Every gardener knows the $200 tomato.

It goes like this: you decide the grocery store tomatoes taste like nothing. So you buy a raised bed. Soil amendments. Stakes. A cage, because last year the deer got everything. Seeds—heirloom, obviously, because if you're going to do this, you're going to do it right. And by August, you're standing in your backyard holding a single, perfect Cherokee Purple, and if you do the math (which you absolutely should not do), it costs you about $200.

I built a social media carousel generator this month. Not because I couldn't find one, there are SaaS tools for this, I tried them, but I didn't like them. The UX was wrong, the templates were limiting, and the pricing was insulting for what you got. So I did what any reasonable person would do: I opened Claude Code, said "I want to build a carousel editor", and spent the next several days testing and breaking and rebuilding the thing across maybe fifty features until it actually worked.

The first carousel took forever because every time I went to create it, I realized I needed another feature. It took me almost a week to get it right.

The second carousel took only a moment to generate. It was, conservatively, 99% less effort than the first. Sure, I had to fiddle with it. But the bones were right. The layout was right. The text flowed into the frames the way I'd designed it to. Claude was baked into the tool, so design help and recommendations were only a click away.

I didn't build that tool the way a programmer would have ten years ago. I didn't write the code line by line. I described what I wanted, argued with the AI when it got things wrong, and iterated until it worked. I didn't build the app so much as I cultivated it. Anthropic co-founder Chris Olah says it himself: "Generative AI systems are grown more than they are built." Nobody sat down and programmed Claude's abilities into it. They fed it data, and it grew. And my carousel grew the same way, one conversation at a time, shaped by what I wanted and what the tool could do. Less architecture, more gardening.

I sat there for exactly one second with a big grin on my face, enjoying the fact that I didn't have to pay for that SaaS subscription, before I considered the fact that between my Claude subscription and the API calls, it had cost me over $200.

I come from plant people. Both of my parents have skills ranging from mastery in horticulture, farming, national park operations, landscaping, and food preservation. I grew up surrounded by people talking, planning, and engaging with growing things. Their enthusiasm and skill must have some sort of genetic component because I seem to be afflicted. I garden, but for me it's a hobby, and I approach it like one, with fantastical planning and an unrealistic budget. But it feeds my soul.

So when I realized I'd just built a new kind of $200 tomato, I laughed out loud and shook my head in disbelief. Of course. Of course, I spent $200 avoiding a $14 monthly subscription that annoyed me. It seemed like the exact, stubborn, principled stance that I’ve seen people take my whole life. It's the same thing that happens with the garden every spring—you don't set out to spend $200 on tomatoes. You set out to grow a better tomato than the one you can buy, because you know you can, and because something in you won't let that go. The budget is often beside the point. The principle isn't.

And then I stopped laughing, because I realized I'd seen this before. Not the carousel. The economics.

I've Seen This Before

The $200 tomato is a joke, but it's an economics joke. You spend $200 avoiding a $3 tomato because you wanted a better one, and you knew you could grow it yourself. The math doesn't make sense on paper. But the tomato isn't really about the tomato — it's about the raised bed, the soil you amended, the knowledge you built, the fact that next year's tomato costs you $4 instead of $200 because the infrastructure is already in the ground.

That's what I was looking at. I'd accidentally built the same economic structure my parents have been navigating my whole life — just in software instead of soil. The $200 carousel wasn't a bad purchase, as much as it was just the first harvest. And the second one took only one moment and a few cents to generate because the garden was already planted.

But here's where the comparison kept going and wouldn't stop.

I started thinking about who owns the land. In agriculture, that's always been the question underneath every other question. Who owns it? Who profits? Who controls the seeds? Four companies — Bayer-Monsanto, Corteva, Syngenta, and BASF — control over half the global commercial seed market. Farmers technically own their dirt. But they can't save seeds, can't replant without licensing, and in some cases get sued for growing crops that blew in on the wind. They own the land. They don't own what grows in it.

So I looked at my carousel and traced it backwards. Whose servers is it running on? Whose API? Whose cloud? Whose terms of service? And I kept landing in the same place. Three companies — Amazon, Microsoft, Google — control 63% of global cloud infrastructure. My app, my business, my photos, my documents — they run on their land. Their pricing. Their terms. They can change the rent whenever they want, and my option is to pay it or burn the crop and start over somewhere else.

A handful of companies control the seed market. A handful of companies control the cloud. Farmers who own the dirt but not the seeds. Users who own the login but not the infrastructure. The economics are identical — consolidation, dependency, and a single point of failure dressed up as efficiency.

And when the single point fails, it fails everywhere at once. The potato blight killed a million people in Ireland because the entire country was growing the same crop with the same vulnerability. The CrowdStrike outage in July 2024 disabled 8.5 million Windows systems in a single update because the entire internet is running on the same handful of machines.

There's a word for this. It's called monoculture. It's what happens when a whole system depends on a single variety of anything. And it keeps working — right up until the moment it doesn't, and then it fails all at once, everywhere.

And when it fails, it's not just someone's photo library that goes dark. Hospital patient records go offline. When AWS went down in October 2025, it took Slack, banks, government services, and gaming platforms with it — 17 million user reports in a single outage window. The food supply chain, the banking system, emergency services — it all runs on the same handful of machines, owned by the same handful of companies, subject to terms of service that most of us never read and had no hand in shaping. When the crop fails, it's not a field that goes hungry. It's everything connected to it.

People have been drawing these lines for decades. The land question, the ownership question, the who-profits question — none of this is new. But the ground is shifting in a way it wasn't shifting before, because the tools that used to require entire companies to build can suddenly be built by one person in a living room for $200. It's a market disruption, and it impacts everyone, all the way up. The pricing models, the labor assumptions, the power structures. They're all built on the premise that building things that work is hard. Not everyone is going to become a vibe coder overnight, but enough people will that the ones who've been growing our fruit for us have to start paying attention.

Who's the Farmer?

Okay, so if the app is my fruit — who grew it? What's the soil, the water, the sunlight? Are we standing in a garden or a field of monoculture?

When industrial agriculture mechanized, the food still grew. The work still happened. But the people doing it by hand got replaced by machines that were faster and cheaper, and farm employment dropped 73%. The power consolidated upward, because a combine harvester costs a fortune, and the displaced workers who couldn't afford the land they worked on also couldn't afford the machines that replaced them.

Anthropic's own research now shows that 75% of the tasks that make up computer programming are already being performed by AI in real workplaces, and the most exposed workers are older, more educated, and higher-paid. This time, the tractor is coming for the office worker.

So who's the farmer now? Who profits?

The growers at scale are businesses like Amazon, Microsoft, Google, Anthropic, and OpenAI. They own the cloud, the models, the APIs. They're landlords. They sell the soil, they have the seed patents, and the water rights, and they profit whether you grow your own or buy from the store.

The farmers are the professional engineering class. The people who've been building software for thirty years, who built the value that made those landlords rich. They're not going extinct, but they're getting compressed the way family farms got compressed. Fewer of them, doing more, with machines doing what used to take a team. That 73% didn't mean zero farmers. It meant fewer farmers on bigger operations, aided by machinery.

The home growers are the rest of us. The person who opened an AI tool because the SaaS subscription annoyed them, the individuals, and the small organizations. We're not replacing engineers. We don't have what it takes to ship code at scale or maintain a customer base. We're doing what backyard gardeners do: growing what we need, because we can, and because nobody was going to do it for us at a price or quality we could live with.

And then there's everyone else. The consumers, who have no direct hand in cultivating their food sources.

And don’t think this disruption won't affect you because AI isn’t coming for your job any time soon, or that you have no interest in building software. When a large professional class gets compressed, the displaced workers don't vanish. They move sideways. They flood adjacent markets. They compete for jobs they're overqualified for, or make career transitions, which pushes the people already in those jobs further down or out. This isn't just a tech story. Every hiring market that has to absorb the displacement feels it.

And while the workers scramble, the wealth consolidates. Economists are calling it a K-shaped economy: the people who own the infrastructure track upward, and everyone else sinks. AI has already delivered a 7% uplift in household wealth, but almost all of it landed with people who were already rich, because they own the stocks, the platforms, and the cloud that AI runs on. The richest 1% now hold nearly 32% of all wealth, the highest on record. The skills are collapsing, but the capital isn't; it's just consolidating to fewer and fewer people.

Science fiction already showed us what happens when the skill collapses, but the infrastructure doesn't. Star Wars is full of people who can build a droid from scrap parts, and they still live in poverty on a desert planet, because the people who own the ships and the trade routes own everything that matters. That galaxy is crawling with engineers. None of them are rich. The skill is everywhere. The wealth isn't. We've been watching this movie for fifty years. The question is whether we're building Tatooine or something else.

So, what can we do?

Not everyone is going to grow their own tomatoes. Most people don't have the interest or the time, and that's fine. And some of the existing harvest is too massive and too entangled to be replaced easily. Enterprise software running on decades of legacy code. Regulated industries where security and liability aren't suggestions. Financial systems where a bug means someone loses their house. Hospital networks. Air traffic control. You can't replant that with a weekend project and a $20 subscription. Those systems will change, but they won't disappear, and the transition will take years.

But every year, more people will figure out they can grow something. Not everything. Something. A tool that replaces the $50 subscription. A workflow that used to require a contractor. A website that their kids' school actually needs. And every year, the tools get easier, the barrier drops further, and the number of people standing in their backyard holding something they grew themselves gets a little bigger.

So the question isn't whether the grocery store survives. It will. The question is whether the farmers' market exists next to it.

People primarily shop at grocery stores for their daily needs. But a lot of those same people also go to the farmers' market once a month to say hello to their CSA guy. Not to replace the grocery store. Just to have a relationship with someone who grows things for them, someone whose livelihood depends on their trust, who lives in their community and answers to it.

Now imagine that for software.

Your friend group has its own platform. Your family has its own stack, photos, messages, and calendars managed by somebody local. A person in your neighborhood handles the digital infrastructure for ten families the way a CSA farmer manages a plot for twenty. Not a startup. Not a corporation. A neighbor. Someone who shares your water, your school board, and your interest in things actually working.

Imagine this, it's a normal Tuesday. A social media platform just got caught doing something terrible to kids again. But this time, you have the leverage to say: screw you, we have our own. Divesting attention and reducing their ad revenue as a way to force policy change, voting with your dollars and your compute. Hosted by your city, your neighborhood, or your family.

This isn't fantasy. There are already over a million workers in platform cooperatives globally, digital platforms owned by the people who use them and the people who build them. Worker-owned tech support co-ops exist right now in Boston and San Francisco with equal wages and democratic management. These aren't just system theories; they are actual businesses. They're just small, and nobody's heard of them, because the algorithm doesn't promote its own competition.

I think about what it would feel like to look my tech person in the face the same way I look at the farmer at the market. To know who built the thing my kid uses at school. To have a name and a phone number, not a support ticket. And it's the thing that disappears first when everything runs on platforms owned by people who will never meet you.

And if you follow the thread far enough, the endgame isn't just the farmers' market. It's owning the land. Right now, every tool you build still runs on somebody else's infrastructure. But the cost of local compute is dropping. The models are getting smaller. The day is coming when a neighborhood server in somebody's basement runs the local stack the way a community garden runs on a shared plot. Not tomorrow. Not next year. But the trajectory is clear: first, you grow your own fruit, then you grow it on your own soil. That's when the landlord stops getting paid.

The future has to be fought for

I'm not naive about this. I know what happens.

The community builds something that works. Proves it's better. And the corporation tries to make it illegal.

Chattanooga, Tennessee, a city-owned utility, the Electric Power Board, launched gigabit fiber internet in 2010. City-owned. Community-accountable. It worked so well that it generated nearly $3 billion in economic benefits. And then the telecom industry, backed by AT&T and Comcast, got laws passed in sixteen states restricting or banning anyone else from doing the same thing.

John Deere locked farmers out of their own tractors, tractors they bought and paid for, with software that made it so only authorized dealers could run diagnostics or make repairs. You own the machine, but you can't fix it. You own the dirt, but you can't save the seeds.

People have been fighting platform landlords since before we called them that. Fannie Lou Hamer, a sharecropper's daughter who started picking cotton at six, bought 640 acres in the Mississippi Delta and built the Freedom Farm Cooperative — housing, a co-op store, a pig bank — all run by the people it served. It became one of the largest employers in Sunflower County. Her approach was simple: "The time has come now when we are going to have to get what we need ourselves."

Dolores Huerta and César Chávez, children of farmworkers, co-founded the United Farm Workers in 1962 and organized a five-year grape strike and national boycott that forced an entire industry to change. Huerta negotiated the first collective bargaining agreement in American agriculture, secured disability insurance for farmworkers, and was instrumental in passing California's Agricultural Labor Relations Act of 1975 — the first law in the country giving farmworkers the right to organize. Her rallying cry:"¡Sí, se puede!" became the mantra for generations of people who were told the system couldn't change. They changed it anyway. And then the system tried to change it back.

If communities start building real alternatives to big tech platforms like local stacks, cooperative networks, and neighborhood infrastructure, every company with a revenue model built on your dependency will fight it. They'll lobby. They'll buy it. They'll change the rules. This is not cynicism. This is the historical pattern, and it plays out the same way whether you're talking about tomatoes or broadband or the right to repair the things you own.

A local, more equitable digital model will absolutely get targeted by those who benefit from the system staying on its current trajectory. The only thing that matters is whether enough people are already participating that fighting it costs more than it's worth.

Slow and steady wins the race

I don't think this is a revolution story. Nobody's deleting their Google account and going off-grid, and I don't think they should have to. I think it's a long-term goal that needs cultivating, and the first step is the same as any: education.

A totally unscientific opinion: The job market is being propped up by a shared agreement that lots of things are still hard. The manager doesn't know how easy AI makes that PowerPoint. The worker doesn't know either, or if they do, they're keeping it quiet. The consumer doesn't know that the $50/month project management tool could be rebuilt for the cost of a few API calls. Everyone is safe as long as everyone stays ignorant. The skills gap is holding up the building.

But the skills gap is collapsing. And every person who figures that out is a Jenga block being removed from the tower we have all been living in together.

The expensive thing is only expensive because people believe it has to be. When they stop believing that, and they will, because the evidence is getting harder to ignore, the structure has to reconfigure. Not overnight. These things start out expensive. They get cheaper. And then they become normal.

The Dirt Under Your Nails

I don't have a ten-point plan. I have dirt under my nails and a direction.

You do not need to learn to code. You do not need a computer science degree. You do not need to wait for someone with credentials to build what you need. The seed catalog is open. The soil is practically free. I built a production tool this month by talking to an AI and being too stubborn to stop when it didn't work the first time. It cost me $200 and a week of my life, and it was worth every penny, the same way the Cherokee Purple is worth every penny, the same way anything you grow with your own hands is worth it, not because of the math, but because of what you learned.

Ask who owns the land you're standing on, and who owned it before. Every tool you pay for. Every platform you post on. Every service that holds your data. If they changed the terms tomorrow, what would you do? If the answer is "Nothing, because there's nowhere else to go", that's the monoculture. Notice it.

Find the people already growing things. They exist. They've been at it for years, mostly without anyone noticing. Community broadband fighters. Platform cooperatives. Right-to-repair advocates. Tribal nations building digital sovereignty. Small technology foundations are building personal servers. They are out there in the dirt, doing the work, and they could use your business, your attention, and your voice.

I'm not telling you to pull the Jenga tower down. Maybe it gets reinforced. Maybe AI literacy stays low enough for long enough that most people never figure out what's possible, and the tower holds because ignorance is structural. Maybe the K-shaped economy keeps spreading, and none of this matters because the people who own the infrastructure will always own the infrastructure. I don't know. I'm not an economist or a futurist. I'm a person who grew a $200 tomato and a $200 carousel in the same year and noticed they had similar economics. I know the soil a little better than I did last season. I know the second harvest was easier than the first. And I know the Cherokee Purples go in the ground in May, and I've got a list of tools I want to build that's longer than the growing season. That's what you do when you're the kind of person who grows things. You keep growing too.