I think as they get better at making these great middle of the road songs, edgy music will reemerge. Whatever AI will be good at, will immediately devalue, in the realm of arts. Just how photography gave way to non-realistic art and how drum machines made sloppy drums (or ridiculous apex twins) hip. Then, the artifacts we see now as flaws will create its own sub genre.

So I see three ways this flows: mediocrity will be even more available, which will make artists who make mediocare music even less succesfull, pushing all human music further in human direction, except those using the unwanted artifacts of this new tech to create new sub genres.

Music, art, fashion is in the end all about changes. What we make now mostly means something in relation to what was already there. It's a big conversation, spanning millenia, and this isn't the last word.

I've been playing around with it for a few days now. While I agree that it seems impossible to create songs with a more "sophisticated structure" (for lack of a better word off the top of my head), you still can get better results by fine-tuning, as is always the case.

If you just request "rock music" or "jazz", you get very dull, generic variants of the requested style. But on a second thought, isn't that exactly what should happen? You throw all the rock music on this planet in a blender, turn it on, and what you get is the most average rock music there is.

If you spend some time spicing up your prompt with flowery language or just a bunch of adjectives, you can get a sound that seems less bland. When you supply lyrics, using square brackets to denote verses, chorus and bridges can also result in a somewhat more structured song, but I found that the AI is pretty lackluster in that regard and you often need several attempts until it follows these inline orders.

So yes, in its current form it has mostly a novelty factor, this is Stable Diffusion for music, but I can easily see this being useful for a small indie gamedev who needs some BGM, or an alternative to the YouTube music library. Instrumental sounds fine, it's mostly vocals that have this clear digital distortion if you pay attention. It's surprisingly good, but still bad.

Hmm, perhaps there is a business model there: an elevator that writes music about the people inside the elevator, and adapts as the situation changes.

“Red-headed lady, gonna eat an apple. She breathes on it first, wipes it on her bloooouuuuse…”

Even if AI gets way better, the one thing that I don't foresee changing is what makes things valuable and or desirable. Sparsity. If everyone has it or can create it, it's not special. I think the GP was referring to this sparsity.

However, if later down the line we create autonomous agents that can initiate the creation of said music themselves then I'd call that enough agency to say that the machine is "making music". Could probably almost do it now; tailored LLMs, image diffusion, music diffusion and you could have an ML agent that acts as a musical artist; posts to instagram etc with images of their persona "working on something new", releasing new tracks, bantering, etc. There are already AI OF stars apparently.

We could say "yeah but a human still set it up and told it to make music" but I would discount that; pretty sure no human has total agency, we are all impacted by our environment, peers, culture and all sorts of other external influences.

And I don't think sparsity changes things (maybe in the material world) but culture certainly does. Things are popular because they appeal to us in bulk, rarity/sparsity always result in higher effort for the payoff and so decreases popularity.

In the field of AI, only game playing / graph search has gotten to that level of superhuman capability.

This will be used for jingles and scoring by low cost studios and marketers. Making money at music was already hard and will only get harder. But maybe this makes sense in a way, music creation just became much easier and accessible to more people.

Alongside this prompt based music creation, we have AI powered autotune and voice masking, which allows even the worst singers to sing perfectly. Popular music was already being retuned during recording and this just makes it easier. In hip hop, old songs beats and verses are getting reused wholesale with little to no modification. See Jack Harlow’s First Class and Tyga’s Bops Goin Brazy. A lot of musical success is now business connections, studio promotion and timing, not being extraordinarily talented. Most rappers now don’t write lyrics down or freestyle, they record a line, pause, record another line, pause , erase the last one, rerecord, and then they’re done. This type of thing an LLM is made for. I think this means less works of art on the ‘radio’ and more garbage music forcefed by studios.

I assume the 2 minute limit has a lot to do with this. Often I find Suno generates anywhere between 40 seconds to 2 minutes, with the shorter generations having less of a recognizable structure as a song. If instead it was 3 to 4 minutes, I think it'd be radically different.

Might also just push us in the direction of less good music, if you can’t make any money as a musician before you are a genius.

Perhaps this has more to do with the input rather than the capability of the AI. The license is somewhat meandering.

Also, there is a neat build-up to the ALL CAPS portion, when the song takes on a much more full and powerful quality. Plus the whisper on (the "Software"). It's obviously responding well to the features of the input itself.

  > drum machines made sloppy drums (or ridiculous apex twins) hip.

I want to joke and say Lars from Metallica but I'd be lying. He was a decent drummer in the first couple albums.

Maybe sloppy isn't the best term here. More like not being beat-perfect. E.g. using slightly off-beat hits or slight tempo variation as a way to emphasize other instruments or lyrics.

Pre drum computers these changes would more likely have been considered to be sloppy drumming, whereas these are more likely to signal authenticity these days.

I should know, I'm a sloppy drummer ;)

Even the most abstract art is not about splashing paint in a canvas and calling it a day. It's about doing something close to this, but creatively and within a specific framework.

It takes much more skill to produce songs with hooks and refrains than to produce random music.

But it takes even more skill to produce something that is creative, not formulaic, but still has some structure that's pleasant/fun/interesting/etc to the listener.

Only time will show if the next step will happen anyway. My gut feeling tells me that AI art will gain acceptance over time, and we will just think of it as "art" or "music", just as we did with recorded mysic and synthetic sounds.

I see little value to humanity in tools that are able to generate an endless amount of music derived from existing music, specifically designed to neatly slot into the place of human artists. We gain little in return from that.

Some people will make an argument like, this lets people generate lots of low-quality music for use in elevators or grocery stores. Well, there is already a massive oversupply of completely free music which can do that. Do people pretend to not know this?

The other weak argument is that it lets people express themselves who haven't studied or practiced music. But, it doesn't, because the interfaces (text prompts or "upload an existing file") are designed to take the place of a human being given instructions for criteria to fill, as if they were a worker, not an expression of the person giving the instructions. If the person giving the instructions were expressing themselves, most of the AI tool would not be redundant. It's as expressive as telling another person to write a song for you with some instructions. Hardly expressive at all.

And the quality of music generated by AI is increasing geometrically. Be careful to consider that any music heard today will be among the least quality generated by AI. Because it will get better with practice, at the speed of light.

That first comment I can get behind, but probably not in the way intended. Recording technology made performers less valuable, so fewer could make a living. The AI composers (and performers) will go further down that road. My conclusion? Get used to it. Just like human-adding-machines got replaced by calculators. And a generation of weavers, replaced by looms guided by punched cards.

This phenomenon is not new, and will go down a well-worn path.

This is nothing new, your examples are great ones too, the benefits we got are, lots more art for everybody. To take the side of the composer/performer/artist is natural I suppose, but that's such a tiny fraction of humanity. To ignore the clear and positive upside, even claim it doesn't exist, is disingenuous?

To criticise AI music for being the same as human music, is a curious thing to say if one doesn't want to distinguish the two somehow? I missed that point entirely I guess. I don't know why that sentence was there at all then.

This is still not relevant. You don’t need to bring it up again. I wasn’t criticizing it for that.

As for the rest of your comment, you didn’t state any upsides for AI generated music, but you did spend time to attack me, so I think we’re done here.

For augmenting comoposers, sure, GenAI can be a tool like others. Musicians have been incorporating rhythms and melodies shipped with their electronic instruments for ages.

Entire genres have been defined by sounds and synth presets, too.

So I do see a bit of the similarities that you describe, but I think this is largely misleading.

The argument is that each technology advance accompanied resistance followed by adaptation. "Recorded music" was arguably as paradigmatically disruptive over "live music" as "AI generated music" will be.

From my perspective the implications of this are dire. AI can completely remove the human element. The skill, creativity, and collaboration required to produce music is a big part of my appreciation for it. Once that's gone, when Spotify can generate exactly what I want to hear, Music as we know it loses its value.

Somehow this nudged me a bit when I switched to YTM, although the bundle with YT background playback on iOS was probably the bigger nudge. There's plenty of AI-generated music on YouTube as well, but for the moment their recommendation algorithm and catalogue is just better for me.

For now, it also doesn't seem to suggest me any AI-generated music whenever I use autoplay or suggestions.

I'm not falling for the illusion that this won't change though.

You're not wrong. A paradigm shift is not an incremental change but a disruption of fundamentals.

> From my perspective the implications of this are dire.

These changes are scary, especially as people try to come to practical terms with the new reality.

> AI can completely remove the human element. The skill, creativity, and collaboration required to produce music is a big part of my appreciation for it.

I still hate autotune. I feel that it ruined music. But, on the other hand, it allowed people who were excellent musicians but terrible singers to make excellent music, even masterpieces. I don't think autotune was a paradigm shift really, but it was pretty disruptive.

People are deeply creative, social, collaborative, musical, artistic, hierarchical and status conscious. These traits will always drive people to make music and share it, and derive meaning from it. People will still pay other people for the music they make.

Photography utterly disrupted the social role that painters held as documentors. No one needed to hire a good painter to have a portrait. They could hire a photographer more cheaply for more accurate documentation. Artists working in the medium of painting really had to grapple with the question of what art is, if academic faithful representation of reality is no longer valued by society. Painting thus began to change. Impressionism led to Suprematism, Dadaism, Surrealism, Abstract Expressionism, Conceptualism, Modernism, Post Modernism.

Artists today will have to grapple with similar questions raised by AI generated art. But humans are creative, indomitable, curious and tenacious. I am absolutely excited to see the art that future human artists will make in the face of all of this.

Another side of it is that it will enable the creation of more music around more topics than before, by non-musicians. The accessibility bar is lower.

For a lot of people, music is a way to express their emotions, and not just by creating/playing it, but by listening to it. Now, you'll be make your own hyper-specific music with lyrics around topics specific to you, without learning any of the underlying skills yourself.

I've certainly wanted some kinds of music/representation in music of some of my experiences to exist, but not enough to go out and learn to make it myself. Now/soon I should be able to do that with AI tools, and I think that's actually neat!

> as giving those instructions to a person to create a song for you

Most people don't have access to such a person.

A tool like this will open up a lot of possibilities.

No, they won't turn into the next Hans Zimmer.

Not a lot of individuals, to be honest. Only a handful of people make any kind of living from composing. Millions try it, but have to be content with performing for their friends. Which will continue unchanged.

So the actual economic impact of AI music will be different from the scenarios being described. What is true is, we will all have a lot more musical listening choices. Which is a net good for the rest of us?

For a normal person, whether something is art or not, is a mix of 1) whether they like it, 2) whether they can, or conceivably could, enjoy it together with other people, and 3) whether they're supposed to enjoy it or call it art, because other people claim they do (social proof).

Examples:

- Pop songs are strongly 1, 2 and 3a (enjoyment), but not necessarily 3b (considered High Art). Most people don't care, or couldn't even tell, if the songs they like were written and performed by actual humans or by machines; they experience them through some machine anyway.

- Paintings. I recently visited a Van Gogh exhibition, and I can't honestly say I liked most of it. Most paintings, in general, are ugly. We call them art because we're supposed to call some paint scribbles on a canvas art, particularly when they're framed and put in a museum (as opposed to bought off the street!) and decreed Art by People In Authority Over What Is or Isn't Art. For this exhibit in particular, my ability to enjoy the paintings was proportional to how much I knew about Vincent van Gogh's life - for those paintings I had some context for, I enjoyed them even though they're otherwise pretty bad to me. But most people, most of the time, don't have any context for paintings they're viewing, and they still call them art.

Hell, arguably, the best "paintings" in that exhibit were a couple that were obviously AI-generated - like Vincent wearing VR goggles, or animated Vincent inviting the patrons to the exhibit.

Nah, what I think is death of art for regular people is quantity and personalization. The most important aspect of day-to-day art experience is that you can enjoy it together with people around you. It's a problem for TV shows and books these days, and even more with "Internet original" videos - there's just so many of them, and with everyone's getting their own personalized feed, it's getting hard to find common creative works you and your conversational partners both seen. Everyone's experience is becoming disjoint from everyone else's (except for occasional superhero or wizard movie) - at which point you eventually realize that enjoying unique art no one else has is pointless waste of life.

For me, one problem with AI art is that there isn’t context for any of it. It’s just there and didn’t come from anywhere.

Eventually, some prompt engineer may build up a coherent body of work that is moving or provocative, but we aren’t there yet.

P.S. I don't mind people using AI as tools, but the human agency needs to be visible and relatable. If it's only a prompt, don't ask why I don't value it.

The prompt can be detailed, creative, and innovative. Kinda like a composer comes up with an idea for a new piece. But now the composer won't need the technical ability to translate it into musical notation.

All our current AIs need prompts to produce content. So, you'd need to enter a prompt so that AI can produce a prompt for you.

But I guess you could develop some variant which just generates prompts in a cycle. But most of the generated prompts will pretty bad, so it will in turn produce heaps of crap.

That's where the human element is still needed - try a prompt, evaluate the result, tweak the prompt again. If this doesn't produce a good result, try a different idea. It's not even that different from how artists produce content these days (the create <-> evaluation cycle). Current AI is not good enough to be able to judge the produced content (that's why it produces so much crap after all).

> I've not heard of any art classes that train you by copying everything that has been produced.

You don't have to learn it, because it's the second nature for humans. We learn by imitation. Babies learn to talk by imitating sounds from parents. Artists learn by imitating style by masters. That's what generative AI does as well.

All of these had immense impact on the way we create (or make art). And despite all this, we still use waterpaint, perform music on ancient instruments, make furniture with minimalistic tools, or use clay to make objects.

I'm not pessimistic about generative AI. If anything, It'll allow more people to create. Allow new and unprecedented art forms. It will have an effect on the way people make money with art. But so did photoshop, digital audio mixers, a table saw, and a CAD/CAM machines.

Like the banana. The bananas isn't art, taping a random banana to a random wall and eating a banana isn't art. A particular person taping a banana to a particular wall for a particular reason is what make the whole thing art.

I definitely agree that you shouldn't. To me absolutist arguments about if AI-generated content is art or not makes no sense. AI is a tool, like for example carpentry, that can be used for many purposes. Most carpentry isn't art, it's craftsmen building useful things and solving practical problems. But there are definitely artists out there using carpentry as the medium in which they express their artistic intent.

> we got the synthesizer. Again we got the same complaints

I honestly don't believe this happened. Citations?

You make it sound like one day there was no recording and then bam! flac quality recordings of musicians, out of the blue. You do realize it started with exceedingly shitty wax cylinders that sounded absolutely atrocious by today standards (and by the past standards as well)

>After the release of The Jazz Singer in 1927, all bets were off for live musicians who played in movie theaters. Thanks to synchronized sound, the use of live musicians was unnecessary — and perhaps a larger sin, old-fashioned. In 1930 the American Federation of Musicians formed a new organization called the Music Defense League and launched a scathing ad campaign to fight the advance of this terrible menace known as recorded sound.

>The Music Defense League spent over $500,000, running ads in newspapers throughout the United States and Canada. The ads pleaded with the public to demand humans play their music (be it in movie or stage theaters), rather than some cold, unseen machine.

>Joseph N. Weber, the president of the American Federation of Musicians, made it clear in the March, 1931 issue of Modern Mechanix magazine that the very soul of art was at stake in this battle against the machines.

https://www.smithsonianmag.com/history/musicians-wage-war-ag...

Algorithmic music has already been around for decades and it never became popular. In the 90's it was of interest only to a small group of academic music nerds. The same is true today. Avant-garde shit for nerds. No one wants to listen to it.

Dunno, but it could probably get published and I'd at least read the comments when it gets posted to HN.

Tricky bit is the design in the signaling and instructions so as to avoid biasing the results while still allowing the desired 'both are human' response. Something like a check box for each chat if that chat was a robot. If you have the budget I'd also run the control Turing test, human x computer, as well as a computer x computer test.

About a third of people are highly suggestible and agreeable, they readily follow authority, fall for scams and can be hypnotised. About a third cannot be easily conned or hypnotised, are vigilant, "disagreeable" and likely to go their own way.

Even of that highly sceptical, independent third, it takes a lot of courage to say "I totally know this is bullshit, and I'm done". Only a handful of Migram's subjects were able, not just to assert a moral objection to hurting a fellow human being, but to question or see through the whole ruse of fake scientists and stooges. Even if you can spot that, it takes rather a rare human to act on it, to call it out and walk away - or rather escape the parameters (get outside the box) you've been placed within.

Maybe there will be a change of feeling, it's starting to come to me, instead of seeing this AI generated content as "soulless" etc I'm starting to see it as an extension of OUR human generated work. It's more like an endless remix of HUMAN talent.

All of that is boring though. The exciting stuff is all that will be displaced, and how we will solve the myth or meritocracy.

There will be a TON of these 60-second tracks, and no one wants to listen for anywhere near that long while skimming a feed, so these will now be ignored.

IMHO this and things like it are basically a sub-class of comedy or satire, so I have time for this sort of thing. It's a joke, and should (can?) only be appreciated as artistic as someone saying "Wouldn't it be funny if ... ?" because now that casual thought can be turned into a pretty instant "well here it is! LOL!".

I don't think you're supposed to appreciate it as music. Maybe I'm calling it wrong though.

(edit - I will admit that upon further thought, I'm not sure how I feel about this when compared to, for example, Nina Gordon recording "Straight Outta Compton" as an accoustic, mildly lamenting singer-songwriter style number 20-some years ago. It's clearly in the same satrical arena but one took a lot more effort and imagination. Kinda, because there was a lot of effort and imagination that went into both the training data and the model, even if this specific output was only a passing joke. It's quite hard to reason about this stuff...)

A lot of satirical songs are absolute bangers, because (for example) to do a send-up of the tropes of X music, you must know all the tropes of X music, and be able to perform them. So a lot of satirical music is actually done by people with a lot of skill and passion for the thing being satirised.

And because satirists don't have to worry about being predictable or unoriginal, they can put in more crowd-pleasing cliches per minute than 'serious' artists, giving them the most intense X of all X artists.

(Not saying the MIT license is a banger though - just that some satirical songs are)

I think it's a bit like everyone who said "black cabs in London have nothing to fear; X years of 'The Knowledge' will always be better than a guy with an App in an old Prius..." Except it was all wrong. Black cabs are probably still superior, but the bulk of people (and especially new customers) are all using Uber, because it's easier and they just don't care.

It sucks that production factory, literally built to make money, AI created junk music will be a thing, but it most likely will. Someone will exploit that they can make a ton of cash with low risk and budget. They'll have the connections to, despite initial (somewhat) faux outrage from the public/press, get radio play/playlisted online ("probably even 'ironic' addition from the hipster music crowd"). The song will be an earworm, and aside from the musicians that hear all the flaws, the passive every day mom and pop listeners will get the hook stuck in their head and it'll just be another great tune like any other.

I don't think we'll replace 'celebrity', I think that'll still happen, and I think maybe greater appreciation will happen to 'real musicians' (with a face!). But 'everyday' disposable pop (like all those one hit wonder tunes that are still played in nightclubs, pubs, throwback radio, wedding discos) - that's going to be disrupted massively.

I guess my point was that I don't hate the linked tune because it's not someone asking me to listen to a song, it's a joke and one that does have some humour to it.

I can absolutely see this tech (if it's allowed to) replacing a lot of working musicians who do music for ads, jingles, tv, film etc. And yes, disposable pop is probably on the chopping block.

> disposable pop ... that's going to be disrupted massively.

I wonder how the disruption will play out - the world is already drowning in content created by humans. If we envision that AI can make disposable pop to the same standard (and I have no reason to doubt it) then that surely creates an absolute deluge, almost boundless in size, of stuff. Promotion will become more or less the only art, to make things stand out from the crowd, and that can probably only continue for situations in which people want a shared experience. For an awful lot of situations the streaming of entirely ephemeral audio would probably do. It could as easily kill 'pop' as a business on the audio side, as it could steal it.

I'm just sorta daydreaming about possible outcomes here. All sorts could happen.

There's a craving for a deeper connection but that's usually the smaller part of our everyday consumption.

If you are a little bit of a critical listener you listen to those songs because you connect to them somehow, that's the power of music, it's a language for emotions that you don't need to know how the artists look like, or their backgrounds, to feel what they try to convey. Having the context/background might help to intellectualise a piece of music but the feeling comes from the art itself.

AI music is pure entertainment, not art.

It's no different than using Photoshop.

Unless you only consider art done with oil and on a canvas to be the "real thing."

And sometimes not knowing is better like the example with Rammstein and row zero were some manager woman was asking other woman if they want to meet their rockstars.

Btw. on a good electronic music set, knowing the arist is probably more a quality sign than a personal aspect. Knowing that i like what arist xy does, means i might like to keep an eye on future work because i like the style of it.

Nonetheless, i do also think that fandom and doing real tours will still be the unique things for bands. There should always be a market for human content.

AI music will disrupt this market.

Spotify for example will try to produce their own AI music (like they already do with regular music). The Christmas playlist will then mainly contain their music. That saves a lot of money for the.

But have also had the non-surprising "Hedonistic Fatigue" that comes with excess access to something originally valued. I have now been able to generate 4 and 5 digit numbers of pictures of awesome colourful steam locomotives and epic dungeon vistas, but now find myself fatigued by "what on earth am I going to use 5000 dungeon pictures for?", coupled with the dread of being forced to CHOOSE from 5000 options. And I learn the known principle, that when you can choose from 4 options, you are happy you picked the best of 4, but when you can pick from 5000 options, you are left feeling inadequate with "I almost certainly was not able to pick the best of those 5000 options, and trying to do so would exhaust me". So suddenly, picking something from your menu of options, feels dreadful and fatiguing.. (I get the same feeling sometimes, when trying to pick a movie to watch out of 16.000 options).

So yeah, no doubt the AI sketch/refinement tool will be merged into our creative process, but for the time being, I feel a second generation of alienation-estrangement with my "available options".

I've been thinking about it the same as you but something that I think we're missing here is what can the person who can already make 5k pictures by hand do with this kind of tool?

It makes me think of the later albums that Frank Zappa produced with a Synclavier. That guy went hog-wild on that thing and banged out some unbelievable albums.

What would he have been able to do with AI generated music technology if it was the Synclavier of his time? What are the Frank Zappas of our time going to do with AI?

https://en.wikipedia.org/wiki/Frank_Zappa#Synclavier_works

And then if it was restricted to a few parties, they still have to compete with human artists, they can't just charge whatever.

What will be worthy is imperfect human-created art. We already went through this decades ago with expensive hand made vs cheap machine made stuff, this is just another iteration.

I am not saying it will be great to be an artist, just like in the past few famous live in limelight and most will struggle to stay afloat.

Yeah, the last two live mega tours (Taylor Swift and Beyoncé) have a tad more personality than the average artist, but the usual stuff that you would hear on the radio might as well be AI generated and live-performed by animatronics and a significant chunk of the audience wouldn't care or even notice.

I mean people watch Love Island on TV for chrisssssakes. Humans can definitely be mindless consumers, where the sugar salt and fat in fast-food is the same as false drama, outrage, sex and violence in media. We all got buttons and they're so easy to push. Just look how popular TT is versus the type of content on there; mostly short-lived mindless stuff.

TBH that's a pretty low bar; "radio" songs have been engineered and polished for a very long time now. I have no hard numbers but gut feeling says radio music only represents 1% of the music industry.

What I would like to see is what happens if you ask it to do in 5/8, 5/4 or any of the 7/* variants and see how that goes (I have no idea, might actually work). There are some examples in common music but not a lot especially for the more obscure ones. Unless they also trained it on a lot of international folk songs.

And yes it can do odd time signatures. It can do all sorts of genres like art pop, cinematic scores, ambient, math rock, avant garde, baroque, etc.

I’ve found it fun to take a coworker’s comment (like in slack, if it’s long enough) then hand-modify it into a song with “[Verse X]” and “[Chorus]” headers on blocks. Sometimes I’ll also run it through an LLM to do like what you did. Anyway it’s been a lot of fun and everyone’s gotten a kick out of it.

Encore, encore!

https://glicol.org/

and the destination is:

https://github.com/chaosprint/RaveForce

Although my progress on AI is slow atm, I found that the copilot in VS code can already help me in live coding performances several times:

https://youtu.be/xzIXzt3hSt0?si=rVihHYiKiAU5IKeI&t=389

Also want to hear your feedback.

Maybe try sub-lye-sense Or similar.

Like many people have said, this tech will only get better.

I use it sometimes in FL Studio when creating electronic music (plugin is called Vocodex).

Presumably they take the AI-generated voice and generate midi notes, and apply a vocoder to the voice, following the notes.

My guess is that they train the vocals and the music separately, the training data is trivial to create from any tracks with tools like with https://vocalremover.org/.

Creativity though? These non-poetic songs (I heard “Lorem Ipsum” the other day) are a fun novelty because a machine creates this in no time. A skilled musician will do something similar.

Innovation? Deepmind showed new creative approaches. Will this happen with music too? Or will they just regurgitate styles from the past?

It’s definitely good for Muzak, where you want a non-offensive, slightly upbeat, and not-too-reconizable background. And perhaps as an idea generator.

That's an often overlooked difference between earlier projects like DeepMinds AlphaGo and the current genAI craze. With something like Go you can dial in the exact parameters to solve for (winning a game of Go) and then let the AI experiment and find novel ways to get there. But with image generation, or music, or language models the measure of quality is too vague to define in a way that the AI can directly optimize towards, so they're instead just fitted to existing examples.

So weird.

(Like the separation of Authors and Copyright holders implies things I needdd to think three times about)

IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM!

Why is it worthless? Because the point of art was to communicate or convey something with other people and the AI has no idea because its not human.

A few words or a sequence of sounds can be enough to transfer great deal of feeling and meaning because we run about the same software and as a result we can generate the same output with a little bit of input. This is all done by looking inside and externalise it, that is someone feels something and makes a song from it and that song can be used to regenerate feelings in other people.

The current AI tech doesn't have a way to do that because doesn't have a way to look inside. At best, it can imitate things within some context but the output doesn't have any meaning at all. The most successful AI content was maybe the "Pope wearing Balenciaga" image but that wasn't because the AI thought it mean something but because someone looked inside and thought this can be interesting.

So no, AI isn't taking over the creative process. AI is taking over the mechanical part of it only, that is the part where the artist traditionally had to master a method of production or an instrument.

The AI evangelists keep pushing short videos or drawings that look "professional" and claiming that Hollywood is done, artists are screwed etc but those are worthless outside of the context that AI made it. No one is interested in paying or even spending time to consume this content, its extremely dull.

On the wall, you will find an Obligatory Art. Sometimes it's just a canvas with 3-4 stripes of paint: you can imagine a purely mechanical process for churning these out; a conveyor belt with brushes hanging over it, perhaps. Other times it's a little more creative. Each room is slightly different. You can also sometimes see these in cheap home decor shops. It may not be much, but it does the job - it really does make the space more pleasant than just blank walls would be.

There are a lot of rooms to fill. Someone has to make all these. It may not be all that creative, but it sure beats working in, say, a produce packing plant. Meanwhile, it's hard to make a living in art - some are wildly successful, yes, but the tip of that pyramid is very small and getting there takes as much luck as skill; and there are a lot of people further down the pyramid who also need to eat while waiting for their big break.

Those are the jobs at risk from generative AI in its current state.

But these jobs never existed in first place, that generic art is done by contractors who charge for the materials.

It boils down to the economic model and the financial and political choices like in every creative industry.

Regarding potential displacement, I would apply the stock photography theory to any creative industry. Ask yourself: is what I do in my creative endeavor the equivalent of stock content for the visual imaging industry? If the answer is yes, you might want to future proof your craft. If the answer is no (as in, your art is more than a simple soulless piece of easily digested and quantity-oriented content) then you will be fine in the long run after the current unsustainable hype cycle dies out.

The stock photography stuff is either documenting event or displaying low effort illustration for low effort productions. I guess AI can be good at churning Apple images for low effort Apple news.

This is mixing up art with the art industry. Artists will struggle just like copywriters are struggling after the arrival of LLMs. Not everything in the art industry is trying to break new artistic ground or communicate some deep emotion to the listsener. For much of the industry, "good enough" will suffice if it's 10x cheaper.

Maybe with exception of strictly technical work like people who remove background in mages or calibrate instruments. Those people are screwed yes.

Personally I think knowing the story behind music makes it better. The music isn't to everyone's taste, but for example Devin Townsend's wiki page / story is a trip: https://en.wikipedia.org/wiki/Devin_Townsend

Like so many people felt defrauded when they discovered that the Milli Vanilli leads didn't actually sing, and that wasn't even AI. https://en.wikipedia.org/wiki/Milli_Vanilli

Edit: I might add that I already suspect any illustration that even superficially looks it might have been generated by AI. This has ruined the enjoyment of so many people's artwork whose style has been co-opted by AI.

If a friend recited a poem, would it matter to do if they read it off the Internet or composed it themselves? Of course it would.

If someone tells you they love you, does it matter if they are a robot or an honest human or a con-artist human catfishing you? YES, THAT MATTERS TO YOU. Yes you "refuse to enjoy" things that have suspicious sources.

But, I barely know any musical artist's name today. Most music is just something to listen to at the gym. It's pleasant enough, but I don't dig in to know who is singing at all. Every source is equivalent to me, as long as it's pleasant to the ear.

Perhaps you're different. The only question is, which attitude is more prevalent?

This song sounds like a disturbing uncanny-valley rendition of a slow Phoebe Bridges song performed by Taylor Swift using a broken auto tuner.

While I think it's technically impressive that this exists at all, I think this tune is still at the stage of "Pope in a puffer jacket".

At first it will be kept secret, and then as it spread more and more among the industry, it will be more or less stated, but by that point people will be already accustomed.

This states from the outset that it is using AI tools, at which point I become more understanding. Someone had an idea, but lacked the singing voice or a friend with a singing voice & free time, so used a tool to fill the gap. This is better, at very least more honest, use of tech as a tool than, for instance, autotune on studio albums, IMO.

If you _really_ want it with a real human voice, perhaps contact some of the many performers on social media to suggest it might be an amusing way to generate some content to monetise. Or, of course, sing it yourself!

--

[1] I've gone from clicking very few of facebook's “recommend for you” articles to clicking absolutely none of them – the number that are, or are indistinguishable from, hallucinations from an LLM that doesn't understand what is actually being written about, already dwarfs things that are worth reading. SciFi TV/film/book reviews and essays seem to be particularly affected, with “local” news links not far behind.

[2] “you won't believe this isn't AI generated!” — no, I won't, because it quite obviously is. I don't know whether to be insulted that you think just saying that will convince me otherwise or sad for the state of humanity that many do seem fooled.

[3] Too many people seem to think that slapping code out of copilot into a stackoverflow answer without nothing to check it for correctness in any way is acceptable, and before that was possible there was already too much bad (sometimes working but blatantly insecure) code out there that people were blindly copying. And that is before the potential licensing & moral issues that mean I have not yet been convinced to use anything like copilot myself, but I'm getting far off-topic here…

Someone making a "proper" song where they just sang these words would be quite boring, then I'd rather spend my time checking out other music.

Why does this being computer generated ruin it for you?

Auto-tune has been around since 1997 so it's not like computers have not been a big part of a lot of music we hear every day.

Listening to songs, as speaking with people, is in large part about enjoying the causes of the song rather than the mere variations in pitch.

Beethoven's 5th even, purely instrumental, is enjoyable because of how the composer is clearly playing with you.

To generate pitch variations identical to beethovens fifth makes this an illusion, one hard to sustain if you know its an illusion. It isnt an illusion in the case of the 5th itself: beethoven really had those desires.

What you describe certainly exists, but it's not the entirety of art, and I would argue that at this point it's not even most of art.

Meanwhile, "To what extent can a piece of art be a thing that is of interest in itself, divorced from its creator, context, or any representation of anything in particular?" is absolutely a valid area for people to explore and one that artists are exploring all the time. There is now one more tool to play with in the toolbox.

I heard the exact same objections to the modtracker scene three decades ago - "it's just computer generated slop, I'm not interested unless it's a real person performing on real instruments". I maintain that not only was it a perfectly valid mode of expression then, but tools like Ableton grew in part from those experiences and are an integral part of much - most? - music now.

AI-generated art may have a bit of the former (assuming the human had enough control over the details of the final output), but has practically none of the latter.

Hence, AI-generated output is not art. But art can be produced using AI tools somewhere in the process.

(1) When I look at art produced by one of those "artists" that commission the actual work to someone else, it's similar (I don't recognize the "artist" as the human I'm connecting to, ideas are a dime a dozen). However, it's still art because I can connect with the anonymous human which actually implemented it.

If we get AGI, I could imagine feeling something towards the art such an entity creates, since a big part of the human experience that we would probably share with an AGI is inescapable death.

But for today's "AI" generated music, I feel the same towards it as I would towards the random step function output of a given tool in Ableton - sounds cool, now what can we do with that to make it into music?

So a human using the tools of sound production is what transforms the function output into music. (Please let me know if I’m misunderstanding you).

I think I see what you’re saying, but that’s already happened here hasn’t it? I mean, it's not as though an AI made the decision to generate this all by itself, a human had an idea to create this piece and wrote a prompt which created this output.

The order is of events is reversed from your Ableton example, but I would contend that this kind of production is no less musical than what someone could create using a DAW, simply that the tools are more accessible

(and I presume there is less direct control over what the end result is going to sound like, but the same could be said of conducting an orchestra versus playing a piano.)

Eta: For example, some people in this thread have complained that the AI generated voice falls into the uncanny valley. I agree, and I think that’s part of the art here.

Like, if Trent Reznor had produced Hurt not by putting his doubts, self loathing and pain into words and music, rather by typing "sad, trending on artstation" into a console then heading for lunch, I don't think it would be any way as meaningful even if it was note for note beat for beat the same output.

I knew nothing about Trent Reznor the first time I heard "Hurt". Often when a song is heard on the radio - perhaps a sentence that dates me, but even so - there is no explanation of where it came from or even what it's called to accompany it; or perhaps there may have been, but the listener wasn't paying attention until after they realised they liked what they were hearing; indeed, there used to be an entire industry for solving the problem of "I heard a song I like and want to know more about it, or at the very least find out what it's called so I can hear it again".

When I first heard "Hurt", it resonated because of how those sounds interacted with my own experience. Everything else came after. Had those exact same sounds any other origin, that first experience would not have been affected - I would have had no way to know.

This is reductionist IMO. The equivalent seems to me to be, "the meaning the reader of words imparts to the meaning is constructed in the reader's head..." but clearly the vast majority of the meaning of the words is derived from the writer's intention. Of course that can be misinterpreted, reinterpreted, co-opted, etc, but regardless, it doesn't mean the author can be simply ignored, or that a psuedo-random generation can be treated the same as human-generated.

This is not clear at all.

Written words are just marks on a surface. Whoever made them may have intended to convey something, but they made them and walked away; they are now absent, taking their intents with them; only the marks remain, and those are not sentient or even alive - they contain no intent. There is nothing about the patterns left behind in themselves that makes them different from any other patterns the universe contains as far as the universe is concerned. If handed a set of marks on paper with no other information, you have no way of knowing for sure how they came to be. You could guess, you could be super confident, but you couldn't be /certain/.

If a reader later comes along who happens to have studied the same pattern-codes as the creator of the marks, however, seeing them will make that reader recall the associations and build up meanings in their head. These may or may not be the same meanings the writer intended to convey.

Children learning to read understand this very well - reading is /hard/, associating meaning with code is /hard/, decoding similar meanings to everyone else is /hard/, aligning the associations spoken words trigger in your head with others around you takes /study/ and /effort/, even realising that you end up with different meaning in your head when presented with some symbol to what others get, though frequent, takes deliberate effort. Reading comprehension questions in elementary school tests are there for a reason.

To a well-practiced reader, the process is natural and seamless, and feels like telepathy; it /feels like/ meaning has been transferred directly from the writer to the reader. But it is not that, and many problems arise when people forget this.

Once you have learned to seamlessly decode symbols into meaning in your head and the process is fully automatic, symbols you encounter in the world will seamlessly trigger meanings in your head regardless of their origin.

This is the human condition: we are all locked inside our own heads, and you can't take a piece of yourself and place it inside another directly. The best we can do is shout into the void and hope something similar inside the person across from you resonates; but it turns out that human shouts are not the only thing that can make those strings vibrate.

When we encounter combinations of symbols in the world that trigger complex meaning inside us, we /expect/ them to have an author who intended to convey something like that meaning, because in the entire history of human experience to date, the only other way for such things to appear in the world has been incredibly unlikely coincidence, invariably accompanied by context that makes it clear it is coincidence. (A certain proportion of social media content is, in fact, people sharing instances of these coincidences!)

However, this is an assumption that we make, and the world is rapidly changing in ways that mean it may, going forward, no longer be a valid one.

More and more, we will encounter combinations of symbols in the world that trigger complex meanings in our heads but originate from no human intent beyond "I need a combination of symbols that will trigger these meanings in the heads of those who encounter them", if even that much is explicit. We are rapidly improving the processes that produce them, and one of the ways we are improving them is removing tells. You will see the symbols, they will decode into stories in your head, and you won't know for sure if a human author was involved or not.

It is vital, going forward, that we all remember that symbols triggering satisfying meaning in our head does not automatically imply deliberate human intent for that meaning in their origin, lest we be entirely unprepared for the brave new oncoming world, just the way a chunk of the populace was unprepared for nigerian prince emails in their heyday, or are still unprepared for telephone calls from "internet tech support" right now.

Conducting an orchestra is an important role but the music is mostly a result of first of all the composer, and then the conductor / arranger's interpretation as well as the skill of the musicians. I really don't see the similarity to a human input of "GNU license, sad, jazzy." The resolution is just way too rough.

In fact, imagine comparing the experience of reading Snow Crash, to reading the sentence, "Cyberpunk story with sci fi elements, VR universe, pizza delivery guy with samurai sword."

I would readily compare the experience of reading Snow Crash (one of the first SciFi books I read of my own volition) to the output that a LLM may produce from such a prompt. My iPhone informs me that I’ve spent nearly 10 hours playing with Characters.ai in which SF storytelling characters are my favorite to interact with. When I first read Snow Crash I felt like “finally, an author that understands that part of the story that _I’m_ interested in!” and my experiences of AI driven creative writing has felt similar. Certainly it feels less “magical” since I’m aware that I’m customizing the author to my personal taste - is that “magic” of feeling connected to the artist *the* art?

While I don't disagree there's a "human touch" to art, I'm not convinced it can't be synthesized to some degree. It may not be innovative, but I think since AI is extrapolating from learned data, it can at least mimic the current pop culture.