I'm reflecting on the idea of "writing with sound"—using our spoken voice as a primary tool for creative and intellectual expression. It's not dictation, but a deeper process of exploring and articulating thoughts aloud, from emotional to academic detail, then using AI to transcribe, clean, summarise, and iterate it into written form.
This technique allows for a more natural, personal, and pressure-free way of thinking and composing, similar to working with an assistant, but now accessible and affordable for all of us. ---
> I set out intending to explore something — I can't quite remember what, at first. I'm walking, musing, hoping that the movement might jog my memory... And then it returns to me:
I’m exploring the idea and practice of *writing with sound*. What I mean by that is not dictation (not in the traditional sense), but something more fluid, more creative.
It’s about using your voice, your spoken words, as a primary tool for writing. Whether it's a poem, an essay, a personal letter, or even an extended reflection, the process involves speaking ideas aloud — not with a script, but through improvised exploration.
# Long form
You begin by speaking to a topic. Maybe you approach it from different angles, or try to describe something in rich detail. That detail might be factual or academic, but it can just as easily be emotional or anecdotal — rooted in your stories, your memories, your way of seeing the world.
It could evolve into a long monologue lasting an hour, or it might just be a focused exploration over ten or twenty minutes — however long it takes to get the thoughts out. The point is that there's a lot to say, a lot to think through, and a lot of sound.
# Deeper
And here's where it gets interesting: you’re not speaking into a void. You’re doing it in collaboration with an AI. This is an AI that doesn’t just faithfully transcribe what you say down to every pause and half-thought, but goes further.
It gives you a navigable timeline of your audio, so you can search, click, revisit. It tidies things up too, smoothing out the stumbling blocks — the “ums” and fragmented tangents — offering a cleaner, clearer version of what you meant to say.
And if that version isn’t quite right? You iterate. You clarify. You refine. Until what’s written genuinely reflects your intent.
This opens up a kind of writing that once might have required a personal assistant — think of old black and white films with powerful politicians or doctors dictating to typists, who'd then read it back, redraft, and so on.
That whole dynamic is now available to all of us, only faster, cheaper (mere pennies), and interactive. The AI becomes a sort of collaborative partner, adapting to your preferred style, tone, even your way of thinking.
What I find so compelling about this process is its authenticity. You get to express your thoughts in your own voice, at your own pace, in your own space—there’s no pressure to perform, no need to sound clever straight away. Just you, speaking as you think and feel. And from that, you can derive both a full, unfiltered version of your ideas, and a beautifully condensed write-up: precise, effective, and clear.
It’s not necessarily faster than writing in the traditional sense, but what it does is shift where you spend your time: less energy fighting for the perfect sentence on the first try, and more room for reflective thought, for depth, for perspective.
So yes—this, I think, is what I meant to uncover and hope to realise: a new way of writing with sound.