This week, my inventory of scavenged cardboard tubes increased from 9 to 36, grew 6x in total length (from 13′ to 84′), and 23x in total volume (0.36 cf to 8.5 cf), thanks to a job lot of free tubes my wife happened to find for me. Here’s the current hoard — the 7 large and 20 medium size tubes are from the new find.

And here’s a spreadsheet view:

Why, you ask, have I been building up a hoard of scavenged cardboard tubing for years? Fair question. And the answer is: because the tubes are there to be scavenged.

But I finally know what I am going to do with it. I’m going to build a Tubeworld!

Elon is building hyperloops and tunnels under LA. Mohammad bin Salman is building The Line, an arcology that’s part of his Neom Ozymandias project, so why can’t I have a Tubeworld? Huh?

So what is a Tubeworld?

It all started in 2019, when I received some long, thin items (thin vinyl strips for blocking out light around the gaps of window shades if you must know) in a 7′ long shipping tube. Like a barbarian, I broke it down to recycle, but at the last minute asked Twitter for ideas for things to do with it. It seemed a real shame to toss such nice material.

Unfortunately nobody had very good suggestions so I tossed the tubes, and regretted it for months. This was the tube that got away.

A year later, in 2020, during my last apartment move, I ended up with a few really nice shorter tubes, from used rolls of plastic wrap of the sort movers use to protect furniture. The movers were about to toss them, but I asked to keep them. This time, I was determined to do something with them. This lovely patch of low-entropy Platonic goodness in the waste stream needed a secondary mission in life, now that its primary mission of holding plastic film was complete.

Once again Twitter had no suggestions that truly spoke to me, but I kept the tubes anyway. They’re the C tubes in the spreadsheet above, and in the second row in the array in the picture.

I ordered an achromatic lens off Alibaba to potentially make a telescope out of them, and designed and 3d printed a couple of parts for the project, but never finished it. The lenses and parts are still sitting in a project box, and I’ll probably make at least one telescope just to learn how, but let’s be real: a telescope project is to a good cardboard tube what a parking lot is to a piece of good urban real estate. It’s a placeholder use, but not a calling. If I wanted to make a serious telescope, I’d use PVC. And I already have a better Celestron telescope anyway.

I acquired the C tubes two years ago, but dreams of doing things with them basically stayed on the backburner until now. I did collect some smaller tubes (front row) and a bunch of nice cylindrical boxes (not shown). The smaller tubes are stiffer and in better condition than tubes of that size usually are. Initially I was saving toilet paper and paper towel rolls too, but they tend to be too flimsy, so I stopped, and decided I would only collect stiff, thick-walled, high-quality tubes in good enough condition to last a long time as part of a project.

Then last month, the dormant 3-year-old mission received a sudden boost: I found a really nice 6’x1.5″ tube (the long one in the picture, E in the spreadsheet) in the trash/recycling room of my apartment building.

Once again, Twitter didn’t have particularly great ideas to offer. I mean, kaleidoscopes and didgeridoos are nice curiosities to make, but let’s face it — like cardboard telescopes, these are uses but not callings. And while artwork based on cardboard tubing is awesome, as is cardboard tube furniture, I am not really an artist or furniture-maker temperamentally. I wanted to do something more engineering-y with the tubes. Something that provides a good excuse to involve 3d printers and Arduinos. Something that brings out the essential tubiness of tubes.

But now, with this unexpected bounty, I’m clearly way outside the regime of one-off random art/design projects. I now probably have the largest amateur collection of cardboard tubing (amateur as in, people who don’t need them for professional uses, like mailing posters and other Long Things or holding Rolled Sheet Things).

I need a solid design meta-idea for 84 feet/8.5 cubic feet of cardboard tubing, in 36 discrete pieces. And no mix of cardboard telescopes, kaleidoscopes, and didgeridoos qualifies as a meta-idea.

As an owner of cats rather than hamsters, a major obvious use — a habitat for small pet rodents — is out.

Which brings me to Tubeworld. But before I tell you what that is, let’s talk about my lab for a bit.


As I wrote a while back, I have been building up a home workshop/lab since the start of the pandemic. The marquee project of the lab is a rover, but the lab itself has a lot more generic capability.

For a while now, though, I’ve felt there was something missing in my attempts to turn Maker. A need that, as you will see, Tubeworld fills.

See, the problem with being a generic Maker of whatever skill level is a bit like the problem with being a prepper. You learn a shit ton of stuff and acquire a great many tools and scavenged materials, but it’s like a never-ending expansion of a capability tree with no obvious end in sight. And there is no natural stream of payoffs, either big or small, unless you do something non-random with it all.

Sometimes I get the sense, reading prepper material, that they almost wish for a zombie apocalypse so all their accumulated prepping pays off (if you can call it that).

Being a Maker is not quite so bad, but it’s similar. I’ve used my Dremel for a total of perhaps an hour in 2 years, using just 2 of the dozens of attachments it has, and part of me itches to do more with it every time I see it hanging there on my pegboard. And this pattern is all over. About 90% of the capability of a general-purpose lab or workshop just kinda sits there 90% of the time. Unused potential yearning for actualization.

Sometimes I do Random Acts of 3d Printing just to use the printer more.

While complex marquee projects like my rover can use a lot of capabilities, they are too focused and integrated. Everything converges on a very specific, tightly integrated thing that works itself more and more tightly into a corner of its design space over time. My rover is now far enough along that only very specific skills, solutions to problems, or materials can advance it (which is perhaps why it has been stalled for months; more on that later). It requires maybe a third of the capabilities latent in my workshop, and most of those capabilities will be used for only minutes to hours over years.

One obvious solution is to have many projects in a lab, which is what actual full-time professional engineers and scientists do in labs. But that’s not really feasible for amateur tinkerers. I now have a dozen project boxes, each reflecting a burst of weekend enthusiasm, and none of them, besides the rover, has progressed beyond a weekend or two of work.

What an amateur makerspace needs is not just a portfolio of projects, but a program. In the sense of the NASA Mars Rover program. A collection of tiny to large projects that interact in interesting and opinionated ways and compound towards an overall larger idea; one that is relatively inexhaustible.

Unlike a single marquee project, a program is not so tightly integrated and convergent that its needs and options are very limited.

Unlike a mere portfolio of unrelated projects, a program is not so random and uncorrelated that you get unbridled divergence that keeps up with an expanding capability tree through sheer time and energy.

A program is to a lab what a theme is to a theme park. Without a larger governing program, a lab is more like an amusement park. I highly recommend Sarah Perry’s excellent old article on this point, Frontierland.

So the Ribbonfarm Lab is acquiring a theme: TUBES!


Tubeworld will be a miniature theme park (potentially also mirrored in VR and distributed in metaspace) built primarily out of scavenged cardboard tubing, where I’m a purist on scavenged, but not on cardboard.

The goal is to build a scale representation of a much larger tube-based arcology (a portmanteau of architecture and ecology coined by Paolo Soleri) to explore interesting questions relating to tube-based habitats and structures. On earth, underwater, in space, or on other planets. My interest in such spaces goes back to the time, a decade ago, when I explored the storm drains under Las Vegas with a friend. That’s a real live tube-based arcology that exists today, complete with a community of homeless (or tube-homed?) people. There’s some great graffiti down there, but also dangerous drug addicts who might attack you. Matt O’Brien, whom I met while I lived in Vegas, has a great book about it, Beneath the Neon,.

Who knows. If the climate gets much hotter, maybe we’ll all be living underground in controlled tubeworlds. If you’ve ever explored the tunnels of a very underground-tunnels type city, like Singapore, Seoul, or London, you’ll get the appeal of such spaces.

Tubeworld will be about exploring such spaces with a speculative eye.

If for example, my 2″ OD H tubes in the spreadsheet at the top of this post represent the smallest tube an adult human could crawl through (like the Jeffries tubes on Star Trek), that would set the scale at 1″ = 2′ (1 inch = 2 feet), or 1:24, and my current inventory of tubing could build a model of a habitat of about 2000 feet in total length. My biggest A tube would be about 12.5′ in diameter, and 63′ long, which would make a rather nice large tubular room, assuming a flat floor set about 2′ up. The B tubes would be 8.25′ OD, which would go down to say 6.25′ if you put in a floor — a nice bedroom or storage space.

So one way to realize a Tubeworld vision with what I have would be as a model of a habitat for about 10 -30 people. I can imagine, for example, a “work” island made out of the A tubes, and an island of storage/sleeping areas made out of the B tubes, connected by the E tube. The basic skeleton would be something like this:

Imagine this structure crafted out of the real tubes in the first photo, with miniature electrical and HVAC systems going, a little transportation system, lighting, climate control, air locks, little storage spaces, mini artworks on the curved walls, maybe some windows, etc. etc.

Imagine navigating a small, remote-controlled camera on a cart through this structure. Imagine the structure in space, or underwater. Or as a system of tunnels excavated on Mars. Imagine shrinking down and navigating this structure yourself, in a VR rendition in the metaverse.

You could explore aesthetic directions (mini models of a realistic setup, like a model railroad) or more systems engineering directions, which is what interests me: questions about distributing power and air, moving things, organizing and access-controlling space, managing the whole thing with a Raspberry Pi (“tribble infestation in Tube A5; seal it off!”), maybe a miniature hyperloop in tube E, why not…

So this model would be something like a 3’x10′ model of a 72’x240′ structure. It would occupy most of the available space in the second bedroom of our apartment, which I currently use as a home office/lab.

But Tubeworld as I imagine it is not a specific structure like this. It would be the space of such structures that could be created with the right set of modular designs for (physical and/or virtual) interconnection and interiors. The idea is to do something with the tubes that turns them into tubular lego-like components that can be assembled into many such structures, with the potential of adding more tubes. Except we don’t just want mechanical connections as in legos, but electrical, mass transport, and compute-fabric connections too. Tubeworld structures must be smart. At any given time, Tubeworld would be in a given configuration, but it could be reconfigured in a variety of ways.

There are of course a ton of interesting questions, starting with the one my wife is sure to ask me: where the hell am I going to put it?

One solution is of course to acquire a mansion with room for this project. Another would be to find a friendly patron or museum with space to host this project somewhere near me in Los Angeles, or wherever I end up long term (once the basic thing is set up and accessible over the internet, I shouldn’t need to be near it physically to work on it too often). For the moment, this is an unwieldy project, but still of manageable size. Broken down, it would fit into a few large boxes for shipping, or in the back seat and trunk of our car. But I can see it getting unmanageably large if I let it. This is one reason it is interesting to imagine Tubeworld as globally connected in a single structure only in VR, but existing in meatspace as a set of smaller nodes connected by virtual links.

But I’ll cross that bridge traverse that tube when we come to it. Until then, I’ll focus on designing and building modules that can be connected and disassembled easily.


I’m still getting my thinking organized around this project, in a dedicated Roam graph that I’ve decided to make publicly readable. So this project, which will be documented in this new blogchain, will be a working-in-public project.

You can check out the Roam graph here.

If you’d like to contribute to this project in some way, stay tuned and I’ll invite collaborators once I get the basic ideas more fleshed out. Anybody who can scavenge (no buying) a single nice tube, and has an idea for doing something with it consistent with the Rules™ (wip) of Tubeworld, will be eligible to join, so start looking for a tube and ideas. If you have a good tube-shaped idea and live in/near LA, I am happy to provide you with a seed tube from my stash to start your own Tubeworld node. Yes, I’m imagining a virtually interconnected community network of tubeworld nodes, minimum node size 1 tube. Why not.

Why tubes in particular? As opposed to say boxes, spheres, etc? Here is the list of reasons I have so far in my page for Why Tubes:

  • Stronger and less failure prone than cuboids
  • Versatile across many architecture patterns
  • Easy to transport underwater, to space, to other planets
  • Can be manufactured relatively easily out of many materials
  • Easier than domes (1d curvature instead of 2d)
  • Easy to create highly controlled environments within
  • Much higher ratio of usable space to material than spheres

What are some interesting things you could build around the basic architectural constraint of working with the interiors of a set of tubes as the medium? Here is my initial list of brainstormed ideas at component level (most of these would go inside the tubes; Tubeworld is an interior world).

  • 3d printable components
    • Section divider wall
    • Sliding door
    • Support struts
    • Rings for hanging/mounting
  • Make from sheet or bar material
    • Floor boards
    • Metal rail for power and transport
  • Electrical
    • Photo-interrupter controlled solenoid door
    • Electrical power bus
    • Comms signal bus
    • Pressure/temperature sensors
    • System of LED lights
    • System of small cameras
  • Mass transport
    • Duct fans
    • Fluid transport tube with pump
    • Tiny rail-powered cart
  • Compute
    • Arduino nano for each tube
    • System level RPi

You can explore other pages on the Roam graph for my initial thoughts on architectural principles, system design ideas, sources of inspiration, etc.


So where do we go from here? The nice thing about having a largish theme mission like this, with clear artistic/engineering constraints, but without a tight design, is that it’s easy to dream up lots of little things you can work on steadily and discursively. Since there is nothing forcing convergence, it should be easier to work on.

So for the first week, I might simply try to get CAD drawings of all the tubes into an OnShape project file. Or design and print a partition, or door, or connector thingie. Or work up the specs for a consistent “flooring” standard to allow for transport conduits to be routed across tubes. Or just draw a bunch of concept sketches like the one in this post. I’ll be aiming to do just 1 small thing per week, taking perhaps an hour.

I have no fixed time table or fixed end-state for this project, or even a vision of any mature state. It’ll be tube-y, that’s all I know.

I’ll share whatever I do on this blogchain, as a mix of reports on new elements, and a changelog.

And you can of course, explore the Roam graph at any time for the current state.

Get Ribbonfarm in your inbox

Get new post updates by email

New post updates are sent out once a week

About Venkatesh Rao

Venkat is the founder and editor-in-chief of ribbonfarm. Follow him on Twitter


  1. Big transport tubes will be a fine thing, but imagine a city wide network of vacuum tubes like they once had in department stores. Left your phone, meds or lunch at home? The family can have it shipped to you within 20 minutes over the intertubes.

    A pipe-dream, of course.

  2. Henry Dotson says

    So a tubeworld is a set of physical tropes that can be combined to form different physical narratives. That would make Tubeworld the TVTropes of the physical space. But the innovation is not “finding” tropes in the wild, rather crafting them using the inherent capacities of the concept-space (the workshop). It’s not even about advancement vs. refinement. It’s bottom-up vs. top-down, where TVTropes is bottom-up advancement based on available materials, and Tubeworld is top-down advancement based on available production capabilities.

  3. Have you got a pet hedgehog by any chance?

  4. Tube uses other than conduits, guns, etc:

    Space station.

    Bicycle frame

  5. Brilliant.

    I too am a collector — in my case, of empty cardboard boxes.

    The most interesting application of which may have been the “brainmaster 2000”, a supercomputer powered entirely by children’s imaginations.