The last post I shipped here was the ConsciousGPT Manifesto in late January. Then nothing. For six months.
Some of you wrote and asked if I was okay. I want to tell you what happened, because the silence wasn’t writer’s block, and it wasn’t me losing the thread. It was the opposite. It was me realizing the thread had outgrown the format I’d been forcing it into, and stopping until I had a different way to hold it.
The publication has three sections you can opt in or out of, and a writing rhythm I can actually keep. The back catalog is still here, three years deep. I’ll point you at the posts that still hold up at the end.
But first I owe you the why.
What I was doing for three months
Some of this you already know if you follow my work elsewhere. A lot of it you don’t.
In late January I started building a thing called a digital brain. Not a metaphor — an actual filesystem of plain-text notes, embedded in a vault that I could query like a search engine, augmented with a pair of autonomous agents named Len and Nel. Len does the work; Nel critiques it. Together they form a feedback loop that lets the system generate, review, and improve its own output without me sitting at the keyboard for every step.
By April I had over 2,000 notes in the vault, a knowledge graph dense enough that picking any node and walking outward gives you a coherent argument three hops later, and an agent system (I call it TONI — Thought Orchestration Networked Intelligence), that runs on my laptop and quietly turns ideas into drafts, drafts into essays, essays into recordings.
The reason I went quiet is that I couldn’t keep writing the way I was writing. The old workflow — sit down, generate a thesis, force it into shape over three days, publish, recover — wasn’t sustainable across the breadth of things I actually wanted to say. Consciousness research alone could fill a publication. So could the book I’m slowly writing about meditation and embodied change. So could the AI engineering work. Trying to do all three on a deadline is what produces the kind of writing where you can hear me holding my breath.
So I stopped, and I built the thing that lets me breathe.
Why it took longer than I expected
I’ll tell you the part I underestimated: the agent system is harder than the writing.
Generating a draft is a ten-minute problem. Generating a draft that sounds like me — that the voice-similarity score against my actual writing centroid clears 0.6 — required me to first analyze my own voice statistically, then build a rewriting layer, then teach the agents to call it before flagging anything as ready. There were weeks where the agents produced volumes of confident, fluent text that wasn’t mine in any meaningful sense. Useful for research notes. Catastrophic for essays.
The breakthrough came when I started thinking about the agent system the way I think about meditation: not as something that automates attention, but as something that can be trained to attend more accurately. The Buddhist analytical tradition has a name for this — yoniso manasikāra, often translated as “appropriate attention” — the practice of looking at what is actually there rather than at what you’ve already concluded. Once I framed the agents that way, the engineering shifted. The voice gate I described above is one piece. The artifact-required gate (no progress claims without a concrete deliverable) is another.
The result, five months in: a system that can ship work without me, and that does it in a voice you’ll recognize.
What’s different about Ordinary, Illuminated
The publication has three sections now. You can subscribe to any combination — including just one — and not be emailed about the others.
Essays is the default — everyone gets these — and it’s the lineage of what you signed up for originally. Personal-philosophical writing about awareness, the examined life, the places where complexity science meets contemplative practice. Approximately one piece every other Tuesday.
Research Dispatches is for the AI work, opt-in — papers I’m reading, experiments I’m running on a small contemplative-trained transformer (yes, really; the corpus is locked and the training is starting this month), and the kind of half-finished thinking that benefits from being said out loud. Shorter than essays, more technical. If that sounds dry to you, leave the toggle off. It will not be everyone’s thing, and the section settings let you sign off without leaving the publication.
Book Workshop is opt-in, a public-draft section for Marshal Your Awareness, the book I’m writing on attention, embodied change, and the practice of staying with what’s hard. Chapters and chapter-drafts will land here, with commentary open to readers. This will eventually be a paid section. For now it’s free; I want feedback before I lock the gate.
The reason for sections is operational honesty. Some of you came for meditation writing and got pulled into reading 4,000 words on transformer architectures because I had nowhere else to put them. That was unfair. The new structure is the apology.
Cadence and tier
Free, biweekly Tuesdays. That’s the contract.
Some weeks one of the sections will produce two essays in close succession; some weeks none of them will produce anything. The Tuesday rhythm is the spine. If you don’t get an email on a given Tuesday, the system is probably working — I’d rather skip than ship something half-thought.
A paid tier will arrive sometime in the second half of the year, gated to Book Workshop and to a quarterly long-form piece. I’ll write a separate post about it before flipping any switch. The publication stays free for everything else, and the back catalog stays unlocked.
The back catalog
If you’re new here — or if you’ve forgotten what we’ve been talking about — three older posts that I think still hold up:
The Unknown Unknown— on the operational weight of ignorance. The argument that information you don’t yet have is the most accurate map of where the work actually is. This essay is the seed for almost everything I’ve written since.
A Case for Eternal Consciousness — a careful, non-mystical case for taking the persistence of awareness seriously as a metaphysical question, including the scientific reasons not to dismiss it as obviously absurd.
Dharma Hot Tub — the most personal piece in the archive. Written on a particular night in a particular tub. If you only read one, read this one.
If you read those three and the new post lands, you’ll have the spine of what Ordinary, Illuminated has always been about: the intersection of complexity, consciousness, and the way actual humans live with both.
What I’m asking
If you’re still here, you’ve read about a thousand words of someone explaining their absence, which is roughly the most self-indulgent thing a writer can do. I appreciate the patience.
Two small things:
One: if any of the three sections sound like the only thing you want, go to your subscription settings and opt in (or out) accordingly. Essays you’ll get by default. Research Dispatches and Book Workshop are toggles you flip — sign up for both, neither, or one. Don’t suffer through Research Dispatches if your interest is the meditation work. The toggles are there now.
Two: if a friend or two might want this — particularly the kind of person who has been quietly thinking about consciousness or examined-life questions and doesn’t have a place to put it down — forward this. The publication grew during the silence anyway, somehow, but the kind of growth that matters is the kind that comes from someone you trust telling you to read something.
I’ll be back with the first essay under the new format.
Thank you for waiting.
— Justin