Brian is already waiting in the loading bay when I pull into the car park. It’s half seven in the morning, and he doesn’t look like he’s slept much. He might be wearing yesterday’s clothes. Who can remember? ‘I told you to get some rest last night,’ I say, manually locking the car door and crossing the blue morning twilight to the warehouse. ‘You’ll look like shit on film.’
‘Can I show you something?’ he asks as I haul myself onto the ramp. He gestures to the parcel leaning against his leg: something boxy covered with black plastic. I turn the warehouse key in the electric box and wait, staring at the parcel, as the door rolls upward. ‘Bring it inside, I guess.’ He picks it up easily and ducks under the rising barrier.
Brian plays the keyboard. He has a huge curly mass of dark hair, but other than that there is little to differentiate him from his bandmates. Out of all of them he’s been the most interested in the project, but it’s a scattered kind of interest. He appears to think almost exclusively in tangents. I am often frustrated by his input, but he’s come up with a few workable suggestions for the video. More than his wall-leaning colleagues, at any rate. But it’s too late for suggestions now.
Inside the warehouse it’s too dark to see, but I can sense the arrangement before me. Like the sitting room floor on Christmas morning before opening the curtains. I slap the light switch, and the florescent bars come noisily awake above us. The great machine is busily arranged before us, primed as we left it last night.
Brian has me hold one end of the parcel while he pulls the bin bag off his side. I pull the bag from my end and the object is revealed. He holds it up for my inspection. He’s built a metre-wide, three-dimensional model of the band’s logo. It resembles a square filled with random-looking smaller squares, like a simplified QR code. He looks to have used cardboard sheets and brass fasteners to put it together. He’s painted it pale pink. It’s actually quite impressive.
‘We can’t use this,’ I say, before he’s had a chance to say anything. ‘It’s too late.’
Brian, to his credit, at least looks prepared for my rejection. ‘It doesn’t matter where we put it,’ he says. ‘I was thinking it could replace the chair just after the halfway point. It would do the same job.’
I picture the chair tipping over, dragging the broom. ‘Absolutely not. It’s not the same weight.’
‘Let’s test it.’
I tell him we’re not testing anything. Not today.
It’s difficult to fully impress on people how precarious the machine is. Even though the whole delight of watching it go is its precariousness. Its absurd overcomplexity. It’s supposed to rattle, it’s suppose to shake. But people imagine that my appearance in a professional capacity brings me some control over the thing. This misunderstanding persists even after spending a week watching me fight with the parts. It’s barely controlled; that’s how it looks and that’s how it is.
From experience I’ve found it’s better to rule the site with complete implacability—not something that’s entirely in my nature. I pretend to be obsessively intransigent, an iron fist creative, whereas really I am just scared. Not scared of failure, because failure is an inevitability in this enterprise. More scared of the scale of the failure. Infinite, boring, expensive, never-ending failure. The machine that can never complete, not because any particular part can’t or won’t work, but because all the maybes conspire to reduce the chance of success to near zero.
Each individual component action is tested dozens of times, then you put them together into larger segments, test the segments. Eventually you line up the segments and test the whole sequence: once, twice, five times. At every stage reactions fail, and must be realigned or reweighted or entirely replaced. The most reliable physical actions will fail when you line them up with two-hundreds others and hope that they fire all in a row. I explain this to Brian as we stand under the warehouse lighting, looking out over two hundred square metres of primed whimsy.
In the end we agree that that Brian’s model will appear in the video, but without direct interaction. He is quite pleased when I offer to put it on the breakfast table behind the paint-filled balloons, where it will be showered with colours when they explode. ‘No guarantees,’ I said. ‘It may not really work, visually. There’s no time to test.’ He nods happily.
My first Rube Goldberg machine was made with office stationary, glass marbles and a few slats of timber. It was not an original idea; I’d been watching similar machines all morning on YouTube. Grainy, shaky videos of garage tabletops. Dominos, golf balls, marbles and string, all lined up to delight with their complexity and their ultimate lack of utility. It appeared to be a popular hobby, but then the Internet makes everything you search for apparently popular.
That morning I was debugging insurance software, but temporary technical difficulties had left me with nothing to do, and I watched dozens of these videos before casting around the empty office and coming up with the makings of a simple apparatus. I laid it out on my office desk, pushing aside my keyboard and monitor to make enough space. The machine was rudimentary, in that it did not move any single marble from one place to another, but instead ran through a series of reactions that each triggered the next, culminating in a butterknife flipping the page of a training manual. A Rube Goldberg machine must always have a function. A task to perform.
I ran it a few times, pausing between to reset the marbles and paperclips to their original positions. Then I filmed the whole sequence on my mobile phone, created a YouTube account, and uploaded it.
The building of such kinetic contraptions seemed to balance some unbalanced part of me, providing a tactile pleasure entirely absent from my working life. Over the following months I continued to create new models in a spare bedroom at home, slowly increasing the complexity of the machine and novelty of the individual reactions. Occasionally I went big, dragging in extra furniture for more variation in height, and once or twice extending through the spare bedroom door into the hallway and beyond. But more typically I preferred to keep the machines small and delightful, cramming as many unexpected interactions into as small an area as possible. The novelty of these components was what kept me interested. How a tea towel might react under weight. How a tethered cup might describe a predictable arc on a table. Each of the notable builders on YouTube was recognisable for their box of tricks, their library of reactions, and I was obsessed with adding new ones to mine. Ones I’d never seen used before. I never envision a machine before starting to build it. It’s at a component level that inspiration strikes.
I filmed every machine and uploaded the footage to YouTube. My channel became steadily more popular, until each new video received hundreds of thousands of views. It’s difficult to describe, but this relative popularity felt entirely inconsequential. It simply bubbled silently behind the computer screen, making no impact at all upon my real life. My coworkers never mentioned it. I was never recognised in the street. My phone never rang. Until suddenly, of course, it did.
My first professional gig was the creation of a machine for a mattress company. The money I demanded seemed to me a ridiculous sum, but they agreed without hesitation, and I took a week off from work to build the machine in their warehouses. The first thing I realised is that the components have to go large. Once you want to include mattresses, everything has to start big or scale up quickly. This has been an enduring feature of client work—small components have been replaced with large components, but the physics is mostly the same. In fact, many of my biggest, most popular contraptions for clients are less complex than my original spare room projects. It’s the spectacle, rather than true complexity, that impresses the viewer.
The mattress company assigned me two warehouse guys, who were more than enough to help move the components into place. I found a professional camerawoman on a freelance forum and hired her. Our working relationship continues to this day. Her name is Fiona. She’s great.
My nerves forced me to keep things relatively simple, and we had the machine pieced together within three days. We filmed on Thursday and, while I’ve never been fully happy with the result, it was a great success. Fiona’s camera and lighting showed my work with a clarity I had never been able to manage myself. The whole sequence was bright, polished, exciting. The video quickly amassed two million views, a good proportion of which was referred from my own channel. The mattress company was ecstatic, but who knows whether it ever sold any mattresses.
From there the jobs came frequently enough and the cheques were large enough to allow me to leave my software job. The clients I work for are typically larger corporations or music groups funded by a record label. Each one has a particular flavour they want brought to their machine, but each one also wants it basically the same. It’s a guaranteed hit. An advertisement most people can’t keep themselves from watching. I make ads, I know that. Sometimes it feels like they’re something else.
The saddest thing about client work is the obliteration of sound. One of my favourite aspects of the early projects was the wood and metal noises as cutlery chimed and snooker balls dropped into buckets. Occasionally I worked a series of moves that was almost musical. For me the aural was a big part of the experience. Most admen have jingles that they wish to play over the footage. A near silent forty-five-second advert would be too strange, they say. It would “interrupt the viewer’s attention”. And for music videos, of course, the song is the reason the machine exists. Often the noises of the machine are entirely absent, or run muted under the score of the piece. There’s little I can do about this. It’s just another thing to distance me from the work.
I spend ten minutes with Brian placing his model delicately in the machine, and by the time we’re done others begin to arrive. There are two-dozen people on set, most of whom we have sourced from “micro-entrepreneur” websites. Cheap labour, essentially. They build up the machine after it has run its course (or failed midway). They re-erect the dominos, reset the flags. Take the marbles and put them back in their slots. Stand up the chairs, wipe away paint, let out the pulleys. Replace the fireworks. The scale of machine we’ve built would take an individual many, many hours to reset. With a few half-trained labourers we can usually reset in less than an hour.
Eventually the other band members walk in. I take them into the warehouse office and spend an hour going over their positions. They are quite receptive today—during previous visits they were almost disinterested. The idea for the music video has been introduced to them by a creative at the record label. But today everything is finally happening, and everyone is engaged.
By mid-morning we’re ready to film. I never expect the first take to be the final one, but Fiona is filming everything today. While she has a small crew managing lights, she operates a single camera herself. The video must be one single take—that goes without saying.
The first run is a dud—a swinging guitar fails to switch on a lamp in the first minute—so we reset and get ready to go again. The sequence begins with the lead singer pulling a euro coin from behind his ear and placing it on a weighing scales. From there the machine continues through its fraught lifespan, the band members moving on cue to appear at key moments in the background. Fiona glides smoothly, holding the self-balancing steadicam with her eye to the glass. The rest of us just hold our breaths.
The components must be recognisable to the viewer. That’s the obvious secret of Heath Robinson contraptions, but one I must explain to many clients. If I were to build a large, overly-complex kinetic machine from anonymous industrial parts, there would be no joy in it. The delight comes from kettles, soup spoons, armchairs, umbrellas. ‘I could build this,’ the viewer must think. ‘This is not industry.’
Two minutes and six seconds into the film and four wooden matts drop simultaneously, exploding four water balloons and showering Brian’s logo with colour. It looks excellent from the balcony, better than I’d hoped. The machine is designed to run for three minutes and eighteen seconds, the length of the studio cut of the band’s next single. The machine can’t be built to run to the second, there are too many jangly variables, but I can approximate to within ten.
I know there’s an endpoint to this kind of work. Audiences get saturated by it, the novelty wears thin. The popularity of such videos has already lasted longer than most Internet lifespans. It’s established now that you can film a production such as mine with minimum cost, minimum skilled labour, and it’s almost guaranteed to go viral (by whatever metric you apply to that word). It’s inoffensive, fun, it demands attention for its full duration. People will share it; the music doesn’t matter, the product doesn’t matter. Eventually everyone will have seen a video like this, a dozen of them, and my escalations in size and spectacle will experience a diminishing return on investment. I almost look forward to this. I can go back to my spare room.
The second take runs smoothly right to the end. The four band members take their position under the corrugated plastic overhang, holding their individual placards by the hinges. Bocce balls fall one by one from the grooves above, knocking the edge of their boards and spinning up the four individual letters to spell the band’s name. Confetti bombs detonate dramatically. Johnathan, the lead guitarist, is holding his letter upside-down. The balcony groans loudly.
We open the doors to dissipate the smoke from the fireworks while we review the footage with Fiona. As the labourers begin to reset I am beckoned into the office by one of the executive producers. An overseer from my true employer. ‘That one was great,’ he says. ‘Really impressive.’
‘Yes, I’m pleased. Except for that final mistake! We should get a clean run this afternoon.’ I want to seem in control. I always feel like I am pretending to be a professional in the presence of these self-assured men.
‘You know, we could fix that,’ he says. ‘Post production.’ He raises his hands in a supplicatory gesture, palms out, fingers flared.
‘Digitally. It would save an awful lot of work. I really liked that take.’
‘The next one will be just as good. It’s a machine.’ I know this probably isn’t true, but I say it anyway.
The door opens and Brian sidles in holding up his model of the logo. It’s more than half covered with thick, dripping paint. ‘What do I do?’ he asks. ‘Can we clean it for the next take? Do I repaint it?’
‘What if we retain the mistake?’ I suggest to the executive. ‘It might be funny. A bit subversive.’
The producer gives it some thought. ‘Can’t. Would undermine the central goal of the whole video. No one knows who these guys are yet.’ Brian’s face takes this blankly. Paint is spattering softly on the grey carpet.
I should be objecting more strongly. Digital intervention is antithetical to the whole artform. But today I just don’t care enough. We call Brian’s bandmates into the office to confer. They shrug blandly. ‘Sure. Let’s get out of here. We can get lunch.’
We gather the taskrabbits together and have them do a false cheer, which they perform with surprising enthusiasm. This is to be edited in over the groan that first greeted the machine’s climax. These videos always end with an unconstrained cheer; it reinforces improbable success to the viewer.
The band and the film crew and the producers leave. Someone starts the labourers clearing the machine components into two large skips out the back of the warehouse. Anything they like, they can keep. The rest is sorted for recycling, or so I’m told.
It saves someone money to be out of the warehouse by the end of today. The machine disappears quickly as I watch from the balcony. Confetti is swept into boxes. Wet mops tackle the water-based paints. After two hours the construction exists only as footage. Fiona will send the raw film to the production company. From there it becomes something else, something beyond my control. It is already beyond my control. A sequence of events both ordered and inherently unreliable. I’m not sure I’ll cheer at the end.