Exclusive Coverage: theCUBE at IBM Think | theCUBE Analysis

theCUBE.net

Clips
News
More from Exclusive Coverage: theCUBE at IBM Think

Shelly Kramer

Tech Analyst

theCUBE Research

Three insights you might have missed from theCUBE’s coverage of IBM Think

IBM Corp. was an early adopter of AI with its “Jeopardy!”-winning Watson computer. Now, Big Blue looks to make a comeback by positioning itself as the “glue layer” for enterprises trying to integrate AI across platforms.At IBM Think 2024, the company announced a series of products based on its AI platform, watsonx, and showcased its dedication to developing a network of partnerships. Its acquisition of Red Hat Inc., a leader in open-source solutions, has continued to pay off. IBM is expanding on that ethos of collaboration by releasing Granite models into open source

IBM opens up to new collaborations in effort to stay ahead in the AI race: theCUBE analysis

Artificial intelligence is in the experimentation stage, with companies scrambling to come out on top, and IBM Corp. is no exception. The AI space is challenging even the most dominant tech companies.During this week’s IBM Think keynote, Arvind Krishna, chief executive officer of IBM, predicted that AI alone would be responsible for a 600 million to billion increase in applications by the end of the decade. With the ongoing success of Red Hat Inc., which IBM acquired in 2018, the tech giant looks to be a dark horse in the AI race by homing in on infrastructure and automation software, such as IBM Concert, an AI-powered automation tool for simplifying app development

play_circle_outline Unleashing the Power of OpenShift and Watsonx: Exploring Game-Changers and the Strength of Partner Ecosystems

play_circle_outline IBM Concert as a tool to manage complexity in operations and reduce risk

play_circle_outline Unleashing the Power of Gen AI and Large Language Models: A Game-Changer for Enterprises

Info

theCUBE Analysis

Shelly Kramer

Tech Analyst theCUBE Research

HOST

search

>> Hello, welcome to this special CUBE conversation. I'm John Furrier with theCUBE. We are in the IBM CUBE clone here on the show floor. This is a creator set built here by IBM to bring all the influencers and the analysts on board here. Guys, this is a similar situation... We've seen this more with our CUBE, of course. Let's do our thing. Shelly Kramer, Sarbjeet Johal, analysts on theCUBE Research team. Let's research and look at the data. Let's find out what's going on. Shelly, you had a great post on thecuberesearch.net on the keynote.

Shelly Kramer

>> I did.

>> Let's get into the keynote review. And of course, Rob Thomas, Dario Gil had the one-on-ones we just had. Let's get into it. What's your take?

Shelly Kramer

>> We've been drinking at the fire hose of all things IBM Think. Arvind started out his keynote with a statement that AI is a fundamental technology. We know that, right? But he said the current estimates are that AI will add an estimated 4 trillion to the annual GDP productivity by 2030. 4 trillion, one technology. That's pretty amazing.

>> Yeah, and he also said a lot of things about, which I thought was great, that IBM's culture is really driving from a Red Hat perspective. And he says a lot of innovation, but a lot of deployments coming, Sarbjeet. And then he said, "We are going to see an era of application sprawl." He's predicting billions of applications coming in the next 10 years. Some big number, I don't know if it was billions, but it was large. It was basically he was saying app sprawl. So, more deployments, which means more production, more enablement, applications. What's your assessment of that? What does that mean?

Shelly Kramer

>> I think that is aggressive number, to be honest with you. We have to take care of our legacy as we create new applications. Of course, there'll be a lot of agents. If you call agents applications, actually his number will be correct I think, or he will be in the ballpark. But creating that many number of applications in the next six, seven years is a... I can't relate to that myself.

Shelly Kramer

>> 600 million.

>> 600 million apps.

Shelly Kramer

>> 600 million to a billion being written by the end of this decade.

>> Yeah, that is significant. So, if you believe it to be true, let's just say half the numbers are right, half a billion new apps with generative AI enablement. That means application tsunami is coming. We've talked about it on theCUBE.

>> .

>> And so, what does that mean to the underlying infrastructure? More observability, they have the automation wheel, Rob Thomas presented that in his session. You got observability. You got to figure out what's going on with the cloud-safe models. You got the network aspect of it.

Shelly Kramer

>> Connectivity, yeah.

>> The whole thing is coming together. You got to make that more automated. So, all those apps need under the covers technology. What is the secret sauce going to be, guys? I mean, what's the research tell us?

Shelly Kramer

>> I think the form factors might change going forward. If we talk about universal application portfolio, that means consumer application as well. The glasses will be coming to the forefront, they'll be lighter. Not at all what Apple has produced right now. It's like monolithic stuff on your head. So, with that, we will get a lot more change in hardware. We will be doing more AI on the device, on the edge. As well as we'll have better models, better cheaper, faster models on the back end. Cloud will flourish with it.

>> I love that. A lot of stuff on your head, monolithic stuff on your head. That's Apple Vision Pro you're referring to. Shelly, let's talk about the go-to-market because on the one-on-ones we just had with Rob Thomas, Arvind and Dario, we asked them about what was their lead ecosystem play. OpenShift and Watsonx are the game changers here.

Shelly Kramer

>> Absolutely.

>> IBM is putting all the wood behind Watsonx and OpenShift, which when asked about will they be infusing Red Hat to promote IBM, Sarbjeet, you asked that question.

Shelly Kramer

>> Yeah.

>> He went, "Whoa, stop right there. We are not going to touch Red Hat." Shelly, give us your analysis. I want to go to you for the Red Hat.

Shelly Kramer

>> Well, I think part of it too, separate from Red Hat in that conversation, is the ecosystem that they were talking about and how committed IBM is to developing that partner ecosystem. And I think it was Arvind, but it might've been Rob who was just talking about the reality today is strong partnerships and strong alliances not only benefits the vendors, but it really benefits customers too. So, when we look at the ecosystem that was announced here today, it's AWS and IBM and Adobe and Meta and Microsoft and Mistral and Palo Alto, I just covered that the other day, Salesforce, SAP. And then, the launch of the Saudi Authority for Data and Artificial Intelligence Authority. So, all of this demonstrates to me the commitment to actually making this work across a broad ecosystem.

>> And what's going to be interesting, Shelly, is to see how they can parlay their, I will say, successes and failures they've had with Watson and AI in the past. IBM's got a deep DNA in data. They know the data business. If they can bring Gen AI to their ecosystem, like they hinted to us about the relationship with Palo Alto Networks. Hey, why being in the threat application business? Partner with the platform Palo Alto Networks, and bring gen AI to that, while having hybrid OpenShift under the covers with Red Hat. Sarbjeet, you asked the question, "Will they promote products with Red Hat?" He said, "No, we'll optimize with Red Hat. We have to leave Red Hat alone because it's an independent company." Quote independent company. It is not, IBM owns it. So, I mean a tough thread to thread that needle right there.

Shelly Kramer

>> It is, actually. You have to read between the lines. His body language was telling something else, but his language was telling something else. That's how I read it. He said, "Look, no, no, we will not use Red Hat to promote our products." But they have started to do that. Actually, we heard that from other senior execs. I will not name the names because they will get in trouble from Arvind.

>> You can name names of people in the IBM theater here. But I mean this an interesting promoting product... Again, he's got to be diligent and police this because if they're promoting IBM product, that's different than integrating, creating a competitive advantage by having access to Red Hat without touching the core code.

>> Exactly. Yeah, and he said that said. "We will have the synergies, but we will not promote our products based on using Red Hat." And we have seen this movie earlier, VMware was EMC company. And I was at VMware when we wanted to be just pure and EMC wanted to push their storage through us, and then we will buy from HP and they'll yell at us. So, I see what's happening there right now. I can feel it.

>> Y'all made a lot of money with VMware. IBM can make a lot of money with Red Hat. The goose that laid the golden egg. You don't make a pillow out of the feathers. We keep the goose laying the golden eggs. That's the old expression.

Shelly Kramer

>> We talked last night with Dario, while at the drinks. That was an interesting discussion, right? So, their proximity to developers has to come through the Red Hat, I think because Red Hat has closed nest developers.

>> I think this is the Trojan horse I think IBM has with Red Hat because the hybrid play right now is so hot. Everyone's infrastructure decision is hybrid. That's a done deal. We've been calling that for over at least a decade now. But now, the past five years, we're seeing as a steady state hybrid cloud, no problem. Next game is open source. CNCF KubeCon is getting boring. Kubernetes has won. That's an orchestration layer in the cloud native world. In comes agents and generative AI, you're going to start to see a lot of innovation in the infrastructure software area, which Rob Thomas was talking about. So, Red Hat will win them new business. Then IBM can bring Watsonx and modernize it. So, the beautiful thing about IBM right now is they're bringing those two together. You do not want to touch this Trojan horse, which is the goose that lays the golden eggs because Red Hat is bringing so much value to IBM that now we add in HashiCorp with the acquisition, which by the way, is going to be a nice compliment to Red Hat for day three operations. Day two, day three. Sarbjeet, what do you call it?

>> Yeah, day two.

>> I mean it's going to be a good fit for them either way. So, right now, they're looking good. I mean middle of the fairway for IBM in terms of where they're positioned. Can they execute? And my analysis and my critique is or my challenge is IBM has not been good at ecosystem. They traditionally have been blue oriented, very blue, if you're in IBM's world, you get know all the end perks. It's an IBM world. You get bluewashed if you get bought in. So, if their acquisitions can come in and be accretive and not be so much blue IBM and make that more open and bring in an ecosystem where there's a real value proposition that isn't services-based or competitive. So, if IBM could share the pie, as Arvind said on this partner keynote, then I think you're going to see a massive success because they have to share the wealth. If they do that, they're going to attract partners, and they're going to have the products too to boot.

Shelly Kramer

>> Well, I think that's what we're seeing though with these ecosystem announcements and these major players in the tech space. And I wanted to touch on briefly the announcement around IBM Concert, which I think is a big deal. And we touched on this just a minute ago when we were talking about this app sprawl and complexity. And every conversation, John, that we have with anybody, whether it's a vendor side or whether it's a customer, complexity is a killer. And so, everybody's trying to get arms around that. And so that's really what IBM Concert does. It's help manage around controlling operations, reducing and mitigating risk, managing compliance with gen AI, which of course, Watsonx is amazing at. So, anyway, I think that what Concert can do is generating analyses, visualization, recommendations that help speed actions, discover gaps, reduce complexity, all of these things that customers desperately want. So, I think that was a really interesting announcement.

>> Yeah, I think the developer angle is big too.

Shelly Kramer

>> Yeah.

Shelly Kramer

>> Yeah, I think open sourcing of the models is very interesting. How Dario describes that how the enterprise data will be married with the models, so that is interesting value prop. It's very clear actually value prop if somebody listens to that pitch and it makes sense. So, that's another thing. And then, the ecosystem, I was impressed when Arvind, in one breath, he listed seven of his biggest partners. That's AWS, SAP Oracle and Salesforce, Adobe and Palo Alto Networks, right? Yeah, they're bringing-

>> They're being a Switzerland. They're going to try to unify the network to layer and the middleware of the stack. I think it's a middleware play for IBM. They know that market. And I think Dario Gil, you mentioned Dario. Just for the folks watching, Dario Gil heads up IBM Research. He also pointed out something I thought was instructive to us is that the LLMs or large language models and foundation models, they're very large. He showed how big they are. And he made a point that there's not a lot of enterprise data in those models. But he also then conflated that with what I think's reality is that the enterprise could be that big too. So, I think what he's pointing out is that the enterprise will have their own large language model. So, his point was technology and the mechanisms are there for enterprises to have as big of a model as say an OpenAI, 20 gig type model. So, he's like, "The mechanisms there. What will be the model? That's the big question mark. We don't yet know that." But he's saying, "That's coming." So, he's just saying, "There's a lot of headroom. It's a real deal. So, it's a real wave for the enterprise."

>> There's a lot of experimentation going on. Right now is not the time where you can be prescriptive. I warn IBM not to be prescriptive, but just be experimental. So, it's like right now, AI is over-the-counter medicine. It's not you go to a doctor and they prescribe you, "You do this." And actually, I raised that issue. It's like you cannot use data as a generic term for enterprise. There's different types of data. There's logs, there's your taxonomy of your products. If you're healthcare, that doesn't change that much, but your transaction data changes. So, you need five, six tiers of data, based upon what's the frequency of change. And then, decide where that data should be. Should you use RAG or you should use other mechanisms to infuse the data into the models?

>> Arvind was talking a lot about how the computing's there, the data's early days. I want to ask you guys a question. How is the generative AI wave impacting IBM's business? And it's almost like they were so excited to do quantum, and then in comes generative AI, they have to stop that. Not the momentum, but certainly the press and all the hype is gen AI, rightfully so. How has gen AI impacted, certainly it puts quantum in the backseat, but-

Shelly Kramer

>> Not for long.

>> Not for long. What's your take on what their impact is for gen AI?

Shelly Kramer

>> Well, I think they're making a considerable impact. I think that we've talked about this a lot before, and I think the challenge for IBM is an image issue. And whether we're talking about gen AI or we're talking about security, IBM doesn't immediately come to mind as a killer player in this space. And so, I think what IBM is laying down here is that, "You know what? We are. And here's our long-term strategy and here's what we're doing for developers, and here's what we're doing for simplification for customers and things like that." So, I feel like they're making some significant inroads. What do you think?

>> What's the impact, IBM-

>> I think impact was going to be big, but they are recovering from that sort of gap, if you will. They quickly figured out that if we open source our models, we can get traction out there. And Arvind, he described that thinking, why they did it this way. So, I think they will do pretty good. But having said that, they are B2B focused. There's no B2C component. Like Oracle has mixed up stuff and Microsoft has their sort of productivity software because-

>> I think gen AI has impacted the company in many ways. I'll tell you that the InstructLab story, how that came about is really a big deal. And I think, that to me is the game changer. That could be the next Linux operating system for AI. If you look at how models interact with each other, and Arvind was running away from that word-

Shelly Kramer

>> You asked-

>> I asked him specifically on purpose, "Will there be an operating system for AI?" And hinting, "Will you be it?" He immediately said, "Yes, distributed computing, but I wouldn't call it an operating system, I'd call it operating workflow." Because look, operating environment, operating system. I mean technically, he's a nerd, so he's parsing it as a definition. IBM has an opportunity with Red Hat to be the operating environment with OpenShift and Red Hat at the bottom of the stack, and then use WatsonX as a unifier as infrastructure software, middle layer, and then use the gen AI goodness at the top of the stack. Rob Thomas said, "We're an infrastructure software company." Now, I find that ironic since they're partnering with all the biggest infrastructure players like Amazon Web Services, Azure, but I think what he means is, "We'll have this stuff under the covers for our own storage, but we'll partner with whatever infrastructure hardware is out there, whether it's an AI PC or an AI server or a cluster. We want to own the orchestration layer and then own the middleware." IBM wants to own the middleware of this next wave. I think it's very clear to me, and we're going to dig this into our research, but...

Shelly Kramer

>> Yeah.

Shelly Kramer

>> I think when Rob said that we don't have the multi-cloud CMDB, I have been saying that for the last seven, eight years. To do the multi-cloud right, we need the multi-cloud CMDB. But the problem with having a CMDB in cloud is the infrastructure is so fluid that you can't pin that I have this infrastructure in my environment right now because we have a serverless in there now, right? We have VMs, we have containers and all these different flavors, and they fluctuate with time and loads. And so, doing a multi-cloud CMDB is an audacious goal. And .

>> basically configuration database for all those-

>> Yeah, .

>> Our little lab visit here on the floor with research team shows that the game's going to be about metadata. You can't make large-scale data addressable everywhere. So, I think the game will be in metadata. We're already doing this on theCUBE Research with-

Shelly Kramer

>> You talk about uptime, right?...

>> separating compute from data. We've got that going on. That's our sixth data platform. Our key research at theCUBE Research. But when you start thinking metadata, think pointers and see. Think spreading pointers all around the network to point to data sets, I think you're going to start to see synthetic data fill in that gap. We don't know, but we're speculating, as we always do here on theCUBE, special appearance Sarbjeet, Shelly-

Shelly Kramer

>> Great to be here....

>> good session. We're here at the IBM Creator Lounge. This is theCUBE special celebrity appearance in the IBM booth. I'm John Furrier, your host. Thanks for watching.

theCUBE Analysis

search

>> Hello, welcome to this special CUBE conversation. I'm John Furrier with theCUBE. We are in the IBM CUBE clone here on the show floor. This is a creator set built here by IBM to bring all the influencers and the analysts on board here. Guys, this is a similar situation... We've seen this more with our CUBE, of course. Let's do our thing. Shelly Kramer, Sarbjeet Johal, analysts on theCUBE Research team. Let's research and look at the data. Let's find out what's going on. Shelly, you had a great post on thecuberesearch.net on the keynote.

Shelly Kramer

>> I did.

>> Let's get into the keynote review. And of course, Rob Thomas, Dario Gil had the one-on-ones we just had. Let's get into it. What's your take?

Shelly Kramer

>> We've been drinking at the fire hose of all things IBM Think. Arvind started out his keynote with a statement that AI is a fundamental technology. We know that, right? But he said the current estimates are that AI will add an estimated 4 trillion to the annual GDP productivity by 2030. 4 trillion, one technology. That's pretty amazing.

>> Yeah, and he also said a lot of things about, which I thought was great, that IBM's culture is really driving from a Red Hat perspective. And he says a lot of innovation, but a lot of deployments coming, Sarbjeet. And then he said, "We are going to see an era of application sprawl." He's predicting billions of applications coming in the next 10 years. Some big number, I don't know if it was billions, but it was large. It was basically he was saying app sprawl. So, more deployments, which means more production, more enablement, applications. What's your assessment of that? What does that mean?

Shelly Kramer

>> I think that is aggressive number, to be honest with you. We have to take care of our legacy as we create new applications. Of course, there'll be a lot of agents. If you call agents applications, actually his number will be correct I think, or he will be in the ballpark. But creating that many number of applications in the next six, seven years is a... I can't relate to that myself.

Shelly Kramer

>> 600 million.

>> 600 million apps.

Shelly Kramer

>> 600 million to a billion being written by the end of this decade.

>> Yeah, that is significant. So, if you believe it to be true, let's just say half the numbers are right, half a billion new apps with generative AI enablement. That means application tsunami is coming. We've talked about it on theCUBE.

>> .

>> And so, what does that mean to the underlying infrastructure? More observability, they have the automation wheel, Rob Thomas presented that in his session. You got observability. You got to figure out what's going on with the cloud-safe models. You got the network aspect of it.

Shelly Kramer

>> Connectivity, yeah.

>> The whole thing is coming together. You got to make that more automated. So, all those apps need under the covers technology. What is the secret sauce going to be, guys? I mean, what's the research tell us?

Shelly Kramer

>> I think the form factors might change going forward. If we talk about universal application portfolio, that means consumer application as well. The glasses will be coming to the forefront, they'll be lighter. Not at all what Apple has produced right now. It's like monolithic stuff on your head. So, with that, we will get a lot more change in hardware. We will be doing more AI on the device, on the edge. As well as we'll have better models, better cheaper, faster models on the back end. Cloud will flourish with it.

>> I love that. A lot of stuff on your head, monolithic stuff on your head. That's Apple Vision Pro you're referring to. Shelly, let's talk about the go-to-market because on the one-on-ones we just had with Rob Thomas, Arvind and Dario, we asked them about what was their lead ecosystem play. OpenShift and Watsonx are the game changers here.

Shelly Kramer

>> Absolutely.

>> IBM is putting all the wood behind Watsonx and OpenShift, which when asked about will they be infusing Red Hat to promote IBM, Sarbjeet, you asked that question.

Shelly Kramer

>> Yeah.

>> He went, "Whoa, stop right there. We are not going to touch Red Hat." Shelly, give us your analysis. I want to go to you for the Red Hat.

Shelly Kramer

>> Well, I think part of it too, separate from Red Hat in that conversation, is the ecosystem that they were talking about and how committed IBM is to developing that partner ecosystem. And I think it was Arvind, but it might've been Rob who was just talking about the reality today is strong partnerships and strong alliances not only benefits the vendors, but it really benefits customers too. So, when we look at the ecosystem that was announced here today, it's AWS and IBM and Adobe and Meta and Microsoft and Mistral and Palo Alto, I just covered that the other day, Salesforce, SAP. And then, the launch of the Saudi Authority for Data and Artificial Intelligence Authority. So, all of this demonstrates to me the commitment to actually making this work across a broad ecosystem.

>> And what's going to be interesting, Shelly, is to see how they can parlay their, I will say, successes and failures they've had with Watson and AI in the past. IBM's got a deep DNA in data. They know the data business. If they can bring Gen AI to their ecosystem, like they hinted to us about the relationship with Palo Alto Networks. Hey, why being in the threat application business? Partner with the platform Palo Alto Networks, and bring gen AI to that, while having hybrid OpenShift under the covers with Red Hat. Sarbjeet, you asked the question, "Will they promote products with Red Hat?" He said, "No, we'll optimize with Red Hat. We have to leave Red Hat alone because it's an independent company." Quote independent company. It is not, IBM owns it. So, I mean a tough thread to thread that needle right there.

Shelly Kramer

>> It is, actually. You have to read between the lines. His body language was telling something else, but his language was telling something else. That's how I read it. He said, "Look, no, no, we will not use Red Hat to promote our products." But they have started to do that. Actually, we heard that from other senior execs. I will not name the names because they will get in trouble from Arvind.

>> You can name names of people in the IBM theater here. But I mean this an interesting promoting product... Again, he's got to be diligent and police this because if they're promoting IBM product, that's different than integrating, creating a competitive advantage by having access to Red Hat without touching the core code.

>> Exactly. Yeah, and he said that said. "We will have the synergies, but we will not promote our products based on using Red Hat." And we have seen this movie earlier, VMware was EMC company. And I was at VMware when we wanted to be just pure and EMC wanted to push their storage through us, and then we will buy from HP and they'll yell at us. So, I see what's happening there right now. I can feel it.

>> Y'all made a lot of money with VMware. IBM can make a lot of money with Red Hat. The goose that laid the golden egg. You don't make a pillow out of the feathers. We keep the goose laying the golden eggs. That's the old expression.

Shelly Kramer

>> We talked last night with Dario, while at the drinks. That was an interesting discussion, right? So, their proximity to developers has to come through the Red Hat, I think because Red Hat has closed nest developers.

>> I think this is the Trojan horse I think IBM has with Red Hat because the hybrid play right now is so hot. Everyone's infrastructure decision is hybrid. That's a done deal. We've been calling that for over at least a decade now. But now, the past five years, we're seeing as a steady state hybrid cloud, no problem. Next game is open source. CNCF KubeCon is getting boring. Kubernetes has won. That's an orchestration layer in the cloud native world. In comes agents and generative AI, you're going to start to see a lot of innovation in the infrastructure software area, which Rob Thomas was talking about. So, Red Hat will win them new business. Then IBM can bring Watsonx and modernize it. So, the beautiful thing about IBM right now is they're bringing those two together. You do not want to touch this Trojan horse, which is the goose that lays the golden eggs because Red Hat is bringing so much value to IBM that now we add in HashiCorp with the acquisition, which by the way, is going to be a nice compliment to Red Hat for day three operations. Day two, day three. Sarbjeet, what do you call it?

>> Yeah, day two.

>> I mean it's going to be a good fit for them either way. So, right now, they're looking good. I mean middle of the fairway for IBM in terms of where they're positioned. Can they execute? And my analysis and my critique is or my challenge is IBM has not been good at ecosystem. They traditionally have been blue oriented, very blue, if you're in IBM's world, you get know all the end perks. It's an IBM world. You get bluewashed if you get bought in. So, if their acquisitions can come in and be accretive and not be so much blue IBM and make that more open and bring in an ecosystem where there's a real value proposition that isn't services-based or competitive. So, if IBM could share the pie, as Arvind said on this partner keynote, then I think you're going to see a massive success because they have to share the wealth. If they do that, they're going to attract partners, and they're going to have the products too to boot.

Shelly Kramer

>> Well, I think that's what we're seeing though with these ecosystem announcements and these major players in the tech space. And I wanted to touch on briefly the announcement around IBM Concert, which I think is a big deal. And we touched on this just a minute ago when we were talking about this app sprawl and complexity. And every conversation, John, that we have with anybody, whether it's a vendor side or whether it's a customer, complexity is a killer. And so, everybody's trying to get arms around that. And so that's really what IBM Concert does. It's help manage around controlling operations, reducing and mitigating risk, managing compliance with gen AI, which of course, Watsonx is amazing at. So, anyway, I think that what Concert can do is generating analyses, visualization, recommendations that help speed actions, discover gaps, reduce complexity, all of these things that customers desperately want. So, I think that was a really interesting announcement.

>> Yeah, I think the developer angle is big too.

Shelly Kramer

>> Yeah.

Shelly Kramer

>> Yeah, I think open sourcing of the models is very interesting. How Dario describes that how the enterprise data will be married with the models, so that is interesting value prop. It's very clear actually value prop if somebody listens to that pitch and it makes sense. So, that's another thing. And then, the ecosystem, I was impressed when Arvind, in one breath, he listed seven of his biggest partners. That's AWS, SAP Oracle and Salesforce, Adobe and Palo Alto Networks, right? Yeah, they're bringing-

>> They're being a Switzerland. They're going to try to unify the network to layer and the middleware of the stack. I think it's a middleware play for IBM. They know that market. And I think Dario Gil, you mentioned Dario. Just for the folks watching, Dario Gil heads up IBM Research. He also pointed out something I thought was instructive to us is that the LLMs or large language models and foundation models, they're very large. He showed how big they are. And he made a point that there's not a lot of enterprise data in those models. But he also then conflated that with what I think's reality is that the enterprise could be that big too. So, I think what he's pointing out is that the enterprise will have their own large language model. So, his point was technology and the mechanisms are there for enterprises to have as big of a model as say an OpenAI, 20 gig type model. So, he's like, "The mechanisms there. What will be the model? That's the big question mark. We don't yet know that." But he's saying, "That's coming." So, he's just saying, "There's a lot of headroom. It's a real deal. So, it's a real wave for the enterprise."

>> There's a lot of experimentation going on. Right now is not the time where you can be prescriptive. I warn IBM not to be prescriptive, but just be experimental. So, it's like right now, AI is over-the-counter medicine. It's not you go to a doctor and they prescribe you, "You do this." And actually, I raised that issue. It's like you cannot use data as a generic term for enterprise. There's different types of data. There's logs, there's your taxonomy of your products. If you're healthcare, that doesn't change that much, but your transaction data changes. So, you need five, six tiers of data, based upon what's the frequency of change. And then, decide where that data should be. Should you use RAG or you should use other mechanisms to infuse the data into the models?

>> Arvind was talking a lot about how the computing's there, the data's early days. I want to ask you guys a question. How is the generative AI wave impacting IBM's business? And it's almost like they were so excited to do quantum, and then in comes generative AI, they have to stop that. Not the momentum, but certainly the press and all the hype is gen AI, rightfully so. How has gen AI impacted, certainly it puts quantum in the backseat, but-

Shelly Kramer

>> Not for long.

>> Not for long. What's your take on what their impact is for gen AI?

Shelly Kramer

>> Well, I think they're making a considerable impact. I think that we've talked about this a lot before, and I think the challenge for IBM is an image issue. And whether we're talking about gen AI or we're talking about security, IBM doesn't immediately come to mind as a killer player in this space. And so, I think what IBM is laying down here is that, "You know what? We are. And here's our long-term strategy and here's what we're doing for developers, and here's what we're doing for simplification for customers and things like that." So, I feel like they're making some significant inroads. What do you think?

>> What's the impact, IBM-

>> I think impact was going to be big, but they are recovering from that sort of gap, if you will. They quickly figured out that if we open source our models, we can get traction out there. And Arvind, he described that thinking, why they did it this way. So, I think they will do pretty good. But having said that, they are B2B focused. There's no B2C component. Like Oracle has mixed up stuff and Microsoft has their sort of productivity software because-

>> I think gen AI has impacted the company in many ways. I'll tell you that the InstructLab story, how that came about is really a big deal. And I think, that to me is the game changer. That could be the next Linux operating system for AI. If you look at how models interact with each other, and Arvind was running away from that word-

Shelly Kramer

>> You asked-

>> I asked him specifically on purpose, "Will there be an operating system for AI?" And hinting, "Will you be it?" He immediately said, "Yes, distributed computing, but I wouldn't call it an operating system, I'd call it operating workflow." Because look, operating environment, operating system. I mean technically, he's a nerd, so he's parsing it as a definition. IBM has an opportunity with Red Hat to be the operating environment with OpenShift and Red Hat at the bottom of the stack, and then use WatsonX as a unifier as infrastructure software, middle layer, and then use the gen AI goodness at the top of the stack. Rob Thomas said, "We're an infrastructure software company." Now, I find that ironic since they're partnering with all the biggest infrastructure players like Amazon Web Services, Azure, but I think what he means is, "We'll have this stuff under the covers for our own storage, but we'll partner with whatever infrastructure hardware is out there, whether it's an AI PC or an AI server or a cluster. We want to own the orchestration layer and then own the middleware." IBM wants to own the middleware of this next wave. I think it's very clear to me, and we're going to dig this into our research, but...

Shelly Kramer

>> Yeah.

Shelly Kramer

>> I think when Rob said that we don't have the multi-cloud CMDB, I have been saying that for the last seven, eight years. To do the multi-cloud right, we need the multi-cloud CMDB. But the problem with having a CMDB in cloud is the infrastructure is so fluid that you can't pin that I have this infrastructure in my environment right now because we have a serverless in there now, right? We have VMs, we have containers and all these different flavors, and they fluctuate with time and loads. And so, doing a multi-cloud CMDB is an audacious goal. And .

>> basically configuration database for all those-

>> Yeah, .

>> Our little lab visit here on the floor with research team shows that the game's going to be about metadata. You can't make large-scale data addressable everywhere. So, I think the game will be in metadata. We're already doing this on theCUBE Research with-

Shelly Kramer

>> You talk about uptime, right?...

>> separating compute from data. We've got that going on. That's our sixth data platform. Our key research at theCUBE Research. But when you start thinking metadata, think pointers and see. Think spreading pointers all around the network to point to data sets, I think you're going to start to see synthetic data fill in that gap. We don't know, but we're speculating, as we always do here on theCUBE, special appearance Sarbjeet, Shelly-

Shelly Kramer

>> Great to be here....

>> good session. We're here at the IBM Creator Lounge. This is theCUBE special celebrity appearance in the IBM booth. I'm John Furrier, your host. Thanks for watching.