Hello everyone,

 

The third episode is dedicated to the last six months.

 

I warn you, it's long!

 

For those who have just arrived, the first episode on the story of the ultimate machine is here and the second dedicated to the makers is there.

 

January

 

So I started working on it in January 2024. In the gray weather.

 

With the vague goal of building a prototype to show the living side of la machine, but with a different, unusual form.

 

Mechanic (Sunday)

 

Regarding mechanics, there is no shortage of 3D files. I used them to understand the mechanism.

 

It's pretty simple, you might say, but mechanics is not my strong suit: during my engineering studies, I remember having to study the plans of a mixer tap for two hours without being able to understand where the hot water was going (however, I had ink stains everywhere with my Rotring).

 

So I restarted the Prusa and, on January 12, I decided to do what I never do: take an online course on Autodesk Fusion. Until now I used TinkerCad, but I felt that I needed to move up a gear. So I spent about fifteen hours listening and watching a certain Nicolas explain the different functions of Fusion.

 

The first results were not glorious, but at least I understood a bit how it was all going to work. And I decided to use only one motor, whereas many boxes have two: one to lift the lid and the other to push the switch. The first tests showed me that it was possible to do both with one motor, while keeping the possibility of creating a language with the opening movements (the living side).

A first attempt at a... strange shape

 

To escape the traditional shape of the box, I designed a first prototype with an appearance, how to say... indefinable. I think I had just figured out how to make objects in revolution with Fusion.

 

I painted it yellow, and the arm pink, because there was enough gray outside. It's all in images in this video .

The first problem with this prototype was that the tape holding the lid inevitably led to the remark "it's nice, but are you going to keep the tape there?".

 

Then came a second problem due to the fact that the wires coming out of the box inevitably led to the remark "but are you going to have all these wires there?".

 

Oh, and I had also started working on a cardboard version , with an idea in mind.

February

 

First task: electronics (the monster)

 

Following the yellow prototype, it was necessary to integrate the electronics into the box. I used boards that already contain the components, breakboards that are connected as best as possible with small cables, making excessive use of the hot glue gun, while praying that it would work.

Second task: the first prototype

 

Then it was necessary to create a real box, with a real lid guaranteed without tape, and which would allow integrating the electronics, the motor, the switch, and the arm.

 

I opted for brass hinges and tested skeuomorphic-inspired sounds – a path later abandoned, but at that moment, I was so happy! You can see it all in action in this video .

March

 

This first version was "good enough" as they say in Yonne, but using brass hinges screwed and then glued with a hot glue gun did not seem like an elegant solution to me. After watching dozens of tutorials dozens of times, I finally designed a hinge without brass and without glue, made of pure plastic. Victory.

 

I also reworked the electronic board to turn a mess of wires into a somewhat neat board of boards. It's very easy to make (I use EasyEDA ) and very cheap to have manufactured.

 

On this board, there is a board with an rp2040, an audio amplifier, a board that converts the 3.7V from the battery to 5V, and a board that allows charging the battery. The whole thing works well enough for a prototype (which means badly).

 

Here's what it looks like. It's better, isn't it?

In any case, all these efforts finally led to a first prototype that could be shown.

Followed immediately by another slightly smaller prototype that was perfect. I was finally able to gather all my ideas into a first video modestly titled "Concepts".

I made another video with other sound tests that you can see here .

 

My sweetheart 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 brand "la machine" (since refused) and bought the domain name la-machine.fr .

 

I also had badges made. If you're interested, I'm giving them away to the first 20 people who let me know!

April

 

Elium Studio

 

Elium Studio is a product design agency: they design Withings products, Freebox, displays in the subway or McDonald's utensils and a host of other things. I have known Pierre Garner for a long time. I wanted Elium to work on the design of the box. After consulting his team, Pierre accepted the project and in mid-April, I rang their doorbell for the briefing.

The workshops

 

Remember the cardboard prototype? I had an idea in mind: 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 they can take home.

 

Last year, it was a small robot made of plastic cups, which lit up and turned its head when touched. This year, they made a useless box.

 

It was necessary to design this paper version more seriously . Fortunately, I was familiar with the world of paper toys since I had already made (with Marc) cardboard objects with electronics inside. It was called reaDIYmate .

 

We had to make an electronic board, buy the materials, take out the Cricut to cut the cardboard and test the assembly to make sure we could do it in 3 hours. Result : the children 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 poet Séverine Daucourt was in residence. The children composed poems which they recited aloud, and the recordings were put "in the box" to allow them to be heard endlessly .

 

Oh, and for the older ones, I teach a class at the WSF on physical design and this year we had a lot of fun making useless boxes .

 

The software (seriously)

 

At the end of April, I had lunch (as often) with my friend Paul Guyot. For those who don't know him, he was the technical director of Violet, the company related to Nabaztag. He has done a lot of things since then, but in 2019 he wrote in Python the emblematic services of Nabaztag for the TagTagTag kit. I showed him the prototype and he offered to help me with the software which, although seemingly simple, needed to be approached professionally.

 

I started with a well-known microcontroller (the rp2040), but it consumes a lot when it sleeps. It snores loudly, so to speak. And consumption is an important issue for the box – I doubt people will think to charge their machine every night. It needs autonomy measured in months. So Paul suggested using another processor (an ESP32 C3 for the experts) which has the advantage of being low energy when doing nothing. The rp2040 is English, the C3 is Chinese. I don't know what to make of that.

 

He also encouraged me to use Erlang but he suggests that for every new project. Erlang is a functional language, close to LISP. LISP was used by AI pioneers, including Marvin Minsky, the inventor of the useless box. Are you following? Anyway, we started with Erlang and now I just have to learn it (and you too).

 

Paul also proposed giving a talk on la machine at the Erlang festival in Berlin in October. If you are in Berlin on October 20, 2024, come listen to Paul: the name of the talk is "La Machine: The Useless Box reloaded with Erlang and AtomVM".

 

May
 

May was partly devoted to shipping the TagTagTag cards from the 2024 production batch. Nothing to do with la machine, although sometimes I feel like I'm making a kind of Nabaztag: it's useless, it plays sounds, and there's something that moves.

Friends

 

I still had time to finalize the call to friends of la machine : a video and a first site made with limited resources that took me an infinite amount of time. By contacting people who were interested in Nabaztag, Minimit, and Multiplié, I quickly had 300 friends. Maybe you too.

Interesting

 

I went to London to attend the fabulous (vaguely annual) conference by Russell Davis: Interesting . I took the opportunity to see the Enzo Mari exhibition at the Design Museum and visit a few shops to spot what useless electronic gadgets were on the market. At Selfridges, I found these Bluetooth speakers with a few keyboard keys and a mini screen, all for £129.95. I thought that with la machine , I might have a chance.

The Electronics (Seriously)

 

It was time to get serious about the electronics. That's where Philippe Debadier came in. Philippe Debadier is the engineer who designed the electronics for Nabaztag a long time ago and hundreds of other electronic boards since. He agreed to help me and created the one for la machine. 

 

The schematic and routing are finished (routing involves moving from a theoretical electrical schematic to a real-life board) and the prototypes are complete. Now it's time to test. Thank you, Philippe!

June
 

Elium Studio (again)

 

On June 17, I had an appointment with Elium for the first work presentation. I won't show you all the options. You might prefer others, and now is not the time to doubt.

 

The chosen design is inspired by Memphis. Not Memphis, Tennessee, but the Memphis movement. The Memphis group was founded in the 1980s around Italian designer Ettore Sottsass (there's a podcast about him here). It's a happy coincidence because two years ago, I saw the exhibition dedicated to him at the Georges Pompidou Center and bought the catalog titled “L’objet Magique”.

 

So here, in preview, is the design of la machine with possible variations.

Elium also made a functional prototype with the electronic elements from my second prototype. A video is available here (it’s a prototype without a battery for those bothered by the USB cable). 

 

We are now entering the industrialization phase, which involves finding a manufacturer and studying all the details of fixing the elements inside the box (you don't want to know how it's held together now).

 

As always, the question of the manufacturing location arises. For now, we are discussing with a manufacturer in Normandy (even though I’m Breton) and a Chinese manufacturer (even though I’m Breton).

 

The Website (Seriously)

 

In May, work also began on the real site – not the friends’ site – work done by Pascale Moise who is, among other things, the author of the Minimit , but who is above all a website developer. She had the idea of a site that opens when you press a button. 

The "Paper Toy" Edition

 

Following the various workshops mentioned above, I refined the cardboard version of la machine to make a very limited series (no more than a dozen items, handcrafted) for those who want to have fun now but also test the various choreographies made of movements and sounds to be designed in the coming months.  

 

This is a do-it-yourself version, based on the same electronic architecture as the final version (but without the battery) and I imagine selling it for around €100 with a 50% discount on the final version of la machine when it becomes available. If you want to know more or if you're interested, write to me! 

 

Here is the kit and what la machine looks like in cardboard (it could become a collector's item, who knows!).

And the Legos?

 

Loïc was busy finishing his Master's (it's done) and started the first tests of the Lego version in June. We spent an afternoon figuring out how to integrate the motor, the arm, and the switch, which will be done with specific 3D printed parts. There will be about 300 pieces in the kit and of course the electronics, the motor, the speaker, the battery, and... the switch. 

Maker Faire

 

With my friend François Xavier Faucher (FXF), whom many of you know through the fabulous blog Toyfab, we decided to take a stand at Maker Faire Lille and create a giant useless box.

 

It’s the weekend of November 19/20 and it’s in Lille. Come in large numbers!

 

Here is FXF's first idea. He knows how to think big.

Next

 

Now that the electronics are finished, the software already has its GitHub repository (all the code will be "Open-source"), the site is well on its way and the design is finalized, there are two very big topics left: the behaviors of la machine and the financing.

 

The behaviors

 

Paul has developed a behavior description language that includes choreography of the arm and the launch of sound sequences, but I haven't had time to work on this aspect yet. I need to have clearer ideas, even if only for the artistic direction. For now, I'm oscillating between the cow box and IRCAM: I think I need to narrow it down a bit.

 

There is a lot to do because each machine is unique: the idea is to have enough sounds and behaviors so that when manufacturing, we can draw lots of a few hundred for each box. We need to find a way to vary movements and sounds by thousands. This is what I would like to test with the "Paper toy" edition.

 

Financing production (a.k.a. Money)

 

In the production of electronic objects, there are two kinds of costs: investment costs, which are called NRE (Non-Recurring Expenses) in jargon, and unit costs related to the manufacture of each machine. In the case of manufacturing electronic boards, there is very little investment to be made; however, in the case of plastic parts, if we use injection methods, molds must be made, the cost of which depends on the design. And cube design requires more complex tooling than average.

 

There are therefore two options: finance the production myself or make pre-sales to cover investments and production. I had long thought that I could launch the product by self-financing the project, as I did for the MiniMit. The marketing would have been really simplified: you buy it, and you receive it right away. But calculating the investments necessary to produce la machine, I realized that this would not be possible (a bit of blindness is always needed to start). 

 

Unless a friend wants to invest or help, I will go for a pre-sales system via a crowdfunding campaign (probably Kickstarter). I know these devices quite well and this will be my sixth campaign! 

 

A crowdfunding campaign means no more and no less than moving from 400 friends to 5000! The key to success is to reach the goal on the first day; otherwise, it bogs down quickly, and it becomes tedious for everyone. 

 

How to move from 400 to 5000 friends? Well, that is the 1000 Francs question, and if there are experts in digital marketing, I'd be happy to call you to chat because if mechanics isn't my strong suit, neither is marketing!

 

☀️

 

Thank you all for reading this far!

 

Let me know if you want a badge if you're interested in paper toys, and don't hesitate to help me grow our community!

 

And let's end with 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