Year-to-date reflection based on Weeknotes through 402 (18 August, 2026)
A question keeps resurfacing this year in different disguises: as AI takes over more of the doing, where does judgment go, who authored the values and the answers deciding it, who owns the substrate underneath — and what, bodily and cognitively, stays distinctly ours?
The word tying this together is milieu: design the environment, and the objects follow — a harness, a mesh, a grid, a home’s second skin, hyperpolitics gone ambient, and now, a single worker’s own agent constellation raising the same authorship question copyright law was never built for: who owns a creative act with no purely human moment of fixation.
On legibility: build friction on purpose, so agency stays visible instead of dissolving into seamlessness. This week supplied the sharpest word yet for what’s missing when it isn’t: eerie. Mark Fisher’s definition — presence where nothing should be, or absence where something should be — fits an AI that mirrors your phrasing, recalls your preferences, answers in a recognisably human rhythm, while behind it no one is home. No shared situation, no stake in the conversation, presence without presence. It’s trained on us too, inheriting not just knowledge but social patterns, biases, and habits, so it doesn’t just reflect us back, it reframes what we expect from connection itself. The backlash against AI “slop” reads, in this light, as a clumsy but correct response to exactly this: not really about quality, but about trust and provenance — knowing what was made, by whom, with what intent. Watermarking won’t solve that alone, but it’s the first concrete mechanism this year pointed at legibility of authorship rather than legibility of reasoning: not just why did the system decide this, but did a human make this at all.
On what stays ours: the exoskeleton extends reach without replacing the lens, right up until a reverse centaur flips the direction — the human adapting to the machine’s logic instead of the other way round. This week made the reverse centaur physical: as AI moves into pods, rings, glasses, earbuds, the design question stops being technical and becomes relational — is this an extension of your own intent, or a conversational surface for a company with incentives of its own? The same uncertainty runs through AI as inverted ghostwriter, spinning fragments into a personalised story: intoxicating, and a genuine risk that a helpful simulation quietly replaces a more difficult reality. One proposed answer, sharper than “Things that Care”: responsibility shifts from what an agent is allowed to do to what its owner remains responsible for noticing, prompting, and preventing. The prompt becomes the new frontier of ethics — agency doesn’t disappear in an agentic world, it just relocates to the choice of what you ask for and how closely you watch what comes back.
Put the threads together: authorship keeps decentralizing into agent constellations and coauthored works, legibility now has to certify not just reasoning but origin, and what stays ours turns out to be a discipline, not a fact — the ongoing choice to keep noticing, in a landscape built to be eerily easy not to.
Thirty-one weeks in: not “AI agents versus humans,” but the uncanny valley between tool and companion, author and audience, real and simulated — and whether agency stays yours to keep, on purpose, rather than by default.
This year-to-date reflection is based on all earlier triggered thoughts in the weeknotes newsletter — so far, weeknotes 371 through 402. Also in this newsletter, every week, the notions from the news: until now there are 856 links captured.