EDGE in Tech: Advanced Manufacturing in the 21st Century

EDGE in Tech: Advanced Manufacturing in the 21st Century

Show Video

- The question here is how can IT balance economic and environmental sustainability in manufacturing? The huge strides in advanced manufacturing in the 21st century have led to complete paradigm shifts in product development and distribution. Technologies such as digital twins and 3D printing are now leading to operational restructuring and accelerated product development, raising a number of associated issues and implications. The panel here was organized and convened by Professor Grace Goo in Berkeley Engineering and the CITRIS team at UC Berkeley.

I understand that Grace is going to be ably represented by her graduate student Daniel Lim today. So with that, I will turn it over to Daniel to introduce the panelists. - Thank you Camille for the introduction. Hi everyone.

Unfortunately Professor Gu has a family conflict today and can no longer make it to this panel. So my name is Daniel Lim, and I'm currently Professor Gu's graduate student, and I'll be moderating the panel on behalf her, and I look forward to interacting with this community. And welcome to the Advanced Manufacturing in the 21st Century panel.

I'd like to start off by saying the landscape of manufacturing has progressed in the past decade. It is a very exciting time to study manufacturing from various angles. And today we are so fortunate to have amazing panelists from various industries sharing their lens of manufacturing, ranging from topics on AI, to virtual reality, to semiconductors. And we have total of four panelists who will be discussing their exciting and diverse areas of work, starting off with Paula Bielski.

And Paula has been with HP and Agilent for 24 years. She's a technology ambassador for her organizations, with knowledge and skill development focused on additive manufacturing, collaborative robot solutions, and adoption of relevant digital technologies. Paula, we look forward to your presentation. - Great. Thank you, Daniel, for the introduction.

As you heard, I work for Agilent Technologies. Manufacturing is a core competency in my organization. Agilent's manufacturing facility in Singapore has been recognized by the World Economic Forum as a Lighthouse.

Could we go to the next slide? This Lighthouse distinction is awarded to facilities that utilize multiple advanced manufacturing technologies. These Lighthouses are helping manufacturers around the world adopt the latest technologies through a shared learning journey. I'd also like to point out that fellow panelist, Andi Morey Peterson, also represents another Lighthouse recipient, Micron Technology. So you're in good hands with folks representing corporations at the forefront of implementing advanced manufacturing practices. Next slide please. So, when you think of advanced manufacturing in the 21st century, what are we actually talking about? I created a table here, and this represents a variety of categories and technology examples that are currently grouped into advanced manufacturing technologies.

It's not a fully comprehensive list; It's a little bit geared towards my own organization. And I didn't wanna talk about everything in the table, but I did want to point out that many of these technologies that are listed here are actually using other technologies to advance. Some examples are that cobots are using artificial intelligence and machine learning to improve their motion techniques, and 3D printing is using cobots to improve productivity. Other examples are that 3D printing is using augmented reality.

So there's lots of cross-functionality opportunities amongst all of these technologies. Each of these technologies is advancing very rapidly, and each will develop its own lifecycle and its own market leaders. The priorities of whichever organization you're a part of will actually dictate which emerging technologies make the most sense for investigations and investments. And the table of emerging technology opportunities will look different to every organization.

Next slide please. So I'm gonna take a couple of minutes to talk a little bit about additive manufacturing. When you think about additive manufacturing, the first thing that you might think of is a 3D printer.

But the process to produce, test, and qualify parts for manufacturing and production using 3D printing as a manufacturing tool actually takes a lot of effort. In well-established organizations, the very first obstacle that you are going to need to overcome is getting your design and manufacturing teams to accept 3D printing as a viable manufacturing option. Then you need to educate teams. When does it make sense to consider additive for part designs? And you'll also need to retrain your existing team for best practices in additive. Different 3D printing technologies are going to be better at different design challenges. So the next thing you need to do is to help folks understand which technologies are available and what each one of those technologies can achieve.

And in many cases, the part requirements will dictate raw material options, and raw material sometimes will dictate your printer selection. Seldom is a part taken directly off of a printer and used in production. So knowledge of available finishing processes, including support removal, achieving desired color, smoothing surfaces, machining, threading, joining parts through various bonding techniques, electro plating options, I mean, the list of finishing processes goes on and on. And additive manufacturing does not always mean that an organization needs to purchase a 3D printer. In-house fabrication is an option, but decisions to in-source or outsource would and should include raw material management processes, personnel requirements, and finishing needs.

No part is released to production without qualification and supply chain stabilization, and you would be surprised how daunting it can be to qualify new processes. No production process is complete without documentation. So it's going to be new for how do you document 3D printed parts, how do you measure the process to understand that it's within acceptable control limits, how do you quantify that the process is ready, and what will your change management process be? And this isn't even a fully comprehensive all of the challenges that you need to overcome.

CAD file preparation is a challenge, 3D printing process parameters would need to be optimized and documented, managing the waste generated by the process needs to be addressed, adherence to new and emerging industry standards needs to be taken into consideration. So the takeaway I want you to understand is that each emerging technology will come with new challenges, and it's never just as easy as buying a 3D printer. So anyone hoping to make a change in your organization will probably face that first challenge. Cultivate a change in mindset. It feels like that step should be easy. I mean, don't the advantages that these technologies offer just sell themselves? Unfortunately that's not always going to be the case, but I'd say don't give up, keep sharing your message, be persistent.

Change is never going to be easy, but the time and effort to adopt advanced manufacturing processes is worth it in the end. And now back over to Daniel. - Yeah, thank you, Paula, for sharing your exciting work on 3D printing.

Next we have Learie Hercules, who is a founder and CEO of Heft IQ. And Herc has been in deep tech most of his life across many industries, including big data, cybersecurity, smart cities, EdTech, and logistics. Herc has also worked across hardware, firmware, software, data engineering, data science, and cloud computing. He's a graduate of the University of the West Indies with a double major in computer science and management.

He currently leads a company focused on helping e-commerce brands grow by abstracting away the complex logistic layers and providing the advanced machine learning driven analytics. So Learie, we look forward for your presentation. - Sure, thank you for having me. Well, we could go to the next slide.

I wanna touch a little bit on smart manufacturing, just because the way the world was constructed, there's now a lot of overlap between different areas. So when we talk about manufacturing, it's actually a combination of a couple things. There's the physical devices, there's the components that make up those devices, there's the firmware that drives the components, and then there's a software application layer, and then there's some connectivity. And there's some type of processing that now occurs on a lot of devices. And so when you take a look at what occurs in manufacturing or device construction, it's really a combination of a set of things that's gonna be important to kind of understand each area as you specialize in specific areas.

So for example, a lot of devices now have embedded connectivity. A big piece of that was the 5G wireless standards that allow the assignment of way more IP addresses. And so we are able to assign IP addresses to devices and ensure that we have a higher throughput from those devices.

And so the whole domain around internet of things, IOT devices, came about. And good examples is you see a lot of medical devices now having embedded, whether it's Bluetooth, door locks, Bluetooth, et cetera. So in the construction of any type of device today, connectivity is gonna be a part of that, and there are a number of connectivity standards, like Bluetooth, ZigBee, BLE, LTE, et cetera. And so, in my own practical experience, we've deployed a number of computer vision towers within warehouses that contains its own connectivity devices. So there's less reliance on a wifi network that might be unstable, and it comes prebuilt with an LTE modem that you can now stream data at high speed to backend. Another piece is of computer vision.

So as we take a look at different technologies, we take a look at the application of the technologies. And so it's important to study problem domains, where technology could be applied, and understand that technology is a tool to solve problems. And so, the landscape of tools evolve, and it's evolving very quickly as well. Okay, next slide. So I touched a little bit about edge computing. Another example from in my use case is, within the fulfillment centers or warehouses, we deploy autonomous mobile robots.

Now, these robots need to be able to do work together with humans. So just beyond physically navigating a space in a warehouse and being able to pick a product from a shelf, there're humans walking around, and so there's avoidance navigation systems. And so one vendor's approach was, okay, we are going to train the mobile robot to detect humans.

And if they detect a human in the path, to stop. And what in fact occurred is productivity went down, because there's just a lot of humans, they're very unpredictable. So they had to go back in and train the humans to avoid the robots. And then productivity doubled. So the application of the technology is also important, and understanding the context in which it's gonna get deployed. On the edge computing side as well, there's a move towards pushing a lot of the processing that occurred in the cloud down to the device.

So back in the day, I would say 10 years ago, we would stream a lot of data into the public cloud, the Azure and AWS. We'd run machine learning models and predictions and push it back down to the edge device. And we had little control over the quality of their connection.

So wifi networks can be unpredictable. And so, there was a movement to push intelligence to the edge. How did that occur? Now there is a lot of containerization happening in edge devices.

So the Red Hat limit landscape you take a look at, now they're deploying containers and devices. So now the development could occur in docker containers and just LXE containers pushed to device. And then there's the deployment of machine learning models through platforms like TinyML, where you can train a model and deploy it in a low processing environment like a device, a Raspberry Pi. And so today, deploying ML at the edge and deploying an application on the edge, it's way simpler than it used to be. And now it's understanding what environment should it be deployed into to create some business value.

Next slide. And then we come to digital twin. Now, let's say in logistics, it's very difficult to understand the layout of a warehouse. You may be able to get a blueprint of the warehouse, but things change all the time. And then in modeling, what is the right mix of robots and humans, or machines and humans.

You need to be able to simulate an environment. And so there's this rise of both the device part of it, like Amazon has a digital twin IOT device framework. And then there's also the mapping piece of how do we instrument and simulate an environment so that when a scenario occurs, you can respond in real time in that simulated environment to predict what's gonna occur.

And so Nvidia, who's, I would say, one of the leaders in GPUs, have created this platform called Omniverse, that is allowing you to create a scenario. And this digital twin landscape, in terms of value, is projected to increase to a 125 billion by 2030. And so, you can see some of the use cases of the digital twin and practical applications. Next slide. Another use case for Omniverse, they've partnered with BMW to create some virtual factory planning. And what Omniverse is is a universal screen description framework, that you can describe a scenario and it constructs a virtual reality.

And the addressable market for that is projected to be $180 billion. So there's a lot of growth in these pieces where, as I said, manufacturing, if we think about it as beyond just the physical device being manufactured, in fact the Omniverse is being used to simulate a lot of hardware form factors as well. You take a look at a company like Digi, they're using a lot of simulations to understand the circuit layout and form factor as well. And so, it's the application of the technology that's important to study, the technology as well as the application of the technology, and understand what species is growing the fastest to kind of align your career towards. Thank you. - Yeah, thank you for sharing your insights on smart manufacturing.

It's a very exciting field that's very been fastly evolving. Next we have Andi Morey Peterson, who has been in the semiconductor industry since 2005, with most of those years she's been at Micron Technology. Andi started as a test probe engineer working on memory technologies such as DRAM, PSRAM, and 3D Cross Point, before going back to get her master's in data science from UC Berkeley. Currently she leads a team of process control system engineers and data scientists to help drive innovative approaches to manufacturing at the R&D facility for emerging memory and storage solutions.

She also is the lead of Micron Women's Leadership Network in Boise. Andi, we look forward for your presentation. - Well, thank you, Daniel, for that introduction. Like he said, I've been at Micron in the semiconductor industry for a long time now. But one of the reasons why I stay here is I'm really excited about the impact that technology can have on a sustainable future.

If you don't mind going to the next slide. So today I'm gonna talk at a high level on what we're doing at my company in particular to power, no pun intended, (chuckles) sustainable fabs for semiconductor manufacturing. Now, we're not just looking at greenhouse gases, but all areas of sustainability. We have aspirational goals to hit, focusing on emissions, but also energy, water, and waste. And while a major part of this work is collaborating with vendors and suppliers to make sure we're purchasing sustainable and ethically sourced tools and materials, but we also use a massive amount of data we collect to provide direction on how we can improve our current manufacturing process. Can go to the next slide.

What's (chuckles) a little ironic is we develop memory and storage solutions, like DRAM and NAND, which go into GPUs, SSDs, cloud computing, et cetera, but we enable our customers to use those tools, but we also use the power of big data through cutting edge technologies, such as what Learie was saying, with artificial intelligence, cloud computing, and intelligent edge connectivity. And we use these to leverage data analytics, smart controlled systems, predictive maintenance, and deep learning to improve automation, but most importantly, reduce energy use and emissions. As a result, just like Agilent, as explained by Paula, one of our fabs was named as a sustainability Lighthouse by the World Economic Forum. And what was exciting, it was the first front end semiconductor manufacturing fab in the world to achieve the honor.

And one of the reasons why it was selected is we were able to reduce overall resources to produce each gigabyte of NAND at that facility by over 50% in the last three years. Can go to the next slide, awesome. So, how do we build smarter fabs and leverage the trends that are driving demand for our products to build even more products? It should be no surprise that we generate massive amounts of data.

Our current fabs will analyze a half a million sensors, 200 million plus images, a quarter billion control points, and over 30 or 40 petabytes of data every day. And these numbers are just continue to explode. And we've been investing in, for nearly a decade, in AI/ML expertise, deep rooting platforms, such as GCP, and all of this is supposed to be generating impact at a scale. For example, sustainability. Faster yield ramps and lower costs. Now, these kinda make sense.

Capital expenditures are at the highest rates ever, so we want a return on our assets. Quality, our customers are demanding the highest quality. They are way less forgiving than they were in the past. And we wanna get to yield ramps faster just simply so we could make money. We are a for-profit company, of course. But sustainability, of course, is my passion.

And I wanna do an example of how we were able to achieve all of these. So we have an anomaly detection program, and we installed several acoustic sensors to various tools on our floor to learn the normal sounds of the semiconductor manufacturing production process. And while events are very rare, the algorithms have found a few events that would have caused major line down events.

So not only were we able to reduce cost by catching these major events early, we were also significantly able to reduce rework, which of course, in turn, reduces engineering time, but more importantly, energy and material consumption if we don't have to rework all those wafers. So in just this one program, we are able to achieve everything on that right hand side of the equation. We improve sustainability, we increase time to yield, we lower costs, and we improve the overall quality of our products. So here is our set of clear and bold aspirational goals that strengthen the link between environmental responsibility and business prosperity.

We're on track to meet our goals for a 42% absolute reduction in scope one emissions by 2030. We're also on track to have 100% of our energy purchase to run our facilities here in the US by 2025. We wanna reuse, recycle, and restore water by 75% in 2030, and reduce waste by 95% in that same timeframe.

But all of this is surrounded with the aspirational goals of reaching to 100% as much as we possibly can on all of these pillars. And we cannot do this without advanced analytics, real-time interdiction, and data science innovations that may have not yet been created yet. So I'm pretty excited for what the next few years have to hold for these technologies and all these aspirational goals, but I do believe that this paradigm shift will only be achieved through the work of the diverse set of talent such as yourselves. And I'll hand it back to you, Daniel.

- Thank you, Andi, for sharing your thoughts on leveraging the AI in the smart manufacturing. So next we have Xueying Zhao. Xueying Zhao has been with Lam Research for two and a half years.

The first 18 months, Xueying participated in Lam Research's rotation program, where she gained knowledge and skills in simulation, additive manufacturing, and hardware. While in the rotation program, she won Unlock Ideas Award and collaborated with our group on a project called Machine Learning Enabled Shower Head Design. Xueying's current role is a supplier engineer. And Xueying, we look forward for your presentation. - Thank you, Daniel, for the nice introduction.

Hi everyone, my name is Xueying. I am a product engineer or supplier engineer at Lam Research. So first I want to briefly introduce what Lam Research does. Lam Research is a company that sells equipment to manufacture semiconductor devices. We know that semiconductor devices play a very important role in our lives now. We use our phone and our computer every day for a long time.

And at the heart of every semiconductor product is a complex microchip. So Lam Research sells the equipment to make those chips. Can you please go to the next slide? Yeah, so in order to make those chips, we need to apply a lot of advanced technologies, like etching deposition. You also need to strip the photo resist and clean it off. So our products cover these areas.

My current role is a supplier engineer, so I interact with suppliers. My role does not involve working in the clean room myself, but I do have clean room experience from grad school. I have a PhD in material science and engineering. So when I was in grad school, I noticed that some of the semiconductor processes are not very environmentally friendly.

So here at Lam, we actually incorporate sustainability into the innovation strategy. So earlier, Andi just gave a very nice presentation about her passion in sustainability. Here I will share a few examples about Lam's initiatives in sustainability.

So we have virtual fabrication for semiconductor sustainability. I actually did one rotation in that area. And the benefits about virtual fabrication is that there's faster time to solution and commercialization. Because now there's less experimental trial and error, and that can save us a lot of time and resources.

So that's one way of sustainability reductions. And other examples at Lam are new system architecture and chamber design for enhanced energy and space efficiencies. And in particular, we have a very unique dry photo resist technology, featured in our Ether tool, that uses five to ten times less chemistry and two times less energy.

Can you please go to the next slide? Thanks. So yeah, last but not least, Lam Research is also accelerating innovation through the use of artificial intelligence and big data. I think that's one benefit about big companies that we should leverage. We're generating a lot of data every day. So as Daniel mentioned earlier, I actually worked on a project with his research group led by Professor Gu. The project is called Machine Learning Enabled Shower Head Design.

So that project is to evaluate the possibility of using machine learning algorithms for hardware design. And the machine learning algorithms developed by Daniel and his research group are five orders of magnitude faster than CFD methods. And this means, which would originally take CFD methods a week to accomplish now only takes six seconds. And our paper was recently published. So yeah, as you can see, this project is in line with the direction to accelerate innovation through the use of artificial intelligence and big data.

Yeah, I think this is the last slide that I have. I'll pass it back to Daniel. - Yeah, thank you Xueying, for sharing your exciting work on semiconductors. That concludes the presentations, and now we will go on the Q&A session. So I'll go with the first question on how can organizations ensure that they are measuring the right metrics to assess the impact of their sustainability and efficiency initiatives, and how can they use these metrics to drive continuous improvements? That was a long question, but yeah, anyone (chuckles) who can answer this, please go ahead. - Yeah, that is a good question, and it's a complex one because, at least here, we have a multitude of ways we collect that data.

We work, for example, with our water usage. We work really closely with the water companies here with on our state, but I think every facility is different. And they directly measure how much water is coming in, we measure how much water we're recycling, we measure how much water goes out, then it gets reprocessed and then goes back into the river as clean drinking water, et cetera.

Energy, some tools come in that automatically give us that information. Sometimes we have to go install meters to measure (chuckles) how much energy usage it's doing. And so, it's a complex where we're constantly trying to figure out how we can do better in collecting that data. Unlike regular process data, like Lam is one of our suppliers and we work with them, and some of the data that we get from them and other suppliers, we're getting it at a very, very fast cadence, where sustainability data is not so much. Maybe we just measure the energy use once every hour, or once every day. And so, it's very complex.

But we do filter all of that up into some dashboards and make goals surrounding every single tool, every single product, all of the tools that it used, what is the product's kind of carbon footprint. And then we make sure that the next generation of products will have even a less carbon footprint total, even though we're making more of it. So, (chuckles) that's very convoluted, but it's just a lot of different ways in which we measure that. - Yeah, thank you Andi. Any other panelists who wants to answer the question? Or we can move on to the next question.

Yeah, so next question is, it's a long question, so it's kinda can be divided into two. What are some challenges associated with implementing AI tools and smart control system in a manufacturing environment, and how can an organization ensure that these technologies are being used in an ethical and responsible way? - I can probably take that. - [Daniel] Thank you, Learie.

- In building AI models, a core piece is really understanding the use case problem to be solved, and the environment in which it's gonna be deployed, and who has access to the model, and how do we iterate the model and ensure that there is no model drift. So if we come back to the design of the model, a lot of the models depend on input and the integrity of the input. And so, from a product driven perspective, it's important to have the relevant stakeholders' input into what should be that training set, and the impact of some of the data that is gonna be used in that training set.

And we see erroneous use cases with some chat bots, that you can actually bias the training set to get to specific outcomes. So in terms of model deployments in real world, it's really around ensuring this integrity of the training set data by engaging the stakeholders. And then once the model is deployed, tracking the model drift, as well as ensuring that there's a ML. There's a whole field called ML ops, just like dev ops. It's putting ML models into operations.

So if you take a look at the ML ops platforms, there's a lot of platforms that track detail lineage, model drift, et cetera, to ensure that the integrity of the model remains as designed. - Thank you, Learie, for the answer. And the next question is from Miriam about what are some of the biggest challenges you face in moving towards more sustainable production? And I think Andi could answer this, or anyone else? - Yeah, sure. I think the biggest challenge is not gonna be surprising to anybody, because I think it's a challenge in anything that we're talking about today, and that's getting clean data that is dependable and reliable, and something you're able to action on once it goes into a model, or even just maybe even a dashboard.

Like I said, we're just getting this data from all these different, and we're piecemealing it together. It just doesn't magically kinda land on our feet and we just go in and we're like, "Oh, this is what we need to do." I mean, all of that background and getting that data up, all the data engineering is probably 90% of the work. Another 8% of the work is actually building the models, and 2% is actually actioning on the data. - Yeah, thank you Andi. And this is actually a follow up question to Andi.

Emmett asks, when you say that your aspiration for water and waste is 100% do you mean that your input is a 100% recyclable materials, or that your leftovers are 100% recyclable by someone else? - Both. So it's that we're just not throwing away anything as complete waste or water is getting not recycled back into either our use or back into the water systems in the city in which we operate. And so, that's both. And that also includes for waste, I don't know if it was just water. No, it said water and waste. Waste also includes leftover chemicals that we use in our manufacturing process.

Sometimes there are chemicals, and I cannot for the life of me remember which one, but there is one, for example, we'll manufacture with a certain chemical, we cannot reuse it again because it has particulates in it, but it is useful for other industries in their processes. And so we're trying to always figure out, okay, if we can't use it again because we have such high standards of what our incoming material quality must be, there are other industries that could use that again, because their standards don't need to be as high. So we'll either recycle it or give it to somebody else. - Thank you, Andi, for the wonderful answer. And this is a question from Arpad to, I think it's Learie.

And the question is, when designing and manufacturing for reliability, how do you model aging effects? For example, this part will still continue to function in 10 years, even under an accelerated product development timeline. That was a question. - Most components, hardware devices, have specs around meantime before failure. And there's a difference between a lab environment where a device is benchmarked versus the real world environment that the device actually lives in.

So what we would do is have partnerships in the environment that we intend to deploy this device into, that we can have test units in that live environment to ensure that the specifications don't degrade, so it doesn't degrade and it's unexpected. All devices will degrade in a new environment, because it's very difficult to hold the real world in inter constraints. And so, having lab specs is just the start. You have to literally put devices out into the real environment and monitor it on a continuous basis. There's really no guarantees. There's things that show up that are very unexpected.

You may have a very warm winter or very hot summer. So there's no way to control the environment, and so you have to deploy test devices into the real world and continually monitor those devices. - Thank you, Learie. Before we move on to any other questions, let's move on to more casual questions to the panelists.

And the question is, any advice for early career professionals that want to enter the field? Learie or Andi or, sorry, Paula, Xueying, could you guys answer? - When I hire, I look for intellectual curiosity and the ability to learn. There are some people who are very, the bulk of their self-esteem is on knowing. And technology moves so fast, within a year, things change.

And so the ability to learn demonstrates that your self-esteem is not so much tied into what you know, but just your ability to keep evolving that knowledge. And so for early career person, I would say it's very important to keep learning and master your craft, but also understand the application of technology in different domains. If you think about computer vision, someone might think about driverless cars. If you think about computer vision, there's whole security applications, there's a whole range of applications. And so, subscribe to blogs and read a lot, subscribe to the standard bodies. I personally read an hour in the morning, an hour in the evening, still, sustained over my career.

And expose yourself to new scenarios so that your brain becomes malleable enough to learn different things as things change. - Yeah, that was an excellent advice even for graduate student like me, just looking at different fields and how we can actually implement the technologies that we make. - Yes. - [Camille] Looks like Daniel may have frozen, so I'll give you one last question to wrap up the panel. If you were to use your crystal ball, what is your prediction about the next big thing for smart manufacturing? Just to close out your final remarks.

Let's go to Learie, Herc. - I think large language models, LLM, like ChatGPT, will fundamentally change how we approach the design of both devices, the manufacturing processes, and optimize the execution. And so, the timeline of concept to MVP is very short. And we have to embrace the technology, and experiment, and build things with it versus reading the news articles and analyst's perspectives. We have to get hands-on to understand how it's gonna actually impact the things that we do, and not just read about it. - [Camille] Thank you so much.

Paula. - Okay, I'll come off mute. (laughs) I guess I'm not very much of a fortune teller, but I guess my perspective is to embrace what's being released today. I think that there's a lot of opportunities. I feel like it's challenging to understand everything that's available to us. And one the things that I try to do is attend as many webinars as I can to learn that next level of what is this technology, what can this technology do.

And then I try to take pieces of that that I think could really be applicable in my business or in my organization, and then shop that around to our technical community just to try to bring visibility to what is out there today. I recently did an a presentation about augmented reality or virtual reality in production environments. And I usually focus on manufacturing, but in this case I was like, well, let me just think about all of the opportunities for my organization. And I was genuinely surprised by the number of people in field services, in marketing, in R&D, that are actively looking at augmented reality or virtual reality for their teams.

And so, the idea of making sure people are aware of what's available, some application examples and ideas. Things are just changing rapidly, and I find it very challenging to stay on top of everything. So whatever I can do to try to distill that down and find the stuff that I think is best for my organization and then communicate that with our technical teams. That's kind of my perspective. - Yeah, thank you. Xueying, any thoughts to share? - Yeah, so in the semiconductor industry, the companies try to pursue innovation to introduce new products more quickly and stay competitive.

And with each technology node, the cost would increase because we're trying to make the devices smaller and smaller. So I agree to what Paula just said, to be aware of what's available, because I think it would be kind of risky to just put all the eggs in one basket. For example, I do believe that 3D printing will be very important for the semiconductor industry in the years to come, but that's probably not the only thing we're gonna focus on. We also look at how we can incorporate the various technologies that are available, like big data machine learning, and see if there can be magic by combining those technologies. But yeah, from my experience, I'm fairly junior.

I do think 3D printing and machine learning would be very powerful techniques for the semiconductor industry. - Yeah, thank you so much. And that was it for our panel and the session. And Camille, I'll give back to you. - That's great.

Well thank you, Daniel, and thank you so much to the panelists for a really engaging discussion. I really appreciate your time and thoughtful responses and presentations.

2023-03-31 01:25

Show Video

Other news