Latest / Technology Now / HPE Discover Barcelona: Analyzing Fidelma Russo's AI and sustainability keynote
Transcript
- Aubrey Lovell (00:01): Hello friends and welcome back to Technology Now, a weekly show from Hewlett Packard Enterprise, where we take what's happening in the world and explore how it's changing the way organizations are using technology. We're your hosts, Aubrey Lovell. Michael Bird (00:14): And Michael Bird. In this episode, we are not quite as live as last week, but we're still bringing you an interview from the floor of HP Discover in Barcelona. Looking at the keynote speech of Chief Technology Officer, Fidelma Russo, we'll be talking about what Fidelma's speech says about the company's direction in the coming months, alongside expert commentary and analysis from HP Chief Technologist, John Frey, who leads the company's practice on consumer collaboration when it comes to sustainability and IT efficiency. Aubrey Lovell (00:44): That's because Fidelma's keynote is keen to focus heavily on sustainability and making tech more efficient, helping both the planet and customers' bottom lines. Clearly it's not one to miss, so if you're the kind of person who needs to know why, what's going on in the world matters to your organization, this podcast is for you. And if you haven't yet, you know what I'm going to say, subscribe to your podcast app of choice so you don't miss out. All right people, let's get into it. (01:09): So as we explained last week, HPE Discover Barcelona is the company's major annual tech event for Europe, the Middle East and Africa. It welcomes thousands of staff, partners and customer delegates from all around the world, to meet, network, learn, and most importantly, talk tech. One of the dominant themes this year was sustainability. It's increasingly becoming a major concern for tech firms, both as they develop their ESG or environmental social and corporate governance policies. And of course, as they keep an eye on where the money flows out in their company, it's safe to say that IT infrastructure costs a bit to run. And reducing energy usage can make a huge difference. Michael Bird (01:49): And that's where today's guest, John Frey, comes in, as a go-to on all issues sustainability and energy saving. When it comes to IT infrastructure, he is perfectly placed to help explain Fidelma's keynote. I met him in Barcelona last week just after the keynote wrapped up. Fidelma Russo (02:03): Last year at Discover, we introduced the Sustainability Dashboard, and today I'm delighted to announce availability of the HPE Sustainability Insight Center. The next generation of the Sustainability Dashboard. Sustainability Insight Center has been integrated to give you the ability to observe and take action, not just on your infrastructure, VMs, containers and workloads, but also on your carbon consumption across multi-vendor and multi-cloud estates. This carbon insight capability is especially important when you're working with a workload like AI. The overarching theme is clear. AI is not just a technological leap, it's a cultural shift. It's a shift that impacts how we work, how we innovate, how we secure and share data, and how we run and transform our businesses. Before today, you may have thought artificial intelligence was going to solve all your problems, but in fact, it's clear you need a trusted partner who can help you solve the complex problems that AI brings. The HPE GreenLake cloud platform helps you manage, observe, and protect your hybrid infrastructure, data and models from edge-to-cloud. HPE GreenLake is the hybrid cloud for all. Michael Bird (03:36): Hi John. So Fidelma covered a lot of ground. What were the key moments for you? John Frey (03:43): Yeah, the key moments for me were of course the launch of HPE Sustainability Insight Center. That's the Sustainability Dashboard we've been talking about for over a year, that actually started as a discussion on one of my staff meeting calls. So it's great to see it come to life. And my colleague, Tiffany Jernigan, who won an HPE internal innovation competition with that idea, is delighted as well to see it finally come and give customers those resources. Michael Bird (04:12): So sustainability, as you said, was a big part of this keynote. Where do you think the next steps will be in going from a speech about more efficient IT to making it a reality? John Frey (04:22): Well, there's a variety of places. One is, as Fidelma noted, we really have to look at the data, the data we use to train these models, the data they collect because there's a huge environmental footprint from that data and the energy consumption to run the infrastructure to get to those insights. A second piece of that is really we need to do a better job of our applications and making sure that the application and the hardware used to train and to do the inference, really come together and are optimized. And in fact, just in the last month, we've released our software efficiency white paper for customers to use, to help them along that journey. And then really it's to pay attention to make sure for the environmental footprint that the solution causes, that we get much more benefit than that. So we make a good decision, is AI the right tool to solve this challenge? Michael Bird (05:16): As someone who's on the ground with this stuff every day, how do you think customers will take these announcements? John Frey (05:22): I think customers are excited. Part of the thing that we find with customers is they're intrigued about AI. They recognize the promise that AI might provide, but they're really unclear both, "How do we get access to these tools?" If they're a smaller customer, so HPE GreenLake for large language models, for example, gives them those capabilities that they didn't have before. I think they're excited that we have a point of view across the company, around the infrastructure, around the training models, around the ethical components, so we can help them avoid a lot of unintended consequences. And then they're frankly excited to see how this can help their business perform better and help their customers. Michael Bird (06:03): Why is it important for sustainability to form part of IT infrastructure planning and investment? John Frey (06:09): Well, it's got to be an integral part of digital transformation, and in fact, we so commonly found it missing that we wrote a customer workbook called, Six steps for developing a sustainable IT strategy, to help walk customers through how to integrate this. It lowers their cost, it drives up their efficiency, it makes their infrastructure more robust. And interestingly enough in this time where all of them, and we heard that in the keynote, are having trouble getting IT technology staff, it helps them improve employee attraction and retention as well. Michael Bird (06:44): So, I'm going to finish on the same question we ask after Antonio's keynote to Matt in the last episode. What would be your elevator pitch for what was said today? John Frey (06:52): Yeah, my elevator pitch is, AI is pretty complicated, but it has so many great benefits. You need to partner with a company that will help you both understand the opportunities, the infrastructure and solutions required and will help you prevent those unintended consequences. Michael Bird (07:08): So John, one of the things Fidelma talked about was how much energy it inference uses. Fidelma Russo (07:14): Carbon insight capability is especially important when you're working with a workload like AI. The immense computational power required for training large language models, often involving vast data sets and complex algorithms, contributes substantially to energy consumption. The same is true for inference as once these models are put into production and they start replying to queries, large amounts of energy are again consumed. Let's just take a look at one frequently used LLM. It's estimated that 1,200 megawatt hours are used to train the model and then further tuning results in even more energy being consumed. And it's estimated that in one day alone, the model consumes 250 megawatt hours, which is actually equivalent to 10 football matches being played at FC Barcelona at Camp Nou. Michael Bird (08:14): So, the explosion that we're going to be seeing in the use of AI across organizations, there's also potentially going to be an explosion in the use of energy consumption. John Frey (08:23): We think so. Michael Bird (08:23): Yeah. John Frey (08:24): And I think the key message is, AI does have a pretty high energy burden both for training and for the inference phase as well. So again, we want to make sure the challenge we're solving or the insights we're looking for, have enough value to account for that environmental footprint. Michael Bird (08:41): Wow, that is so quite mindblowing, isn't it? Because I think AI does feel to a user like it's effectively free, doesn't it? You go onto these generative AI tools, you type in your request and it just happens magically and you don't even think about the energy usage, do you? John Frey (09:00): Yeah, absolutely and it- Michael Bird (09:01): There's a cost there. John Frey (09:02): ... The user, it's on the web. If they put in a question, it will give them an answer so they don't see the environmental consequences, but know that there's a tremendous amount of horsepower and compute power behind answering those questions that come at an environmental cost. And so part of that challenge is, do we get more benefit as a result of the answer than it costs us from an environmental consequence perspective? Aubrey Lovell (09:28): Great to hear from Fidelma and John again on the podcast, and we'll be back with Michael and John in a moment, so don't go anywhere. All right. It's time for Today I Learned, the part of the show where we take a look at something happening in the world, we think you should know about. Michael Bird (09:44): And since I'm back in the studio, I can take this one Aubrey. Aubrey Lovell (09:46): Yay. Michael Bird (09:48): Yay. So if you're anything like most five-year-olds, you love dinosaurs and you love robots, right Aubrey? Aubrey Lovell (09:54): Yeah, sure. Michael Bird (09:56): Well, how about combining the two and creating a whole new offshoot of engineering in the process? So a group of mechanical engineers in the US have been working with paleontologists in Poland and Spain to create a robotic model of a marine organism that existed 450 million years ago. Here we go, pleurocystitid. Try saying that five times quickly, is believed to be one of the first creatures to have moved using what we now know as muscles. The team used soft biotics, which is robotics with soft materials and flexible electronics to create the movement of the planet's earliest organisms to better understand how life evolved on earth. (10:39): They even gave the robot a cute little name, Rhombot, that's after the family the organism belonged to, which is lovely. Now they are calling this new field of study paleobionics, which sounds pretty cool. And by studying the simplest organisms and how they evolved over time, researchers are hoping to be able to explain why certain features became dominant and even map out our evolutionary path with robotics. The team are now looking to recreate the movements of the first creatures that went from water to land. It is exciting because it means paleontologists and researchers don't just have to look at fossils and imagine how these organisms moved. Half a billion years later, they can now physically recreate it. How cool is that? Aubrey Lovell (11:27): Wow. That is amazing. It just blows your mind. That's wonderful. All right, I think it's time to transport you back to Barcelona through some audio editing magic, to continue your interview with HPE Chief Technologist, John Frey, talking about the keynote of CTO, Fidelma Russo. Michael Bird (11:45): John, of course, one of the other aspects of artificial intelligence is that sort of ethics, the unintended consequences. What are your thoughts on that? John Frey (11:53): It's very important. And in fact, Fidelma really talked about ethical AI and why that has to be an integral part of any AS solution. Fidelma Russo (12:01): As I talk with customers who are all under pressure to deliver AI workloads to their business, several common questions emerge. How do I ensure that I am using GenAI in an ethical and safe way? How do I pick the right infrastructure? What is the right management and operation software stack? We must start first with how to ethically and safely use GenAI. To adopt GenAI at scale, innovation is not just desirable, it's imperative. And companies need forward-looking solutions to address the complexities of data privacy. Then you want a hybrid deployment because we want to be able to use both public cloud models and also models that we train on-prem. We also want the ability to integrate these GenAI models into our business processes. Our innovation team has been obsessively focused on creating a HPE GreenLake cloud application to enable our internal developers to safely use generative AI models. (13:13): And I want to give you a little peek behind the curtain on a project that they have been working on. Project Ethan is an application that is equipped with robust guardrails. It ensures compliance with data privacy regulations, and it facilitates smooth and secure AI interactions, from simply using chats, to using APIs to create new use cases. It enables the use of various models, either on-prem or in the public cloud. We deliver authentication, authorization, and metering out of the box. Project Ethan is the tip of the spear where the AI journey begins. So with that, I'm happy to announce the HPE GreenLake Hybrid Operations Software Suite. This is a world-class hybrid cloud platform-based SaaS portfolio, and enables you to manage, observe, and protect your data and your AI models from edge-to-cloud. John Frey (14:20): In fact, often when customers are thinking about an AI solution, they miss the ethical implications. Let me give you a real-life example. We had a grocery customer into the Houston Customer Innovation Center, and they were thinking about using a robot that went up and down their store shelves looking at low stocking situations, particularly for high-margin products. So they wanted to prevent that. So they thought, "Well, we could put cameras on a robot, have it stroll through the store looking at the shelves and figuring out if we're low and restocking those shelves." What they didn't think about is, what would be the privacy implications of their customers about having a robot that continuously wandered the store and got video imagery of all of these customers? So we helped them think through the ethical implications of that. On the positive side, we help them think through, "Well, if you're going to implement that solution, and if you can take care of some of those privacy concerns in a proactive way, notify your customers that that's in use, but also make it where they could interact with the robot." (15:24): How many of us have been looking in a store for something we can't find? We're sure they sell it, but we don't know where it is. What if the customer could interact with the robot and instantly be told, "Oh, that's on shelf this, on aisle two." And instantly go find it. What a positive benefit that would be. And then one last challenge in that environment is each store has one server in it. So you've got to be able to do that inference work and got to be able to handle that technology in a very compact way so that it doesn't overburden the technology infrastructure in the store as well. Michael Bird (16:00): Thank you for that. John Frey (16:01): Yeah, no worries. Aubrey Lovell (16:03): All right. We're getting towards the end of the show, which means it's time for, This Week in History, a look at monumental events in the world of business and technology, which has changed our lives. Michael Bird (16:15): Now, the clue last week was, in 1975, this place really took a bite out of the computer market. Did you get it? Well, it was quite niche. It was the opening of the Byte Shop in Mountain View, California, in 1975. As well as being one of, if not the first consumer computer shops, it did something very, very important. Founder Paul Terrell was tired of selling kits for early computers that his customers struggled to build and gave up on. So he got in touch with a young man called Steve Jobs, don't know if the name rings a bell, 21, who was planning on selling computer components. Terrell convinced him to sell whole, ready to use computers by putting in a hefty order straight off the bat, which it's argued kickstarted the fledgling personal computer revolution. Now, the stores are no longer around, but they had a lasting legacy and proved that a computer really was something you could just go out and buy in a store, bring home and plug in. Aubrey Lovell (17:17): And next week the clue is, it's 2003 and this bill is junk. Michael Bird (17:23): Ooh. Aubrey Lovell (17:23): This is a tricky one. Michael Bird (17:24): I don't think I know what it is. Aubrey Lovell (17:27): I don't either. All right. And that brings us to the end of Technology Now for this week. Thank you to our guest, HPE Chief Technologist, John Frey, and to you. Thank you so much for joining us. Michael Bird (17:40): Technology Now is hosted by Aubrey Lovell and myself, Michael Bird. And this episode was produced by Sam Datta-Paulin and Al Booth with production support from Harry Morton, Zoe Anderson, Alicia Kempson, Alison Paisley, Alyssa Michery, Camilla Patel, Alex Podmore, and Chloe Sewell. Aubrey Lovell (17:57): Our social editorial team is Rebecca Wissinger, Judy-Anne Goldman, Katie Guarino and our social media designers are Alejandra Garcia, Carlos Alberto Suarez and Ambar Maldonado. Michael Bird (18:08): Technology Now is a Lower Street production for Hewlett Packard Enterprise, and we'll see you next week. Aubrey Lovell (18:15): Cheers.