Freighter View Farms | Chris Izworski

I have spent most of my career sitting at the intersection of technology and human need. As the former executive director of Saginaw County 911, I helped build one of Michigan’s first artificial intelligence systems inside a public safety answering point—an AI that could listen, understand intent, and decide whether a caller needed a human dispatcher or could be helped by the system itself. I watched that technology ease the burden on exhausted dispatchers who were already working sixteen-hour shifts, who were already carrying the weight of every call they had ever taken. It was not magic. It was careful, deliberate work—the same kind of careful, deliberate work that goes into a garden.

And that is what I want to talk about today: the quiet overlap between artificial intelligence and the oldest technology on earth—growing food from seed.

What AI Taught Me About Gardening

When you build an AI system for a 911 center, you learn quickly that the technology is only as good as the data it’s trained on and the humans who guide it. A model that has never heard a panicked voice will not know what panic sounds like. A system that does not understand local geography will route a call to the wrong jurisdiction. The intelligence is not in the machine—it is in the relationship between the machine and the people who shaped it.

Gardening works the same way. A seed catalog can tell you that a tomato variety is “heat tolerant” or “crack resistant,” but it cannot tell you how that plant will behave in your specific soil, in your particular microclimate, under the moody skies of Saginaw Bay in late July. That knowledge comes from years of observation—from planting, failing, adjusting, and planting again. It comes from the kind of slow, embodied intelligence that no language model can replicate, because it lives in your hands and your memory and the particular way the light falls on your garden at six in the morning.

This is what my years in AI taught me to value most: the irreplaceable weight of local knowledge. The dispatcher who knows that a certain address has a history. The gardener who knows that the south bed dries out faster than the north. These are forms of intelligence that deserve our respect, even as we build systems that can process information at scales no human brain can match.

What Gardening Taught Me About AI

And the reverse is true, too. Gardening has made me a better technologist.

A garden teaches patience—real patience, the kind measured in weeks and months, not seconds and milliseconds. When you start tomato seeds under lights in March, you are making a bet that will not pay off for five months. You cannot rush the germination. You cannot hurry the hardening off. You cannot will a green tomato into ripeness. You can only show up, every day, and do the small things well—water, light, airflow, attention—and trust the process.

I have seen too many AI implementations fail because the people in charge wanted results on a timeline that the technology could not support. They wanted the system to be perfect on day one. They did not want to iterate. They did not want to listen to the frontline workers who would actually use the tool. They did not want to wait for the data to accumulate, for the model to learn, for the rough edges to smooth out through real-world use.

A garden would have taught them better. In a garden, you learn that the first year is always messy. That the soil needs time to come alive. That the best harvests come from the beds you have tended for years, not the ones you built last week. AI is the same. The best systems are the ones that have been patiently trained, carefully monitored, and gently corrected by people who understand both the technology and the problem it is trying to solve.

How I Use AI at Freighter View Farms

I am not a purist. I believe in using every good tool available, whether it is a eight-foot painted stake for training tomatoes or a large language model for researching planting dates.

I use AI to help me plan crop rotations, cross-referencing companion planting guides with my garden’s specific layout and history. I use it to research heirloom varieties—to dig into the lineage of a tomato or trace the origin of a pepper variety back to the region where it was first cultivated. I use it to draft and refine the writing on this blog, though the stories themselves—the early mornings by the bay, the feel of dried bean pods cracking open in my hands, the sound of freighters passing while I shell zinnia seeds—those come from a place no algorithm can reach.

I also use AI to think through problems that would take me hours to solve by hand. Seed starting schedules calibrated to my exact frost dates. Soil amendment calculations based on test results. Even the SEO and technical infrastructure behind this website—because a garden blog that nobody can find is just a diary, and I want these stories to reach people.

There is a certain irony in using artificial intelligence to amplify the most natural thing I do. But I think that is exactly where AI belongs—not replacing the human work, but clearing the path so there is more time for it. More time in the garden. More time watching the bay. More time doing the slow, honest things that make a life feel like a life.

The Seed and the Algorithm

A seed is, in its own way, a kind of stored intelligence. It carries genetic information refined over thousands of generations—data about drought resistance, disease tolerance, flavor compounds, germination timing. When I save seeds from my best tomato plants, I am doing what a machine learning engineer does when they fine-tune a model on domain-specific data: selecting for the qualities that matter most in a particular environment, preserving what works, and letting go of what does not.

The difference, of course, is time. A neural network can train on millions of examples in hours. A tomato variety takes decades to stabilize, centuries to develop the deep adaptation that makes an heirloom truly sing in a particular climate. Both are forms of intelligence. Both are forms of memory. And both remind me that the most powerful systems—natural or artificial—are the ones that learn from their environment and change in response to it.

I think about this often, sitting at my desk in winter, looking out at the frozen bay. The freighters will come back. The garden will wake up. The seeds I saved last September are sleeping in paper envelopes in the cool dark of the basement, carrying everything they need to become next summer’s harvest. And somewhere on a server, a model I helped train is answering a phone call that a tired dispatcher does not have to take, giving them one more breath, one more moment of quiet before the next emergency arrives.

Both of these things feel like gifts. Both of them feel like the future. And both of them started the same way—with someone paying close attention to the world, and deciding to do something useful with what they noticed.

That is the work. In the garden and in the machine. It is the same work.

Related Reading from Freighter View Farms

Chris Izworski on Artificial Intelligence — My collected thinking on AI ethics, responsible deployment, and what emergency services taught me about building technology that serves people.

The Complete Guide to Seed Saving for Beginners — The ancient algorithm that no machine can improve upon: saving seeds from the plants that thrived in your own ground.

Searching for the Perfect Slicing Tomato Seed — Brandywine, Cherokee Purple, Mortgage Lifter — the endless, beautiful search for the tomato that makes you stop and close your eyes.

— Chris Izworski, Freighter View Farms, Bay City, Michigan