FLDI Research — The cost of the endless feed Checked through 9 November 2026 { "units": { "daysPerYear": 365.25, "minutesPerDay": 1440, "fixedWakingHours": 16, "lifetimeAgeEndExclusive": true, "currency": "illustrative constant USD, no PPP adjustment" }, "formulas": { "lifetimeCalendarYears": "minutesPerDay * (endingAge - startingAge) / 1440", "wakingShare": "minutesPerDay / (16 * 60)", "annualGlobalPersonYears": "uniqueDailyAudience * combinedDailyMinutes / 1440", "platformMinutes": "combinedDailyMinutes * weight / sumOfWeights", "equivalentOpportunityYears": "annualGlobalPersonYears * assumedNetValueGap", "annualOpportunityValue": "uniqueDailyAudience * dailyMinutes / 60 * 365.25 * displacedWorkShare * realizationShare * hourlyValue", "individualPresentValue": "sum(annualOpportunityValuePerPerson / (1 + realDiscountRate) ** (yearIndex + 0.5)), yearIndex=0..66", "hypotheticalQalys": "uniqueAudience * assumedCausalExcessAffectedShare * assumedUtilityDecrement * durationYears" }, "scenarioPresets": { "Low": { "audienceBillions": 1, "minutes": 30, "gap": 0.05 }, "Central": { "audienceBillions": 1.5, "minutes": 60, "gap": 0.2 }, "High": { "audienceBillions": 2, "minutes": 90, "gap": 0.5 }, "Net benefit": { "audienceBillions": 1.5, "minutes": 60, "gap": -0.1 } } } All age windows exclude the ending birthday. Global population is a single assumed unique DAILY audience, not summed platform reach. Platform weights allocate the same combined foreground time. Reels and Shorts minutes are not observed separately in the cited usage data. Low/central/high are illustrative scenarios, not confidence intervals. Negative net value gaps allow net benefit. No scrolling time is converted to deaths or health-adjusted life-years. Supplemental QALY arithmetic requires separate hypothetical causal inputs; it is not an interactive article feature. The article has one lifetime calculator; global attention, remaining exposure and welfare examples use fixed assumptions. The opening historical superlative is an editorial argument, not an established ranking or a quantified atomic-bomb comparison. Opportunity years, dollars and QALYs may overlap and must not be added. Raw time is undiscounted; only monetary lifetime valuation has a discount rate. Source code is provided in /research/fldi-research-source.tar.gz. Model implementation: src/model.ts. Regression tests: tests/model.test.ts. Full register: /research/data/sources.json.