Paper Portfolio
Decision (2026-05-26)
This wiki now feeds two papers, not one. The descriptive paper (P1) and the prescriptive operator-paper (P2) are siblings — different audiences, different questions, same evidence base, same hypothesis surface. Spawning P2 is the consequence of H5_ai-as-operational-not-product, which is written from the operator’s seat and surfaces a question the original brief does not answer.
Background
The original brief (00_initial-brief, 02_purpose-and-justification) framed a single paper aimed at capital-rich, conceptually-light operators. Its question is descriptive: where does economic value land across the agentic AI stack? The wiki’s hypothesis surface (H1–H4) and its concept layer were built to answer that.
H5_ai-as-operational-not-product (added 2026-05-26) reframes part of the same evidence base from the operator’s seat: given where value lands, what should an operator-builder without a distribution prior actually do? The descriptive answer to H1–H4 is not a sufficient guide to that question, because the descriptive paper is a map for capital-rich readers who can play any position on the U-curve, and H5’s reader is the capital-light operator-builder who can play almost none of them.
Rather than try to braid both audiences into a single paper — which would dilute both — the project now hosts two papers and lets the shared wiki feed both.
P1 — Where Value Lands in the Age of Agentic AI
The existing paper. No change in scope.
- Question. Where does economic value accrue across the agentic AI stack over the next 2–5 years (and the 5–10 year horizon)?
- Audience. Capital-rich, conceptually-light operators: founders, VCs, CEOs of large enterprises, family-office and sovereign-fund decision-makers (see 01_audience).
- Stance. Descriptive inquiry. Opinionated by construction; not a defense of a pre-baked thesis (02_purpose-and-justification).
- Hypotheses load-bearing for P1. H1_L0-L7-ladder, H2_u-curve-of-value, H3_orthogonal-axes-under-priced, H4_rl-specialization-value-pocket. H5_ai-as-operational-not-product appears as a boundary condition — what the descriptive map looks like from outside the VC-megafund track — but is not the spine.
- Working title. Where Value Lands in the Age of Agentic AI.
- Length & shape. ~20 pages, structure per 03_structure.
- Distribution. Per 04_distribution.
P2 — What to Start Now (working title)
The new paper, founded by H5.
- Question. What should an operator-builder with no distribution prior start now, given where AI value lands?
- Audience. Capital-light operator-builders: solo founders, small partnerships, traditional-industry second-generation operators (the author’s home market in particular — Italian / Veneto SME-adjacent), aspiring founders who are reading P1 and concluding that almost none of its positions are reachable from their starting point.
- Stance. Prescriptive inquiry. Treats H5 as the founding hypothesis: AI captures durable value as an operational input, not as a delivered product, for builders outside the VC-megafund race-in track. P2’s job is to make that hypothesis actionable.
- Hypotheses load-bearing for P2. H5_ai-as-operational-not-product is the spine. H6_industrial-ai-rollup-captive-suppliers (added 2026-05-26 from the same daily-thought reflection that produced the YC five-layer-loop concept) is the rollup variant — H5 at the single-firm scale, H6 at the multi-firm holding-company scale, both Track B. P2 also leans on the buyer-side material the wiki already has — non-tech-digital-core-synthesis, wef-ai-in-action-2025, deloitte-ai-dossier-eri, oecd-sme-ai-adoption-2025, fusion-skills, digital-core, taker-shaper-maker — and treats them as operator-builder inputs rather than as map features. New operator-side concept pages anchoring P2: ai-first-company-loop (the five-layer operating model that gives “managed by AI” a concrete architectural definition) and software-as-temporary-artefacts (the cost-base advantage mechanism that compounds with firm count in H6).
- Working title. Provisional: What to Start Now: An Operator’s Map for the Agentic Era. Italian working subtitle for the Veneto-distributed version: Cosa avviare adesso. Revisit after P1 ships.
- Length & shape. TBD. Likely shorter than P1 (~15 pages) and more concrete — the deliverable is a decision framework, not a market map. Final structure waits for H5 to harden under at least 2–3 ingests.
- Distribution. TBD. Likely a follow-on to P1; same channels but different framing. Italian-language manifesto from 05_open-questions is more naturally a P2 derivative than a P1 derivative.
The two tracks (the structural reason P1 and P2 are different papers)
| P1 reader | P2 reader | |
|---|---|---|
| Capital | Has it | Does not have it |
| Distribution prior | Has it | Does not have it |
| Default posture | Choose a position on the U-curve | Build outside the U-curve entirely |
| Useful answer to “what should I do?” | The map | A different question |
| Track on H5 | Track A (VC-megafund race-in) | Track B (operator track) |
The clearest way to think about the split: P1 is the map; P2 is the playbook for readers who cannot use the map directly. Both papers depend on the same descriptive findings; they differ in what they ask the reader to do with those findings.
How the shared wiki feeds both
The wiki is the substrate. Both papers draw on the same hypothesis pages, concept pages, sources, and data. No content is forked.
- Hypothesis pages carry a
Used by:annotation indicating P1, P2, or both, added by/lintduring the next sweep. - Concept pages are not annotated — the conceptual surface is genuinely shared.
- Source pages are not annotated — the evidence is the evidence.
- Open questions (05_open-questions) stay shared, with new P2-specific questions tagged
[P2]in their bullets. - Paper structures live in two separate pages: 03_structure (P1) and a future
03b_structure-p2.md(P2), to be drafted after H5 has had at least two ingest-cycles’ worth of evidence around it.
Timing and sequence
- P1 ships first. Per the original 4–6 week synthesis horizon (02_purpose-and-justification). The map is the prerequisite for the playbook.
- P2 drafted after P1 ships. A 4–6 week interval is the working assumption — long enough to let H5 mature, short enough that the two papers feel like a coordinated release rather than two separate projects.
- Manifesto and Italian-language derivatives are P2-flavoured by default; they ride P2’s launch, not P1’s.
Open questions promoted to 05_open-questions
- What is the right operator-builder Weber-classifier home? H5 surfaces this gap; see weber-taxonomy-2026-gaps.
- Where does P2 live commercially? P1 is a free PDF + GitHub-Pages wiki (04_distribution, 06_collaboration); P2 might want a different surface — a Substack, a small cohort-based course, an Italian-language landing page. Decide before P1 ships so the P2 surface can be staged.
- Italian-language P2 vs English-only P2: the operator-builder reader the author has most direct empathy for is Italian. Is the natural primary language for P2 Italian (with English secondary), or the same English-primary stance as P1?
Related
- 00_initial-brief — the original single-paper framing this note succeeds (does not replace).
- 01_audience — P1’s audience; P2’s audience is distinct.
- 02_purpose-and-justification — P1’s rationale.
- 03_structure — P1’s structure.
- 04_distribution — distribution channels; both papers will share most of them.
- 05_open-questions — joint open-questions list; P2-specific bullets added 2026-05-26.
- 06_collaboration — contributor model; same model for both papers.
- 07_analytical-vocabulary — shared vocabulary across both papers.
- 08_contributor-charter — shared contributor standard.
- H5_ai-as-operational-not-product — founding hypothesis for P2.