What Happens When Every Producer Can Build Their Own Tools?

I opened ChatGPT this morning and it suggested I build an interactive mix bus compression demo.

Which was strange.

Because that's basically what I've spent the last 9 months doing – and the demo it described was to go with a post from a few days ago.

Underneath the suggestion was a line about creative professionals using it to develop educational content, build interactive music tools and make visual assets.

Which is part of my working week.

Whether it has picked up on what I've been doing, or whether a lot of people have simply arrived at the same place at the same time, I don't know.

Either way, I think it points towards something bigger.

People who work creatively in the digital world can increasingly build the tools they want to work with.

And for music producers, I think that's going to change the kind of software we use.

We've Been Able To Do This For Years

Musicians building their own software isn't new. Max/MSP has been around for decades, and so have Reaktor, Pure Data and SuperCollider. People have built extraordinary instruments, sequencers, effects and entire performance systems with them.

But there was always a threshold.

You had to learn the environment. You had to understand its logic – objects, signals, patching, routing, variables, whatever the particular system required.

You didn't necessarily have to become a traditional programmer. But you had to become a programmer of sorts, and you had to be prepared to put the time into learning it.

What's changing isn't that musicians can suddenly build software.

It's how little you need to understand before you can start building.

I can begin with an idea – I want something that demonstrates subtractive synthesis, a harmonically rich waveform that lets you hear what happens as you filter parts of it away – and start from there.

That's a very different entry point.

The Tool Can Start With The Idea

This is the part that interests me.

I've built tools to help me write music better. I've built sequencers around chords, basslines and melody. I've experimented with ways of demonstrating compression, rhythm and synthesis.

But the important bit isn't really the software.

It's that the tools contain my way of thinking.

Imagine I'm working and realise I want something very specific.

Not another compressor.

I want something that can give me fresh fills whenever I want one. Maybe a rhythm tool that approaches groove the way I've learned to approach it. Or something that generates basslines the way I like to work, giving me ideas to choose from as starting points for a track.

Some of these might become products. Others might only ever be useful to me.

But they don't need a market before they're worth building.

And that's where I see us getting more personal software.

Commercial music software has to justify the cost of development, testing, support, compatibility and maintenance. So there has to be enough demand to make building it worthwhile.

EQs, compressors, reverbs, delays, synths and limiters are the fundamental tools we use to make records.

But when software becomes quicker and easier to experiment with, an idea doesn't need to appeal to everybody to be worth building.

Someone can build a device around one peculiar production technique they've developed. Someone else can make a sequencer based entirely around the way they hear rhythm. A mix engineer can turn one of their working methods into an interactive tool.

Some of them might make absolutely no sense to 99% of musicians and become indispensable to the remaining 1%.

And some weird little tool made by one person to solve one problem could find its way into somebody else's setup.

They might use it completely differently.

That's where it gets interesting.

A producer can express an idea through a record. They can explain it through an article.

And now they can build the idea into software.

From A ZX81 To This

This is probably why the whole thing blows my mind as much as it does.

One of my earliest experiences with computers was programming a ZX81. Back then, getting a computer to do anything at all felt like an achievement.

Then computers became part of making music: MIDI, sequencers, samplers, DAWs, plugins, software instruments. I've watched each stage open another door.

But this one feels different.

I don't need to know exactly what it is when I start.

I can explore the idea by building it.

That's something I've dreamed about being able to do for years.

There's Still A Gap

None of this means making reliable music software has suddenly become easy.

Getting an idea working in a browser is one thing. Making something I'd trust with real-time audio, plugin formats, different operating systems, DAWs and updates is another.

AI can produce code that looks convincing and fails in practice.

There's a huge difference between an interactive compression demonstration and a compressor I'd put across my mix bus at four in the morning while finishing a record.

But the most interesting ideas don't have to begin as finished products.

They can begin as experiments.

The Producer Becomes The Developer

I don't think plugin companies are going anywhere.

But there's now another layer underneath that industry.

Individual software.

Producer software.

Artist software.

They won't necessarily be better because they’re made with AI.

They'll be better because they carry the flavour of the person who made them.

We've already seen versions of this with the Waves Signature Series, and most recently for me, Jaycen Joshua's God Particle. It isn't sold on its circuit topology. It's sold on the fact that it's his chain – the one he actually uses – his quirks, his preferences, his order of operations, baked in and handed over. You're not buying a compressor. You're buying a way of working.

Anybody whose craft already happens in the digital world is in the same position – photographers, writers, designers, editors. Small tools, made by the people who actually do the work, shaped around the way they happen to do it.

The tools can start coming from us now.

So perhaps the interesting question isn't:

What will AI do to music production?

It's:

What happens to music when every producer can build the tools they wish existed?