 And then welcome to theCUBE pod episode 50, Dave. I'm John Furrier, Dave Vellante in our studio. We're here all week with SuperCloud. Dave, we're going to do the pod in the Palo Alto Cube studios. Great to see you. I love being here, John. This place looks great. We'll get the stage set up. We just had our live stage performance here this week, SuperCloud six. The theme was AI innovators, which will be an ongoing series with theCUBE. Obviously there's a lot of AI innovators to highlight and people trying to be AI innovators. So we're going to continue to cover AI like a blanket. But just to, you know, an amazing week. I mean, we just had like great stuff in studio. We'll go back to the anchor desk and get some stories out there and a few conversations. Of course, you get the Boston and the Malboro mass studio and just keep on getting that content. But it's been an interesting week. You got the whole Tic-Tac vote. How about that? That was a huge, you know, government start getting the policy of banning. Did you see Mnuchin? Mnuchin says he's putting together a group to buy Tic-Tac. Of course, they're never going to let that happen. That's never going to happen. He's never going to let that happen. But it's kind of interesting. Yeah, it's just a former treasury secretary. It just goes to show you how the tech platforms are now becoming like what looks like utilities. We've said on the previous pods that, you know, this could be regulated, but there's so much going on. I mean, that's a big story on SiliconANGLE. We have a lot of, you know, funding news going on. You're seeing a lot more. What's going on? What's next? We had some great startups on here in SuperCloud Six. We had Kyle Weller from One House AI. We had Venkat from Rockset. They're doing stuff that's unique. It's not just like one trick pony kind of solution. They got everything in there. They got this new systems that are building. But just overall, a great week. Uber and Walmart, I thought, were really, you know, practical uses of AI. You know, what came out of that is this is going to be the year where experimentation gets into, you know, reality. And I think actually, on the one hand, Uber and Walmart are actually applying AI in a big way, but most customers are still in the experimentation phase. Just tons of funding around together. AI closed another huge round. Databricks is still red hot. Unstructured raise, 40 million. I think it's series B. Another $50 million financing round. Sorry to the systems. Debutes of water. A wafer scale AI training chip. Anthropic releases cloud three. And that's now available on Amazon. That was pretty much big news yesterday. And then vast data continuing to show their partnerships. And you got GTC around the quarter. You got KubeCon. And we're going to be at the Broadcom financial analyst meeting next week during the same week as GTC from NVIDIA. There's a lot of activities going on. A lot of people going to be in Silicon Valley in town. You're coming back out. So that should be a great week. Zscaler bought Avalor, Israeli based company. They kind of consolidate and curate security data. Going to tie it into metadata. Doug Merritt was on that board. I think he still is. They raised 30 million. I heard the number was like 300 million. Yeah, I mean, it was just a great week. And of course we had last Friday International Women's Day. Savannah did a great job. Amazing job down there. That's the one we've ever had. And we've been into every single WIDS. Yep. She did really outstanding work. One of the things she did is she asked at the end of each for every guest. Like what advice would you have for young women? And so we got some good content coming out of that. Savannah did the best. I will say Savannah's WIDS was the best we've done there. Yeah, yeah. She really has the mojo when it comes to doing women in tech interviews, her alignment. She understands it, but she really probes the personal side of it, which I like. And also she ties in the news and she ties in also the thematic of it was. But she kind of brings out more cultural themes I like. And I think that has more legs. As we do more programming out, it'll run longer. It has more meat. So it's good to get those extra story nuggets in the interview style. And I think that's great for us. She's really good at active listening and engaging with the, not only the guest, but also the audience. So I really like her. She makes it fun too. What else is going on? Just a lot of AI discussion around open AI's video. Sonora with Wall Street Journal had a great article. I thought Joanne Bell was probably one of the best products I've seen to like great video, 10 minute video and story with the CTO of open AI. And they went into how good it's getting, but also the flaws and diffusion models are getting better. And it's going to come up with the whole, what is real? Of course, they pivot to the elections. You know, what's real, what's fake. And you got to watch out these days. My son Tyler was saying on our family text, when you answer the phone from your bank, don't use your original voice because when you answer the phone, they might be capturing your voice and do voice activation. I'm the same. I go, hello, hello, hello, hello, hello. You have a muffled voice. Who is this? Yeah. But you know, John Chambers was on theCUBE, one of our earlier Q-pods said voice will be the killer app. Remember he said that? Yep. So he nailed it. And companies that are actually being able to authenticate. So Microsoft had big news with the co-pilot for security. I saw Sean Bice, who's been working on that project for my Amazonian, Sean Bice is in the security with Charlie Bell. He's been working on co-pilot from the beginning. He was at RSA last year, almost came on theCUBE, but the PR police said, no, you can't go on theCUBE. So he promised me, Sean, if you're watching and listening, you got to come on theCUBE at RSA this year. We definitely want to have you on and tell Frank Shaw, we said it's okay. I'm sure he's going to be like, yeah, no problem. But just the funding, right? I mean, you know, you got a series B for unstructured $40 million series B from Menlo Ventures at a $230 million valuation. And the company just started, okay? So, you're starting to see the valuations to move up a bit. And I think you're starting to see a speed game. And again, Dave, I want to get your thoughts on this because- Because Wiz? No. Wiz had a $800 million round at a $10 billion valuation. Wiz is a cloud security company. Very hot. Remember last year, the line outside their party at RSA, right? So they're to your point. I mean, it's just valuations in AI and cyber still. I saw Wiz two years ago when we were at an Amazon marketplace event. Wiz. Wiz, Wiz. And like, they were just growing like a weed. I talked to some of the early employees who were there representing them, the partner alliance side that they're like, oh my God, it's crazy. And they were the fastest growing startup ever. And then one year they started went really, they went straight up- And they are really high in the ETR spending data. And then you saw the, did you see the Hammerspace meta deal announcement? I don't know what that's going to do to Hammerspace. That's Dave Flynn's company. I don't know what it's going to do with their valuation, but it can't hurt. Yeah. Very technical blog. I started to read it. It looked like a good pivot to try to get into the vast swim lane. I mean, vast success is looking really good right now. So I'm thinking if you're in that market, you look at vast data, you say, wow, I should get into that. So I think you're going to see this pattern and it happens every time. You and I both talk about the historian kind of view of the cycles. Once this emergent embryonic market hits, the patterns become clear, clear visibility on the swim lane of the opportunity. And then everyone snaps in line. So it usually takes one or two players to see it. And then they go, boom, they go right on it. And then everyone goes, okay, that's the model. Let's pivot behind that trend. That's the market. And once the market appears, it's a race. And then it's going to be full throttle. And then it's going to be like the NASCAR, you know, who's in lead and then ultimately someone will take it away. The interesting thing about that, that Hammerspace, Facebook or Meta announcement was maybe not so much even Hammerspace, but the Meta piece where they're going to basically, you know, build open source tooling. And so Charles Fizzi and I, we're always debating. He doesn't believe that Alibaba is actually a hyperscaler. So we debate about that. He obviously doesn't believe Oracle's a hyperscaler. I don't count them hyperscalers. He calls them a clown, not a cap. I don't count Oracle as a hyperscaler either, but I think they have a great business. And so we debate about that a lot. But I think, so I asked them, can you concede that Meta is actually a hyperscaler? So I think Meta could actually, you know, in this AI race come out with, I mean, if they choose to have an enterprise class open source, you know, managed service for AI tooling and you know, compete with the big three US hyperscalers. What do you think about that? I mean, I think Oracle's in the mix because if you go back to Oracle and think about some of the words they used just years ago, engineered systems. When you look at all the GPU action with the clustered systems we've been calling them, they're basically engineered systems and they're built for high performance. And at the end of the day, when you start getting into this next wave, it might come down to, hey, I'll put all my cloud dogma aside, I want performance. I want ease of use, I want performance. I'm going to throw workloads out a bunch of big iron, give me the performance I need, that's all I want. And if someone's got a better performance mouse trap, that might work and I think that's a viable scenario. We were there when Larry announced, the cube was there at the Oracle headquarters when Larry announced the original Exadata, you remember that? I do. And it was fast as hell. It was a while ago. It was fast as hell. It was like early last decade, but so it's proving out to be the correct move and Oracle, you know, all time high for stock and they had great earnings, solid earnings I should say, but what's getting people excited about Oracle is their capex spend. And you know, Charles fits again, he poo-poo's it. Look at Oracle compared to the hyperscalers of course, but Oracle's capex is starting to grow and they're forecasting more capex so people get excited because that means that along with their RPO forecast, which are growing, and indicate a lot of cloud demand. And the thing about Oracle's cloud is it's really, really friggin' profitable, okay? So they've got the database, they've got the autonomous database, they have their applications that are being powered by that autonomous database and they own all the infrastructure underneath it because they build it themselves. So Safra, cats years ago realized, wow, after we bought Sun, we can actually integrate this hardware. We don't need HP hardware anymore. Remember HP Oracle, how famous that was. We can just do our own. So they captured all that margin and they get software like margins, not hardware like margins, much more like, you know, Amazon and Microsoft than, say, Dell. And so it's an incredible business. It's pretty interesting. I mean, at SuperCloud Six, we had a lot of conversations around like what will be the infrastructure of the future? And the big takeaway from SuperCloud AI Innovators this week was the infrastructure work that's going on right now is pretty significant. A lot of people are grinding away and working on how to make that faster scalable but also stable because there seems to be a lot more momentum on the app side. I'll say open source and the proprietary foundation models are getting a lot of attention with Anthropic now running on our cloud three out and Amazon's now got it in bedrock and it's available. You're going to see a lot more action. So as developers start cranking away on the AI apps, they got to run it on something. So managed services will take the first hit of those. So you'll see a pop, open AI will get a pop, Anthropics will get a pop. So I think open AI is the Microsoft bet, Amazon's going anthropic, that'll be kind of a managed service to run in beds and all kinds of things through cloud and or open AI. And then as companies start using the AI, the next question is, what do I do? I got to host it. So do I host it on premise, which they will? How do I, what do I host it on? Do I host it on rack and stack servers or server under the table? What do I put a clustered or purpose built engineered system together or HPC kind of vibe? Or do I go to a specialty GPU cloud like a core? So these are going to be decisions that are going to evolve. And I think when you look at some of that middleware glue code, you're seeing companies come up saying, hey, just buy us and we can do all these things. And some of the old moves might not be available. So like observability is changing radically because there's new needs. You got AI, for instance. AI is very network based nowadays. So like you got to need network traffic. If you're going to use AI if you're in the infrastructure. So, you know, at KubeCon next week, Rob Streche was talking with me last night, our cube of research and leading on the whole cloud native side. The AI at in the infrastructure is going to be really DevOps like cloud operations, AI ops. You know, what's going to connect systems? How do we make it stable? How do I provision microservices? How do I manage it at scale? And that's data, that's data driven. So you got to have the data, right? So, and you can do that stuff on-prem now and at the edge. Yeah. And so you always point out the cloud operating model. You know, you're going to see a lot with AI it's going to change the game to basically you got to be in control of your own data, whether you're on-premise or whether you're a publisher, I mean, the story of Google's coming out with a new generative experience. I want to get that whole segment later, but like Google's got this new search experience that's going to cut traffic to sites. You got to go direct. So, I wanted to pick up on something you said about the sort of AI deployments and what's happening on-prem at the edge, if I can. This was not a good week for semiconductor stocks. They kind of got hit and we're going to see next week at GTC and we have the Broadcom Financial Analyst meeting, you know, semiconductors obviously at the heart of AI, but there's a big discussion going on in the industry about what you're going to be doing at the edge with do you need GPUs or can you use, you know, basic, you know, simpler, more cost effective CPUs. And I've always felt like, okay, you know, when you look at what's happening with ARM and iPhones, these are GPUs inside of that. So, Vikram Joshi who was here, he didn't come on SuperCloud 6 but he was here hanging out. We tried to get him on, we just couldn't. He's the, you know, founder of Compute AI and he was telling me that, look, these workloads, this matrix math that you do at the edge, you know, it's actually pretty simple math but there's a lot of it. So, you're actually going to need GPUs. So, this is a big debate now. I saw an announcement with ARM and this company doing stuff with Intel and blah, blah, and people are hoping that you can use, you know, general purpose CPUs because they're way cheaper but he's basically saying, you know, to your point, with the amount of data that you're going to need to process, that simple math, but there's so much of it, you're going to need GPUs, this is argument. So, we'll see how that plays out. You're going to need GPUs. You need to have them working in a new kind of system. I mean, this is where, I mean, remember the old days of the PC, right? You know, as the PC motherboards got better and better, they had to put more chips on and make the processor got memory so it was a system architecture, it was a motherboard. And you put it in a shell, that was called a PC. If you think about that same concept today, it's not the PC anymore or a server, it's all the other machines. So, it's logically like a PC, you got to have components and you're going to have a central processing union, GPUs, all kinds of processors, all have to work together and be connected. Okay, so that connection is network based and that's where Broadcom comes in. And NVIDIA. And that's why Broadcom, we had at MWC, we wanted to get Charlie on because he was a featured interview because what Broadcom's doing is the future. The connected systems have to be, are going to be large scale. I mean, I don't want to say the word mainframe, Dave, but like, remember, when I worked at IBM right out of college, they had the term called big iron. Remember that? This is big iron. Okay, big liquid cooled. It weighs like 70 pounds. Big iron meant a mainframe honking machine, takes up a whole room, glass room, you know, AC's on, all kinds of power. And it did limited things compared to now, but now the big irons kick ass, it's so huge and it's so high-performing. That's going to be the future. And by the way, the constraint is not the size of the motherboard, like in the PC, the constraint is power and cooling. And so, okay, you're going to have to really make these big iron systems engineered in a way that's going to have to maximize the resources available. And that's why you're seeing cottage industries pop up like cloud insurance. Startup we interviewed this week is doing basically GPU insurance. I'll ensure your capability and capacity and if I don't deliver, I'll pay. So that's like flood insurance, Dave, it's like, you know, it's like, hey, let's get some flood insurance, all right. We'll talk to you about Charlie Cowis. This is where he was, he's talked about chiplets and I've sort of been, you know, okay, help me understand where chiplets fit, but this is where power comes into play. And the other thing you, a couple of times this week referenced the Jensen conversation at Stanford. And you remember what he said there? He said, you know, in the future, every workload is going to be accelerated. What is he saying? He's saying every workload is going to need GPUs. Now he's so biased, but he's been right a lot. And so, you know, basically, you know, he's always talking about Moore's Law is dead. Well, one of the things he said that was interesting and I want to bring this up because no one in the industry is talking about it in the Gen AI space. Everyone's going gaga over Gen AI, prompt engineering. Right. Okay, he was saying prompt engineering has prompt engineering under the covers, but he was essentially saying there's going to be, the two waves are engineering prompts and responses and then reasoning. So to your point about these workloads needing, you know, horsepower, you're going to have the use case where there's going to be stuff going on in the workload or application. We say, okay, I'm getting answers. I'm doing some stuff and computing stuff. And then say, oh, I want to reason, send that workload out to a, probably a central cluster and go crank it away and give me the answer as fast as possible. So that reasoning is a little bit deeper, system two thinking he called it, which we called, you know, first principles kind of thinking that we would do as humans, like, okay, I'm going to go think on a problem, not just say, hey, Dave, how's your data? Great, John, thanks for asking. How's the weather looking good outside? This is really interesting what you're saying because his point was, look, when you ask ChatGPT a question and you want an answer pretty quickly and you get it pretty quickly, but what if you said, hey, I'm going to ask a question, you know, help me think through this business problem and take a day, you know, or spend this, spend a thousand dollars, spend $1,500 on solving that problem. I don't really care how long it takes. If you come back in a week, that's cool. Do your best. So, yeah, do your best. Imagine if you were able to, you know, have that, you know, patient patients and let AI, you know, run for a week. So I thought that was, you know, pretty amazing. Now, I wonder where, you know, Grock plays in all this because they're the super low latency chip company, right? So I wonder how that all fits in here. I don't know. I mean, we need to think about it. Well, we're going to continue the pound on that. There's a story on ad week today. It's all over the net about Google search. This ties into a panel I'm doing for Ray Wang. Ray Wang's hosting this thing this week around ethics in marketing. It's talking about some of the analysts, firms out there that aren't kind of doing an ethical things or, or... Or it's disclosure, really, isn't it? Yeah, it's a disclosure. But you're speaking on that. Yeah, I think you're going to talk about how to build organic audience, which is what we've been doing. Yeah, I mean, I think there's more of, I think there's a lot of confusion and desperation around how to make money. And, you know, I predicted this 20 years ago that ads going to crumble. Ads will drop, banner ads, and then that would be more reputation. I'm going to talk about reputation. But I think a lot of people are flat footed with the fact that they can't make money, so what they do is they'll do whatever it takes. And sometimes they won't disclose anything. So we're going to talk a lot about that. But the big impact is that the market's evolving. So you've got the role of influence relations and PR and AR, that panel's going to be about that. But the Google news is about the AI impact. And this is kind of ties to why that panel exists because there's so much fake news, so much fake information. So you have the rise of fake analysts, right? So that's what I'm going to talk about, that some people just really aren't analysts. They're just like either marketing specialists or PR people. So you can't say I'm an analyst and not being an analyst. So I call it the rise of the fake analysts. Well, you could say, anybody can say they're an analyst. So what's an analyst? Analysts gets access, analysts, this is Ray Wang's point, analysts get access, analysts get paid a bigger piece of the pie because they have IR impact. But you can't, if you're faking it, then you're basically one getting paid, that someone else doesn't get paid, Andor, or... You're right about that. An analyst gets paid between $10,000 and $20,000 a day for a really top analyst, maybe the top-top ones. I mean, analysts are the high-priced service. So if you're a high-priced service, you have to have an elite level of service. You can't say I'm a service provider at the elite level. It's like juicing yourself up in baseball, steroids. I mean, I'll tell you this, when I was at IDC, we knew what an analyst was. An analyst went deep into a domain. He or she, at IDC, of course, you quantified those markets. And I knew more about the markets from a quantification standpoint than virtually anybody on the planet, because that's what we did, that we did that every day all day and we were the best at it. You think about Gartner analysts. Gartner analysts are very good. Everybody likes to shit on Gartner, but they're really deep experts in their particular domain. Now, over the years, some of that's morphed, right? It's sort of been, some of them, what is an analyst that's been watered down? I mean, you've seen that. Well, the thing is that there's a bigger spectrum of services available that analysts can provide, right? So you have hardcore, old-school, pure analyst model. Independence. Independence, doing independent research, tons of work, and that research, that analyst research stuff that has results and presents it to the community at large and the community will debate it and or digest it. Versus doing. And pay for it. Versus doing advocacy. Now, or getting advice. Well, companies that hire, hey, they're an expo, hire you, help me with my business. I'm going to a market and my products position properly. They get paid for that service and get paid handsomely. Then there's other services, like marketing services. Like, hey, help me amplify, do an ad for me. Hey, I'm an analyst and I love this product. I don't think there's anything wrong with it. As long as you just say it. As long as it's disclosed. Yes, this is what we do. And so I was talking to Jason Bloomberg about this. He is a great server. He provides advice. He's a great researcher and he writes with SiliconANGLE. Well, he's got a marketing service and he's really clear. This is what I do. He's not hiding the ball. He's not saying, hey, I'm a marketing specialist, but I'm going to act like I'm a real analyst and get paid and be positioning myself that. And that's the difference. If you're just, there's no scoreboard. So again, my point is, I do think, actually, I had one thing. I think it's okay as long as it's disclosed. But I also think just from the ethics of an analyst, you don't ever want to change your opinion because you're getting paid. And I hope people don't do that. I think that you can't put lipstick on a pig and say, okay, I'm getting paid to say this. I mean, that'd be a spokesperson. You can't do that as an analyst. You can't can say every company's going to be a next big company. No, you have to be, and I was right about that. You have to be true to yourself. And that's why I like to use data. That's why I love our partnership with ETR because we have data and we have our own data as well that you got to bring to the table. So opinion without data. You got to call the balls and strikes right. You can't say everything's a strike or everything's a home run. I mean, I saw analysts out there, they get out there and they say every company's going to be the next big thing. So statistically, when it becomes a big thing, they say, you know. And you can't say, okay, I'm going to say nice things about a company that pays me and I'm going to say negative things about a company who doesn't pay me just to get them to pay me. I mean, I know why personally as an analyst don't play that game. I mean, I'll tell, I mean, Snowflake's a great customer of theCUBE. We know that, but we were the first to report some of the challenges that we see in Snowflake and pricing. Oh, but also they were doing a good job. It's also, they were doing good work. So it's easy to compliment someone who's actually doing a good job. Well, that's true. But when someone hits a home run, you say, hey, they just hit a home run, it's a nice one. But the point I want to make, and sometimes this happens is when you have a customer who's paying you and then you say something that they perceive as negative, but you know you have data. I mean, in this particular case, we're talking about, we've talked to many, many practitioners say, wow, it's just too expensive to do that data engineering and data pipelining inside of the Snowflake. So we do it outside of it. We either do it in Databricks or we do it in Amazon, you know, with Elastic MapReduce, et cetera, EMR. It's okay. So Snowflake hears us, hey, let's talk about this. So you have an honest conversation about how they're addressing it. Other companies, well, I've had this happen to me say, hey, I thought we were friends, or, hey, we're paying you. I'm like, I'll give you a, if you're paying me? So what, don't pay me then. You know, take, here's your money back. This is my reputation. So I think as an analyst- Well, this is the reputation. This is why I'm brought this up. I didn't want to go into that tangent and go too much rant on that, but we'll talk about it on that webinar today at one o'clock. But the issue of generative AI and fake information and hallucinations brings up and also the ability to do content better and faster can will arm someone who's not as deep to look deep. You've said it many times that ChatGPT takes an elementary school student and turns them into a Princeton Ivy League grad. At least from a writing standpoint. From a writing standpoint. But now you have other things, coding. So I brought this up because in our business, the media business, which I'm going to be talking a lot about on this panel, Google, Google's using search engines changing. They're using their generative AI experience could cut organic traffic to publishers as much as 20 to 60%. So they have this new generative experience where you type in something and they'll create the content for you from sites that they've been crawling. So it's pretty mature to create, understand the impact according to EdWee. But what this means is that imagine if search traffic goes away. I mean, we know how much search traffic comes to our site. And that's why I think with AI you're going to see a shift towards reputation. And again, not to tie the analyst piece into it, that's only one dimension. It's happening in every industry where you have vanilla and or fake content masquerading as truth. And you're going to start to see attention and influence become a really interesting power dynamic when you start to think what's reputable. So the question isn't so much who's the best and you can judge content by its content. Maybe it's a watermark or whatever, but you're going to start to see a trend towards direct response, direct navigation to sites, direct navigation to communities where reputation will be defined by the quality of the work and the people around it. So if Google takes the search traffic and stay with me, the point is going to come. The search traffic goes away then you're going to have to have direct navigation to the site. It means people have to go as a destination saying siliconangle.com. And so you have to change the entire behavior of the franchise. So it's a huge blow to journalism. Or you're going to have to put the content, you like to term watering holes, you went to Rob Stretcher, you went out to put that content into those watering holes. Yeah. And the problem that's by the way is always, oh, we've done it. Yeah, that's, it's been, it's been our model. But think about the people who make good content. It costs money. So if you don't, if you can't monetize that, you get desperate and you do desperate things. So they spend on their news desk. Well, they have subscription business, but if they take the New York Times Wall Street Journal out on the big, the big outlets, they got revenue streams. They got events, they got subscriptions and they got some advertising. But that's going to kill the one B tier and below content. And what's going to happen is you got to figure out how to monetize so what people are doing that they're, they're groping for anything. I'll do video and I'll bump up the views or I'll do this, I'll do that. So the journalism are, they're the ones paying for it. Now Google's enriching themselves off the work of the journalists. So, so this is a huge, huge thing. So, you know, back to who, how to make money to whatever way they're talking about analyst relations, whether you're talking about journalism, whether you're talking about digital marketing, money making is going to change. We predict, again, I predicted this 20 years ago that when this crashes, it's going to be about reputation. And I think the good news is we prepared ourselves, but it's going to be a complete shock to the system. And we're already seeing it play out like these panelists today. It's going to be talking about why people are doing this, who pays for what and what service. And I think you'll see it normalized, but into the day, it's going to come back down to if you can have good content and you can attract an audience. Okay, that audience can be merchandised and monetized into a way for advertising type thing where that advertising will be community-based. So to me, it's clear reputation will be number one. If your reputation is, I just want to make a quick buck, then you can't fund content because it's not sustainable. Yeah. So, by the way, Charles Fitz responded to me. What do you say? He said, the whole, just to set it up, I said, can we agree that Meta is a hyperscale cloud? And he said, came back, he said hyperscale in their infrastructure, but not a cloud. They have a history of pulling the rug out from under the devs, which by the way is true. So it would be an uphill climb. Then he says, their AI strategy is very unclear beyond open source software. Well, well, okay. But I think Fitz, he still owns a lot of Microsoft stock because it's clear to me that they see what's happening with open AI, which was supposed to be open, you know, open source, and they closed it up. And so I think Meta is saying, hey, we're going to contribute to the open source community. But he did make some good points about how they have pulled the plug up from underneath their developers as has Twitter. We've seen that. We were at the butt end of that as a Twitter developer with our crowd chat. He wasn't really high on Oracle. Who? Charles? Yeah. No, he's down on Oracle. He's like negative on Oracle because he's all things Microsoft, right? Cause he used to work there. He bleeds Microsoft and obviously, you know, smart guy. He's like Kubernetes is next. He came for TikTok. He's very clever. I love fun. I like Charles. It's like Saturday Live on Twitter for Tech. It's really funny. Yeah. But he's smart. He's on point too. But the thing is that I think, I disagree. I think Meta is well positioned. I think their scale could be a great resource. I think they could be the next AWS if they play their cards right relative to AI and I think as well as Google because they can just use their open source devs and use their power with the models that they got. So if they do that, then they could rock and roll. Yeah. And I don't want to beat on the whole open AI structure. All in guys unpack the structure. We did that a year ago on a breaking analysis. But it's very timely now with Elon suing them. But the TikTok thing's interesting, John, if I could for a second, I did a Twitter poll. You saw that. And I asked people, if you were in Congress, how would you vote on TikTok? Would you force divestiture? Would you ban the platform or leave it alone? You said leave it alone, I think. You said leave it alone. I said leave it alone. I said force divestiture. Most people are saying ban the platform, which is astounding to me. I mean, essentially force divestiture and ban the platform are the same. So it's like 67, about two-thirds of the audience said that's what they want to see. I mean, essentially this bill, if I understand it, is basically a ban the platform. Because the Chinese government is not going to let Steve Mnuchin buy TikTok. But my point has always been about TikTok is we should insist on reciprocity. If our social media platforms can't operate in China unless they're divested, then we should say the same thing here. I don't know. Brendan, producer, send a text. Check the demographics on that bowl. Probably not under the age of 30. So probably that would be leaving it alone. They love the product. As a father and someone who wants to see people be more enriched with their mind and conversations, I'd say kill the app because it sucks you in. Next thing you know, two hours are gone because it just feeds you like a pellets of content. And I know how my kids feel about it. They'd freak out if they do. But they love it. So the question is, is it harmful? And if you look at it from the perspective of, yeah, they're couch potatoes. But you can say that about TV when we were growing up. Oh, yeah. These kids, the minds are going to get fucked on TV because they sit there and just stare at the tube. And the boob tube they used to call out. So like, well, OK, whatever. What happened on TV? They did other things. So TikTok will evolve because you'll start seeing it evolve with ads. Now the question about TikTok is not so much the medium itself. It's that the Chinese are using it for surveillance. And for other nefarious activities. So to me, I would only look into it, and this would be more of an off-book black ops operation for the government, is just get in the weeds and just unpack the app and just quarantine it so it's not a national security problem. So to me, you can't take the app away. People like it. But you can manage that. I hear you saying. But I don't think the argument is that I don't think that's the right argument to make. I think you have to make an economic argument. What economic argument? One of reciprocity. The economic argument is, look, if China is not allowing our social media, US-based social media platforms to operate freely in China, unless they are 41% owned by the Chinese company, then the same economic model should apply here. There should be reciprocity. But we're not dictators. Well, but yeah, we're not dictators. But China is dictating that we don't copy them. Social media. They should copy freedom. But they won't. I know. So then why should we allow a Chinese-owned company to operate freely in our markets? Well, I mean, to me, my point is we should put it back on China and say, here, here's your choice. You can either let our social media platforms operate freely in China, and then we will give you reciprocity here. Or we will treat your social media platforms the same way you treat ours. I think that is a much longer argument. I would say take the whole thing over and just make it US-based. Well, they won't get the algorithm that way. They won't ever give you the algorithm. But my point is, put it back on China, make them defend that awkward position of reverse engineering of the algorithm. Well, why hasn't anybody done that? I don't think anyone's tried. At scale. You don't mean meta has tried? Of course. But they don't have the volume that they have. Meta does. Not on TikTok. They have threads, and they have reels on Instagram. Doesn't meta have more active users than TikTok? I don't think so. Check on that. Meta's got like 2 billion active users. Well, the point is, I mean, I'm not sure meta wants to copy TikTok or with Instagram. I mean, who uses Facebook? But Instagram, what's that? I guess here's the point. Is yet again, the government's too late in coming in because it's Brendan's point. The young audience is not going to stand for this. They're going to ignore some of the big investors. What if you were an early investor in ByteDance? This is also underscores the perils of investing in Chinese companies. What's going to happen if they actually force the vestiture? What's your TikTok value going to be worth? What's that valuation your investment going to be worth? It's going to be cut at least in half. Chinese officials say the US has shown, quote, robbers logic towards TikTok. And Washington must stop unfairly suppressing foreign companies. They're so full of shit. I mean, what are you talking about? China suppresses foreign companies. That's what they do all the time. And so, I mean, I think you can't. And that's why I'm saying you have to make it about an economic reciprocity argument as opposed to, well, it's bad for kids or whatever. I mean, not to go back to one of our early pods. We're on pod 50 right now, our podcast 50. I think on our first 10, we talked about Twitter being the next TikTok. Remember, we talked about Elon's moves and him saying, hey, if the government shuts down Twitter, TikTok, then X could step in there. And then reels obviously came out, threads, I mean, with Facebook. So you started to see the Americans trying to position themselves, but, you know, the question comes up is, is it a foreign issue? This comes up also with AI. This is not well talked, discussed in the mainstream yet, but the node Koshla is on the side of AI's like the Manhattan Project. So we all seen Oppenheimer, right? So the movie, very secret. He's not too biased. He thinks that it's a national secure that we hit AGI first as a society. And that is the equivalent of the A-bomb, the Manhattan Project, that we should not open source it and keep it as a national security treasure and asset and not open it up in the Chinese end. And then we're like, well, come on. I mean, you think they're not going to have access? And we reported last week that a Google Chinese person stole AI source code and was convicted by the DOJ. So there's always been espionage. So the question is, can you even stop that? Or just make everything open and hope it's democratized? It's a good debate. Well, it's an interesting discussion. I never thought about it. I don't know how I feel, but, you know, my gut says always go, stay open. Well, but there's an argument for like the Manhattan Project. Good to point about open innovation, but the other point about meta and what they're doing, you know, with that Hammerspace announcement, I was Chris Mellor's, you know, right up. You did a great job. But they're pursuing AGI on an open platform, right? So to your point, that's, I don't know how I feel about that. But the related, it's somewhat related, kind of adjacent here is the EU AI regulations. Did you see that? No. It basically, you know, once again, EU is leading in regulations like they did with GDPR. This one is basically, you'll love this. Basically they're going to say, okay, this AI platform, this AI software is, it's either what I call red, like it's not safe. So it's not allowed. It's yellow, which is, there's some concerns, but, you know, we're going to manage them. Or it's, I wouldn't say green. I would say it's not red or not yellow, okay? So, and so that'll be allowed. So, I like the simplicity of that, but I don't know how you, first of all, define what is dangerous and red and not allowed versus what is, let's call it like light green. I don't know who adjudicates that. I don't know who's smart enough to figure that out. And how do you evolve that? Like, is it something like, okay, if something does a deep fake that has to be watermarked and that's mandated, is that capability there? Can it be unwatermarked or is it immutable? I mean, I suppose there's things I can do like that, but there's going to be a thousand other things or maybe more that have to be adjudicated. So I don't really, Let me get your thoughts, Dave, on something that we brought up in a couple of pods ago. I want to come on that point about, the government issues going on. We always talk about that, but we predicted that startups will be falling out of the sky. The economy would be. What's happening? The economy would be terrible. Unless the peace fiber is a multi-mile high. But it's a weird market. The economy, spending, IT spending, outlook. We're getting some early signals from our cube research data and the startups. So let's start with the startups. What I'm hearing is there's like a collection of companies on the verge of basically having no runway to add into this year. So 2024 is going to be the vapor, the vapor fume date, as we say, the fume date is when you run on fumes. It's like when you're driving your car, it's like how many miles? Lights on. Red lights on. 10 miles. How could, I can make it 855. Conserve gas, turn the AC off, you know, push. So when you're on a fume date, weird stuff happens. You do think weird things. People leave, boat takes on water, as they say. So you can see a lot of companies on that fume date. A lot of down rounds, a lot of mergers and that was in the startup world, a lot of failure coming. At the same time, the AI's still booming. So that's happening. IT spend, I'm catching wind that IT spending might be down, might be changing. What's your take? What's the research tell you? I know you've been digging into this. It's really preliminary. I talked to the ETR guys this morning and so they're still gathering the data. We're still hopeful that it comes to remember last year, December, January, IT decision makers were predicting a 4.3% increase in spend, but very back end loaded. Okay, so far that's, it's a little softer than that, but there's still, the data's still coming in. So we're kind of hopeful that, hopeful it gets back up there, but you know, it's got a ways to run. Gartner, earlier this year, came up with a forecast that IT spending would increase by 8%, which sounds really robust. That would be, what, double GDP or something like that, maybe even more, but I think Gartner might be including spending on internal staff. But so it's early right now. I do think I would say this. Anecdotally, when I talk to customers, they are definitely squirreling away money for AI projects and they're not sure what to do with it. They're doing some stuff and they're doing some experimentations, but they're not going into production in a big, aggressive way, other than things like, you know, code generation and some chatbot stuff and, you know, summarizing text. I mean, all the basic stuff that you can do with ChatGPT, but in terms of bringing it in, building RAG, you know, like we're doing, you know, I would say we're still in the experimentation phase, John, right? I wouldn't say we're in, you know, high volume production for our RAG-based QBAI. And so I think what's happening is people are squirreling away money because their top line budgets are not growing. And that right now is the fundamental. And we'll be reporting on this, you know, within the next... And you've always said that the money always is moving from one project to the other. We've been hearing that top talent has been moving off, you know, cloud-native projects like Kubernetes and whatnot into the AI side. So you're starting to see some of those Alpha engineers being pulled into AI. So you've seen teams change. So I can see that little squirreling away cash spend. Yeah, hold back. Especially as the big vendors like the Cisco's of the world are too retooling their observabilities and the needs are changing as we report on SuperCloud. So I think we're going to be in this like tide shift, like mid tide, the tide's going to come in and out and you're going to see people, you know, in between kind of this, you know, stuck in this shifting of the winds. What do you think about, you had mentioned several pods ago, that a lot of the crypto developers moved into AI. Crypto's now, Bitcoin hit 73,000 this week. And I was saying, you know, hopefully there'll be a convergence of crypto and AI will come together. What are you hearing in terms of the developer momentum now that Bitcoin is, you know, back up and Ethereum's rising, people are talking about Ethereum ETFs. I think the Bitcoin momentum has nothing to do with the developer action on crypto. I think that's an indicator that Bitcoin has a viable currency, it's legit. I think you're going to see more and more impact from the Bitcoin market, because let's face it, at 73,000, whatever it is today, there's people who made a lot of money on Bitcoin who have Bitcoin and they need to put it to you. So I think you're going to see the Bitcoin billionaires become a very important part of the ecosystem and they'll probably come in and disrupt key markets around what developers do in terms of investment. I think you're going to see that new entrants come in, not just the classic these ease, but you're going to start to see a global landscape of new players from crypto, Ethereum and mainly Bitcoin, who have made billions. And by the way, this is currency exchange. So it's like, okay, it's Bitcoin. So the question will be, can they get liquid? Can they put that in the fiat? Can that convert it into asset classes like venture and real estate? So I think you're going to see a robust economy develop around crypto conversion Bitcoin to asset classes. So I'm envisioning new VC funds coming in to deploy crypto value into an asset class venture deploying into developers. That's one. The other one is going to be the fact that now that AI is becoming a new infrastructure, I think you'll see those distributed crypto developers think about the problem space differently. I think they'll stay decentralized and that's going to needs to fake out a little bit, but I'm still bullish on smart apps, D apps on top of a blockchain for certain transactions is definitely going to happen. No doubt in my mind that what we saw in 2018 and what we were building with Kubecoin, the idea that you would take the middleman out and create a immutable transaction ledger is absolutely compelling. I mean, it's compelling infrastructure. The question is the timing of it, what's the killer app and can you get critical mass? I think things like Google taking away search revenue from publishers will maybe create a new opportunity for market there. Analyst relations thing we were talking about that could be completely reputation coin based. Let the market decide who's reputable and then we get these fake firms that do fake leaderboards that are out there that don't really understand the value. Real reputation. Well, valuation of reputation is something that's never been done before. So interesting things happening with Bitcoin right now. There's a halving coming up in mid-April. So that's probably why people are so enthusiastic meaning. So what happens is every, I don't know, I think four years the algorithm that Satoshi, whoever he or she is or they basically cuts in half the value that you get when you mine a Bitcoin. Okay, so that's part of it. And I don't even think of Bitcoin as a currency. I just think it is a store of value where it gets interesting is to your point is Ethereum. So for those of you who don't know, Bitcoin really is not programmable, right? That's so the Ethereum founders, Anthony Diorio has been a cube along and others developed Ethereum with this programming language called Solidity, which is in programmable, which allows you to build Dapps, distributed apps, which is to your point. And so what I really like about what you're saying is as these big social media platforms like Google, screw with the publishers, there could be a democratization of control over who gets rewarded. Right now it's backlinks and things like that and who spends the money on Google for AdWords. In the future, it could very well be some kind of open source like reputation algorithm that is really combines software engineering, cryptography and game theory. There were the community decides who has the reputation and democratizes that whole dynamic. Yeah, I think that's a great use case and that's going to happen. The question is when, when it's going to happen. And I think the thing about crypto and decentralized infrastructure is the one problem that I see that's maybe stalling this is the fact that if you look at the growth of AI and all the cloud native work, the role of data is changing so rapidly, the velocity of it, how it's stored, how data is streamed. So the volume and variety and velocity of data is so fast but it's so compelling as input into AI where there's real ROI on the table that will trump any kind of theoretical decentralized solution because the speed of that we kill it. Look at Snowflake, they popularized separating compute and storage, right? That's going to change things like Chronosphere out there and Rockset. We talk about these startups. So as this next level comes, that has to get set and that's driving more change at that infrastructure level. That will beat the benefit to switch to decentralized. Because they can't handle the volume of data fast enough. So again, member of the old days of crypto. Do I need a database or is blockchain a database or is it like a system of record? And we argue that the databases, you don't need to use crypto for databases. So the database market right now is under major flux and so that's going to probably take the precedent over. So it's slow to use crypto or... AI is moving way too fast, moving way too important with the value it's creating and the database market is changing. So that's going to drive faster and that's going to beat the crypto infrastructure in the short term. Now Bitcoin is a currency, that's about wealth. And that's going to come out of the market. I don't think you're ever going to be going to buy a cup of coffee with Bitcoin in this. No, but you can move a billion dollars into a fund and contribute to a asset wealth. So you know what one of the most interesting angles I've seen recently, and this has been now well over a year old is when MicroStrategy basically became a crypto company. Like a crypto properties investor, right? MicroStrategy is an analytics firm, right? And this guy, Michael Saylor, who's the CEO of MicroStrategy decided, hey, we're going to take our free cash flow and put crypto with Bitcoin specifically on our balance sheet. And you look at what that stock has done. So the other thing about Bitcoin right now is all these ETFs that got approved by the government. And the SEC was fighting it, Gary Gensler was fighting it. But look at MicroStrategy is actually a great way to own Bitcoin without paying fees to these ETFs. But he's actually a really thoughtful individual. And I think that's a company to watch. They're no longer a software company in my mind. They take their software to throw up free cash flow so they can invest in Bitcoin. Well, I think one thing I walked away from the Jensen speech last week at Stanford was, again, this reiterates our theory. The power law of specialized models happening. Software is the game. Resources are going to go down to zero. There are constraints, though, going to manage the systems game. They call it whatever, AI factory, but it's really a systems game at this point. And that's going to be the AI conversation for the next year, I think. Infrastructure, it's like the still classic three layers of stack, Dave. Infrastructure, middleware, applications. So you're going to start to see that kind of three phase approach really form for AI and AI native. And I have yet to see what a real AI native stack looks like. You got software like CUDA for Nvidia, but what is an AI stack? Native stack looks like for developers. So we're going to continue to do that. And, you know, pod, maybe pod 70, we'll, we'll hit there. We'll be there. But we'll be 60, 10 more pods, 10 more weeks, we'll see. But yeah, let's wrap this up. We got to, we got to go start our day today. Next week's CUBECon. The CUBE team will be there. On the Broadcom Financial Analyst Meeting, GTC. We'll be out here for that. This never ends. Great stuff. And then the event season is kicking in. So then April is just. Yeah. Stay with us here in the CUBE pod. You know, we're going to bring you all the content we've been seeing, all the analysis. Check out siliconangle.com and check out the CUBE research.com. That site is formerly wikibon.com. That's now the CUBE research. That's where you're going to start to see all the new analysts come on board, putting their content there. That's where the deep dives are. Check out the CUBE research.com. Of course, Silicon Angles, we're all the millions of users a month go to. And of course, youtube.com slash Silicon Angle for all the videos. That's our firehose library. And of course, Twitter, Silicon Angle, Twitter, the CUBE. And DM us if you have any questions or comments. Thanks for watching.