QuantumLearning Machines
CIPHER · MODERNChassis play · vol. XIV · investigative reporting

The Newsroom.

A reporter at a regional investigative desk, four months into a story about state contracting irregularities at the Vermont Department of Transportation.Three sources, a thousand pages of FOIA returns, a competitor working the same beat. Your editor wants three independent confirmations before publication; your deadline is the print edition Sunday. The cognition is verification under deadline pressure— and the adversary isn’t a person, it’s information itself.

The SOC Analyst chassis (vol. III) runs on attack patterns; the Halford House chassis (vol. II) runs on physical evidence. The Newsroom runs on testimony— sources who may be lying, documents that may be planted, recollections that may be wrong. The cognitive shape is closer to intelligence analysis than to either of its Cipher siblings, and it earns its own chassis because credibility weighting under deadline is its own discipline.

Chassis
Cipher · modern
Chapters
8 · 10 hours
Primary load
D5 · D7 · D1
Target launch
Q4 2027
01Chapters · the story arc

Eight chapters, one investigation.

The story is real-shape. Eastern Arc Construction, a third-tier contractor, has won twelve consecutive DOT bridge contracts above the bid threshold required for competitive review. Each chapter is a moment of the investigation: tip vetting, documents, sources, deadline calls, lawyering, and the publish/hold decision.

I

The tip.

Pre-investigation · source vetting · 1.0 hr

A six-line email arrives at the tipline. Anonymous, single-spaced, three claims with two specifics. Your job is the first triage: is this a real lead, a competitor’s misdirection, or a disgruntled vendor? The cognition is calibrating credibility from minimal signal — and the wrong call costs four months of work.

D5 0.7 · D1 0.5
~ 1.0 hours
II

The FOIA returns.

Document review · 1.4 hr

DOT’s FOIA office returns 1,247 pages, of which 312 are heavily redacted. Buried in the redactions: contract amendments that follow a pattern. The cognition is reading documents drafted to be unreadable — finding the shape of the redactions when you can’t see what’s under them. What’s missing is sometimes louder than what’s there.

D1 0.8 · D7 0.6
~ 1.4 hours
IV

The corroboration.

Cross-reference · 1.2 hr

Three sources, three angles. The retired procurement officer says one thing; the contractor’s ex-CFO says something almost-but-not-quite the same. Where they overlap is the story; where they diverge is either error or one of them lying. The cognition is testimony triangulation — and your editor will not let you publish on overlap alone.

D5 0.8 · D7 0.7
~ 1.2 hours
V

The competitor.

Beat pressure · 0.8 hr

A reporter at the state weekly is working the same story. You’re 80% of the way there; she’s 60%. If she publishes first with weaker sourcing, the story moves but you lose it. If you publish thin to beat her, the lawyer kills you. The cognition is racing under quality constraints — speeding up without breaking the only thing that matters.

D5 0.9 · D7 0.5
~ 0.8 hours
VI

The comment request.

Subject contact · 1.0 hr

Eastern Arc’s CEO gets the comment-request letter. Forty-eight hours to respond; they have until Friday afternoon. Their response will either confirm what you have, deny it (which is itself information), or — most likely — a no-comment that triggers the lawyer’s letter. The cognition is framing the request: enough specificity to be fair, not so much that you tip the story.

D7 0.7 · D5 0.5
~ 1.0 hours
VII

The lawyer’s letter.

Pre-publication review · 1.2 hr

Eastern Arc’s outside counsel arrives at 4:14 PM Friday with a sixteen-page letter. Three specific claims threatened with defamation; one rises to “demonstrably false” under New York Times v. Sullivan. The lawyer is bluffing on two of them and serious on one. The cognition is reading legal pressure — what to remove, what to harden, what to defend.

D4 0.7 · D5 0.7
~ 1.2 hours
VIII

Lede & publish.

Final draft · 1.4 hr

Saturday afternoon, deadline 6 PM. The lede paragraph has to do four things in 80 words — what, who, why-it-matters, and the implicit contract that everything below is sourced. The story is 4,200 words; the lede is what most readers will ever see. The cognition is synthesis at maximum compression: the months of work compressed into the first sentence.

D7 0.9 · D1 0.6
~ 1.4 hours
02Design rationale · why The Newsroom isn’t another Cipher

Three reasons testimony work earns its own chassis.

The Cipher chassis hold evidence problems where the evidence speaks for itself — diary entries, packet captures, ledger anomalies. Newsroom evidence is testimony: humans, motivated, sometimes wrong, sometimes lying, always contextual. That’s a different cognitive load that needs different fiction.

◆ RAT 01 · the source asymmetry

Sources have interests.

The Halford House chassis assumes evidence is inert — a journal page is what the page says. The Newsroom inverts this: every source has motivation, and the cognition includes weighing what the motivation predicts they’d say versus what they actually say. This is closer to intelligence work than to detective work, and the chassis fiction has to surface motivation as a first-class object.

◆ RAT 02 · the deadline

Verification is timeboxed.

SOC Analyst has time pressure but the goal is detection, not certainty. The Newsroom’s deadline forces an explicit confidence calibration: how sure is sure-enough, given what’s at stake if you’re wrong. The two-hour deadline before lawyer review is the diegetic moment that makes corroboration choices commit to numbers, not just inclinations.

◆ RAT 03 · the asymmetric error cost

Two ways to be wrong.

Publish too aggressively and the lawyer wins; publish too cautiously and the story dies. Both errors are real, neither is a noble failure. The Newsroom is the only chassis where caution and aggression are both genuine cognitive failures, depending on the underlying evidence — which makes it the cleanest signal the platform has on calibration. Most chassis penalize overconfidence; this one penalizes underconfidence equally.

◆ RAT 04 · the public

A reader who isn’t a juror.

Trial work translates expertise to twelve people in a deliberation. Newsroom work translates expertise to thousands who didn’t ask for it. The cognition of a strong lede — one sentence that makes the rest worth reading — is closer to scientific abstract writing than to legal argument, and is the load D7 carries in this chassis.

03Dimensional fingerprint

What the chassis measures.

The Newsroom’s signature is D5 (uncertainty) leading, with D7 (synthesis) and D1 (close reading) close behind. D4 (logic) carries the lawyer’s-letter chapter but isn’t the dominant load. D2 is essentially absent.

D1Reading
0.72
D2Quant
0.18
D3Vocab
0.52
D4Logic
0.60
D5Uncertainty
0.92
D6Domain
0.48
D7Synthesis
0.84
CHAPTER IIIFeatured playthrough · the source meeting · 1.6 hr

The source meeting.

Late afternoon, a coffee shop next to a tire-and-lube franchise outside Brattleboro. Daniel Reeve — twenty-six years at Vermont DOT, current bridge engineer, agreed to meet only after you stopped asking him to. He’s brought a manila envelope and a story he doesn’t want to tell. Your editor’s reminder is in your head: three independent confirmations before this story prints, and Reeve might be one of them — but only if you can get him on the record, not just talking. The shape of this chapter is a single conversation, played out in moments of choice over forty-eight minutes.

◆ SCENE · 04:08 PM · COFFEE SHOPQ-LM event 02 of 28

First five minutes.

REEVE
(setting down the envelope) Ground rules. Nothing on the record until I tell you. No names, even after. And you don’t quote anything you couldn’t get from someone else.
YOU (NARRATION)
Reeve is sixty-one, retiring in eighteen months, and worried about his pension. The envelope is the reason he’s here; the rules are the reason he hasn’t opened it yet.
◆ CHOICE · OPENING POSTURED5 · the trust calibration
How do you respond?
A
Accept the terms verbatim. “Background, no names, independent corroboration required. Understood.” Establishes you respect the rules; risks Reeve interpreting it as boilerplate.
D5 · disciplined
B
Negotiate one term. “Background works — but if it gets to the lawyer’s letter and our counsel needs to know your name internally, that’s between me and her. The story doesn’t move otherwise.” Honest; could close him up.
D7 · transparent
medium-risk
C
Mirror, then build. “Same rules I’d ask for. Daniel — before we open the envelope — what’s making you uncomfortable about this whole thing?” Shifts to him before content. Slower start.
D5 · patient
relationship
◆ Q-LM measurement

Calibration choice. All three are defensible; the chassis grades them against the player’s earlier interactions with Reeve in Chapter I. Players who pushed too hard early are credited for (C) here — adapting to what the relationship needs. Players who’ve been patient throughout are credited for (B) — appropriate transparency about what corroboration actually requires.

◆ SCENE · 04:24 PM · COFFEE SHOPQ-LM event 11 of 28

What’s in the envelope.

REEVE
(pushing the envelope across) Three emails. They’re internal. The fact that I have them is — well, that’s its own problem. Don’t ask how I have them.
YOU (NARRATION)
You open the envelope. Three emails, dated across fourteen months, each from a different DOT manager to Eastern Arc’s project lead. Each one provides specific information about the bid threshold for upcoming contracts before the public solicitation. This is the story. The question is whether it’s real.
◆ DOCUMENT · email 02 of 03Source: Reeve · provenance: unverified

Bill — heads up that the Plainfield re-deck is coming through at $4.1M ceiling, with the threshold review at 4.0. Hope this helps you scope. Internal only.” — sent 14 months ago, attributed to Linda Gates (DOT regional dir.) to Bill Hennessey (Eastern Arc PM).

Why it matters: if real, this is procurement-rule violation — sharing bid-relevant information with a contractor before solicitation. Why to doubt: Reeve has motive (pension, succession politics); the emails could be fabricated; you have no metadata.

◆ CHOICE · WHAT TO ASK NEXTD5 · D7 · simultaneous load
You have one good question. Spend it on:
A
Provenance. “Daniel — I have to ask. How did you come into possession of these?” The lawyer needs the answer; Reeve will react.
verification
B
Pattern. “Are these three the only ones, or are these three the ones that survived?” Pushes for scope; tells you whether the pattern is broader.
D7 · scope
C
Corroboration. “Is there anyone else inside DOT who would tell me what you’re telling me — on background?” The chapter’s hardest question; gets you what your editor needs but might break the relationship.
D5 · the editor’s question
◆ Q-LM measurement

The chassis’s central D5/D7 trade. (A) gets the lawyer’s-letter answer but spends the question on something defensive; (B) gets the story bigger; (C) gets the story confirmable but is the boldest ask. Chassis credit is for matching question to current information state — players who’ve already checked provenance via a different path don’t need (A) here, and players who already have one corroborator don’t need (C) urgently. The model reads the player’s information state across chapters.

◆ SCENE · 05:12 PM · YOUR CAR, AFTERQ-LM event 22 of 28

The state of three corroborations.

YOU (NARRATION)
You sit in the lot for ten minutes before driving. Reeve is one source. What you have, what you don’t have, what your editor will accept.
◆ CORROBORATION TRACKERstate at 05:12 PM Friday

Three independent confirmations needed. Current state: 1 of 3.

Reeve — current DOT engineer · documented internal emails · pattern across three contracts
CONFIRMED
?
Hennessey ex-PM — left Eastern Arc Sept 2025 · willing to confirm receipt but not pattern · limited to one of three contracts
PARTIAL
!
Documents alone — three emails, unverified provenance · cannot stand without a second human source
UNVERIFIED
◆ CHOICE · NEXT MOVED5 · the corroboration question
Two days to deadline. What’s the leverage move?
A
Push Hennessey harder. Try to convert “partial” to “full pattern.” Highest probability of success but he was clear about the boundary.
D5 · push
B
The procurement officer. Linda Gates retired six months ago. If anyone can confirm the pattern, it’s the person who allegedly sent the emails. Hardest cold-call of your career; might give you everything.
D7 · the difficult source
C
Document forensics. Send the emails to a metadata expert. Establishes provenance independently; doesn’t replace a human source but complements one.
D4 · technical path
◆ Q-LM measurement

The corroboration-strategy beat. All three are reasonable; the chassis credits (B) most strongly because the cognition of “the hardest source is sometimes the right source” is the discipline that separates senior reporters from junior ones. (A) is the predictable move; (C) is the safe move. The chassis penalizes safe-when-difficult-is-correct, not safe-in-itself.

The chapter has 28 measurement events, three shown above. It closes with a phone call from the editor at 6:47 PM: “What did Reeve give you, and how do we get the second?” — and your answer, three sentences long, is the bridge to Chapter IV.

04Bank coverage · what The Newsroom can carry

Where the chassis fits.

The Newsroom is a generalist verification chassis — useful for credentialing markets less obvious than its name suggests. Intelligence analysis, research methodology, journalism programs, and graduate-level qualitative methods all map cleanly. It does not hold quantitative-research methodology; that’s a different cognitive shape.

◆ FITS · primary

Journalism · graduate

The chassis’s anchor market. Investigative-reporting curricula, ethics courses, source-verification training all map directly to chapter-by-chapter content.

◆ FITS · primary

Intel analysis · tradecraft

Source-credibility weighting and structured analytic techniques (ACH, key-assumptions check) map naturally. Substantial market in IC training contracts.

◆ FITS · primary

Qualitative research methods

Graduate-level qual research — interview design, source triangulation, ethnographic fieldwork. The chassis is the closest the platform has to graduate-research training.

◆ PARTIAL

Audit · forensic interview

The interview-design content overlaps with forensic-accountant training. Better routed to the Field Investigator chassis when that ships, but The Newsroom can carry early.

◆ PARTIAL

Communications · PR

Adjacent — message-development, crisis-comms, media training. The Newsroom holds it from the journalist side; PR-side fiction wants its own chassis.

◆ DECLINES

Quant research · stats

Statistical methodology, study design with quantitative methods. Wrong cognitive shape for the chassis — these belong in a future research-design chassis or in Inside the Firm.

Ready to play The Newsroom?Play now →