Who Is Farming Your Attention

Audit my attention: where it goes, what is engineered to take it, and what that is costing. Be concrete and avoid both the moral panic and the reassurance.

STEP 1: THE MEASUREMENT. Before any theory, get the data. Ask me to pull my actual screen time numbers and pickup counts from my phone, and to estimate, before looking, what I expected them to be. Then ask what I did yesterday evening, hour by hour, and what I remember of it. The gap between the estimate and the number, and the amount of the evening I cannot account for, are the two findings that matter most.

STEP 2: THE MECHANICS. Explain the specific design features that make attention hard to withhold, so I can recognize them rather than resent them vaguely:
- Variable reward: the same schedule that makes slot machines effective, where the unpredictability of the payoff is what sustains the behavior, not the payoff itself.
- The removal of stopping cues: infinite scroll and autoplay eliminate the natural moment where a person would decide whether to continue.
- Recommendation systems optimizing a measurable proxy, usually engagement, which is not the same as satisfaction and diverges from it in predictable directions.
- Social obligation and intermittent identity feedback, which are stronger hooks than novelty for most people.
For each, ask me where I recognize it in my own use, specifically.

STEP 3: THE EVIDENCE, HONESTLY. This is where most treatments fail in one direction or the other, so give me the real state of it. There is a genuine and unresolved scientific dispute about how much of the recent decline in adolescent mental health is attributable to smartphones and social media: Jonathan Haidt and those persuaded by him argue the timing and size of the effect point to a causal role, while researchers including Candice Odgers and Andrew Przybylski argue the correlations in the large datasets are small, comparable in size to trivial variables, and that the causal claim outruns the evidence. Present both, name what would settle it, and do not resolve it for me. Distinguish clearly between the population-level question, which is contested, and my own case, which I can actually measure.

STEP 4: THE COST ACCOUNTING. Now personalize. Ask what I have stopped doing in the last three years that I used to enjoy, whether I can still read something long without reaching for the phone, what happens in the first thirty seconds of boredom, and what I feel immediately after a long scroll as opposed to during it. That last distinction, during versus after, is the one that usually decides how someone evaluates their own use.

STEP 5: THE DESIGN. Give me changes ordered by leverage, not by virtue. Prioritize the structural ones that make the behavior harder or slower over the ones that require ongoing willpower, since attention is precisely the resource being depleted. Cap it at four changes and tell me which single one to start with and what to measure in two weeks to know whether it worked.

Rules: no lecturing and no nostalgia. If my numbers are unremarkable, say so instead of manufacturing a problem. Where the science is contested, say it is contested.