IoT Coffee Talk: Episode 338 - "It's Just A Tool!!" (The Obvious Secret of AI)

[music] We could we could do air guitar. Yeah. >> Yeah. Come on everybody. Air guitar. >> Yeah. [music] >> Yeah. Steve, there you go. I'm just a purist. >> You're just a purist. >> Hey everyone, welcome to IoT Coffee Talk, the greatest greatest show. We're number one, right? >> This is the number one podcast in the universe >> in the world. In the world, you know, >> continues. >> Yeah. Yeah, I mean, as we mentioned before, you know, a few weeks ago, it was exciting to see the news that we had leapfrogged Rogan and then we leaprogged the Moonshot guys and we leapfrogged the All-In podcast, you know, Sachs and all those guys. >> And we're number one. It's because we're talking about what people want to hear about most, which is IoT, >> who knew, >> right? Yeah.

>> Right. Right. It's so hot right now again. >> Yeah. you know, I just heard so I'm here at uh in Las Vegas for AMD's um indepth thing. It's like a technical deep dive on their epic portfolio of, you know, data center CPUs and stuff. >> Yeah. >> And IoT was coming up, man. People talked about IoT, you know, and >> is that is that epic with a Y? >> Yes, [laughter] that's what I thought. >> Did you physical did you hear any physical AI talk? >> No. No, surprisingly. Interesting. Yeah, even though I I can't talk about any of that stuff because embargoed, >> but uh yeah, it it it's cool. >> I'll take that as a yes. >> Yeah, [laughter] it was like um >> well, you know, you have to remember I mean AMD's been in supercomputing for, you know, ever. You know, like the frontier supercomputer, you know, all these >> these [snorts] super computers, they they are the leaders.

>> That's that's a good AI stuff. >> You're right. We totally forget about that. There used to be that race of who what who had the fastest supercomputers. You always need like the top 20 or whatever. >> Yeah. >> And you're right, AMD was always up there for sure. Prior to this whole AI data center thing. >> Yeah. >> Um, so how many people lost their ass in the stock market yesterday? >> Oh my god. >> Ouch. >> Remember to take this seriously at your own risk. This is forformational purposes only. >> Entertainment. Entertainment. We are not actual stock brokers. Um, >> yeah, especially Rob. >> Especially me. But if you've been riding the AI wave, you felt the pain yesterday. So, sorry. I did, too. So, it's shared pain. Shared pain. We're feeling it. Semiconductors. [laughter] >> Neoc clouds. >> Neoc [laughter] we need to be we need to be reminded that the stocks market are not always things that go up. No, >> you me down.

>> What? >> Yeah. >> What? I thought it's always >> Oh, d >> rock and [ __ ] >> So, are you the kind of guy that believe that you can win at a casino? >> Yeah, that's Yeah, that's a good question. I have to admit cuz we go to Vegas often. It's kind of our second home for all of us analysts. And I I've noticed that I like I haven't gambled in Vegas in a long time, even though I'm there like every month almost. So, I wonder what that's all about. >> Well, you just prefer You just prefer to gamble outside of Vegas because you just said you're you're in the stock market. You just gamble. >> I'm gambling somewhere else. Yeah, exactly. >> And then prediction markets. >> So, >> yeah, that's right. But, but I got a call from Hawk Pan and he said, "Brocom's going to the moon, baby." [laughter] Oh, wow. Sorry. [snorts] >> Is it the new goal not Mars more than the moon? >> Yeah. Yeah. Yeah, that's true. You know, data center data centering moths. That's a great >> Exactly. Yes. Well, welcome everybody to IoT Coffee Talk. Look at this all-star cast on the world's number one podcast, Debbie Reynolds, the data diva.

>> You want to know about data? You know, I heard a rumor that without good data, AI is useless. Is that true? >> That's so true. >> Oh, wow. So, we still have to care about >> we still have to care about data. Is that what you're saying? >> Yeah. can't turn fresh into treasure. It doesn't work that way. [laughter] >> Being Dimitri, I would I would argue that it is not useless. It just reinforces the bad data. >> Yeah. >> So, do we still >> how does So, like if I was doing fine tuning or whatever, let's see. I'm trying to make my model smarter on my stuff. Is the data I'm giving it are people just slamming whatever data they got? Do we still have to like do data labeling and all that kind of stuff to make it work? >> There has to be some curation there, right? So, if you're throwing a lot of crap crap in, >> you're going to confuse the models, right? And so, >> um, I think I just read an article about, you know, these people like doing an example at Claw. They're having all these skills that they're trying to, >> um, you know, bolt on to stuff and it's just flattening the it's flattening >> the responses, you know, it's like it's just too much. That's too much. I haven't when when AI was more purposebuilt and it was very is more narrow in terms of what it could do and so that to me that's like the best result that you can

ask for but I think people >> everybody wants an easy button so they want know they want the thing to do every single thing it's like yeah it can do some things you know my in my view AI is a thing that helps you do a thing >> so it's not it's not the thing and I think that's what people are selling. That's where the problem is. They're think, okay, this is like the savior. It's going to do everything. It's like, no. >> Yeah. >> No, not you. >> Well, there's an I think there's two levels in the question you're asking, Rob. You so obviously there's there is, you know, ways to fine-tune the model itself. But this is the bar is becoming so high that this is more the the advanced lab that are going to do that. Now the way you apply the model to the question you ask is what I've actually been experimenting quite a lot in the past months and weeks and that so if you start building a corpus of all the things you know in a very structured way like you know you have your own competencies you have your own references you have your own data and you explain to the AI how all those files are organized this is where you start to have your very good value because otherwise the AI just compress based on your discussions you have with it you don't really know what it memorize of you And more importantly, you have no control over it. So >> the point is you have to be very organized in your thinking, your thought and your own personal documentation or your own personal digital brain and then when you point those model at it, you get very very good results. If you just expect to ask a very simple question and

the system give you a magical answer. Nah, [laughter] that doesn't work. >> Yeah. So >> and then sometimes sometimes and this has been true forever. Sometimes when you enter stuff in first, they think that that's the most important thing. >> Yeah. >> So, [clears throat] it's sort of categorizing things in ways that you may not be thinking about. You're thinking, "Oh, I'm just going to throw this in or that in or whatever." It's like, "Oh, I thought I thought because you threw this in first that this was more important." So, that's kind of where a lot of the weight goes. >> Yeah. And it's requires a big discipline because you have to be very organized in your thinking to to extract all over you. But I find very effective if you implement it well. It is it is actually very pretty fascinating. But >> yeah, as we said earlier, people going to go for this stuff just launch code and say, "Hey, tell me about this." And you're going to get some crap. >> Yeah. And people people don't people don't um appreciate as much the the impact that holes in data have. So you can have an expectation of an AI doing something for you and if you have holes in your data that means you have holes in the context and so the AI can never do if it never do simply because the data doesn't exist what you expect it to do. So, I mean, you know, here's the sad thing. You we have so

many people going around, you know, uh they're just going around, you know, making platitudes about data. It's like, well, you know, it's all about data. Have you ever worked with data before? Do you even know data? Do you know >> Yeah. It's it's a lot more complicated than that. And you know, other >> Yeah. And like this ontology stuff. I mean, >> yeah, >> when you start to get up at that level. Yeah, I know. And you know, >> um, you're you're at a different level there with knowledge management >> and, um, not dealing with knowledge. So, when you have like these gaping holes, >> yeah, the the AI is going to be limited. It's just not going to do what you expect it to do. And that's why their stuff works a lot better, like you said. >> Yeah. will make inferences for things for the gaps, right? And so you may not know that. So I tell people [snorts] an absence of data is also data in in models right?

>> So you're not thinking of it that way, but yeah. And then that and one thing I always tell people, you know, a lot of the data systems aren't good at not >> so you have to really come at it with really affirmative thing. You know, do this, do that, and say like don't do this, don't do that. It's like it's not >> very and has never been. >> Yeah. >> Yeah. [clears throat] And that's why IoT is now becoming such a you know topic of books because people are looking at signaling the gaps, right? So taking uh vision models and uh images language models and trying to increase their um you know uh the data that they're collecting and fill and fill in right >> you know >> I was going to say just this week uh the international standards organization announced a new standard for sharing data between manufacturers and different equipment and the big reason was uh if all of this data is fragmented, the AI never is aware of the gaps um let alone the common uh commonalities between different pieces of equipment in different locations. And so they just announced it this week, but you think that that would have been the norm.

>> You would have thought. Yeah. Wow. I I'm having an epiphany right now >> looking at all of us. We're recording live from our respective senior living centers that we live at right now. And [laughter] and the reason I the reason I'm saying that is I'm starting to feel like we're just like all those people at Y2K, all the Cobalt programmers that they had to bring back from the senior living center and they had the walkers and stuff like that to come save the planet from Y2K with all the cobalt with the twodigit dates instead of the four. >> Now that IoT is back, it's hot. It's the new cobalt, baby. Yeah. And and and so the space cowboys and cowg girls are here to come back. They pulled us out of there. We get some shots kind of like they give NFL players at halftime to keep them going. And so we are back, baby. >> Yeah. >> The the sad thing, Rob, is about this era is that yeah, we we we we had to get a few people out of their cobalt retirement, but unfortunately the majority was moving to SAP or a ERP.

[snorts] So that's I mean the RP vendors made fortune around Y2000 selling new solution as opposed to fixing the old ones. So >> question doesn't totally works. But [snorts] >> I have a question. So you all seen in the news uh lately about how Europe they had the cyber attack where they were attacking the refrigerators, the stores and stuff. >> And I don't know about you, I'm like waving my fist at the sky. I'm like, do you not listen? Like we've been telling people this forever. That was a bad idea. >> My opinion deb on that specific incident is that if you're stupid enough as a manufacturer of fridge to leave a back door like that in your system, you deserve to be hacked. I'm sorry. You deserve to be hacked and you're just the biggest idiot on the planet. >> And since this show where we assume that you're an idiot on this show, cuz we're all idiots calling in and so I'm I grant you all that. I'm assuming that the whole planet is running unpatched old [ __ ] >> and and it's because I heard something about humans the other day. What is it?

>> Oh, they're lazy. Yes. And they don't want to do stuff. [laughter] So therefore, that's why we're here to rescue all you lazy ass people who are using old [laughter] [ __ ] and you don't patch anything and you expose that. So what what we're going to do though is we're going to use this magic of IoT and we're going to give you some air cover. And so with IoT, all those idiots at those grocery stores across Europe, if they had like I don't know what would it be like some kind of temperature sensor monitoring >> sensor. Oh yeah, a temperature sensor like going to IoT and we send that message. Even if they were unpatched, they would have had something we call early warning, which is different than predictive maintenance, but it's going to be good enough in this case. What if it was sending a temperature, you know, every 15 minutes or five or 1 minute? They could have nipped it in the bud as we say here in America and their lazy asses would have rescued all I mean Steve do you have any I don't know if you've ever heard of this scenario before Steve I wonder how much >> I read about it pulse checker pulsechecker.com I read a little bit about it >> really oh >> yeah it's a remote remote monitoring of equipment legacy equipment there you go really and and and Dimmitri the big fear that that I think will come out from this is this was at the manufacture level that there was a back door and that the that the businesses who owned it had no idea that it totally and the worst on top of that is if you had pointed code to the firmware it would

have found that in a millisecond and what we're actually trying to do is we're trying to prevent researcher from pointing AI to firmware. It's becoming a big controversy. We have a the group of hackers friends I have you know we we're all playing fun with you know connecting our TV to code and figure out what's in the firmware and they're trying to prevent us from doing that because of course protect protect they want to protect their IP and on the other side they have an idiot that could back door whose password is so predictable that any embassy on the planet can find it out so there's a problem there so there's a think about think about what we're discussing right now this week and two weeks ago we were talking about the Same thing for PHP, right? PHP, >> the $50 billion a year in commerce that is at risk because of >> something that nobody had a clue about. >> Wow. >> You know, so I mean I I would argue that in this case, you know, it is kind of your ancient language, poor memory [clears throat] management, poor protection, blah blah blah blah. So it's a different vulnerability from a developer knowingly putting a back door with a stupid password. I mean that that's pure human stupidity. That that's mean it is pure.

>> We got a lot of that. We got a lot of that. >> Your memor memory overflow are known problems. Yeah, they could have been detected as well. But I mean I give them a little pass on the PHP side. I don't give them a pass on the fridge side. >> Well, it's like I I almost bristle at the idea. They call it a air quote cyber attack. It's like it's just like stupidity. Like it's just dumb. I mean, but but I'll say I'll say this. Um, >> we have made a conscious effort in the last year to move to um non-Chinese technology because of different intel that we pick up of things that can be done remotely to take down equipment, right? And you're looking at a great example of it. And and so yeah, the fear is that so many things are manufactured in China, different components, one out of a hundred shipping containers are actually physically inspected.

And if if somebody wanted to start doing some really dastardly things, you could cause some havoc if you took down all the refrigerators simultaneously. And that's >> stupid refrigerators. Think about some of the other things >> people can starve. >> Massive starvation. Wow. >> And and you know it's a very complex problem because uh I give you another example of where the complexity of the issue is you you might remember that at some point Apple started to say that they were going to sign some of the subcomponents in the phones and now the whole community of repair people say oh this is unacceptable. >> Yeah it might. But it guarantees you that when you're plugging this new camera which has been signed by Apple it's been inspected. It's been validated. you plug in a camera you bought on Amazon, on Alibaba, Express and you don't know where it comes from. You don't know what software is in it. So, it's a very I think it requires a very deep rethinking on the way because I was thinking when we started this show about IoT, I think because we we keep on saying the internet of things.

>> It's kind of becoming the internet of everything because there's >> love you. [laughter] >> Wait, wait, is is that Marbel Lopez calling in from Cisco? [laughter] >> I don't know. No, but my point is my point is there is software everywhere. There is microcontrollers everywhere now. So everything has a piece of software running. >> Is it time to rebuild? Is it time to rebuild everything? Like use use AI to build better this time. Rebuild >> you know where where are the humans? Where are the humans? This is what I'm saying. So let's say your refrigerator is going down or whatever. Like disconnect from the internet and plug it back into the wall. >> Yeah. What happened? >> No, no, it's got it. It's initiated an embedded It's initiated an embedded command on the device as long as the device is plugged in. >> It says run at full power until this thing dies. >> Well, what the thing they did is that you know that all fridge have a defrost function. So the defrost function heats to remove the ice. That's what they turned on >> because you can ask a fridge doesn't do heat. Yes, a fridge does heat to defrost. You know, Steve, when you were saying there run full blast until it dies, that remember when you invented that whole stuckset deal that took down the Iranian centrifuges >> and you said run it fast until it dies.

Yeah, >> we can't talk about that. But by the way, >> sorry, top secret. But it was but we know enough that it was not running as fast as possible. It was variating the speed. So it creates stress and break. Yeah. It was fast, slow, fast, slow, fast, slow until it breaks. That's what it's >> like thinking fast and slow. All right, folks. >> What What's going on? I just I just got on an alert that uh the is the market crashing again? >> OH, SHOOT. NOT AGAIN. OH MY GOD. GET BACK IN THERE AND SELL. SELL >> SELL. BYE. BYE. BYE. >> PLEASE. I just got it. >> Elon's going to look like it's up. >> Yeah, Elon's going to lose billions again so bad. >> OH MY GOD. ALL MY BITCOINS are gone. It's the blue wave. We're talking Oh my god. >> Talking about bitcoins, we had another big >> What's up? Okay, a false alarm. [laughter] >> Anyways, going down >> false positive. Talking about bitcoins, we had also the Binance thing recently which people should realize that if you imagine that you have a bank when you have a safe and usually in bank in safe in banks there are two keys. One key by the bank and there is your key. Now what if you leave your key in the vault of the bank?

>> Yes. >> Which is exactly the way those crypto exchange work. So yeah, it's comes easy to be >> I'm still I'm still scared of digital money. And I say this mean you know it was a hundred years ago or so that the US government went door to door and they asked for people's gold you know and they took it. Imagine if they didn't have to knock on the door. >> That's what is that's scary. Unless you have one of these where your keys are. >> Yeah. You are at risk if you leave your keys because this is what happens with all those crypto exchange. They have your private keys. So if somebody finds a way to go around the private key because the binance that even needed the keys because they just attack the thing that was doing the transfer automatically. So yeah. >> Wow. So has anybody >> this is actually why I believe we we don't have the technology to make crypto at scale for everybody because either you rely on centralized systems that manage your keys and you're risking that or you need to have those things which are this is difficult to use.

>> Yeah. Yeah. Yeah. You know and you know this is this is I mean this is why people have to listen to IoT coffee talk. We brought that up like >> the first year we started this show >> when everyone was going bonkers with Helium and all this other crap and we got into debates about whether or not uh uh you know crypto is secure and >> I remember how it happened. I remember Leonard called me one morning. He's like, "Rob, I had this dream last night and I woke up in a cold sweat and the dream's like, "You've got to warn the people. You got to tell them. you need to create a podcast that will become the number one pod in the world and get the word out, >> you know, and he's manifested. >> It was Yeah. And it was like, we got to do this. We need to get on the interwebs and tell the people >> and we did. >> I think it's I think it's interesting that the things that we've predicted have are literally happening. It's like, are you all listening at all? At all? >> Yeah, >> I know. Well, luckily we have tens of billions of people. Do we have I mean not billion.

>> Well, tens of billions of agents are listening to our show. [laughter] >> Tens of billions of people. >> Let me let me just let me just double that, Rob. I'm I'm running the agent. I'm d >> Hey, you know what? Everyone's everyone's making [ __ ] up anyways with TAM. So, go ahead, Rob. [ __ ] yourself. >> I've been I've been hearing some crazy lately. >> Billion humans on the >> Somebody said something like, [laughter] you know, whole planet, everything Tam. Yes, exactly. Hey, so people are making stuff up, man. They're just stuff out. >> On that point, on that point, I'm sure you all know that, but I don't know if our viewers that there's this whole discussion about how much the AI is driving the US economy. >> Yeah. Which is your super big. Well, we're building the data. We're building the data centers. So, it's the building chips that are driving it, not the use of the AI. [laughter] >> That's right. Just the the invest. What do you think, Steve? I was just going to say, Leonard, I don't think that that people would make things up, especially about our podcast because our international fan mail just keeps climbing week after week after week.

>> We never have time to do real work because we're going through fan mail. >> Listen, if it wasn't for AI, I don't think we could translate all of the internet languages. >> So, speaking of AI, have you guys heard this? Here's the hot new thing. I started hearing about it maybe a month ago. Jev J V Jev. Everybody on the pod know what Jev is? >> Wow. Okay. >> Rings a bell. >> It's a hot new thing. It's it's it's like the succinct way of doing kind of like quick yes and no or number responses with AI and it's super fast. Also, we are still looking for >> we're still looking [laughter] for some people who want to go ahead and check your baggage because we're not going to have enough room. >> Okay, so Jev is an AI decision model developed by Typesafe. I love type safe code as you do. Unlike text generation models, it responds directly to situations with yes, no probabilities, multiple choice selections or scores, and just designed to be faster and cheaper than traditional LLMs. I've been hearing about Jev for maybe about a monthish. Um, and so >> I don't know, I don't know, a bunch of smart people, I guess.

>> San Francisco Bay, >> of course. Of course, it's some from San Francisco. You know what they don't realize is the real intelligencia comes from >> Louisiana. >> You didn't know that? You didn't know that? >> Try. Yeah. >> I'm going to move to you. >> She go Tigers. >> Go Tigers. >> A lot of times that a lot of times the things that people are doing are asking are simple. So, is is the is is uh the Kimbe Matumbbo taller than Spud Web. Yes or no? >> That's right. That's a great example. You're right. I love that. That's a that's a good point. And so, I think there's it's growing in popularity. Some person built it and now all the big AI players are basically copying it, right? Is what as they do. >> Um so, yeah, you know, >> well, the the real innovation would be if these models start saying no, I don't know. That would be >> Wouldn't that be a godsend? I'd love that. Wow.

>> That would be lovely. Wow. Am is my printer out of toner. >> Yes, it's out of toner. You know how I know that? Not through AI. I tried to print a page and it didn't work. >> And probably the most the most intelligent answer I'm waiting from an LLM which I've never had is >> no. I don't know. Go read the book. >> Oh, >> yeah. I think as well the with the AI the issue with AI is everything that we talk about is gloom and doom and in the amount of time that's passed since AI came to be. Uh there's there's no increase in the uh the effectiveness of our public school system. There's no there's no outpouring of medical benefit that's you know live and real and coming down uh soon. Uh it's not attacking problems. That's what I think is hold what makes it so fragile in the marketplace right now. It's all just so much hype >> and like yeah real problems. So it's like anything that was really effective.

You have to really be able to solve real problems that people have like the start the smartphone would not have taken off the way it did if it didn't literally help people do things they wanted to do. So like if you're not literally helping people even like this data center thing it's like okay >> these communities they're like okay so we have already seen that you're finding all these humans and you want to have robots for everything then you want to come into my community and steal my resources and you know you're not bringing jobs right cuz all robots are going to uh you know run these factories and stuff and it's like you want tax cuts that we're as a consumer we're going to end up paying for. So, how does this benefit me? It does. >> Oh, I am feeling the rage all the way through this internet big wire [laughter] that we're connected by here. And gosh, it feels like some kind of wave or a tsunami. I don't know. But people are >> I would agree with that. I think people are realizing exactly what you're saying, Debbie, which is, you know, it is actually public funded, but it benefits an extremely narrow portion of the population at this point. And people >> Well, we could say the same thing about the internet. The internet was publicly funded and all it's been all it's mainly used for is monetizing us.

>> Yeah. >> It's gonna connect the people and bring us together. >> But I think it's a it's a question monetized. It's a [laughter] question. >> I think it's a question of vision and leadership because Steve I I would actually I would argue about your your statement about schools. I do believe it because I do believe that AI if implemented if implemented properly for education. I mean not replacing teachers but replacing the boring portion of teacher which is just the basic download of raw information that could be automated and made actually much better than teachers and then have time to encourage critical thinking in very small group with smart teachers. So if we were to rework the whole education system leveraging AI for what it is good at, I think we could see an impact because the point I was wanted to make is that I do believe people are dealing extremely wrong the way they're using AI because they think they can throw complicated problem at it and they're going to get a simple answer. I don't use AI like that. I use AI as an assistant and I work very collaboratively with it and I am much much much much more productive than I was before. For me, it's probably 10x, but I don't just drop one sentence and expect it to do the whole messaging and positioning document. No, I spend enormous amount of time doing the context and iterating with it. So, it's kind of this people are trying to get it be lazy using it. That's the problem I see. It's like you have an intelligent smart professor you have 20 minutes to

talk to and you're just asking two stupid questions. >> Yeah. I mean, I might even go so far as to say, how could we use AI to reinvent the public education, you know, system? >> Make it how how might we use it to recommend ways that are interesting to the students that have no interest in >> totally. But that's that's different than educating the kids. The problem is is you have these folks who think that you need to use AI to educate the kids or have them get educated on how to use AI and >> it's no not really. Um I think learning there's friction in learning and if you take it out then you get you get dull people. Um when you make learning too convenient you don't act I don't think you actually learn. You don't retrain stuff. I mean, think about how we grew up and all the friction we had to deal with, how difficult it was to uh, you know, uh, do large scale analysis. Well, you had to number one think through uh, a lot of stuff, but also figure out how to scale it, right? Maybe the scaling part eventually AI can do, but going through the cognitive exercises and then burning, you know, knowledge and um, you know, expertise into your brain, whether it's for math, science,

>> like the Matrix. >> I know kung fu. I can fly hell. >> Yeah, but but I but I would say like when's time when's the last time you hand calculated something on a piece of paper? >> Well, of course not. I'm I'm not going to do that. But, you know, when you know, if a teacher uh was uh telling my uh six-year-old, "Here, take a calculator and learn how to do arithmetic." I told I'd have a conversation with that teacher. >> Oh, true. >> And it would basically be, "What the what the hell are you doing?" >> Yeah. >> Teach my more of the 16year-old. Oh, you know 16 year at that point. Yeah, you can you can start to bring in tools but based on uh you know the where the where people are in their learning uh continuum, right? >> Yeah. >> But there there are people literally who think that these K through 12 kids need to learn how to use AI today. And it's like look, no, this is a tool. It's going to change. It's going to be different. Right?

>> Focus on having your kid learn how to learn stuff, not how to learn AI. >> Focus on your kids going out and riding bikes all afternoon and coming home when it got dark outside. >> Yeah. >> It's learning to But we're not suggesting any of that. I totally agree with you, but I think >> No, but but my point my point is there's a lot of people who think that. >> Oh, yeah. And and there's a lot of people who should know better who think that. >> And that's the thing that bothers me and because people are so becoming so AIDS and upset. >> They think that it should be everywhere and everything. And that's it's absolutely delusional. >> And that's how you get to idiocracy. That's how you get to idiocracy. And we're going getting there really quickly. >> Yeah, we're getting I would agree. We're getting there faster. And and to some of our earlier points, I mean, the fact that the stock market's going and making so much wealth around it is also making people believe more. It's you get a big get a big bubble crash and people will start questioning if they need to use AI or not. But as long as it's going up, oh yeah, we did more here, more there. My stock is going up.

More AI, more AI. Well, the thing that kills me is people now promoting um prediction markets as predictive tools. It's like no, they're not these are just people thinking gambling. For all we know, it's just bunch of people b getting betting on the side and they're completely ignorant of the topic that they're betting on. I mean, it's gambling market. >> That doesn't mean make it a prediction market. Poly Market is the new is the new Poly Market is the new cocaine. >> Those those guys on the All Allin podcast, they are always I think they're sponsored by Poly Market actually. Poly Market told me that D. So blah blah blah blah blah. You know what? Here's what I learned a long time ago since I'm so smart is that I at one point I decided >> you are Rob, you are >> I really just want to focus on the big big problems and I determined that I don't need AI to do that. So, I remember I'm gonna say this is 2010 or something like that.

You may not have noticed it. There was this big drought kind of in the Midwest. I remember the cover of like Time magazine and all burn stuff like that. And then as someone who grew up in Houston, even though at the time I was living in the Seattle area, I remember all these companies down here like Enron and others and Buo and a million others that have built pipelines for natural gas all over the country. And I remember sitting there thinking, huh, could we solve this really big problem of not getting enough water to grow crops all around the country? And I was like, "Wow, I think we figured out some basic stuff. We know how to desalinate ocean water. And I know it's energy intensive, so maybe we'll have solar and wind and maybe little nukes or something like that. And then we know how to pump water because we can pump oil and natural gas all over the country because we do it every day." And and I remember just drawing a picture of pushing all that water from the coasts into the interior of the country over the Rocky Mountains and everything and writing out how you do it. And to me, I was like, that's a big problem to solve that could really help a lot of people. I don't know if AI is helping people in the way that something like that would help. I don't know. I'm not seeing it. I'm not >> problem is you Yeah. Then then you have to introduce the the real people factor to do it.

>> Yeah. Because there's no AI for that. >> But the flow in your think the flow in your thinking, Rob, >> of course, here comes the flaw from Dimmitri. >> No, but the No, [laughter] >> come on. Bring it. Come on. Bring it. >> Who do you think I am? No, >> I know who you are. No, the reason I'm not pointing at you, Rob, but I'm pointing at, you know, pretty much everybody that Leona mentioned is not necessarily you thinking deep about it. When you say, "Does AI solve this problem?" AI is not something that will think on its own and going to wake up one morning and say, "Oh, there's a water problem. We're going to have to solve it." Now, I bet you that if and actually I would encourage you to do that, Rob. Do the same exercise. try to erase what you did 10 years ago and do the same exercise with Clude on your side as an assistant. >> Okay? >> And I bet >> I don't doubt that it'll give me some >> and I bet and I bet you'll get you'll get something as good as what you had before. Plus, you might have even more ideas on how to make the point valuable to more people and carry your ideas away. So, this is back to what I was saying earlier. I think people would think, oh, we're just going to ask AI, oh, go fix word anger and hope we're going to get an answer. No, it doesn't work that way. But now we have access to a certain form of intelligence which I believe if well used by human can increase our ability to make impacts.

But it's not going to fall out of the >> it's not a data center wake up one morning send you a prompt. Hey, we're going to solve that for you. But >> let me let me finish real fast. I'm almost done. Then you guys talk away. At the time I remember putting this out there on the internet and I had some people like, "Oh, that's really neat." and everything and they're like, "But this would be the most expensive undertaking in the history of America. It may cost a trillion dollars and we would never spend that kind of money on anything." Fast forward to today. [laughter] We're spending trillions to put water sucking, energy sucking data centers everywhere. And it's like, >> yeah, well, you [laughter] know, you know, it's funny. >> They wish they had that water there. >> Oh, they wish they had that water. You forgot inflation. Rob, you forgot inflation. >> Oh, you're right. That solves it. You're right. Yeah. Yeah. >> You know, actually, actually, Rob, you can see that the people with resources, they don't want to put it on things like that. So, I think Earth, they're trying, the message is that Earth is for poor people cuz they want to colonize.

>> Oh. Oh my god. It's going to be like, remember, Debbie, do you remember that movie Alisium or whatever where they had the orbiting space station with uh Yes. with Jodie Foster. And you're right, all the masses of people around the earth just like scavenging for being alive and whatever. >> You might be. >> It's actually a pretty good point. It's a pretty >> I think this is the problem. >> Scary, but good. [snorts] >> This is the problem. We call it intelligence. Um, it's not really uh it's a tool. And think about it. Um, without the questions, it's completely useless. It doesn't do I mean you could you could train it with all the world's knowledge right if you don't have the prompt there's no output it's just a tool and even with the reasoning models even with the reason model same thing reason models is just like an agentic framework it's just >> is it really reasoning is it really reasoning it's trying a few things a few times and going oh okay it's >> and then it'll go through a chain of thought and then sometimes it'll confuse itself Actually, a lot of times it confuses yourself. And I keep hearing that, by the way, which they haven't solved that problem >> or it goes in the wrong direction.

>> Hallucinations. >> Compound hallucinations. >> What do you think? >> I have a simple question. I have a simple question. >> Is a hammer smart? >> No. >> There you go. >> A hammer is already there. >> Capable, right? It has capability if you use it right. >> Right. So it must be so right. So AI is not smart >> but if used in the right way it could be capable but people are trying to make it something that is never going to be >> right about the question. Yeah it should be okay. >> But Leonard it should be okay for us to say AI is a tool. We are getting beaten up anytime you say AI is a tool or it's a research assistant or whatever. the so-called intelligencia on Sand Hill Road is like, "No, no, no. You're just taking baby steps." You know, they're like, "You need to be AI native. Your company needs to be AI and the whole thing needs to be powered by AI." If you're just using it as assistant, that's just like using it like search or whatever. And they've been >> that goes back that that goes back to what I was saying. You said San Hill Road, which validates the hypothesis that the problem is we're trying to figure out how to monetize humanity, not help humanity.

>> Yeah. Exactly. Exactly. Yeah. >> Yeah. I at that point I think you would backlash because I do believe that the jobs that I would be the easiest to be replaced are actually the VC job in the seauite. I think this is the one that going to go once people. I mean that's that that sounds and feels really comfortable and nice for a lot of people but I don't I actually don't think so. >> The reason I'm saying that is because if you the majority of the the work at this level is execution. >> It's not vision. You still need the vision and the leader to drive the system, but all the rest is pure execution and can be totally done. [clears throat] >> Tell me what a CFO does that AI cannot do. Tell me give me an example. >> No, it it it's experience and uh discernment. >> Yeah. But it's the but it's the person it's the driver and the leader that gives the direction. Then I mean no, let me let me tell you something. There's people who are saying that analysts are going to go away because of AI. Guess what? I personally believe no no but >> but still I'm when I'm I'm using this as an analogy people say that analysts are going to go away because you have AI.

>> I don't think so. >> I agree. >> Why? Because AI generates so much [ __ ] You know who what you know buyers are more now. >> What about software developers, Leonard? >> I don't think they're going to disappear either. We need more. We need more software developers. >> That's not what Microsoft is doing. >> No. No. But they're because they they've screwed themselves. >> But you can know I mean they're in trouble guys >> software engineering not you know coding is only one activity in software development right. Yeah, that's that's what I wanted to say. >> And then also um the the fact that we actually analysts who are engaging with organizations leadership and the technical technical folks within these uh organizations, we we get grounding through our interactions, right? And we have the opportunity to figure out what's re real and what's not.

>> AI doesn't get to do that. It's very difficult. You can you and the and the biggest crap that people feed into AIS are press releases now. These things are complete [ __ ] you know, >> and and you need you need people with discernment that can go and figure out all right, what's real and what's not. And so that's why I'm, you know, we start off this conversation with uh how IoT is back is because there's gaps in gaps in the information and and so people fill the gaps and there are going to be roles in organizations >> and you know who fills the gaps are going to be critical in filling those gaps. >> Bill Pew, he's the one who fills the gaps. The gaps are the ugly. >> Yeah. [laughter] But I I think there's a I think there's a pattern [clears throat] between this discussion and the what we had on education. And I think that what Steve and I suggested because people look at the way we work today and they think, "Oh, AI is going to replace that. AI going to replace that." I don't think that's what's going to happen. But it's going to redefine the way or give us the opportunity to reconstruct the way we do things because I want a challenge on software. We talk about developers and software engineering guys. We talk about PHP, we talk about Binance, we talk about the fridges. You tell me that we have mastered the art of making software. Hell no. So, and I think that AI is going to give us the opportunity to reshape the way we build software.

>> Improve the quality. No, I mean improve the quality. >> That's the point. So, what what what some companies are doing which is just purely replacing developer by AI, this is going to be totally stupid. But the one that on the other side said, "No, I don't want you to use any AI because humans are better at doing software." are wrong, too. Because we've proven that can do shitty software as >> and and and you come to that real collective realization when you realize it's a tool. >> Yeah. >> More people more people are doing that which is a bright spot. That's >> but there's more and more consumers that are connecting their banking information, >> username and access to >> I mean I'm just saying like people are doing things that are >> Yeah, but that's that's we also start Yeah, we also started with that topic too which is you know most people are quite stupid and lazy. >> So >> they're lazy so they do a lot of stupid stuff. You know what? And we're going to get stupider, more stupider, and more lazier if we keep letting AI do other things for us. So, here's >> for your uncle Rob here. Here's my recommendation for all of humanity. When you're sol when you're faced with some kind of problem you need to solve, don't just go quickly to the AI to get the answer. I encourage you to try it yourself first. Give it a try. Whether that's mathematics, I know we don't use calculators ever, but try it. Try the hard thing. Yeah. Yeah. Because I I know I've allowed myself to get sucked into,

you know, when I'm writing code and Claude's super powerful, but sometimes it goes in a direction it's like that's not really what I was thinking. >> And I remember getting exactly what I wanted when I, you know, it I' I don't want to call an epiphany, but lately I was like, I kind of like the way I was doing it better than that that sounds egoish but >> try it. No, I want everybody to try being hard things because your brain will atrophy and you'll be like Wall-E on that spaceship, you know, like, >> you know, and I don't want us to be idiocracy wall-ally people. So, please try first before going to AI. And yes, AI is just a tool. It's just a freaking tool, right? It's just sitting there. It's not conscious. You know, when they try to act like it's alive like an employee, which is what they want. It's cuz it's an agent running in a loop. It's called software. Remember, AI is software. Agents are software, okay? And they run on a computer just like stuff we've done for decades, right? >> And increasingly on CPUs. >> CPUs. CPUs are going to be the darkhorse winner in all this because agents are software that runs on CPUs. Yes. So, um, much love.

>> They all they all need water, but they can't swim in it. >> Oh, genius. Wow. That's a that's a good quote. You know, the robots are going to be capable of swimming. So, >> wow. >> Really? Yeah, probably. >> Maybe. Yeah. But >> look like a boat. >> At first they won't, though. They'll throw a humanoid robot and you'll see sparks flying and all that kind of stuff. Um, so anyway, please try. Always try first. Whenever you got a problem, try first. Do you know now I get though when part of like if you're doing research to get the answer to problem and you don't know the answer and maybe it's something you need to search and in the past you would go open up your encyclopedia bratannica and then after that you'd go to the library and then Google came along or altista and you'd search the internet to find the answer and now today you know when AI came along with these chat bots lots of people just use AI chat bots as a replacement for search and you know what that's fine as long as it gives you the sources and stuff like that of where it's coming from >> and it's actually we are still we should enjoy the moment right now >> because we are still you know let me explain what I mean we are still like in the early 90s of the internet where you can go search for knowledge and you're not bombarded by advertising >> I mean whenever I follow a sport competition now I don't go on the website I ask GPT because he gives me the raw result and the status without any of the freaking stupid advertising pop-ups that goes everywhere. I mean,

you go on ESPN website to try to find a game result one nine. So, this is not going to last long. So, enjoy it while it takes >> I do that for the I did that for the World Cup. I did that for F1. It's fantastic. >> Oh, yeah. >> Leonard, >> take a five Take five bucks or a 20 out of your wallet. Walk up to the next machine. $20 [laughter] max bet. All the winnings goes to Elevate Communities. Hey man, you were you were you were you were talking about uh elect digital payments, man. I'm paranoid about >> Nobody knows better than about those machines. Oh, they've already scrapped you. You've walked >> you've walked past enough RFID readers, they've got all your info [laughter] >> and and your and your inbox is getting filled with advertising that >> right I I do have to admit I do love being over in terminal E and Terminal 3 and G in Vegas. I'm still going to call it McCarron Airport. You're not going to ever make me start calling it the Harry Reid airport. whatever it's McCarron.

Um, but yes, I love the new terminal. And boy, even if you're not flying out of that terminal, Terminal 3, but you got to get to the airport from Vegas, I encourage you to go to terminal 3 because you all goes to the same place, except one is lightweight and you get through fast. The other one has 10 billion tourists. >> Anyway, anyhoo, anyhoo, >> maybe five billion. >> Yeah. So, we will all be there in a few more weeks, right? Uh the last day [clears throat] of November, no VME to go see AWS. That should be fun. [clears throat] >> All right, I got to go. >> Well, that's what I heard, you know. Yeah. When you got to go, you got to go. I get it. I noticed you're near the restrooms. [laughter] >> Yes. Yes. Right there it is. Right there. >> Oh, you got to go. >> Yeah. There's the [laughter] There's the sign for the men's room, you know. I hope I'm really hoping you're going to turn off the camera and and that the worst thing is is [laughter] he thinks he's going to do it. He'll turn off the camera, but the sound will still be on and it'll still gross everybody [laughter] out.

>> And anyway, the the bathroom sensors the bathroom sensor is streaming the data anyway. And I got it on my Discord. So, [laughter] >> all right, guys. See you guys later. >> All right. Adios. >> See you, buddy. [laughter] >> Have a good one. Crazy, crazy, crazy. Oh my gosh. Wow. Good times. What are y'all doing this weekend? >> It's my birthday tomorrow. >> No way. >> Happy birthday plays Alabama. >> Wow. Happy birthday. I think there's a good chance that the Tigers can beat the Tide. >> Hey man, it's been a weird year. Um, you know, I know we started out with LSU is going to just own everything >> and then and then it didn't. I can tell you that I can tell you the LSU team to it's the dark horse right now. The basketball team, Will Wade, >> they fired Will Wade and then they found out that all of college basketball was doing the same thing. And so now they've rewritten the rules.

>> Oh. >> And uh he's back at LSU and putting together a pretty uh pretty good team. >> Outstanding. Outstanding. Yeah. I don't know if Bama has it in them anymore. I think that I think they're ready to move on from Dbor. >> Yeah. I never I don't know why they hired the guy. >> I said the same thing about Brian Kelly. >> Yeah. Yeah. Yeah. True. I feel you. I feel you. Yeah. >> I said the same thing with Less Miles. Never hire a coach whose name is less. You want the guy whose name is more >> more. [laughter] >> Oh my Oh my gosh. What are you doing this weekend, Dimmitri? Anything? I am uh yeah actually I'm doing a certification for the next batch of cert community emergency response team students we had in the past seven eight weeks I do that with Sanville California >> I love it >> that's always good to get some new people trained on what to do when there is a disaster because stuff happen [snorts] >> that's good thing that you are trained up on that stuff because we need you at Elevate Communities because we're going to come save the day >> absolutely [snorts] disaster response. What are you doing, Debbie? Anything crazy this weekend?

>> Nothing really fun. Trying to get some rest. It's been The last few months been kind of crazy. So, for the first time in months, I probably take two days off on the weekend. >> I love it. I love it. That's a good idea. That's a good idea. That's awesome. I did do something interesting last night. Uh with my oldest son, I went to see the new Godzilla movie in IMAX. and >> Godzilla minus Zero. It was good. It was good. It was a good movie. >> Was it a good IMAX movie? >> Yeah. Yeah, I thought so. You know, did you did any Did you guys see that other Godzilla movie minus one? You know that the guy the Japanese guy it's the best Godzilla movie ever made? This last one he did a few years ago >> on a budget of $10 million. First time Godzilla ever won an Academy Award. It it showed how somebody lowbudget actually can make a better after all these million of Godzilla movies since the 1950s. This one blew them all out of the water. Um they even have you can you can watch it online. They even have a black and white version. Um and so well it also because it gives you different perspective. It was it [snorts] it's Godzilla right at the right after the end of World War II. A lot of people, you know, he really recreated what Tokyo actually looked like. Then it was like a shanty town and people just living in total poverty. There was nothing. Um and so it kind of you you get a feel for what life is like just as a human and then this this creature comes along and

really makes things no bueno. Um but anyway, I yeah, I'd highly recommend seeing that first one he did. Um and then yeah, this new one. It's great. You know, it's a real human story in addition to monsters doing their thing, right? So, uh, >> it shouldn't be. It's the way it should be. Coexisting. >> There you go. We should coexist with Godzilla. Godzilla can be >> coexist with Godzilla, >> you know. >> I would argue I would argue we already coexist from Godzilla. [laughter] Not giving you any names, but >> Yeah, we have enough Godzillaas. >> Yeah, we do. Yeah, we do. >> And they are not AI. [laughter] They are very human. >> That's Wow. Good point. Good point. >> Yeah. Creature from Jackal Island. There is a creature. >> Go read that book. >> I never read that book. I heard about it. >> That is the book. >> Yeah. Yeah. Maybe I should >> Did you go watch the movie Chef Chef Gordon?

>> No. I But I'm going to watch it. Is it Is it streaming everywhere? Probably. >> Yeah, you find it on YouTube. You can find it on I think it's on Prime. >> Yeah, you sent me the text. Yeah. Actually, I'll probably just watch it right here on this laptop. You should >> like like you know >> if you like food and you like music and you're a little nostalgic, it's a great movie. >> People are nostalgic. Nostalgia sells. >> Yeah. >> What's the name again? Sorry. >> Uh Chef Gordon Super Mench. >> Okay. >> But it's it's a it's a documentary. >> I love documentaries. >> We love that. >> You know, Debbie, >> that sounds cool. >> It is cool. >> It is very cool. >> I read the synopsis. It looks like he was a heck of a guy, actually. Yeah, he's still around. >> Do you know it's just kind of like how we're the same people coming out of our senior living centers to come save the world with IoT like the Cobalt people did. >> I also noticed that maybe people of a certain age seem to be into documentaries.

[laughter] >> Yes, [laughter] >> you're so right. You're so right. >> It's like, wow, how do these documentaries take off? Um, so I think I may, we'll see, use Claude for something useful. Um, so our illustrious co-host, uh, Pete Bernard, you know, he and I did the nine-part documentary on Windows Phone WTF and had all these guests and everything like that. We actually have the list and all the steps we need to do. There is a a path on Amazon Prime to get your documentary onto Amazon Prime. And so I've got all the steps and everything and we've been talking about it and everything and maybe sew it together. >> I hope you do it. >> Yeah, I think it [laughter] would be cool to be able to watch it on Prime. Absolutely. That'd be good. I I looked on it for Netflix and you still have to have like middleman intermediate people in Hollywood and obviously we don't bet but but Amazon >> just have cloud fake one.

>> Yeah. Yeah. >> He said send a bunch of agents do that thing. >> Yeah. Exactly. I'm sending my agents to talk to your agents. We'll do lunch. >> Mhm. Anywh who Anyho, well, it's good to see all you guys, you know. Much love for >> all of you. Have a great weekend, everybody. >> Remember, try it before you let AI do it. Give it a shot. Don't let your brain atrophy. Keep that brain going. All that kind of stuff. >> Humanity first. # humanity first. All that kind of stuff. Water is good, as it turns out. Electricity is good. So, all that good stuff. Uh, >> and um, you know, much love. Have a great weekend and we'll see you on the other side. >> See you folks. >> Adios. >> All right. >> All right. All right. [music]