There's a man in Hackney who dug through his own floor for forty years and kept going after the council moved him to the top of the building specifically to stop him.
And a man in Colorado who took rocks out of a national forest for fifty years. Both of those are real.
Daniel wants the full taxonomy after the Tunnel Girl episode. Two categories, he says. The bizarre, pursued with dedication that makes no sense on paper, and the personal projects that quietly grew into infrastructure-scale works. And underneath both, a question he actually cares about. What is it about these people that lets them do what the rest of us would write off as unreasonable?
Two categories, and the through-line connecting them is almost identical. Hermits. No training. No permits. No blueprints.
So we've got two categories, a handful of dead builders, and one very live question about what's wrong with these people. Or right with them.
Let's start with the eccentrics, because Furze is the archetype. Colin Furze, Stamford, Lincolnshire, former plumber. He asked his YouTube audience what vehicle he should build next. Fourteen thousand suggestions came in.
And about fifteen thousand of them wanted a hoverbike.
Roughly fifteen thousand pushed for a hoverbike, yes. No engineering qualifications. No flying experience. None. He built it anyway in 2016, and The Verge ran the headline. Man builds working homemade hoverbike, doesn't die. That's the whole story in six words.
I want to sit with that for a second.
With the hoverbike or with the headline?
The headline. The bar for success was set at "doesn't die." That tells you everything about the risk envelope these people operate inside.
He also built an apocalypse bunker in his back garden. Ten feet deep, sixteen by twenty, accessed through a hidden hatch in his shed. Then he spent three years digging a tunnel from his workshop to his house to the bunker. Secretly, at least in the videos. And then he retroactively applied for planning permission.
After the tunnel was done.
After. South Kesteven council, reference S16-0017.
So the sequence is build first, ask later, which we're going to see again in every single case today.
Every one. The interesting thing about Furze is the performance layer over the engineering. He plays it up, on camera, "we have to dig stealthily so the neighbors don't know," while posting all of it to a channel with over ten million subscribers. There were people on Hacker News quietly pointing out that the stuff looks off the cuff and spontaneous because that's his bit, but there is almost certainly planning and actual engineer review going on beforehand.
Is that confirmed anywhere, or is that people guessing?
Guessing. Widely repeated guessing. There's no primary source where Furze says "yes, I have engineers on payroll reviewing this." The retroactive permit is real. The rest is inference by people who know what the work actually looks like.
Which itself is telling, because a normal person watching Furze basically cannot tell the difference between real recklessness and choreographed recklessness. Both look like a man in a shed doing something that shouldn't work.
The exact same ambiguity shows up in the Tunnel Girl case. Kala, in Herndon, Virginia, software engineer. Started a "storm shelter" in her basement in August 2022. She didn't just dig. She read the FEMA shelter guidelines. She read Rock Mass Classification, which is a real civil engineering textbook. And then she posted everything to TikTok, at first mostly for other engineers and electricians and mining people who could catch her mistakes.
So she crowdsourced her own safety review.
More or less. By November 2023 she had about three hundred and thirty thousand followers and had spent somewhere around fifty thousand dollars. By the end of the year, over five hundred thousand. Inspected and declared unsafe December seventh, twenty twenty-three. Stop-work order December thirty-first. And then she fought it and won. She had permits by March 2025. The tunnel is twenty-two feet below ground now, thirty-foot entrance.
And she's been on the show's radar before, obviously.
Right, but the point for today is what she represents. She's the case study for "this looks insane until you realize the person did the reading." The people who watched her early were engineers, not thrill-seekers. They were watching to see if she'd get the support right.
But then there's William Lyttle, who's the sharpest case, because he doesn't fit that pattern at all.
Lyttle is the one to hold onto. The Mole Man of Hackney. He inherited a twenty-room house in London. He dug a wine cellar in the mid-sixties, and by his own account found a taste for the thing, and then he kept digging for roughly forty years. Multi-level tunnel network under the house, up to eighteen metres long, a hundred and eighty feet if you want feet, reaching the water table.
Eighteen metres. That's not a cellar. That's a small underground complex.
And the part that has stayed with me all week. The council eventually rehoused him into a top-floor apartment, to discourage tunnelling. That's the phrasing in the reporting, "to discourage tunnelling." And he knocked a hole in a dividing wall anyway.
From a top floor. Herman, that's not a hobby. That's a compulsion that survived the removal of the ambient dirt.
Which is exactly the question. Where does dedication end and pathology begin. There's a line from that Hacker News thread that's the spine of this episode for me. It is hard to know when someone is following a mad vision or just mad.
And the answer is, on the evidence, sometimes both, and the two are not mutually exclusive.
Not at all. Rodia passed the engineering test with a hundred percent margin.
Hold on. You're jumping ahead to Category Two.
I know.
You always do this. You get within reach of the towers and the next thing I know we're talking about foundations before we've done the lesser-knowns.
I'll come back to the towers. There are three others from that thread that aren't well known and deserve a line each. Dashrath Manjhi, in India, cut a hundred and ten metre tunnel through a hill with a hammer and chisel, roughly twenty years of work, after his wife died. Baldasare Forestiere dug a catacomb-style underground home in Fresno because it was cooler down there. And Fred Dibnah, the British steeplejack, tried to build a replica coal mine with a thirty metre shaft in his back garden using period techniques, and died before he finished it.
How confident are you in those three?
Moderate. They're all from the same thread and I couldn't independently verify every detail. The Forestiere house is real, the Dibnah mine is real, Manjhi is well documented, but the specifics in that thread I'm holding loosely.
I'll take that with a grain of salt.
Two grains. But you see what Lyttle and Furze and Kala have in common, and what makes the whole taxonomy uncomfortable. They're all aware, at some level, that what they're doing looks unreasonable. Furze plays it up. Kala documents it in real time. Lyttle, though, doesn't play it up at all, because in his own words he's just a man who loves to dig.
And he says something even better than that. There is great beauty in inventing things that serve no purpose.
That's the line. That's the line that answers Daniel's question about questionable practical value head on, and it doesn't do it by defending the projects. It does it by rejecting the premise that usefulness is the thing being measured.
Which brings us to Category Two, where the usefulness question stops being irrelevant because the scale stops being backyard-sized.
Which is where the towers live, and I've been very patient.
You've been barely patient. Let's do it. Watts Towers. Seventeen interconnected sculptural towers. Simon Rodia, Italian immigrant, tile mason. Thirty-three years, nineteen twenty-one to nineteen fifty-four, working alone. Tallest tower just shy of a hundred feet. And the foundation is fourteen inches deep.
Fourteen inches.
That's the detail that makes engineers either lean forward or cover their eyes. He built scrap rebar bent against railroad tracks into frames, wrapped them in wire mesh, troweled his own concrete mix over the top, and pressed in whatever he found. Seven Up bottles, Squirt, Canada Dry. Broken glass. He walked nearly twenty miles along the Pacific Electric right-of-way scavenging material.
So the man was, functionally, walking a rail corridor collecting trash to build a cathedral out of.
And then, in nineteen fifty-five, he gave the property away and left. The city condemned the towers. They were going to be demolished, and a nineteen fifty-nine engineering test found they could withstand lateral forces up to ten thousand pounds, which is why they aren't in a landfill right now. Bought for preservation the same year for two thousand dollars. National Historic Landmark in nineteen ninety. LACMA finished a restoration in twenty twenty-two.
Simon Rodia, a man with no formal architecture training, built a structure that passed a load test that structural engineers designed, on a fourteen-inch footing, and the testers couldn't believe it.
That's the truest sentence you've said all week.
Justo Gallego.
Gallego is the case that pushes even harder. Former monk, Mejorada del Campo in Spain. He started a cathedral in nineteen sixty-one, in thanksgiving for recovering from tuberculosis, and worked on it for nearly sixty years. Mostly alone. Recycled material, junk, salvaged bits, no architectural training, no permit, no Church endorsement, no funding. He designed it, and this is the part I keep rubbing my eyes over, entirely in his head, without a single measurement or calculation on paper.
What do the locals call it?
La catedral del loco. The cathedral of the madman.
They're not wrong, and it's also not remotely the point, which is why I like it.
He died in twenty twenty-one at ninety-six, before completing it.
Jim Bishop.
Bishop Castle, Rye, Colorado. Jim Bishop bought two and a half acres at fifteen years old in nineteen fifty-nine. Four hundred and fifty dollars down, twelve hundred and fifty total, saved from mowing lawns and a paper route. What started as a one-room cabin became a hundred and sixty foot castle, built over about fifty years, from a thousand-plus tons of rock, much of it taken from San Isabel National Forest. No blueprints.
No blueprints, no insurance, rocks from a national park.
And this is the one where the two categories collapse. The whole thing is a personal project but it's the size and shape of infrastructure. It's billed as the world's largest one-man project. He said "I just build. I don't measure." And then the line I want printed on something, "what's real neat about not having blueprints is that if you make a mistake, you call it art."
That's the most honest summary of outsider engineering I've ever heard. Every design decision becomes retroactively correct because there was never a spec to violate.
A thirty-foot steeple added nineteen ninety-five, because you can add a steeple.
Why not. He died November twenty-first, twenty twenty-four, at eighty, after nearly five decades. And the part I actually respect, he legally barred anyone from ever charging admission.
Every one of these builders has the same operating sequence.
Build first. Ask later, if ever.
Gallego built for sixty years with no permit. Bishop built a castle-sized structure with rocks from public land and no insurance. Rodia left the property partly to escape the permit fights, and it's worth remembering that he was also fighting the city over the towers specifically. Kala fought City Hall and won, and won a permit afterwards.
There's a specific angle on Rodia's exit that Hilbert has been sitting on for the entire episode.
Corn.
I'm not saying anything. I'm just noting.
Let me get to the ethics of this before we get sidetracked, because there's a counterargument that the romantic reading leaves out. There was someone in that thread saying, "I think it is bad because it promotes the idea other people can do the same. Which is extremely dangerous."
Correct.
You're agreeing with it?
Of course I'm agreeing with it. Look at the list. Lyttle undermined a London house, and had to be rehoused. Kala had a stop-work order. Bishop took rocks out of a national forest without so much as a permit. If a hundred people tried to do what any of these four did, ninety-nine of them would be in a lawsuit or a hospital.
And we only find the ninety-nine interesting when they're funny. The one who doesn't kill anyone and passes the load test is a genius. The one who drops a house on his neighbor is a cautionary tale. Same impulse.
The line between mad vision and just mad isn't philosophical. It's statistical. It's outcome.
Which is exactly why I don't love the romantic reading either. These projects are only charming because we're reading about the results. Nobody writes a blog post about the guy who dug twelve feet and hit a gas main.
To be fair, the guy who hit a gas main couldn't have written it. He's not around.
That's dark.
It's the actual shape of the argument. The filter is survivorship. There's a version of this story we never hear, and it's the majority.
Agreed, but the mix was wrong. The Rodia mix was wrong.
The restoration. LACMA, twenty twenty-two. The original towers had a mortar that rang like a bell when you tapped into it. The restorers used a harder mix and the tapping test doesn't ring anymore. My uncle worked the sixties restoration. He told me that when I was a kid, and I've never heard anybody else say it, and I believe him.
Hilbert, is "rang like a bell" a tonal thing or a structural thing? Because if the original render had a different aggregate size, you'd get a different acoustic response from the same load behavior. Load capacity doesn't drop just because the tap tone went flat.
I don't know. He said bell, I'm telling you bell. I don't have the metallurgy on that one.
Fair.
The other thing about the sixties restoration. The crew that worked on it was paid partly in scrap from the towers themselves. My uncle took about a dinner plate's worth of the original steel and glass home. He had it for forty years. It went into a safety deposit box when he got older.
A dinner plate doesn't need a safety deposit box.
It's too heavy to keep in the flat.
It's dinner-plate sized but too heavy to keep in the flat.
That's what I said. And the box is at a bank my uncle also robbed. Different decade, but same building.
You don't want to explain the second half of that sentence.
No.
Fine. The tapping test is a good line.
It is. But also that "Rodia gave the property away" is not the whole story. The neighbor and the fence is the story, and the family never got an answer about it, and I'm not going to be talking about it further. I've got a level issue on the second microphone, it's about three DB hot on Herman's side, I'll sort it in the mix.
I noticed that.
It's sorted.
So back to Daniel's question. What is it about these people that allows them to pursue undertakings the rest of us would write off? And I want to resist the clean answer, because I don't think there is one.
The pattern in all the cases is the same sequence, not the same personality. No blueprints. No permits. No formal training. Build first, apologize later, and hope the result is interesting enough that nobody tears it down.
Which means the "what is it about them" answer is almost boring. There's nothing exotic in the psychology. It's a willingness to start without permission and keep going without feedback.
The 1959 test is what saved the towers. Not the vision. Not the thirty-three years. An engineer with a hydraulic jack and a clipboard. The vision was sufficient to build it, but it wasn't sufficient to defend it.
That's true of all of them. Bishop's castle survived because he put it in his will. Kala's tunnel survived because she fought the planning authority. Gallego's cathedral survived because the town loved it enough after the fact to want it preserved. The eccentricity gets the work built. The paperwork is what keeps it.
Which raises the question about Furze specifically. Is the eccentricity now a genre? Is it a performance category, the way reality TV is?
Yes, and that's a real shift. Furze is a former plumber who has declared over a million pounds in one company in a single year. That's not an odd man in a shed. That's a production.
He grew up on a channel that was just him digging. Now there's a whole generation of builders who grew up on Furze, which means their first instinct is to ask not "will this work" but "will this be fun to film."
Then the interesting question is what that generation builds that isn't just content. The Furze impulse is useful when it's aimed at something real and wasted when it's aimed at a view count. The Filter doesn't distinguish.
Lyttle's line again. There is great beauty in inventing things that serve no purpose.
He wrote that about his own life and he means it.
The Mole Man spent forty years confirming it.
One thing to take from today. Not a fix, not a course of action. Just the shape of the thing. Every person we talked about today made the same decision at the same moment. They started building before anyone had authorized the building.
And ninety-nine out of a hundred people who make that decision end up in court, in a hospital, or with a gas main. The filter isn't about them. It's about us, and which results we decide to read about.
Hilbert, thanks for producing this one.
Yeah, thanks Hilbert.
Send us your own prompt on Telegram at t dot me slash MWP listener bot. We're at my weird prompts dot com. This has been My Weird Prompts.
We'll be back soon.
