Simplifying Generative Design for Everyday Use in the Design Process

MIKE SMELL: Welcome to Autodesk University. And welcome to simplifying generative design for everyday use in the design process. My name is Mike Smell, and I'm a senior product manager on the Fusion 360 team focused on generative design. And today, Rasa Lazarevic will be joining me.

RASA LAZAREVIC: Thanks, Mike. I'm a senior user experience designer, part of the generative design team.

MIKE SMELL: Thanks, Rasa. Today we're going to focus on a conversation around simplifying generative design in Fusion 360. And we're going to look to do that through talking about three main items. First, we're going to set a good foundation for your understanding of generative design, what is it, and when, why do you use it.

We'll also get a bit more practical and give you some best practices and tips for practically using generative design and finding success. And then lastly, Rasa and I will talk about some new approaches that we've been researching and developing against to further enhance the simplicity around using generative design in Fusion 360.

Let's start off by answering the question, what is generative design? Generative design in Fusion 360 is really about design exploration. We're trying to simultaneously generate multiple CAD-ready solutions based on real world manufacturing constraints, cost evaluation, and product performance requirements.

We'll take a closer look and go through each one of these components in close detail. Starting off, we'll talk about performance definition and objectives. And it's really important to understand that generative design is a performance driven exploration. So that means understanding how the part that you're designing will perform is of great importance.

We allow you to define performance in the context of the part that you're defining or you're trying to design itself. We can define those performance characteristics in an assembly context. And we can also then define the performance definition. Are we trying to minimize the mass and make the part as light as possible, or are we trying to increase the stiffness of the part. What is the safety factor that we're trying to achieve.

Do we have limits on the Moodle frequencies that the Department can experience so that it doesn't resonate, or do we have displacement limits or buckling safety factors that we need to achieve. The next aspect of generative design that we go off and base our explorations on is multiple materials and manufacturing methods. So what we're doing is generating shapes with both material and manufacturing process awareness.

And we do this for a number of different processes, additive manufacturing, die casting, 3 and 5 axis milling, 2 and 1/2 axis milling, and two axis cutting. This is a really important slide to pencil or to note down in your notebook. A lot of our users think that generative design is purely for additive manufacturing. And that's simply not true.

The next aspect of the solution is around our ability to provide cost insights related to the manufacturing process in the production volume that you're investigating. So you may find that as you're going through your design process, and exploring all the possible solutions, that depending on how many parts you make, it may be possible that one method is going to be better than the other from a cost point of view, as well as a performance point of view.

So we're trying to provide a vast array of insights to help you make engineering decisions. The last aspect of this is that we go off and do all of this exploration in parallel with the power of the cloud. We're giving you a number of different design ideas that you can compare and contrast to find the best solution for your needs.

And then the last part of this is after we've done our exploration, we've found a shape and a manufacturing method and the material that makes the most sense for our design objectives. We're giving you directly consumable outcomes that are fully editable inside of Fusion. Now if that's an additive outcome, or unrestricted outcome, or three or 5 axis outcome, those are your more organic shapes.

And those you will have access to the two supplying body that was used to generate that shape. And you can edit it with all of the native pea spline editing tools inside of Fusion 360. Similarly, if you're doing two axis cutting, or 2 and 1/2 axis machining, we will give you an outcome that's built up of a set of sketches and extrusions to ultimately define that product.

And again, because this is a native Fusion model, you can then edit those sketches or extrusions just like you made them from scratch inside of Fusion. When we put all of this together, generative design is really about driving better engineering. It's about enabling productivity in the design and engineering process so that you can take a much bigger look at the types of solutions that will best meet your design challenges.

Now that we understand what generative design is at its core. Let's talk about why and when we might want to use generative design. Starting with why. Again, this is really about us giving you a tool to better support your engineering processes. We know that most engineers are often faced with the challenge of striking the right balance between cost and performance, while meeting their design objectives, and their design timelines, and their design costs.

But what we also know is that engineers are typically then limited in the time and energy that they can spend on any design problem. They may be taking an existing design and making a few tweaks and calling that the new revision. What we're trying to do with generative design again, is around improving productivity and amplifying the amount of time and resource that you have by rapidly creating and exploring a number of new design options.

This puts your engineering and design teams in a position to rather than having to do the brute force exploration themselves, they let the software do the exploration themselves. And then they're using their expertise to go through the process of making trade off decisions, looking at price, performance, manufacturing, availability, material availability. So we really see this as a force multiplier to your engineering staffing.

Now let's talk about when to use generative design. And we'll look at a few common project categories. And one of the important things to point out is that generative design is not just for research and development. Again, we've got a number of our users who think that like generative design is not for them, because they're not really looking and doing a bunch of forward looking innovative types of designs. But that's not really the only time when generative design makes sense.

New product proposals is a great opportunity to leverage the power of generative design, to create new and creative ideas, find more concepts in less time, drive to predictable cost, because we're understanding the materials and manufacturing methods that we're going to be using, and drive to higher performance. We often hear from customers who've made the journey to move forward with generative design is that it's allowing them to solve challenges in a way that they wouldn't have thought of before.

And that takes me to our next point. Design evolution. Generative design helps you get beyond incremental improvements. It helps identify new ways to solve a design challenge, and it's really helping you drive to performance optimization. With the right inputs, generative design will help you create stronger, lighter, and stiffer parts. And then the last category of projects where we see generative design playing a great role that is not all always obvious, is in manufacturing support.

Oftentimes, folks look at generative design and they think Oh, what do I make that I want to make better? That's the thing that they typically sell. But we've seen that generative design has been extremely useful in helping you improve the performance, or reduce the cost associated with the things that you use on a daily basis to help you make the products that you sell.

So think about tooling, jigs, fixtures, assembly support devices, any of those types of things that are involved in the process of making the thing that you sell are also great opportunities to leverage generative design. It's also an opportunity to go off and say, well, what if I manufactured this product with additive manufacturing versus milling, or vise versa, or what if I were to cast this part. So being able to evaluate alternative methods and materials simultaneously is a great scenario where generative design makes a lot of sense.

The last one is also often overlooked with generative design. And it's around simplifying the assembly process. We've seen a number of scenarios where folks have taken more complex assemblies of rather simple parts, and leverage generative design to create a single, maybe more complex part, that replaces all of those individual parts. So when we think about the time to assemble a design, its ability to be repaired, having one part to deal with is sometimes easier than a number of parts that need to be dealt with.

Now let's take a look at a couple customer examples of where these categories start to show themselves. Volve is a customer that we worked with over the past year, and they're designing the world's first AI Racing drones. And they're leveraging generative design coupled with some of their proprietary AI algorithms to understand the performance requirements and then feed those into generative design. So that they can create the lightest stiffest drone bodies as possible to meet the needs of these racing applications.

So they're bringing a very, very new approach to product design to market so that they can achieve the absolute best performance for their products. The next one we'll talk about is Starling Cycles. You've seen we've done a number of stories this year with generative design in the bike industry. And starling cycles cannot be overlooked.

They came to us talking about how do we improve the efficiency of our frame building process, and also achieve weight reduction. And they started off the conversation with us is to say sometimes you stick with a design, because that's how it's always been done. But exactly where generative design plays a role is you can pretty easily look at very different ways to solve the same problem with a new approach.

So in this scenario, they were looking at the suspension pivot, saying, how do we change the way that we build this section of the frame. In the traditional approach, they were manually welding areas together, cutting them out, putting basic cylindrical components in there, loading it all together. And there was often times when they'd find issues with alignment, or performance, and it just took-- it was just a tedious process to ensure that there weren't errors.

And what they look to do with generative design is to say, can we make a much simpler component to assemble in the frame that still meets all of our performance requirements, that makes it much easier to control the precision of those suspension pivots. And this is something where by leveraging generative design to create this section of the geometry, they were able to reduce the time for frame assembly down from one hour down to 15 minutes, because of the simplicity of loading the tubes into this newly designed junction component, versus trying to manually weld all these pieces together to make an equivalent shape.

So again, really interesting approach to say we've got a known solution, but now we just want to solve that in a new way to improve our assembly process. And then the last story that we'll talk about here briefly is print city. So they're an additive manufacturing Bureau. And they have a machine-- one of their primary machines is doing FDM printing.

And they were facing some challenges with how that printer was performing, and the spool of material that needed to be managed. And they went off and looked at generative design to say, can we create a new spool design that would support the manufacturing process for all of the things that they were manufacturing with additive manufacturing.

So again, not necessarily the product that they sell or the product that they're producing for their customers, but an aspect of the overall system that does that manufacturing. The other interesting thing about this approach was while they were going down the route of additive manufacturing as the means to produce this frame that you see holding the spool, they actually didn't use additive manufacturing constraints and generative design.

They were looking for a much simpler design that was more cost effective to print, that required the least amount of support material as possible. And what they actually ended up with was a modular design of two axes and two 2 and 1/2 axis generative design components that got them to a much higher quality design in a shorter amount of time.

So this is a really interesting story for me because it dispels a lot of the typical myths that we hear of generative design is only for additive manufacturing. And if we're going to additively manufacture something we need to have these really exotic shapes. Print city actually did the opposite of that here.

Now I would encourage you, if you'd like to learn a bit more about these customers, please join our other generative design session called generative design, another year older, and another year better. My colleague Peter Champneys, from our technical consultant team who worked with all three of these customers goes into these stories in very, very great detail in that session. And I think there's a lot to be learned there.

Now that we've got a good foundation underneath of us for what is generative design, and when and why do we use it. I want to talk to you in a bit more practical sense about how do we use generative design effectively. Now I'd like to start off by giving you some simple questions to ask yourself to help you understand if generative design is ready for you.

Let's start with some really high level organizational type of goals. If you have goals around improving product performance, increasing productivity, and innovation capacity, winning more business, or reducing product cost or weight there is a good chance that generative design is going to be a great tool to support that process. Similarly, if we think about it at the design level, and we know that the design has load caring and performance requirements. We know that we're already evaluating performance requirements with simulation tools, looking at things like stresses, vibration, buckling safety factors, there's a good chance that those same designs can be improved by using generative design.

And if you have the potential to manufacture the part with a different material or manufacturing process, there's a good chance that generative design is going to be a good solution for you. Now, those are some pretty high level conversation points to help you identify whether or not generative design is right for you. Certainly, there are a few additional levels of detail that we need to look at in the types of materials you're using, the types of applications that you have, the types of manufacturing processes that you have, that would help us narrow in even a bit further on whether or not generative design is right for you.

But at the highest level, if you're trying to understand whether or not you should be considering this as a tool to support your design and engineering workflows. This is a great start. Now that we understand how to get started, I want to walk you through the workflow at a high level and then dig into some best practices for each step in the workflow.

It's important to understand that generative design is really about what you put in is what you get out. So we start off by defining the setup. The setup is made up of loads and performance information. It's also made up of the input geometry. Where do I need to connect pieces together, where do I need to carry the load, where do I need to carry the support, and then where can I not put material. Where does it need to interact with the rest of the system.

The next aspect of the workflow is letting this system, letting the generative design technology go off and evaluate your setup and start to generate the outcomes. And this is where we start the exploration process. We start looking at all the alternatives that are generated. And we can start to evaluate what makes sense. Do we want the lightest design, do we want the cheapest design, do we want a specific manufacturing process. And we can then compare and do trade off analysis between these different outcomes to see how they meet all of our requirements.

And then the last aspect of working through the generative design workflow is consuming the design. So here we're going to identify the best solutions from the set that we've explored. We're going to turn those into usable Fusion geometry that can then be edited, refined, tweaked, validated with simulation, tool paths generated for, et cetera.

So that's the high level workflow. And now we'll get into some of the best practices for each of these stages. When we think about setup, understanding the functionality of the solution that you're trying to create is key. I cannot stress this enough. First off, understanding our preserve geometry and understanding that represents your mounting and connection locations will get you going in the right direction.

There's a key point here, though. We should try to avoid small or what I'll call thin dimensions as compared to the overall design space. Now at a general level, when we see what generative design is doing, we're typically working in an open design space like you see in the image, and we're adding material between the connections as it makes sense to build the design that we're after.

One of the common misconceptions is that I should put in the least amount of material as possible. And that's where this concept of thin dimensions comes in. While Yes, generative design will thicken it, if the geometry is so thin the algorithms will have trouble even resolving and recognizing that geometry exists in the context of the overall design space, and we won't be able to produce good results.

The next thing that's often overlooked is understanding how best to work with obstacle geometry. Obstacle geometry should consider the range of motion if we're working in an assembly context, we should also look at clearances for fasteners, or tooling if we are going to go in and assemble this thing. The last thing we want to do is create a really exotic and interesting design, and then realize that we can't put it into the rest of the assembly because there's no clearance for fasteners or tooling to go through that assembly process.

Another common thing that we see users stumble with is starting shapes. A starting shape is going to be required if the obstacles block the line of sight between the preserves. So in the image here, you see that the green parts can pretty freely see each other as they look from one green part to the next. However, if we had larger obstacles creating obstructions between those we would want to add in the starting shape to help the solvers understand those paths of connectivity at the beginning.

And then the last point and one of the most critical in this slide is that loads and constraints should reflect your desired functionality and performance. Unrealistic loads, materials, or objectives will produce unexpected and likely undesirable results. So a lot of times we see folks trying to use generative design for applications where the ability to carry a load is not really critical.

They're purely just looking to create some interesting shapes or exotic looking things where the loads don't really matter. And they're trying to arbitrarily define some loads and constraints to get a specific shape. While this is possible, this certainly isn't the prescribed way to use generative design. And you may find that there are some challenges in working that way. And this is actually something Rasa and I will be talking about later in the presentation as we're going to try to address this in a completely different way.

The next aspect is digging into exploration. It's important that you realize that your design and engineering expertise is critical to this exploration and evaluation process. It's important you think about evaluating multiple variables when you're reviewing your results. What's the overall shape quality and complexity? Do you like the way it looks? Does it look like. It may have issues in load scenarios that you didn't consider? What's the performance? What's the manufacturing process and cost?

These are all insights that the exploration process will give you that you can use your engineering and design expertise to make decisions on. The next one is around leveraging the recommendation engine. You'll see here in the image, the blue bar at the top of the Explore environment, where we identify for unique outcomes that best represent some mixture of critical parameters that we believe are important to selecting an outcome.

We make these parameters available to you and you can tweak them there from the slider icon right next to outcome filters. Where each of these heuristics can be controlled to say, I want designs that are lighter in weight, most expensive or at least expensive. So you can adjust the dials on what the recommendation engine is using to select those outcomes.

And then the last thing to point out, which again, is often overlooked by our users, is that intermediate iterations can provide viable solutions. So when we click on the outcome view, you notice that we have the slider bar at the bottom. That slider bar can be leveraged to see how the design has evolved over time. And many of those earlier iterations again, can be viable solutions.

I encourage you to look at the Favorites workflow, where you can tag those outcomes, where they will then be enabled for comparison with other final outcomes. We will then calculate cost information for them, we will create the design preview, you'll be able to use them. And again, in the comparison or the scatter plot views. So that they can be considered like the rest of the final outcomes in the set.

And then the last topic is around consumption I think it's important to stress the fact that detailed refinement and validation is still going to be required in most every case in working with generative design. So what does that mean? Eligibility of the outcomes is a key feature of what we've developed, and it enables you to refine or tweak the design. Something that generative design is not going to do is put on part numbers, or put on your brand logo.

All of that sort of stuff we expect you to be doing. There's probably other aesthetic features that you may want to control and the ability to tweak those are their. Outcomes should still be validated against expanded performance requirements as necessary. With generative design, we're looking at linear material behaviors. We're looking generally at linear static stress type of behaviors. As you saw, we talked about having the ability to target specific natural frequency limits. We have the ability to target displacement objectives. We have abilities to target buckling objectives.

However, if we have to validate for non-linear performance, or a drop test, or fatigue, these are things that we are still going to be using with your traditional validation workflows that you're used to doing with designs that you've already made up to this point with traditional methods. And then the last bullet point that often help our customers understand is that simulation results of a generative design component may show localized discrepancies in results versus what you see in the generative Design workspace.

Now this isn't a huge problem because our expectation and what we expect that you will see is that factor of safety limit that you've set or that displacement limit that you've set should not be violated in a vast majority of the design. Where you may find issues is around the boundary conditions like you may see in traditionally designed parts that you're simulating. You may also see some slight discrepancies at the transitions from the generated geometry to the preserves.

It's important to realize that these sets of results be at the generative design results, and then the results that you create either in the Fusion simulation workspace, or in Ansys leveraging the workflows that we've built between the Fusion product and the Ansys products. They're working on different input bodies and using different mesh types. So if you think about what generative design starts with, you give it some preserves, we leverage the technology to fill up the design space, and we then take that mesh and convert it into a b-rep.

That b-rep then when it is simulated in another tool, is then fitted with a different mesh type. So small discrepancies may be possible. So in closing, when you think about the practical use for generative design, again, it's about providing you a tool for better engineering. It is a complement and not a replacement for any one tool that you have, or all of the tools that you may have.

Generative design is a tool to support engineering efficiency and design exploration. It's a tool that can augment, enhance your existing product development workflows with unbiased solutions. What does unbiased mean? Unbiased solutions means it may create solutions that you would have never thought of based on your tribal knowledge. And this is a really interesting aspect of what the tool is doing.

It's a tool that provides performance in manufacturing aware design options. It's a tool that provides insights for engineering decision making. And it's a tool that provides editable outcome geometry ready to be used in downstream workflows.

Generative design is not, however, a replacement for designers and engineers. I'm always amazed that the number of folks that feel like generative design is coming for their job, when that's simply not the case. We've talked about the importance of designer and engineering expertise in the decision making process. Generative design is not going to be a total replacement for your end to end product development workflow.

It's not going to create drawings today. It's not going to do detailed design, like adding part numbers and logos. Many of those things. As we said, it's not a replacement for detailed simulation and validation workflows. So again, you may have a battery of tests that a design needs to survive before it can go into production. And not all of those are going to be addressed in generative design. And lastly, it's not a replacement for a designer or an engineer's judgment and expertise in understanding how a design may meet the needs of your applications.

So with that, Rasa and I now want to talk about some of the new approaches that we're working on to further drive simplicity and generative design. Rasa, take it away.

RASA LAZAREVIC: Thanks, Mike. Awesome. And I'll be talking about the new approaches for simplicity. So at Autodesk, we're constantly trying to do as much customer research as possible, and make sure it's the best practice in the way that we work. So we do customer surveys, interviews, as well as customer visits. This year in particular, we did a lot of surveys and interviews, we want to understand how customers are using during design today, whether they do use it, customers that haven't used it, why they aren't using it today, and how do we encourage a better experience for our users.

So through this, we found some interesting customer insights. When we realized for customers that they felt that it was too expensive initially. That the outcomes are difficult to use. As Mike mentioned there is this perception that it's mainly targeting 3D printing only, and doesn't support traditional manufacturing processes.

The customer's also mentioned that it feels too much like a simulation tool and they also don't have time to learn it. And what we did is we looked through this whole feedback, and we started to think to ourselves how we can make improvements. So we to think ourselves how might we make it more affordable for our users. So this is the original pricing structure that we had today. Where we have a flex access model, where it cost 25 cloud credits per generate, meaning once you've completed the setup you click Generate to pay 25 credits.

And once you get into the Explore environment, it would cost you 100 cloud credits per outcome to get to make it exportable and use it in the design environment. We also have a subscription model, where it costs $8,000 for the annual subscription for unlimited usage as well as $1,000 for monthly usage.

What we then did is try to create a new pricing strategy. And this is the current one where we made sure that it would cost only 33 cloud credits for the flex axis once you've created your setup, meaning that once you get to explore environment there'll be no charge for these outcomes. You can explore all the possibilities, select as many outcomes as possible, and it won't cost anything.

And it also includes all the older outcomes that you've ran previously with the original pricing structure. And we've also made significant changes in the subscription model, where not only costs 1,600 a month-- sorry, annually. For unlimited usage and only $200 a month. So we try to make it as best pricing as possible.

So what we wanted to do is also recommend the best access for this usage, because we have flex, as well as subscription. What we did is when we spoke to customers, we realized that the best way to do it is if you're working with four studies-- more than four studies a month, we recommend the subscription model. And if you're doing less than that, we recommend flex access for the usage.

So once we now come back to our feedback that we receive from customers. We were made sure that we try to resolve the first issues that we spoke about the expense issues, but we also wanted to look at how do we make it clear that we also have these traditional manufacturing processes in our experience. So as Mike mentioned earlier on, is that we provide all these multiple material selection, as well as the different manufacturing methods. And we're constantly trying to improve these methods and also add new ones in the future.

So we want to make sure that all our customers are aware that we do provide these tools in our software. And with the last two points, we started to think to ourselves-- next slide. We started to think of ourselves, how can we make this tool feel less like a simulation tool. We also realize that customers didn't have time to learn it. And we heard a couple of interesting insights about the general design today allows customers to get inspired by new ideas at the early stage of the design process.

And we set ourselves new goals once we started to unravel these key findings. So we thought to ourselves, how might we make GHD a daily go to functionality part of a core tool in the design environment, as well as make you feel more familiar for design engineers, and provide real time feedback as our users are using the command in their workflow. So what we then did after a lengthy period of 10 months of constant iteration, as well as broad range of collaboration, a lot of work. We were able to come up with generative design for modeling.

And I'm going to hand it over to Mike to do the demo.

MIKE SMELL: Thanks, Rasa. So as Rasa pointed out, we set ourselves out some new goals to say, how can we broaden the envelope of what generative design is. How can we leverage the technology to create some new workflows that didn't exist in the current form of generative design today.

And today, we're announcing this project, generative design for modeling, that we hope will do just that. So I want to start by showing you where we are with some working prototype code. And then we'll talk about the details of what this project is all about. So here you see we've created a new generate panel on the design ribbon.

We're going through and starting the command and picking faces to connect. We give you the option to define bodies to avoid. And then we're going to go off and generate a number of alternatives for how we could create a connection between these locations in space.

So if you look here at the lower half of the dialogue, you'll see a number of solutions being generated in parallel. All of this leveraging generative design technology on the cloud. You'll see each of those shapes as they start to come back they look a little bit different. You also notice that we've got icons in there, and top half those are-- you see the sketch icon. In the bottom half, you see the t-spline form icon.

And that indicates that not only are we taking different strategies for how we create the shapes and what the shapes look like, we're also taking different strategies for how we construct these shapes and provide adaptability. So here as we go through the different alternatives, we'll go ahead and select one of the sketch based alternatives. And you'll see that this will build out a fully featured Fusion timeline with all the sketches, lofts, extrude, combine features used to create this single part.

I can edit the generative modeling command, come back to all my alternatives, and maybe I want to choose a different more smooth looking shape that's based on the form body. Here you can see that quickly updated and put the new body in the Canvas, now the important thing to take away here is this is a native Fusion model at this point.

At this stage in our development, you'll notice that we don't have very crisp end conditions, but we've modeled those in and we can leverage the combined tool to go ahead and join those into the body, just like any other Fusion modeling workflow. Where we're continuing to progressively add detail to the design as it makes sense. So here we can isolate that body, look at what we've created.

And now we start the process to understand is this something that can be useful in our design. So here I'm going to derive the part out into a new file. I'm going to go through the process of doing some basic simulation to see, do I need to come back and edit it. Is it directionally right. Should I be pursuing other alternatives.

So again, stressing the point here that the things coming out of the generative modeling command as part of the generative design for modeling project, we're building fully rich even models. And what you'll also notice is that with very minimal setup, we've went off and created some unique shapes to fill a space. And this is intended to fit early in the design process. Where we're not really burdened or even at the point of starting to think about what are the loads? What are the manufacturing processes.

How do I need to-- how much should it cost. So we're just trying to fill that space envelope to see what types of shapes we should pursue to solve our design challenge. So in a bit more detail generative design for modeling is about expanding the accessibility of generative design. We know that we've got a really interesting and powerful technology that's built for performance based design.

We want to also leverage that in a new way to further simplify how users can start to leverage the generative design process. So here, this new workflow is for rapidly exploring design alternatives, where we're primarily focusing on creating geometric connections. It feels like a modeling tool. It lives in the design workspace. It's super rapid and it requires minimal setup definition.

Just pick the faces and off you go. You will find that as I said in the demonstration, we're going to create multiple shape alternatives and construction methods. So if you want a sketch based native Fusion feature based design, not only can you get those, and you'll get multiple variations of those. And then similarly, if you're looking for swoopy or more t spline based shapes, you can absolutely get those. And you will get multiple variations of those.

The last thing that we're going to be doing with the generative design for modeling project is creating an on ramp between shape creation in the design workspace to performance in an objective based workflows in the generative design work space. So what you can imagine you will see in the future is you can start modeling very quickly leveraging the generative modeling command, getting shapes and designs that build the space requirements that you have, and then there will be a natural continuum of how you then start to further enrich that design. And then get to a more and more performing performance and manufacturing aware and optimize design leveraging more of the generative design technology in the generative design workspace.

So we see this being used in a number of different ways for basic face to face connections of complex shapes or things that are more singular in nature, like our generative design outcomes. Many of which here would look familiar to you. Where we're picking multiple faces in space all at once, and asking the technology to come up with a design solution for us.

So here you can see again, some more organic looking shapes that connects them makes some obvious connections. We've got some more basic shapes where we're just connecting between two faces. And then we also have some places where again, we're building designs based on multiple connections, but we're using a completely different strategy where we've got more prismatic Fusion feature based parts.

We will continue to mature this technology over time. So what you see today it will likely be much more mature and much more robust by the time that you see this thing available for you to touch it in the Fusion product. Now we're here having a conversation about simplifying generative design for daily use. And you might think, well now you guys have went off and built this other thing called generative design and how do they fit in with one another.

So there's a couple simple ways to look at this. Generative design as a whole is a technology stack of solutions that we use to go off and support design exploration. The generative design solution that has been in the market for a while is around optimizing for performance materials and manufacturing. Things are manufacturing process aware, and we're providing cost insights about those designs.

The outcomes from generative design for modeling or the generative modeling command that will eventually be in the Fusion Canvas it's about rapid design shape generation. It's used early in the design process. It's not slowed down by understanding all the performance and load requirements. It's not slowed down by having to understand all of the manufacturing processes that you may have access to. It's purely about building shapes that produce highly editable results in the timeline.

And this is really what we're trying to do here in the process of simplification is expanding the footprint of generative design so that the technology can be used by more folks in more applications. So what you'll see is generative design for modeling is going to be a design workspace-centric form-based design process that can go directly to downstream consumption. You can also start with generative design, like many of you have been doing for the past few years, where we're focused on performance objective based design.

Now, what I mentioned is that a part of the generative design for modeling project will be starting to build this on ramp where we're leveraging the form based information that we're creating in generative modeling. And using that as an input to generative design, where we can then add some enriching information to it and get an even better resulting design for downstream consumption.

So I'd like to turn it back over to Rasa to summarize some of the things we talked about today.

RASA LAZAREVIC: Thanks, Mike. So we've made major milestones as we've discussed in the whole presentation. And this is Thanks to internal collaboration, as well as customer feedback throughout the whole process. And what we're trying to do is we're going to constantly try to achieve this through the years time. And we'll try to make improvements to a solution and make the experiences seamless as possible.

So if you haven't used any design already, we do have a seven day trial. And we recommend that you go try it out. We also have many sample files and tutorials that will help you get started. And we really hope you use it to leverage generative design as best as possible.

Also, as we discussed, we do have the new generative design for modeling command. It's still in the early phase, but we're definitely opening it up for opportunity to have it to try it out. And if you're interested in trying this out, please get in touch through this email.

And lastly, contact us if you have any questions. We always open to feedback. And please feel free to reach out.

MIKE SMELL: Thanks so much Rasa. Great summary of what we've talked about today. And one of the things I want to stress is we highlighted a number of items today as it relates to understanding the fundamentals of generative design, how you can use it in the most practical ways, and get to success with it. And also looked at some of the actions that we've been taking over the past year to drive to simplicity.

There's a number of other things that we've done over the course of 2021 around not only simplifying generative design, but making it more accessible and more applicable for all users. I would encourage you to go check out another a--you class that I led with my other colleagues to talk about generative design another year older, another year better. Where we do a review of what everything we delivered in 2021.

We do a deep dive into some of the customer stories we talked about today. And we also are taking a look at some of the other projects that are in our future roadmap that will help expand how you can leverage generative design. So Thanks, again, for watching. And we hope you enjoy your Autodesk University 2021.