12/16/2025 - Transforming a Bad Song Idea Into a Great Track Using Suno AI

Description

Chapters

Transcript

0:00 [ __ ] supper. Champy, come here.
0:05 [whistles]
0:08 [music]
0:16 [music]
0:18 [singing]
0:26 >> [music]
0:39 [music]
0:48 [music]
0:50 >> Champy, what do you want to do? What do
0:51 you want to say?
0:54 [music]
1:01 What? What are you barking for?
1:04 What? We don't rap around here.
1:07 [music]
1:09 [laughter]
1:12 [music] You all right there?
1:19 [singing]
1:21 >> [music]
1:24 >> Is this place I can rest my forest?
1:31 Gather my thoughts in sweet silence.
1:38 The bus is [music] place where the
1:40 feelings are dead.
1:44 Running over [music] exposure to
1:46 violence.
1:49 This is place I can slowly face. The
1:52 only one I truly can know. [music]
1:56 These are tears from a long time ago.
1:59 You're done. You're done singing.
2:03 You You got half half a verse and you
2:06 stopped. You just stopped.
2:09 This is not in your contract. Your cont
2:12 You've got obligations.
2:14 You think You think cheese grows on
2:16 trees.
2:19 No, the chicken does. I know. I know. I
2:21 know that. Smart ass.
2:26 I don't have a singing dog tonight. I
2:28 don't know what to tell you people.
2:29 Let's see if an A minor gets him.
2:34 [music]
2:36 No, Champ.
2:39 [music]
2:44 >> [music]
2:50 >> Well, I heard there was
2:53 secret
2:56 David played and it plays a lot.
3:00 You don't really care [singing] for
3:03 music, do you?
3:07 [music]
3:08 Goes like this.
3:09 >> [singing]
3:09 >> The fourth, the fifth, the monof
3:13 [music]
3:17 king composing.
3:19 Hallelujah.
3:22 We got no dog tonight. Hallelujah.
3:27 Hallelujah.
3:29 [music]
3:31 Hallelu.
3:34 [singing]
3:35 Hallelu.
3:39 Oh yeah.
3:42 [music]
3:47 I guess we're going to have to deal with
3:48 all the crap that was released.
3:51 [laughter]
3:53 Those those Frontier model companies
3:55 went apeshit today. We got a 100
3:57 seconds. We got literally 100 seconds
3:59 from Champ. We got We got Let's see.
4:03 Here's his place. I [music] can rest my
4:05 forehead. He was like, "Oh,
4:08 gather my [singing] thoughts in [music]
4:10 silence."
4:14 [singing] There's this place where the
4:16 feelings are dead [music]
4:20 from an overexposure to violence. Right
4:23 there, he just walked away.
4:27 And this is a place [music] I could
4:29 slowly face. The only one I truly can
4:32 know.
4:34 These are tears from a
4:35 [singing and music] long time ago. It's
4:37 all silent. This all dog free. Got these
4:40 tears from a long time ago.
4:43 I need to cry 30 years
4:45 [singing and music] or so. These are
4:47 tears from a long time.
4:51 No champ. [music]
4:53 [singing]
4:59 He is not singing. He's not having it.
5:02 Got nothing. We got no dog tonight. I
5:04 don't know what to tell you, man. They
5:06 say never work with pets or kids.
5:09 [laughter]
5:11 Oh man, what's happening everybody?
5:16 So, we had a a lot happened today in the
5:19 world of AI. We got a new image model
5:20 from chat GPT
5:24 that's in the neighborhood of Nano
5:25 Banana. I from what I can gather
5:29 nano banana is better at remember when I
5:32 said that nanobanana is not an image
5:34 generating tool that it's it's a visual
5:36 reasoning model. Um I think chat GPT is
5:41 still an image tool. I don't think it's
5:43 as good at reasoning. Now we can
5:45 experiment with that tonight. Like we
5:46 can turn on we can turn off automatic um
5:50 model picking and just pick the highest
5:52 one and see if it gets better.
5:55 You you don't have enough to howl while
5:58 you play guitar. Oh, wait. [music] Oh,
6:01 you don't pay me enough to howl while
6:02 you play guitar. I don't know what's
6:04 going on with him tonight. He's just
6:07 he's he's uh [music]
6:12 Oh, let me flip that down there. Let me
6:15 flip this over here. Go the black bar.
6:18 Pull this up here. Like this. Like that.
6:22 Do that a little like that. Do this a
6:24 little like this. Do that. Do this.
6:28 Everybody calm down. Everybody
6:29 Everyone's going to be just fine.
6:32 [music]
6:40 [music]
6:44 Image isn't on your phone. I think the
6:46 phone one. It's just they they don't
6:48 have the extra what you call it. They
6:51 don't have the extra bells and whistles
6:53 like the image pane, the the image page,
6:59 but I think the model's still there. If
7:01 you have it in your web account, you
7:02 should have it on the phone.
7:06 [music]
7:11 [music]
7:18 >> [music]
7:22 [music]
7:29 [music]
7:31 >> Um, I can't say who it is yet until
7:33 we're final final finalized, but we got
7:36 someone really cool for AI Festivus.
7:40 It's going to come talk to us. Um,
7:43 really really exciting. a a a relatively
7:47 big name in the world of AI. Um, so I
7:51 don't want to jinx it, so I'm not gonna
7:53 say it, but really good. We thought we
7:56 had lost them and now we think they
7:57 might be back. So we will see. Um,
8:03 so that's super cool. Welcome everybody.
8:06 It's Tuesday night. Um, isn't it? Yeah,
8:10 it's the 16th. Next week is Festivus.
8:15 >> [music]
8:20 [music]
8:28 [music]
8:33 [music]
8:37 >> It's not Sam. Or is it?
8:41 [music]
8:47 >> [music]
8:53 [music]
8:59 [music]
9:04 [music]
9:10 [music]
9:17 [music]
9:18 >> Um, so we have a new image model. Um,
9:22 there's some new [ __ ] going on with
9:23 Lovable.
9:26 There's a bunch of new stuff with
9:27 Google. Google's just dropping all sorts
9:29 of stuff. I feel like like uh Google
9:32 right now is just kind of carpet bombing
9:33 the the uh the competitive play field.
9:36 And you know, it's funny. It it these
9:39 things these es and flows of corporate
9:41 power are really interesting. Like when
9:44 Chat GPT first came out, like they were
9:48 mopping the floor with anyone else for a
9:51 year and a half. And especially Google,
9:53 Google just they really dropped the ball
9:55 on the generative AI thing. And then
9:57 man, when they started to turn that ship
9:59 around, especially this year, 2025,
10:01 Google woke up. And it, you know, Google
10:04 probably woke up in 2022,
10:07 right? When they knew that Open AI was
10:10 on to something and so they started
10:13 putting energy toward all this
10:15 generative AI stuff,
10:18 they [ __ ] invented it, right? Uh, and
10:21 it took them two years to [music]
10:24 two years to turn the ship. But holy
10:26 [ __ ] man. When you have whatever
10:28 they've got, 400,000 people, 200,000
10:30 people, it's a lot of people.
10:32 and all the money
10:34 they can put their best people on stuff.
10:36 Amazing.
10:39 Evening. Been out of the loop for a bit.
10:41 Welcome back, Narrator Forever. What do
10:43 you think is the best creative writing
10:45 one these days? I'll tell you what,
10:47 Narrative Forever. Um
10:50 I for me, Chat GPT40
10:54 was the best um creative writing tool
10:57 and and they're probably going to
10:58 deprecate it. I think that the entire
11:01 five series, 5.1 got better and 5.2 is
11:05 not bad, but they just they're just kind
11:08 of cold and they're just not not good
11:11 writers. Um, apparently GPT4.5
11:16 was a really good creative writing tool,
11:18 but I never used it. I was I was always
11:21 happier with 40. Um, a lot of people
11:25 swear by Claude. So apparently Claude is
11:28 really good at creative writing. I It's
11:31 one of those funny things where you kind
11:33 of imprint on, you know, like like baby
11:35 ducklings,
11:37 [laughter] you kind of imprint on on on
11:42 a tool, right? And so I've occasionally
11:46 tried Claude, but I just I haven't gone
11:48 over and done it. I can tell you that um
11:53 Gemini,
11:56 I think it was 2.5 Pro.
12:00 Gemini 2.5 Pro and they're up to three
12:03 now. Gemini 3 um
12:08 was surprisingly good at creative
12:10 writing.
12:11 And I say surprisingly because Google is
12:14 so engineering centric. They're so
12:16 they're so focused on STEM and not
12:19 steam.
12:20 um that I didn't expect Gemini to be
12:23 good at all ever at creative writing and
12:26 2.5 Pro was really good at it and
12:29 three's pretty good at it. So, but
12:32 here's [clears throat] the thing.
12:36 If you're a decent communicator, like if
12:38 if you're a decent
12:41 collaborator, say working with another
12:43 human writing partner, um any of these
12:47 tools quite frankly are so good now,
12:50 they've got so much training data in
12:52 them. They basically have everything
12:54 that's ever been written that's been put
12:55 on the internet, which includes all of
12:57 the worst writing in the world, but it
12:59 also includes all of the best writing in
13:01 the world. Right? So, if you're good at
13:03 prompting and sort of narrowing the
13:04 context window, any of these models can
13:07 be incredible at writing. Like one of my
13:10 tricks with writing, if I'm doing
13:11 serious creative writing where I want I
13:14 want storytelling and I want structure
13:16 and and [ __ ] like that is I'll do a
13:19 whole session of have have chat GPT tell
13:24 me about five different um storytelling
13:27 frameworks and then I'll pick two of
13:29 them just whatever floats my boat that
13:32 day and I'll say create a custom
13:34 storytelling framework that combines
13:36 these two you know like like Joseph
13:40 Campbell's Heroes Journey and and Save
13:43 the Cat and combine those into some new
13:45 storytelling framework and then I'll say
13:48 I have an idea for a story about XYZ
13:52 you know give me 20 ideas that would
13:55 work well in that framework and then
13:56 it'll do that and then you know
13:59 so if if you do that kind of stuff where
14:01 you prime the pump um you can probably
14:03 get some decent stuff done. Joy party
14:06 okay GPT 5.2 do is doing a pretty good
14:09 job with my novel. Okay, great. Does
14:12 take a lot of direction though. Yeah,
14:13 that's that's been my experience is it
14:16 it it tends to [ __ ] the bed first where
14:19 you're just like it's just not good
14:20 writing. Um although I will say that it
14:25 was GPT5,
14:27 not even 5.1.
14:30 It was GPT5 that wrote the lyrics for
14:33 Before the Lights Come On, and those
14:34 were banger song lyrics. But I gave it a
14:38 lot of context going in. I like Claude's
14:41 writing, but the context cut off makes
14:44 it absolutely infuriating. I I'm hearing
14:46 so many people complain about fraud
14:49 claude right now because it they just
14:51 they just run out of tokens and they're
14:52 like, "Sorry, you're out." Which sucks.
14:55 Pseudoite is a good rapper app that
14:57 gives you a huge amount of background
14:58 prompts. That's interesting. All right,
15:00 that's good. So, listen to Archetypal.
15:03 He's he's working on a novel. Been
15:04 working on it for a while. Robbie, when
15:06 will chatbt replace lawyers? You would
15:09 think that law would be easy for it.
15:11 Well, Robbie, it's a that's a it's a
15:13 really interesting point you bring up.
15:14 So So here's here's where I think that
15:17 we go.
15:20 I think that law clerks and and junior
15:23 lawyers, people that do the the research
15:26 and the initial drafting of, you know,
15:30 law stuff, positions and cases and
15:33 things like that. I think they get a lot
15:37 of that stuff gets automated. So, I
15:38 think law firms are going to get more
15:40 efficient.
15:42 You don't hire a lawyer because they
15:45 know the books. You hire a lawyer
15:48 because you trust that they've been in
15:51 the world long enough and they are smart
15:53 enough and they are influential enough
15:56 and they are unflapable enough
15:59 to get you the [ __ ] you need. Right? So,
16:06 like like for example,
16:09 right now I'm working on a musical and
16:11 it's starting to get serious. We've got
16:12 some producers that are interested and
16:13 we've got to start thinking about
16:15 contracting and things like that. I
16:17 could ab well I actually have I could
16:20 absolutely use Chat GPT put it into
16:23 lawyer mode, right? You're you're a
16:25 famous Broadway lawyer, you're a
16:27 producer, you're a this, you're a that.
16:29 Put it into that mode,
16:32 which I did. And then I had it help me
16:36 understand some stuff. And then I had it
16:38 put together a proposal for me. And I
16:40 the proposal didn't feel right. And I'm
16:43 like, can you go look and see if this is
16:45 standard practice? It's like, oh, no,
16:47 it's not standard practice. I'm like,
16:48 oh, [ __ ] Okay. So, I went had it go off
16:51 and do deep research. And it went off
16:52 and did deep research and it came back.
16:54 And so, I've now got this very
16:56 sophisticated understanding of how
16:57 Broadway financing works and how um
17:00 Broadway contracts work. and you know,
17:03 if you're going to bring in a producer
17:04 early in the project, here's how you can
17:06 compensate them and motivate them but
17:07 not give away the farm and all that sort
17:09 of stuff, right?
17:12 I've also got in my back pocket a lawyer
17:15 who's been working on Broadway for, you
17:18 know, 35, 40 years.
17:22 They're like a thousand bucks an hour.
17:25 They're going to require a retainer, so
17:27 that's probably, I would think, a
17:29 minimum of 10 hours, right? So that's
17:33 $10,000. It probably be $7500 is my
17:36 guess
17:38 that I have to fork out upfront
17:41 to be able to work with that lawyer,
17:43 right?
17:46 I don't want to rely on chat GPT cuz I'm
17:50 [ __ ] clueless about this. This lawyer
17:52 is going to be able to look at what Chat
17:54 GPT did and immediately go yes, yes, no,
17:57 no, yes, yes, maybe. That's a choice,
18:01 right? [laughter] They're gonna be able
18:02 to look at it and just go, um, here are
18:04 the six things that Chat GPT got wrong.
18:06 Here's how we fix them. Right? So, what
18:09 I think Chat GBT is going to do is it's
18:12 going to save me instead of me having to
18:14 pay for 4 hours
18:17 at $750 an hour to him have an
18:20 associate,
18:22 you know, write up a base contract and
18:24 then we discuss it. I'll do the base
18:27 contract in chat GPT and I'll give it to
18:30 him and have him review it. But that
18:32 trust that trust of of you know someone
18:35 who's been in the world, someone who's
18:37 cut deals, someone who's licensed
18:39 properties, someone who's done contract
18:41 law,
18:43 I think it's about the trust more than
18:44 the than the skill in in a lot of cases.
18:48 I think consulting, that's the way where
18:50 where consulting is going to go is I
18:52 think you're going to have smaller
18:54 consulting firms that are using AI to do
18:58 a bulk of the thinking and a bulk of the
19:00 writing, but you're going to have
19:02 partners, you know, 10 or 20 partners
19:05 that are all highly trusted, highly
19:10 accomplished,
19:11 baller kind of people because you're not
19:14 going to hire you're going to hire them
19:16 because of the people and their
19:18 relationships and their knowledge, their
19:20 real real world knowledge. So anyway,
19:23 [music]
19:28 did you just call Claude fraud? I
19:30 probably did. Um, unless your lawyer has
19:33 an inflated ego
19:36 and downgrades GPT info, that's
19:38 possible. Um, they need humans for
19:40 nuance. That won't be there that long.
19:43 Like I I really do think I really do
19:45 think that the that the the large
19:47 language models, make make no mistake,
19:51 I think we'll start to see in 2026
19:54 um the systems getting so good that
19:59 you know
20:01 they can do they can do full-on jobs.
20:05 Um it might be the end of 2026, it might
20:07 be 2027, but you know, within the next
20:10 year and a half. Um we're going to see
20:12 systems that can absolutely do the job.
20:14 It you won't even need it for nuance,
20:16 Corey.
20:17 But I but humans are not going to adapt
20:20 to it that quick, right? Humans are not
20:23 going to trust it because we're going to
20:24 have I Let's say this. By the time large
20:27 language models get good enough to
20:29 trust, we'll have had five years of not
20:32 being able to trust them, right? So,
20:35 you're going to have a mix of people
20:36 who've never used AI before and a pix
20:39 and a mix of people who've used it and
20:41 understand that it hallucinates and it's
20:43 shitty and it needs humans involved,
20:46 right? So, we're going to have five
20:47 years of that
20:49 and then the tools are going to get good
20:51 enough that you could actually use them,
20:53 you can actually trust them, right? But
20:55 we're not going to trust them for
20:57 probably another five years, maybe 10,
20:59 because we're idiots, right? We're
21:02 idiots with ego. We're monkeys with
21:04 language,
21:06 right? Not gonna take my brain away from
21:08 me. [laughter]
21:18 Elon said AGI is 2026. I know, but
21:20 Elon's been Elon said cars would be
21:23 driving by, I think, 2023 or four or
21:26 two.
21:28 Um, and they're just now starting to
21:29 drive. So he was about four years late
21:31 on that one. I think that we will see
21:34 hints of AGI in 2026. I think we will
21:36 see the beginnings of systems
21:39 that maybe instead of getting us to 80%
21:42 good where they are right now, maybe
21:45 they'll get us to 95% good. And if if
21:48 you've got systems that are getting you
21:49 to 95% of what you want, um you're
21:52 pretty [ __ ] close to AGI. Right now
21:55 we're just not there. Right. Right now
21:57 it's still 80%. Like right now, we can
21:59 still have human ego to go, "Yeah, well,
22:02 yeah, uh AI is smart, but uh still needs
22:05 old Mr. Charlie here to uh make sure
22:08 everything's okay." Okay. All right,
22:11 little Quinn. Quinn, my little AI
22:14 girlfriend. Sorry, you're such a dumb
22:16 dumb, [laughter]
22:17 right? We can do that for I don't know,
22:20 two years. We can we can still be
22:23 obnoxious idiots. And then I think And
22:26 then I think at some point we're going
22:27 to go, "Yeah, yeah, it it can pretty
22:29 much do my [ __ ] now.
22:35 Technology moves so much faster than us
22:37 humans. The humans slow us down." Okay,
22:40 Cindy Coon's in the house. What's
22:42 happening, Cindy [ __ ]
22:51 Um, I I want to I want to talk about
22:53 some of the new stuff, but let's see how
22:55 many folks we got here. We got a few
22:57 folks here. Um, I want to show you I
23:00 found I was looking for I've got a song
23:03 I wrote about my grandmother that I
23:05 absolutely loved and I can't remember
23:08 the chords to it. I never wrote the
23:10 chords down to it because I played it so
23:12 much. I'm like, I'll never forget these.
23:14 And I forgot them. And so I was looking
23:17 back through some old lyrics I've got.
23:19 I've got the words written down of a
23:21 bunch of songs and no chords to them.
23:24 And one I figured out tonight. Um and
23:27 then I found a recording of a song that
23:30 is so bad. Like the the recording and
23:34 singing is so bad that I got like flop
23:37 sweats listening to it because I thought
23:40 I thought, "Oh, maybe I'll do something
23:42 with this on my AI learning lab." and
23:44 [laughter] and I listened to it and I
23:46 got flop sweats because I'm like, "Oh my
23:48 god, if they hear this, it's it's
23:49 embarrassingly bad, right?"
23:52 So, I thought what might be fun, you
23:55 know, for you is to is to watch me
23:57 squirm. I'll play a little of it. I'm
23:58 not going to play much of it because
23:59 it's so bad. Um, but
24:04 can we take a song that maybe has an
24:07 idea in it, but is so bad and turn it
24:10 into something good? I thought that
24:12 might be kind of fun.
24:14 No cassettes of that song for reference,
24:16 Kyle. I don't think I ever recorded it,
24:18 Corey. I don't think I ever recorded it.
24:21 Um, the only thing I could think of is I
24:24 wonder if I did it in Oh, there was a
24:27 there was a Wasn't there a
24:31 an iOS
24:35 um
24:39 memo memo like music memo app?
24:44 Yeah, music memos.
24:51 Did I
24:53 my idea?
25:01 No,
25:03 nothing there.
25:06 And then there's voice memos.
25:18 No.
25:21 I got nothing.
25:24 I don't know. I don't know. I don't I
25:28 can't remember how it goes. I can't
25:29 remember the chords.
25:31 Um, but it was a heartbreaking song. It
25:34 was really good.
25:36 Sounds like Kyle is minutes away from
25:39 sitting in the ick. Yeah, exactly. Did
25:42 you put it in Garage Band? I don't think
25:44 I ever recorded it, Danielle. The one
25:46 I'm about to play you was the one
25:47 sitting in Garage Band. It was actually
25:49 sitting in a file called um
25:53 um Love Shot Eyes, which I did I did a
25:55 cover of Love Shot Eyes in Sunno here.
25:58 But I found this file and it's called
26:01 Love Shot Eyes, but it has this other
26:02 song in it called the clouds,
26:05 which [laughter] is this really sappy
26:08 song about watching the clouds drift by.
26:11 [laughter]
26:12 It's so [ __ ] bad. It's so [ __ ]
26:14 bad. [laughter]
26:18 I I looked in the GarageBand file. I
26:20 couldn't find it. I mean, it might be in
26:21 there. I didn't I didn't have enough
26:23 time. But anyway, all right. So, I'm
26:25 gonna I'm gonna let you listen to this
26:27 song. This is [ __ ] humiliating.
26:29 [laughter]
26:33 Oh, Christ. What am I doing? What am I
26:36 doing, people?
26:38 Um,
26:40 [laughter]
26:41 okay.
26:46 Love shot or bloodshot? Love shot.
26:49 Love shot eyes. Um, here I'll let you
26:52 hear that one because that one that one
26:54 I did a cover of. It's pretty good.
26:58 Uh, let's see. So,
27:01 [clears throat and cough]
27:09 love shot.
27:11 Love shot eyes. Let's see. Let's filter
27:15 by public.
27:17 None.
27:20 Liked. Focus. Publish by liked. Is this
27:24 the one?
27:30 Oh, wait.
27:43 >> [clears throat]
27:47 >> down.
27:54 [music]
27:57 >> I think that's it. [music]
28:05 >> No, it's not. Hang on. I'll give you
28:07 love shot eyes. Hold, please.
28:10 Love shot eyes.
28:19 [clears throat]
28:22 Huh?
28:30 Oh, by the way, if you're here last
28:31 night, remember I I looked for like 15
28:33 minutes in my photos to find those old
28:36 black and white photos and I said I bet
28:40 it's right above where I am. It was
28:42 literally two scrolls away. [laughter]
28:46 I found them and they're pretty cool.
28:48 I'll show them to you.
29:10 Is this it?
29:31 I thought I did Love Shut Eyes. Maybe I
29:34 just wasn't happy with it. Maybe I
29:35 didn't publish it. I don't remember.
29:37 Whatever. Whatever.
29:40 All right, let's go back to Sunna.
29:45 Okay.
29:47 All right. We're going to share this
29:49 horrible song. All right, prepare.
29:51 [laughter]
29:53 It's going to suck and I'm not going to
29:55 play much of it because it's
29:56 embarrassing. So, this is so the here's
29:59 the experiment. Can you take something
30:02 from your life that was maybe had the
30:05 seed of an idea in it but is horrible
30:08 and use AI to like sprout that seed and
30:12 turn it into something not horrible?
30:14 That's the experiment.
30:16 [laughter]
30:17 So, here's the horrible.
30:20 Oh my god, this sucks.
30:25 [music]
30:31 [music]
30:32 The clouds [singing] were crawling to
30:35 the east. Hey,
30:39 but they were screaming to be seen.
30:48 The Rockies held their ground.
30:55 Wind passed by [singing] without a
30:58 sound.
31:00 Oh,
31:02 the clouds.
31:05 The oneline chorus. Oh, the clouds.
31:09 [laughter]
31:10 Okay, that's horrible. I I will give you
31:13 that.
31:14 >> [laughter]
31:15 >> The Beatles. It's Beatlesike. Yeah, it's
31:17 very Beatlesike. Uh, minus the
31:20 sophisticated chord [laughter]
31:22 progressions and decent vocals and
31:25 melody. All right. [laughter]
31:31 You ever do something where you're just
31:32 like, why did I just do that? That's
31:34 [clears throat] how I'm feeling right
31:36 now. [laughter]
31:39 Let me have some coffee so I can get a
31:41 little more jacked up on Panic.
31:44 I like it. Mary Mary likes it. Well,
31:46 let's see if I like I did two pseudo
31:48 covers of it. So, let's we'll we'll
31:50 start with the pseudo covers I did just
31:53 as let's see what we can do and then
31:55 maybe we can actually My ears are
31:56 bleeding. [laughter]
31:58 Maybe we can turn it into something
32:00 interesting. Robbie, is the new chat GPT
32:02 image model any good? It's quite good,
32:04 Robbie. And I'm going to play with it in
32:05 a bit. So, this won't just be about my
32:07 shitty songs. We're we're gonna go play
32:08 with what I what I thought would be fun
32:10 tonight is we'll do sideby-side
32:12 comparison with Nano Banana and uh and
32:16 and ChatgPT image. It's quite good.
32:18 Here's the thing where I think what I
32:20 think Chat GPT image is really good at
32:24 is you know how you take a picture of
32:26 yourself and you put it in these things
32:28 and you say, you know, make me in, you
32:30 know, whatever a hotel outfit or
32:33 whatever. Um,
32:36 every time I've ever done that, it looks
32:38 kind of like me, but it looks like my
32:40 cousin. It doesn't look like me to me,
32:42 right? It might look like me to someone
32:44 else, but it just doesn't look like me.
32:46 This thing from chat GPT, it looked like
32:49 me. Like the couple of things I've done,
32:52 I'm like, "Yeah, that looks like me."
32:54 And I've never had an image thing do
32:56 that. Um, so that's kind of fascinating.
32:59 Um, I don't think it's as good as the
33:00 visual reasoning stuff as Nano Banana.
33:03 So, like things like flowcharts and and
33:06 you know, infographics and things like
33:08 that, I have a feeling Nano Banana's
33:10 better at. Um, anyway, back to my shitty
33:13 song, The Clouds. Okay, so here's the
33:14 first uh the first cover it did.
33:25 [music]
33:35 >> [music]
33:38 >> Yeah, it does want to be a folk song. I
33:40 went Americana 12 string complex vocal
33:43 harmonies. That's what this is. Let's
33:44 see it then.
33:49 >> [music]
33:49 >> The clouds were crawling to the east,
33:54 [music]
33:56 but they were screaming to be seen.
34:02 [music]
34:05 The Rockies [singing] held their ground.
34:10 [music]
34:12 Wind passed by without a sound. [music]
34:17 Oh, the clouds.
34:21 [music]
34:27 [music] The sun was setting [singing] in
34:31 the west.
34:36 Those clouds in their beds.
34:40 >> A great lyric.
34:42 [music]
34:43 Rocks looked on without [singing] a
34:46 care.
34:51 [music] Grasses swayed [singing]
34:53 and we all stared
34:56 at the clouds.
35:00 [singing]
35:01 [music]
35:06 The clouds they covered up.
35:10 >> All right. It's not bad.
35:14 >> [music]
35:22 [music]
35:28 >> The clouds were crawling to the east.
35:32 Well,
35:32 >> this is cool.
35:36 But they were screaming to [singing]
35:39 be seen.
35:44 >> [music]
35:45 >> The Rockies held their ground.
35:52 Wind passed by without a sound.
35:57 Oh, the clouds.
36:03 That one's not bad. All right, let's do
36:06 uh let's do um
36:09 [sighs]
36:13 some kind of folk. Some kind of trippy
36:15 folk.
36:17 Um
36:21 mythic. No, mythic will get too [ __ ]
36:24 weird and heavy metal, won't it? Um give
36:26 me Who's got ideas? Come on. You're all
36:30 psychedelic folk. Okay, that's good.
36:32 actually psychedelic folk. It's got um
36:36 it's got some um
36:40 like velvet underground Lou Reed Turkish
36:43 folk. That's funny. Um
36:47 psychedelic folk complex vocal
36:50 harmonies. Let's do And you know what
36:52 else could help is maybe this is meant
36:56 to be sung by a female.
36:59 psychedelic f complex vocal harmonies
37:01 offbeat percussion. I I like the offbeat
37:03 percussion. Um and then let's because
37:07 Corey Sandler's here. Let's throw in a
37:09 little cello.
37:10 [laughter]
37:11 Trap folk. Okay, good. Trap folk. I like
37:13 that. Wait. Oh, minimalist
37:15 orchestration. I love that. Oh, okay.
37:19 We're going to do that instead of cello.
37:21 Minimalist
37:23 orchestration.
37:25 Okay, this is going to be good. I think
37:27 this is actually going to make a really
37:28 cool song.
37:32 [clears throat]
37:34 Oh, and then let me do one where I go
37:37 audio. Let's pull the audio influence
37:39 down.
37:41 I'm going to pull audio influence down
37:43 to 15.
37:46 And then I'm going to pull weirdness up
37:48 to 50. And I'm going to pull style
37:50 influence up to 70.
37:53 Then we'll do another batch. All right.
37:56 So, this is psychedelic folk, complex
37:59 vocal harmonies, offbeat percussion,
38:02 minimalist orchestration.
38:04 All right, here you go. Corey Sandler.
38:07 Here's your weirdass song. Kyle, weird
38:10 team. We are. I want to create my own
38:13 pop star NPC. [ __ ] do it. I'm telling
38:16 you, man. Right now is absolutely the
38:20 time. I I was thinking about this today.
38:23 Um I I forget. Oh, Kim Kardashian just
38:26 signed a contract to do micro dramas.
38:29 So, in in in
38:31 uh in Japan, is it is it Korea? I think
38:34 it's in Korea. In South Korea right now,
38:37 micro dramas are like a $4 billion
38:39 business or 40 billion, I don't know,
38:41 some stupid amount of money that they're
38:43 making on these shitty, horrible 90
38:47 second soap operas that are filmed in
38:48 vertical format. Um they're just they're
38:52 atrociously bad. They're making a
38:55 [ __ ] fortune. So they're So US
38:57 executives are like, "Well, that there's
38:59 money to be made. Let's do those." So
39:01 Kim Kardashian just did one of those.
39:04 Like I was thinking today
39:06 Oh, no. Not this again. Yeah. Well, I
39:08 like I was thinking today um you could
39:11 do a Gen X You could do a Gen X micro
39:14 drama that didn't suck that Gen Xers
39:17 would love, right? They just they'd just
39:19 be bitter. They'd just be all [ __ ]
39:21 bitter. And then you would have like
39:23 supporting cast of like Jenzers coming
39:25 in and being triggered by them
39:28 [laughter]
39:29 and just the the Jenzers are always
39:31 offended. The millennials are always,
39:34 you know, hopeless, not motivated. Jed X
39:38 is like just get off your ass. I'm
39:40 triggered.
39:41 End the episode.
39:44 [laughter]
39:46 Oh man. How do we maintain vocalist over
39:49 days and months? Um you mean like so in
39:54 Sunno one of the things that you can do
39:56 is you can create a persona. So if it
39:58 generates like that th this this singer
40:03 wait not that one this this singer here
40:06 [music]
40:18 [music]
40:22 the clouds were crawling. [singing]
40:26 So, I can now with that song like I like
40:30 that that that
40:32 vocal a lot.
40:34 I can How do I do this?
40:38 Remix. Edit.
40:42 Is that it? No. Create. Make persona.
40:46 [gasps]
40:47 So, oh, you don't see the popup menu,
40:49 but in the create menu, am I in the
40:52 create menu? Yeah, in the create menu I
40:54 can say make persona
40:56 based on the clouds. Uh this this is
41:00 shown on the songs. Upload a photo.
41:04 Prompt suno to create an image. No.
41:06 Describe the I'm just going to say
41:08 grally
41:11 um uh
41:14 folk hero or not no a folk rock hero.
41:21 Um,
41:24 avatar. Okay, create the image.
41:27 Is it creating it? Is it doing it?
41:31 Whatever. It doesn't matter. So, I can
41:33 save it.
41:35 Make the persona public.
41:38 Sure. I don't give a [ __ ] Okay. But now
41:42 that I've got a persona, I can add that
41:44 persona to any song I want. Maybe give
41:46 it a name and say each time you ask for
41:48 that name, it's the same singer. Yeah,
41:50 it's exactly what it is.
41:53 Okay. All right. So, let's go to our
41:55 psychedelic psych psychedelic versions.
42:04 [music]
42:17 >> [music]
42:21 [singing]
42:26 >> The clouds were crawling.
42:27 [music and singing]
42:29 >> Yeah. See this work,
42:33 >> but they were screaming to be seen.
42:41 [music]
42:42 The Rockies [singing]
42:43 held their ground.
42:48 [music]
42:49 Wind passed [singing] by without a
42:52 sound.
42:54 Oh, [music]
42:55 the [singing] clouds.
43:04 >> The sun was setting [singing] in the
43:07 west.
43:12 List those [music] clouds [singing] in
43:14 their Sunday best.
43:16 [screaming]
43:20 [music] Rocks looked on without
43:22 [singing] a care.
43:27 [music] The grasses swayed [singing] and
43:29 we all stared
43:32 at the cloud.
43:42 Um, by the way, Grimes's new song is
43:44 called Artificial Angels. It's about AI.
43:47 Gro is is in the music video. You know,
43:50 it's funny, Robbie. Um, I just one of
43:52 the things that's that's happening in
43:54 Nashville right now is a lot of
43:57 musicians are not hiring studio m studio
44:00 musicians to do their their comps. you
44:02 know, they're writing songs where they'd
44:04 normally go into a studio to do exactly
44:06 what I'm doing here, right? You take you
44:09 take a shitty recording that is not even
44:12 listenable, right? And you figure out a
44:15 way to take this the seed of what that
44:17 thing was and turn it into something.
44:20 Um, I think one of the dirty secrets
44:22 going on in music right now is I have a
44:24 feeling that there's there's a whole
44:28 there. You remember when when uh
44:32 Ethan Mullik wrote the article called
44:34 Secret Your Secret Cyborgs? Yeah, I
44:37 already got that. I already got that,
44:38 Brandon. Um [clears throat]
44:42 um Ethan Malik wrote this article called
44:44 Secret Cyborgs. And basically what it
44:47 was was in companies that don't allow a
44:49 AI, there are employees that are using
44:52 AI, but they're not admitting it, right?
44:54 So they're your secret cyborgs. I think
44:56 that's happening right now in the music
44:58 business a lot. You've got a couple of
45:00 people like Timberland and and Grimes
45:03 and um some other artists that are
45:08 leaning into it. They're going, "Fuck
45:09 it. This is just another production
45:11 tool. I'm in. Let's go. Let's go. Let's
45:12 go make music and then we'll figure out
45:15 what we want and we'll go produce it,
45:16 right? We'll produce it how we normally
45:18 produce it." Or maybe not. Like like
45:20 they don't seem to care about the the
45:23 function of how you get there. they just
45:24 want to get to the music. And then
45:26 there's others like, I'll never use it.
45:27 And I think there's a big gray a gray
45:31 swath of the music industry right now
45:33 secretly using the stuff because these
45:35 tools are just too good not to use. Um,
45:37 you can choose not to use them, right?
45:40 because you like playing and you like
45:42 writing and you But if you're like me
45:44 where like I occasionally write a song
45:47 and most of the time it's
45:50 like that one was um
45:54 and I can turn that into something
45:56 listenable. [ __ ] yes.
45:58 [ __ ] yes. So anyway, all right, let's
46:00 listen to another
46:07 [music]
46:12 >> [music]
46:18 >> So, so Robbie, I have a playlist called
46:20 Cyber Tunes, all music about technology.
46:22 I don't know if you know this, Robbie,
46:24 but I've spent the last year and a half
46:25 writing a musical about AI. It's about
46:29 an AI chatbot that falls in love with a
46:31 tech reporter that's reviewing her. It's
46:33 called Sydney: An Artificial Love Story.
46:35 And if you go to Soundcloud and look up
46:38 Sydney and Artificial Love Story and
46:40 there's there's two there's two versions
46:42 of the album there. One is the original
46:45 one with full orchestrations and then
46:46 there's there's one called Sydney an
46:49 artificial love story 2025 update.
46:52 That's the latest uh soundtrack. So So
46:56 yeah, take a look if you want to uh if
46:58 you want to hear the uh the soundtrack
47:01 to the the musical I'm in the middle of
47:03 producing right now. Um, we're in the
47:05 middle of producing it. Anyway, [music]
47:14 >> this definitely needs to be a woman
47:15 singing this.
47:17 [music]
47:25 >> The [music] Rockies held their ground.
47:27 [singing]
47:32 That had some weird ass [ __ ] in it.
47:33 Okay. So, this is what did we add to
47:35 this one? We added Oh, this is
47:36 minimalist orchestration. This is the
47:38 one I think I'm going to like or this
47:40 prompt
47:43 already. Love it.
47:51 [music]
47:54 [singing]
47:56 The clouds are crawling.
48:00 >> All right. like the beginning, but it's
48:02 too complicated.
48:05 [singing]
48:09 [music]
48:12 [singing]
48:14 about them. [music]
48:24 [music]
48:30 The clouds were [music] crawling to the
48:33 east, [singing]
48:37 [music]
48:38 but they were screaming to be seen.
48:44 >> [singing]
48:45 [music]
48:48 >> The Rockies held their ground.
48:55 [music]
49:01 [singing]
49:02 >> Close. Let's do another one with that
49:04 same prompt. And then let's go. We're
49:06 going to take weirdness down a little
49:08 bit.
49:11 And I'm going to take style influence
49:13 down a little bit and we'll create
49:15 another one. Um, that was close. That
49:17 was close.
49:19 Thera the table. I don't know what that
49:21 means. Is that Should I write that in as
49:22 in the prompt?
49:24 Let's just see what we get here. See if
49:26 we can get one. That's That's
49:42 By the way, if if you don't know where
49:43 we are, we're at sununo.com.
49:48 >> The clouds were crawling,
49:55 [music]
49:55 [singing] but they were screaming.
50:00 >> This lost the soul.
50:03 D
50:13 [music]
50:20 [music]
50:26 cloud. [singing]
50:29 >> [music]
50:35 [music]
50:37 [singing]
50:41 >> up.
50:44 [music]
50:48 The clouds [singing] were crawling to
50:51 the east. [music]
50:56 That was close.
50:58 have bad bad words. [music]
51:02 I added the carnatic carnetic carnetic
51:06 scale. I added that for the last two
51:09 after these two. [music]
51:19 [music]
51:27 >> [singing]
51:28 >> that
51:32 >> the [music] clouds were crawling to the
51:34 least
51:35 >> like this,
51:39 [music]
51:40 but they were screaming to be seen.
51:48 [music] The Rockies [singing]
51:56 passed by without a sound.
52:02 [music]
52:05 [singing]
52:11 The sun [music] was setting in the west.
52:17 >> [music]
52:19 >> those clouds in their [singing]
52:22 best.
52:24 >> You know what's funny? That line, those
52:26 clouds in their Sunday best
52:29 in my version and in the other version
52:32 with the male, I hated that line. That
52:34 just sounded like organic and cool.
52:38 setting in [music] the west.
52:43 [singing]
52:45 The [music] sun was setting in the west.
52:53 List those clouds in their [singing]
52:55 Sunday best. [music]
53:00 >> Rocks [singing] looked without a care.
53:08 The grasses [singing]
53:10 and we all stared
53:15 at the clouds.
53:22 [singing and music]
53:27 [music]
53:28 [singing]
53:42 >> [music]
53:43 >> The clouds they covered up that sun.
53:48 [singing]
53:49 [music]
53:50 His work today was [singing]
53:53 nearly done.
53:58 [music]
53:59 The shadows cross the land.
54:03 [singing]
54:06 Sun broke through for one last stand
54:09 [singing and music]
54:12 at the cloud. [music]
54:16 [singing]
54:24 [music]
54:26 It was getting dark, you see. [music]
54:33 Nowhere [singing] else I'd rather be
54:41 sitting [music and singing] still and
54:42 trying to breathe.
54:48 I think I may [singing]
54:51 never leave
54:54 the cloud.
54:57 [singing and music]
55:05 >> [music]
55:09 >> That That [ __ ] works. [ __ ] works.
55:14 >> [ __ ] works. All right, I'll publish
55:16 that one. [laughter]
55:18 Publish
55:19 the clouds. Um,
55:22 from a
55:25 small
55:29 Dry,
55:32 small, dry,
55:35 unimpressive
55:36 seed
55:40 emerged.
55:45 This [laughter]
55:47 add a style summary.
55:50 psychedelic
55:54 folk.
55:57 Uh, no, I'm not going to show my re
55:59 remix origin. [laughter]
56:02 No, nobody's going to hear other than
56:05 you guys hearing that song for 20
56:08 seconds. That's it. [laughter]
56:12 Add comments. Allow remix. You can remix
56:14 this one. Pin song to profile. Okay,
56:17 that's good. The clouds cover. We're
56:19 gonna get rid of the cover.
56:22 The clouds now. Let's animate it. Oh,
56:25 no. That's going to take me
56:28 time.
56:29 All right. Publish.
56:33 So, if you go look up the clouds. All
56:35 right, that's published. I want to I
56:36 want to listen to three more and then
56:38 we'll we'll go. Although, wait, we
56:41 should [laughter]
56:42 of course we're not going to go. Let's
56:43 do um
56:46 we'll do um um
56:50 Gothrock
56:52 um [clears throat]
56:56 angry. [laughter]
56:59 Gothrock angry
57:02 um angelic. [laughter]
57:06 Okay, those are going to be good.
57:09 [laughter]
57:13 Oh my god. The first fe female was also
57:16 great. The first female. Let's see.
57:18 Where did we do the first female?
57:23 >> Psychedelic. The very first one, right?
57:27 [music]
57:32 [music]
57:38 Oh. [music]
57:40 Oh,
57:45 [music]
57:46 [singing]
57:54 [singing] the clouds were crawling
57:55 [music] to the east.
58:02 >> But they were screaming to be seen.
58:12 You know what's funny, Joy? You would
58:14 have stopped at that version. When I
58:16 watched Corey Sandler do art, I I felt
58:20 the same way. Corey ran down
58:24 visual rabbit holes
58:27 for like 40 or 50 iterations where I
58:32 would have stopped way up there. It's so
58:35 funny how that how that stuff works.
58:37 Just check
58:39 if here's one. But that's the one I
58:41 think it was. I think that might be the
58:43 one. Maybe it was the second. I think
58:46 the second
58:48 [music]
58:55 [music]
59:01 clouds were crawling to the east.
59:08 >> [music]
59:08 >> But they were screaming to be seen.
59:12 >> The other one.
59:18 [singing]
59:20 [music]
59:22 The clouds were crawling to the east,
59:27 [music]
59:30 but they were screaming to be seen.
59:34 >> [music]
59:39 >> The Rockies [singing and music] held
59:40 their ground.
59:45 [music]
59:45 Wind passed [singing] by
59:48 a sound.
59:51 [music]
59:52 >> That was more straightforward. I'll
59:53 publish it. Um, add a caption. Swell
59:58 song. [laughter]
1:00:01 Publish. And then I want to go hear the
1:00:03 rock and roll ones. And then we'll go
1:00:05 we'll go do something else. All right.
1:00:08 [clears throat]
1:00:10 [music]
1:00:13 This is goth rock angelic. Wait.
1:00:15 Gothrock angry angelic.
1:00:22 [music]
1:00:33 >> [music]
1:00:38 >> The clouds were crawling to the east,
1:00:44 [music]
1:00:45 but they were screaming to be seen.
1:00:52 [singing]
1:00:53 [music]
1:00:54 The Rockies held the ground.
1:01:01 >> [singing]
1:01:02 >> pass by without a sound.
1:01:07 Oh, the [music]
1:01:13 [laughter]
1:01:14 Cy not even close.
1:01:17 [music]
1:01:22 [music]
1:01:28 CLS [music]
1:01:38 [music]
1:01:46 couldn't live.
1:01:50 But they were screaming to be seen.
1:01:55 >> So that still work.
1:01:58 >> The Rockies [singing] held their [music]
1:02:01 ground.
1:02:06 >> Wouldn't pass by without [singing]
1:02:08 a sound.
1:02:22 You can't hear you can't hear the song
1:02:25 with that's thing. I just don't like
1:02:28 that kind of music. All right. Um, let's
1:02:31 move on. Let's move on. Vicky comment.
1:02:34 All this stuff is throwawayable and
1:02:36 there will always be something better.
1:02:37 Always something worse, something
1:02:38 comparable. Yeah, I mean Vicki, your
1:02:41 point's a good one and I think
1:02:42 ultimately I mean this this gets into
1:02:45 the whole idea of you know my contention
1:02:47 that our job in the future is to
1:02:50 essentially be Rick Rubin, right? Like
1:02:53 the reason Joy, you were like, "Oh, I
1:02:55 would have stopped at that first one and
1:02:56 I kept going." Is I don't know. There
1:02:58 was something about the first one that
1:02:59 just felt a little thin. It felt like
1:03:01 there was somewhere deeper we could go.
1:03:04 But it's like that's the that's the
1:03:09 that's ultimately [laughter]
1:03:11 what what art is. Ultimately art is you
1:03:15 having an idea and then you choosing
1:03:20 when to stop, right? [laughter]
1:03:22 Like it's it's at the point at which you
1:03:24 know when Corey Sandler's making her
1:03:27 pottery and she's doing glazes and
1:03:28 experimenting with glazes and
1:03:30 temperatures and things like that. You
1:03:32 know there's probably 50 versions of
1:03:35 something until she gets it where she's
1:03:36 like, "No, that's me. That one's me."
1:03:39 And then that becomes a series, right?
1:03:42 It's like, "What is what is the thing
1:03:45 that made her stop at that 50th one or
1:03:48 probably the 47th one, right? She's
1:03:50 she's like, "Ah, that that one's really
1:03:51 good.
1:03:53 it. That's just who she is, right?
1:03:54 That's her values. That's her aesthetic.
1:03:56 That's her moment where she goes,
1:03:59 "That's that's it for me." And this is
1:04:02 one of the things that Rick Rubin says
1:04:04 all the time is it's not for them. He
1:04:08 he's he's famous for saying the audience
1:04:10 comes last
1:04:13 because
1:04:15 all he's trying to do is get the song to
1:04:18 sound good to his brain.
1:04:21 And when it works, it works. And then he
1:04:23 goes and he puts that out and he talks a
1:04:26 lot about
1:04:28 there are songs that he's put out that
1:04:31 he thought were incredible,
1:04:34 that were absolute flops,
1:04:38 but he doesn't give a [ __ ] about it.
1:04:39 It's like that's not his job. His job is
1:04:42 just to figure out what he wants, put
1:04:45 his stamp on it, and put that in the
1:04:47 world.
1:04:49 um knowing what to keep and what to
1:04:52 throw away. And we may not all keep and
1:04:54 throw away the same thing. I think
1:04:56 Vicki, we would not like what would be
1:04:59 fascinating what would be a fascinating
1:05:01 exercise is to
1:05:04 I I mean you could you could do it in
1:05:05 such a a number of ways. You could say
1:05:10 you could say you take you take a song
1:05:12 and you do I don't know 60 different
1:05:16 versions of it.
1:05:18 And then you let a hundred people go.
1:05:20 Actually, you take 60 people. You take
1:05:23 60 people [laughter] and have them
1:05:25 listen to 60 songs and everybody can
1:05:27 only pick one. It' be fascinating to
1:05:30 look at, you know,
1:05:32 you know, is there some sort of
1:05:33 consensus? Do they do they converge
1:05:35 around something or is it just
1:05:37 everyone's got this slightly different
1:05:38 taste? You know, I like to compare AI
1:05:41 music today to to Beasty Boys. Yeah. Um
1:05:45 Paul's Boutique at the time. Yeah.
1:05:47 Exactly.
1:05:48 It this is,
1:05:51 you know, just just like early hip hop
1:05:54 caught [ __ ] for. You're not real
1:05:56 musicians. All you're doing is sampling
1:05:58 other people's music. You're, you know,
1:06:00 you're hacks. That's not real music.
1:06:02 You're just talking. That's not real
1:06:05 music. Real singers sing. Real musicians
1:06:08 play violins, right? And you're playing
1:06:11 a tape recorder. No. Sorry. That doesn't
1:06:14 count, Skippy. Right? Except that it
1:06:18 changed [ __ ] everything, [laughter]
1:06:20 right? It changed the music industry
1:06:23 because they were musicians. They just
1:06:25 had a different way to express
1:06:26 themselves. There's going to be a whole
1:06:28 new generation of people that use these
1:06:29 AI tools in really creative and
1:06:31 compelling ways. And here's what I can
1:06:34 promise you.
1:06:36 We will not be able to tell the
1:06:38 difference in in the hands of good
1:06:40 crafts people whether something was AI
1:06:43 generated 100% 50% or 0%.
1:06:50 We won't know because the AI tools are
1:06:52 going to get good enough that it's just
1:06:54 up to the person generating it to say,
1:06:56 "Hey, here's my song I wrote. I mean, I
1:06:58 wrote this song. I wrote the lyrics. I
1:07:01 wrote the music. The chord changes are
1:07:02 mine, but this crazy ass version that we
1:07:06 just kept and published.
1:07:10 [music]
1:07:11 >> Yes.
1:07:16 [music]
1:07:26 >> How how how are record labels reacting
1:07:29 to AI music? Well, uh, Universal Music
1:07:33 sued UIO and now they are in partnership
1:07:37 and you can no longer download your own
1:07:40 songs from Ude. So, so how they're
1:07:44 dealing with it is they're suing the
1:07:45 living snot out of these companies and
1:07:48 then taking over creative control.
1:07:50 Warner Music did the same thing to
1:07:52 Sunno. So, right now, you can't download
1:07:55 your music if you're not a paying
1:07:57 member. And then Sunno also just
1:08:00 announced, hey, we're going to be
1:08:02 deprecating our model to remove any
1:08:05 copywritten work in it, which the music
1:08:08 industry is going to go great. But then
1:08:10 what Warner's going to do is Warner is
1:08:12 going to take all of their licensed
1:08:14 music. They're going to put it in there
1:08:16 and then just like the billiondoll deal
1:08:18 that Disney just did with Open AI
1:08:22 where you're going to be able in Suna in
1:08:24 in uh Sora to be able to generate Disney
1:08:29 characters in Sora videos.
1:08:32 Disney paid Open AI a billion dollar.
1:08:38 They didn't charge Open AI to be able to
1:08:40 use Disney IP.
1:08:43 They paid them a billion dollars to
1:08:46 access their 800 million customers a
1:08:49 week
1:08:51 to say make [ __ ] with our with our IP
1:08:54 because we're going to own all that [ __ ]
1:08:56 and and our customer base is going to
1:08:59 make stuff with our IP
1:09:02 that we're going to exploit the ever
1:09:04 living [ __ ] out of. So that's what the
1:09:06 the the music companies right now are
1:09:08 attempting to do the exact same thing.
1:09:10 So Warner and Sunno have a deal.
1:09:13 Universal and UDIO have a deal. There's
1:09:15 producer AI hasn't been touched yet. And
1:09:18 there's uh there's a third one. I can't
1:09:20 remember what it is right now that
1:09:21 hasn't been touched yet. Um here's
1:09:24 what's going to happen.
1:09:26 Most of these big mergers, the big
1:09:28 company is going to absolutely [ __ ] up
1:09:30 the product. And what's going to happen
1:09:32 is the people from these companies are
1:09:35 going to quit and they're going to go
1:09:37 build open-source versions of these that
1:09:39 that decimate the big labels. [laughter]
1:09:43 So So we we ain't heard the end of it
1:09:46 yet, but it's it's like listen,
1:09:50 Hollywood is doing this, too. Hollywood
1:09:52 is investing in this um uh who was it
1:09:55 was it Lionsgate that invested so much
1:09:57 money in runway and then they you know
1:10:00 what they discovered if if they limited
1:10:02 the training set to too narrow the model
1:10:05 sucked. So So that's what they're in the
1:10:08 middle of right now is they tried to get
1:10:10 in there and put their IP attorneys
1:10:13 around it and they [ __ ] up the model.
1:10:15 So I I don't know. I I mean
1:10:19 in every single industry, what's going
1:10:21 to happen is your centers of power
1:10:24 are first going to laugh at the
1:10:26 technology,
1:10:27 then they're going to sue the
1:10:29 technology, then they're going to own
1:10:31 the technology, right? It it's it's how
1:10:33 it's always gone. It's how it's going to
1:10:35 go here. And what's going to happen is
1:10:37 most of those big companies are going to
1:10:39 [ __ ] it up. And then what's going to
1:10:41 happen is someone's going to squirt out
1:10:43 the other side and they're going to do
1:10:45 something that's actually that actually
1:10:47 becomes the industry standard that is
1:10:49 outside of the power structure and
1:10:51 they're going to become the new power
1:10:53 structure.
1:10:54 So
1:10:56 um we could have had this 10 years ago
1:10:58 if Pandora didn't sell out to YouTube or
1:11:01 copyright enforcement. Yeah. I mean but
1:11:03 there's a point at which listen if if
1:11:05 I'm a if I'm a um let's say I'm an
1:11:08 entrepreneur which I am. I've started I
1:11:10 don't know 15 20 companies something
1:11:12 like that. I'm 14 years into my current
1:11:14 one. Um
1:11:18 let's say I started Suno
1:11:22 and what I'm looking at is I've got a
1:11:26 company that's making $100 million
1:11:29 in revenue after whatever it is three
1:11:31 years something like that. I'm making
1:11:32 some decent amount of money and I get
1:11:35 sued for I don't know what the lawsuit
1:11:38 was. It was 100 billion
1:11:40 100 billion or I don't know 200 billion
1:11:44 that if I lose
1:11:46 not only do I lose my company I
1:11:48 personally am going to owe a hundred
1:11:51 billion dollars [laughter]
1:11:53 to to this case right um or they're
1:11:59 going to come to me and say well you
1:12:00 know we'll take this to court and we'll
1:12:02 use our high-powered attorneys to grind
1:12:04 you to dust or we can do a deal.
1:12:09 And so that's what's happened with two
1:12:10 of these companies right now is they've
1:12:12 cut deals with the big companies. So
1:12:14 that's just that's just how it goes. I
1:12:16 don't blame anyone, man. If you're out
1:12:17 there trying to make it like and they're
1:12:20 like, instead of you being a hundred
1:12:22 billion dollars in debt, how about we
1:12:24 give you, I don't know, $40 million for
1:12:26 yourself personally, you can go have a
1:12:29 nice little family. You can do whatever
1:12:30 you need to do. You sit here long enough
1:12:32 to teach us what you did and how you did
1:12:34 it, and we'll take it from here. Thank
1:12:36 you very much. and go [ __ ] it up
1:12:38 because we know how to run the music
1:12:40 business, which this [ __ ] quite frankly
1:12:43 doesn't fit in very well. [laughter]
1:12:46 But it does, right? Because
1:12:50 the listeners of music,
1:12:53 if they like a song, they like a song.
1:12:57 They don't give a [ __ ] if it was AI
1:12:59 generated or not. Right? You you've got
1:13:04 we're in a weird [ __ ] up transition
1:13:06 period where
1:13:08 where we're putting moral judgment
1:13:10 judgments on technology that doesn't
1:13:12 give a [ __ ] and on a populace that
1:13:14 frankly doesn't give a [ __ ] The
1:13:16 populace just wants good music. And if
1:13:18 it's good music that was sampled like
1:13:21 hip-hop style or music that was put into
1:13:24 a loop generating machine like Garage
1:13:27 Band, that's not real music. Real
1:13:29 musicians don't use loops.
1:13:32 Some do,
1:13:34 right? Or if it was AI, if it's a
1:13:38 banger, it's a banger. If people like
1:13:40 it, they like it. So anyway, uh but the
1:13:43 big companies are going to try to
1:13:44 swallow them all up.
1:13:46 Um something tells me Apple Music isn't
1:13:48 going to embrace AI music artists. I
1:13:51 completely disagree with that. I think
1:13:53 everyone I
1:13:56 Okay, [laughter]
1:13:57 there's
1:14:00 Apple and everyone will embrace wherever
1:14:04 the [ __ ] money is.
1:14:11 If there are 50 hits on the radio, three
1:14:14 years from now, there's 50 hits on the
1:14:16 radio. Well, we don't even have radio
1:14:18 anymore. There are 50 hits on the
1:14:19 streaming platforms.
1:14:21 big banger hits and half of them are AI
1:14:24 musicians and half of them are not. Do
1:14:27 you think Apple is gonna refuse to to to
1:14:32 host the AI AI any AI assisted music? I
1:14:36 mean ASCAP and BMI and the and the third
1:14:39 licensing entity have all said that AI
1:14:43 assisted songwriting is fine as long as
1:14:46 there's been human contribution in it.
1:14:48 It's fine. Well, what's human
1:14:50 contribution? I mean, prompting is human
1:14:52 contribution.
1:14:54 Doing iterations of something is human
1:14:56 contribution. Me uploading a shitty song
1:14:59 that was the foundation of one of these
1:15:00 things is human contribution.
1:15:03 But it's an AI generated song. Does it
1:15:05 really [ __ ] matter? It doesn't. And
1:15:09 so I think I think the big companies go
1:15:10 where the money is.
1:15:13 You know, you can hate it. You can hate
1:15:14 the game, but that's the [ __ ] game.
1:15:18 It's the game.
1:15:20 Now, you want to you want to counter the
1:15:22 game here? Here's where where I think
1:15:24 you could counter the game. Let all the
1:15:26 big companies go chase all the big
1:15:28 dollars and do a counterrevolution of
1:15:32 analog music, right? You [clears throat]
1:15:35 know, you're only going to hear my
1:15:36 songs, you know, in live venues. I'm not
1:15:39 going to put anything on Spotify. You're
1:15:41 only going to listen to me at concerts.
1:15:44 and we're going to confiscate your
1:15:45 phones so you can only experience us
1:15:48 analog. I could see that absolutely
1:15:50 being a trend.
1:15:52 [clears throat and cough]
1:15:53 But but AI music is that distinction is
1:15:59 it's only because
1:16:02 we know what was here before.
1:16:05 Five years from now, no one's going to
1:16:06 give a [ __ ]
1:16:09 Um
1:16:11 didn't realize it was still Monday.
1:16:14 I want AI music pressed to vinyl just
1:16:17 for the irony. Why not? Why not press it
1:16:21 to vinyl?
1:16:23 I I'll tell you what, this song.
1:16:29 [music]
1:16:31 >> Someone said it before. I would go see
1:16:33 those women sing that song. Put to And
1:16:36 Corey said, I'd like to be on stage
1:16:37 singing with those women. like put
1:16:41 together a band to go sing [ __ ] like
1:16:43 this,
1:16:46 [music]
1:16:49 >> right? You got your little hand pan,
1:16:54 right, dude? The with his man bun, you
1:16:58 know, sitting in the corner. You got the
1:17:00 the beautiful the beautiful women in
1:17:02 their in their wispy dresss and their
1:17:04 long hair and the beads, right? And
1:17:07 they're like swaying [singing]
1:17:10 >> [laughter]
1:17:11 >> name
1:17:13 your little yogi in the corner.
1:17:15 [laughter]
1:17:16 Why not?
1:17:21 [music]
1:17:22 But they were screaming [singing] to be
1:17:25 seen.
1:17:30 The [singing] Rockies [music] held their
1:17:33 ground. [laughter]
1:17:35 Got people ordering kombucha. Come on,
1:17:37 man. You can see it. [laughter]
1:17:43 Oh, Christ. Okay, so let's see. I've got
1:17:47 I got to talk about some [ __ ] So, let
1:17:48 me just go to full full me.
1:17:52 [clears throat]
1:17:53 I got it. Okay. Um,
1:17:58 if you don't know about it, AIFest is
1:18:01 next Friday. It starts next Friday. Go
1:18:03 to aifestivist.com right now. If you're
1:18:05 like, Kyle, what the [ __ ] are you
1:18:06 talking about? Is this a music thing?
1:18:08 No.
1:18:10 I thought we were doing pictures
1:18:11 tonight. We were. But then welcome to
1:18:14 the AI learning lab. This how we do
1:18:16 things here. Sometimes I complete
1:18:18 sentences. Sometimes I do what I say I'm
1:18:20 going to do. Other times I yell about
1:18:23 folk music. [laughter]
1:18:30 AI Festivus is a remarkable event
1:18:34 that takes place Friday and Saturday,
1:18:37 December 26th and 27th. So, next Friday
1:18:39 and Saturday from 9:00 a.m. Pacific to
1:18:43 900 p.m. Pacific on Friday and from 900
1:18:46 a.m. Pacific to 9:00 p.m. Pacific on
1:18:49 Saturday. 24 hours of programming over
1:18:52 two days
1:18:54 with with a 12-h hour nap [laughter]
1:18:56 because the first year we did did this,
1:18:58 we did it for 20 24 hours straight. And
1:19:01 Ann and I both realized my my co-host
1:19:04 Ann Murphy from She Leads AI um we
1:19:07 needed naps because we're older.
1:19:09 [laughter]
1:19:10 So AI Festivus is that. And then what
1:19:12 happens in those 24 hours? We've got
1:19:15 remarkable speakers from all over. We've
1:19:18 got filmmakers. We've got IP attorneys.
1:19:20 We've got people that work at big tech
1:19:22 companies. We've got all sorts of cool
1:19:24 stuff. Um there's people that are going
1:19:27 to be teaching you stuff there. We're
1:19:28 going to have a mini AI film fest. We've
1:19:31 got um Kelly Bosch is going to be
1:19:34 showing work there. Kim Offford, Joy
1:19:37 Perie.
1:19:39 Um
1:19:40 we're going to be doing that. I'm going
1:19:42 to be previewing my musical. One of the
1:19:44 sessions is going to be the the second
1:19:46 half of act one of my musical that
1:19:49 you're going to be able to get to to
1:19:50 hear actors read the words uh and then
1:19:53 hear the songs that go that go with
1:19:55 those words that I'm really excited
1:19:57 about. Um so anyway, and it's free. It's
1:20:00 [ __ ] free. Free free. Free as in not
1:20:04 three. Free with an F.
1:20:08 You're like that's stupid. I know.
1:20:12 So you want to support it? There's a
1:20:13 couple of ways you can support it. You
1:20:15 can go right now, you can pre purchase
1:20:18 the um deluxe um recording bundle, the
1:20:22 replay bundle um for 27 bucks. 24 hours
1:20:27 of AI education for 27 bucks if you go
1:20:29 buy it right now. So go do that. That
1:20:31 supports us. We also have a friends and
1:20:34 family sponsorship thing where if you
1:20:36 just want to sponsor it and like uh you
1:20:38 guys are up to some really remarkable
1:20:39 stuff here, you can do that, too. Um, if
1:20:42 you go to aifestivist.com,
1:20:44 you can see who's going to be speaking.
1:20:46 Um, not everyone's in there. We're still
1:20:48 still getting some people's information
1:20:50 in.
1:20:52 Um, you can do that. If you want to
1:20:54 sponsor us, if you're a company or you
1:20:56 know a company, there's a sponsorship
1:20:58 page. There's also a virtual trade show
1:21:00 booth. So, if you want to do virtual
1:21:02 trade show booths, you can do that. All
1:21:04 right. Um, so go check that [ __ ] out.
1:21:07 Okay.
1:21:12 Oh, there's also a lull this Saturday.
1:21:14 Oh, this Saturday, Vicki, is that right?
1:21:21 Got a tiny discount code, like $3 off.
1:21:25 [laughter]
1:21:28 It's free. We We put out a couple of
1:21:30 joke ads that were like buy one get one
1:21:32 free. The secret is the first one's
1:21:34 free, too. [laughter]
1:21:37 Um,
1:21:41 okay. AI festivist. So, that's good.
1:21:45 Salons's Got Talent. If you go into the
1:21:47 AI salon, if you remember the AI salon,
1:21:49 go into challenges and competitions.
1:21:52 Brandon's got a post there for Salons's
1:21:55 Got Talent. So, over the past, I don't
1:21:57 know, two weeks or so, we've had people
1:21:59 submit work. You can go in and vote on
1:22:01 work with emojis.
1:22:03 And then we're going to take the top
1:22:05 three, I think it is three or four,
1:22:06 Brandon, and we're going to um narrow
1:22:10 those down. We're going to pick a
1:22:11 winner. And then during Festivus, we're
1:22:13 going to share the finalists work and
1:22:16 we're going to award um a one-year
1:22:18 mastermind subscription to the winner.
1:22:21 Okay. So, that's happening. You can go
1:22:23 vote through tomorrow, I think. Is that
1:22:25 right?
1:22:27 The other thing you can do if you want
1:22:29 to see some of the people that are
1:22:31 speaking tomorrow and Murphy and I are
1:22:33 doing the AI readiness project podcast,
1:22:36 we've got five guests that are speakers
1:22:39 for AI festivists. So come hang out with
1:22:42 us. So if you go to a readiness
1:22:44 project.com,
1:22:46 you can see where to stream us and watch
1:22:49 us. And you can that's tomorrow at 400
1:22:51 pm Mountain time. Okay. And actually
1:22:54 what what's cool for my day tomorrow is
1:22:57 I go from AI readiness project podcast
1:23:00 where we get to talk to our cool
1:23:01 speakers and then I go into rehearsal
1:23:04 for the Sydney preview.
1:23:07 So I'm meeting with my coowriter and the
1:23:10 two actors that are going to play Sydney
1:23:12 and Kellen in the uh in the thing. All
1:23:15 right. Tonight at midnight voting ends.
1:23:17 Okay.
1:23:19 And then the mastermind.
1:23:22 The other thing is if you're a part of
1:23:23 the AI salon and you're not part of the
1:23:25 mastermind through the end of the year
1:23:27 it's 20 bucks a month and that's ongoing
1:23:29 as long as you keep your subscription.
1:23:32 We just launched the AI salon mastermind
1:23:36 practice which is a a uh a framework for
1:23:40 creating a daily practice around AI. And
1:23:43 if you're a member of the mastermind you
1:23:45 can join us on Thursdays at noon Eastern
1:23:48 for the practice lab. And what we do in
1:23:51 the practice lab is we talk about what
1:23:54 we're doing, how we're doing our daily
1:23:56 practice. We're walking people through
1:23:58 this framework of things like how to
1:24:01 play with purpose and how to, you know,
1:24:03 go beyond your comfort zone and and
1:24:05 learn different things. Okay? So all of
1:24:08 that stuff is happening.
1:24:11 So So this is just like [laughter]
1:24:15 like whipped cream on top of that. Okay.
1:24:20 Um,
1:24:24 tomorrow I promise we will go play with
1:24:26 pictures. Um, oh man, I'm so late.
1:24:29 That's okay. Side hustle Mimi, I told
1:24:31 people we're going to play with the new
1:24:32 OpenAI image model, which is quite good.
1:24:35 Um, I can show you some images I made
1:24:37 with it. Hang on, let me let's do that.
1:24:39 I'll at least give you a little a little
1:24:41 nugget.
1:24:44 [clears throat and cough]
1:24:45 What's this? Oh, that's Festivus.
1:24:48 Yeah, there's all of our speakers. Oh,
1:24:50 wait. Am I sharing this? I'm not. Share
1:24:52 this tab instead.
1:24:57 Look at all these speakers.
1:25:05 Kate Rivera did the design of the site.
1:25:07 It looks great. So, there's all that.
1:25:11 Um, all right. I'm going to go to chat
1:25:14 gate.
1:25:20 So, the one thing that's new, well, it
1:25:21 it was here before. It was called
1:25:23 library. I think they changed the name
1:25:25 of library to images. If you click on
1:25:28 it,
1:25:30 um, it's got a weird interface. I'll
1:25:32 give them that. So, you've got your chat
1:25:34 box up here. You've got different styles
1:25:36 down here. Here are some pre-prompted
1:25:40 things that you can play with.
1:25:42 Redecorate my room. Create a
1:25:44 professional job photo. Remove people
1:25:46 from the background.
1:25:48 Um, restore an old photo, give us a
1:25:50 matching outfit.
1:25:52 Um, and then
1:25:57 like here are
1:25:59 Can I do I can go up? Yeah.
1:26:04 So, I told it to give me a
1:26:07 Can I zoom? No.
1:26:11 I told it to put me in front of a GTO
1:26:13 jud a 69 GTO judge in Grabber Orange.
1:26:16 And I don't know if you can see it, but
1:26:18 right there behind my leg is the judge
1:26:20 badge and that says GTO. Normally these
1:26:23 image models [ __ ] that up. I said I
1:26:25 wanted to wear red Converse, Chuck
1:26:27 Taylor's, a garbage pale t-shirt, a New
1:26:30 York Yankees hat, and have the Rock and
1:26:32 Roll Hall of Fame in the distance. Um,
1:26:35 then I said I want that guy not to be
1:26:37 quite so fat and to be a little taller
1:26:39 because I'm vain.
1:26:41 [laughter]
1:26:42 And so it did that. And this is, I got
1:26:44 to say, this is one of the first models
1:26:46 that made an image of me that to me
1:26:49 looks like me. Um, all the other models
1:26:51 haven't, but I said I wanted him in a
1:26:53 dark blazer with a coke, but it took
1:26:56 away the car and took away the t-shirt.
1:26:59 So, I said, "No, put the car back and
1:27:01 put the t-shirt back." And there it is.
1:27:04 So, there's the version of me sort of
1:27:07 leaning on the car in my black jacket
1:27:09 with my hand in my jeans with my garbage
1:27:13 pale t-shirt, rock and roll in the
1:27:15 background, Coke in the foreground.
1:27:18 Pretty pretty crazy.
1:27:20 I had it do 26 children's blocks with
1:27:24 letters or images.
1:27:27 Um, that one it [ __ ] up. Then I had it
1:27:30 do happy holidays and it just took my
1:27:33 corporate photo and kept it and I said
1:27:35 no no I I want me in a smoking jacket
1:27:39 and it made me really fat. So I'm like
1:27:42 no I'm not that fat. At least in my
1:27:44 mind. So So there's me smoking a pipe.
1:27:48 That's bad. That's bad. That's bad. None
1:27:52 of these look like me. That but there
1:27:54 you go. There's me as what's her name.
1:27:55 the what's her name lady. [laughter]
1:28:00 So anyway, it's it's it's pretty good.
1:28:04 It's pretty good. So anyway, I would go
1:28:05 play with it. Just go to images, but you
1:28:07 can also just call it in the uh in
1:28:09 regular chat GPT. You can just say make
1:28:12 me an image of
1:28:14 All right. So we'll play with that
1:28:16 tomorrow. We'll we'll do some sidebyside
1:28:18 comparison stuff. All right, cool. Groov
1:28:21 groov groov groov. All right, peace out.
1:28:24 Um, don't stay up too late. Too late.
1:28:28 [laughter]
1:28:31 All right. Peace out then.
1:28:34 Bye.
1:28:37 They're [clears throat]