 Hello, everyone. Welcome to the Cube Pod episode 36. I'm John Furrier, Dave Vellante here, extracting the signal from the noise on a Friday when we do our recordings. I'm in Palo Alto days in the Boston office in Marlboro, Massachusetts. Dave, great to see you. Great stuff. Just landed from Austin. It's not really in Texas. A morning eye, not a red eye. I got my Bruins shirt on again. I got a lot of great fan mail on the Bruins shirt here. So go Boston Bruins. Big fan, as you know, even though I live in California. I love the Sharks too, by the way. So, like the way the Bruins thing, but a lot going on. Let's just get right to some main news. So let's get right to the big story in Silicon Valley, which is and reverbing the crypto world is Sam. Bankman free goes down in a hallmark landmark. I shouldn't say, I shouldn't say landmark fraud case. This guy is a total thief. It took literally record time, no deliberation, done guilty as charged. No defense. That's done in the books. Finally, that and you got lucky too, by the way, the press didn't could have been more more on this, but with the Israeli war and all the other distractions, he got away easy by not being totally publicly fleeced on this thing because he's just a tool built everybody. So we'll see how that turns out. We'll see what the sentence says. More regulation conversation. Biden put out an executive order around guidance around AI and a scaring people. And a lot of people come out of the woodwork. Andrew Ning, who's the founder of Coursera and part of Google Deep mind basically said, regulation is scary. People will get into that. Kube cons coming up next week, which is a big cloud native conference by the cloud cloud native foundation. Linux foundation will be there super computing after that. One of our cube analysts, Dustin Kirkland got a new job came out went back into the to the start up world to be VP of engineering at chain guard. So shout out to our cube guest analyst, Apple earnings, not as good as they hoped services carry in the day for Apple. And then finally, my rant is going to come up Dave solar wins. CISO gets sued by the SCC. And that's going to be a huge discussion putting shocks throughout the security system. Why him? Why not the data engineering guy is the whole wire is even a private company, even responsible for what should be the responsibility of other people. People are freaking out more AI conversations. You were at tech del del tech world. We're going to unpack what's going on inside the ropes. I know it's an MDA del tech summit del tech summit. Yeah, del tech summit. Yeah, I should say. And then finally, what I thought was the most coolest thing out there this week is the new Beatles song. Okay, John Lennon had recorded some stuff in the seventies and a software engineer had decoupled his voice from the piano on tape. And it's just a remarkable story of archaeology around musical archaeology, the creativity of the Beatles, the shared fascination with technology. I'll see if jobs were alive, he'd be totally stoked because you know, he loved the Beatles and it really marks the completion of the last recording that John Paul, George and Ringo did together. Okay. And with AI, this kind of reboots one of the most popular influential bands in the history of music Dave. So this Beatles song is pretty awesome. And it's very cool. So I give that the cool factor big time. But it's been a really interesting week. I was also at Palo Alto, Jeremiah O'Yan and Chris Ye with Blitzcale Ventures, a new firm they're putting together at a Palo Alto meetup with demos, all the hottest AI startups called the Lama Lounge. And it was packed. I was giving a little demo of our Cube AI. Saw a lot of old timers and OGs coming together. The young guns are out. Dave, it reminds me of the early days of the web and web 2.0, but much faster incubating. So just so much going on this week. Was it the Hanna House? It was amazing. Hanna House, was it? Is that called? The Hanna House, which was the old varsity theater. And SAP sponsored that back in the big day-to-days, remember? We did a big there one year. Yeah. It was great. It was great. It was awesome. And you should have seen the excitement. I was checking it out, the demos and people were just showing their product. I brought my laptop because I thought it was like the old school meetups where people just hanging out. No, it was it was rocking. High-paced, much faster. And I was talking with Ross Mayfield, founder of Social Tech. He works at Zoom now. He's a web 2.0 legend as well. And he and some folks who from the industry were talking about how slow it was in web 2.0 compared to this. It was a slow, slower boil in terms of the innovation than it is now. It's much faster. So these meetups, it's just faster. And it's interesting. I mean, you were at the Dell tech thing. What was going on down there? I mean, because they just had a big announcement with Meta. And they're putting in there. I mean, I'm really reticent to say anything because they must have said like 100 times. Remember, this is NDA. And then it was funny. I said, you know, when you're asking questions, of course, you know, we love to ask questions. So they say state your name and your firm. So I said, this is Dave Vellante from the Cube Research, formerly Wikibon, and that is not NDA. And so, you know, that got a few chuckles. But it was really good. That's what I could tell you. There was a lot of analysts there. It was in Austin. There was definitely executive presence. You know, no surprise, right? Dell always brings the execs to these things. And but this was a three day deep dive. It's still going on. I had to leave early because I got a commitment and home tonight. So I, you know, going from Austin to Boston, it's not easy to get back here. You either got to leave like, you know, 7am and get back at noon or you got to leave later in the afternoon and get back at midnight like Stretchy's doing. So Robin holding down the fort. And they're doing deep dives. Like, like really, they went deep. They unveiled roadmaps. We debated. We aligned on a lot of stuff. They're probably, there were a lot of analysts there. I won't even say the number. And a lot of smart people. The thing about Dell is they have the end to end spectrum. So you get people that really know client and devices. Guys like Bob O'Donnell, who's like, and he's more general than just devices. Like he knows, but he knows Apple really well. He knows client stuff really well. All the way to guys like us, you know, Stretchy, Data, myself, Enterprise, a bunch of Forester and Gartner guys there, all the IDC guys saw Matt Eastwood. It was, it was a packed event. And really, really good content. I again, I'm hesitant to say like exactly the format because there was they were really emphasizing NDA, but I came out impressed. It was one of the best analyst events I'd ever been to. And they had customer input, partner input, awesome gen AI conversation. They actually, they actually doing a lot more than people realize, I think. Well, I wasn't under NDA since I wasn't there. So I'll tell you what I think's happening. And of course, I can connect the dots because I wasn't there, but I could figure out from what I saw was there and I figured and seeing what's going on the way. I saw some posts from on the on the llama news. It's clear that Dell sees AI is coming fast. And, and, you know, Kelsey Hightower, who is a great cube alumni legend in the cloud native world, wrote a post that says if laptops keep getting faster, cloud based developer tooling is going to become less appealing. These machines are so powerful that most the value add can run locally. Maybe we'll get a hybrid model SAS running locally, but it's clear where things are heading. I had said and then someone wrote a MacBook Pro can do all those llamas. Okay. Here in the pod, we said this. I also said it to Dell is that they have an opportunity to bring a generation of new users that will use powerful laptops and servers to do all their local programming and then also train models and then put it in the cloud. So there's going to be a surge of on-premise. And if you look at Matt Baker's post on LinkedIn and talking about the llama announcement with meta and their commitment to that, you got to be thinking that they're on this because Dell, as you pointed out, successfully transitioned from the web. We debated this on the cube. I mean, oh, Dell went from, you know, when the web was mail ordered, when mail ordered to the web, they transitioned that. I think they're poised to transition. I think, yeah, they were probably really low key because they don't want to telegraph their moves, but I can tell you right now, I bet Dell's doing that. So to me, you're seeing a lot more processing and compute. And remember how developers posted on the cloud, Dave, local host, code on the machine, put it to the cloud. So I think you're going to see a renaissance in hardware like we've never seen before, especially as the silicon game gets better, especially as people want more GPUs, as they want more TPUs and compute. We had a quote on the cube. Remember in SuperCloud, Vikram said compute should be oxygen. It should be free. Free. So I can share this with you, which is not NDA because I got this at the Dell financial analyst meeting. Sam Bird was there. He's the head of Dell's client division. And he talked a lot about processor diversity, a.k.a. not just, I mean, you're talking about Intel, AMD, ARM, dot, dot, dot. And he also talked about NPUs inside the processor. And of course Apple's been doing this for years. So they're kind of catching up to Apple on that front. But look at what Apple's done with the M series, right? I mean, and so, of course, the whole industry is going to follow. Apple's advantage is they design their own processors, right, with their ARM based systems. But so, yes, I think you're making some really good points there. And the whole Lama two thing is interesting, John. I have talked to other sources at Facebook. They don't really know how, like what percentage of the downloads actually are running Lama two on prem. Like we use Lama two, but we're using it in the cloud. But they don't really know, well, I say we use it in the cloud. It's actually sort of an alternative cloud. But so, so they don't really know that, you know, what percent is on prem, but they can like look at who's downloading and say, well, these are like financial services companies and insurance companies who have a lot of data, like sensitive data, healthcare companies on prem. And they got data centers. So they can infer that a large proportion of the downloads are actually being applied to models, oftentimes with retrieval augmented generation, on prem. And so that's Dell's play, right? Dell guys like Dell, HPE, Lenovo, IBM with Watson X, that's kind of hybrid AI. Yeah, it runs in the cloud, but it's also going to run on prem. And of course, the edge. So Dell is, I think, very well positioned there. And I, you know, that's again, that's not NDA. They've talked about that a lot at Dell Tech World and all the financial analysts meeting. I mean, apps got to run somewhere and they're going to run on computers, they're going to run on laptops, they're going to run on edge devices and, you know, data and computer moving to the computer is going to be the compute engine. And this, this retrieval rag, you call it, retrieval automation generation. That's a huge trend, mainly because it treats data differently and for retrieval. And I had a long chat last night with some original, you know, web guys at this meetup and we were talking about how the idea of keyword navigation in search, like Google, is going to make way for math retrieval. If you look at rag, it's all math based and they do all kinds of these embeddings. It's a good way to get data and then get these similarities together and allows you to pull in things that are similar. But again, you got to have good data. I think this data hygiene thing is going to be a huge part of that. So I think, you know, Dell's position, all these companies are positioned on prem because the cloud's going to be expensive. Yeah. And, and, and it's going to be a huge cost. So well, at re-invent will be very interesting in a couple weeks. I'm going to ask Adam Szelewski that question around cost. Is that the blocker? So, yeah, but so just to stay on Llama 2 for a minute. So they, you know, they, they announced Llama 2 in July, I think, and they announced a 7 billion parameter, 13 billion, and I want to say 70 billion parameter version. Now, just for comparison, POM from Google is probably about 500 billion parameters. And I think open AI, you know, people speculate it could be a trillion. So the point is models are shrinking and you can do a lot with smaller models, as we know, with what we've done with CUBE AI. And these on-prem models and RAG, basically what they do is they take a highly clean, cleansed corpus of data like our CUBE data and they complement the LLM and they vector in that high fidelity data and then they send it back to the foundation model. So the results are much better, they're cleaner, there's less elucination and more accurate. And then obviously you got to train them, you got to put in the right guardrails and governance, but that's the trend. I mean, it's relatively straightforward to do if you've got a good engineering team. Yeah, I mean AI, we talked about the last time, AI wrappers versus, you know, more co-pilot, you know, native apps. We'll see what comes up at CUBE com, be very interesting to see that. And if you look at the AI earnings, you saw earning season is kind of in full swing, I don't know if you noticed, Dave, but there's no sign of re-acceleration in the software business, but you saw some winners in like certain areas like some cloud, the cloud players, Shopify was up, JFrog was up, Palantir was up, Fastly was up, but you know, all the other software guys are down, zoominfo, paycom, bill.com, Confluent, Alasian, Procore, all the traditional software SaaS people getting killed. Microsoft, Microsoft was the most notable to me of earnings because they're, they really specifically, I mean, other than NVIDIA, nobody's really come out and say, we have AI, generative AI revenue. Microsoft said it was a 300 basis point tailwind, so three percentage point incremental tailwind to Azure revenues. So, you know, you could do the math on that, you know, whatever you think. Is that because they were well positioned with the products? I mean, they had, was it because of their office suite and integration, or was it open AI subscriptions? What do you think that was? I think all of the above. I think definitely open AI subscriptions, open AI usage, and they're pushing pilot, co-pilots at $30 per user per month to all their, you know, 365 users, which is expensive as hell. But actually, have you, have you used that at all? Have you played around with Microsoft's co-pilots on 365? No, I haven't used it, but I've seen it in action. And it's, it's pretty frigging impressive, John. I mean, the things you can do and, you know, drafting notes and setting up doing, doing things in Excel that used to be so complicated. And it's, it's pretty powerful. You know, on the one hand, you're like $30 per user per month. Wow, that's expensive. But it's powerful. What did you use it for? What was your use case? So I didn't use it, but I, somebody showed it to me. They were like, look at this. It was like watching VMware for the first time, you know, spin up virtual servers. I was like, holy shit. I'm gonna take a look at it. Yeah, they were, you know, basically writing documents. They were, they were doing spreadsheets, you know, sometime. I mean, they have templates, but, you know, the templates never really looked the way you want them to look. You're basically setting this up with natural language. It, it was very powerful. I want to, I think, I mean, I think, I think it's a no brainer. I mean, we're doing ton of interviews, leading up to reinvent and we just had super cloud. So to me, clearly, you know, I got a little essay. I'm doing a little, I'm doing a little research brief, a kind of a quasi e-book on this day, but I'll tell you right now, and even my daughters are in who are in their, and just getting in their careers. There's conversations around, do we, do we limit AI, right? Do we engineer it first? Do we limit it first? And there's a lot of young people who are like thinking, hey, why don't we just, we got to regulate it. And they don't even know what they're doing, right? So, you know, there's, there's all kinds of approaches. And the key, the key that's coming out of it is that the personal information space that you're mentioning with this co-pilot is really important. Help me do my job better. Help me do my documents or PowerPoints or set up meetings, AI-driven personal assistance. It's going to be a big deal. And that's going to be an instant low-hanging fruit. Those chatbots that were lame will become better. And then you mentioned the data retrieval and augmentation generation. That's where I see the business value, right? You used to see businesses. So you have the personal, you know, the four areas are personal and then business to run this retrieval. People have all these data reserves. They have exhausts. Remember the term exhaust? And so you're going to see, you know, better retrieval methods, better use of data. That's like what people are working on now. And I think that's what you're talking about with Dell. And I think Dell's right on that. And then I think the third area that's coming out of this is where that what happens next is once you start getting the value in the business side and that consumerizations happen, you start to get into new business lines or new data sets emerge. I saw one company I interviewed, they said they're in the aggregation business. Now they're actually aggregating data from other sources. They never could do that before. So you're seeing new competitive advantage features coming from their app or service around new, new untapped potential around these data sets. And then finally, the fourth area of the AI is cognitive reasoning. That's where it starts to get kind of like squishy. So and that's like, like doing things like high stake drug discovery, doing molecule impact analysis high end stuff that was unattainable that you could do the AI. And I think those are the key areas that are coming out of all of our research as the key landscape areas of use cases and opportunities. And I think the legal stuff, the regulation jumps to the fourth one. It's we're not there yet. The reasoning and all that stuff it's happening is that people are working on stuff, but we're so in the we're so in the early innings or not even game time around this because just now it's about personal, the co-pilot helped me do a better document, write a press release, helped me write a story, do a PowerPoint. So I think, you know, let's get through one, two and three. And then once the reasoning and the so-called, you know, machines are taking over, that's down the road. So I am anti-regulation and people are like, what do you mean you're anti-regulation? Don't you want guardrails? What guardrails for what? Okay, this safety issues, I get that. But I think that's just, you know, something that people just thrown out there to scare people. And this is an issue. Andrew Ning, who's the founder of Coursera and DeepMime wrote a tweet and said, he's not so much against regulations about how people are freaking out and scaring young people and getting into the field, Dave. Well, like, oh, that's interesting. So, you know, several things on that. So first of all, the couple of things that came out of SuperCloud 4 we'll have in SuperCloud 5 coming up soon in November. But that whole idea of exhaust and model collapse when the LLMs are creating their own data and generating their own data, then, and it's a probabilistic model, it increases the chance of over-rotating on the highly probable and it misses the less likely and you might think, well, that's no big deal. Why is that a problem? Well, in healthcare, that could be a huge problem. If it misses the outlier. The other thing is on regulation. I think a lot of the big guys actually, they don't mind regulation because they know that's going to be a barrier of entry because they can afford to work their way around it. The old term regulatory capture where the rich get richer with government. And the last thing, when we talked about this, I think last week on sentiment AI, artificial general intelligence where the machines actually take on human consciousness and I'll repeat it briefly, but somebody who was the guy from Intel at SuperCloud 4 said, well, what if we're already there just to mind a mind of assume we're already there. Wouldn't the machines be smart enough to fake us out and hallucinate and make themselves look stupid and lull us into a state of relaxation? The only reason to bring that up again, just not to repeat it, but you know, in talking to technologists, John, all the building blocks are there to actually create artificial general intelligence. If you'd like draw the curves, you basically need more compute power and more data. And, you know, people believe and I'm kind of beginning to believe it too. I used to not worry about it, AGI, but I start to worry about it now that all the foundational elements are there, but you need more compute and you need more data, which we know we're going to get more data and more compute at some point, quantum is going to be here and quantum just keep getting better and better and better. So it's actually in who knows how long it's going to take? We have no idea. Look at, generative AI came out of nowhere. Yeah, but I just saw the founder of Signal. She's really technical. She's like, this is all bullshit. She was poo-pooing generative AI. So in the tech circles, it's all like, it's been around for a while. This is just a kind of new window dressing around seeing something on the consumer side that's an interface. That's just chat GPT. So there's not a lot of in the elite circles like, come on, it's a yonor. Come on, let's get real now for us mere mortals where people in the mainstream, it's magical. So if you actually look at generative AI today, what is it actually doing? And I just laid out the four areas I think are the most important, your little buddy sidekick, discover those data treasures that you have in your company, unearth the value, taking on big tasks, decoding business data, that's going to be value. And that's not threatening. There's no there there in terms of threat of AGI because no one's actually going to use generative AI to run things because of the hallucinations. And I think your power law that we're putting out there that we built is on point because the power law that we are talking about the size of the models, we're going to see an alchemy of that models. And what's going to happen is no one's going to put up with hallucinations. This is why the fear comes in. So the hallucinations has caused everyone to go, whoa, it's not ready for prime time. Well, that's not true. What you're going to see is special specialized models. And this was again, validated for multiple interviews this week I had with people that specialized models will emerge. I interviewed facts that big financial services coming day. They have generative AI in production. And you know what they're using it for? To make their data sets freely available in Amazon's marketplace as data as a service because they have good data. So when you have good data, you can manage hallucinations. So that's going to be the engineering trick. So my prediction is AGI is fantasy. That is never going to be an issue at least for a long, long time because the reasoning the reasoning is not that good yet. To be that good. So are people afraid that people are going to use AI to make decisions? That's a human decision. That's different. So are you guard railing the tech or you guard railing the expectations of the user who thinks it's real? That's going to be the big challenge. And everyone's going to focus on this problem and it is about one simple thing. Is it real? Or is it not? That's going to be the question and that's going to be the focus and not is AI going to hurt us and what's tech? Because a lot of these guys the big guys want to stall and prevent the startups from getting in. That's what Andrew Nning was saying is like it's and we called it out in the queue. Cue many pods ago. I said it's all a scam, Dave by the big guys to hold the market down. Okay. But so kind of ranting too early here. No, it's good. My other rant is going to be the CISO. But we did say that. We did. We definitely said that at the time. We thought Elon in fact, you know, who started that letter was sort of trying to slow down the market because he just announced this week. He's going to do something with X AI or whatever he's going to call it. But so okay. So your friend from Signal who's a super alpha geek, I have a really, I have a really great contact, a friend of mine who's I honestly agree with her. I honestly agree with her by the way. She went way over the top but she's kind of right. I heard her commentary. She was just poo pooing it totally. It is it is remarkable what chat GPT can do. So I mean, regardless of how come, how come you didn't invent it? So anyway, but but nonetheless, she's super smart. We kind of did. We kind of did with the QBI. But wait, wait, wait. So I have this contact. He's deep, deep deep inside the government. And and he's not he's not in in the government but he's a contractor. And I meet with him pretty regularly. We talk and he's like super into AI. Has been for years. He's the one who told me about Eliza. If you look up Eliza on Wikipedia, it's like in the 1960s. It was like, he said, this is just a better Eliza. Back then, it was like a mainframe. It was probably an IBM mainframe. People were interacting with it and they were like, holy cow, this is like magical. And his thing is, and he really knows AI, but he's, you know, biased, right? Because he's got a perspective. He said that this is basically pattern matching in a database and some powerful search. It's not a learning system. And that is what, you know, he sees as true AI. And that's why he's really down on full self-driving. He's like, full self-driving is like, you know, Elon says it's next year, next year of a year. It's, he says, he's used a line to me. He says, there's a reason why we don't start driving, allowing people to drive until they're 16 because the cognitive development takes some time. And so having said all that, when I was texting with him the other day, and it was somebody had made the statement like, like you just said, is don't worry about, you know, AGI. It's not anything to fear. And he's like, oh, yeah, it is actually. And because if and when it comes, and it will come, there's a lot to fear. And so that's why, John, I'll just share with you. That's why every time I talk to the AI, I say please and thank you. Just in case. Well, the over the top, the brain, Andrew was saying, the first of all, people are saying AI is going to cause our extinction. That's like overblown. But here's my point of that. So, I think it's it's an evolutionary tech trend. So the reasoning is just not good enough. And I agree with your friend there. I think it's just, it's going to get better. So here's an, here's an example. I was talking about this last night with a bunch of old and young, smart people at the Palo Alto AI event at the Lama Lounge at Jeremy O'Yan, Jeremiah O'Yan put on with Chris, just for Chrisie. And I said, back on the web in 1996, HTML was the format for web pages. Okay. Now go back, David, the time machine in your mind and think, 1996, 1997, where were you? What would you have? You had a PC. We all had PCs. We had Windows. And then how did you get to the internet? How did you, how did you actually get on? AOL. How? How? How? Which, what was the mechanism? Yeah. You had a modem. You had dial-up. Modem. Yeah. What was the speed of the modem? It was like 96. 19, 2,400, 9,600, 14, 4, K. So people don't, young people don't remember this. They probably heard it. It's like that. You know, the phone connecting. And it was like, it was so slow. I mean, it was like ridiculous. It was K kilobytes, K, B, not M, B, or G, B. And we were like loving it. Like, oh, it's amazing. They weren't gigamillionaires. They weren't gazillionaires. Yeah. All right. So, okay. So we weren't gazillionaires like that AT&T commercial, which I love. So bandwidth was a dial-up and web pages loaded slowly. And HTML was simply character-based. There was not a lot of imagery because of the graphics were slow to load. So the question at that time, and I remember I was working with on my first startup, which was keyword navigation. You type a keyword, go to a web page, which was a new concept. People were like, why would I type a keyword? I don't like a type of URL. Okay. That's now standard. One of the side projects we were doing in parallel, so besides keyword navigation, the name server was HTML, reduced instructions at HTML. So in other words, making HTML load faster, Dave, on dial-up. And so you say, okay, that's a really good problem to solve because if you looked at it, Dave, it said, okay, at that time, the problem was the page loads slowly. So I'm going to ask you a question. You're the investor, Dave. You're the VC. I got a business proposition. I have technology that helps web pages load faster on modems. Do you invest or not? Is that a good problem? I don't want to put you on the spot. Horrible. I mean, there's a horrible investment. I mean, looking back. Why? I'm just trying to think. The speed, my point is, the speeds get better. So that problem goes away because the next iteration modem is faster than you get broadband. So what happens is the entrepreneurs went to solve bigger problems, not that immediate problem that goes away with evolution of the technology. So the bandwidth gets faster. There's more faster pipes. There's more data centers. So as people came onto the internet, it was faster to load. It was more graphics, full motion video. The processors got better on Intel machines from 386 to 486 to Pentium. So my point is with AI, this is the same exact thing is going to happen. It's nascent and early. And you're going to see the problems that people are bitching and moaning about are not the real problems. The real problems is scaling with the onboarding of AI. It's a whole nother mindset. So these young kids are like, oh, that's a good way to look at it. Well, don't solve the wrong problem. What will be solved by the evolution of the wave of the category? Is compute going to be faster? So yes. Yeah, definitely compute. You're going to have more bandwidth. You're going to have more GPUs. Rags going to get better. More data. You're also going to have just a compendium of alternative processing power in terms of CPU, GPU, NPU, AI accelerators. All that is that the combination, the combinatorial factors. This is something to think about. We've talked about this before, but I'll repeat it. The combination of those factors that I just mentioned are blowing away Moore's Law. They're 2.5 to 3.5 times greater performance per annum improvement than Moore's Law. Now, power per watt, performance per watt becomes an issue, but that's why I love ARM. But I mean, we're talking about processing power that's 2.5, 3X that of Moore's Law's curve. So the curve's bending. Look at what happened with Moore's Law, right? Everybody used to talk about it all the time. Thomas Friedman, oh, Moore's Law, Moore's Law, Moore's Law, Moore's Law. Every time he came on TV, it's amazing. And it was amazing. But this is like, order of magnitude, more amazing. Well, I'll tell you right now, to me, the rise of the super cloud architect and operator that we're talking about is this foundational infrastructure. The middleware is going to be a big deal. Understanding AI architecture is going to be probably the biggest competition. I interviewed a company this week. They do data engineering, the fastest growing category in platform engineering, which is managing data pipelines. Okay. You did a story on Uber, right? This is the conversation again we had last night. Databases have to be integrated based upon what the use cases are. So you got time series, you got SQL, you got unstructured, you got object store, open source solutions are surging. So you're going to have a road ahead that's going to look a lot like multi-cloud, multi-AI environment. So an AI system has to emerge, Dave. It has to emerge. So to me, it's very clear that there are things that will solve this fear that AI is going to take over the world and destroy society. So I'm bullish on AI for that reason because the same thing happened with the web. Early on, it was very nascent. People poo-pooed it. Oh, the web's a toy for kids. It's not real. It's so slow. Look how elementary the graphics are. And then what happened was the utility of it was so powerful. As you pointed out with ChatGPT, it cannot be ignored. The consumerization of AI has made it a thing and expectation. And I believe every company and Snowflake just announced something this week too. I'm looking at our feed here and it's looking at angle. They're now into AI. I mean, a year and a half ago, like, AI is not really important for us right now. They're now got it, right? So you're seeing that, Dave. Big time. AI is everywhere. Why? Because everyone expects it. Call it AI washing. What do you want to call it? It's coming, just like the web did. So it's, again, this whole argument, it's a no-brainer. And I get nervous that the fear-mongering is going to come for it. Now, that being said, I want to ask you a question because we always rant about the government. Who's faster? The government, industry, or entrepreneurs? Let me think about that. Actually, I led the witness on that one. All right. So the government's not fast at all. So they shouldn't be even me involved. Even industry itself, this is why I'm kind of like watching these big so-called tech titans say, like Sam Altman, oh, yeah, we got to do it. I think he's full of shit. I think he's totally blowing sunshine up everyone's, you know what? And just trying to slow things down. So industry wants a stall. The entrepreneurial side of it's booming. So the question is, who should be responsible for the social aspect of how AI gets brought to the market? If you assume things are going to get better, like the bandwidth problem with the web, I think the entrepreneurs should drive this. I think let the chips fall where they may and have some sort of entrepreneurial code that says, hey, do the right thing, which is don't go out of business, make some money, but don't do AI for bad. So this is going to be very interesting to watch. I mean, that's my take. What's your take? I mean, I don't think I have the answer, but I do know this, that the government is not, is not going to be able to predict what's going to happen. And whatever regulations they try to put in place, I can almost guarantee there will be unintended consequences that will end up favoring some, more likely the rich, either rich get richer. And generally speaking, if not done properly, which oftentimes when it comes to big tech regulation, the government has failed miserably in my opinion, it makes the U.S. less competitive. I think it did that when it broke up Ma Bell. The European mobile companies, technology, telecommunications companies dominated for the longest time. Bell Labs is owned by Nokia as an example. IBM became a shell of itself. I mean, I don't think the government's track record has been good at all. Now, having said that, you definitely want the government to be in conversation with industry and with entrepreneurs and to have a clue, but they have to be very careful, in my opinion, as to how they act. This is the other thing about democracies, John, is that China can just say, you can't do this, boom. Jack Ma, you're in the penalty box. No ant IPO, done, boom, by decree. That couldn't have happened in the United States that quickly. So the question I have back to you is, is that an advantage for China that they can make proclamations and determinations like that? I don't know. China, of course it's an advantage when you have an entire government's orchestrating together, colluding with it as one state, is definitely an advantage. It's not, it's called communism, dictatorship. That's what they're doing and they're running everything. Yeah, but does that accelerate innovation or does it potentially start from innovation? They steal innovation. They don't do any innovation. They're stealing it all. That's what's proven. Look, the government should not be involved in anything. They're just, they just sue, SEC just sued, the solar wind, CISO, Chief Information Security Officer. Let's rant about that. You go, do your rant, because that's unbelievable. Well, let's hold on. Okay, let's rant on that now. We'll come back to the news on the AI regulations. But the SEC suing the solar wind CISO for knowing that he could have stopped the fraud. That is unbelievable. It's unprecedented. And by the way, why the CISO? Why not the data engineers? Why not everybody? And by the way, it was a private company. They're not really the government. It's in the private sector. And so this is just creates massive problems. It was no fraud. There was no other action. There was a hack. Maybe he didn't do the patch. He's got operational challenges. This opens a can of worms. And by the way, every company in security has to hire their own, quote, militia to defend their companies from cyber warfare. We've been talking about this in the podcast. I've been ranting for everyone who knows me. Knows I've been ranting howling at the moon for over a decade that every company has to defend themselves against foreign adversaries fighting a digital cyber war that our government is letting happen. And so what's even worse now is the government's flipping itself on its own people and companies. So it just opens up way too much problems in my mind. Because then why him? Why is he accountable? Why isn't the SREs who's in charge of the infrastructure? Who's in charge of platform engineering? Who's in charge of buying the endpoint protection? What's device? Who does only threat detection? Is it a threat detection problem or is it the data protection problem? Why not the CEO hired the CISO? It's just, I mean, it's just, it's so stupid because the government's job is to protect companies that are in the sovereignty of the United States of America. So it's like ridiculous. And so it's just, it's just so counterintuitive. It's just, I'm so like weeded out by this because it makes no sense to me. If there's some smoking gun in there, you know, maybe that he was complicit in it. Maybe, I don't know, but based on what I know and can see from our reporting, he was just a CISO. He didn't patch it. Okay. He missed the patch. All right. He ignored the flaw. But what does that mean? Right. I mean, he misled the public. You kidding me? The companies go out of business when they get hacked. He misled the public. It's not his job to be a public relations. Okay. It's jobs to protect. You know, and by the way, why would you share more information to the outside world when you're going to know you're going to get screwed by the hackers? Okay. They, they, they alleged that he concealed security fee failures that led to nearly two year long attack called Sunburst. Okay. It was carried out by Russian hackers, inserted malicious code into SolarWinds Network management software used by thousands of customers, including the U.S. government agencies and private companies. The SolarWinds hack was one of the most sophisticated hacks that was the, the, the, the supply chain hack heard around the world that look, think about what CISOs and SecOps teams have to deal with. This is a, this is a good rant, John, because the, the partnership between government private private industry and government is broken in so many ways. And when you, when you challenge government on that, you get a bunch of lip service, but just, you got to just got to look at the actions of, of government. And I'm not saying they should give free passes on this, but, but wow, that, that was a shock. So that, first of all, so I'm going to quote, I'm going to quote some of the, some of the text here from our story from our SiliconANGLE reporting. From the time of the initial public offering in October, 2018 until January, 2021, SolarWinds and Brown. Okay. Timothy Brown De, quote, defrauded SolarWinds investors and customers through misstatements, omissions and schemes that concealed both the company's poor cybersecurity practices and it's heightened and increasing cybersecurity risks, the lawsuit says. It goes on to say SolarWinds public statements about its cybersecurity practices and risk painted a starkly different picture from the internal discussions and assessments about the company's cybersecurity policy violations, vulnerability and cyber attack. Okay. They're under attack. Okay. He's freaking out. You know, loose lip sync ships. He's probably didn't want to say anything. Again, public company, time of kind of war, wartime. You need a wartime conciliary when you have these kinds of situations. Okay. This brings up a good point, Dave. When I was at the Mandion conference and you were at CrowdStrike, one of the things that came up was, was AI could be used for all this compliance stuff. Now, think about the compliance the CISO has to go through. Imagine getting attacked and you got to go in and report like file paperwork to someone. Imagine that your house is getting attacked and you went defending yourself, but you got to get on the phone and deal with some agency bullshit. It's like you got to take care of what's happening. Your house is on fire. Bombs are dropping. Bombs are dropping cyber warfare. And so what Mandy and I were talking about was, and this came up big time, you could actually automate a lot of the compliance. So I think this is where I was getting at the some of the AI things in my, in my essay. This is where, you know, the hidden treasures of data, you have the policies, you have all this stuff, you just automate the reporting. This was a human problem. He didn't report it. Okay. It's a public company. Maybe the investors should know they're under attack. Okay. Well, what do you do? I mean, this is where the security industry is really challenged because they were already having a hard time defending themselves on threats and protecting their data, right? Well, plenty of companies doing, trying to do it. And they're constantly fighting an uphill battle, like you've been reporting, you know, from the events. The bad guys are winning. And this puts another rock on their shoulders to carry in a war they're already losing. This is a great topic. I mean, the SuperCloud 3 covered gen AI and security. We have on December 12th, we have basically a SuperStudio event with Dell support where we're looking at the, what used to be an adjacency between data protection, like backup and recovery and cybersecurity, those two worlds are colliding. We've got some great guests to talk about this stuff. To your point about automation and AI, one of the areas where LLMs and generative AI can be super helpful is like just simple like reporting to your point, automating runbooks. People hate writing. Remember, like after a Zoom meeting, people would type up the notes. This is what we talked about. Now you just push a button. Boom. And there you go. And by the way, Microsoft's got actually some pretty good capabilities there as well that I saw. But so, you know, those are some By the way, Zoom has a highlight feature of your meetings. If you press a button and you're in the paid version, it gives you the highlights. Yeah, highlights, it summarizes the meeting. And I actually from what I saw on teams, I'm not a fan of teams. I trashed Chime the other day, but I used, I had a meeting. I used Chime the other day. It actually was better than I thought it. They must have made some improvements. I used to hate Chime, but it was actually really good. It was much more intuitive. Yeah, back on the Chime bandwagon. Nice. Yeah, I wouldn't say I'm on the band. I mean, you know, here's the thing. You get used to one, right? I'm on the bandwagon. I mean, I didn't say I love it, but I wasn't hating it. It's much better than I remember. And in teams I don't like, it's just not intuitive to me, but the summary of the meeting was kick ass. But well, that's a good rant, John. It's pretty scary. Yeah, we're going to put a pin in that. We're going to come back to this. It's going to be an ongoing thing. Again, that whole security thing brings us back to AI. We got there from the AI regulation. So let me just give you the rundown. I want to get your reaction to this. So here's the current AI regulation news for the week. It hasn't arrived. Biden signs an executive order directing AI companies to develop safer AI, which I have no problem with. In fact, one of our CUBE alumni, Reggie Townsend, Reggie Townsend from SAS, was in the White House. So shout out to Reggie Thompson at SAS and SAS team. I just impressed with the people over there at SAS. Reggie Townsend, right? Yeah, Reggie Townsend. Yeah, awesome guy. He was featured on SuperCloud. He's on the ethics side of AI. He's got a very interesting perspective, great video. Search Reggie Townsend at CUBE and watch that video if you're interested. He's really got a good voice on this. We're going to do more with him. David Strom on our team put up a great analysis on this executive order. It's promising. He writes, but it's going to be tough for the U.S. to govern AI effectively. And there's a lot more detail that he links to an Ernst and Young report. I think it's toothless in my opinion, but I don't want to put a wet blanket on all the good work that the people in the industry are doing. So I kind of support it. It's an executive order. Okay, just leave it there, whatever. Okay. The Office of Management and Budget put some meat on the bone as well. They released a draft of AI guidance for federal agencies. That's out. So we'll see what that looks like. And then in the EU, a parallel effort that actually started earlier, Dave, 28 countries signed Let's Lead Declaration of AI Safety. Okay, so this is kind of a whole another ballgame around the European side. The U.S., the U.K., China, and 25 other countries signed a declaration stressing the need to address the potential risks posed by artificial intelligence. So now it's during a high-pile summit in the European Union. So that's interesting there. And then finally, it's just we had the AI event last night with the start-up. So, you know, as far as the start-ups are concerned, it's more productivity stuff, Dave, to your point. I would say that I watched all the start-ups like it doesn't start as presenting. A lot of them were like about productivity, not a lot of game-changing stuff. So, maybe I just had a bad view on it, but I thought it was great from that standpoint. The thing that jumped out was the Google brain founder Andrew Ning who said the threat of AI is overblown. The threat of AI of human extinction overblown also went on to say that. I love the Kelsey Hightower comment around the AI being used on local machines as well. That's kind of the highlight of the news. What else is happening here? Oh, Google Cloud Vertex Search adds new enterprise-ready features. Snowflake added a slew of announcements. LinkedIn's got an AI coach, chatbot coach, they're getting a new job. The earnings run down. It was amazing. AMD was initially looked like they were losing ground, and then they came back. Lisa Sue was amazing. Samsung says its chip business is going to recover. Jayshree Ulaos got Arista Networks, Stocks Poppin, Commvault, Sanjay Merchandani doing a great job down there. Qualcomm had their thing, the Snapdragon Summit. You know, that's kind of interesting. Rapids 7 got hit negatively. Jayfrog beat. Confluent got hammered. Extreme Networks beat. Rapids 7 beat. Didn't think they had a revenue check. No, hammered up. Sorry. Rapids 7 did really, really well. Extreme Networks was the one that was saying that they beat, but they got hammered. Palantir is kicking ass. I thought Apple's earnings were actually better than I thought they were going to be. They had massive services. Do we have revenue decline? I know, but the services, to me, sets them up for the future. I love that. I agree, yeah. Oh my God. I do. Yeah, exactly. You know, Fortinet, that's a story of the first disappointment is rarely the last. You know, Informatica, saying things are good, but they did a layoff. Western Digital splitting in two. Broadcom still expects there's a VMware acquisition. That Western Digital story was very interesting to me, that Western Digital. Western Digital splitting into two companies, right? That, let's talk about that. Yeah, I mean, I just caught it, but they got an activist investor in there, Dave. And remember, they were a customer of ours and we haven't heard from a while. So it's funny when the cube customers go dark, something's wrong. That should be like a bell weather in our little cube index. Stock following. Elliott management, right? Same company that basically took down EMC and enabled Michael Delt. Think about it. I got to go to the quickest side. Can you imagine if there's no way that Dell could have bought EMC in this interest rate environment? Never could have happened. I mean, talk about like having the balls to react, to act when interest rates were low, which that's kind of my rant is, you know, it's Druckenmiller ranting on the Fed, but we can come back to that. But yeah, Western Digital splitting up, splitting the hard drive business from the flash business. Now, one would think it would make total sense to have those two businesses together, but Elliott thinks they can maximize the value of shareholders if they split them up. Kind of interesting. I think, you know, you know, that's a good point about the zero interest rate thing. Because if you think about that's almost like very junk bonds like 80s, what they did, what they're doing, right? So it's interesting how they could never pull that off here. But I mean, again, if you look at the earnings, Dave, look at the list here. It goes on and on. Well, by the way, we'll come back to that Western Digital. Yeah, it's splunked. There's a huge private equity gutting going on. I don't know if you noticed, but yes, the bond market, the interest rates that are due. I saw a bunch of graphs this week on Twitter from Trent Griffin, who's been tracking a lot of the bond maturity dates coming. It's going to be a bloodbath. And people don't think this recovery is going to come out soon. If you look at the stocks that we just talked about, the ones that are hurting in those software companies and the ones that are winning in the picks and shovels for the next generation. So again, if you squint through the landscape here, you're seeing the rise of the cloud collision that we're going to cover in SuperCloud 5, which is battle for AI supremacies on point. The web was a shift. AI is a shift. And the stakes are high, Dave. This is a massive wealth creating opportunity. This is a massive entrepreneurial opportunity. And I'm telling you right now, this is the beginning of a reset. This is going to be the great reset of this generation. You're going to see the winners and losers start to form and you're starting to see it right now. The old software companies will die. Shopify is doing great. You know why? Because everyone's using Shopify. Their headless system is booming. And so you're seeing the picks and shovels. Amazon doing well. Broadcom. I think the Broadcom is going to take a little bit of punishment. Maybe they take the medicine, but they're taking a massive combination with VMware. I think they're going to come out of that stronger. Apple, of course, they got services going to float the boat there. That could even go to a zero margin. Revenue model of services can continue the marketplace there. So I'm not saying zero, but like if services can you to grow, it's going to offset their hardware margins. Obviously services are high gross profit, as you know. Splunk is an acquisition with Cisco. I mean, all the signs are there, Dave. Right. Look at the winners. Storage. Chips. Observability. Software development tooling. Platforms. Anything that's got a platform that enables value and sets up automation with AI is going to rise. Well, it's interesting. Both Dell and HPE. I was physically at the Dell financial analyst meeting. I watched the HPE one remotely. Didn't get the invite to the in person or I would have gone, but Matt Eastwood went and I was talking to him about a little bit. Both companies are projecting stronger momentum in compute than they are in storage. And it used to be storage was the much better business. And I guess it still is from a gross margin standpoint in profitability, but compute with AI is now the bell of the ball, you know. But you were mentioning the interest rates and the bonds. So this week, it's not really a rant, but it's semi-rant. Stan Druckamiller was like ripped. Janet Yellen and the Fed basically saying, look, they had this opportunity. And he said, basically Druckamiller said, it started with Trump and Mnuchin. They basically, you know, I think ran a, I think he said a trillion dollar budget deficit. So they cut taxes. Economy was good. They could have, you know, paid down some of the debt they chose not to. So that ran up to debt. And then you get the COVID hit. So of course, they had to spend money. And then the Biden administration came in, spent more money. His point was that there were several periods of time when the Fed could have issued, you know, longer term debt, 30, 40, 50, maybe even 100 year debt. I can't say. But I asked somebody who would, who had an opinion on this, who was in the know, you think there would be demand for a 100 year bond? And the answer was, this person would know. Yeah. Absolutely. Governments used to want to, you know, rest and vest. Just park it. And so and so the whole, the point is that, that we miss this opportunity. So, okay. So now what do we do? And this is, this is kind of my rant. And Drucker-Miller was so right on that you absolutely have no choice but to attack entitlements. Nothing else matters. You know, it's defense, social security, and Medicare or the big three. You know, how much can you cut defense? You have to have strong nation, but you can cut some. But social security and Medicare absolutely have to be cut. You got baby boomers coming in now. It becomes insolvent in nine years and it's going to be just a massive portion of the federal budget, you know, by 2030. And nobody's willing to talk about it. Not Democrats, not Republicans, no senators, no congressmen. Nobody wants to talk about taking entitlements away. Instead, they're doing cost of living increases. It's madness. Yeah. Yeah. And so. I'll tell you right now, the counter that cyclical cycle around what's going to emerge on the debt for bonds and interest rates. Loans and bonds are coming due. Big time that's financing all this stuff. On the startup side, Q3 was a terrible numbers for startup valuations and round sides. And so, valuations and cash was bad. Down rounds were dominantly 20%. Down rounds were being prominent 20% of all rounds in the quarter. Nine straight months of high down round frequency means a reductive in valuation. Worst quarter on record, on startups in the card database which manages like over 1,000 rounds. And then median C rounds are down significantly. So here's the data, Dave. So if you look at the seed round, the body called price seeds, that's kind of the new series A. It's back to 2004 or five levels, Dave. You're looking at 50%, 50% title of all cash raised this past quarter for startups. It was 50% title, amount of cash raised 3.4 million. So 50% of the 500 to 1000 companies raised at the 50% title, 3.4 million in the pre-seed. On the 75th percentiles, 5.0. Valuations at the 50% title was 13 million. 70% was 20 million, pre-money. That's kind of like what it was back in the day, classic VC. Almost a full reset. Again, price seed was the old series A. Series A is the new B. So series A funding, 50% title of cash raised was 10 million. Dave, that's it. 10 million. High end was 17 million. So it's in the valuations with 39 and 60 million, respectively. We're seeing a very, very bad funding cycle for startups. It's really the worst I've seen. I mean, when you go back and take it to 2004, the last time I did a financing, I did a 5-on-8, okay? Or at that time, I could have probably got 10, but you know. A 5-on-8 post? No, that's, post-money was like 17 million. Okay, 5-on-8 pre. No, no. No, no, I think it was a 3-on-5. It was roughly 30%. It's in these numbers, basically. I think it was, it might have been like 10-13. I got to look at the numbers, but I can't remember. But this is old school, old school VC. So it's going to be very interesting to see how the angel market reacts to this, given that the AI startups don't match up. That's why the scuttle button, the valley is go small, small ball. Yeah, but I mean, but companies made money, VCs made money back then. It just got so stupid, right? And then the other thing that you mentioned, you sort of referenced as PEs, private equity firms. You know, as you know, VCs want to make, you know, 100X, 1,000X, 10,000X. I mean, absurd, right? The private equity companies, I mean, if they're getting 2X with late-stage investments, they're kicking ass because they're putting in so much dough. But, you know, if you put in money, you know, late in 2021, right? You're seeing valuations get cut in half. It's that sort of UiPath syndrome. Remember, UiPath at one point, and I think at $30 plus billion valuation, people were putting money into the series, whatever it was, series F. And, you know, now it's worth like $10 billion. So, you know, they're feeling the pain. Snowflake investors who came in late, like the Warren Buffets, you know, they got in prior to the IPO. The IPO went out at $120. Stock's trading at, I don't know, some of the $140s, $150s, you know, it's hovering around there, bouncing around. So, they're still making money. But who knows? That could change, too. I mean, you know, that's a stock that's still priced for perfect execution because of the Slutman-Scarpelli factor. Who knows about Databricks? You know, Databricks supposedly had this, you know, nice up-round. But who knows if they're even marking to market? You know, it's hard to tell. All these private funds that are in there, they're marked at basically either par with what the investors put in. I mean, I look at my investments and they're all like, oh, yeah, we get this much value and it's only down like half of 1%. And, you know, it's basically flat or it's up a little or it's down a little. But if they mark that to market, it'd be down by half. But they get paid on the current mark, which is not to market. Yeah. I mean, the economy is going to be a very rough ride. I'm very interested to see how these next earnings sees us and what comes out of re-invent, Dave. So, let's just quickly wrap up what we're doing on KubeCon, Supercomputing, SuperCloud 5. We got, really, I'm going to be essentially on the road this entire month. I think I'm home one week. You too. I'm going to Miami next week. Five days this month. I'll be in Palo Alto. I'll be in Chicago for KubeCon, C&C, Aff Linux Foundation. That's going to be always great. We've been to every single KubeCon, every single one since the beginning. I'm doing, yeah, that's awesome. And DockerCon, I'm doing two events next week. One in Miami, the Cisco Partner Event, and they're having a partner, I mean, an analyst summit as well. I'm going to see Ziaz will be down there. Carolina Melanasi said she's going to be down there. And then I'm going to IBM. Bob O'Donnell's going to be at Cisco. I'm going to see him at IBM. He's coming in a little later. I'm going to go in early for dinner because I haven't been to IBM's analyst thing in a while. So, I lost touch with them. So, and I'm excited about IBM, John. The data shows Watson X's. I think they finally got their shit together on Watson after years of pain. So, I'm kind of excited for that. And then, you know, re-invent's going to be big with SuperCloud 5. We'll get super computing. It's in Denver, super computing, where all the chip action is happening. That's the same week as Microsoft Ignite, which will have team coverage from Cube Studio. So, that's the same week as super computing. And then we're HPE Discover, which is the same week as re-invent. So, we basically have SuperCloud 5. We'll be pumping content from Barcelona. We're taking our previous Microsoft Ignite coverage and we'll be pumping in our editorial content from re-invent. We got Live Studio and Palo Alto, SuperCloud 5 people coming in. We just, I don't know if we put it out yet, but we're doing a call for guests. I know we've already got some guests lined up. We got a lot of questions on this. I just want a quick clarification and a little PSA. So, Cube will be at re-invent, okay, capturing editorial content. We have no sponsorship on locations. All our team's investment to go on site with the press area, getting all the top executives, but exclusive with Adam Sileski. We're going to have all great content, no paid sponsorship work on site, all editorial, our investment. We're going to put money out of our own pocket to be there. At the same time, we're having an in-studio program. Savannah Peterson, Lisa Martin will be hosting in Palo Alto a live stage performance for two days, Tuesday and Wednesday. Chuck Alley and the team will be putting that together. We're going to be broadcasting more content in Palo Alto. And what thanks to our great network of SuperCloud supporters, that's happening. And what's going to happen is you're going to see a ton of content. So if you want to come into the studio or be part of the program, we have a request for speakers, request for panels. We want to bring your own panel together. We can do that. How we should do that is going to do it again. If you want to talk about the battle for AI supremacy, that's the title of the SuperCloud. It's going to be a special edition. And we are super excited. We're going to have so much editorial content. We're going to be on the ground in Vegas, reporting back to the studio team. They're going to put a stream together in studio commentary and reaction from experts in our network, in our community. So it's going to be a unique first for theCUBE, a monster event. There'll be more content raining down from theCUBE than ever before. It'll be a reinvent like you've never seen. It's going to be chock full of content. Of course, our cubeai.com where it's going to be tuned up to help you figure out what the best stories are. So go to theCUBE AI and check that out, little plug. But type in things like what is platform engineering or platform consolidation? Is FinOps relevant? How do I do cost optimization? Things that have been set on theCUBE from our experts are now indexed and with fast retrieval with our RAG system, our automation generation. It's super great right now. It's getting better. And it's only in alpha. So check it out. So David wanted to clarify because they got a lot of questions. Are we streaming live from the show floor? No, we'll be sending content back to Palo Alto. And we're paying on our own pocket to be there because we want to get the most important stories and we'll do whatever it takes to get those stories. So quick clarification and plug for theCUBE AI. Next thing you know, you're writing songs for theCUBE. Like the Beatles. So my breaking analysis was this week. I'm digging in deep to the Gen AI power law that you, Rob Streche, and I have been evolving and getting feedback on. So I'm digging into that, getting into to retrieval augmented generation, RAG models, really looking at some of the things that we've learned, bringing in some of the ETR data, talking about sort of on-prem and hybrid AI. So look for that for sure. I'm going to send you some notes, Dave, because I have a long set of, I've been doing a ton of research on vector databases, vector and RAG. So here's some stuff there. I'm going to send this to you right now. Folks, thanks for watching episode 36, Dave. Any final comments you want to make before we break this down? I just, you know, I continue to believe that AI, I believe the hype that this is going to be, Gen AI is going to be super transformative, maybe not so much, you know, in the form of chat GPT. I mean, but that was the catalyst that has awakened everybody's sense of the potential of this new generation of technology. And I think I guess what I would say is, the key is going to be finding the use cases, making sure you got the right business model, and then showing value, because if you do that, you're going to get money to refund and reinvest, because right now the Gen AI stuff and the AI stuff is stealing from other budget items, and that can't last unless you can show value and then do share gains. I'm just excited. I'll just continue to reiterate, I'm drunk on AI still every day. This has been an amazing run so far. I think it's not even just scraps and serves. I think we're going to see an explosion of innovation if it doesn't get thrown the wet blanket of regulation on it. I think we have to let the, let the, let the chaos rain and then rein in the chaos to quote the famous Andy Grove. So I think this is going to be a seismic shift like the PC revolution and the web combined, bigger than those two forces. I see generation shift. And again, front row seat, Dave, like we've had for 13 years to queue, that's going to be great, great run. So we'll continue to bang it out. And oh, by the way, it's now public. Dave put it out there that the wiki bond is being renamed as the cube research and advisory. We're going to rebrand that site and continue our legacy of providing great content. So Dave, I know you've been working hard on that. Can't wait to see the news next week when it comes out, but a little preview there for the folks who made it this far. All right. Thanks, John. See you next time. See you guys.