**Artifact** from Bead: F5b21ea4 · [canonical source](https://redfish.acequia.io/guerin/.agents/f5b21ea4-2b73-4c8b-96f0-892f63ad86cf/2026-06-04/artifacts/intelligence-blindness-and-the-digital-acequia.md) · session 2026-06-04 · discussion: Talk: F5b21ea4
## The wildfire-to-social-services analog of ground-up intelligence and the substrate that makes it possible **Bead:** `f5b21ea4-2b73-4c8b-96f0-892f63ad86cf` · **Date:** 2026-06-04 **Status:** Working draft. Captures the design move Stephen named on 2026-06-04: the digital acequia is, operationally, an *intelligence substrate* — what realtime.earth and alert.live are being built to be for wildfire, applied to social services. **Companions:** [intermediation-and-selective-power.md](intermediation-and-selective-power.md) (the funding-intermediation problem), [user-stories-call-to-action.md](user-stories-call-to-action.md), [ostrom-principles-corporate-vs-acequia.md](ostrom-principles-corporate-vs-acequia.md).
## The opening claim > *"It's like fighting a wildfire with only satellite imagery."* > — Stephen, 2026-06-04 The current paradigm for crisis response and the current paradigm for social services share the same structure: centralized institutions collect coarse data, do centralized analysis, decide resource allocation, and push minimal information back to citizens. The result is *blindness* — a structural inability to see what's actually happening on the ground in the time windows when seeing matters. We are blind on: - **The first 3 hours of a wildfire.** Paradise. Palisades. The window when the fire is establishing and direction shifts decide whether a town survives. - **The first 30 minutes of a post-fire flash flood.** When burn-scar hydrology is fundamentally different from any prior model. - **The first 3 minutes of a school shooting.** When situational awareness for first responders, parents, and witnesses determines who lives. And we are blind in social services every day: the first 24 hours after a homeless camp loses its informal community-policing capacity; the first week after a foster placement falters; the first month after a new substance-use crisis emerges in a neighborhood; the first season when a food-pantry network's actual capacity shifts. Bureaucratic information is *coarse* — quarterly reports, KPI dashboards, intake forms standardized across counties. Ground-level reality is invisible to it. But ground-level reality is not invisible to the people on the ground.
## What's available but unused — wildfire In the wildfire case, the ground-up intelligence already exists and is unused: - **10,000 camera feeds** scattered across the affected geography: residential security cameras, traffic cameras, gas-station cameras, school cameras, doorbell cameras, social-media live feeds, cellphone video. - **Residents' phones**, each carrying a GPS, microphone, camera, and an exhausted local who knows the back roads. - **Local fire-watch knowledge** held by people who have lived through prior fires. - **Aircraft imagery** that the government collects but largely does not release. The current paradigm cannot process this. The system is *one-way*: information flows from government to citizens (when it flows at all), not back. The government controls the aircraft imagery and doesn't provide it. The 10,000 camera feeds exist but no infrastructure aggregates them in real time for use by the people whose neighborhoods are burning.
## What's available but unused — social services Stephen's example: > *A local homeless person knows where to get a warm meal, blankets, informal medical attention. They have mobile devices that can use cameras and microphones for local camp security via community-based policing to complement governmental law enforcement. They can produce maps, indicate where need is, and also map resources — e.g. this homeless guy has a chauffeur license and can pick up 10 people through the day that need medical care if an extra Santa Fe Trailways medical bus is available. Or better yet, a philanthropic donation of a car and funding for insurance.* This is *high-resolution local intelligence.* It includes: - **Where the resources are** — specific warm-meal sites, blanket sources, informal clinics, parish halls open on specific evenings. - **Where the need is** — specific people, specific shortages, specific hot spots. - **What unrecognized capacity exists** — the chauffeur license; the spare bus capacity; the elder with a working pickup who could be paid to drive on demand. - **What the matching problem actually is** — *this* person could drive *these* ten people if someone gave them a car and insurance. Currently invisible because no institutional process has any way to surface it. None of this is in any state agency's database. None of it is in any nonprofit's intake form. It is in the heads and the phones of people who live in the situation.
## The substrate Stephen's team is building Stephen, Simtable, Redfish, and others are building infrastructure that solves the wildfire version of this problem: - **acequia** — the substrate name and the governance metaphor: community-hosted, locally-stewarded, peer-to-peer between local nodes with selective coordination through wider channels. - **realtime.earth** — real-time geospatial intelligence aggregated from many sources, with local sovereignty over who sees what. - **alert.live** — the alerting and distribution layer for time-sensitive events. The technical pattern, in Forman's landscape-ecology vocabulary already used by the Groundworks proposal: - **Patches** = symmetric host servers (community-owned or community-trusted) running cameras, compute, storage, bandwidth, edge-based AI/ML inference. The local node *is* the patch. - **Corridors** = routes between patches — peer-to-peer where possible, cloud-mediated where necessary, with the cloud serving as fallback rather than centerpiece. - **Matrix** = the underlying network substrate that determines whether information can flow at all. The matrix today is hostile to ground-up intelligence: federal data systems siloed, vendor APIs gated, bandwidth controlled by carriers, compute centralized, AI/ML inference happening in data centers a continent away. The matrix Stephen's team is building is *peer-to-peer plus cloud*: distributed processing at the edge, lightweight aggregation where genuinely needed, no central choke point.
## The central claim **The same substrate solves the social-services blindness.** The technical pattern is invariant across domains. Wildfire and social services differ in *what* the cameras and microphones are recording, *who* the parciantes are, and *what* the intelligence is used for. They do not differ in the structural fact that: - Ground-level intelligence already exists. - Devices to collect it already exist. - The centralized paradigm cannot process it. - A peer-to-peer-plus-cloud substrate, locally governed, can. For Groundworks: **the digital acequia is operationally an intelligence substrate before it is anything else.** Communities cannot govern allocation if they cannot see what they're allocating. The "selective power" articulated in [intermediation-and-selective-power.md](intermediation-and-selective-power.md) — communities choosing which nonprofits serve them — requires high-resolution visibility into what the community actually needs, who actually has unused capacity, and what the matching problem actually is. Without the substrate, governance reform is rearranging old data flows under new banners. With the substrate, governance reform has something to govern *with*.
## The chauffeur example, fully traced Take Stephen's chauffeur example as the canonical case. **Under the current system:** 1. State Medicaid contracts with a transportation broker. 2. Broker contracts with a vendor. 3. Vendor schedules rides through a centralized dispatcher. 4. Eligible patients call a number, wait for confirmation, are picked up — or not. (No-show rates in rural NM Medicaid transport are widely reported as high.) 5. A homeless man with a chauffeur license, available time, and existing relationships with ten people who need rides is *invisible* to this system. He has no contract with the broker. He has no vehicle. He has no insurance. He has no formal way to participate. **Under the digital acequia substrate:** 1. The local community-led acequia (at neighborhood scale) maintains a community-hosted registry of *known capacity* and *known need* — built by parciantes who actually know who needs what and who can do what. 2. The chauffeur-license-holding parciante is in the registry. The ten people who need rides are in the registry. The matching problem is visible in the acequia's records. 3. The matching surfaces a specific resource gap: one car + insurance + fuel = ten people get to medical appointments reliably. 4. A philanthropic funder, looking through the acequia's published needs (which are no longer abstract "rural transportation need" but specific named gaps with named capacity), funds the gap directly — a car, insurance, fuel — to the named parciante, accountable to the acequia. 5. Total intermediation: one funder → one acequia → one parciante. Total bureaucratic overhead: minimal. Total selective power held by the community: complete. This is what the digital acequia means *operationally*. Not an abstraction. A working intelligence-and-allocation substrate.
## What this means for the Groundworks proposal The published proposal at [acequia.org/groundworks/](https://acequia.org/groundworks/) currently frames the intervention as governance reform — *communities decide how to allocate.* This is necessary but not sufficient. Without an intelligence substrate, "communities decide" reduces to "communities receive whatever information the centralized institutions choose to share with them." The same one-way flow, just under friendlier rhetoric. The proposal should name the intelligence substrate as an *operational component* of the intervention. Stephen's existing work on realtime.earth and alert.live is the demonstration. The same substrate extended to social services is the digital acequia in practice. Concretely: insert a short **§The substrate** section between §The intervention and §A proposed strategy focus, naming — - The wildfire-and-social-services blindness analog - The substrate components: peer-to-peer plus cloud, community-hosted, locally-governed, multi-modal sensing, edge-based AI/ML - Forman applied at the substrate level (patches = hosts; corridors = routes; matrix = network), in addition to the existing social-level reading - The connection to actually-existing realtime.earth and alert.live infrastructure - A commitment that the social-sector extension is specific scoped work, not an aspiration This avoids the "vague tech utopia" failure mode by pointing at infrastructure that exists.
## The Forman bi-level reading The Groundworks proposal already uses Forman's landscape ecology at the *social* level: - Patches = organizations (schools, clinics, mutual-aid groups) - Corridors = referral relationships, trust, collaborative routines - Matrix = funding rules, government procedures, data infrastructure Stephen's framing extends Forman to the *technical substrate* level: - Patches = local host servers (cameras, compute, storage, bandwidth, edge AI/ML) - Corridors = peer-to-peer routes plus cloud paths - Matrix = the underlying network substrate Both levels exist and they interact. A strong social patch (a community nonprofit) without technical substrate is blind to its own constituents. A strong technical substrate (a community-hosted server network) without social patches is data without governance. The proposal's intervention works only when both levels are addressed together — and the same Forman vocabulary makes the connection legible.
## What the substrate is NOT - **NOT a centralized data extraction system.** The substrate is community-hosted; data stays at the edge unless explicitly shared. - **NOT a single database or consolidated platform.** Many local nodes, federated through P2P plus selective cloud. - **NOT an elimination of institutions.** State agencies, foundations, nonprofits continue to exist; they engage with the acequia substrate as participants rather than as gatekeepers. - **NOT a cryptographic argument.** The earlier draft of [intermediation-and-selective-power.md](intermediation-and-selective-power.md) gestured at zero-knowledge proofs, multi-party computation, and decentralized identifiers. Per Stephen's 2026-06-04 redirect, the substrate framing replaces the crypto framing for board audiences. The operational claim is *"it's how you actually know what's happening instead of bureaucratic and coarse info."* Cryptographic detail belongs in a deeper technical companion for a different audience.
## Process notes - Stephen's wildfire examples (Paradise; Palisades; 3-hour / 30-minute / 3-minute blindness windows) are his own domain expertise from Simtable. They should remain attributed. - The chauffeur example is fully developed here for the social-services case. Stephen could swap in or add other concrete examples for the published proposal: late-frost crop loss in Chimayó; homeless-camp community policing; food-pantry capacity matching; foster-placement support networks; post-disaster mutual aid coordination. - The "same substrate, different domains" claim is the central design move. If this artifact gets integrated into the published proposal, that claim should land sharply — probably as the keystone sentence of the new §The substrate section. - This document does not propose specific technical architecture (protocol choices, hosting models, identity systems). Those belong in a technical companion. The substrate framing here is at the level the board needs: *what it does* and *why it is the operational shape of the digital acequia*, not *how it is implemented*.