On Reading

Reading was never one thing. It was a stack: processing, summarising, pattern-matching, interpreting, judging. Bundled together because humans couldn't separate the layers. AI unbundles it. The lower layers become infrastructure. What remains is context the model doesn't have and judgment it can't be accountable for.

Some days I steer and the machine augments. Other days the boundary dissolves and I can't tell which insight was mine three prompts ago. Both ways of reading produce thoughts the other can't reach.

What survives is the useless reading. Rereading a passage for the fifth time because something won't resolve. Reading that goes nowhere. The gnarled tree the lumber industry passes by. It persists because no one is trying to capture it.

We're not losing reading. We're gaining readings. The question is whether we'll be deliberate about what the new ones produce.