I Wrote This Post With My Voice Only
It's 11 p.m. I'm walking alone in the forest next to my house, phone in hand, talking to my agent. While I speak, this post is being written, structured, and pushed to my website.
You're reading the proof.
TL;DR
A post about how I write this blog with my voice: thoughts recorded on walks, an agent that remembers them, an interview, and a pipeline that ships. Three lessons from building it and a way to copy the whole thing. All below.
The graveyard of voice notes. I think best when I walk. Alone, no screen, no one around, mind free to wander wherever it wants. For years I did what most people do: I had the thoughts, and then I lost them.
At some point I told myself: it's a shame to lose all of this. So I started recording. Short notes, long rambles, up to three hours a day of raw thinking: concepts, personal stuff, work, plans. Thousands of hours of it, sitting in my phone.
Which created a new problem: a voice-note graveyard. Recorded, never opened again. Thousands of letters to my future self, none of them delivered.
The shift. The shift was one question: what if my thoughts didn't have to die in a recordings folder?
Now every voice note goes to my agent (on idapt, the platform I'm building). It gets transcribed, and the agent distills it. Decisions, ideas, open questions, context about my life and my work, all filed into a shared knowledge base that every future conversation can draw from.
The more I talk, the more it knows me. And the more it knows me, the better it helps me think, because the questions it answers are mine, with my history in the room. A stranger gives you guesses. Something that knows you gives you plans.
The pipeline. This is how everything on this blog gets made now:
- Capture. I talk. Walking, in transit, whenever something is worth keeping. Voice notes for storage, or a live voice conversation when I want to work on something right now.
- Distill. My agent turns the raw transcripts into structured notes in our shared knowledge base: decisions, lessons, ideas, facts about what I'm building.
- Create. When I want to write, the page is already full of me. I ask my agent to pitch me ideas mined from everything it knows about me. Then it interviews me with good questions, based on everything I've already told it. My spoken answers become the material. We shape the structure together.
- Ship. I write the final version in my own words, the agent edits and pushes it to the site's repository, and it's live minutes later.
That last step is why I can write from a forest at 11 p.m. No laptop. No editor. Just my voice.
Three things I learned.
1. My best thinking happens walking. I thought that was just me, but it's a known thing. Nietzsche wrote that all truly great thoughts are conceived by walking. And Stanford researchers found that walking boosts creative output by about 60% on average, compared to sitting. My experience matches: sitting, I think in straight lines. Walking, my mind makes connections I'd never make at a desk.
2. Speak, and it happens. There's a line in Genesis where creation happens by speech alone: a thing is said, and the thing exists. I don't mean that arrogantly, but that is what this feels like. When an idea crosses my mind now, I don't write it down for later. I speak it, and by the time I'm home it's either done, in progress, or planned. I stopped maintaining a mental stack of things to remember. My brain used to be a task manager. Now it's a mind again. It made me roughly five times more effective at turning thoughts into reality: nothing gets lost between thinking and doing.
3. Memory is what makes an agent a partner instead of a tool. A chatbot with no memory can help anyone, and that means it helps no one in particular. My agent has thousands of hours of my thinking as context. So when it proposes a structure, a priority, or a decision, the advice is shaped by everything it knows about me. The human stays in charge. It's a lever that works the way I work.
One thing worth knowing before you try this: the skeleton of this website existed before this pipeline , a base I'd built earlier. What's new is everything that lives on it. Every post, every update, created the way this one was: spoken, structured, written, shipped, all by voice. The site is the gallery; the pipeline is what fills it.
You can do this too. The barrier is lower than you think:
- Record your thinking. Whatever you'd say to a friend about what's on your mind, say it to your phone instead.
- Give it to an agent with memory, not just a chat window. The whole value is the accumulation.
- Ask it to interview you when you want to create something. Answer in voice. Let it structure.
- Say "ship it" and read the result.
Two account connections (GitHub, hosting) and you never open an editor again.
I'm still in the forest. This post exists because I spoke. If you've been keeping thoughts in a folder somewhere: your graveyard is an asset. Talk to it.
P.S. I am building all of this in public. If you want to watch it happen, I am @richardmorel0 on X. Or start tonight: go for a walk, say one true thing to your phone. That is step one.
No other posts published yet.