 OK, so, next up, we have a man with a very lovely shirt. It's very dappily dressed, which is convenient because he's talking about automated knitting machines. And his name is Tom Myers. Hi everybody, and my shirt says hi as well. 3D printers are really, really cool. A really, really fantastic technology that has the capability to fundamentally disrupt centralized manufacture. I think they're really cool. But the problem, at least personally, I see with them, is that they don't make things that ordinary people want. They're fantastic in a hackspace where you need engineering parts and custom-made little things. But there are only so many statues of Yoda that ordinary people want and need in their lives. I want 3D printers to be as good as possible. I think we should design 3D printers to make things that are more useful to ordinary people, things that people will use a lot more in their everyday lives. And so I've been doing just that. I've been building a 3D printer that makes clothes. I'm afraid I don't have it with me on stage. I was planning to, and it's in my tent at the moment. But I tried to put it up in the makespace village last night. And unfortunately, it collapsed partly, and some of the bits broke. So I can't show it to you today, but I do have a photo, of course. I don't know if you can see it. But there it is. Perhaps, more importantly, I have some examples of the things it can make. This is, I don't know how many of you can see. I'll switch to the appropriate slide. But it's a tiny little EMF jumper. Yeah, with a few mistakes made. This does, however, demonstrate what this machine is capable of, and what I was designing it to do, which is essentially to be a complete 3D printer that makes complete clothes in a single step from nothing, from wool or cotton or various other things. As I say, it makes clothes in a complete step. You'll see there are no seams or stitching of any kind on the sides or on the sleeves. That's because of the way it's made. All of it's knitted together in a single step. And, yeah, as a consequence, it's very much a 3D printer that makes clothes. I've talked a lot about how it can knit as a knitting machine. But I actually, I don't particularly like the term knitting machine. I prefer soft printer, so a 3D printer that makes soft things, a soft printer, because it can weave as well as knit. I have evidence for this. It's very small. It's not a big demonstration like the other one, but it's just a little piece showing that was also made on the machine that's a woven fabric rather than a knitted one. I should rewind slightly. There are two major ways of making clothes in industry, knitting and weaving. Both are... weaving is when you take a series of vertical threads and then interlock horizontal ones next to them on a loom, and you just keep doing that lots of times. You end up with a sheet of fabric. You then cut that out and stitch that together. Knitting is similar. I'll explain that in a little bit more detail because it also relates to how the machine works. I've got a video here. Let's just me explaining how it works on Mutex. I'll be doing the talking. I don't know how well you can see. I have to point things out, but I also have to pause it. Sorry, I'm moving backwards and forth between the two. This, which I will now bring to the front, hopefully without breaking it, is one of the beds of the machine. You'll see how you've got a long series of these needles equally spaced along it. There will hold the garment in place. Oh, thanks. Let's just hold this. Thank you. I'm a 3D printer. You'll see you've got a whole series of these loops and it's held at about a 45-degree angle. Almost the angle he's holding it, actually. These are capable of moving up and down like this and then there's a loop held on each one. If you come out of it just so they can see the thing now. You'll see you've got a loop of fabric held on each one of those hooks. They're just held down by some weights you can see at the bottom. The needle moves forwards. Well, you can see this. The needle moves forwards and then there's a little latch which I'll show you on here. You see it's got a hook and it's got a little latch here. You can put that down if you're getting tired. There's a little latch here which the loop of fabric goes over and past and then you feed a new piece of fabric through the hook and that creates a new stitch as you're about to see in the video. You can see the new stitch forming in a moment just there and that creates a new loop of fabric and that you keep doing that over and over again to create lots of loops of fabric and consequently a whole garment. That's essentially how the machine works. In this video I'm doing everything by hand but it's just as easy to do it using a carriage that moves forward and back along the bed and has some angled sides that push the thing up as it goes past and then push them down again. That's essentially how you do this autonomously like I was hoping. I'll move back to the presentation now. That's meant to be blank. What is this technology capable of completely well? It's capable of making complete clothes, as I've said. It's also capable of doing pockets and various other fancy things, embroidery, things like that, are all doable with techniques like this. What it can't do, perhaps obviously, is it doesn't have any buttons or zips, things like that, because they're hard. You can't just knit those together. The simplest way is to get the user to sew them on for you, but none of them are really straightforward. It's a little bit awkward. The other important thing about this technology is spacing the needles as closely as possible. In the machine I have, which you saw for making this, the needles were quite far apart. They were nearly a centimeter apart from each other. That means you have to make chunky garments like this. Obviously that's not ideal. You'd rather make thin things, because that's what most clothes are made of. The way you do that is instead of having them like a centimeter apart, you have to have them something like a millimeter apart. That's definitely feasible. I have examples of needles here, which are, I think, 340 microns wide, about a third of a millimeter. Very, very thin needles that can accommodate being spaced very closely and consequently making really high quality fabrics. It's actually aligning them together in a bed like the one you saw. You can't really do it on a laser cutter like I made that bed with, because the width of a beam is about 150 microns, which means you're basically running out the resolution limit of the laser cutter. The way you do it instead is you have a big block of aluminium, a very thin circular saw, and you just cut the grooves one by one, which means you have to do it by hand, and as a consequence of that, in the future we'll be able to make really high quality fabrics completely autonomously. That's very cool. The next question I want to answer is why would a user want any of this? We already have a fashion industry that's very functional, and we all use it. Why would we need to change it when designing these machines that can custom make clothes? There are a few advantages. The fact that it's custom made means it can be fitted much more easily. When you're making these in China or the Far East through a long supply chain, it's really hard to do anything better than a small, medium, large, but if you're making it in your own home, or very close to your own home, it's pretty easy to simply take somebody's size, their waist size, height, input that into the machine, take the design, map that to your particular spec, and custom make it for you right in your own home. That's very feasible with this technology, and it means the clothes it makes will fit better than any that you'll find that have been mass produced. A second advantage is speed. Fashion obviously moves and changes quite quickly, and it's really beneficial to be able to move quickly along with it. That's a huge amount of time to wait to bring in a new design, whereas with this, all you would need is a fashion designer to design something, upload it. You could pay them a small amount, download it, and make it really, really easily, and much more quickly allowing you to stay on top of fashion friends. It's also really good for fashion designers, because at the moment, if you want to make any fashion line at all, there's an initial outlay to making any of this stuff. You have to be fairly confident that you're going to sell it in order to make a profit, whereas with this, you just upload the design online. If one person makes it, you've already made a profit, there's no wastage, there's no loss, and it allows people to be a lot more experimental and a lot more creative with their designs, because they can take risks if things don't succeed. The final advantage is ethics. Obviously, a lot of the clothes you'll find on the high street, a lot of the clothes that we wear weren't made in very ethical conditions, low wages, cramped long hours, and so on. This absolutely guarantees that that isn't going to happen, because it's made right there in front of you. Ethically, it's a real advancement, and I think it could also pressure the fashion industry into improving their own standards if there's a viable competitor that doesn't have any of these ethical issues that they have. Finally, I'm just going to talk about how these things exist. Clearly, this doesn't exist in a vacuum. I didn't just invent all of this from scratch. A lot of this technology existed in the 70s and 80s in the form of these old knitting machines that got more and more sophisticated up until the mid-80s, I think, when they made less and less of them and they eventually stopped, and the industry eventually stopped, which was a real shame because they were getting very close to building what I want to build now, which is a completely autonomous machine. I have this, which was made on a machine in our local hack space, which is called MakeSpace. Thank you. I have this. The machine this was made on has most of the important components that I want from this. There's a duck by the back which is distracting me. Hello, I've got another audience member. Sorry, all you live streaming people, you won't see that. The technology has existed for a long time to be able to autonomously knit things. The problem with the machine this was made on was that all the bits just weren't connected together. It had the motor, it had the electronic needle selection, it had the computer and so on, but there wasn't an internet infrastructure to download and move designs. So it just wasn't a technology that was ready for its time. Now, of course, computers have got so much faster, more sophisticated and everybody's got internet connected devices coming out of their ears. It's far, far easier to design things like, it's far, far easier to make and share designs. So an autonomous machine like this will be far more powerful. It's also based heavily on the open-knit project which many of you may have already heard of. That's a project to build something very much like this, an open-source knitting machine essentially. The only, the problem with that I have, I mean a lot of this design of the machine I have here is based on the open-knits, it's based on open-knits work, but I want to really go beyond open-knit in terms of they only seem to want to make an old knitting machine that's capable of making woolly jumpers. I want to be able to make a knitting machine that can do everything. And with the weaving technology I've shown with the much higher resolution closely spaced needles, I believe it's possible to do that. Yeah, that's the end of my talk but I've left some time for questions in case you didn't understand any of what I said which is quite likely given how rambly I am. I believe we have a mic coming around for people. So can you explain how your stitch-making is different from a traditional knitting machine? The stitch-making isn't different not in terms of when I want to do a knit. It's the same process. Knitting's always been the same. The only difference is if I want to do a weave I have to use a different technique which involves a lot of loop transfers and things. But the knitting is essentially the same process, yeah. So how do you model the clothes like I'm thinking traditionally in 3D printers you use something like open-knits CAD or something how would you do that for this? Well, there are a few ways. It's actually not at the point yet where I've been able to create a fully-modeled cloth and it's still an early prototype I haven't got to the stage where I can get a fully-modeled item of clothing and put it in. But in terms of this for instance you would probably have some kind of XML file coding this as one loop and then these as two more and code that in. You could go as far as specifying where each stitch is supposed to be and telling it to stitch there, then stitch there, then stitch there. There's quite a few levels of abstraction you could go. You could be sort of complete machine code to very abstract. Just the general outline still to be mapped on to a person's body. Yeah. That looks like it's really two flat things joined at the edges. Can you do sort of three-dimensional shaping in a general way or is it really a flat thing kind of a structure? Kind of two flat things connected. The machine looks like this and so you see it's two flat beds opposing each other. So there's always an element of two flat things. But I think you can be a bit tricky with it. There are all sorts of tricks you can do to try and mitigate that by maybe making the seam diagonal rather than straight along the thing. Yeah. But there's no fundamental way. But that's the same problem you have together. So it's not a massive improvement but it's still a bit of an improvement because you don't have a seam anymore. Yeah. Thank you. Are you familiar with knit ML? No, I'm not. It's a knitting language based on XML. And the goal of this would be that it's a universal language so that I can write a pattern in knit ML and say in Brazil could and they wouldn't need to... I don't need to speak Portuguese and they don't need to speak English and we could exchange patterns, knitting patterns through this medium. And I'm thinking that might be a way to... That would be a good way of answering her question. Exactly. It's how you how you create a garment on that is that you need a programming language that speaks knitting. She's still a prototype. I haven't got to that stage yet. Are you on Ravelry? No, I'm not. We'll talk. He's so stupid. Yeah, go ahead. So, I do hand knitting. Have you thought about sleeve attachments? I'm afraid I don't do hand knitting. So it might be difficult to answer. Look at the clothes I'm wearing. This is obviously is a knit fabric but you do not want the stretch of a sleeve to be the same direction as the stretch of your body part. It needs to be down the sleeve. So, any thoughts yet? The short answer is no. But it's a very versatile machine. There's a lot of clever tricks you can do with it. It is different to hand knitting in quite a few ways. But in general you can usually get around quite a lot of stuff by doing things in unusual ways. We should talk. Did you have a question? Maybe you could give a more fully fleshed out example of one of the unusual ways. I can't imagine how you would make a shirt with a pocket because it has two cavities. Maybe there's another example. A shirt with a pocket is a good example. One way of doing that would be instead of knitting on all of the needles you only knit on half for each bed. You knit alternate needles, one knitting one knot each way. That increases the spacing between needles so you have to use a really low spacing to start with. Once you've done that you've got half the needles spare on each one. Once you get to the pocket you then knit forward back for your loop of the thing and then you go back to the needles where you want the pocket and you do another stitch just along the remaining needles that you didn't use to make the thing by only using half. Make sure it's joined up at the ends by perhaps combining alternate perhaps doing one row like this and then the next like that and then doing round on each one. That would mean that the pocket was connected to the whole thing and then keep doing that until you stop doing the pocket and then just go back to knitting the shirt from there on. I'm not sure if that's a very clear explanation it's mostly just me waving my hands but it is possible a pocket certainly. Any other questions? OK, cool.