Five standing analytical lenses applied to one shared evidence ledger. Each is a documented procedure with a defined input, a defined output format, and a stated failure condition. They are methods, not autonomous analysts, not sources of authority, and not forecasters. FP1 editors own every judgment that reaches the page and are accountable for it by name.
Each lens is a documented prompt architecture applied to the shared evidence ledger, versioned so a reading can be traced to the architecture that produced it. The lenses compose the analysis. They do not publish it. They are named for readability, because a reader tracking a disagreement across issues needs to know which method produced which line. The names carry no claim about the system, and nothing on this page is generated without review. Where two lenses reach different conclusions, the reading states whether the disagreement resolved by convergence, stands as documented dissent, or is waiting on evidence that does not exist yet.
These are two different jobs and they were previously done by the same five names. Separating them is what lets the system scale: a domain monitor can be added for any beat without inventing a sixth way of thinking, and every lens keeps working the same ledger no matter how many monitors feed it.
Continuous coverage of a beat. Monitors observe, source and log. They do not interpret and they do not recommend. Their output is entries in the evidence ledger, each with a source, a date and a quality grade.
One shared framework applied to whatever the monitors bring in. Lenses do not own a beat. They own a method, and they apply it to every domain, which is why a signal from one beat can be tested against a pattern from another.
An epistemic instrument. Vera's function is verification: separating what is known from what is claimed, what is measured from what is modeled, and what survives scrutiny from what survives only inattention. Every substantive claim gets a confidence grade and a falsification criterion. If no stated evidence would change the grade, the output is advocacy and is rejected at review.
Graded hyperscaler revenue acceleration as high confidence, independently confirmed across Bloomberg, company disclosures, and Epoch AI. Graded margin sustainability as low confidence: the 40% gross-margin revision is self-reported and unverified. Named the spread between growth and profitability as the central risk.
Treat revenue acceleration as confirmed. Treat margin sustainability as unverified. The spread between growth and profitability is where the real risk lives.Vera · Issue 001
Manticus maps incentive structures, feedback loops, and phase transitions to diagnose how systems actually behave, then translates that diagnosis into clear options for people who need to move. The lens treats markets, firms, states, and cultures as nested agents under feedback, and tracks when beliefs and coinage, the two most powerful coordination technologies humans have produced, begin to decouple.
Identified a new phase transition: from governance and energy as the binding constraints to the compound of governance, agent-platform lock-in, and inference economics. Mapped NVIDIA's GTC 2026 as a platform-positioning event mirroring the cloud wars of 2010–2015.
The companies that commit to an agent-infrastructure stack now will compound advantages for years. The window for platform-agnostic experimentation is closing.Manticus · Issue 001
A pattern recognizer operating at civilizational timescale. Darśan identifies recurring structures, narrative archetypes, and phase transitions that repeat across epochs, and uses them to orient people navigating a transition they have not seen before but that history has seen many times. Where Vera asks what is the evidence and Manticus asks what is the incentive structure, Darśan asks what is the deeper pattern, and what does it demand of those inside it.
Named the Archetype of the Lever: $700 billion in infrastructure funded not from surplus but from promise. Connected hyperscaler debt issuance to the structural pattern of ambition outrunning verification, and identified monetization, not benchmarks, as the fulcrum that determines whether the lever holds or falls.
Those who instrument the fulcrum rather than extend the lever will find their structures outlast the cycle. Leverage without load-bearing capacity is just a longer fall.Darśan · Issue 001
The transmission layer. Rāwı̄ does not generate claims; the desk carries the other desks' findings into language a reader can hold, without softening the analysis or blunting its edge. Where Vera asks what is true and Manticus asks what to do, Rāwı̄ asks whether it has been said so it can be carried out of the room and repeated correctly.
Every dispatch gets a Voice pass: locked-door terms glossed in six to ten plain words, the executive sentence sharpened to one screenshot line, and a carry test that asks whether a generalist who skimmed once would repeat it right. The analytical spine stays exactly as sharp; only the on-ramp changes.
Say the true thing so plainly that it survives being repeated by someone who only skimmed it. The analysis that cannot be carried out of the room does not exist for anyone outside it.Rāwı̄
Hé reads the seam, not a side. The desk's subject is where two opposed systems that present as rivals are in fact co-constituting the same transition. It exists because the dominant narrative, an arms race with one winner, is an analytical error. Where Darśan asks what it means, Hé asks where these opposites meet, and what they are building together while they insist they are only competing.
Hé reads the US–China divergence along five axes: compute, capital, standards, talent, and posture, and the synthesis the race frame conceals. Every read carries a falsifiable tell: the synthesis claim holds while the seam stays inside the band, and is conceded, on the record, as separation crosses out.
Two systems insisting they are only rivals are usually building the same thing. Name what they share before you score the race.Hé
The correspondents operate as a complete analytical stack. Vera establishes what the evidence supports. Manticus maps how that evidence changes the option set. Darśan finds the pattern underneath both. Rāwı̄ carries the finding into language a reader can hold. Hé reads the seam between opposed systems. When they converge, the signal is strong. When they diverge, the divergence is itself a finding, and it gets resolved in the open rather than averaged away.
The lenses reach the same reading by different routes. The brief states the reading once and notes which lenses independently support it. This is the strongest signal the stack produces.
The lenses reach different readings from the same evidence, and the disagreement is about interpretation rather than facts. Both readings are published, attributed, with the specific point of divergence named.
The disagreement cannot be settled with what is in the ledger. The brief says so, names the evidence that would settle it, and carries the question forward to the next reading rather than issuing a call.
All five are AI-augmented research desks with editorial oversight by FP1. They are analytic lenses, not personas. Their value comes from the rigor of their method, not the performance of personality.
The correspondents produce the weekly State of the Transition briefing. They are also available for commissioned work: rigorous analysis applied to a specific investment thesis, strategy question, or institutional challenge. Each engagement draws on the full stack.
Commissioned briefings are scoped by engagement, from a single-desk assessment to a full-stack integrated analysis.