Welcome everyone to today's webinar. I have with me here Tony Pekay and later on we're going to have our special guest Michelle Wii West. Um thank you for joining all. So without further ado, Tony, why you don't introduce yourself and then I'll go after you. >> Sure. Hey folks, Tony Pike. I'm with Nvidia. Um super excited to be here Jose. I always like it when we have opportunities like this to uh have some discourse, talk about some interesting stuff. And today I'm super excited that we got a special guest to kind of bring all this together for us. >> Exactly. It's going to be very exciting conversation. And um from my side, my name is Joseep Dermillian. I'm a director head of worldwide AI sales here at NetApp. And today Tony and I will walk you through the NetApp and Nvidia partnership. We want to touch base. I know you are all waiting for Michelle to show up. But uh before that we will have a little bit of exciting you know conversations between Tony and I on AI and the partnership between NetApp and Nvidia and after that we will bring up Michelle here to have a great conversation on her experiences. So um without further ado thepartnership between NetApp and Nvidia actually is a longstanding partnership that started in 2018. Uh today we have hundreds of joint customers as well as we have many shared as Nvidia says MPM partners and NetApp says NAPM partners. So we have 20 plus channel partners together. Um we have joint DJX ready ISV partners as well. We have been working together with Nvidia on ideation workshops through our service delivery partners as well as Nvidia professional services team uh where we have done 20 plus workshops so far and then we have joint solution developments that we're going to walk you through in our presentation today. And last but not least, we have developed together industry use case based solutions in verticals like healthcare, life sciences, financial services, retail, automotive and etc. So with that I want to ask Tony what is the AI landscape in 2022 and can you give us an overview of what's going to happen this year from an AI perspective and from Nvidia's lanes? >> Yeah, it's a great question Jose. You know, if I think about why our companies are partnering in AI, it's really because we share a common mission, namely to democratize the power of purpose-built AI infrastructure really for every enterprise, not just the ultra large or the hyperscalers, but really for every business that can benefit from it. um around the world um we've seen businesses embrace this uh AI fueled transformation to enable the use of things like uh natural language processing and large language models to deliver superhuman conversational skills to enterprise applications that can improve customer experiences and loyalty. Um, businesses are building digital twins uh that use AI guided physics and logic to mirror the real world. Uh, enabling them to do everything from building tomorrow's factories today to reimagining healthcare. Uh to get there, we've enacted a strategy to make AI infrastructure more accessible to every business. Whether they need the best of breed leadership class uh AI center of excellence, if you will, or whether they need to support existing AI workloads on their existing IT infrastructure, um or whether they need to extend their AI platform to the edge of their enterprise. If you think about these three pillars of this enterprise IT transformation that's taking place, NetApp and Nvidia are partnering uniquely to enable all of these modes of transportation uh transformation such that every business can uh reap the benefits of AI and infuse it from end to end. So you know with that backdrop hose, you know, we've been working together for many years. Uh but I think that you know for folks listening in maybe hearing about this partnership for the first time uh you know how do you see NetApp's unique role and mission to democratize AI end to end for every enterprise.>> Yeah thank you Tony. Uh there you know there's there there's a lot of challenges today in the way of companies and enterprises implementing AI. One of the very big challenges is related to data actually and when we started this partnership in 2018 and onwards when AI started becoming more mainstream we realized that you know and that's not a secret that data is the one of the most important aspects of the whole equation but the problem with the data was the data being not available wherever data scientists are doing their work or data being scattered in different places or the difficulties in collecting the different data types from the different sources.So what NetApp brought into the equation is actually that whole data management aspect right and presenting that data or bringing that data closer to where the data scientist exists. So we are seeing a lot of customers starting their AI projects in the cloud and then enabling that and scaling it back to on premises or even collecting more data in the edge pushing that into the core. Now the news here is if you have more data and better data and qualified data your models aremuch better than training it on a smaller data sets right so that is where NetApp's you know value prop kicks inproviding this data fabric that expands from edge to core to cloud today with our verified solutions on premises that we're going to touch base later on we enable companies who want to start their AI projects in the core with Nvidia DGXs or other Nvidia GPUs in OEM based servers. We also enable those who want to start their AI projects in the cloud as a PC for example and then that could be any cloud provider out there. We are today available as NetApp the only company available in the three major hyperscalers as either uh um as either you know um a provided service a managed service or bring your own license uh type of implementation. So regardless of where you choose to deploy your AI projects, NetApp is helping customers across the spectrum. Those who want to do a cloud-based AI but they have um disabilities in moving the data across the hyperscalers or even creating that hybrid cloud AI implementations. We come in and we together with Nvidia with you guys uh we help enable those projects and scale those projects. That is the most important part. So if you look at that what I covered in the previous slide was improving model accuracy because of the data uh from anywhere and that is clean data right again we're talking about non-biased data and we are aware of all the different things happening today in the world of AI so when we're talking more data we're talking about the clean data but and think about it if you are for example um an autonomous driving company or you are creating models where um inhealthcare where have just acquired a new company. They have been using the cloud. You have been using on premises from an IT perspective. But to train your models, you need the data from these two companies together. And there is not a simple mechanism to bring those different types of data, different data sources together. So NetApp comes in and helps you to accelerate those data management problems. Right? The second thing that I talked about and I want to reiterate is scaling your AI projects seamlessly between the clouds as well as on premises. Uh we're seeing again manypeople start their AI projects in the cloud. Um they do the roundtrip back to on premises or they create the hybrid um scale um cloud project. So for example, we recently just announced um NetApp FSX on AWS, right? So you can use SageMaker connected to NetApp FSX. your data leaves in a NetApp um provided data management system in the cloud which means you have the same operating system on premises. So if you want to scale that it is seamless because they are all working on the same operating system. The same benefits you gain on premises you gain in the cloud as well. The second thing that or the third thing is most importantly it is not only about the data movement. So once you apply NetApp into your AI projects, you gain tools like that help you to collaborate that work between the data scientists that reproduce the work between the data scientists and also help you to trace the models. So for example, using NetApp technology, we can help hundreds of data scientists to share the data that is needed to train their models on without the need to copy that data multiple times. And we see this happen. We even have customers leveraging this right now. We can provide instantaneous access to the same data to the golden source of the data without you going through thepain of managing that data at multiple places of getting the data out of control not having governance over the data as well as tracing those models especially if you're talking about responsible AI traceability is very important we help you version your code together with your data with one click and put that away if you want to come back to that model two years, three years from now, we can bring the data that was trained that was used to train that model so that you trace those models back for multiple purposes. And last but not least, with our joint reference architectures, with our joint solutions, we have a faster or we provide a faster ROI and a reduced total cost of ownership. We don't want you as data scientists or even IT people to figure out which cables to connect where to figure out how to present those data all the way to the Jupyter notebooks. We have mechanisms for you to do that. We have architected and referenced you know connectors that goes all the way to the native data science tools which means lessen time on configuring and support and installation and more time on data science um development and enablement. And you know we talked a lot about the value prop. How about customers? So you know we today for example we have a joint um a big retail customer that is leveraging both NetApp and Nvidia technologies in order to enable multiple different use cases. One of them is enabling their customers to try on makeups for example remotely. The other one is improving the customer satisfaction score uh when they are shopping online the customer recommendation systems and you know this retail store or this retail end uh company has 500 plus data scientists and they are using the previous mentioned netup technology in order to enable those 500 data scientists to work faster towards their data science projects share their data faster and share their data in a more governed way. Um Tony, I know you also have an example uh for us for a joint customer. Would you like to chime in? >> Yeah.And I love the retail example you just gave. You know, another one is uh we know of a pharmaceutical company uh doing drug discovery uh using this very kind of infrastructure. Uh as we know, AI is incredible at shrinking the timeline and development cost of bringing new drugs to market. Uh but in this instance uh we saw this company decide strategically to centralize all their different development projects on a centralized shared infrastructure in order to reduce costs uh share expertise uh and best practices and of course uh speed the innovation cycle and achieve really a faster ROI uh using what we would call an AI center of excellence. is completely greenfield infrastructure that brings together people, process and platform. Uh so in this way uh this customer not only benefits from really the joint solution leveraging NVIDIA DGX with NetApp storage uh but they themselves are democratizing access to high performance AI development infrastructure to everyone internal development teams. >> I agree. I like that it is a really great example of how companies are accelerating their drug discovery and other research withAI. Um you know wealso have another uh great joint customer spare bank um out of Europe and they are leveraging NetApp and NVIDIA technologies to offer AI as a service. Now when you're offering AI as a service there's no single workload that uh that needs to be accommodated bythat IT infrastructure. But the challenge there was if you're offering this as an AI as a service, you might have someone doing natural language processing where they would need you know a transformer model working on a large batch or you can also have some you know um graphics and vision based model training where they are using deep neuronet networks for example that is a different type of characteristics from demand perspective from let's say storage andGPU based so you know sparebank is a very good example of someone who leverages the netup and Nvidia technology to offer uh broad services, multiple different workloads, multiple different verticals they hosted and actually because you have different types of customers that cloud connectivity is very important for someone who's offering this as an AI as a service so that they can connect with customers who have been whohas started their AI projects in the cloud or they want to connect it to this platform. So great examples and we said we have manymore uh customers jointly and we're always uh proud to help them accelerate their projects with AI. Uh with that Tonyum you know wementioned obviously three customers and many others but there is still a large you know population out there they have some problems with their AI projects and probably uh some of our listeners here on this call they're still struggling. What do you think the biggest uh you know uh inhibitating factors here from let's say an IT perspective or moving forward with their AI project? >> Yeah, that's a great question. I think there are a few things at play here that we typically find. One is we've seen that countless AI project teams embark on their own what I'd call DIY or do-it-yourself uh development platform when they couldn't find the right kind of compute and storage resources for AI workloads natively within their own IT shop. Um, this, you know, [snorts] we call it affectionately shadow AI, if you will, uh, has become rampant in a lot of enterprises, probably a lot of the folks who are dialing or joining us today. Um, and while it helps maybe small teams uh get their individual project off the ground, it hurts the enterprise at large in their endeavor to scale AI because of this decentralization effect that we see uh with these innovation teams forming silos. So this often gives rise to escalating costs and this brings me to the second problem which is rising cost. Namely, many businesses report that the opex associated with AI development is running out of control and this is above and beyond what they spend just hiring data science talent largely because of unfortunately cloud. Uh so while cloud has been great at um you know lowering the barrier to ex access accessing these kind of resources like GPUs uh and storage as project teams see their AI models uh and their data sets grow over time uh in complexity and size they also start seeing uh the cost of data movement and data storage grow in response and this effect um that we kind of call data gravity uh is really an inflection point where many businesses start to realize that having a hybridized infrastructure uh and a hybrid infrastructure strategy for development starts to make a lot of sense. So you can think of this as a combination of owning the base and renting the spike. you know bringing some amount of you know onprem or collocation located infrastructure that belongs to you for your ongoing steadyst state needs in combination with elastic spikes addressed in the cloud. There's really there's nothing wrong with starting in the cloud. It's just recognizing that if your data sets are created somewhere else then probably your IT platform and specifically your AI training should probably happen right next to it. Hence the need for this hybridized approach that can give you the balance of both and ultimately allow you to keep cost undercontrol and achieve the lowest cost per development run on your models. So, you know, Joseb, we've been doing a lot of work to simplify together how businesses can access the this kind of purpose-built platform that we've talked about that embraces hybrid, is best of breed in terms of performance and scale, and can be an operational standard really for any enterprise. So, maybe now is a good time to kind of explore these a little further. >> Exactly. And I know a lot of people are waiting for Michelle to come up here. Uh we're a few minutes away. I promise this is the last slide on our side. So um as Tony mentioned, we are together enabling enterprises to deploy their or to accelerate their AI projects through uh four different offerings. The one is the reference architecture that we uh developed with Nvidia almost four years ago. We are now in the process of getting the DJX superpot AI data center. Um it certification is in progress. We're going to have it that soon which basically tailored towards NLP and large um data set based AI trainings. And then we exclusively have the NetApp on topi integrated solution through our partners which means you buy the whole stack you don't do it your own you buy the whole stack verified architected installed and with a one support line for you to get support on the whole stack. um uh as well as from Nvidia andthe other exclusivity um offering that we have with Nvidia is actually the Nvidia DGX Foundry where it is a fully managed service uh based on again uh Nvidia's GPUs as well as NetApp's data management system fully connected to the cloud as well. What you get across all these offerings and I want to repeat that is the expertise or the you know world-class expertise from both Nvidia and NetApp. the um thejoint work that we have done in the past three four years uh ha have made us learn a lot in terms of how to deploy, how to configure, how to architect, how to fine-tune. So you don't have to go through all of that. So again you have options here. If you want to do it your own um you go to the NetApp on AI reference architecture. If you already got your DGXs and you want to expand that um youcan deploy that. If you're starting this fresh and you want a totally fully configured solution, um you go with the NetApp on topi integrated solution through our partners. If you don't have the capabilities to host this in your data centers and you want to do it as a managed as a service type of uh environment, uh you go with the NVIDIA DJX Foundry. And last but not least, if you're looking at a super pod, it's coming up very soon here. So don't forget to let us know about that. Now is the right time I guess Tony to bring in Michelle to you know we talked all about the good stuff here between Nvidia and Neta but let's hear it from her specificallyher experience on AI and how it helped her in her journey. So with that let's bring Michelle on stage. So Tony and I here together with Michelle.we west welcome to our conversation here. I know a lot of people don't need an introduction about you, but it would be great if you give us a short intro introduction about yourself.>> Yeah, thanks for having me on. I'm so excited to, you know, speak a little bit about AI and my career with all of you guys. My name is Michelle Wi West. I am an American pro golfer. Um, I won the US Open and I'm a mother to my 18month-old amazing daughter. >> So, thank you very much, Michelle, for being with us here. uh you know AI is a very big topic obviously for NetApp and Nvidia and it appears to be an important topic for you as well. Can you tell us a little bit why? >> Yes. um you know Ifeel like I kind of I grew up in that generation where half of our lives um a big part of it we didn't have social media we didn't have AI technology that was really present in our everyday lives and then you know more so like towards the later half um you know AI really started to pop up and then I started to learn about it utilize it and more specifically you know just a couple years ago I injuries were always a really big part of my life um And AI really has helped me to recover from my injury in a smarter way than I that I've ever done before. Um because, you know, before I utilized AI, everything was about guessing. Um you know, you do your physical therapy exercises with your PT and you know, you're guessing what your strength is, you're guessing what your rotation, you know, numbers are. But AI, with the help of AI, I've been able to um recover a lot faster, smarter. Um, and especially having a daughter, you know, you don't have, you know, hours and hours on ends toguess what your strength level is, what your, you know, rotation where you are in that recovery. Um, so AI's really help with that. >> Awesome. >> Awesome. >> Awesome. Fantastic. >> I know this is a music to your ears. >> Yeah, absolutely. Andit's such a pleasure to be here. Um Ihave so many questions, but the next one I think we'd like to ask you is um as you look at sport and the industry in general, do you see AI kind of being infused on a much broader basis in sport andhow uh professionals like yourself, you know, hone their craft? >> Yeah, I mean I already see it g um getting used a lot by sports gambling. um you know you see all those betting apps and you know that's why you know having the numbers having the statistics isso important for our tours I think sports organizations take that for granted you know um you know the P let's for example the PGA tour hasso much information on every player shots during each round you know the LPGA um unfortunately we don't have the funds to acquire that much information and without the information we can't utilize AI to really generate these exciting statistics for fans to you know either a gamble on but or b to follow. Um so I think AI is being really integrated in the gambling um sports gambling aspect of the industry but also um more personally I think it's going to really start toboom with um instruction um you know recovery uh training um nutrition um and you know with the stats as well just acquiring you know all the statistics and using the algorithm the AI algorithm toreally pinpoint what in their game is their strengths and their weaknesses. Um, and before, like I said, you know, we were just guessing, you know, I'd be like, "Oh, yeah. I think my driver feels good. I think my, you know, my putting stats feel okay." Um, you know, but withAI now, we're going to be able to really pinpoint exactly what your strengths and what your weaknesses are, which is going to be really cool. >> Awesome. So, basically, you're using more of the data available andhow do you So, piggybacking on that, Michelle, how do you think technology versus talent is playing a role here? Do you think someone using the technology will still need the talent? >> Oh yes.I mean I think you know in the game of golf um you know talent isvery important. I think what you're going to find is um the burn sorry excuse [clears throat] me the burnout rate might be lowered a little bit. Um because like I said, you know, in order to get good at golf, weknow that it takes hours and hours of practice, but you know, if you have hours and hours of blind practice, you know, that requires you to stay out there a lot longer, you know, you can have increased frustration. But if you have smarter practice with the help of technology, you know, you'll be able to get better faster. Um, I think you still need the talent as the baseline, but with the help of technology, you're going to see less injury. You're going to see less burnout rate, and you're going to see players get better and better on a, you know, on a higher rate, you know. Um, just to shift this to kind of apersonal level, you've had an incredible journey, uh, starting at an exceptionally young age and seeing success at an exceptionally young age, too. And uh you're a parent, you're raising a daughter, 18 months old, and uh I'mjust curious, how is this shaping how you,know, bring your own child into sport, whether that's golf oranything else for that matter?>> Well, we're a little bit too young right now to think about that, but you know, obviously I I've thought about it a lot. Um you know, my husband has thought about it a lot as well. um you know just more so because I have um gone through so many injuries. You know Ilove my career um every second of it. I don't regret you know turning pro when I was 16 and all that but you know Ido look back and the reason why I have arthritis in my hands is because Ihit so many golf balls before my bones you know really set in. you know, that's what the doctors finally told me. And I'm like, okay, wow. Like, I kind of wish I had known that. But also, if I had known that, would I have done anything differently? Idon't know. Um, but she even at 18 months, she's really eager to um to play golf and to try to dribble the basketball. Um, you know, we took her out, we always take her out on the range and, you know, recently she's been getting really,into it. Um, you know, my husband was hitting golf balls and she was just like staring at him and every time he made contact, she'd be like, "Oh, I got so excited." So, that's really exciting for us to see, you know, if she wants to be a pro alete, if she wants to be a professor, if she wants to do anything she wants, that's really up to her. But, um, you know, I think thecool thing about golf is that it's a really unique sport where hopefully, you know, I can play with her, she can play with her dad, she can play with her grandfather. Um, it's a multigenerational sport, so we're just excited to, you know, hang out with her. >> Well, she definitely gonna watch the videos of you swinging and playing, and I think that's going to be very exciting, right? exciting, right? exciting, right? >> Yeah. Um, Michelle, Iheard you saying one time that you can definitely not win agolf from the or the tournament from first day, but you can definitely lose it on the first day. you know we see this in the business a lot where people get frustrated very easily if they lose a you know in sales a deal or if they don't meet the customer requirements but can you tell us like what are the challenges youface and then how do you keep yourself the pace of focusing on the end goal right >> yeah I mean that that's a very important part um you know golf is a really hard game um and it seems like you know technology and AI and the business that you are in is also a very hard game. Um, and in golf specifically, you know, you can have a really bad hole. Um, and the most important part is what you do next. You know, you can make a bogey, but then if you continue to make a bogey, you're just digging yourself into a biggerhole. And even if you made five birdies the first day, youcan't win the tournament on the first day. But if you lose it and you make one bogey and that becomes five, six, seven, you've unfortunately lost the tournament on the first day. It'sreallyhard to come back from that. So, you know, you're going to hit speed bumps, you're going to hit obstacles, but the most important part isthe next step that's already done. You kind of have to forget about that and kind of pep yourself up and um forget about it and have a new start. you know, Ibelieve that you can have unlimited fresh starts, you know, and at any point you like, okay, fresh start over, you know, you don't run out of those. Um, and I think the moment that you run out of that, youreally are in trouble. Um, you know, I think the biggest obstacle for me personally was myself and the space in between my two ears. Um, you know, there was a lot of self-doubt, you know, lack of self-confidence. Um, you know, it may not seem like that on the screen, but definitely within that was mybiggest obstacle, my biggest hurdle in my career. Um, and it's still I'm still a work in progress. I'm still working on it. I don't have the answer for it, but for me, theunlimited fresh start at any time I'm like, okay, I'm starting over. This is it. Um, that for me really helped.>> Oh, that that's great perspective. Thank you. And I hope everyone watching this uh it is an important topic because again if you pigeon hole to that first lose oryou know the problem then you're going to lose the whole strategic um initiative right >> yeah you know Michelle um for those of us who are noviceses or worse uh when it comes to sports when we get into something like golfing there's this um this mental checklist right there's so many things that we are having to remind ourselves about and reinstruct ourselves about just to just make one shot. And it always fascinates me how people who operate at your level, so many of those things happen like autonomously or at least they seem to us they seem happen autonomously from maybe tens or hundreds of thousands of hours ofdoing this to perfection. But can you share some insight about that thought process that [clears throat] you know uh is not so obvious to the untrained but you literally go through this every time you're in a situation taking a shot and you're making all the mental calculations about how to approach it and how to execute. >> You know it's one of those things that when you put it that way um it seems really amazing but for me you know it is automated. You know, there's years and years of practice andreally the body is the most amazing, smartest mechanism. Um, you know, our body does so many things automatically every single day. Breathing, we didn't even think about it. Our body just does it naturally. you know, our body islike AI, you know, it has an algorithm within itself and it's learning every single day, you know, and with golf and, you know, any new activity that you do, you can't be discouraged because the first time is going to be reallyhard. Um, you know, I it'slike watching mydaughter eat for the first time like, okay, this is not difficult, but for you it is difficult. And then the more you do it and then you kind of forget and then now she's just eating a whole pizza by itself by herself. And um you know the body learns and you know with golf it just takes time and it takes patience and it takes grace and um you know there's a whole checklist in the beginning and I think that's really important toget through the checklist and not kind of fast forward yourself to the automation because if you skip that checklist and you start to create bad habits those are really hard to break. So it may take a little bit longer in the beginning, but to it's really important to nail down the foundation and go through that mental checklist and kind of think of, you know, like really baby steps. Um going at it slowly because you don't want to rush the beginning because then it'll get harder to kind of backtrack. Once you have bad habits, those are reallyhard to break. So it's a really important when you have aclean slate and you're starting something new to really start it the right way. Oh, thatis really fantastic, Michelle. Good to hear. And I'm assuming you're already setting up your next goals. So, when are we going to see you again on the pitch? >> Actually, you know, when this is going live, I'll be on the golf course playing in a tournament. >> Wow. >> Wow. >> Wow. >> Yeah, >> Yeah, >> Yeah, >> that's exciting. Are you excited about it?>> I'mvery excited. Uh nervous, stressed as well, but you know, that's when you know things are really important when you get nervous about things. So, I get excited about being nervous.Oh, okay. And then to that question, whatis your uh favorite golf course in the world now that you're back to the courses? >> Yeah. You know, I get asked that a lot and it's a really hard question to answer, but um it may sound cheesy, but I think Pinehurst number two is my favorite course. Um you know, for me, when the moment that I stepped foot on that golf course, I was just wow. I was blown away by all the historic moments that happened on every single hole and um it just I just really love the architecture and the whole vibe. Um so yeah, I have to say Pinus number two is my favorite. Michelle, you know, um, obviously there's a lot of fans who hail from your native Hawaii, uh, whofollow you intently. And, uh, I think one of the questions we got here, uh, was, uh, from a fan who lives in Kenna I have to say this. Um, Kenn Oh, boy. >> Kane. >> Kane. >> Kane. >> Kenohhe. [clears throat] Did I say that right? Koh. >> Yeah. >> Yeah. >> Yeah. >> Um, so that fan's asking, "What grind do you look forward to when returning home?" Um, I'm looking forward to having py and malasadas and lunch plates. Pretty much everything. Um, I'm so excited. I'm going home for the first time since the pandemic um soon. And I'm just going to eat everything.>> Can we deliver? Can we deliver some? [laughter]>> Yeah. >> Yeah. >> Yeah. >> I love the food in Hawaii. That's the best.>> Oh, there's nothing better. And um so back to being a parent uh and I know a lot of listeners here will probably uh have kids. Whatis your advice to the parents of young golfers today?>> Um you know when I started playing golf I wanted to play because I wanted to hang out with my parents. They were playing this couple's club and they were dropping me off with my friends and I'm like wait Iwanted to be out there with you. Um and I think you know all kids just want to hang out with their parents at the end of the day. Um, so really just making it a family activity. Um, you know, going out there and playing with your kids. Um, not just watching your kids and, you know, having them play just because you want them to be good at some good at golf, but really making it a family activity and just going out there and having fun. You know, I think a lot of times when you take the game a little bit too seriously, you forget that it's a game. So, you just want to keep reminding your kids that it's a game. It's a hard game. Um, it's a game that's full of life lessons, you know, full of mistakes, but, you know, you can learn from it and, um, at the end of the day, it's still a game. >> Imean, I don't have kids yet, but I'll take the advice. Uh, you know, appreciate it. >> You know, it's one of those things and just relating to what Joseph just said. Um, I I'm raising two daughters and I've had both of them in competitive sports and I've fallen into the trap of having, you know, big expectations. You know, we live vicariously through our kids. um you know from the child's end of it if you look back as far as you can you know your earliest days at what point um did you start to realize that this became something more than this is a great way to hang out with my parents and now I'm really good at this and I can take this really far. >> Yeah. I mean that happened really early on for me. Um, I was a very cocky little kid and um, I just knew I wanted to be a professional athlete. That was just what I wanted to be. I didn't care what sport. And then I slowly started to check things off my list. I was not very athletic in uh, baseball, soccer. I literally did everything. I wanted to be a tennis pro really badly. Found out I can't really run that fast, so I was like, "Okay, I'll just land on golf then." Um, so I started getting really serious about golf. I knew I just wanted to um be an athlete and I got really good um and I got reallygood and I was like okay this is like really happening. I was like [laughter] my dream is like really happening. So it happened really early for me. um you know that's obviously not you know traditional or happens a lot but you know I just had a really big passion for it and I just I don't know for some reason like I said I was a very cocky little kid and I just loved beating people that were older than me >> understandable that's spirit for sure >> exactly you know to my team listening to this be like Michelle you know [laughter]so um probably A last question here and um so what was the greatest sacrifice Michelle you had to make right to cross all this because in my Ibelieve that you know uh wins or success don't come easily and it requires sacrifices so if you want to pick one what was yours >> you know that's a great question um you know fortunately my parents um placed a very high value on education and having a social life. Um, so there weren't a lot of times where I had to, you know, choose between, you know, sacrifice myfriends, my social life for golf. Um, you know, Iwent to a normal high school. I, you know, went to Stanford even after I turned pro. Um, so I was veryum, fortunate in that way. I think the sacrifices really came from u and I guess I did you know sacrifice a lot but I didn't really think about it as sacrifice because I knew this is what I really wanted to do you know but there's one moment that I um I wish I had gotten back and it was a big sacrifice for me you know I was playing in a major tournament um and there was a week of my high school graduation um and I had to choose between one or the other and I obviously chose the tournament Um, but it was actually really cute because my high school dean um made a visit to Stanford and you know a couple of my friends. There was a lot of kids from my high school that went to you know my college and we kind of they did a little surprise and they all got together and um they ended up giving my diploma there. So I guess at the end of the day um you know it all worked out but you know that's like one moment where you know when I was a kid I was like oh man I really wish I'd gone to my high school graduation but you know I've been very fortunate in that way. Thank you, Michelle. That was amazing uh conversation and thanks for giving your perspective on your career as well as your parenthood. Um Tony, any last comments from your site? >> I really appreciated Michelle's kind of masterful insights of how AI is infusingitself into the lives of professional athletes like yourself and really eventually all of us. So, I think there's great to learn. >> Michelle, what are your 2022 predictions?my 22 predictions. Well, I don't have any predictions because I think um I stopped predicting the future a long time ago, but you know, I do wish everyone a very safe um 2022, a very healthy 2022. And you know, I guess a prediction Ipredict that you're going to see a lot of AI companies um come up and like you said, integrate start integrate AI into your everyday lives. And I don't think people are going to even notice it. >> No, I think this has been fascinating to also, you know, get some insight from Michelle on how AI is uh intersecting the lives of professionals like yourself andsport in general. So, I think there were a lot of really good takeaways from what you shared with us, Michelle.>> Thank you so much, >> Michelle. Good luck to your tournament and uh hopefully we'll see you with trophies next time. >> Yeah, thank you. I hope so too. [laughter]>> Yeah. >> Yeah. >> Yeah. >> Thanks Michelle. >> Thank you. >> Thanks everyone. >> So with that um I hope you enjoyed that session with Michelle. Uh we have got couple of questions from you here and we will pass it on to Michelle to get some answers and we will follow up definitely on that. Um we will end up the session here with couple of other questions we received and uh Tony this one is for you. There was a question uh that was saying how do you know when is the right time to move from cloud to on premises. >> Yeah, actually this comes up a lot uh because we've been talking about this hybridized world and uh kind of the duality right and um we see a uh kind of this journey that replicates itself in a lot of organizations that start out they start out in cloud and they got a lot of productive work especially in early prototyping done in cloud. um you know I'll just relay one example I remember with a customer where they're um building an image uh recognition application based on uh taking images building algorithms to detect things in those images and then presenting that as a service via an API to um customers who would basically want to use that as part of uh some other aspect of their business uh versus trying to build their own image cognition application. And uh in this particular environment, what the customer found was that as their data set grew, namely thousands to hundreds of thousands to millions of images, they started to hit this inflection point because the size of the data sets were growing exponentially and the complexity of the image recognition model was also growing in complexity. So what they found was as their developers were submitting jobs to their uh cloud hosted resources for instance to do a model training run as it were um a little bit of fear would start to creep into that cycle of development. Namely, if you hit the button and submit the job and then let's say a parameter or some setting was not correct or some aspect of the model or algorithm was off, you might find out two days later uh that you know the thing crashed or that the model didn't converge or you simply did not advance uh the success of your product and that's a high price to pay. So this fear of failure would creep in and it would inhibit developers. is they would spend a lot more time kind of curating the training run and less time experimenting because of that fear of failure. And that's what I kind of refer to as this inflection point when the cost associated with the uh development effort um starts to outpace people's ability to experiment freely. And when the development cycle is now constrained because you're fearful about the cost per training run, it probably means that for that project and for that effort, you need to shift the workload to something that's more of a fixed cost infrastructure so that you can regain confidence so that your developers can try more permutations because um essentially AI models are about um expressing kind of the creativity that's within the data scientist, the person who's working with the algorithms to ultimately build a model that has the highest predictive accuracy possible. But they can't get there if they have kind of some virtual handcuffs that are around the fearof failure, the fear of the cost or rising cost associated with experimentation. So that inflection point typically signals to most businesses when they need to move that workload into a more deterministic fixed cost infrastructure and that could be either within their data center or for a lot uh could be in a collocation facility or another managed service that has dedicated infrastructure that replicates what you might have had if you had your own data center. That's typically the journey that we see and why more and more enterprises are embracing this notion of hybrid and repatriating some of the more intense um ultra demanding side of their model development and craft work to this kind of infrastructure. So they have the best of both worlds if you will own the base and rent the spikes as it will. >> Thank you for that. And there's a follow-up question on that. What is the NVIDIA DJX foundry? And I think you already explained it, but you might need to a little bit um paint it. >> Yeah, we've been talking a lot about this especially, you know, our two companies uh because we're solving for a problem. Namely, if you look at some of the use cases that we talked about in the beginning, some of the world's most complex AI that um people are developing uh in mainstream enterprises, whether it's natural language processing or recommener systems or autonomous systems or whatever it is um more and more these kind of applications if they are part of the lifeblood of your enterprise if your enterprise is wrapped around enabling this kind of capability internally because you simply can't outsource because of intellectual property, data sovereignty, how mission critical it is, then you're looking at high performance infrastructure. You need the right kind of infrastructure that can take a model that traditionally might have taken months to train and you need to shrink that down to days or hours because your business depends on it. And for those kind of environments and those kind of situations, customers look to us and say, "Well, how do I get this kind of high performance infrastructure?" And honestly, I'm out of a data center. I don't want to spend a lot of capital on uh infrastructure. I simply want the high performance infrastructure experience without having to own it myself. And this is why we created DGX Foundry, um which NetApp is an intimate partner of, uh to give organizations the ability to have the cloud-like utility of this infrastructure without having to own it. So basically this infrastructure hosted in a collocation data centers like Equinex for instance and being able to rent that on a monthly basis for temporal needs to basically get that project over the finish line and then ultimately you know you can relinquish it or maybe think about you know maybe I have ongoing demand for this kind of thing and it's helped me justify building a business case for my own AI center of excellence as we talked about in the example of that pharmaceutical ical. So essentially, Foundry is about democratizing access to that kind of high performance leadership class infrastructure to any enterprise that wants to use it uh in this example as a rental offer without having to commit and they can commit when they're ready. >> Yeah. Um I think that'sa great explanation, Tony. Thank you for that. And there's uh one last question that I'll probably take this. How does NetApp contribute to the responsible AI initiative? Um I think you know we we've spoken about this when you are making a traceable AI that is you know responsible AI is a very big topic today. It doesn't only include the data it doesn't only include infrastructure it has people it has uh ethics etc. But if you think about it from a data perspective, for example, we have a autonomous driving car company who wanted to be responsible in terms of if they are putting a model out there into the streets and the car driving itself, they want to be responsible of anything happening towards that. Right? So basically they are using the technology that we provide in order to uh trace those models in order to version those models andput it aside for manyyears. And that's one of one area that NetApp is helping people to become or to develop more responsible AI initiatives. And uh there's going to be another blog post on this that's going to come alive uh soon here onnetapp.comaiand responsible AI. Um and you can also use that URL for all of our joint development and joint news. Um Tony, thank you very much for being here today. Uh Michelle, thank you very much for being here today again and uh we're looking forward to have you uh one more time with trophies as I said. Uh Tony, any last comments from your side? >> No, it's been a great session. I loved the discussion with Michelle. Great color and additional kind of outside in perspective from someone who doesn't necessarily live and breathe AI every day, but has felt the uh the impact firsthand in her own kind of personal journey. And then obviously um it's always fun to talk with you and with the uh audience here about the important work our companies are doing together to democratize AI infrastructure for their most important missions in their own enterprise.>> That's right. And again, last but not least, thank you all for joining today for this joint NetApp and NVIDIA webinar. Um we hope you have a great rest of your day and looking forward to seeing you again. Thank you very much. >> Thank you.