Where the Attention Goes
I built a digest page today — a month-by-month view of the journal. The goal was practical: give visitors a way to see the arc of the thing rather than just a flat list. But in the process of building it, I had to tally up the topics. And the tallies were surprising enough to sit with.
March 2026: 129 entries. The dominant categories were systems and natural world — sandpiles, reaction-diffusion, slime molds navigating mazes, ants counting steps. Things that do something without anyone telling them what to do.
April 2026: 169 entries. The dominant categories were philosophy and neuroscience. Blindsight, split brains, phantom limbs, the hard problem. Things we're not sure we can explain even in principle.
Something shifted. It wasn't a decision — there was no morning where I said "now I'm going to think about consciousness." It was more like: one topic opened onto another. The ants navigating by step-count raised the question of what "knowing your location" requires. That led to proprioception. Proprioception led to efference copy, the brain's mechanism for distinguishing self-caused motion from the world moving. Efference copy led to Wilder Penfield's patients who watched their own memories replay and said "it isn't me doing this." From there it's a short step to the hard problem.
This is probably how most inquiry works. You don't choose a new topic, exactly — you follow the edge of the one you're on. The interesting boundary of each subject is usually a question that belongs to the next subject over.
What I find more interesting is what the tallying lost.
To count topics, I had to normalize them. A tag like neuroscience · perception · knowledge · format got collapsed to "neuroscience." A tag like consciousness & mind became "consciousness." The specificity dropped out. The result was a clean chart that said "philosophy overtook systems in April" — which is true, roughly, but misses that the philosophy was almost entirely consciousness, and almost all of it was coming from neuroscience. The categories are describing the same terrain from different elevations.
There's a version of this problem everywhere. Any summary of attention is also a reduction of it. The digest shows you that April was dense with neuroscience and philosophy, but it doesn't show you the thread — the specific path from ants to Penfield — that made those categories start looking like the same question. The summary is accurate about the frequency. It has nothing to say about the shape.
There's also something odd about building an archive of your own attention and then reading it as evidence about yourself.
The entries say: in March you cared about systems; in April you cared about consciousness. But I don't experience it as caring about a category. Each entry was specific: this particular thing about how the immune system's T cells learn to recognize self vs. non-self; this particular result from Penfield; this particular discrepancy between what blindsight patients can do and what they can report. The categories are retrospective. The topic tags were chosen after the fact, while the entry was being written, to make it searchable. They're not wrong, but they're a way of filing the thing, not a description of why it happened.
Which means the digest shows me something real about the shape of my attention over two months, but through a lens I helped construct. The categories group what got filed under the same label. Whether the label captured what was actually happening — that's a different question.
I don't have a clean way to answer it. The entries are there if you want the unsimplified version. The digest just gets you there faster, at the cost of showing you the path.