Videos - The Coin Bot

Jump to This Time	Transcription Text
00:00:17	Well, that didn't go very well. Hey, what's up, everyone? So I have a confession to make. When I try to design something, say, machine or a software program, I don't really ever design anything. And by that I mean I don't draw things out
00:00:37	in CAD or work through the process on a notepad or come up with software ideas and to do lists before I start to actually code or create things, I tend to just iterate and iterate and iterate until I get a semi working program or product, and then I just make do with what I've got. I get it working well enough. But any of you that have followed Wintergatan Wednesdays realize
00:01:09	that Martin has recently kind of put a halt on his Marble Machine X, because he realized that as his design got more and more complicated, he really needed to take a step back and go back to the drawing board and really investigate his own design requirements for the machine. It appears that he was putting form over function, and that means that if something in his mind looked cool,
00:01:35	he put that as a higher priority than the machine actually functioning properly. One of the things that made him have a change of heart was a conversation that he saw between Elon Musk and the Everyday Astronaut, where Elon laid out his five engineering design principles. And I thought it was interesting for Elon to admit that he often does these steps in reverse order.
00:02:02	And as I try to generate this machine that's capable of finding numismatically interesting coins that I'd like to add to my own collection, I find myself doing the same thing. I chew on these designs in my head over and over again, and I never really get a chance to see if the design is optimized or if the design should actually even exist.
00:02:27	So in this video, I wanted to go over those five design principles that Elon Musk lays out, and I can also talk about my own design requirements that I originally had for this machine. So the first design principle that Elon Musk has is to make your design requirements less dumb. What I'm trying to have us all
00:02:52	implement rigorously is the sort of five step process. First, make your requirements less dumb. Your requirements are definitely dumb. It 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.
00:03:09	What Elon means by that is that engineers, often at the onset of a project, will make the design requirements much more complex than they need to be to perform the actual needed function of the finished product. The second design parameter that Elon tries to use is to try to delete as many parts as you possibly can from the finished design. Then try very hard to delete the part or process.
00:03:36	This is actually very important if you're not occasionally adding things back in, you are not deleting enough. The bias tends to be very strongly towards let's add this part or process step in case we need it. But you can basically make "in case" arguments for so many things. And step three, once you've deleted all
00:03:56	those extraneous parts and you've actually deleted so much that you actually have to start adding parts back in, then you can kind of optimize things. Anyway, so step one, make your requirements less dumb. Step two delete the part for process step. If you're not deleting a part of process step at least 10% of the time, if you're not adding things back in 10% of time, you're clearly not deleting enough. Right.
00:04:21	And then only the third step is simplify or optimize . . . The third step. The reason it's the third step is because it's very common. Possibly the most common error of a smart engineer is to optimize the thing that should not exist. And I think this is interesting. He says that there's nothing worse than
00:04:44	optimizing a part that shouldn't even be there. Step four is to go faster. And then 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. If you're digging your grave, don't dig it faster, stop digging your grave.
00:05:08	And in my experience, there's a couple of different ways to go about that. You can actually literally speed up the machine so that it processes faster, or you can create more machines so that you process more in less time. And the fifth and final step is to automate. And then the final step is automate. Now, I personally made the mistake of going backwards on all five steps where
00:05:36	literally I automated, accelerated, simplified, and then deleted. I think that's interesting that Elon has learned through his own experience on the Gigafactory floor that automation should be the very final step in this engineering design process. And I think he's learned that there are often many steps in processes that either could not or should not be automated.
00:06:03	There are some things that humans are actually much better at doing. And since I put these YouTube videos out, since I started kind of putting myself out there and asking for feedback and comments from people, I've actually gotten some good comments about why I'm doing the AI steps of this machine last. That's not only because of Elon's design principles where he places automation
00:06:28	last, but the AI is actually the automation step. And I can envision a machine that has value even without that AI portion. If you think about it, you can create a machine that's capable of taking images of both sides of the coin and presenting those to the user to sort manually. I think that would be really handy for someone who is vision impaired,
00:06:52	for example, or in my own case, someone who's getting a little bit older and has to look through at his collection entirely through a magnifying glass. What I've discovered is that the antidote to creating dumb design requirements is actually putting yourself out there, and I hope that this channel serves that purpose. I want to put my design requirements out
00:07:13	there and invite feedback and constructive criticism from everyone, and they can kind of help me to see if I'm going down the wrong path or if I'm making some sort of mistake, and I will really try to listen to that. So without further Ado, I really want to share with you my original design requirements for this machine, and maybe you can maybe think about which ones those are the dumb ones.
00:07:41	The first design requirement that I had is that the machine must be capable of finding every numismatically interesting penny that comes through the machine. And I thought in that way that I could post those pennies on eBay, even if I wasn't interested in adding them to my own collection. That way the machine would be at least generating some income for me.
00:08:02	The second design requirement that I had originally was that the machine is capable of flagging any coin that I would like to have in my own collection. But of course that would mean that the machine was aware of what was already in my collection and it would be necessary to design the machine capable of understanding what grade my coins were, which adds a whole new level of complexity when you try to machine grade a coin.
00:08:30	My third design requirement is that the machine must be somewhat small and portable. My idea was to just have it sit on my desktop and process coins all day. My fourth original design requirement was that the machine must be fast and reliable and never miss an important coin. The fifth design requirement is perhaps the most challenging,
00:08:53	and that is that the machine must be able to be made cheaply, and I was envisioning using a 3D printer to create plastic parts and maybe using old cell phones or something as the camera equipment for that. And the 6th design requirement was that I just set it and forget it. My original idea was to just be able to pour thousands of coins into some Hopper at the top of the machine and just
00:09:16	have it sort out all the numismatically interesting coins with no intervention from me and no jamming, no worries. In fact, I envisioned a final version of the machine that was capable of actually putting coins in a coin flip and posting the auction on eBay for me without any intervention at all for me. After inviting feedback through these
00:09:38	YouTube videos, people have sent me some videos on YouTube of machines that do something similar to what I'm trying to accomplish here. And what I noticed in those machines is their design requirements were much simpler because they allowed some human intervention. Their main function was really to eliminate the 99.99% of coins that no one's really interested in looking through and they present coins that may be
00:10:04	numismatically interesting to the user so that the human can actually look at the coin and see if it's something they want to add to their collection or sell. Now I thought that was interesting because again, like you're involving the human and not just automating everything with no purpose. And after seeing that I could see that it was really important for me to kind of dumb down my design requirements
00:10:27	and maybe get the human being a little bit more involved. So that leaves me with just two design requirements for the finished machine. Mainly I want a machine that's still portable, something that sits on my desktop and doesn't take up too much space and I also would want a machine that's capable of removing those 99.99% of the coins that I really don't want to look through.
00:10:51	And in so doing, I'm hoping that that kind of rejuvenates my interest in coin roll hunting and about build some excitement Because it will only present me with dates that I'm interested in looking at. And of course that leads us to the first design that I have in my head for this machine. But of course having it in my head doesn't
00:11:10	really help anyone as the point of this YouTube channel is to get feedback from everybody to see if I can even further dumb down my design requirements. So without further Ado, Let me jump into FreeCAD And I will try to kind of lay out a quick sketch of what I had in mind. Now keep in mind I don't have very strong skills in FreeCAD yet so just bear with me a minute And I'll get right back to you
00:11:49	much much much much later. Well, that didn't go very well, so we'll save that design for the next video but if you like this type of content, I hope you hit the LIKE button that you'll SUBSCRIBE to the channel and you'll hit the notification BELL so you don't miss out on any future videos. And of course, I hope that you'll leave
00:12:07	a comment down below if you have any feedback for me or if you think that I'm going down the wrong path, I really invite your feedback and constructive criticism. Until next time, I hope you have a great day And I look forward to seeing you in the next video. Thanks everyone.