Imagine having to replace every plugin in your last project with a piece of hardware. Find the money. Find somewhere to put it. Wire it all together. Then see how much enthusiasm you have left for making the track.
For plenty of producers, that would be the end of the session before it started. Software let more of us get on with making music. I struggle with the idea that making the next person’s starting point easier should suddenly be a threat to the culture.
In 2007, I wrote my BA Music dissertation on the symbiotic relationship between technology and dance music. Nearly twenty years later, AI has brought me back to that relationship. My instinct is to welcome the possibility that more people will get to make something worth hearing.
The price of admission has changed before
Hardware has a physical price of admission. Every additional synth or processor takes money and space. You can love the experience of using it and still recognise how many people that excludes.
The history of dance music includes people finding their way around those costs. Roland’s account of the TR-808 describes how falling second-hand prices helped put it into new hands. Those producers heard possibilities in its electronic character and gave it a life beyond the job of imitating a drum kit.
That is the relationship I mean by symbiotic. A manufacturer builds a machine. Musicians discover a use for it, or push one characteristic until it becomes the whole point. That sound creates demand. Developers respond, and the next generation of producers starts from somewhere different.
Software widened the opening again. VST arrived in 1996; effects, and later instruments, could live inside the recording environment. A growing collection of plugins could take the place of purchases that once needed another shelf and another set of cables.
A computer still cost money, and learning to make a decent record still took work. But you could get further with less equipment. You could try things you would never have been able to justify buying in hardware.
Would we really want to give back the music that became possible because of that?
A studio on a laptop is only part of the answer
You can own all the right tools and still have no idea how to get the sound in your head out of the speakers. That gap can be miserable. An evening disappears into menus and settings, and the idea you were excited about starts to feel like a chore.
I can see AI helping at that point. Someone might hear a bass rhythm clearly but struggle to program it. Someone else might need a few musical variations to work against before they recognise what they actually want. Getting something into the session gives them a chance to respond to it.
This approach predates the current argument about AI music. Magenta Studio’s 2019 tools could generate or continue MIDI phrases inside Ableton Live. The output was material a producer could edit. You could keep a few notes, change the rhythm or throw the lot away.
That kind of interaction interests me. A suggestion can provoke a decision. You hear it and realise the phrase needs to arrive later, or that the last note is the only bit you like. You have found a direction by having something to push against.
And producers are unlikely to use AI only in the ways its developers expect. Someone will obsess over an odd artefact or feed a result through a process that changes its character completely. I suspect some of the most interesting music will come from that sort of misuse. Dance music has plenty of room for an experiment that sounds wrong until somebody makes it work.
There is more at stake than making more tracks
I do draw a distinction between working with a musical suggestion and asking a system to deliver a finished record. AI can make far more of the musical decisions than a software synth does. That changes the conversation about authorship and involvement, and the comparison with earlier tools only takes us so far.
I’m wary of treating the difficulty of the old process as evidence of its artistic value. Learning to hear what a track needs is worth the effort. Spending an evening fighting a workflow may simply be an evening you never get back.
Making music easier will not make every result interesting. There will be work you have no desire to hear. There will also be people who surprise you, including people who would never have got started without the help. I think that possibility matters more than protecting the obstacles they would previously have faced.
The part I want producers to keep is their judgement. In a dance track, moving one sound slightly can change the way everything leans into the next beat. Leaving a bassline out for another bar can make its return feel completely different. Having more ways to reach those decisions could be a good thing.
That is why I’m interested in where AI goes. The relationship I studied in 2007 is still moving, and it will be shaped by the people who get their hands on the tools. Somewhere among them is a producer whose first record I would like to hear. I don’t need them to have had the hardest possible time making it.