0:01 All right, hi everybody, my name is Daniel, and I’m very excited to be co-presenting today here with Marek Moffett from General Lattice. This presentation is in two parts, I’m G to talk for a bit about the metap side of things, and then Amer’s going to talk about the General Ltis side of things. And it’s, you know, the title was about deploying differentiated IP before the heat death of the universe, and that that is, you know, what what this was about. So General Ltis has this amazing IP for their digital materials platform, and Metafold brings, you know, we solve critical geometry problems in manufacturing, and we do this through our full solution development platform that handles design, simulation, automation, and commercialization. And, you know, our our role here was to maybe do better than the heat death of the universe, maybe, you know, a couple quarters. And also just to mention that this first started, this this collaboration, at this conference last year, in 2023. And so just to, you know, to emphasize how important these events are for for creating these types of projects and and work.

1:23 So maybe zooming out a little bit, in this day here in 2024, with a lot of, you know, software has been largely commoditized. There are endless clips about, he, how AI can produce software in minutes, autogenerated code, and so on, why build new software at all in 2024. And I I was really struck by this concept, the concept of a longtail need, this is not my idea, there’s an amazing article by Maggie Appleton, you should all read this, it’s fantastic. But the the concept is this, that if your needs as a user of software are common, are you, you know, maybe you’re doing the things that Solid Works and Rhino do really, really well, then the idea is don’t build software, use those tools, your needs are well met by software. But the story continues, especially when you change different software platforms together, and you get into what are called longtail needs, where only you really need this functionality, maybe, so the software does, you know, one thing for one person or group, and that’s, that’s a longtail need of software. And I truly believe this to be kind of the case for, I would say, like most people in in this room. And here the diagram shows the long tail is being kind of vanishingly thin, so from a commercial opportunity the diagram suggests it’s not great. But but I actually believe this is a, you know, a fat tale, that there’s, there’s always going to be more work in this area, that that the needs are increasingly uncommon.

So bringing it back to the project that we were working on, General Ltis had, you know, they were trying to deploy their digital materials library, and they had built out a proven, robust workflow with Rhino and Grasshopper, which, you know, also a shout out to Sphering for maybe the neatest Grasshopper definition on screen at this conference, that was, that was immaculate. And the the first problem here is that, you know, deploying a Rhino Grasshopper workflow is is hard, and it’s hard because desktop technology stacks do not translate to the web. You want to use the web for accessibility, for deployability, for ease of use, you want to really control how the the interface to the software works. So moving a desktop workflow to the web is is really difficult. So fortunately the Metap asset pipeline is something that does this, it translates geometry from one format into another, gets it onto our system, and we were able to recreate the Grasshopper workflow entirely, enely, with with our asset pipeline.

4:32 The second problem is, and you’ll see more of this in the second part of the presentation, is that the digital materials platform is one that leverages luses, and, you know, complex geometry, to achieve these amazing mechanical properties, and representing these as meshes does not scale. And it doesn’t scale, you know, partially from a computational point of view, but also just like a a download point of view, I, sort of, user experience point of view, you might wait a really long time to see the result if you’re waiting for a very large mesh file. So the Metap platform at its core is our own implicit geometry kernel, and we, also, that’s one part of our IP, and the second part is a geometry streaming technology that lets you transmit volumetric data from a server to a client in 2 to 500 milliseconds.

5:25 The third problem is, to create a final product using lattice geometry and sort of infill type structures, you need to, you will inevitably need to CSG, or, you know, union, these structures back with an original design space, and these are notoriously brittle operations to complete. The Metap geometry kernel, this has been talked about before, that the one of it shares with other implicit libraries, the concept of a never fail geometry operation, volume metric and implicit operations will never not work. They might not give you what you want, but they will never not work, and just having that functionality means that you can really build these automated workflows and get robust results.

6:18 So that was the, that was kind of like the glossy version of this collaboration. In reality, and I wanted to spend the slide to talk candidly about this, expectations in reality often didn’t quite line up. As we work together with the General Ltis team on the Metalold side, our expectation was that the geometry would be the hard part, in reality it was all the other stuff, it was the authentication system, the job system, just getting their devs talking to our devs on a development environment, and just overall API architecture. These were things that we hammered out and worked together on over the course of the project. A slightly more nuanced one, you know, the team at Metap were nothing if not math and geometry nerds, and so our expectation was that if we can build it then you should be able to just build it. In reality, you, documentation, training, examples, this this idea of of of knowledge transfer is so important on these projects, and this was such a critical part to empower the General Lis team and our other clients, is to to use our API effectively. Our document, documentation site, is gone through a whole overhaul, and it’s excellent, you should all check it out.

7:40 So as a result of this collaboration, the Metafold product has made huge strides. So whereas it took, you know, couple quarters to develop the functionality for General Ltis, which is a very nuanced application, we’ve been spinning up applications in two days. And so this is a demo app that takes scan data, I I could do the scan with a phone like Polycam or something, and then generate the silicone shield for radiation therapy on the web, couple clicks, any clinician can use this. And apart from what it does, you can build this in a day or two.

8:22 Another functionality that has come up a couple times yesterday is the concept of design of experiment. So Metap Fold offers a meshless simulation solver, it’s batch runnable in the cloud, so you can run hundreds of concurrent simulations. And we stood up a little design of experiment web app that, you know, configures all the different parameters, lets you visually inspect the structure, and then gives you back the full stress strain curve at 25% strain and some other metrics. Again, it’s not so much what it’s doing but how you can build this and integrate this for for your own needs, and this took us, you know, a day to stand up.

We quickly realized also, through the collaboration with General Ltis, that we need to reinvent the client side experience. So if there are any web devs in the house, you will probably be using 3JS, it’s a mess in there. And so we built this high performance client, it’s a WASM module, it look, it’s full, like it’s about under two megabytes in size, and it speaks directly to the volumetric data, and does some hybrid rendering as well. So you can get really nice results and create these, you know, configuration apps, and sort of really polished visual experience applications with the the Metalold graphics engine.

9:51 Using AI where it really works, I think, is another theme, and here we stood up a quick chat bot that can interact act with the job system, and it also interacts with the Metap geometry kernel, and lets you run things like a CNC check. So if you take the time to read through this text, you’ll see that it lists the models that this person has access to, and then it checks to see if you can CNC mill the teapot with a 5mim bit, and that’s a very specific example, but it’s not faked, so it actually looks at the volume loss having done the milling, and that volume loss is significant, so it concludes that this is not a good fit for CNC mailing. It can also do similarity checks between models that exist on the platform as well.

10:38 We’re ramping up our, you know, these these application development efforts, and so in the works right now is a plugin for Ansys, so getting implicit modeling working directly in Ansys Discovery, this is something that’s working very well right now and should save everyone a huge amount of time. We’re also doing some open source CFD work using the Palabos solver, this would result in an open source, like, Docker container, which you can run, like, on Google Cloud Batch or something like this, and this will, you know, really level up our ability to run these CFD solves. Bringing it back to General Ltis, these were applications that we stood up to demonstrate the speed of development, but in reality the a full application like the Frontier one is much more nuanced.

11:29 Yeah, so thanks Daniel, and justold is, and, know, talk about some of our other platform partners as well, from hardware to material to software, but every single one of the partners we work with is completely critical to our digital material platform. So I’ll just hop into it there, and hop into it with a meme here. And so at the highest level, what I want, we want Frontier to be thought about completely different from a lattice design soft ware. And so we want our users to feel like their person, the image on top here, not the person on the bottom, person on the bottom is, looks like me for the past six or seven years. But to better understand this, I I I want to explain what General Ltis considers a digital material, and we consider that an architectured structure with a combination of geometry, hardware, and the raw material itself. And and Frontiers aim to be a platform that has all of that data in a one stop shop.

12:33 So if you may ask, okay, well, why, why is that useful? So if we wanted to, say, take a given foam application, maybe a simple foam pad, a traditional application engineer, somewhere along that process is probably going to look like sketching to 3D modeling, quickly prototyping. But pretty early on in that stage they’re going, I have to start thinking about material properties of that application, and what and what they need to meet. And so what that process looks like for an application engineer is, pH pad, I’m going to go to a material database such as Matt Match or Syval, and I’m going to go and look up all of the raw material properties that I need to meet for my application. So this happens, probably happening right now, all the time.

13:22 So if you wanted to do this for additive manufacturing, if you wanted to integrate a meta material, aterial, something, not just looking at the raw material but adding a geometry layer, how, how do you go about that? As we’ve heard, there’s so many parameters, there’s so many different expertise rabbit holes that you can get lost in, is a gyro, what cell type you’re using, what thickness, what’s the density. And if you really think about this, and compare this to how a traditional application engineer does something today, the application engineer wanted to go and use a lattice structure, they’re going to have to go and collect all of those material properties like a Mat Match would be done today. And so that’s really what we want to recreate for Frontier, that is what Frontier is, is a digital material solutions platform. So you don’t have to do the design experiments, you don’t have to be an expert and spend tons of money to go and collect all of this data.

14:22 And so why doesn’t, why doesn’t everyone do this? And it’s an intersection of of several technical fields, 3D modeling, computer science, manufacturing, material science, and it goes on. And so really what we want to do is streamline that for the user. And so if you were to go and create a library of your own, which I know a lot of the OEMs may have, lots of studies that may exist on this part of the business or this part of the business, and even if you have have that data, how do you sort through it, and how do you show it to internally or to your customer that that’s something that’s actually useful. And so you’re dealing with all of the data, material property data, upfront, you’re dealing with software, so even if you have the raw material properties, can you actually design the shape, and not in just a puck, can you get that into a design space, and and ultimately, if you checked all those boxes, can you actually order more than 10 of those.

15:29 And so the idea is, what if these digital materials were accessible to everyone, what if that application engineer could go onto an online database, query based off what’s needed for their application, have an intuitive way to put that same property in their part, and then finally, not just download the part or, you know, figure out, pass it to their manager to figure out how they want to order that internally, can we take care of that final step as well, is supply chain production, and and get that that part in hand.

16:05 And so how we do this with Frontier, our three main pillars, and these main pillars are digital material library, our integrator, which is our a design tool, careful when I say design tool, the key of the integrator is to truly be a oneclick lising solution. The idea is that you’re not messing with the cell size, you’re not, you’re not messing with thickness, you truly get your result within your uploaded CAD bere design shape. And then, ultimately, once you get to an application that looks pretty good for testing, ordering, we have a supply chain built in. So, and from one all inone platform, you can make leaps and bounds in your new product development cycle.

16:56 And so just going into, in a bit of a detail here, of each of the pillars, the first pillar being the material library, validated mechanical properties, currently we’re physically testing these for a compression set of data. Our goal is to grow our database as wide as possible, so thermal, drop testing, everything under the sun is the idea. And something really exciting that Metap and Daniel was alluding to, their batch mass analysis, our internal system for collecting this data, we call Gloss, or General Lus Operating System, and being able to pair our operating system with real physically tested data, with new exciting analysis coming out, is really going to help us grow our datab base. On top of that are awesome partners from a material and hardware point of view, so you can see at the bottom, AM Forward, EOS, Photoc Centric, and and Form Labs are extremely critical to to making this platform form work, connecting all the industry leaders, all the material leaders, all in one spot, so the user doesn’t have to go and do all that work themselves.

18:07 So we’ll go into the integrator here, and again, this super exciting for us, and to go back and talk about Metaps dynamic pipeline, asset pipeline. Over the years we were doing lattice as a service for the past five or six years, that was me sweating on the the computer down there, and at the end of the day, really, what we came to was a really great solution, an automated way to actually create lates in your parts, but we had no way to scale that. So that’s where Metapo came in, and having the ability and all their APIs, and and their implicit kernel, allowed us to stand this up from an internal piece of software that was really powerful to get it out to the masses, and actually have people doing this, and not just one or two.

19:00 And so then three, again, just on the supply chain, working with the Wilson Ball, it was a really exciting project, and there was tons of challenges along the project, not even on the design side, which there were, there was also problems on supply chain, right? It’s very difficult, production parts are very difficult to get correct repeatability, how are you going to ensure when you get part A printed from one supplier and part B that they’re the same. So going and finding industry-leading partners that we can develop SOPs for for these intricate structures is our goal, to get repeatable production parts across the industry. And again, just highlighting all these partners across all all three of our pillars, so material, hardware, software, in in supply chain. And and what’s great and exciting to us is, we we have no business going and being able to do a material, hardware, all software, and all supply chain, and so we’re really trying to gather the industry together to have these leaders, and really have a platform that can give you a solution, so you’re not on your own from square one to to part in hand.

20:27 Yep, so just qu, quickly talk through this is just a quick demo of Frontier here. I’m prepping a file I got from GrabCAD, in Fusion, in Rhino, I’m going to upload this step file into Frontier. As we can see here, parts, I’ve gone and identified what digital materials meet the relevant properties for my application, going to add those to my folder, going to select what part I want to l in, slaps the lates in. And now what’s really exciting is, recently added a Boolean features, so now we can import step geometry from Fusion, from Rhino, we’re not trying to recreate the modeling process, that’s a very important key of this. If you don’t want you modeling features, model that in CAD, what’s good at it, and upload to Frontier. So if you prepare your model appropriately, with skins and added geometry, and you selected geometry you don’t want, you can get a very nice final result, all in the web, very quick.

21:38 This is showing our ordering system and download abilities, and ultimately can bring your geometry back into your CAD workflow. So here’s showing our lattice back in Fusion, and and back in Rino here. And what’s really exciting is, for the user, solid modeling and implicit modeling, it’s multiformat modeling, it’s very new, it’s very interesting, and until the, it’s the state is clear of how the industry goes, we’re super excited about supporting all of these file types, supporting implicit, supporting explicit geometry, and and allowing the user to ultimately have the most amount of decisions they want. And so being able to export as implicit data, as mesh data, and being able to import D data, allows the users to choose the correct data type for what for what they’re trying to do long term. This is just some mye candy here, bringing our our mesh back into a C4D render.

22:44 And so, yeah, looking forward for General Lattis, we’re really excited about growing our platform partner network, so anything under the sun, from material to hardware, the more partners we can have, hopefully the more applications we can support. And as added design functionality for the integrator, we’d like to add some more geometries, more unit cells, gradient functionality, different import options, so not just accepting step or VRE, and then expansion into rigid data sets. Right now we’re focused on foam replacement applications, we wanted to start small, focus somewhere, and ultimately start building this database into wider, wider data, data sets, to accept more and more applications. And I think that’s it from J, thank you. Thanks Daniel.