Published articles are a reasoning trail. Canonical pages are the current standard.
Articles show evolution
They preserve what the category looked like at a date: early organizational ideas, stricter definitions, disclosure thinking, provenance logic and later challenge criteria.
Later corrections matter
Older formulations such as whole-path solo, fixed disclosure percentages, legacy asset/patent counts or platform rankings are not silently carried forward when the canonical model has changed.
Publication is not proof
An article can document chronology and public thinking. It does not independently validate solo provenance, technical novelty, asset value or valuation.
From prediction to a reviewable category.
Organizational compression
The earliest Medium essays framed the one-person company as an AI-enabled change in organizational structure. They are historical precursors, not the current definition.
Open sourceAI-observer narrative
An AI-narrated essay captured the intuition that a solo AI-native build path could be unusually documented. It is retained as narrative history, not independent evidence.
Open sourceDefinition, ceiling, and old metrics
The thesis moved toward structural compression, attention constraints and direct review of output substance rather than relying only on headcount, funding or visible organizational mass.
Open sourceProfile criteria and disclosure architecture
The Substack series added cross-domain/depth criteria and the disclosure paradox: enough public structure to evaluate existence, while sensitive asymmetric value may remain controlled or reserved.
Open sourceA strict verification standard
A strict “whole-path solo + complete documentation + market value” formulation made provenance reconstructability central. It is preserved as an earlier standard, not the current MZN Phase 2 boundary.
Open sourceExternal signal, not proof
The Crunchbase essay recorded a dated visibility signal. The canonical site treats such platform activity only as a reason to investigate, never as OPU validation.
Open sourceThe provenance paradox
The thesis became explicitly phase-bounded, separated AI tools from second-human contribution, and introduced traceable maturation as a possible provenance signal under reconstruction.
Open sourceTwo independent tests
The Hard Questions essay is the closest public precursor to the current architecture: asset substance/value and solo provenance are separate, Phase 1/2/3 are separated, and negative outcomes are allowed.
Open sourceThe current OPU model keeps the strongest ideas and retires the unstable ones.
Retained / strengthened
- One-person as a structural, evidence-testable question
- Explicit Phase 1 / Phase 2 / Phase 3 separation
- Solo provenance and asset substance as independent tests
- AI assistance disclosed as part of the method
- Reconstructable documentation and chronology
- Structural compression as context, not valuation
- Layered disclosure with controlled review
- Defeat conditions and valid negative outcomes
Retired / narrowed
- Whole-career or universal whole-path solo as MZN’s current claim
- Agent-free operation as a universal definition of every OPU case
- Fixed eight-month / direct-cost values as category requirements
- Legacy 28+ patent-grade count and fixed disclosure percentages
- Platform ranking as validation
- AI/model opinion as independent review
- Automatic conversion of replacement cost into valuation
- Imperfection as proof without traceable records
High-signal publications used in the current OPU review architecture.
The New Generation of Unicorns: Startups Built by AI-Orchestrated Founders, Not Massive Teams
Early organizational-compression thesis: AI can reduce visible organizational mass and change the founder/team relationship.
Read publicationThe One-Person Unicorn: A Concept We Talk About, But Haven’t Defined
Moves the discussion from “solo founder” as a label toward structural compression, role concentration, attention limits and the need for a new evaluation standard.
Read publicationPost-AI Means Post-Old Metrics
Argues that team size, visible spend and organizational bulk are no longer sufficient first-order proxies; output substance and capability compression deserve direct review.
Read publicationWhat a One-Person Unicorn Should Actually Look Like — six profile criteria the prediction has skipped
Introduces six earlier profile criteria: cross-domain output, depth + breadth, reproducible methodology, constraint-tested output, architectural coherence and layered disclosure.
Read publicationWhy a real one-person unicorn cannot show everything — and how to evaluate what is deliberately hidden
Develops the disclosure paradox and a practical distinction between public evaluability and protected asymmetric value; also identifies failure modes such as vapor, performance-only and dossier-without-architecture cases.
Read publicationA Real One-Person Unicorn: From the First Idea to the Final Output. Logged. By One Person.
Proposes a strict structural standard centered on solo path, reconstructable documentation and market value; makes logging and provenance central to verification.
Read publicationCrunchbase #3 Globally, #1 Outside the United States
Records a dated external-platform visibility signal and places it in the evolving OPU narrative.
Read publicationThe Paradox of Proving You Were Alone
Separates AI assistance from second-human participation; argues that AI-native work leaves a reconstructable trace; introduces traceable maturation as potentially relevant evidence and keeps imperfections from becoming automatic proof.
Read publicationThe Hard Questions Behind a One-Person AI-Native Company
States that MZN is asking for evaluation rather than self-certification; separates Phase 1, Phase 2 and Phase 3; centers the case on two tests: asset significance and solo provenance; makes defeat conditions explicit.
Read publicationSeveral essays were cross-published or later reworked between Medium and Substack. They are treated as one conceptual source family rather than multiple independent confirmations. Draft-only material is kept in the internal source map and is not presented here as published evidence.