If AI replaces musicians, does the entire plugin industry die with them?

[BBFG#]

Most of our "country" stars in the States are now "all hat, no cows" now. And most of their music is not really country, but American Pop/Rock from the late '60s reborn.

[pdxindy]

But it's probably around 25% of the global music market. If you can't break the US market, you're never going to reach the really big time. That's just a fact.

I don't think that's true. Do you have a source for that "fact"?

I don't know what "really big time" means but there are massive stars with billions of streams in India and China who are virtually unknown in the US.

[jamcat]

How is that relevant to anyone here?

Obviously we're talking about western music as it applies to western audiences and western musicians. Because that's what we are here.

But that does now make me wonder if Suno can churn out some north Indian raga.

[ghettosynth]

OK, guys.

So now the criteria to prove that AI can write a competent catchy song is that it has to be #1 on an international chart, and Zeisner has to admit to personally buying a copy of it.

Well, I don't know about all of that, but, the main hook to this catchy number has been dancing around my brain all afternoon. I get that not everyone likes RnB.

[guitarzan]

did we ever get jesus riding a lobster?

IMG_0924.png

[BBFG#]

Naw, that's Kenny Loggins!

[Zeisner]

Being catchy is what makes it popular.

A catchy tune is only one factor of many. It helps but it's not necessary to make music or an artist popular. Example: Kylie Minogue (before "Can't Get You Out Of My Head"). Popularity is way too complex to reduce it to catchy tunes. You can make the catchiest music ever and it will never become popular because you're not a good entertainer or you don't have the looks or you don't align with the current trend/zeitgeist/hype or you don't find the right label etc. Reality is complex.

This is also one of the biggest issues with AI. Fans think it's going to be the solution to everything, as if there was just a single barrier that keeps them from becoming popular. And AI companies encourage all those simple minds. You only need this piece of software and all your problems will be solved! Because that totally worked out with any other "key to everything" product before...

[wagtunes]

Being catchy is what makes it popular.

A catchy tune is only one factor of many. It helps but it's not necessary to make music or an artist popular. Example: Kylie Minogue (before "Can't Get You Out Of My Head"). Popularity is way too complex to reduce it to catchy tunes. You can make the catchiest music ever and it will never become popular because you're not a good entertainer or you don't have the looks or you don't align with the current trend/zeitgeist/hype or you don't find the right label etc. Reality is complex.

This is also one of the biggest issues with AI. Fans think it's going to be the solution to everything, as if there was just a single barrier that keeps them from becoming popular. And AI companies encourage all those simple minds. You only need this piece of software and all your problems will be solved! Because that totally worked out with any other "key to everything" product before...

This is nothing new. This has been going on since the early days of popular music.

"Get an agent. He'll get you signed to a major label"

"Submit your songs to publishers. They'll get a big name to sing it "

"Take my songwriting course. It'll make you a hit songwriter."

I could go on and on and on and on and on. I know because I've lived through all these promises and more. I've been in the music business since 1977, probably before you were even born.

So how is AI any worse than any of these other things? It's not. It's not the thing that's the problem. It never was. It's what you do with it.

For example. Submitting songs to publishers. Make sure it's professionally done FIRST. Crappy demos will never get you anywhere. I wish I knew this when I first started. I had no chance with my horrible demos. And I do mean horrible.

The point is, it's not the tool. It's the person using it. And any tool, I don't care what it is, can be used for good OR bad.

It's up to the person to decide which it will be.

[DCrown]

@jamcat
and you would listen to it without someone forcing you by pointing the barrel of a gun at your forehead?
It always sounds like something I already heard before, no originality, no surprise and super overpolished.
I even know from what original song AI stole most things (billboard country no 1 song), but can't remember title and artist any more, maybe I will remember again. I didn't waste time on wasting time to analyse lyrics of ai song, are they good?
I am not surprised it is no. 1, the charts are full of ai and no one will remember any of these songs in two years.

[havran]

@jamcat
and you would listen to it without someone forcing you by pointing the barrel of a gun at your forehead?
It always sounds like something I already heard before, no originality, no surprise and super overpolished.
I even know from what original song AI stole most things (billboard country no 1 song), but can't remember title and artist any more, maybe I will remember again. I didn't waste time on wasting time to analyse lyrics of ai song, are they good?
I am not surprised it is no. 1, the charts are full of ai and no one will remember an of these songs in two years.

Perhaps as life goes one might begin to appreciate the quotidian and the familiar before one croaks.

(Oh! I see that you expanded your post while I was typing.)

[havran]

Maybe it could be said that the hope for novelty is a good thing, but let's not leave by the wayside an appreciation for the things we might take for granted.

[DCrown]

To make that country AI song, tell AI to make a song similar to "Human" (Rag n bone man 2016) with a deeper voice. The fake country song is a combination of two 2 songs I clearly hear, might find the 2 nd one too.
I am not into country music that much, but to categorize that song as country must be a joke, they would even categorize Little Red Corvette as a country song today.

I understand what it is all about.
The more musical knowledge you have the more different examples of different existing songs you could tell ai to do.
I dare say I could be very successful by creating ai songs, even though I don't know a lot about music between 2005 - now knowing I missed almost nothing anyway.

I am not a Beatles fan, haven't listened to any of their songs for many years, but could immediately sing and play The fool on the hill, Nowhere Man, why don't we do it in the road, Rocky Rackoon and many more and these songs were not even their hits (maybe Nowhere man only). Every ai song I listened to I forgot after 20 minutes (Alzheimer's disease effect).

Thanks for posting jamcat

First Beatles song I listen to after many years and I can't even express how much more I like it than all that ai crap



I mean that song is based on traditional blues progression. You might think what sense does it make to create another Blues song, so many blues songs already existed before that song.
It is about human creativity, passion, talent and soul and with a bit of weirdness the way Lennon sings and how cool the drums by Ringo and Paul's bass at 1:17 haha. Unique, original!

Another famous blues song example many years after Lennon. KISS by Prince. Prince added some funky vibe to a blues and dared sing a blues with his falsetto voice - a blues with such a high voice???! Third verse is a example of his screaming talent. There was a discussion of how tight and dry his mix was, not sure whether it pointed in a new direction . Unique, original!

Four more famous blues examples
Still got blues Garry Moore
Red House Hendrix
Belly Button Window Hendrix
Voodoo Chile Hendrix
How come you don't call me any more Prince
I love the blues, she heard my cry George Duke
Hard times Ray Charles

Every song a blues , every song unique and different, cuz humans are unique and different.
Starts at 0:48




omg, these horns and the piano.
A good example what makes a song great,
great voice, no one could sing like Ray, fantastic piano playing, fantastic horn arrangement and sad lyrics (that I remembered after first time I listened) what you would expect in a blues song.and natural not overcompressed and overpolished mix.

Ray Charles released it in 1954,
Prince released Kiss in 1986 or 85?
More than 30 years of creative (blues) songs in 4 decades.
Then I compare it to 2011 -2026, 15 years, the itb, autotune, social media and now ai era - I am speechless.
You will also notice that all singers could play one or even more instruments, it is not a must for a singer, but!

I would be happier about 50 new songs on that level every year - not just blues, of course - than 200.000 new uploaded tracks every day, the vast majority just boring noise.

@Wagtunes
Btw Prince sang it using a Sennheiser MD441, a microphone I recommend for your voice.

[jancivil]

how hard is it to accept that you can´t control what is smarter then you?

A machine that can't even count? That is your definition of smart?

I watched something yesterday or the day before regarding Apple's branding called the Large Reasoning Model. I first heard of it last year, though not under that name, strictly to do with the frustrations of its failings; it's still struggling today on the same problem I saw last year, they don't know how to get it to reason through anything very involved. The more it's fed in terms of the tokens applied to a task, the worse it does. It doesn't do what a human being does when things get more complex, which is spend more time studying a problem. It's essentially learned to look for shortcuts, so it's wasting everybody's time; to equate what it does with thinking is a sort of category error. It's not even a good calculator.

It's not intelligence. And to assert 'experts' have said there's no way to determine if it's conscious is baseless, who knows how you get there. It simply isn't, and what we have today is def. not going to pass the Turing test.

[jancivil]

blah blah blah blah blah as long as the results are good, using Ai to do this is not really any different

AI has no way to discern the quality of anything. It has no taste, no opinions, it has no way to experience anything at all. Confer Mary's Room, aka the Qualia Argument: Mary has never experienced the color red. She has all the information there'll ever be about it but until she ventures outside her room where there is the color red to experience, she has no actual clue. Qualia in the simplest terms is 'knowing what it's like'. Not even Buddy Rich is going to take it outside and show it what it's like. "No more clams or you're fired" is lost on it.

AI doesn't know what tone color is; it can be trained and directed to a result where the person operating it has something in mind, but it has nothing on its mind. There is no mind there.
A session musician doesn't have to be guided to every result, one is employed due to the success rate of knowing what to play right TF now. The equivalence is way beyond a stretch. I don't think anyone with any experience here will make such a blunder.

Ask Google "Can Suno be said to understand" and see what its AI has noticed.

I did the other day, today I got similar but not identical answers. One of them today is "Suno is a stochastic parrot at music, rearranging learned patterns rather than innovating or creating from a place of lived experience." I'd have to check to see if AI Overview's source was me or not ("Reddit +4").

But it's not wrong.

[AsPeeXXXVIII]

Plugins can be vibe coded. The human touch can't.

Make of that what you will.