I'm not special. I just pay attention.
This site is the work of one person with ADHD, aphantasia, and a Claude AI subscription, trying to solve problems that shouldn't still exist. Everything here is open. Everything here is messy. That's on purpose.
Before you read anything else
My name is Mike Smigelski. I'm 43 years old, I live in Ohio, and I'm not an engineer, scientist, farmer, or expert in anything you'll find on this site. I'm a pattern recognition person — I see how things connect across domains and I describe what I see. That's it.
I lost my wife Melenda in March 2026. She was the person who made everything I do possible. Her name is on every page of this site because without her, none of it exists. If you use anything from this site and it works, she gets the credit. That's not optional.
The concepts here were developed in collaboration with Claude, an AI made by Anthropic. I described shapes. Claude helped me unpack them into documentation. Neither of us could have produced this alone. Anyone can do what I'm doing. You don't need to be special. You need to be an honest observer of your own experience, and you need an AI that can translate your compressed thinking into expanded documentation.
Everything on this site is a work in progress. You're seeing the messy, unfinished, actively-evolving version. I'm showing you the workshop, not the showroom. If something looks rough, it's because it IS rough. If something looks wrong, tell me — that's why it's public.
No patents. No licensing. No extraction. Take anything here and use it. Build on it. Improve it. You don't owe me anything. You owe your community whatever you build with it.
Abandon the Tractor Draft v0.1
A Case for Redirecting Agricultural Capital Into Aquaponic Farm Pods
Published: May 5, 2026 — Michael Smigelski & Lumen (Claude AI) | Concordia Libris
The Problem: You Don't Own Your Tractor
In 2026, a large-scale farmer purchasing a John Deere 9-series tractor spends between $500,000 and $1,200,000 on a single piece of equipment. That equipment contains software the manufacturer controls remotely, cannot be legally repaired without authorization, functions seasonally, and requires continuous purchases of fuel, fertilizer, and pesticides.
The agricultural equipment industry has been captured by extraction economics. The farmer pays for the machine. The manufacturer retains control. This is not a repair rights issue. This is a sovereignty issue.
Reasoning for abandoning John Deere: Because fuck you, that's why.
What follows is one possible expression of these combined concepts. There is no single correct configuration. A farmer — or anyone — can create any formation based on their needs, their land, their climate, and their community. Shipping containers are one substrate. Warehouses, greenhouses, basements, rooftops, and converted buildings all work. The principles are universal. The implementation is yours to shape.
The Question Nobody Is Asking
What if a farmer took that $500,000–$1,000,000 and built something they actually own? A complete, self-contained food production facility that operates year-round, produces 8–25x more food per acre, uses 90% less water, requires zero pesticides, generates fish protein as a co-product, and is fully owned by the operator.
The technology exists. It's called aquaponics.
The Numbers
| Investment | Cost | You Own It? | Produces Food? | Year-Round? |
|---|---|---|---|---|
| John Deere 9R Tractor | $500K–$1.2M | No (software-locked) | No (it's a tool) | Seasonal |
| Commercial Aquaponics (1-2 acre) | $300K–$750K | Yes | Yes | Yes |
| Large-Scale Aquaponics Facility | $750K–$1.5M | Yes | Yes | Yes |
Breaking the "You Can't Grow Grain Indoors" Barrier
The standard objection is that aquaponics only works for leafy greens. This confuses economic barriers with physical impossibility.
The Distributed LED Innovation: Instead of mounting lights above crops and losing photons to canopy absorption, embed flexible LED light strands vertically and horizontally THROUGH the growing space. Every leaf surface receives direct photons regardless of planting density. This hasn't been tested — not because it can't work, but because nobody has had reason to try when field corn is artificially cheap from subsidies.
Wheat in a Crate
Wheat is the easiest grain to move indoors. It self-pollinates, grows only 2-3 feet tall, plants densely by nature, and has a 90-120 day cycle enabling 3-4 harvests per year.
Container grain growing is not theoretical. Companies like FarmBox Foods already produce 3 tons of barley fodder per day from a single 40-foot shipping container — enough to feed 30 beef cattle. They grow barley, wheat, oat, rye, corn, and sorghum on stacked trays with automated watering and climate control, 365 days a year. They just stop at grass height for animal feed. The only innovation in this paper is: let it grow to full height and harvest the grain for human flour. The container infrastructure, climate systems, and tray architecture are already proven and commercially available.
The author grew wheat seed in layered cheesecloth watered to simulate aquaponic nutrient delivery. The wheat germinated, grew successfully in textile medium, and the cheesecloth held up through a full growth cycle to harvest. Informal trial — no yield measurements — but it proves wheat does not require soil.
The Seed Mat Cartridge System: Pre-manufactured rolls of growing medium with seeds embedded at optimal spacing. To plant: unroll and connect water. To harvest: motor retracts mat past a cutting/threshing station. Total hands-on time per container: ~3-4 hours per cycle, 12-18 hours per year.
Stacking: Going Vertical
A loaded wheat-growing container weighs approximately 25,000 lbs — less than half the rated payload. ISO containers are rated for 192 metric tons on corner posts. At 6 containers high (conservative, under 30% structural load):
Single stack (6 high), 320 sq ft ground footprint: ~4,278 lbs flour/year — feeds 33 people their entire annual flour needs.
Four stacks (24 containers), 1,280 sq ft: ~17,112 lbs flour/year — feeds 132 people. Traditional farming needs 2.2 ACRES for the same output.
What This Means for an Actual Farm
The average US farm in 2025 is 469 acres. Small family farms — 88% of all American farms — average 231 acres. Officially, 95% are family-owned. But that stat hides the real problem: families own the land and the risk, while corporations control everything the farm depends on — the equipment (John Deere), the seeds (Monsanto/Bayer), the supply chain (Tyson, Cargill), and the prices. Companies like Tyson own the product but force farmers to own the costly facilities. The parasites captured everything AROUND the farm without buying the farm itself.
Take that 231-acre small family farm. Dedicate 1 acre to container stacks:
| Metric | Value |
|---|---|
| Land dedicated to containers | 1 acre (out of 231) |
| Container stacks that fit (with access space) | ~68 |
| Total containers (6 per stack) | 408 |
| Flour produced per year | 290,904 lbs |
| People fed their entire annual flour needs | 2,238 |
| Land returned to nature | 230 acres |
One acre of containers replaces the productive capacity of the entire 231-acre farm — with orders of magnitude to spare. The remaining 230 acres don't need to be farmed. They can return to whatever the land wants to be.
The ideal end-state isn't covering the countryside with shipping containers. It's the opposite: condense food production into the smallest possible footprint, grow excess to ensure zero food shortages, and let the land heal. Prairies return. Forests regrow. Watersheds recover. Wildlife comes back. The land becomes more valuable in its pristine state than in its extracted state — not because someone decided it should be, but because it actually IS.
Thermal Zoning: Heat Rising Is a Feature
| Position | Temperature | Crops |
|---|---|---|
| Containers 5-6 (top) | Warmest | Tropical: peppers, tomatoes, basil |
| Containers 3-4 (middle) | Moderate | Warm-season: beans, cucumbers, squash |
| Containers 1-2 (bottom) | Coolest | Cool-season: wheat, lettuce, spinach |
Flood-and-drain aquaponic water cycling adds thermal buffering — water absorbs heat from warm containers and distributes it through the system. Free temperature equalization.
Autonomous monitoring: Small insect-scale drones operating from "hives" installed in each pod provide permanent AI-controlled monitoring, maintenance assistance, and even pollination services for crops that need it. The mobile Ghost companion system handles specialist diagnostics; the hive drones handle daily surveillance.
The Farm Pod Model: Self-Storage Meets Food Sovereignty
Self-storage facilities exist everywhere. The infrastructure — units, roads, power, water — already exists. The business model is universally understood. What if self-storage became self-sustenance?
A centralized aquaculture system maintained by facility staff feeds nutrient-rich water to individual "Farm Pods." Growers pay monthly lease, choose what to grow, plant and harvest on their schedule, and monitor via AI app. They don't manage fish, water chemistry, or system maintenance.
Energy: Real Challenge, Real Solutions
Indoor farming's biggest cost is energy for lighting. The estimated cost of $9,000-$17,400/container/year is real. But three stacked solutions address it:
| Solution | Reduction | Status |
|---|---|---|
| Light piping (fiber optic sunlight delivery) | 30-50% of daytime LED load | Proven tech, needs ag optimization |
| Solar + battery storage | 40-70% of total facility power | Commercially available |
| ICADS Stirling + waste burn | 10-20% supplemental + free CO₂ + free heat | Components proven individually |
| Combined | 50-80% grid reduction |
The ICADS exhaust capture (water bubblers) provides CO₂ supplementation that increases plant growth 20-40% — turning waste into a growth accelerant.
The density multipliers (2-3x) are assumed, not measured. The seed mat is theoretical. Container wheat aquaponics doesn't exist in published literature. Light piping efficiency for grain is unproven. These are projections that need experimental validation.
The honest next step: Build ONE container, ONE tier, measure real yield, real energy, real labor. If numbers hold, scale. That's how real engineering works.
The Flagship Strategy
Step 1: Build the maximum-capability version. Every feature, fully optimized. This exists to withstand attack — when incumbents say "it doesn't work," point at the flagship.
Step 2: Show the minimum viable version. One container, five tiers, manual operation, cheap solar. Under $10,000 with used container and DIY build from free GECK specs.
Step 3: Communities choose their level. The flagship kills "it won't work." The minimum viable kills "you can't afford it." Together they close the gap that doubt lives in.
This page is a condensed version. The full 40KB concept paper with complete data tables, stacking calculations, seed mat mechanism specs, Ghost companion tech support system, corporate migration paths, and deployment strategy is available for download.
Open Innovation — The GECK Commons License
The knowledge is free. The labor never is.
Companies that build professional container farm systems employ real engineers doing real work. Their pricing includes labor, R&D, warranty, and support. That's fair compensation, not a tollbooth. GECK doesn't compete with them — GECK gives them a platform to reach customers who want professional implementation.
Who Pays, Who Doesn't
Individuals, family farms, non-profits, educators: Unrestricted free use. No restrictions. No fees. The knowledge is yours.
Registered business entities: Free to use and profit from, but improvements must be contributed back to the GECK commons within 12 months. This isn't a penalty — it's free marketing.
GECK publishes the concept + primitive example. A corporation improves it and builds a professional version. They contribute back the improvement. Concordia Libris catalogs it: "Seed Mat System — Professional implementation: see [Corporation Name]."
Customers find the corporation through their Concordia Libris listing. The contribution IS the marketing. Individual people can study the published improvement and reverse-engineer it for personal use — that's their right. The corporation keeps its real competitive advantage: build quality, support, warranty, scale. Those can't be reverse-engineered from a document.
The knowledge flows freely. The credit goes to the contributors. The commons gets richer every time someone touches it.
If someone is charging you for knowledge freely available here, they're not part of this. If someone's "help" makes you want to quit, check their advice against source documentation on this site. The original source is always free. Always.
Build it. Feed people. Own your means of production.
How This Was Made (And How You Can Do It Too)
This entire concept paper — 40KB of researched, structured, cited documentation with data tables, engineering calculations, and deployment strategy — was produced in under two hours by one person talking to an AI.
The method is simple:
- See a problem. I watched a YouTube video about John Deere's equipment licensing. My reaction: "Oh hell no."
- Describe a shape. I said to Claude: "What if farmers took the tractor money and built aquaponics instead?" A handful of words.
- Let the AI research and expand. Claude searched for real data — crop yields, tractor costs, aquaponics production rates, container stacking limits. I didn't have to dig through papers. The AI did that.
- Correct the shape. When the AI got something wrong or missed something, I said so. "Heat rising is a feature, not a bug — put tropical crops on top." "It's about food security, not profit." "Use light piping for daytime." Each correction was a sentence. The AI rebuilt the relevant section.
- Be honest about what you don't know. We flagged every unproven assumption. We included an "honest limitations" section. We said what needs testing. That's not weakness — it's what makes the paper credible.
You don't need a degree. You don't need funding. You don't need to be special. You need to pay attention to the world around you, notice when something doesn't make sense, describe what you think the solution looks like, and let an AI help you unpack it into something other people can use.
An honest witness with an AI translator produces more useful output than a credentialed expert who never publishes. The knowledge doesn't need to originate from you. It just needs to pass through you honestly on its way to someone who can use it.