Am I staring extinction in the face?’: classical music and AI

[Martin Walker]

A thought-provoking article from the Guardian today, illustrating that even classical composers are beginning to worry about AI:

‘I’m a composer. Am I staring extinction in the face?’: classical music and AI

https://www.theguardian.com/music/2025/ ... er-radio-3

[SafeandSound Mastering]

There have been computer programs that attempt to make classical inspired music for decades.

It is not news. 15 years old already.

https://www.slate.com/articles/arts/mus ... _bach.html

And I think it has been happening since the 90s.

Automastering happening for a decade. It's not mastering, it's utter rubbish. It is not even close at any level. It is merely an artificial pretend service for people who know no better.

I find ay aye really boring. Isn't that the elephant in the room ?

Involving yourself in ay aye is surely negating your own capacity of expression, surely ? It's just so incredibly dull.

It is a vast waste of time and energy in my personal opinion. My personal views is not from fear. I don't fear ay aye. It is just you can barely go through a day before you read or see something related to it so it is in your face at the moment.

I think people will see through the vacuity.

I do use a stem sep tool, but never for the job itself (did it today to advise on a snare level) I ask for the tweak which is the right way to do it. No shortcuts.

I spend time on more wholesome activity and values.

I hope ay aye helps for medicinal breakthroughs.

[Hugh Robjohns]

For centuries, Humans have developed skills, and their offspring have learned those same skills and improved upon them and shaped them, and so on and so on with each generation.

Every composer today creates music based on what they've learned about all those who have gone before them... Just as every mastering engineer shapes music based on the skills and preferences they've learned from those that went before them.

The point is that AI does all that learning incredibly quickly, and the algorithms that create true innovation and inventiveness are also being developed just as quickly. The 'vacuity' of early systems won't last long...

And I can foresee a world where 'professional creatives' really do become 'hobbyists' creating for pleasure while most paid jobs go to AI systems which are easier, faster and cheaper to work with, delivering as good or better output!

[Drew Stephenson]

I am increasingly of the opinion that if you don't have either an existing fan base or an act that you can perform live, then there is very little scope for any kind of career in composition or song-writing.

I had expected it to take a bit longer to hit the classical world though.

[SafeandSound Mastering]

I feel that is merely technical and creative conjecture. A guess.

I think the missing aspect in your post is authentic living value.

Numbers are not life, they are not the universe, they are merely an abstract, a derivative, a breaking down of. Clearly a useful one but the use is never independent of funding sources.

I maintain that the vacuous nature of ay aye will rekindle and reinvigorate what being a human being is, and beyond that.

Obviously this is related somewhat to music and creativity less so rather more mundane data related activities.

From personal experience I have made life changes recently that in part, were influenced by ay aye, however not in an accepting of ay aye.

Through negation.

There were other deeper factors just to be clear. Changes that respect what life is in actuality.

For myself thus far ay aye is entirely meaningless.

I did not bury my head in the sand or become involved. I don't feel I want to give my energy to fuel its momentum, certainly not in musical related activity.

That's a feeling based intuition. Which is a very important form of intelligence in itself.

Maybe I'll change my view if it saves my or a loved one's life in the future.

[RichardT]

I think it won't be long till AI generates music that really touches peoples' emotions, and then the game may change.

In the past, some manual skills have survived automation, but the manual market becomes a smaller share of the total. I expect the same may occur here, but who knows for sure?

Maybe there will be hybrid bands, where AI and humans play and improvise alongside each other. That could be quite interesting.

[Hugh Robjohns]

I think the missing aspect in your post is authentic living value.

I disagree. I think we naturally like the believe that there is something 'special' and 'indefinable' about the exclusivity of human creativity... but I don't think that's really the case. It's just sufficiently complex that it has appeared that way... until now.

After all, our brains really are nothing more than logical machines in the computer sense. Incredibly complicated ones that build unique neural nets... but we now have technology that can do the same things and the power behind AI and its exponentially increasing complexity will soon allow digital machines to appear just as creative and innovative (if not more so) as humans — if that's what they're programmed to be.

I maintain that the vacuous nature of ay aye will rekindle and reinvigorate what being a human being is, and beyond that.

I'm not convinced... but time will tell. Thankfully, my career and pension don't depend on the outcome!

[merlyn]

Creative activities are the last thing that humans need help with. Creative activities are fun. Useful technology would empty the bins, or make a cup of tea, doing mundane stuff to free humans up to do creative stuff. But nope, Silicon Valley tech bros have targeted the fun bit.

The article linked above centres around a vulgar display of wealth -- an opulent house in San Francisco. The point that is maybe being missed is that we're getting LLMs whether we want it or not, and whether it is any good or not.

I came up with a bunk analogy, which seem to be popular. It's like instant coffee. Instant coffee is not the same as coffee brewed from beans but it's in the ballpark. In the ballpark enough to sell successfully, and fill supermarket shelves. It's not about whether instant coffee is good, the same as brewed coffee or anything other than that is what we get. Similarly, we'll get LLM output whether or not it's as good as the real thing.

Perhaps a bit bleak but "LLMs give the wealthy access to skills and deprive the skilled of access to wealth."

[Albatross]

It might be good for toothpaste ads etc so the library composers are likely going to have problems. People will always seek out the authentic because the creator attaches meaning to the creation. Its why the artist's back-story is so important to the marketing of the work.

[RichardT]

Nobody has a clue how humans produce music, but I agree with Hugh there's no reason to suppose there is anything involved other than brain cells and sense organs.

The value we put on human music is due the fact that we relate to and value people, and relate to the expression of human ideas and feelings.

But for most purposes, we don't get music of that quality even now - and AI can surely replace that. I would guess that film, TV music and jingle composers are under severe threat.

[SafeandSound Mastering]

I think the missing aspect in your post is authentic living value.

I disagree. I think we naturally like the believe that there is something 'special' and 'indefinable' about the exclusivity of human creativity... but I don't think that's really the case. It's just sufficiently complex that it has appeared that way... until now.

After all, our brains really are nothing more than logical machines in the computer sense. Incredibly complicated ones that build unique neural nets... but we now have technology that can do the same things and the power behind AI and its exponentially increasing complexity will soon allow digital machines to appear just as creative and innovative (if not more so) as humans — if that's what they're programmed to be.

I maintain that the vacuous nature of ay aye will rekindle and reinvigorate what being a human being is, and beyond that.

I'm not convinced... but time will tell. Thankfully, my career and pension don't depend on the outcome!

Human creativity is about 700,000,000 years in the making.

Our brains did not evolve independently of our heart. Or the rest of our beautiful and utterly amazing immensely intricate biochemistry. All of which supports actual intelligence.

Our brain is not a neural net, it is a intrinsic element of a complete, interdependent, indivisible biological organism based on existencially evolved intelligence beyond anything known in the universe.

A slight difference.

[Hugh Robjohns]

You may be right... I'm still not convinced. But we'll find out soon enough!

[SafeandSound Mastering]

I feel there is a possibility that human beings may have a reorientation of what real intelligence is.

There are about 360,000 baby humans being born today, and as far as can be predicted, every subsequent day into the future.

The entirety is worth some deep consideration.

[The Elf]

I think it won't be long till AI generates music that really touches peoples' emotions, and then the game may change.

I was already there a couple of years ago. I gave it a prompt and the music that returned I absolutely wish I had written. It was fabulous and I listen back to it often. Even the lyrics were inspired, albeit it the AI misunderstood part of my prompt, yet managed to make something clever of it.

But I know I can't use it...

[OneWorld]

Surely the question should be “Will my music become extinct?” Because sure as the sun rises in the East, we’ll become extinct. Bach et al live on through their music, why can’t we?

Is the point of writing music an attempt to live in perpetuity? Don’t look for fame, it’ll find you anyway if it’s meant to.

And anyway, depends what a writer’s objective is, write music for oneself and that would satiate one’s creative lusts, or seek to make and sell a product, and therein lies the rub, one thing to create the product, then the hard work starts, marketing it, and I reckon a live presence is critical. Look at Taylor Swift, in reality, her music isn’t novel, usually going round Cmaj, Gmail, Amin, Fmaj etc, good melodies, a canny business head, and a grafter extraordinaire.

Taylor Swift during her Tiny Desk Concert, speaking on the subject of songwriting, she gets up in the morning, goes to work, she always turns up, but the ideas don’t. Then other times, she’ll find herself getting up at 3am, stumbling about in her pjs towards the piano, an idea has just come into her head and she “just has to get it down” we all know the feeling, I don’t think that will ever become extinct.

We should remember AI is a tool, one that can access an amount of data increasing at an exponential rate, but i5 isn’t doing anything novel, when it can bring people back from the dead, or make teleporting possible, then I might be impressed.

It is disappointing that the same old circumstances prevail, DeepSeek is already being dissed, because of its origins, chip supply restricted, to throttle progress, for political reasons - we all know the issues and the players, we have this supposedly amazing world changing technology, lumbered with the ante-deluvian psychology of adversity, division and xenophobia, what’s so intelligent about that? Those issues don’t seem as they’ll become extinct any day soon.

[OneWorld]

I think it won't be long till AI generates music that really touches peoples' emotions, and then the game may change.

I was already there a couple of years ago. I gave it a prompt and the music that returned I absolutely wish I had written. It was fabulous and I listen back to it often. Even the lyrics were inspired, albeit it the AI misunderstood part of my prompt, yet managed to make something clever of it.

But I know I can't use it...

Surely that’s the whole point, AI is. tool, and can write alongside us, but you created the original idea?

Say for example “using the string quartet format, write a piece, based in bEmaj, modulating to Amaj, use chromatic changes and throughout an ostinato…etc etc etc” it would take me eons, even if I had the basic structure done, taking into account each instrument’s idiosyncrasies, trying to avoid clashes, but I believe AI could knock that of stuff out all day long and my aspirations to become the next Bach, if not that Debussy will do, till then noodlizing on the guitar will do

[amanise]

I think that the future for humans lies in live performance. For example, being a pianist and taking bookings, like a sort of Uber sessionista, to go round to a person's house who is wealthy enough to own a piano, and a house to put it in, but hasn't the talent to play it. And never will. Then sitting and playing Rachmaninov hits while they and their friends eat dinner. Like a pet for the night.

[Eddy Deegan]

I partly feel this is the second spin of the roundabout for me. I've always written and recorded music because I want to but other than my time in the '90s bands and duo (where the money was made from live performances albeit selling a few tapes and t-shirts here and there) I've never sought to do it for a living. When I was a performer I did it for the fun and the fact I loved the music we were playing although most of it had been written by other people. Indeed the few times I did anything creatively commercial (jingles and a soundtrack) left me gasping for less.

Back then, and for decades previous, the industry had plenty of great bands and their output was always going to be better/shinier/more commercial (delete as appropriate) than my efforts were. I knew it at the time and I never had any illusions otherwise. I kept creating my own music though because I wanted to.

My efforts back then were in "competition" with well-established groups that dominated the market. My efforts today are in "competition" with AI, changes of public taste or anything else one cares to mention. I put competition in quotes because to me it's not really about competition. It's about doing something I like to do. Some people love playing golf but 99.9% don't do it to win the Ryder cup. They do it because they like doing it.

AI is just another form of A&R to me. They used to publicise a lot of junk alongside the good stuff too but I didn't care, I'd just not listen to it. I didn't want compete then and don't want to now but I'll never stop endeavouring to write music I like to listen back to.

That's my selfish perspective though. I do feel for those who have forged a position in the industry only to have AI take their clients in the way AI can and does. I suspect that many consumers of AI-generated products are less discerning than one might wish.

[Sam Spoons]

Consumers of popular music have always been less discerning than musicians and composers might prefer but the chase for the lowest common denominator seems to have been accelerating since the '50s. 'Serious' music (classical, jazz, prog, etc.) has been niche for at least 200 years, possibly much longer, as long as there is a core of people who enjoy 'real' music AI will struggle to be considered as 'proper' music.

[merlyn]

It's worth noting that GPT 5 didn't live up to the hype. Before GPT 5 was launched Sam Altman was hyping it with statements to the media like "GPT 5 scares me" and "GPT 5 is now at PhD level in everything". After the launch he said " sorry, we made a mess of that".

OpenAI, the company behind ChatGPT, hasn't been floated on the stock market yet. Hype, hype, hype, have an IPO then retire to a private island or maybe the moon. So much money has been sunk into LLMs now that they're going to get their money back somehow. And that's the usual 'give it away for free until people depend on it, then start charging for it'. Then, when they've got their money back, run away and let the whole thing collapse.

LLMs have their uses, but the hype is so overdone it's beyond belief. We're getting into Russ Andrews territory.

The flaw in OpenAI's logic was that scaling an LLM up is always going to make it better. There was a huge step up from GPT 3 to GPT 4 that was obtained by scaling up. People in the field have pointed out that pure scaling isn't going to get to AGI (artificial general intelligence -- parity with a human) or ASI (artificial superintelligence -- more intelligent than a human). Looks like Gary Marcus was right.

[OneWorld]

It's worth noting that GPT 5 didn't live up to the hype. Before GPT 5 was launched Sam Altman was hyping it with statements to the media like "GPT 5 scares me" and "GPT 5 is now at PhD level in everything". After the launch he said " sorry, we made a mess of that".

OpenAI, the company behind ChatGPT, hasn't been floated on the stock market yet. Hype, hype, hype, have an IPO then retire to a private island or maybe the moon. So much money has been sunk into LLMs now that they're going to get their money back somehow. And that's the usual 'give it away for free until people depend on it, then start charging for it'. Then, when they've got their money back, run away and let the whole thing collapse.

LLMs have their uses, but the hype is so overdone it's beyond belief. We're getting into Russ Andrews territory.

The flaw in OpenAI's logic was that scaling an LLM up is always going to make it better. There was a huge step up from GPT 3 to GPT 4 that was obtained by scaling up. People in the field have pointed out that pure scaling isn't going to get to AGI (artificial general intelligence -- parity with a human) or ASI (artificial superintelligence -- more intelligent than a human). Looks like Gary Marcus was right.

Yes, that's all it is, hype. They are very skilled at building a story around something, no matter how inane.

However it gets them talked about, free publicity. Mentioned in every news bulletin (The bloke that owns Oracle and has now bought TikTokUSA, and CNN and many more, expect more 'breaking news' relating to tishtosh)

There is that much that the plebs presume that whatever app/product/whatever is being mentioned, is the go to product, it becomes ubiquitous, the makers know the Average Joe won't read past the headlines and go research that product, and conclude "Hang on a min, I'm being sold a pup, do they think we are that gormless?"

"Yep"

Same as we hear "Coca-Cola" anywhere and every where on an infinite loop, people buy it when they think "Thirsty" not aware it has buckets of sugar in it.

Same with AI, mention it and ChatGPT springs to mind, despite there being many others of its equal and better.

Will robots ever make babies?

------------------------------------------------------------------------------------

Quip of the day - Bono, in his trademark melodramatic mood, at a fund raising jamboree, encourages people to go into a slow clap, and he says, with a grave voice.....

"Every toyme I clap me hands, a Choyld in Africa Doys"

And a bloke in the audience quips.....

"Well stop clapping then"

[Arpangel]

And I can foresee a world where 'professional creatives' really do become 'hobbyists' creating for pleasure while most paid jobs go to AI systems which are easier, faster and cheaper to work with, delivering as good or better output!

Easier, cheaper, faster, do you want a world where they are the most important criteria and reasons for living and doing anything?
AFAIC if that’s the case then just stand back and let it take over, I’m glad I’m not going to be around to see this.
All I see these days are people staring at screens like zombies, parents give phones to their kids to shut them up, when they should be out playing, discovering nature, enjoying the real world, and communicating with real people, no wonder so many kids are "diagnosed" these days.

[SafeandSound Mastering]

I have zero interest in AI for music personally, but others of course have freedom to do as they wish to. If they find it somehow enriching.

With the broader debate.

Contemplating this, in part, with other life events I have started reflecting on real intelligence as being that which serves nature. Of which human beings, in the main, psychologically seem to have seperated themselves much to their detriment.

Nature is intelligence. Evolved intelligence, evolved in the sense that we rely on an intelligence that we barely even consider on a daily basis, our blood cells, mitochondria, motor proteins, the mind bogglingly complex biological activity that supports consciousness. Anyone seen a motor protein ?

Actual:

https://www.youtube.com/watch?v=gpjcW-ltOFo&t=11s

Animation:

https://www.youtube.com/watch?v=mBo_o0iO68U

A form of real intelligence is born of consciousness, something that a bunch of copper is fundamentally incapable of. There is a scientific statement known as the "Hard problem" of consciousness. In short, how does sentience appear from a coalescence of molecules arranged in a certain way.

On a tangent I am sure many people watched the Brain programme that was on the BBC. (Secrets of the brain) The first thing I considered after the end of programme 2 was that I never once heard the word 'consciousness' through 2 hours of talking about the evolution of the brain. Remarkable.

I feel only biological organisms have true intelligence, evolved and brought about by consciousness or just their existence like a tree is part of broad biological existential intelligence.

It is vey apt it has been named artificial, it is the ideal distinction.

It seems the lack of control means that ay aye does not seem to be in servitude to humanity. It will be predominantly in servitude to funding sources and that which can be most efficiently exploited.

[RichardT]

A form of real intelligence is born of consciousness, something that a bunch of copper is fundamentally incapable of. There is a scientific statement known as the "Hard problem" of consciousness. In short, how does sentience appear from a coalescence of molecules arranged in a certain way.

Yes, the ‘hard problem’ really is a hard one. If someone said to every skilled system designer: ‘create a system that has an experience of the world like humans do’, I don’t think a single one would have a clue how to go about it.

There seem to be a few possible answers.

It’s possible that experiencing the world arises, by some miracle, from systems of high enough complexity.

Or there is some kind of ‘soul’ that has the experience - but that simply begs the question of how a soul, whatever that might be, has an experience.

Or maybe some quantum effects underlie this - this is Roger Penrose’s theory of consciousness.

At the moment, we simply don’t know. I don’t think there’s any way we can rule out 1 or 3, so it may well be possible that in the future a bunch of copper will have an experience!

[OneWorld]

When ever there is a radical change there are tales of doom and disaster, when the steam train came out doctors said travelling on a train will weaken your heart, or whatever. When margarine came out and all those polyunsaturated things I was told they made your kids bi-sexual. OK not as bad as becoming extinct but still quite concerning, so we'll get over AI just like anything else, we only have two choices, no choice and no choice at all.

Half the world has to wake in the morning and wonder where the next meal is coming from let alone shatGPT