Skip to Fulmerica links
Worth Building · Chapter 24 October 2026A draft for public use

What food technology could be for

Food may be the most natural home these ideas have. Meals are how people gather. Agricultural cooperatives are the proposal's own example of an alternative that lasted for generations. And the dominant trajectory in food, solo delivery, ghost kitchens, personalized nutrition, platform fees, is the "individually well served and collectively diminished" future at dinner-table scale. This chapter applies the ideas to food directly, then maps twelve territories with current facts, specific things to build, who is already doing them, scores, and where a reader could start.

01The idea, at the table

More capable together, not more convenient alone

Technology should make people more capable together, not just more convenient alone.

In food, the things that protect people have names already: the shared meal, the farmer who knows the eater, the grocer who knows the town, the pantry volunteer who knows the family, the daily visit that comes with the delivered lunch. Each is being optimized away by a technology pointed at the individual. None of it requires less AI to protect. It requires AI pointed at a different target.

SceneThe default pathWhat we could build instead
DinnerEach person orders alone from a phone; a ghost kitchen cooks; a contractor delivers; nobody meets.A household or a building coordinates one meal; the agent handles the ordering and the schedule; the people cook and eat together.
The restaurantPlatforms take 15 to 30 percent and own the customer; the owner becomes an approver of someone else's app.Locally owned delivery cooperatives; the agent runs the back office; the owner is on the floor greeting regulars by name.
The farmData flows to equipment and platform vendors; traceability compliance locks small producers into whoever sells the software.Producers own their data cooperatively; a shared traceability commons meets the regulation without a landlord.
The grocery storeThe last store in town closes; a dollar store opens; an app promises delivery from fifty miles away.A community-owned store with a shared back office stays open; the owner has a succession plan; the town still has a place to meet.
The benefits officeA chatbot explains the rule; the family still misses the deadline; nobody records that they did.The notice is read, the proof is assembled, the care circle is reminded, and the loss becomes visible to the people who can change the rule.
The grandmother's lunchA drone or a locker; a sensor confirms she ate.The volunteer still knocks; the system planned the route, handled the paperwork, and told the daughter she looked tired.
02Seven commitments, seven food examples

What each commitment means when the subject is food

CommitmentIn food, it meansA concrete instance
1 · AI is a means, not the measureJudge food technology by whether people eat better together and farmers and grocers stay in business, not by orders per minute.A delivery cooperative's success metric is restaurants still open and drivers' take-home pay, not gross merchandise volume.
2 · Keep people in chargeLegible rules, human judgment at the expensive moments, and exit: portable menus, portable customer lists, portable farm data.A restaurant can leave a platform with its menu, its orders, and its regulars intact. A farm can switch traceability software without re-entering ten years of lots.
3 · Build capability, not dependencePeople learn to cook, shop their benefits, and run their stores better, not just less.A WIC shopping helper that explains why an item is or is not covered, so the second trip needs no helper.
4 · Make drift visibleSurface the slow changes: platform fees creeping up, benefits quietly cut, stores closing one at a time, waitlists lengthening."Your effective platform cost rose from 24 to 31 percent in six months." "In your county, 2,100 households lost SNAP in June; pantry visits rose 40 percent."
5 · Design for groupsThe household, the building, the pantry's client families, the CSA's members, the town are the users.A CSA share split across three households with a pickup rotation; a building that orders one dinner.
6 · Protect the friction that mattersRemove the paperwork; keep the cooking, the visit, the market conversation, the volunteer's knock.Meals on Wheels routing and compliance automated so the volunteer spends the saved minutes at the door.
7 · Refuse inevitabilityDelivery apps, dollar stores, and vendor lock-in are choices made by someone. Name them, and point to the working alternatives.Locally owned delivery platforms are profitable at lower fees. Community-owned groceries work when the back office does. Agricultural cooperatives have lasted a century.

The three questions, applied

ThingMore capable or more dependent?Closer to each other or to the system?See, steer, and leave?
Third-party delivery platformDependent; the owner loses touch with their own customersCloser to the systemFees opaque; customer data not portable; exit means starting over
Personalized nutrition assistantDependent by designAlone with a screenUsually can leave
Community supported agricultureCapable: people learn to cook what growsCloser to a farmer and often to neighborsFully
Community-owned groceryCapable: a town runs a businessCloserFully, and owned
A benefits kit with a care circleCapable: can explain their own caseCloser: the circle is the userExport everything
Vendor traceability software for a small farmDependentCloser to the vendorRecords held hostage
03Where to push

Where to push in the food system

Place to pushWhat it meansIn food
ParadigmAI as a meansFood as relationship and shared capacity, not as logistics to be optimized. The meal, the market, and the farm are social infrastructure.
Goals and metricsCapability, trust, connectionMeals eaten together; repeat relationships between eaters and producers; local dollars retained; stores and farms still operating; benefits kept, not just issued.
Information flowsMake drift visible; the most practical leverWhere benefits are lost, where stores close, what platforms really charge, which waitlists grow, where donated food fails to arrive. Most of this is unmeasured.
RulesLegibility, portability, interoperability, exitTraceability standards that small producers can meet without lock-in; menu and order portability; farm data ownership and equipment repair; cottage food rules legible by state; stolen-benefit protections.
Self-organizationCooperatives, civic groupsGrocery, delivery, farm, worker, and purchasing cooperatives; food policy councils; mutual aid fridges; the shared kitchen.
Why food, specifically

The proposal says cultural orientations change when people see an alternative work, and offers credit unions and agricultural cooperatives as proof. Both were born in food and farming communities. The grange, the farm co-op, the rural electric co-op, the food co-op: the food system is where America's cooperative institutions were invented. The examples the proposal needs already have deep roots here. The task is to add AI at their core.

04Twelve territories

The food map

Scored on the catalog's rubric (shared layer nobody owns, buyer or deadline, outcome data, the three questions, fit). Each card names which of the nine reusable parts it reuses.

How to read a card: its score adds five judgments, each from 0 to 2 (shared layer nobody owns, buyer or deadline, outcome data, the three questions, fit), for a total out of 10.

F1Score 10 of 10Reuses: notice reader, deadline engine, place config, care circle

Food benefits as notices: SNAP, WIC, summer EBT, and stolen benefits

The 42 million people on the Supplemental Nutrition Assistance Program (SNAP), the 6.9 million on the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and the families whose $120-per-child summer benefit depends on which state they live in.

Why now

SNAP's new rules have cut benefits since June 2026. WIC participation rose from 6.24 million in 2021 to 6.87 million in 2025, while two state agencies moved to online electronic benefits transfer (EBT) and remote-issuance waivers run only through September 2026. Twelve states (Alaska, Florida, Georgia, Idaho, Indiana, Mississippi, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Utah) are not running summer EBT in 2026. Federal replacement of SNAP benefits stolen by card skimming ended 20 December 2024; only a few states reimburse, chip cards are rolling out slowly, and a bill to restore replacement authority was introduced this year.

Plays

  • SNAP as the Kit's second program, already pick one in part five.
  • A WIC shopping and enrollment helper that teaches the rules rather than hiding them; plain-language shopping lists exist in six states and nowhere else.
  • A stolen-benefits claim kit: notice, claim, state-by-state replacement policy, escalation, and an outcome commons that shows which states leave victims unpaid.
  • A summer-EBT tracker and enrollment path for children not auto-enrolled.

Already there

Propel (the Providers app) reaches millions of EBT households and publishes summer-EBT guides; Code for America on the state side; mRelief for screening. Nobody holds the person-side outcome record.

Score

Shared layer 2 · Buyer or deadline 2 · Outcome data 2 · Three questions 2 · Fit 2

F2Score 9 of 10Reuses: standards cohort, place config, member-owned data

Small farms and the traceability deadline

Producers, packers, food hubs, and co-ops that will have to produce lot-linked records within 24 hours of a request, and who could not stand up the systems in time once already.

Why now

The food traceability rule's compliance date moved from January 2026 to 20 July 2028 because small and mid-size producers, co-packers, and distributors said they could not build compliant electronic systems in time. The rule requires traceability lot codes, key data elements at critical tracking events, and 24-hour production of sortable records. Exemptions cover only the smallest direct-to-consumer farms. Meanwhile the average farmer is 58.1, about 300 million acres are expected to change hands in the next decade, Colorado's agricultural right-to-repair law has been in force since 2024, and parts-pairing bans took effect in three states in 2026.

Plays

  • A traceability commons: shared lot-code and record infrastructure owned by the producers who use it, with exports ready for the Food and Drug Administration (FDA), so compliance does not mean a vendor landlord. The standards cohort, exactly: fund a cohort of small-farm software tools to adopt shared formats, then win the buyers, who are food hubs, co-ops, state agriculture departments, and the deadline itself.
  • A farm data cooperative in the Driver's Seat pattern, so equipment and platform data return value to the farm.
  • A land-transition navigator alongside the Land Transfer Navigators program and the employee-ownership work in the Work and Ownership chapter.

Already there

Commercial traceability vendors; ReFED and GS1 on standards; the National Center for Appropriate Technology's (NCAT's) land-transfer program; state agriculture departments. No producer-owned commons.

Score

Shared layer 2 · Buyer or deadline 2 · Outcome data 1 · Three questions 2 · Fit 2

F3Score 8 of 10Reuses: care circle, records pipeline, outcome commons

The daily visit: Meals on Wheels and the people waiting for it

Nearly 13 million older Americans hungry or at risk, and the volunteers whose knock on the door is the point.

Why now

More than 46,000 older adults are on Meals on Wheels waitlists; one in three providers has one; the average wait is nearly four months and some run to two years. Federal nutrition funding under the Older Americans Act was cut in fiscal 2024 for the first time in a decade, and deeper cuts have been proposed. Providers report making meals for less than the state pays.

Plays

  • The proposal's aging-parent row, literally: AI does routing, compliance, and paperwork so volunteer minutes go to the doorstep; the system tells the daughter what the volunteer noticed.
  • A care circle around the meal: family, neighbors, and the provider coordinate so the person is seen daily even off the route and off the waitlist.
  • A waitlist atlas by provider and county, from records requests and the providers themselves, so the wait becomes a legislative fact.

Already there

Meals on Wheels America and its roughly 5,000 local programs, area agencies on aging, routing vendors. The data layer and the circle are open.

Score

Shared layer 2 · Buyer or deadline 1 · Outcome data 2 · Three questions 2 · Fit 1

F4Score 8 of 10Reuses: evidence file, place config, outcome commons

The last grocery store in town

The 5 million rural residents who drive ten miles or more for groceries, the 76 counties with no store at all, and the owners who have no one to hand the keys to.

Why now

More than 100 rural groceries closed in Kansas alone between 2008 and 2018. In a recent survey 24% of rural grocery owners planned to leave within five years and 73% had no succession plan. Community-owned stores work in some towns (Axtell, Kansas, population 400, rebuilt its store) and fail in others (a nonprofit store in Kansas is closing in 2026 on thin margins, high costs, and no staff). The proposal's café owner is a grocer here.

Plays

  • A shared back office owned by the stores: ordering, inventory, pricing, and the paperwork of being a SNAP and WIC retailer, drafted by models and decided by the owner on the floor. A purchasing-and-services cooperative for stores too small to have one.
  • A succession and conversion navigator (community ownership, employee ownership, municipal ownership), configured by state, with an outcome commons on which models survive.
  • A grocery-access atlas by county, pairing closures with the pantry and benefits data from F1 and F9.

Already there

Kansas State's Rural Grocery Initiative, state healthy-food financing programs, wholesaler programs. No shared back office, no conversion data.

Score

Shared layer 2 · Buyer or deadline 1 · Outcome data 1 · Three questions 2 · Fit 2

F5Score 8 of 10Reuses: outcome commons, standards cohort

Food as medicine, and who owns the referral

Patients prescribed food, the clinics prescribing it, the community organizations delivering it, and the vendor databases that currently sit between them.

Why now

New York's Medicaid waiver funds six months of medically tailored meals; other states are applying; a national Food as Medicine program was proposed as a Farm Bill amendment in 2026. Studies claim medically tailored meals cut emergency visits by 70% and admissions by 52%. The referral between clinic and kitchen runs through closed-loop referral platforms whose data terms community organizations have contested.

Plays

  • A community-owned referral and outcome commons for food-is-medicine, so the record of who was referred, fed, and helped belongs to the community organizations and patients, not a platform.
  • A produce-prescription redemption standard across the incentive programs (Double Up Food Bucks runs at 91 markets, 71 community-supported agriculture programs (CSAs), and 42 groceries in Oregon alone) and the Medicaid waivers.
  • Prescriptions filled by local farms, tying F5 to F6.

Already there

Unite Us and findhelp (referral platforms), the Food is Medicine Coalition, Fair Food Network, state Medicaid agencies. The ownership question is open.

Score

Shared layer 2 · Buyer or deadline 2 · Outcome data 2 · Three questions 1 · Fit 1

F6Score 7 of 10Reuses: care circle, standards cohort

Eaters and farmers in relationship

Community supported agriculture members and the farms that depend on them, and the incentive programs that double SNAP dollars at markets.

Why now

Several thousand farms sell through CSA shares, and across the country they report declining enrollment and low retention. The dominant fix is the meal kit: convenience alone. Nutrition-incentive funding ($36.3 million to fourteen awardees in the latest round) is fought over at every renewal, and each market and farm builds its own redemption plumbing.

Plays

  • CSA retention as a between problem: shares split across households, pickup circles, cooking together, with the agent handling the logistics. The household row, at a farm.
  • A shared SNAP and incentive redemption layer for markets and CSAs so a farm does not need a developer to accept benefits.
  • A market data commons on incentive redemption and SNAP sales for the renewal fights.

Already there

Fair Food Network, farmers market associations, CSA software vendors (Harvie, Local Line), the Gus Schumacher Nutrition Incentive Program's (GusNIP's) technical assistance center. The social design and the shared layer are open.

Score

Shared layer 1 · Buyer or deadline 1 · Outcome data 1 · Three questions 2 · Fit 2

F7Score 7 of 10Reuses: standards cohort; new: open ordering stack

Restaurants and the platforms that own their customers

Independent restaurants paying an effective 30% on platform orders, and the drivers paid by the same platforms.

Why now

Commission caps from the pandemic have mostly loosened; New York City's 2025 settlement allows total fees up to 43% with "enhanced services." Locally owned delivery cooperatives (LoCo Co-ops, grown from Iowa City and Fort Collins) charge lower fees and are profitable; The Drivers Cooperative in New York has more than 6,000 members and reports pay gains up to 30 percent. This is the proposal's "dependence without obligation" in its purest commercial form.

Plays

  • An open menu-and-order standard so a restaurant can leave any platform with its menu, orders, and customers intact: the exit lever.
  • A fee-drift monitor for owners: "your effective platform cost rose from 24 to 31 percent in six months," the proposal's own example sentence.
  • Shared technology for locally owned delivery cooperatives, so each city does not rebuild the stack.

Already there

LoCo Co-ops, The Drivers Cooperative, city fee-cap ordinances, Toast and Square on ordering. The standard and the monitor do not exist.

Score

Shared layer 2 · Buyer or deadline 1 · Outcome data 1 · Three questions 2 · Fit 1

F8Score 7 of 10Reuses: place config, evidence file

School and summer meals: the data-match that feeds children

The 27 million children in schools offering universal meals, the districts deciding whether they can afford to, and the cafeteria staff who feed them.

Why now

Eight or nine states now guarantee free school meals to every student. 27.2 million children attended a community-eligibility school in 2024-25 and 74.9% of eligible schools adopted it for 2025-26. Eligibility rests on "identified students" found by data matching to SNAP, Medicaid, and other programs, which means paperwork capacity decides whether a district qualifies.

Plays

  • Help small districts maximize direct certification: the evidence file and place configuration, pointed at a district office instead of a family.
  • The education row at the cafeteria: AI takes the meal paperwork so food-service staff feed kids; food-service worker cooperatives where districts outsource.
  • Track the twelve states out of summer EBT and what it costs their children, for the next legislative session.

Already there

The Food Research and Action Center, School Nutrition Association, state agencies, point-of-sale vendors. The small-district capacity gap is open.

Score

Shared layer 1 · Buyer or deadline 2 · Outcome data 1 · Three questions 1 · Fit 2

F9Score 8 of 10Reuses: outcome commons, records pipeline

Food banks under the cuts, and the data they already hold

Pantry volunteers and the families in line, during the largest shift of hunger from public benefits to charity in a generation.

Why now

Food bank demand is up 11 to 300 percent by region in 2026. Feeding America estimates the SNAP changes equal up to 9 billion meals a year, more than its entire network provided last year; for every meal a food bank gives, SNAP gave nine. About 65% of the network uses one intake system, which means a person-level record of need already exists, held by a vendor and the network. Community fridges hold no records at all, by design.

Plays

  • A consented outcome commons that links benefit losses to pantry demand by county: the single most persuasive "make drift visible" dataset in food this year.
  • Pantries as distribution for the Kit's SNAP program; the intake conversation is where the notice gets read.
  • Logistics for mutual aid that never asks for a name: fridges, gleaning, neighbor networks, where the human exchange is the point.

Already there

Feeding America and its 200 food banks, Link2Feed and Oasis Insight for intake, Freedge and local fridge networks. The linkage to benefit outcomes is open.

Score

Shared layer 2 · Buyer or deadline 2 · Outcome data 2 · Three questions 1 · Fit 1

F10Score 7 of 10Reuses: place config, evidence file

Home kitchens and shared kitchens

Immigrant and rural food entrepreneurs selling from home, and the step to a licensed kitchen that most never take.

Why now

Nine states expanded cottage food rules in 2026, allowing refrigerated items under direct-sale conditions with revenue caps from $50,000 to $250,000 before a licensed facility is required. Labeling, allowed foods, water tests, and handler cards differ by state, which is the fragmentation cause in a kitchen.

Plays

  • A compliance kit for home food businesses configured by state: what you may sell, to whom, how to label, when you cross the cap.
  • Shared commercial kitchens as cooperatives, the shared infrastructure for the step up, with scheduling and compliance handled by the agent.

Already there

Forrager and state extension offices for rules; kitchen incubators (La Cocina, Hot Bread Kitchen) in cities. Rural and small-city coverage is thin.

Score

Shared layer 1 · Buyer or deadline 1 · Outcome data 1 · Three questions 2 · Fit 2

F11Score 6 of 10New: donation matching

Surplus food and the handoff that fails

Stores and restaurants with surplus, pantries and fridges that need it, and the liability and logistics in between.

Why now

About 29% of the U.S. food supply, 70 million tons worth $380 billion, went unsold or uneaten in 2024, while 48 million people were food insecure and less than 13% of donatable food was donated. The Food Donation Improvement Act extended liability protection to direct donation to individuals.

Plays

  • Donation matching that keeps the relationship local and the data with the community, not a platform that owns the match.
  • A liability-safe documentation kit for small donors who still believe they cannot give.

Already there

Too Good To Go and Flashfood (commercial), Food Rescue US, 412 Food Rescue, Copia, ReFED's data. Crowded; enter only as the data-ownership alternative.

Score

Shared layer 1 · Buyer or deadline 1 · Outcome data 1 · Three questions 2 · Fit 1

F12Score 6 of 10Reuses: civic legibility (part five)

Local food governance

The 358 food policy councils and state networks counted in the last census, deciding school procurement, market zoning, and SNAP outreach in public, mostly unheard.

Why now

The civic-life row: tools that make local decisions legible and help neighbors organize. Food policy councils are the existing civic body; most have no staff and no shared data.

Plays

  • Meeting documentation and plain-language decision records for councils, in the Documenters pattern.
  • A shared data layer across councils so one town's procurement standard is another's starting point.

Already there

Johns Hopkins' Food Policy Networks project, state networks. Tooling is absent.

Score

Shared layer 2 · Buyer or deadline 0 · Outcome data 1 · Three questions 2 · Fit 1

The household row, and what not to build

The proposal's household row, a shared system that makes everyone more available to each other, has an obvious food expression: cooking and eating together. It also has an obvious wrong turn: another meal-planning or nutrition assistant for one person. The pattern worth building is the one that coordinates a household's or a building's shared meal, learns nothing about the individual it does not need, and schedules the human part rather than replacing it. Offer it to the CSAs, the pantries, and the senior-meal programs rather than selling it to consumers.

05An ordinary day

A Tuesday, when the technology is pointed at the right target

In a town of 1,400 the grocery store opens on time because its ordering ran overnight and flagged one decision for the owner, who spends the morning talking to customers. At the pantry a woman photographs a letter about her food benefits; the volunteer sees the deadline, the exemption she qualifies for, and the two documents she needs, and spends the rest of the visit asking about her kids. Forty miles out a farmer ships lettuce to the school district with lot codes the producers' shared system generated for nothing, and the district's cafeteria serves every child free because a clerk's software matched the enrollment data in an afternoon. At noon a volunteer brings lunch to a man of 84 and stays twelve minutes, because the route and the forms took none of her time; his daughter gets a note that he seemed tired. A restaurant two streets over runs its own delivery through a cooperative the owners share, and the owner knows her regulars' names because the app does not need to. In the evening three households split a farm share on a porch. None of this needs less technology. It needs technology that is for the people at the table.

Signs we are getting there

  • The number of counties without a grocery store falls, and the stores that remain have succession plans on file.
  • Stolen food benefits are replaced in every state within two weeks of the report.
  • Pantry intake conversations end with a deadline met, not only a form filled.
  • Small producers meet the 2028 traceability rule with producer-owned systems, and none report losing records when they switch software.
  • Meal-delivery waitlists shrink while volunteer minutes at the door rise.
06Where you could start

Three sizes of first step

A weekend

  • Map grocery access in your county: every store, its hours, whether it takes SNAP and WIC, and the distance from the nearest town without one (F4).
  • Photograph and transcribe your own benefits notices, identifiers removed, for a shared test set the notice reader learns from (F1).
  • Drive one Meals on Wheels route and write down every minute spent on paperwork rather than at a door (F3).
  • Attend and document one food policy council meeting: agenda, decisions, promises with dates (F12).

A season

  • With one pantry, run the SNAP case type for thirty households and record what the state did with each case (F1, F9).
  • With one Meals on Wheels provider, automate routing and reporting and measure minutes at the door before and after (F3).
  • With one food hub, prototype shared lot codes and records for five farms, a year and a half before the July 2028 date (F2).
  • With one farmers market, build the pipeline that reports incentive redemption and benefit sales without a developer on staff (F6).

A career

  • Convene the producer-owned traceability cohort and carry it to the state agriculture departments and food hubs that will buy it (F2).
  • Build the shared back office cooperative for rural grocers, with Kansas State's Rural Grocery Initiative as the first partner (F4).
  • Lead a community-owned referral and outcome commons for food as medicine inside a state Medicaid waiver (F5).
  • Run the observatory that links benefit losses to pantry demand, county by county, and puts it in front of legislators every quarter (F9).

Who is already doing it

  • Fair Food Network runs Double Up Food Bucks, the dollar-for-dollar produce match now in many states.
  • Kansas State University's Rural Grocery Initiative has worked on rural store survival and community ownership since 2006.
  • LoCo Co-ops builds locally owned restaurant delivery cooperatives city by city.
  • Meals on Wheels America and its local programs deliver meals and the daily visit.
  • Feeding America's network and the intake systems most of it uses hold the person-level record of need.
  • Propel reaches millions of households on benefits with balance and program information.
  • Code for America builds the state side of benefits enrollment and, in 2026, caseworker tools.
  • The National Center for Appropriate Technology runs the Land Transfer Navigators program for farmland succession.
  • ReFED measures food waste and the donation gap each year.
  • Johns Hopkins' Food Policy Networks project supports the hundreds of local food policy councils.
  • Community fridge networks such as Freedge keep food moving between neighbors with no paperwork at all.
07Scorecard

Twelve food territories

Score key: SL shared layer nobody owns, BD buyer or deadline, OD outcome data accrues, 3Q the three questions, FT fit for a builder who convenes. Each is scored 0 to 2, so the total is out of 10.

CodeTerritoryShared layerBuyer or deadlineOutcome dataThree questionsFitTotalHorizon
F1Food benefits as notices2222210Now, inside the Kit
F2Small farms and the traceability deadline2212292027 cohort for July 2028
F3Meals on Wheels and the daily visit2122182027 with a provider
F4The last grocery store2112282027 with the Rural Grocery Initiative
F5Food as medicine referrals2221182027 with a state waiver
F9Food banks and benefit-loss data222118Atlas slice, 2027
F6Eaters and farmers1112272028
F7Restaurants and platforms2112172028
F8School and summer meals1211272027
F10Home and shared kitchens1112272028
F11Surplus food handoff111216Only as data-ownership alternative
F12Local food governance201216With part five's civic work
08Picks and declines

Six food plays that fit the plan, and four to decline

  1. Food benefits inside the Kit now (F1). SNAP is the first pick in the Where Scale Failed chapter; add the stolen-benefits claim kit and the WIC shopping helper, because both are notice-shaped and both teach the rule.
  2. The producer-owned traceability commons (F2). The clearest standards-cohort play in food: a hard federal deadline in July 2028, small producers who already failed once to comply, and no shared layer. Convene the small-farm software tools in 2027.
  3. The daily visit (F3). The proposal's aging-parent row with a 46,000-person waitlist. Partner with one Meals on Wheels provider; route and file so volunteers knock; build the circle around the meal.
  4. The last grocery store (F4). The café row in a town of 400. A shared back office and a succession navigator with Kansas State's initiative.
  5. Benefit loss to pantry demand (F9). The Atlas slice that makes this year's drift visible, from data the network already holds, with consent.
  6. Who owns the food-is-medicine referral (F5). A community-owned outcome commons offered to the states now writing waivers, so the record between clinic and kitchen does not become a vendor's asset.
Personal meal-planning and nutrition assistants. Convenience alone, for one. The proposal's warning in app form.
Another food-rescue marketplace. Crowded; enter only as the alternative that leaves data and relationships with the community.
Precision-agriculture data platforms. Venture-owned and vendor-locked by design; the cooperative is the only version worth building.
Ghost kitchens and delivery-first restaurant tooling. They optimize away the exact thing worth protecting: people eating together.

Assumptions and limits. Scores are judgments. The CSA count is stated loosely because published figures range widely; the health-outcome claims for medically tailored meals are from program advocates and should be read as such. The intake-system share of the food bank network is as reported in trade coverage. Universal-meal state counts differ between eight and nine depending on source and date.

Sources added in this partChecked 4 October 2026