Hey everyone! I’ve been thinking a lot recently about the idea of people being able to build their own tools using AI. I think it’s a really extraordinary thing.
We know very well that a piece of technology shapes the things it produces. We’ve been thinking a lot about how genAI is creating a whole bunch of “AI art”. In this case, AI technology is very much shaping the products (it’s making it directly!)
But another way AI will influence creative work in the future is through its ability to allow people to build their own custom tools. In this setup, human users can decide what they would like in their tool, deciding ultimately how they would like the technology to influence their outputs.
This is pretty cool because for most of human history, artists have mostly used technologies built by others to make their work. In music, DAWs, plugins, hardware and software are just the latest versions of that. But now, any artist can dream up an idea for their ideal workstation, decide its constraints and limitations, and its functions and possibilities, and make the best artistic workspace for themselves. I believe this will result in such a diverse range of outputs. The modern DAW is great, but it’s not the only option for a space to make music, and I’m sure that now regular non-software-engineers can start build their own workspaces, we’ll see some really creative forms of tools to make music in.
The fact that people can now build tools without having to rely on developers to build them, or learn programming themselves may have huge implications.
I’ve also been thinking about this in terms of slightly different angles from the usual discourse surrounding AI and tool-building. It’s tempting to immediately think about how these tools could be sold to others, or how they could become a business opportunity. But right now, the main thing I’m interested in is how we can make tools that do exactly what we want them to do, in the way we want them to do it, laid out in the exact form we want them to be in. All of this steps around the usual question of what the market wants or expects.
At the moment, I’m more interested in people building tools for themselves—and I think it’s worthwhile that people explore this themselves too.
Because of this, I’m much more interested in the philosophical implications of making tools for ourselves than I am in the discussions around implications of productivity or commerce.
A lot of people outright dismiss AI because of the obvious, negative uses of it: getting it to write essays, generate tonnes of content, flood the internet with slop. But those are the simplest applications of it: get AI to make ‘the thing’.
What feels more interesting to me is using AI to help people build the tools they actually want, with the feature set they actually need, in order to make ‘the thing’ themselves.
A disclaimer: I know this is contested territory, but I should note here though that I believe there is a place for AI to generate material itself. I’m a massive fan of some of the innovative works in the creative fields that have come from AI directly making audio or visual material itself, or drastically processing it in new ways. There is absolutely so much garbage being generated, but I believe there is still some pretty extraordinary uses of it.
I’ve found it common to have tools only do about half of what I want. Because of this, I use a little part of one tool, then another tool to cover its gaps, and another one after that. It becomes a patchwork of tools. But what if more people could build tools for themselves that did 100% of what they wanted, in the exact way they wanted them to, laid out and designed in a way that matched how they think?
Imagine if more people had their own DAW, custom built for themselves, with custom effects, sequencers, and timelines. Some people might want traditional DAW layouts, while others might want more modular workflows. Some might stick with coding interfaces, while others might want to draw shapes to make music. The point is not that one of these is the ‘correct’ way. The point is that the layout of a tool is related to the mind using it. The tool is correct for whoever designed it, but if it’s passed to another user, it might go against how that person thinks about things like workflow and function.
This feels especially important to me. Software markets tend to accommodate some minds more so than others. Some people are forced to work inside tools that only sort of fit the way they think. They patch together tools and workflows, deal with compatibility issues, and rack up subscription costs, all because there isn’t a tool that really matches their mind. I think we’re entering a time where that person may be able to build their own custom tool with the exact workflow and feature set they’re after.
And because tools shape outputs, more unique tools should lead to more unique outcomes. That’s why I don’t think the creative fields are simply going to wither under AI. I think there’s a real chance they could thrive through these means.
It all ties back to the nature of technology. Technologies are things that do things in specific ways. In software, they manipulate data toward specific ends. When we rely on a DAW like Live, we’re accepting the engineers’ and designers’ techniques for how those ends are reached. That’s not necessarily a bad thing. But we’re moving toward a point where the average user has more say over how those ends are reached. We’re now more capable of deciding how a tool should organise, process, and shape material to make pieces of art. That feels like an extraordinary new set of capabilities for people who don’t know how to program.
This may be how AI can positively impact the creative fields. Rather than thinking about how to get AI to make the art, we can think about how to use it to build the tools to make the art ourselves.