI avatar of myself on
1:01:01 purpose and have it do something good.
1:01:04 Is that still called a deep fake? No,
1:01:06 it's just that's an avatar, right? But
1:01:09 so so right now AI is evil until proven
1:01:14 innocent. And for some people, they're
1:01:17 like, "It's evil, and nothing you can do
1:01:19 will convince me it's anything but evil.
1:01:22 You can't there's no there's no battle
1:01:25 there. There's no argument to that,
1:01:27 right? You can't you can't win that
1:01:28 argument." And it's like, and I have no
1:01:29 interest in winning that argument, quite
1:01:31 frankly.
1:01:32 I'm just like, "Here's what this [ __ ]
1:01:34 does. Were you aware it could do this?
1:01:38 And if not, you better get [ __ ] on
1:01:40 it."
1:01:44 Oh my god, Achilles. Totally agree. I'm
1:01:47 72 and love it. Yeah, quick click fall.
1:01:49 I'm liking it, but I'm a weirdo. I also
1:01:51 love Bulgarian bag pipes. Um, okay.
1:01:56 Let's see here. I I feel like the the
1:02:00 cover of this is just like we're not
1:02:03 breaking out of Let's go back to mail.
1:02:09 Let's pump up the weirdness. Let's
1:02:13 keep the style influence high and let's
1:02:17 lower the audio influence to 15.
1:02:22 All right. And then let's go.
1:02:27 How do I get some more styles?
1:02:32 Um, house Latin
1:02:35 electronica
1:02:37 sol soft humming slow tempo
1:02:40 soaring mb boob
1:02:43 I don't even know what the [ __ ] that is
1:02:45 vibrant energy R&B have it right the
1:02:49 prompt
1:02:52 a lively groove anchored in syncopated
1:02:54 drums bright claps and a deep baseline
1:02:57 lush electronic keys smooth w guitar
1:03:00 lines and a subtle R&B flare
1:03:04 layered harmonies. All right, let's
1:03:05 let's see if we can get it to do
1:03:06 something where it actually sounds like
1:03:09 a different song.
1:03:22 Better.
1:03:24 [Music]
1:03:27 Come on.
1:03:30 [Music]
1:03:35 [Music]
1:03:41 My my
1:03:44 [Music]
1:03:49 fall down.
1:03:52 [Music]
1:03:56 down.
1:03:57 [Music]
1:04:13 [Applause]
1:04:18 They just made an instrumental. That's
1:04:20 funny.
1:04:23 I love
1:04:25 yeah
1:04:27 I
1:04:29 [Music]
1:04:31 down
1:04:34 down
1:04:35 [Music]
1:04:43 down.
1:04:45 [Music]
1:04:55 Cedar Hill is a quiet town
1:04:59 where the oak trees sway.
1:05:03 The whistle of the train echet.
1:05:14 That's kind of cool. Give that a thumbs
1:05:16 up. All right.
1:05:19 Yeah,
1:05:22 [Music]
1:05:26 yeah, yeah.
1:05:27 [Music]
1:05:52 Cedar Hill is a quiet town
1:05:56 where the oak trees.
1:05:59 [Music]
1:06:04 >> All right, I think we got the I think
1:06:06 these settings for like high weirdness,
1:06:08 high style influence, low audio
1:06:10 influence are good. When I had the audio
1:06:13 influence too low, it didn't that's when
1:06:15 we lost the vocals alto together.
1:06:18 That's pretty cool. Um,
1:06:23 let me let me move weirdness to like 60
1:06:28 and style influence to like 90.
1:06:38 [Music]
1:06:40 I
1:06:43 I
1:06:44 [Music]
1:06:51 la
1:06:55 [Music]
1:07:06 w
1:07:10 [Music]
1:07:14 That's cool.
1:07:19 [Music]
1:07:21 >> All right, let's go to LA.
1:07:27 [Music]
1:07:29 [Applause]
1:07:31 [Music]
1:07:41 Cedar Hill is a quiet town.
1:07:44 >> Oh, this is cool.
1:07:45 >> Where the oak tree sways,
1:07:50 the whistle of the train echoes through
1:07:53 the day.
1:07:57 People greet you with a night.
1:08:02 Nothing ever seems to change.
1:08:06 But the story still unfolds, each unique
1:08:09 and strange.
1:08:14 >> She grew up tough
1:08:16 in a town we call home.
1:08:24 >> See, the chords there are they don't
1:08:26 match the melody. The chords are off
1:08:28 right there. That's very strange.
1:08:31 I love how I love the clarity of the
1:08:33 vocals, though.
1:08:35 That's pretty slick. All right, we'll
1:08:37 come back to that. I'll give it a thumbs
1:08:38 up, but it's got some problems.
1:08:40 [Music]
1:09:04 [Music]
1:09:12 [Music]
1:09:15 is a quiet town.
1:09:19 >> Pretty cool.
1:09:22 [Music]
1:09:46 This is like what you'd listen to if
1:09:47 you're like at a hip doctor's office.
1:09:50 It's still It's still shitty music, but
1:09:53 it's like funk shitty music.
1:09:56 [Music]
1:10:00 It's awful.
1:10:04 Here's Gregorian chant.
1:10:21 [Music]
1:10:33 [Music]
1:10:34 Cedar Hill is a quiet town
1:10:38 where the oak swist
1:10:44 of the train echo. goes through the day.
1:10:50 People greet you with a nod.
1:10:54 Nothing ever seems to change.
1:10:58 But the story still unfolds,
1:11:02 each unique and strange.
1:11:06 She grew up tough
1:11:09 in the town we call home.
1:11:14 Folks never took to her no matter her
1:11:19 tone.
1:11:20 [Music]
1:11:22 Longing to be on the inside.
1:11:26 And Murphy, did you fall asleep on your
1:11:29 keyboard?
1:11:33 What is that last post?
1:11:38 Ann's been burning the candle on both
1:11:40 ends. I think it's gotten to her typing
1:11:42 ability.
1:11:44 [Music]
1:11:48 founders.
1:11:53 >> All right, that one's cool. Like that
1:11:55 one. All right, what's this one? This
1:11:56 one is
1:12:00 the song of uh this is this is another
1:12:02 Gregorian chant one.
1:12:05 We got three more of those.
1:12:10 >> That's good.
1:12:12 Who will
1:12:14 [Music]
1:12:20 [Music]
1:12:30 >> Okay, we definitely got out of it
1:12:32 sounding like Weird Mary at the
1:12:34 beginning.
1:12:36 [Music]
1:12:43 Walking
1:12:44 [Music]
1:12:44 [Applause]
1:12:48 we
1:12:51 walking
1:12:55 walking strong.
1:12:57 >> Cedar Hill is a quiet town
1:13:01 where the oak trees sway.
1:13:05 The whistle of the train echoes through
1:13:09 the day.
1:13:13 People greet you with a nod.
1:13:17 Nothing ever seems to change.
1:13:21 But the story unfolds each.
1:13:27 [Music]
1:13:32 >> Well, that was wild.
1:13:35 [Music]
1:13:43 All right, these are all kind of the
1:13:44 same. All right, the the Jesus. The last
1:13:47 ones I did is ac cappella.
1:13:50 Let's see.
1:13:53 Modern crisp ac cappella.
1:13:57 [Music]
1:14:03 >> Okay. Holy [ __ ] Really?
1:14:06 I still can't believe you can just do
1:14:08 this.
1:14:13 [Music]
1:14:22 D
1:14:27 [Music]
1:14:37 is a quiet town.
1:14:39 >> It's totally pentatonic.
1:14:44 So I did two versions with the male
1:14:46 singer and two versions with the female
1:14:48 sing lead singer.
1:14:50 [Music]
1:14:56 [Applause]
1:14:59 Oh,
1:15:04 [Music]
1:15:08 [Applause]
1:15:11 beautiful.
1:15:20 >> You know, the more we're cranking
1:15:21 through these genres, the more I'm
1:15:23 thinking like an album of this is
1:15:24 actually a really interesting exercise.
1:15:26 Maybe this something we do over the next
1:15:28 week. I know it's going to drive some
1:15:30 people [ __ ] crazy, but that's all
1:15:32 right.
1:15:33 >> Screw them. Like they don't have to
1:15:35 watch.
1:15:36 [Music]
1:15:45 >> City Hill is a quiet town
1:15:49 >> where the oak trees sway.
1:15:54 The whistle of a train echoes through
1:15:57 the day.
1:15:59 [Music]
1:16:02 >> People greet you with n.
1:16:06 Nothing ever seems to change,
1:16:10 but the story still unfolds. Each unique
1:16:14 and strange
1:16:15 [Music]
1:16:24 [Music]
1:16:29 Go home.
1:16:33 [Music]
1:16:39 [Music]
1:16:45 [Applause]
1:16:51 [Applause]
1:16:58 Cedar Hill is a quiet town
1:17:02 where the oak trees sway.
1:17:06 The whistle of the train echoes through
1:17:10 the day.
1:17:12 People
1:17:16 with
1:17:18 >> nothing ever seems to change,
1:17:22 >> but the story still unfolds.
1:17:25 Each unique and strange
1:17:30 grew up
1:17:34 we call
1:17:36 we call home. hopes never to talk to
1:17:40 her.
1:17:46 Longing to be on the inside,
1:17:50 >> you know, it's funny. Sharon Crawford,
1:17:51 that thing, they're starting to sound
1:17:52 like they are singing backwards. I think
1:17:56 a lot of times those five-part AC
1:17:57 capella groups sound like that. They
1:17:59 they do that thing where they go,
1:18:03 that's funny.
1:18:06 So, this is nice. Nice beginning.
1:18:10 I got an idea.
1:18:13 >> If anyone has ideas for styles, ju we'll
1:18:15 just keep playing here. I'm going to I'm
1:18:17 going to try a um let's see. Um
1:18:23 four part country.
1:18:28 Oh, wait. Four part
1:18:31 harmony.
1:18:34 Allah
1:18:37 Crosby
1:18:39 stills Nash and Young.
1:18:42 But let's re let's rewrite it and see if
1:18:44 it rewrites it and pulls out their
1:18:45 names.
1:18:48 Um, yes it did. Okay, cool.
1:18:54 Upbeat PA. Okay, let's do upbeat PA for
1:18:58 producer Brandon. Upbeat poke.
1:19:08 Spritly accordion leads with cris crisp
1:19:11 staccato melody propelled by a punchy
1:19:14 tuba bass.
1:19:16 [Music]
1:19:21 Oh my god. Okay, let's let's let's go
1:19:23 back to this. This a capella sounded
1:19:26 pretty good actually. Let's listen to
1:19:27 this one.
1:19:29 [Music]
1:19:43 Okay, wait.
1:19:46 Brandon just got shown up. Uh, I think
1:19:48 Brandon meant emo pa. I think that's
1:19:51 right. Emo pa.
1:20:00 Okay. So, we're going to get some emo
1:20:02 pula in there. Let's go back to this
1:20:05 whatever we were playing.
1:20:10 [Music]
1:20:24 Cedar Hill is a quiet town
1:20:28 where the oak trees sway.
1:20:39 [Music]
1:20:43 That one's if if you had never heard the
1:20:45 song before, that one would be that
1:20:47 would make you stop and listen. All
1:20:49 right, what do we got here? This one is
1:20:54 So, this is one inspired by Crosby
1:20:56 Steels's Nash and Young.
1:20:59 [Music]
1:21:06 Yeah, for those of you just joining us,
1:21:08 the the exercise we're doing, we took a
1:21:10 song that we created on this channel
1:21:12 about a year and a half ago in Udo.
1:21:16 So V5 is out. So we're playing now in
1:21:19 Sunno V5. Um, generating a bunch of
1:21:22 different covers of the same song just
1:21:25 to just to do it and poke the edges of
1:21:29 this thing and see how good it is. It's
1:21:30 it's it's quite remarkable and it's I
1:21:34 don't even like I haven't even started
1:21:35 getting into creating songs that are not
1:21:38 based on another song, right? This is
1:21:40 this is just we're just doing covers of
1:21:42 this one song. So that's what we're
1:21:44 doing sharing on YouTube. I mean, how do
1:21:46 they do that? It's well, it's
1:21:51 the way all transformer models work is
1:21:55 absolutely [ __ ] insane.
1:21:59 When you write a prompt, it creates
1:22:01 these probab. So, so what happens is
1:22:04 when when music or images get get
1:22:07 embedded into these models, they
1:22:09 actually fragmentize they they they
1:22:12 fragment the components. Basically, they
1:22:15 they they diffuse it into sort of in
1:22:19 this case sonic noise or in the case of
1:22:21 images visual noise using a noise
1:22:24 algorithm. And then there are tags that
1:22:26 tag the music. And when you
1:22:30 give it a prompt, it creates
1:22:32 probabilities where it like smashes all
1:22:34 these noise images together into a
1:22:39 single noisy image and then it doises it
1:22:42 over like 80 steps and it creates an
1:22:44 original tune that's some amalgamation
1:22:47 of like, you know, however many 10,000
1:22:50 songs.
1:22:51 It's absolutely insane how it does it.
1:22:54 Um, and then I like how they're getting
1:22:56 it to produce higher quality sonic
1:23:00 separation and things like that like
1:23:01 that. I don't know. Like what I don't
1:23:03 know is how they're actually training
1:23:05 the models. What like what is what is
1:23:07 the input tagging and are they actually
1:23:10 training this on multiple tracks and
1:23:11 then they're building it in multiple
1:23:13 tracks and then stitching them together?
1:23:15 I that I I don't know any of the the
1:23:17 technical magic behind Sunno in
1:23:21 particular because I know that a feature
1:23:23 that's coming out tied to V5 is what's
1:23:26 called Sunno Studio. Do I have studio by
1:23:28 the way? No. Um where it's it's
1:23:32 basically going to be like a digital
1:23:33 audio recorder where you can, you know,
1:23:36 mix and you know do remastering manually
1:23:40 which is kind of crazy. So,
1:23:45 um,
1:23:51 oh, wait. I just created something
1:23:53 rather than listen to something.
1:23:56 All right. What's this one? This one is
1:24:00 Oh, we didn't listen to this one. This
1:24:02 was the uh Crosby Stills Nash and Young.
1:24:05 [Music]
1:24:12 here.
1:24:17 [Music]
1:24:26 Cedar Hill is a quiet town
1:24:30 where the oak trees sway.
1:24:34 The whistle of a train echo through the
1:24:38 day.
1:24:41 Uh, that's okay. I forgot to I should
1:24:43 have turn turned this into male male
1:24:46 singers for for that style, but that's
1:24:47 okay. Um, the the one thing that I
1:24:50 should talk about here that I'm doing,
1:24:52 so
1:24:54 especially if you're doing commercial
1:24:56 work, you should avoid using other
1:24:58 artists names. So, one of the things you
1:25:00 can do is you can go to like chat GPT
1:25:03 and you can say, "Hey, describe to me
1:25:05 Crosby Steels's Nash and Young Music."
1:25:07 and it'll describe it in abstract and
1:25:10 then you can ask for that style of music
1:25:12 without doing that. What Sunno has added
1:25:15 in here is this magic button, the the
1:25:19 the rewrite it for me button. So Ryan
1:25:22 earlier said do do something like TLC
1:25:25 and I said like TLC and then I hit the
1:25:28 magic button here and it wrote laid-back
1:25:30 90s R&B groove with crisp drum drum
1:25:33 machine beats, bouncy bass, bright.
1:25:35 Right. So, so it's it's not using their
1:25:39 name, but it's pulling out that style.
1:25:41 So, that's a a best practice, especially
1:25:44 if you're doing commercial work, is do
1:25:46 not use known artists names other than
1:25:50 use it as inspiration to get a
1:25:51 description that you can then modify.
1:25:54 All right. Um,
1:25:58 okay. Let's keep let's keep rolling
1:26:00 here. Here's more four-part harmony.
1:26:06 [Music]
1:26:20 [Music]
1:26:24 Oh, you know what would be fun?
1:26:25 Irregulars.
1:26:29 What if we So, you know that there's
1:26:32 that that cosmic
1:26:35 uh contest. What's that contest,
1:26:37 Brandon? The one called that's got 11
1:26:39 labs behind it.
1:26:42 Do you remember? It's like I think it
1:26:44 was Cosmic or Comet Comet. I don't know.
1:26:47 Whatever.
1:26:49 But they have a music video section.
1:26:52 Chroma the Chroma Awards.
1:26:56 What if we did a whole bunch of these
1:26:59 pseudo covers of Weird Mary and then
1:27:02 everybody just grabs a different version
1:27:04 and then we all go off and make music
1:27:06 videos that are all covers of Weird
1:27:08 Mary. We we like we like bomb the Chromo
1:27:11 Awards with covers of Weird Mary. That
1:27:14 would that would get some that would get
1:27:15 some attention. That'd be kind of cool.
1:27:19 So maybe we'll do that. Maybe we'll put
1:27:21 it we'll do it as an irregulars we'll do
1:27:24 it in the irregulars channel and and I
1:27:27 can post a bunch of songs and then
1:27:29 people can claim them if you want. That
1:27:31 could be kind of fun, you know. Get some
1:27:34 get some press for the AI salon. Get
1:27:36 some press for Weird Mary. Get some get
1:27:39 noticed in a contest.
1:27:46 [Music]
1:27:52 City Hill is a quiet town.
1:27:56 [Music]
1:28:03 >> That's a little muddy.
1:28:08 Where the oak trees sway,
1:28:12 the whistle of the train echoes through
1:28:16 the day.
1:28:20 People greet you with a nod.
1:28:24 Nothing ever seems to change.
1:28:29 All right.
1:28:31 What's this one?
1:28:34 This is still folk rock.
1:28:36 [Music]
1:28:39 All right, that sounds the same.
1:28:41 This is spritly accordion leads. This is
1:28:44 uh this is upbeat PA. Here it is.
1:28:47 Producer Brandon, here's your jam.
1:28:55 [Music]
1:29:02 Listen
1:29:05 [Music]
1:29:10 to that tube for baseline.
1:29:14 [Music]
1:29:23 [Music]
1:29:27 Cedar Hill is a quiet town
1:29:31 where the oak trees sway.
1:29:35 The whistle of the train echoes through
1:29:39 the day.
1:29:42 [Music]
1:29:46 >> I I gota I I really hate to admit it,
1:29:49 Brandon. I like I don't want to admit
1:29:51 that you had a good idea. I really
1:29:53 don't. But this is kind of genius.
1:30:00 >> But the story still unfolds.
1:30:03 [Music]
1:30:08 >> She grew up
1:30:10 in the town.
1:30:13 >> Yeah, it screwed up that section.
1:30:18 [Music]
1:30:23 [Music]
1:30:32 [Music]
1:30:37 Cedar Hill is a quiet town where the oak
1:30:41 trees sway.
1:30:43 The whistle of a train.
1:30:48 People greet you.
1:30:53 But the story still unfolds each and
1:30:59 >> all right. This is good.
1:31:01 Um,
1:31:03 it's certainly weird. It's weird, Mary.
1:31:06 It It can be weird. Um,
1:31:09 okay. Let me show you.
1:31:14 Wait, what's this one?
1:31:17 Spritly accordion leads. I think this
1:31:19 was the the one. Was this emo?
1:31:25 I don't know. I don't know if this was
1:31:26 emo or not.
1:31:29 [Music]
1:31:40 Oh, it's got an owl. It's emo. Okay,
1:31:42 good.
1:31:44 [Music]
1:31:59 See the hill is a quiet town
1:32:04 where the oak trees sway.
1:32:07 The whistle of the train echoes through
1:32:11 the day.
1:32:15 >> Unbelievable.
1:32:20 Come on.
1:32:22 Why is that not playing?
1:32:25 [Music]
1:32:36 Oh, good. Hair '8s hair band. That's
1:32:38 good.
1:32:40 '8s hair band. Good. This is This is
1:32:44 really good. Um band. And then I'm gonna
1:32:46 I'm gonna do
1:32:48 a magic magic prompt. Opening with
1:32:51 soaring electric guitar riffs and punchy
1:32:53 gated drums. This 80s hair band layers
1:32:57 powerful vocals over driving bass. Oh,
1:32:59 let's make this um male singer. We'll do
1:33:02 like screaming male singer. Beautiful.
1:33:04 All right, there's that
1:33:07 [Music]
1:33:12 archetype. Okay, I got to sleep. This
1:33:15 has been a night.
1:33:20 Have you done country? Well, I feel like
1:33:22 the song itself is kind of country. I
1:33:25 think it would be like I there there's
1:33:28 like whole other
1:33:30 worlds of genres we haven't touched. And
1:33:33 we don't have to do them all. I mean,
1:33:34 we're we're probably about tapped out on
1:33:36 this at this point, but I want to hear
1:33:38 the things that we've built in the style
1:33:40 of Young Blood. Oh, that's interesting.
1:33:42 Let's try that.
1:33:46 Um,
1:33:49 in style of Young Blood,
1:33:56 I don't know Young Bloodood's music
1:33:58 enough to know. High high energy trap
1:34:00 influence punk rock track. All right,
1:34:03 we'll try that. Okay, that's going to be
1:34:05 the last one. I promise.
1:34:09 [Music]
1:34:12 through the town
1:34:14 [Music]
1:34:20 hill
1:34:22 [Music]
1:34:24 where the oak trees sway.
1:34:28 >> I love the idea. I'm loving the idea
1:34:30 more and more that we we do an AI salon
1:34:34 um weird Mary bomb on on the comp on the
1:34:38 competition.
1:34:40 They're like, "Why does this song keep
1:34:42 coming up?
1:34:48 [Music]
1:34:50 [Applause]
1:34:51 [Music]
1:34:56 Wait, this one's emo pa. Emo poka punk.
1:35:04 [Music]
1:35:19 Cedar Hill is a quiet town
1:35:22 where the oak trees sway.
1:35:26 The whistle of the train echoes through
1:35:29 the day.
1:35:30 [Music]
1:35:32 People greet you with a nod and nothing
1:35:36 ever seems to change.
1:35:39 But the story still unfolds.
1:35:43 [Music]
1:35:57 That's crazy.
1:35:59 Punchy accordians.
1:36:03 That Easter B.
1:36:07 [Music]
1:36:11 >> Wait, that sounded like a crazy That
1:36:13 sounded like weird Mary in the nervous
1:36:15 hospital.
1:36:19 That is all right.
1:36:23 [Music]
1:36:25 >> Silverf Fox AI salon gets banned from
1:36:27 future AI contests. I would love that
1:36:30 actually. I mean, you know, no PR is bad
1:36:32 PR, right?
1:36:37 Oh, Buork. All right. Uh, it wouldn't
1:36:39 rewrite in the style of Lincoln Park for
1:36:41 some reason. Um, sounds like Buork.
1:36:46 Although that sounded that last one kind
1:36:48 of sounded like Borc, which is I assume
1:36:50 that came from.
1:36:52 Uh, all of a sudden the uh the magic
1:36:54 button's not working anymore. So, I
1:36:56 think it's a sign that we should we
1:36:59 should stop.
1:37:05 [Music]
1:37:30 I think my browser's dying.
1:37:34 [Music]
1:37:46 >> This is R&B neo soul.
1:37:51 [Music]
1:37:54 Hill.
1:37:55 [Music]
1:38:00 >> Yeah. I mean that one. Imagine the video
1:38:02 for this one, right? Like really sultry
1:38:04 in a jazz club.
1:38:07 [Music]
1:38:16 >> People greet you with a n
1:38:20 Nothing ever seems to change,
1:38:24 but the story still unfolds.
1:38:26 Each unique and strange
1:38:30 [Music]
1:38:33 tough
1:38:36 home.
1:38:37 >> That was cool. All right, what do we got
1:38:40 here? Laidback 90s. Same thing.
1:38:45 [Music]
1:39:04 All right, this one is Those are all the
1:39:06 same. I'm going to skip that one. All
1:39:08 right, soaring electric guitars. This is
1:39:11 um this was 80s hairband.
1:39:16 Actually,
1:39:19 [Music]
1:39:43 this is fascinating Because
1:39:46 if there's one thing '8s metal hair
1:39:48 bands didn't do, it was like a lot a lot
1:39:50 of minor chords with weird chord
1:39:52 progressions. And it sounds like the
1:39:54 model can't quite figure this out. It it
1:39:57 like it's it's going it's janking up a
1:39:59 little bit here. The diffusion model
1:40:04 [Music]
1:40:06 [Applause]
1:40:07 [Music]
1:40:30 No, that one's awful.
1:40:38 [Music]
1:40:39 [Applause]
1:40:42 [Music]
1:40:54 [Music]
1:41:00 [Music]
1:41:08 That's awful. That's not listenable.
1:41:10 All right. This is high energy infused
1:41:13 punk rock.
1:41:16 [Music]
1:41:47 Yeah.
1:41:57 Hey. Hey.
1:41:59 [Music]
1:42:02 See the hill as a quiet town.
1:42:07 [Music]
1:42:11 The whistle of the train.
1:42:15 That's interesting. But eh.
1:42:18 All right. Uh, one more of those.
1:42:29 [Music]
1:42:31 [Applause]
1:42:31 [Music]
1:42:36 That's awful. Okay. All right. Upbeat
1:42:39 pop production. I think this is TLC. Was
1:42:41 this TLC or do we already do TLC?
1:42:46 [Music]
1:42:52 >> Oh, this is KidsBop, right? Pop, dance,
1:42:54 pop.
1:42:56 [Music]
1:43:05 Whenever
1:43:08 [Music]
1:43:22 I think I I think I'm officially sick of
1:43:24 Weird Mary at this point.
1:43:29 You did it. Yay. You hate that song you
1:43:31 loved.
1:43:36 [Laughter]
1:43:40 Oh man, you should post the original
1:43:44 remaster in a regulars for us to remix.
1:43:47 Yeah, I could do that. I think I can
1:43:49 just publish I can just publish one of
1:43:52 these two or just publish a pile of
1:43:54 them. I'll I'll figure out how to do
1:43:56 that. Um, okay. Let's let's be done
1:43:59 here. Um
1:44:04 um today's Thursday, so tomorrow we've
1:44:07 got office hours at 11:00 am on
1:44:09 LinkedIn. Just go to my LinkedIn uh
1:44:12 profile and you can see events there.
1:44:14 And then um tomorrow night, Friday night
1:44:17 date night. It is not first Friday, so I
1:44:20 don't think I've got gallery hopping
1:44:22 duties, but you never know. So should be
1:44:24 normal time. Um,
1:44:28 if you're in the salon and haven't
1:44:30 joined the mastermind, you should do
1:44:33 that because there's a whole bunch of
1:44:34 good stuff going on in there. We're
1:44:36 doing a thing with Cindy [ __ ] right now
1:44:38 with um with uh
1:44:41 uh pro the prototyping club doing micro
1:44:44 courses. We're on our fourth week next
1:44:46 week and then we're probably going to
1:44:47 start something new. So, there's all
1:44:49 sorts of cool stuff going on in there.
1:44:50 We've got the uh life hacks club which
1:44:54 the the Wednesday after our salon uh
1:44:58 presents meeting which is October 7th.
1:45:00 The life hacks club is doing a thing
1:45:02 there. So you should join that and uh
1:45:04 yeah that's that. And go to uh pop up
1:45:08 the banner for are you ready for or yeah
1:45:12 are you ready for a.com
1:45:15 if you're trying to figure out this AI
1:45:17 stuff. I know some people were in here
1:45:18 earlier going like how do I start? what
1:45:20 do I do? Um, go check out this course
1:45:22 that we put together um over the over
1:45:25 the course of the year called Are You
1:45:27 Ready for AI? It's it's an AI readiness
1:45:29 training program and it's tool agnostic.
1:45:32 It's it doesn't have anything to do with
1:45:34 specific tools because they're all
1:45:35 changing. So, this is how do you shift
1:45:38 into the mindset and how do you kind of
1:45:40 professionalize yourself with AI? That's
1:45:42 what that course is all about. All
1:45:44 right, G. And with that, um, I'm gonna
1:45:49 drag my 80s metal hair band out of here
1:45:55 and and, uh, try to get Weird Mary out
1:45:59 of my head, or at least the beginning
1:46:02 minute and a half of it.
1:46:08 That's not going to go well. But anyway,
1:46:10 uh but I'll probably, you know what I
1:46:12 think I'm going to do is I'm going to go
1:46:13 back through now that I've got this big
1:46:15 pile of those things. I'm gonna go
1:46:16 listen to them all again with this idea
1:46:19 of um like making an album and making a
1:46:23 a playlist of things that we could make
1:46:25 videos from. We could all sort of opt in
1:46:28 if you want to if you want to make a
1:46:29 video for a contest. Um we should
1:46:33 probably go look at the rules of the
1:46:35 contest and make sure we're using the
1:46:36 right tools. It's like whatever you do,
1:46:39 don't use sunno. Oops.
1:46:44 All right, everybody. Have a good night.
1:46:46 Hope you had fun tonight. Bye.