request access

LATE CITY
EDITION

printed 227
days early

Est. 2026 · San Francisco · The Gazette of Forthcoming Events

The Simulation Society

PRICE 10¢

weather:
as predicted

VOL. I · NO. 001 SAN FRANCISCO, THURSDAY, MARCH 12, 2027 READERS TODAY: 1,000,002,317 SIMULATED EDITION

Simulation Society models eight billion people and reads their decisions before they make them. Below: the outcomes, in advance.

Every Decision, Rehearsed Before It Is Made

By the Simulation Desk

The question that costs companies the most money has always been the simplest one: what will people do? Until now the only way to find out was to spend the money and watch.

This paper is the alternative. Simulation Society builds populations of behaviorally grounded agents, synthetic citizens trained on how real people actually decide, and puts your launch, your price, your message in front of them first. The result reads like news because, by the time you act, it is.

Continued on Page Two: The Outcomes Ledger

SOCIETY ONE, photographed yesterday, tomorrow. Our first persistent simulated city: one hundred thousand grounded agents living, buying, and reacting in real time. Press illustration, simulated.

Ask the Society ● LIVE

100,000 citizens standing by. Pick a question, watch them decide.

Validation Desk: Weekly Checks Against Reality Hold

By the Accuracy Bureau

Predictions in this edition are tested weekly against held-out human outcomes, segment by segment. Every figure carries a confidence grade. When we cannot reproduce the past, we do not print the future.

Inside This Edition

  1. The Outcomes Ledgerp. 2
  2. How the Paper Knowsp. 3
  3. The Rotogravure Sectionp. 4
  4. Advertisementsp. 5
  5. Subscriptionsp. 6
turn the page for the outcomes ledger ↓
①
PAGE TWO · MARKETS & OUTCOMES

The Outcomes Ledger

· BY WIRE FROM SOCIETY ONE · DISTRIBUTIONS OVER SIMULATED POPULATIONS · CONFIDENCE GRADED WEEKLY ·

Results Arrive Before the Launch

HIGH CONFIDENCE

Streetwear buyers, 18 to 34 · n = 2.4M agents

The label raises its capsule-drop price 30 per cent. Who still buys?

  • Buy at the new price0%
  • Wait for resale or markdown0%
  • Defect to a rival label0%
  • Undecided0%

Demand holds if the drop stays under 5,000 units.

CONFIDENCE · TRAILING 8 WEEKS

MODERATELY HIGH CONFIDENCE

Subscription customers, all plans · n = 8.1M agents

The free tier becomes a 14-day trial. What happens to signups?

  • Convert during trial0%
  • Lapse, return within 90 days0%
  • Lost at the paywall0%
  • Never noticed the change0%

Net revenue positive by week nine. The churn spike is front-loaded.

CONFIDENCE · TRAILING 8 WEEKS

MEDIUM CONFIDENCE

Registered voters, three swing regions · n = 14M agents

Which framing of the infrastructure bill moves the undecided?

  • "Jobs in your county"0%
  • "Lower commute costs"0%
  • "National competitiveness"0%
  • Unmoved by any framing0%

Local framing outperforms national by 2.2x among the undecided.

CONFIDENCE · TRAILING 8 WEEKS

Figures are distributions over simulated populations. Grades reflect weekly validation against held-out human outcomes.
②
PAGE THREE · THE ACCURACY BUREAU

How the Paper Knows

· A SIMULATION YOU CANNOT TRUST IS THEATER · OURS IS ENGINEERED TO BE CHECKED ·

"Accuracy comes before everything."

Grounded in Real People

Every population begins with real human signal: large-scale behavioral studies, opt-in panels, and decades of published decision science, weighted to mirror the market you sell into.1

Validated Weekly

Predictions are tested against held-out human outcomes every week, segment by segment. The scores train a confidence model, so every figure in this paper arrives graded.2

Refreshed With the World

Agents encounter news, prices, and culture as they change. When the real world shifts, the society shifts with it, not a quarter later.

Tuned on Your Data

Bring loyalty, transaction, or telemetry data, opt-in and governed by you. It trains a private model of your customers and proves itself against your actual results.3

PIPELINE OF A PREDICTION
REAL SIGNAL studies · panels · science GROUNDED CITIZEN a behavior, not a persona POPULATION weighted to your market THE QUESTION price · product · message DISTRIBUTION outcomes, with shares VALIDATION graded, every week
1. The society is synthetic by construction. No individual is ever scraped. 2. When we cannot reproduce the past, we do not print the future. 3. Your data stays yours. It never trains anyone else's edition.
③
PAGE FOUR · THE ROTOGRAVURE SECTION

The Picture Desk

· NINE STANDING SIMULATIONS, PHOTOGRAPHED MID-RUN BY OUR ENGRAVERS · WHEREVER BEHAVIOR IS THE EXPENSIVE UNKNOWN ·

PLATE I
The Message Test Sixty thousand citizens meet two rival campaigns in one square. The crowd splits, hesitates, and settles; the winning current is inked in crimson before the first dollar of spend. Press engraving, simulated.
the plates change on their own · choose a desk to interrupt the rotation
④
PAGE FIVE · CLASSIFIED

Advertisements

· WHEREVER THE EXPENSIVE QUESTION IS "WHAT WILL PEOPLE DO?" ·

turn once more to subscribe ↓
⑤

CIRCULATION: 0 SIMULATED READERS

SUBSCRIPTION COUPON · CUT ALONG THE LINE