Hello everyone,
The third episode covers the last six months.
I warn you, it’s long!
For those of you just arriving, the first episode on the history of the ultimate machine is here, and the second on makers is here.
January
So I started working on it in January 2024. In the dreariness.
The vague aim was to build a prototype that would demonstrate the living side of the machine, but with a different, unusual shape.
Sunday mechanic
As for the mechanics, there’s no shortage of 3D files. I drew on them to understand the mechanism.
It’s pretty straightforward, you may say, but mechanics isn’t my strong point: during my engineering studies, I remember having to study the plans for a mixer tap for two hours without being able to figure out where the hot water went (on the other hand, I did get ink stains all over the place with my Rotring).
So I fired up the Prusa again and, on January 12, decided to do something I never do: take an online course on Autodesk Fusion. Up until then, I’d been using TinkerCad, but now I felt I had to step up a gear. So I spent about fifteen hours listening to and watching a guy called Nicolasexplainthe various functions of Fusion.
The first results weren’t great, but at least I had some idea of how it was all going to work. And I decided to use just one motor, whereas many boxes have two: one to lift the lid and the other to push back the switch. Initial tests showed me that it was possible to do both with a single motor, while retaining the possibility of creating a language with the opening movements (the living side).
A first attempt at a… strange shape
To escape the traditional box shape, I designed a first prototype that looked, how shall I say… nondescript. I think I’d just figured out how to make revolutionary objects with Fusion.
I painted it yellow, and the arm pink, because there was enough gray outside.
You can see it all in this video.
The first problem with this prototype was that the scotch tape holding the lid inevitably provoked the remark “that’s nice, but are you going to keep the scotch there?”.
A second problem arose from the fact that the wires coming out of the box inevitably provoked the remark “but are there going to be all those wires there?”.
Ah, and I’d also started work on a cardboard version, with an idea in the back of my mind.
February
First job: the electronics (the monster)
Following the yellow prototype, it was time to integrate the electronics into the box.
I used boards that already contained the components, breakboards that we connected as best we could with small cables, making immoderate use of the glue gun, all the while praying that it would work.
Second project: the first prototype
The next step was to make a real box, with a real lid guaranteed to be tapeless, and which would integrate the electronics, motor, switch and arm.
I opted for brass hinges and tried out some skeuomorphic-inspired sounds – a track later abandoned, but by then I was so happy! You can see it all in action in this video.
March
This first version was “good enough”, as we say in the Yonne, but using brass hinges screwed on and then glued with pistocolle didn’t strike me as an elegant solution. After watching dozens and dozens of tutorials, I finally designed a hinge without brass or glue, in pure plastic. Victory!
I’ve also reworked the circuit board, transforming it from a magma of wires into a somewhat clean card. It’s very easy to do (I use EasyEDA) and very inexpensive to have made.
On this board, you’ll find an rp2040 board, an audio amp, a board that converts the battery’s 3.7V to 5V and a board that charges the battery. The whole thing works well enough for a prototype (i.e. badly).
Here’s what it looks like. It’s better, isn’t it?
In any case, all these efforts have finally resulted in a first prototype that can be shown.
This was immediately followed by another, slightly smaller prototype that would fit the bill perfectly. I was finally able to put all my ideas together in a first video, modestly named“Concepts“.
I had made another video with other sound tests, which you can see here.
My lover then suggested a name for the useless box: la machine. I immediately finalized the logo (well, I found a suitable font for it), registered the trademark “la machine” (since refused) and bought the domain name la-machine.fr.
I’ve also had some badges made. If you’re interested, I’m offering some to the first 20 who let me know!
April
Elium Studio
Elium Studio is a product design agency: it designs Withings products, Freeboxes, subway displays, McDonald’s utensils and a host of other things. I’ve known Pierre Garner for a long time. I wanted Elium to work on the design of the box. After consulting his team, Pierre agreed to the project, and in mid-April, I rang them up for the brief.
The workshops
Remember the cardboard prototype? I had an idea: to use it for workshops with children. Every year, I run a workshop at the Arbalète school in Paris (with my son Armand), during which the children build a small electronic object that they can take home.
Last year, it was a little robot made of plastic cups, which lit up and turned its head when touched. This year, they made a useless box.
Thispaper version had to be designed more seriously. Fortunately, I was familiar with the world of paper toys, since I’d already made (with Marc) cardboard objects with electronics inside. It was called reaDIYmate.
We had to make an electronic board, buy the equipment, bring out the Cricut to cut out the cardboard and test the assembly to make sure it could be done in 3 hours. The result: the kids were able to take their useless box home and, for the nerdiest among them (or their brothers, sisters or parents), reprogram it.
A second opportunity also arose in a kindergarten class where the poetess Séverine Daucourt was in residence. The children composed poems which they said aloud, and the recordings were put “in the box”, so that they could be heard again and again.
Oh, and for the older kids, I teach a course at the WSF on physical design, and this year we had a lot of fun making useless boxes.
Software (seriously)
At the end of April, I had lunch (as I often do) with my friend Paul Guyot. For those who don’t know him, he was the CTO of Violet, the company linked to the Nabaztag. He’s done lots of things since then, but it was he who in 2019 wrote the iconic Nabaztag services in Python for the kit TagTagTag. I showed him the prototype and he offered to help me with the software, which, although it looked simple, needed to be approached professionally.
I had gone for a well-known microcontroller (the rp2040), but it consumes a lot of power when it sleeps. It snores a lot, so to speak. And power consumption is an important issue for the box – I doubt people think about recharging their machine every night. Autonomy is a matter of months. So Paul suggested that I get another processor (an ESP32 C3 for the experts), which has the advantage of consuming very little power when it’s not doing anything. The rp2040 is English, the C3 is Chinese. I don’t know what to make of it.
He’s also encouraged me to use Erlang, but that’s something he recommends for every new project. Erlang is a functional language, close to LISP. LISP was used by AI pioneers such as Marvin Minsky, inventor of the useless box. Do you follow? So, we’re off to Erlang, now all I (and you) have to do is learn it.
Paul has also offered to give a lecture on the machine at the Erlang festival in Berlin in October. If you’re in Berlin on October 20, 2024, come and listen to Paul: the name of the talk is “La Machine: The Useless Box reloaded with Erlang and AtomVM“.
May
Part of May was spent shipping TagTagTag cards from the 2024 production batch. Nothing to do with the machine, although sometimes I feel like I’m making a Nabaztag of sorts again: it’s useless, it plays sounds and there’s something moving.
Friends
I still had time to finalize the appeal to friends of the machine: a video and a first website made with the means at hand, which took me an infinite amount of time. By contacting people interested in the Nabaztag, Minimit and Multiplié, I quickly had 300 friends. Maybe including you.
Interesting
I went to London to attend Russell Davis’ fabulous (vaguely annual) Interesting conference. I took the opportunity to seeEnzo Mari ‘s exhibitionat the Design Museum and visit a few stores to spot what useless electronic gizmos were occupying the market. At Selfridges, I found these Bluetooth speakers with a few keyboard keys and a mini-screen, all for £129.95. I thought, with the machine, I might have a chance.
Electronics (seriously)
It was time to get down to business with the electronics. That’s where Philippe Debadier came in. Philippe Debadier is the engineer who designed the Nabaztag’s electronics long ago, and hundreds of other electronic boards since. He agreed to help me and designed the one for the machine.
The schematic, the routing (routing means going from a theoretical electrical diagram to a real-life board) and the prototypes are complete.
Still to be tested. Thanks Philippe!
June
Elium Studio (again)
On June 17, I had an appointment with Elium for a presentation of the initial work. I’m not going to show you every track. You might prefer others, and now is not the time for doubts.
The design is inspired by Memphis. Not Memphis Tennessee, but the Memphis movement. The Memphis group was founded in the 1980s around the Italian designer Ettore Sottsass (there’s a podcast about him here). It’s a happy coincidence because two years ago, I saw the exhibition devoted to him at the Centre Georges Pompidou and bought the catalog entitled “L’objet Magique“.
Here’s a sneak preview of the machine’s design and possible variations.
Elium also produced a working prototype using the electronics from my second prototype. A video can be seen here (it’s a prototype without a battery, for those of you bothered by the USB cable).
We’re now entering the industrialization phase, which involves finding a manufacturer and working out all the details of how to attach the components to the inside of the box (you don’t want to know how it fits, do you?).
As always, there’s the question of where to make them. For the moment, we’re talking to a Norman manufacturer (even though I’m from Brittany) and a Chinese manufacturer (even though I’m from Brittany).
The site (seriously)
May also saw the start of work on the site, the real one – not the friends’ – by Pascale Moise, who, among other things, is the author of Minimit, but who is above all a site developer. She came up with the idea of a site that opens when you press a button.
Paper toy edition
Following the various workshops mentioned above, I refined the cardboard version of the machine to make a very limited series (no more than a dozen objects, hand-crafted) for those who want to have fun right away, but above all to test out the various choreographies of movement and sound to be devised over the coming months.
It’s a do-it-yourself version, based on the same electronic architecture as the final version (but without the battery), and I imagine it will sell for around €100, with a 50% discount on the final version of the machine when it becomes available. If you’d like to know more, or if you’re interested, drop me a line!
Here’s the kit and what the cardboard machine looks like (it could become a collector’s item, who knows!).
And Lego?
Until now, Loïc has been busy finishing his Master’s degree (that’s done), and in June he began the first trials of the Lego version. We spent an afternoon working out how to integrate the motor, arm and switch, which will be done using specific 3D-printed parts. There will be around 300 parts in the kit, including of course the electronics, motor, loudspeaker, battery and… the switch.
Maker Faire
With my friend François Xavier Faucher (FXF), whom many of you know from the fabulous Toyfab blog, we decided to take a stand at Maker Faire Lille and make a giant useless box.
It’s the weekend of November 19/20, and it’s in Lille. Come one, come all!
Here’s a first idea from FXF. He knows how to do things in a big way.
Next
Now that the electronics are complete, the software already has its Github repository (all the code will be open-source), the website is well underway and the design is set, two very big issues remain: machine behavior and financing.
Behaviors
Paul has developed a behavioral description language involving arm choreography and sound sequence launch, but I haven’t had time to work on that yet. I need to have a clearer idea, if only of the artistic direction. At the moment, I’m wavering between the moo box and IRCAM: I think I’m going to have to tighten things up a bit.
There’s a lot of work to be done, because each machine is unique: the idea is to have enough sounds and behaviors so that, at the time of manufacture, we can draw a few hundred at random for each box. So we have to find a way of creating thousands of movements and sounds. This is what I’d like to test with the “Paper toy” edition.
Production financing (a.k.a. Money)
There are two types of cost in the production of electronic objects: investment costs, known in the jargon as NRE (Non Recurring Expenses), and unit costs linked to the manufacture of each machine. In the case of electronic board manufacture, there is very little investment to be made; on the other hand, in the case of plastic parts, if injection methods are used, molds have to be made, the cost of which depends on the design. And the cube design requires more complex tooling than the average.
So there are two options: finance the production myself, or sell pre-sales to cover investment and production costs. For a long time, I thought I could release the product by self-financing the project, as I did with the MiniMit. Marketing would have been really simplified: you buy it and receive it straight away. But when I calculated the investment needed to produce the machine, I realized that it wouldn’t be possible (you always need a little blindness to get started).
Unless a friend here wants to invest or help out, I’m going to go for a pre-sales system via a participatory funding campaign (Kickstarter a priori). I’m pretty familiar with these systems and this will be my sixth campaign!
A crowdfunding campaign means going from 400 friends to 5,000! The key to success is to reach your goal on the very first day, otherwise you’ll get bogged down pretty quickly, and you’ll have to pull out all the stops, which will be a real pain for everyone involved.
How do you go from 400 to 5000 friends? Well, that’s the 1,000 franc question, and if there are any digital marketing experts out there, I’d love to give you a call and have a chat, because if mechanics aren’t my forte, neither is marketing!
☀️
Thank you all for reading this far!
Let me know if you’d like a badge, if you’re interested in paper toys, and don’t hesitate to help me expand our community!
And finally, the soundtrack:
“Oh, I get by with a little help from my friends
Mm, get high with a little help from my friends
Oh, I’m gonna try with a little help from my friends“
With A Little Help From My Friends – The Beatles
Olivier