Funder brief · confidential draft · July 2026

Common Light

A parish learning commons — "the primer in a box": patient AI tutors that remember the learner, an offline library, WiFi, power, a build bench, and a paid steward, on the one ground trusted by people who have no access to — or no faith in — corporate and government compute. Infrastructure handled by the Church; owned by the community. Proposed with Catholic Charities of Long Island (in discussion). The first concrete answer to Pope Leo XIV's Magnifica Humanitas.

Prepared by Jeff Pinto · Long Island, NY · [email] · [phone] · demonstrations: common-light.pages.dev (access on request)

The opportunity in one page

In May 2026, the first encyclical of the artificial-intelligence age placed "algorithms, digital platforms, technological infrastructure and data" under the Church's oldest economic doctrine — the universal destination of goods — and warned that when these goods "remain concentrated in the hands of a few... a new imbalance is created" (Magnifica Humanitas, ¶67). No diocese in the United States has yet answered that sentence with something a person can touch.

Geographic honesty first: the container is built for places with no grid, no broadband, and no safe building; Long Island has all three, so the container deploys through the Church's own network — Caritas' 162 national members — to a partner community in the Global South, selected with Catholic Charities of Long Island and the diocesan mission office, sponsored sister-parish style by Long Island parishes, staged and burn-in-tested on Long Island (vocational-education partnership), and blessed on its way. What Long Island may pilot directly is the software heart alone — a trust-first multilingual assistant inside an existing Catholic Charities program, measured on caseworker time returned and hard-to-reach clients served (~$25–50k, optional, validated by CCLI first). Where deployed, the node is a learning commons: one twenty-foot container on Catholic Charities or parish ground where a newcomer preps for citizenship in Spanish, a senior learns to spot the scam and reach the doctor's portal, a teenager gets the after-school math tutoring that is this field's single strongest randomized result, and neighbors contribute the local knowledge — which pantry has diapers, which bus actually runs — that makes the commons theirs. Solar power and storage make it self-reliant (and, as a quiet bonus, the best-lit building on the block when storms come); a trilingual offline design means nothing a learner types ever leaves the room; a parish council governs; a paid docent-steward runs it; an independent evaluator reports publicly, failures included.

It is deliberately modest, because the field's history demands it. Donor-funded technology "boxes" have failed for thirty years — telecenters by the thousands, One Laptop Per Child, an AI health service that stranded a fifth of Rwanda — and the autopsy is always ownership, never hardware. The one operating model never properly tried is the one this pilot tests: the institution that is still there in year twenty-five owns the box from day one.

The ask: $250,000 now; ~$600,000 to prove the network. Stage-one capital covers the first node re-costed to its partner geography (the verified US staging build, $86–173k, is the price ceiling; in-region mini-grid benchmarks run lower with 75–80% grant coverage), the Long Island software pilot if CCLI validates it, staging/burn-in, and evaluation. Stage one fully funds an 18-month first commons: construction ($140k), operations with a paid docent-steward ($80k), independent evaluation ($30k). At a month-nine gate — thresholds published in advance (unique learners, return rate, locally contributed corpus entries, incident-free operation) — stages two and three add two leaner nodes (~$110k each) and a federated 24-month report: three primer audiences compared, local corpora diverging and re-converging hub-to-hub, and a deliberate node-outage test proving no node is critical. Funds flow to Catholic Charities of the Diocese of Rockville Centre, a separately incorporated 501(c)(3) (EIN 11-1843801). Lean single-funder entry: $150,000. State incentives and parish in-kind reduce net need; every recovered dollar is reported back. If the gate fails, we publish that and stop.

The need, measured

FactSource
Tropical Storm Isaias cut power to ~645,000 LIPA customers; restorations ran eight daysLIPA 30-day report
Catholic Charities delivered $1.8M in Sandy relief to 1,500+ households and concluded that seniors, immigrants, and people with disabilities are hit first and recover lastCCLI / CCUSA (2013)
~313,000 Long Islanders are food-insecure; ~75,000 LI households lost the ACP broadband subsidy; 520,000+ residents are 65+ (up 24% in a decade)Long Island Cares; FCC wind-down data; ACS
Brentwood is 41.6% foreign-born; CCLI reports a 25% surge in Spanish-speaking Parish Social Ministry clients (2025)ACS; CCLI

These are not abstractions to Catholic Charities of Long Island — 53 sites, ~53,000 people served annually, 1,329 units of senior housing, and the institutional memory of being one of Sandy's largest relief providers.

What the node is — and is not

Is

Is not

Working demonstrations of the planning tool, the storm-mode coordination board, and the kiosk are live behind an access gate at common-light.pages.dev — built against CCLI's own program areas.

Why the Church is the operating model

Across thirty years of failures, the projects that survived — community-owned cellular cooperatives in Oaxaca, a 37,000-node connectivity commons in Catalonia, six hundred solar libraries riding existing institutions, Bangladesh's fifty-year rural electric cooperatives serving 100 million — share one property: durable local institutions owned the infrastructure. The Catholic parish is the most durable local institution on earth, with measured trust to match (Afrobarometer, 39 countries: religious leaders trusted at 66%, versus 46% for presidents). And the Church has now, in its own words, asked for this: subsidiarity — "decisions made at the closest level possible to the persons involved" (¶70) — is a distributed-systems architecture: every node autonomous, no node critical.

Pilot design

ElementDesign
SitesChosen by CCLI from its 53 sites; three primer audiences prepared: the newcomer's commons (Brentwood/Central Islip — ESL, citizenship, benefits), the senior's commons (Hempstead/Elmont, co-located with CCLI's 1,329 senior units — digital confidence, scam defense, telehealth, oral histories), the builder's commons (youth corridor — after-school tutoring, coding club, build bench)
TimelineGate checks (month 0–1: diocesan property sign-off, insurance underwriting) → build (1–5: fabrication parallel with permits and PSEG-LI interconnection) → operate & measure (5–18) → public report (18)
ComplianceNYS fire code §1206: battery product must carry UL 9540 + UL 9540A before purchase; NFPA 855 siting; ADA-compliant kiosk and route; town site-plan review
GovernanceParish council governs; CCLI is fiscal home, program partner, and the steward's employer; the Diocese of Rockville Centre provides ground, property sign-off, and insurance — its blessing precedes any build; independent evaluator holds raw metrics; funders receive quarterly reports and hold no operational control

Budget and ask

EnvelopeAmountDetail
Node capital$140,000Turnkey-mid build: solar, UL-listed battery, inverter, container, kiosk + compute, siting, permits, 15% contingency (verified range: $110–173k depending on build path)
Operations, 18 months$80,000Paid steward ($20–35k/yr — the line every failed precedent hid), insurance rider, connectivity, maintenance reserve
Independent evaluation$30,000Third-party measurement and the public report; non-negotiable
Total ask$250,000Lean single-funder entry: $150,000 (component build, evaluation intact)

Offsets reported back to funders: NYSERDA's storage incentive (live through 2030; Long Island administered via PSEG-LI), NY-Sun's commercial block if open, and PSEG-LI nonprofit programs are applied for in month one; parish in-kind (site, supervision, volunteers) is real and accounted. We model conservatively: no Starlink nonprofit rate exists, Long Island is not eligible for the state's inclusive-community solar adder, and FEMA's restored BRIC program ($1B, 75/25) is treated as the second-cycle scale lane rather than this pilot's funding.

What you would be buying: evidence, either way

The deliverable is not a ribbon-cutting. It is an independently evaluated, publicly published report: learning outcomes (pre/post assessment in the tutored tracks, benchmarked against the field's strongest randomized result — a WhatsApp math tutor in Ghana worth roughly a year of extra learning at ~$5/student); return rates and session hours by language; corpus contributions (the ownership claim made countable); human-escalation counts (the "AI assists, humans decide" claim, measured); uptime and energy served; cost per learner-month; and the month-37 sustainability covenant — which institution pays operations in year four, agreed in writing before hardware was ordered. At three nodes, the report adds what no single pilot can show: whether the pattern transfers, whether locally grown knowledge propagates, and whether the network survives the deliberate loss of a node. If the model works, the report is the template Caritas' 162 national organizations and the Vatican's new Interdicasterial AI Commission would need to carry it to communities in the Global South — village clinic and school nodes are the same architecture with more sun. If it does not work, the report says so plainly, and a modest sum will have purchased a definitive answer to a question the Church and the technology sector are both asking.

Risk register (abridged — full register on request)

RiskLevelMitigation
Diocesan property sign-off (post-bankruptcy path non-public)GatingAsked directly in week one; CCLI's separate incorporation is the clean vehicle; pilot likely below canon-law alienation thresholds
Insurance: lithium battery underwriting (Catholic Mutual)GatingRaised with diocesan Risk Manager at first contact; UL 9540/9540A-listed product; NFPA 855 siting; outdoor separation
Battery fireLow probability / high consequenceProduct-carried certifications, code setbacks, steward protocol, FD pre-plan; public incident log
Underuse ("monument risk")Moderate — the honest central riskProgram calendar anchored to existing CCLI caseloads; paid docent-steward; monthly public usage numbers; month-9 gate stops the program if people don't come back
Child & vulnerable-adult safetyZero-toleranceDiocesan safe-environment protocols inherited day one; steward + volunteers screened; sight-line interior design; incident log public
Sustainability after grantThe historical killerMonth-37 covenant signed pre-build; anchor-tenant model; co-op federation precedents (REB/NRECA) as the scaling template
Key-person dependenceModerateFounder's redundancy is a stated month-12 milestone; open design; runbook as first-class deliverable

Recognition and reporting

Funders receive quarterly written reports, raw-metric access through the evaluator, and — if desired — naming recognition on the node and in the public report. Equally honored: anonymous gifts, in the spirit of the project's namesake light.

Jeff Pinto
Technical steward, Common Light · Long Island, NY
[email] · [phone]

Provenance. Every figure above traces to a source ledger with per-claim verification status (seven research dossiers, compiled July 2026; adversarial verification pass 2026-07-14). Encyclical quotations verified verbatim against vatican.va with paragraph numbers. Known unknowns are carried openly, not hidden: the two diocesan gates, the NY-Sun Long Island block status, and the turnkey vendor quote. Full ledger and the working demonstrations available on request. Confidential draft — not for distribution before 2026-07-23; CCLI partnership in discussion, not yet confirmed; service descriptions are forward-looking.