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

[Zeisner]

Finding somebody who is skilled enough to make a difference (another me certainly won't do any good) is not easy to find.

I never said it's easy. You need to do plenty of networking.

Then, they'd have to be willing to work with me.

That's the biggest issue. You can't handle criticism and you get aggressive fast just because somebody has better ears than you do. Most audio engineers would avoid working with you, even if you paid well. You need to work on that.

Then they'd have to be willing to do it for a piece of the royalties because as I am now drowning in medical bills after my accident, I have no money to pay anybody.

Second biggest issue. But there are ways to earn more money by considering opportunities like making soundtracks for video games. Comes with the advantage that you usually don't need vocal performances. But you need to do networking to get a foot in the door and work for free with some indie developers (who also want to get a foot in the door) for their free games.

You certainly wouldn't work with me for nothing and according to you, you have great ears.

I used to work with people calling me names for doing my job for far too long, I'm not doing that anymore no matter how well they pay me. It's bad for my mental health and money can't fix that. I rather make less money (and sometimes even none) if I can work in a civil and productive manner. I can't work productively if I'm angry, it ruins the result.

Even if all of that wasn't an issue I would recommend to not use AI enhancements because editing all those artifacts out would take way too long, it's not worth the effort. I would instead try my best to work with your natural voice by using the right microphone, EQ and saturation. And suggest to attend workshops for voiceover artists (You can learn tricks there you won't see in singing courses, great for weak voices to sound more powerful). And I would ask you why you're not adjusting your compositions to your voice instead. That cracking tremolo part I already mentioned was the only interesting part of your performance.

I have good ears but not great ears. A colleague of mine (who does mastering) has great ears, he can hear amplitude changes of 0,1 dB which is crazy (I tested him). If he tells me that my guitars sound wobbly then I pay attention and say thanks.

[BONES]

We don't "pump out music for charts", we just pump out music we like and let the chips fall where they may. e.g. We went against the label on what the single should be, because we didn't particularly like the version of the song they wanted to put out, so we insisted it be the version we liked. Consequently it failed to chart but that's OK 'cause we still don't like the other version. We always knew there was a song in there somewhere, it just took a while to find it.

Good reviews and chart performance are just confirmation that what you're doing is maybe as good as you think/hope it is. At the end of the day, though, you can only do it for yourself. There's no mileage in trying to guess what other people might like or respond to, certainly not for us.

Yes, I'm aware what a Fairlight is and I was aware in the 80s that Jan Hammer used it. Nothing to do with AI though.

Are you aware that Fairlght was also the site of the second McDonalds in Australia, just a couple of streets away from where the fellas created the Fairlight CMI? Of course, like the CMI it is long gone now, turned into a block of flats with (presumably) stunning views of North Harbour and Manly Wharf.

I think you should start by listing your computer and operating system and how much RAM it has and then other gear and plugins in your signature.

You do that because it's helpful information for other users, especially if you do a lot of beta-testing. e.g. Every time I see someone whinging that Studio One won't work properly or it does this or that, I assume they are on a Mac but I don't usually know because no-one bothers to put that info in their signature. So instead I have to ask, which means a 24 hour delay before I know whether or not I might be able to assist. But I assume you never think about anyone but yourself.

[Tubeman]

He used Dr. T’s Algorithmic Composer software on a Commodore 64 with his Fairlight to do the Miami Vice stuff.

That has nothing to do with AI.

DR T’s Algorithm Composer is the conceptually same thing without prompting — it’s basically still the underpinnings used in AI music composition.

No, it's not.

[sellyoursoul]

We don't "pump out music for charts", we just pump out music we like and let the chips fall where they may. e.g. We went against the label on what the single should be, because we didn't particularly like the version of the song they wanted to put out, so we insisted it be the version we liked. Consequently it failed to chart but that's OK 'cause we still don't like the other version. We always knew there was a song in there somewhere, it just took a while to find it.

Good reviews and chart performance are just confirmation that what you're doing is maybe as good as you think/hope it is. At the end of the day, though, you can only do it for yourself. There's no mileage in trying to guess what other people might like or respond to, certainly not for us.

Fair enough. I still think we hold very different positions on what the creative process is. I'm of the mind of your buddy, where finishing something isn't necessarily the end goal. Not to say that it isn't good to shoot for finishing things (sometimes), but that is an aside. I guess I would say that it's a nice bonus when things arrive at a place of being finished. But to me, the creative process is it's own point, the joy of it (and insert many other descriptors here), the part that makes you feel alive in the moment of creating the thing that is coming out, playing with your own hands and internal self and bouncing things back and forth when playing with other people who you have good chemistry with. I can't imagine that process ever being equaled by AI generation.

[guitarzan]

He used Dr. T’s Algorithmic Composer software on a Commodore 64 with his Fairlight to do the Miami Vice stuff.

That has nothing to do with AI.

DR T’s Algorithm Composer is the conceptually same thing without prompting — it’s basically still the underpinnings used in AI music composition.

No, it's not.

Jan Hammer used Dr. T’s Algorithmic Composer software with his Fairlight. Algorithmic Composer evolved into Dr. T’s Tunesmith…

Dr. T (Emile Tobenfeld) just received the MIDI Lifetime Achievement Award at the 2026 NAMM Show. In the video about the event put out by The MIDI Association, toward the end, around 1:20, they say about Tunesmith:

“… its influence is seen today in generative music tools and AI assisted composition platforms…”

So I’m not that far off. Underneath it all AI is taking your prompts and manipulating the parameters of something that works quite like an algorithmic generative composer.

Besides, I thought everyone was against the idea of generative composition in general, otherwise what makes choosing a key and pressing an onscreen button to generate a piece of music any more intellectually involved than having a conversation through prompts with an LLM to arrive at the music?

What I think is overlooked is AI composition isn’t locked down to the automatic song generators.

There are AI composing tools that have MIDI in and out… so they are really just like the old algorithmic composers in that way, but now you have all the say you want in how those arpeggios or whatever are executed versus the old way of choosing a key and hitting a randomize button.

So if you’re not against all types of generative music, then I think the AI is actually much more engaging for the user.

[liquidsound]

Someone sent me this link trying to make me depressed...
Maybe...
https://shumer.dev/something-big-is-happening

[wagtunes]

Finding somebody who is skilled enough to make a difference (another me certainly won't do any good) is not easy to find.

I never said it's easy. You need to do plenty of networking.

Then, they'd have to be willing to work with me.

That's the biggest issue. You can't handle criticism and you get aggressive fast just because somebody has better ears than you do. Most audio engineers would avoid working with you, even if you paid well. You need to work on that.

Then they'd have to be willing to do it for a piece of the royalties because as I am now drowning in medical bills after my accident, I have no money to pay anybody.

Second biggest issue. But there are ways to earn more money by considering opportunities like making soundtracks for video games. Comes with the advantage that you usually don't need vocal performances. But you need to do networking to get a foot in the door and work for free with some indie developers (who also want to get a foot in the door) for their free games.

You certainly wouldn't work with me for nothing and according to you, you have great ears.

I used to work with people calling me names for doing my job for far too long, I'm not doing that anymore no matter how well they pay me. It's bad for my mental health and money can't fix that. I rather make less money (and sometimes even none) if I can work in a civil and productive manner. I can't work productively if I'm angry, it ruins the result.

Even if all of that wasn't an issue I would recommend to not use AI enhancements because editing all those artifacts out would take way too long, it's not worth the effort. I would instead try my best to work with your natural voice by using the right microphone, EQ and saturation. And suggest to attend workshops for voiceover artists (You can learn tricks there you won't see in singing courses, great for weak voices to sound more powerful). And I would ask you why you're not adjusting your compositions to your voice instead. That cracking tremolo part I already mentioned was the only interesting part of your performance.

I have good ears but not great ears. A colleague of mine (who does mastering) has great ears, he can hear amplitude changes of 0,1 dB which is crazy (I tested him). If he tells me that my guitars sound wobbly then I pay attention and say thanks.

I want to thank you for your kind reply. It was more than I deserved, FWIW, I can sing very simple songs. In fact, here's a song I sang a few years back. I think it's one of the better things I've done. The song itself has been covered by a fairly successful band and I am earning modest royalties from it.

Here it is. If you still think it's utter garbage (vocal wise) please let me know. Because this is about as good as it gets for me.

https://soundcloud.com/steven-wagenheim ... o-tomorrow

Any comments are welcome.

[Teksonik]

A.I. can never kill human "innovation"?

https://youtu.be/Te1HkBx7rDw?list=RDTe1HkBx7rDw

[4damind]

I actually think we might see more plugins being purchased, not fewer. It’s just that the tech is evolving. What sold 10 years ago won't cut it tomorrow. Instead of traditional circuit modeling, the next generation of "emulations" will likely be AI-driven models. We’re entering a completely new and exciting era.

I also don't believe AI will replace music entirely. We’re moving toward a hybrid workflow. Musicians are already using AI for lyrics, ideation or to fill gaps in instruments they don't play. However, the soul of the song - the composition and final production - will still happen in the DAW, guided by human intent.

As for music "AI slop" I expect it to remain a niche. Platforms will likely label or even restrict low-effort AI content to protect "real" artistry. Overall, I see this as a positive evolution rather than a reason for deep pessimism

[Tubeman]

I actually think we might see more plugins being purchased, not fewer. It’s just that the tech is evolving. What sold 10 years ago won't cut it tomorrow. Instead of traditional circuit modeling, the next generation of "emulations" will likely be AI-driven models. We’re entering a completely new and exciting era.

I also don't believe AI will replace music entirely. We’re moving toward a hybrid workflow. Musicians are already using AI for lyrics, ideation or to fill gaps in instruments they don't play. However, the soul of the song - the composition and final production - will still happen in the DAW, guided by human intent.

As for music "AI slop" I expect it to remain a niche. Platforms will likely label or even restrict low-effort AI content to protect "real" artistry. Overall, I see this as a positive evolution rather than a reason for deep pessimism

Except certain music genres make money mainly from touring live if even from that. You think those genres are gonna die away? I don't think so. People still want to see/hear good songs and musicianship live. I personally can't even stand the autotune-singing that has plagued most chart music nowdays. I suppose AI can sing better than those though

[4damind]

Except certain music genres make money mainly from touring live if even from that. You think those genres are gonna die away? I don't think so. People still want to see/hear good songs and musicianship live. I personally can't even stand the autotune-singing that has plagued most chart music nowdays. I suppose AI can sing better than those though

Definitely... but that's already a reality? The big money in the music business isn't made with streaming. Apart from commissioned work for TV and film, that's already the case today. AI won't play a major role there either, apart from controlling the technology on stage.
A laser avatar and an AI voice from a computer that sings? Well, something like that will certainly exist, but I don't think it will be the entire industry.

[Neon Breath]

What I wonder is what will happen when AI don’t have any new fresh food to feed on? Right now, fine it scrapes music out here previously created by artists, bands and musicians but after that, it will feed on its own stuff on and on and on? A photocopy of a photocopy of a photocopy of a photocopy….

[Teksonik]

What I wonder is what will happen when AI don’t have any new fresh food to feed on? Right now, fine it scrapes music out here previously created by artists, bands and musicians but after that, it will feed on its own stuff on and on and on? A photocopy of a photocopy of a photocopy of a photocopy….

A.I. can emulate but can it innovate? Can it make whole new genres or sub-genres? I don't know, maybe so and perhaps as it matures it will be able to create instead of recreate.

[BBFG#]

What I wonder is what will happen when AI don’t have any new fresh food to feed on? Right now, fine it scrapes music out here previously created by artists, bands and musicians but after that, it will feed on its own stuff on and on and on? A photocopy of a photocopy of a photocopy of a photocopy….

A.I. can emulate but can it innovate? Can it make whole new genres or sub-genres? I don't know, maybe so and perhaps as it matures it will be able to create instead of recreate.

The thought has crossed my mind also on whether it can F-up like a real human? That is, I remember many times where one member of the band made a glorious accident that took the production into an unexpected direction that went far above the original. AI can do it by a conflated consensus, but not by happenstance. There's also the question of synergy. Which can happen only when everyone suddenly plays better than they knew they could. The sum far exceeding the total of its parts.

One more thing. Back when I started this electric music life around eleven-twelve years old, I was continuously warned to learn my instruments without effects first. Lest I become dependant on them more than playing the instrument. It's easier to catch up on the technology of tools after than it is to catch up on the instrument after.
If you need a tool to compensate for what you lack, then use it.
But understand that a person who always uses a wheelchair when they have good legs will make their legs atrophy. That is where the creative process will suffer. And those it happens to may never even notice.

[guitarzan]

What I wonder is what will happen when AI don’t have any new fresh food to feed on? Right now, fine it scrapes music out here previously created by artists, bands and musicians but after that, it will feed on its own stuff on and on and on? A photocopy of a photocopy of a photocopy of a photocopy….

A.I. can emulate but can it innovate? Can it make whole new genres or sub-genres? I don't know, maybe so and perhaps as it matures it will be able to create instead of recreate.

If you can innovate the prompts to give to AI it will attempt to create a tune in the new genre you wish to create — then you keep hammering at it with prompts until it produces what you had in mind. You create the genre and innovate, AI does its best to comply with your prompts. It will keep trying to shape the project until you either decide you have what you want or you give up. AI doesn’t decide any of that.