A couple of days ago, a post popped up on my feed on Bluesky, the social network, that caught my eye and made me want to dig into the arguments.

It was from Peter K. G. Williams, an astrophysicist at the Center for Astrophysics | Harvard & Smithsonian. He was sharing an academic paper he had just posted to arXiv, the open-access scientific archive, and he introduced it with words that carried genuine weight:

Likely the most important thing I've ever written — if you're an academic and the AI mania unsettles you, here's my best shot at presenting some intellectual tools that might help you articulate why.

The paper is called "AI and The Enclosure of the Scientific Commons." As I read through the summary, I could see why it was resonating with so many people, but I also knew right away that I wanted to explore the nuance that was missing. In it, Williams argues that artificial intelligence isn't just another shiny gadget; it’s an engine of enclosure.

If you haven't brushed up on pre-industrial British history lately (and who hasn't? Well, me for one), "enclosure" is shorthand for something very simple and very ugly: wealthy landowners taking the shared village green—the pastures where everyday folks had grazed their sheep and gathered firewood for generations—and putting up ditches, hedges, and barbed-wire fences to declare it private property. It was the moment shared community survival was turned into private landlord profit.

Williams’ argument hits a nerve because it gives voice to the foreboding feeling that a lot of working developers, artists, and writers have been carrying in their chests for the last three years. When people feel unsettled by the AI gold rush, it’s not because they’re backward "Luddites" who hate machines. It’s because they’ve watched the technology pull off a classic pro-wrestling heel turn: it won over the crowd by promising open research and democratized knowledge for all of humanity, and then, the moment it got big enough, it slammed the steel cage door shut and demanded a monthly subscription to see what was inside.

I’m not an academic. I’ve spent my life as a self-taught software engineer and developer educator who worked retail sales and learned how to listen to what people actually need rather than selling them sizzle without the steak. But as someone who learned to build things without a computer science degree, I know what a commons feels like.

Because without one, I wouldn’t have a career. And I’m just one of millions of people who can say the exact same thing.

The Sacred Ground of the Public Library (and the Tool Shed)

When you grow up with few resources (money and mentors, to name two), the public library isn't an abstract economic theory. It is sanctuary.

It is free heat in the winter. It is a quiet table where nobody asks you to buy a coffee to sit down. It is rows of shelves filled with books that would cost more than your rent, handed to you on a plastic card with a simple, beautiful, unwritten agreement: treat this with respect, and bring it back so the next person can learn.

I learned how to code because people left documentation, tutorials, and open-source code out in the open. They left it on forums, on SourceForge (Git was barely getting started when I first turned pro, and SVN was all the rage), and on public wikis. The open web was the biggest public library humanity ever built.

In recent years, many public libraries have started expanding into something even cooler: tool-lending libraries, or community tool sheds.

Imagine a sturdy wooden shed behind the town hall. Neighbors donate their old tools. A retired machinist drops off a heavy-duty socket set. A carpenter donates a miter saw. A plumber leaves pipe wrenches. If your porch step has a squeaky hinge, or your lawnmower won't turn over, you don’t have to drop $600 at the big-box hardware store.

Or take my own garage. When my first electric unicycle suffered a flat tire, I had to take the machine down to its bones—disconnecting the motherboard, unhooking motor phase wires, and splitting the outer casing. I put it back together slightly wrong (I’m still not entirely sure what I did), powered it up, and watched in horror as an angry, 50-pound motorized gyroscope ripped out of my hands and thrashed crazily around the garage, shattering its plastic shell against the walls.

I bought my second unicycle specifically because the engineers designed it with maintenance in mind—changing the tire was a much safer, modular operation. But even on a machine built for repair, maintenance is never frictionless. Right now, that machine is sitting in pieces in my garage because I need to replace a battery pack. I got through dozens of bolts until I hit the very last screw, and the head stripped clean out.

Suddenly, my weekend project wasn't about batteries anymore. It was about learning what a screw extractor is, how reverse-threaded flutes work, and having the patience not to smash the casing in frustration.

A standard household toolbox doesn't come with screw extractors. That’s the exact moment you need a tool-lending library, or a neighbor down the street who has the specialty bit in their pegboard and ten minutes to show you how to use it without ruining the casing.

In his 1974 classic Zen and the Art of Motorcycle Maintenance, Robert Pirsig called that stripped screw the ultimate "gumption trap." It’s the moment the romantic fantasy of modern technology crashes into physical reality. Corporate marketing sells the illusion that technology is effortless, disposable, and frictionless. But real life—and real software—is maintenance. It is greasy hands, patience, and learning how to gently extract a stripped screw without destroying the chassis.

And you can’t do that without a community tool shed. The shed works because of mutual care. The tools belong to everyone, so everyone has an incentive to keep the pegboard organized.

The Laser Scanner and the Factory Across Town

Now, imagine what happens when corporate capital discovers the tool shed.

A commercial landscaping conglomerate doesn't back up a flatbed truck in the middle of the night and steal the physical wrenches. That would be illegal, and besides, physical theft is too clumsy. The tools are still hanging on the pegboard. Anyone walking past can still see them.

Instead, the company brings in an array of high-precision 3D laser scanners and metallurgical sensors. While the neighborhood sleeps, they scan every single tool on the wall. They measure the exact taper of the carpenter’s chisel, the precise tooth count of the pipe wrench, and the hand-polished balance of the machinist’s socket set. They ingest centuries of collective craft and trial-and-error—knowledge refined by generations of working people who shared their discoveries for free.

Then, they take those scans across town, build an automated, robotic assembly factory, patent the resulting manufacturing processes, and launch "SmartYard AI."

A week later, glossy flyers hit every mailbox in the neighborhood: “Why waste your Saturday doing dirty yard work? For just $40 a month, our autonomous robots will trim your hedges, edge your lawn, and tune your engine!”

The original tools are still in the shed, sure. But the economic reward of that collective heritage has been vacuumed up. The corporation captured the value of a gift economy, turned it into a proprietary subscription moat, and didn't leave a single nickel in the donation jar to fix the shed’s leaking roof.

That is the modern enclosure. It’s not breaking in with bolt cutters; it’s capturing the collective intelligence of the commons, commercializing it, and leaving the upkeep to unpaid volunteers.

The 3D-Printed Plastic Wrench

If corporate extraction were the only problem, the community could rally together to defend itself. But modern technology introduced a second, far more chaotic disruption.

Call it the flood of the 3D-printed plastic wrench.

Make no mistake: 3D printing is an incredible technology. When used with intention and care, desktop 3D printers do miraculous things. They let you print custom brackets that save a discontinued washing machine from a landfill. They print bespoke replacement gears and specialized jigs that give everyday tinkerers superpowers. The technology itself is not the villain.

The trouble begins when a mobile app starts awarding gamified badges and karma points for "contributing to the neighborhood tool shed."

Suddenly, dozens of people with cheap desktop 3D printers start churning out hundreds of plastic wrenches, sockets, and pliers overnight using cheap, brittle PLA craft plastic. They dump boxes of them into the tool shed bins every morning.

Here is the insidious part: from five feet away on the pegboard, a 3D-printed plastic socket looks almost identical to forged chrome-vanadium steel. It has the right shape. It has the size markings embossed cleanly on the side. To an automated visual check, it looks like a real tool.

In the software world, this is the flood of AI-generated code submissions (known in programming as "pull requests"), synthetic blog posts, and hallucinated security vulnerability reports. It has pristine formatting, passes every automated syntax spellchecker, and includes comprehensive-looking comments. It looks like senior-level engineering.

Until you put real-world torque on it.

On a freezing Tuesday morning, a neighbor who works an hourly shift rushes into the shed because their car's battery terminal is loose. They grab what looks like a 10mm socket, hook it to a ratchet, and give it a hard pull.

SNAP.

The plastic socket shears in half. The ratchet slips, stripping the bolt head and smashing the neighbor’s knuckles against the sharp steel frame of the engine bay. Not only is the car still dead, but now the neighbor is bleeding, missing their morning shift, and stuck with a stripped fastener.

Meanwhile, back at the shed, the retired machinist who volunteers on Saturday mornings to keep the tools sorted walks in to find fifty plastic wrenches dumped into the forged steel bin. He has to spend his entire weekend hand-testing every single tool under load to figure out which ones are real steel and which ones are brittle fakes.

After three weeks of this, exhausted and demoralized by an endless flood of zero-effort plastic, the volunteer packs his toolbox, shakes his head, and walks away for good.

The marginal cost to extrude that synthetic plastic wrench was fractions of a cent and ten seconds of unattended machine time. But the cost to verify its load-bearing integrity burned forty-five minutes of human care—or shattered a working person's knuckles.

The Master's-Degree Padlock

When the volunteer steward walks away, the neighborhood association panics.

They call an emergency meeting. The steering committee—mostly credentialed folks who don’t work with their hands and buy all their tools new from specialty catalogs—is horrified by the plastic shards littering the lawn.

Their solution is swift, decisive, and catastrophic: they slap a heavy brass padlock on the shed door.

They pass a new neighborhood ordinance: From this day forward, no one may enter the tool shed without presenting a master mechanic’s license, a certified engineering degree, or a verified corporate commercial liability policy.

Notice who this actually hurts.

The big commercial landscaping company across town? They don’t care. They already have their own private warehouses and multi-million-dollar automated factories. They never needed the community shed in the first place.

The people running bot farms to print cheap plastic? They’ll just move on to the next neighborhood shed down the road.

The only person left shivering in the cold is the kid down the street trying to fix a bicycle.

It’s the self-taught tinkerer trying to figure out how an engine works, or the working-class neighbor who just needed an Allen wrench to assemble a baby crib.

In the real world, this is what happens when academia and enterprise software panic over AI slop. Academics retreat behind thousand-dollar journal paywalls. Open-source projects close their public discussion boards and lock their project archives behind invite-only doors. Well-meaning regulators propose licensing frameworks that require millions of dollars in enterprise compliance teams just to publish open-source code.

In their desperate bid to protect the commons from low-effort noise, credentialed gatekeepers end up enclosing the shed just as permanently as the corporations did.

The Rocky Path of New Technology

So where does that leave us? Are we doomed to choose between corporate strip-mining on one side, and elite credentialed lockdowns on the other?

I don't believe we are. But navigating this requires admitting an uncomfortable truth: the path of new technology is always rocky.

New tools don’t arrive smoothly. They arrive like an earthquake. They disrupt comfortable routines, they break fragile social contracts, and they make a colossal mess. Working toward making new technology comfortable and safe is hard, gritty, unglamorous labor. But it is also how society grows beyond its current limitations and learns to solve problems in completely new ways.

We have walked this exact rocky path before.

When I was a young adult, everyone in education and academia lost their minds over Wikipedia.

The immediate institutional reflex was pure prohibition: "Kids will use Wikipedia for homework—how do we catch them and fail them?"

Almost nobody in authority asked the constructive question: "How does this tool challenge our teaching, and how do we guide students to verify what they find?"

We are trapped in the exact same reflex today: "People are using AI—how do we ban it, run plagiarism detectors, and lock the gates?"

We rarely hear the question that matters: "Here is a powerful, brittle tool with real load-bearing limits. How do we teach people to spot those failure points and use it with rigor?"

It was chaotic back then. In the early days of Wikipedia, there was plenty of vandalism, bias, and low-quality nonsense dumped onto pages overnight.

But society didn't padlock the internet. We didn't pass laws saying only tenured PhDs were allowed to write digital reference articles.

Instead, two things happened:

1. The community built living, bottom-up governance. They built edit histories, automated anti-vandalism bots, dispute resolution channels, and that famous, ubiquitous little tag: [citation needed]. They created friction against bad actors without locking the door to honest learners.

2. Society developed a new kind of media literacy. We learned how to use Wikipedia properly. We stopped treating it as an infallible oracle and started using it as a starting map—a launchpad to scroll down to the bottom of the page and find the primary sources, the books, and the research papers that we could read and verify for ourselves.

Today, Wikipedia is recognized as one of the greatest collaborative monuments in human history. It is a cornerstone of the open commons—and ironically, the very foundation of high-quality data that modern AI systems rely on to function.

The Shop Floor Code

We are in the messy, uncomfortable, plastic-wrench phase of artificial intelligence. We are living through the panic.

Generative AI, just like a 3D printer, can be an astonishing tool for a self-taught mind. It can act as a patient, 24/7 tutor that explains an obscure compiler error at 2:00 AM to someone who can’t afford a private coding tutor. It can help bridge the gap for neurodivergent thinkers who struggle with blank-page paralysis. It can help people who didn't grow up with institutional privilege learn in ways that were previously impossible.

To keep that doorway open for the next million self-taught kids, we don't need ivory-tower padlocks or corporate monopolies. We need a simple, grounded Shop Floor Code:

1. If you didn’t torque it, don’t donate it. If you used an AI tool to generate code, text, or a pull request, you are personally on the hook for its load-bearing reality. If you didn’t run the tests, trace the logic, and verify that the wrench won’t snap under load, keep your hands off the submit button. Never make an unpaid volunteer maintainer pay the cognitive cost for your zero-effort curiosity.

2. Call out extraction for what it is. We have to stop accepting the fairy tale that mega-corporations scraping centuries of public generosity without reciprocity is "inevitable progress." When companies build fortunes off the commons, they have a moral and financial obligation to support the human stewards and infrastructure keeping that commons alive.

3. Never lock the door against the learner. Maintainers need better defensive tools—requiring contributors to demonstrate their code runs on a real machine, cryptographic identity signatures to prevent burner accounts, and community vouching systems. But we must fight tooth and nail against credentialist walls, expensive paywalls, and regulatory moats that demand institutional pedigree just to participate.

Technology—whether it's a plastic library card, a desktop 3D printer, or a frontier artificial intelligence model—should give everyday working people more agency, not less.

The goal isn't to burn the machines, and it certainly isn't to padlock the shed. The goal is to keep the tools sharp, the library open, and our own hands firmly on the work.