Marble Machine X - A Lesson in Dumb Design
The best kind of videos that I've made are videos where I'm trying to teach myself new stuff, and you gets the chance to learn with me. And in this very video, I'm trying to teach myself some of the most profound stuff I've ever attempted. And, if I manage to learn just a tiny sliver of what I'm trying to teach myself, this could be one of the most valuable videos I've ever made in my humble opinion. Okay, so what's going on with the Marble Machine X right now then? During this summer I thought long and hard about this, and I have identified four major issues holding me back from succeeding with the Marble Machine X. Issue number one: I didn't make the design requirements less dumb.
As I will show you in this video, this is is causing major issues with the functionality of the Marble Machine X. Almost every persistent problem on the machine can be traced back to my dumb design requirements. This has been such a big learning experience for me. And, by making this video I hope that some of you get the chance to learn from my mistake.
In a brilliant three-part video series, Tim Dodd, the Everyday Astronaut, toured the SpaceX factory grounds with Elon Musk, and... in the middle of the tour, Tim asked Elon why the grid fins no longer folded in. Elon replied by giving a 5-minute lecture about design processes that tells us exactly why the Marble Machine X is not working. [see captions in the video] Elon Musk: -Does not matter who gave them to you, it's particularly dangerous if a smart person gave you the requirements.
because you might not question them enough. Tim: - Yeah, you might take it as like gospel, like "I have to do this thing". - Yes. Everyone's wrong, no matter who you are, everyone's wrong some of the time. So make your requirements less dumb. [see captions in the video] This is actually very important.
If you're not... occasionally adding things back in, you are not leading enough. Only the third step is simplify or optimize. The THIRD step. Not the first step. The reason it's the third step is because it's very common, [see captions in the video] - Right.
- Okay, and say, well, how do we get... Why would you do that? Well, everyone has been trained in high school and college to... that you gotta answer the question. Convergent logic.
So you can't tell a professor, your question is dumb. You will get a bad grade. You have to answer the question. [see captions in the video] That is: [see captions in the video] - Right. - Finally you get to step four, which is accelerate cycle time.
You're moving too slowly, go faster, but don't go faster until you have worked on the other three things first. [see captions in the video] And then the final step is automate. And uh... now I have personally... made the mistake of going backwards on all five steps, multiple times. Where, I tell you, I automated, accelerated, simplified and then deleted.
But like one example I've talked about before, is like the, they were these like fiberglass mats, on top of the Model 3 battery pack, they were in between the floor pan and the battery. And it was at one point choking the battery pack production line and I was like, basically living on the battery pack production line, like probably fix the line. But finally I said, what the hell are these mats for? And... I asked the, the battery safety team,
because I was like, are these mats for... what are these mats for? I said "are these mats"... I said "are they for fire protection or something"? They said, "No, they are for noise and vibration. So you don't get that."
And I said, "But you're the battery department." And I asked a NVA, noise vibration analysis, team, what's it for, they said fire safety. So literally it was like being in a Dilbert cartoon. [see captions in the video] [explosion] Okay, so, what actually makes the design requirements for the Marble Machine X dumb? I have put aesthetical requirements, meaning how "cool" something looks, at the same importance level as functional requirements.
I... underestimated the difficulty of the project so much that I thought that I could afford to prioritize looks over function. This shows you how extremely naïve I was, going into this process. I underestimated the problem, and I overestimated myself. Looks have to come second, after the thing is actually working. Who would have thought? [laughs] And, yeah, I've been aware about this mistakes for quite some time.
This is what I said about the exact same topic in a video from long ago: Today I've been contemplating giving up on the machine, and I also been contemplating how amazingly awesome the machine will be when I finish it. I'm fighting this fight, you know, between looks and mechanics. It's all this eternal war, between the artist and the engineer.
I'm playing both sides and I'm losing on both sides, basically. While the machine is made up of all these arms, the mechanics is really the foundation for everything else. If the marbles get stuck, or if they fall out of the machine, nothing else really matters, like Metallica once said. I've made a demonstration here to show you exactly how these dumb design requirements result in failure after failure on the Marble Machine X.
So the goal here, is for the marbles to go from point A up here down to point B down here. On the table, we have two competing solutions. So, here on the right side we have the constrained solution, where the marbles are going through this PMMA tube.
On the left side, we have the unconstrained solution, where the marbles go in a group straight down, hopefully into the bucket. Let's see how well these designs are doing their jobs, so... I'm gonna start over here with the constrained path.
[marbles released] We have two marbles in the bucket, and... Oops! Looks like the marbles got stuck there. Okay, that's not so good. Let's see the unconstrained path.
[marbles released] Aah! All the marbles reached the goal, so... Okay. Turns out this unconstrained path actually did a really good job. Who would have thought? [chuckles] Which one of these two solutions do you think I used for the Marble Machine X? Of course this one, the constrained one.
So why did I choose this solution, since it doesn't seem to really work so great? Because it looks cool. [chuckles] So from the beginning of the project, I just wanted the marbles to go in one line, like a factory assembly line, and... that was kind of a darling, so... What do you expect? Do you want me to just kill my darling, just because it doesn't work? [chuckles] Na na na na. This old trope "form from function", it's some kind of antiquated boomer design advice.
I'm more into the millenial version "form and perhaps function". Yeah, let's take that one. And then let's accept donations from like 11 000 people, who believes in my abilities, and use the funding to put it into a "form and perhaps function" project. Yes, let's do that Martin, way to go. So this design that gets stuck, we have it in the machine itself.
Right down here. So, before the marbles enter this tube right here, they have to be collected and go into a single line. And, the marbles has never got stuck ever in the history of the Marble Machine X here, until recently when I finished the cyber bass, the marble throughput was increased, and guess what, the marbles got stuck. Who would have thought? [chuckles] [audience cheers] Let me tell you about what's happened just before the last livestream I made. before going to Sweden over the summer.
It was the very last day before my trip. The cyber bass was ready. Sun was shining, the floor was free from marbles and I felt fully prepared for a livestream. Where I wanted to impress on you all how well the machine was doing, and... Normally before these livestreams I was always late, doing last-minute fixes under huge stress, but not this day. [chuckles] I was ready like, one hour before we were supposed to go live.
I just had to shower, and then crush it together with Hannes2000. Felt great. But... just before heading into the shower I made like one final test to prepare the machine and suddenly I heard the heart-breaking sound of marbles falling on the floor.
I stopped the machine, and the constrained marble path had flooded for the first time ever. At this time I didn't panic. I cleaned up all the marbles from the floor. Added a little bit of tape and tested again. The marbles flooded again. Slightly more panicked, I repeated these tests. And the marbles flooded like three more times.
Every time resulting in me crawling around like Gollum, picking marbles from the floor. So in the end I got it right, and the marbles seemed to stop flooding. Good, all was good. Now it was twenty minutes until the livestream was set to go live.
I was a little bit stressed but not too much, so I just needed to crank the machine one more time, in preparation for the live stream. So I started to crank the machine and I saw all the marbles starting to go up in the first gear. And as in slow motion, I saw the marbles fill up the whole gear and then I saw the PMMA tube here explode. [chuckles]
With marbles exploding out all over the floor as a result. So, the lower gear moves faster than the upper gear, so when the marbles are pressing too hard on that pipe, they had nowhere to go and the pipe exploded. Perfect! Now it was ten minutes until the livestream was to go live.
The floor was filled with marbles, and the transportation pipe was broken, so I taped the pipe with masking tape, I crawled around like Gollum yet again, and got all the marbles up from the floor. Four minutes before the livestream was going to go live, I went into the shower in full panic. Seven minutes later we started the livestream. Only three minutes late.
And even though the machine hadn't been fully calibrated for timing at this time, the machine actually played pretty well. Unedited, live mixed music. So, here are some clips from that very livestream. [♪ unreleased / 80000 Marble Test ♪] [Cyber bass riff] [Drumming cats and dogs sounds] Hannes2000: - [laughs] Yes! So, why did my fully prepared livestream turn into chaos and panic? Well, because I didn't my design requirements less dumb.
That's why. Because I prioritized looks over function. But wait! Don't the marbles actually need to be constrained to fall into the magnet lift correctly? Yes.
But, that's exactly the type of wrong question Elon thought us to not answer, but instead question the validity of the question itself. The marbles only need to be constrained if I'm using this type of magnet lift. Okay, so, why did I use this type of magnet lift? [chuckles] Anyone care to guess? Waiting... Yeah, because it looks cool. [chuckles]
In fact, I think it looked so cool. So I accepted that with the magnetic lift system, I needed a complex demagnetizer as well. Which in itself became an endless point of debugging and optimization. So, if I had skipped the magnetic lift, not only could I have simplified the marble transport, I could also have skipped several complex moving parts. But, I didn't do that, because hey, I thought it looked cool.
Ow, I hurt my eye. [chuckles] Oh anyway. So when I started to hangout with engineers, they screamed at me to not use the magnet lift. I thought that engineers were generally narrow-minded, negative and lacking in imagination.
I thought it was up to me, the Jesus of optimism and "yes can do" attitude, to save these lost souls from the dark paths of skepticism. [chuckles] And now after three years of crawling like Gollum, the pain in my snowflake knees have finally made me go full circle. And now I think that the skepticism of a weathered engineer is actually a thing of beauty.
It's a driving force and not a stop block, so... There... I guess there's a reason why it's called "form from function". And there's a reason, why it's not called "form and perhaps function". Who would have thought? [chuckles] Am I just going to scrap this machine? No. My plan is to finish the assembly of this machine, warts and all. And then record one song and make one music video with _this_ machine because I think the Marble Machine X deserves that, and I think that you and everyone invested in the Marble Machine X deserves that.
But, what about the world tour then? Has all this actually been a complete waste of time? Absolutely not. Queue optimistic music. [Optimistic music] So let's zoom out. Shift our perspective, and think about the machine that builds the machine. The machine that builds the machine is in better shape than ever. I would never have the skills I have today if I hadn't gone through this whole process.
We built the Marble Machine X and it's nearly working perfectly. Nearly is not enough, but the fact that we got so close tells me that if we can make a new machine, with ten-X improvements, using what we have learned, we will have a machine that can go on world tours, so... For example when SpaceX test flies a new design, they don't really care if the rocket crashes on impact, as long as they are getting the data they want. Dave Lee: - I was talking to this guy Mac... Marcus House,
he has a SpaceX channel. Warren: - I know Marcus, yeah. Dave Lee: - Yeah, great guy, and I was asking "Why is SpaceX making like SN21, and SN22, like you know... rockets when they even haven't tested SN20. And his answer was really insightful, he's saying: "It's because SpaceX thinks the machine that makes this machine...
the process of production is more ... nice. Warren: - The machine that builds the machine. Dave Lee: - Exactly. ... is more important than actually the product,
so they're willing to scrap the product. They don't even have to ever use whatever they made for SN21-SN22, it's just the learning that they've gotten from actually trying it, it is worth it." So how can we use what we have learnt, and make something beautiful out of this new knowledge? The answer is Marble Machine X-T.
"T" standing for "touring" version. And, that would be the third generation marble machine. And, the MMX-T will be brutally simplifiied, both design-wise and process-wise. And, you can read my early thoughts on these simplifications on the Wintergatan.net blog.
I can't wait to get another chance, I can't wait to knock this one out of the park. Don't underestimate the man who overestimates himself. Third time's the charm everyone. So... [click]
The obvious solution to issue number one is starting over with Marble Machine X-T. [♪ unreleased / 80000 Marble Test ♪] [click] So far we've looked at the first of the four major issues holding me back from succeeding. with the machine.
Before moving on to issue two, three and four, I would like to make an update regarding the crowdfunding project for the Marble Machine X. The Marble Machine X is from today officially fully funded. The MMX crowdfunding campaign is a complete success. You did it! I would like to meet each and everyone of you personally and say "Thank you", face-to-face, in the real world. In case the Wintergatan world tour ever happens, we are offering Marble Machine X contributors a special VIP + 1 ticket. So each VIP+1 ticket would grant you: entry to a Wintergatan concert for you plus a friend, meet up with me and other MMX contributors.
You also get the chance to meet the Marble Machine up close for you and a friend. So you can inspect what you have supported. On Wintergatan.net you can read about the different groups that are elligible for these VIP+1 tickets. If you belong to any of these groups, please register your VIP+1 ticket on wintergatan.net.
I have to say it feels kinda awkward to give out a reward that might never become real. But on the other hand, by now I think that we've reached a common understanding about the uncertainty of this project. The risk factor is part of what makes it great.
The Marble Machine X is right up there with Star Citizen at this stage, so... [chuckles] If the world tour happens, I want all MMX contributors to be able to visit the concerts, not as guests but as VIP members of the team that made it happen. Hang on here, Martin. We're not really with you. How can the Marble Machine X crowdfunding project be completed before the machine itself is actually finished? Thanks to your support, I reached a point where I have everything I need to finish this project.
With your support, I bought materials and music instruments for the Marble Machine X. And now I have all the parts I need to finish the marble machine. With your support, I've been able to collaborate with Olivier from id-acoustique to build my dream music studio. And the music studio is now completed. With your support, I have built my dream over-the-top workshop.
In this workshop I have all the tools, machinery and organization I need to finish the Marble Machine X. And on the workbench right here, I'm working on a marble demonstration for later in this video. Let's do a sneak peek. Strange [chuckles] With your support, I've even built an outdoor gym. So these snowflake arms don't melt into water.
So I'm actually transfering your funds directly into muscle mass into the very biceps that are actually building the Marble Machine X. Let's check this epic gym out. So yeah, here it is and you can see the progress is absolutely brutal. I've already started with like these extra things on the side here.
Let's see, what's the weight of these. One point twenty-five kilo. One on each side, remind you. So, we're like already gone from like twenty kilo to like twenty two and a half. So yeah, progress is real Here's my chin bar, and yeah... From snowflake to flaming fire, yo.
With your support, I could move to a house on the French countryside where the sun is always shining. We have grapes, and olives. So, I could build my studios here, right in the middle of nowhere. In the nature I love. With your support, I've made an over-the-top video studio. I have light and cameras, and with this equipment I made over two hundred youtube videos.
With your support, you have given me an abundant amount of the most scarce resource we have: Time. What a rare gift for someone to be able to accept. Thanks to your support, I've been able to build a team of colleagues, that means the world to me. You have given me the perfect conditions, it's now up to me to finally execute. So, wouldn't more money make it easier for me to execute? It did make it easier, until now. But today, when I can't really find efficient ways to utilize the money to speed up this process, I can't keep accepting it with good conscience.
If you end your financial support today, I want you to know that the Marble Machine X project will not suffer. Martin, Martin, Martin. You still don't get it, do you. We want to continue no matter what. We actually don't really care about all these snowflake ramblings of yours, we want nothing back. Shut up and take my money. [laughs] I know you're out there, and I appreciate the hell out of you.
You're more than welcome to continue to support Wintergatan, but, I want to be clear that what you would be supporting is changing. From the first of October 2021, your support would be for all Wintergatan work in general. "Carte blanche". The support would not be earmarked for the Marble Machine X or for making youtube videos.
When I'm building and making videos at the same time I'm constantly multitasking. Multitasking makes us less efficient and more prone to errors and this leads us directly [click] to issue number two: I have been working on too many non-essential things, so to improve I need to work on fewer essential things. [♪ unreleased / 80000 Marble Test ♪] In the book "Essentialism: The Disciplined Pursuit of Less", Greg McKeown argues for the importance of finding out what is actually essential, and then eliminate everything else. A "Non-Essentialist" he writes, "thinks almost everything is essential" and "views opportunities as basically equal". An "Essentialist" on the other hand "thinks almost everything is non-essential" and "distinguishes the vital few from the trivial many".
"The difference between the way of the essentialist and the way of the non-essentialist can be seen in the figure opposite. In both images the same amount of effort is exerted. In the image on the left, the energy is divided into many different activities. The result is that we have the unfulfilling experience of making a millimetre of progress in a million directions." Kind of relatable [laughs].
"In the image on the right the energy is given to fewer activities. The result is that by investing in fewer things we have the satisfying experience of making significant progress in the things that matter most. The way of the essentialist rejects the idea that we can fit it all in. Instead it requires us to grapple with real trade-offs; make tough decisions.
[see video] In chapter eleven, Greg writes about the power of a graceful "No". [see video] "[...] the disciplined pursuit of less is just lip service. We have good reasons to fear saying no, yet, as hard as it can be to say no to someone, failing to do so can cause us to miss out on something far more important."
Okay, how does this relate to my situation and to the Marble Machine X? Well, I've been spending my energy in a million directions. Documenting my work to make youtube videos and livestreams has been essential for this process until now. To make progress from here, I have to grapple with real trade-offs and make tough decisions. And going forward I need to deem most youtube videos as being non-essential. And only make a few really essential videos. Also the work that remains on the machine now is tedious and repetitive.
I have nothing new to share with you. The Wintergatan videos have already become tediously repetitive. I don't want them to become a forty season soap opera, where me and Wilson keeps dealing with our relationship issues, season after season. [see video] I strive to become one of those people. [click] The solution to issue number two is to practice essentialism.
[♪ unreleased / 80000 Marble Test ♪] Issue number three: I have been doing a lot of shallow work to improve. I need to do deep work. If essentialism can teach me, what to work on the question about how to work on the essential things still remains. Everything beautiful I have ever created, came from long periods of uninterrupted concentration. It came from working alone in an enclosed space out of view from the outside world. Since my creations became popular, my way of working has changed.
I´ve been working in public, multitasking on to many projects and my productivity has gone down. In Essentialism, Greg McKeown calls this "the paradox of success". [see video] "... and led to our success in the first place."
And since I fell into the paradox of success, I´ve released zero new songs on spotify, I built zero functional Marble Machines. For the past three years, I´ve worked harder than I´ve ever worked. I´ve been busier than ever, and yet I´ve not been productive. The work, I´ve done for the past three years, is what Cal Newport calls "shallow work". In his book "Deep Work - Rules for Success in a Distracted World", Cal Newport defines shallow work as: [see video] Cal then goes on to find the alternative: [see video] To exemplify the value of deep work, Cal tells a story about Carl Jung, who in 1921 created large waves in the field of psychology. Jung was famous and busy, but he wasn´t satisfied with busyness alone.
So, he built a tower out of stone in the woods to promote deep work in his professional life. The resulting Bollingen tower has become a legendary symbol for concentration and Jung later went on to become one of the most influential thinkers in history. I already have a stone tower in the forest. In fact, I´m sitting in it right now.
But what good is a stone tower in the woods, if I squander my time in doing shallow work. As Neal Stephenson, the cyberpunk author puts it: [see video] Kind of relatable. Why do I fall into shallow work so easily? According to Cal one of the many answers is, that in our current hyperactive internet culture many of us are tending to work increasingly visible business. Because we lack a way to demonstrate a better way our value. "Busyness as Proxy for Productivity: [see video] This explains why I always retour to making videos despite having said several times that I would make fewer videos.
It feels like I´m being more productive, when I show you in a visible way what I´m working on. But fate loves irony: the more I focus on showing you, what I´m working on the more I do shallow work and the less I get done on the machine. Inspired by deep work and time blocking, I made a schedule for my autumn, that looks like this. Instead of doing unorganized shallow work 12 hours per day I am going to do organized work 8 hours per day. And leave time for exercise and rest. When I build the first machine, I spent hours cutting plywood gears manually by hand on the bandsaw.
One slip in concentration could ruin several days of work. And I remember being completely happy standing by the bandsaw deeply focused those days. " 'The best moments usually occur when a person´s body or mind is stretched to its limits in a voluntary effort to accomplish something difficult and worthwhile', Professor Csikszentmihalyi calls this mental state flow." "Deep work is an activity well suited to generate a flow state".
"Flow generates happiness." "The act of going deep orders the consciousness in a way that makes life worthwhile." "Deep work is a proven way to deep satisfaction" Okay. So, this is a picture of my brain during the past year, and this is a state, I´m trying to enable by the removing what is non-essential and going deep. Solution to issue number three for me, is to practice deep work.
[♪ unreleased / 80000 Marble Test ♪] Issue number four, that has held me back from succeeding with the Marble Machine X is, that I haven´t been able to delegate to a team that can execute. Most importantly, I´m not talking about my current team here. I´m talking about the potential larger extended team, that I should have build. It turns out, hiring and managing people is a full time job in itself.
And it also turns out, that I am really, really, really bad at it. Navigating social webs of expectation takes me out of deep work. I could perhaps improve a little bit, but not enough.
So, first solution is, that I´m hiring a Wintergatan COO. Chief Operating Officer, full time. If you go to wintergatan.net, you could find the job listing there. In short: "I am looking for someone who can act as my right hand, and together with me, make Wintergatan 100x more efficient.
Your exact role would be tailored to you, but one of the most important aspects is that you are plugged in to the Cultural Matrix of our world and feel the macro currents of how value propagates in the digital era. We are transitioning from web 2.0 to web 3.0 and I am looking for a partner who sees this big picture and have ideas about how to operate Wintergatan in this exciting future." If you are interested in working with me, please read the job listing and reach out to us. Thank you!
The second idea is to form some kind of decentralized autonomous organization with a governance token. I think, this could be extremely suitable for the MMXT project. A DAO would create a new way of collaborating where participants get to share the value created. So, the solution for issue number four it to first hire a COO and then later potentially also creating a DAO. This is my vision, this is the direction of progress I want to pursue.
Every step I can take in this direction, will be a great improvement. [♪ unreleased / 80000 Marble Test ♪] Okay. So, what are my near term plans? The very next project for me is, that I´m going to finish a very special project that is inside this box. I´ve been working on a secret project in the background for the last two years. I wanted to keep it behind the scenes, because I wasn´t sure how it would panel. But it´s nearing completion now.
And don´t get me wrong. I´m not starting a new project. I´m finishing a project, that was started two years ago. I have collaborated on this project with one of the best mechanical engineers there is.
Florian from the fhuable youtube channel. Florian has made the prototype inside this box, and Florians execution on this project have been over the top amazing. Crushing all my expectations.
But, I was kind of naïve to start this project alongside the Marble Machine X. But if we succeed, I´m sure that you are all going to love it, and understand, why I went for it. It´s going to be a project, that you will be able to interact with yourself. And yeah, I can´t wait to show you this. Justin: - I mean i'll say that I... I also have kind of a bit of a hacker mindset, you know. I try to find the vulnerabilities, I try to find a crack, I try to take a route which no one expects and is kind of hidden.
Oftentimes you have a problem and it looks like there's some sort of impossibility result. Like you can't do something. And like the way that you escape in possibility results is... is kind of by redefining the problem or redefining the assumptions. Um, you... you really don't want to take for granted things that other people took for granted.
And... and basically relax the problem in very subtle ways, where... where suddenly you... you open up a design space. Um and so yeah, this is something that I like doing is, um... kind of taking something where people seem stuck. You know they're not... they're no longer making progress, they thought about it for a long time.
And they believe it's impossible and then kind of coming in with fresh eyes and saying - and a new perspective - and saying actually if you tweak your... viewpoint just a little bit. There is a way around it. By relaxing the problem in very settle ways and by shifting my perspective, I now see how I have been tying myself down to the ground, with a million tiny ropes like Gulliver, the giant.
I have tied myself down with dumb design requirements, non-essential commitments and shallow multitasking. And I have stayed there on the ground for too long, laying in my comfort zone of busyness. It´s time for me to stop paying only lip service to these ideas. It´s time for me to stand up, shatter the non-essential bones and go towards my long term goals. Thank you for watching.
[lights switching off] // Subtitled by Wintergatan Writers. Join our team on discord. //