IoT Coffee Talk
Welcome to IoT Coffee Talk where we talk tech and all things related to the Internet of Things. This program was started by a group of friends who just happen to be thought leaders, experienced practitioners and business leaders in the technology and communications services industries and more. This pod will feature weekly recordings of our casual conversations on the tech topics of the day with the hopes of delivering engaging and thought-provoking insights to our audience. We hope you enjoy program.
IoT Coffee Talk
#304 Old Wrinkled Brains
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Yes, yes, yes. That was good.
SPEAKER_03Did that work? It worked. We heard it. We heard it. Awesome. We figured it out.
SPEAKER_04We figured it out.
SPEAKER_03We also figured out that Apple's um voice isolation is very, very powerful.
SPEAKER_04Did you buy some new Apple gear?
SPEAKER_03No, no.
SPEAKER_04It's just Did you go get you didn't go get an M5 something or other?
unknownNo.
SPEAKER_04That's it.
SPEAKER_03No, I still I'm still running an M15.
SPEAKER_04You gotta be careful as far as you get farther away and we don't need it anymore, and it's like choppy.
SPEAKER_05Oh, okay. Yeah. Yeah.
SPEAKER_04Yeah, you know what? I think you need to trade in your M1 Max Ultra on an M5 something later. Pro Express Service Pack 2. Yeah, you need to get that stuff, man. Tim Cook called me and he's like, I need all you guys to go to the Apple store today and buy all our new, you know, our new chips, man.
SPEAKER_03Neo. That little tiny little DM.
SPEAKER_04So let's talk about Neo. Let's talk about MacBook Neo. What do you think? Are they going after a down market, a different demographic? Is that what that is?
SPEAKER_03Uh I think it's a services play. They and they, you know, they've lost out on the education market for years, and they're gonna get back in.
SPEAKER_04After owning it a long time ago, right?
SPEAKER_03Yeah, they owned it.
SPEAKER_04They lost it to Chromebooks. Yeah, yeah, yeah.
SPEAKER_03I think they're kind of uh uh borrowing from the Facebook or the Meta playbook, going with a cheaper um device, but um trying to tap in, expand that services market. I mean, they have huge margins on their services, right? Yeah. And then they have all the sports stuff, uh, their own Apple TV, their own content.
SPEAKER_04Um it's gonna be pretty I didn't think of any of that stuff that you said. All I thought was they want to compete with those crappy ACES laptops, you know. No, you know, or the $200 HP laptop that you can get, you know. Yeah.
SPEAKER_02Um and so I mean it's gonna be a hard job to get back into education because I mean my kids, they kids get handed a Chromebook, you know, at the door along with their you know, ID card and their first homework. And they're so pervasive in schools. I think uh both Apple and Microsoft were sleeping and Google just flooded the zone. I mean they're right, the hardware's got three or four years life on it as well, so they're not gonna they're not gonna have any chances for us.
SPEAKER_03Yeah, and I was uh surprised in education the touch screen actually matters. So that's where I think Apple may have to consider uh introducing touch screen for a Mac, which is maybe sort of yeah, but I outside of that, I don't see anyone actually using a touch screen on a two and one. You know?
SPEAKER_04Uh they go, I've got a touch screen, I've never touched it, but I've got one. Yeah. Yeah.
SPEAKER_01I mean that sounds like a strange strategy because their position has always been that you need more power locally to run intelligence and all this kind of stuff. So if they go into this direction of those you know lighter cloud-dependent services systems, and especially in education, you're then creating a generation of people that you cannot pass the message that the power should be run locally. So that sounds very um I mean, I haven't had time to watch the announcement this week, but that sounds very weird to me.
SPEAKER_04Dimitri's trying to take us back to the 90s when we were gonna buy a Sunray system, you know, network computer that taught that's all it has is a Java runtime. I think that's what we should do, otherwise we're gonna be or or when we got into this century, remember the netbooks? Yeah, yeah, like the really cheap low-end laptop running XP. I remember that.
SPEAKER_01I remember Larry Ellison and Oracle World destroying a PC and saying, Here's what happened if your PC is destroyed, but you just need a browser and buy Oracle because I have the data, blah, blah, blah.
SPEAKER_04I remember Larry Ellison going, It's just bits. Why? And you're buying plastic with bits on it and putting it into a CD-ROM. You could just download this from this internet thing that we just created. Yeah, yeah. You know, um, that's yeah, I love it when people zero client computing. Zero client computing about that. Absolutely. And remember Sun even built chips that were optimized for Java. Yeah, that was their whole thing.
SPEAKER_01In a ring, I remember, I remember Java ring. Yeah, the Java ring.
SPEAKER_04I still have it, I still got it. Burn it, burn it. Yeah, and I have the connector, it plugs into a serial port, RS232 serial port.
SPEAKER_01Yeah, I think Bill is about to leave.
SPEAKER_07I I mean, I'm listening to this. I said, I'm like, what in all the hell are we talking about here? Because I'm and look, I don't care what any education system says, I don't care what anyone else says. We're straight old school over here, all right? It's Linux, it's command line, and that's what my girls learned.
SPEAKER_05Oh, there you go.
SPEAKER_07Wow, there you go. That's it. When when they went to school and then and the the each of the classes, they were like, Oh, you have to use a Chromebook. I was like, you take your Mac and you put that Chromebook next to it, and you do all your work on the Mac.
SPEAKER_04Wow. You're one of those parents, aren't you? So you know what? Yeah, you were at the school board meeting, weren't you?
SPEAKER_07Hell yes.
SPEAKER_03Yeah. And you know what? Claude uh OpenClaw is learning uh to use your CLI as well.
SPEAKER_07That's what's up. I mean, the rest of this, the rest of this, I mean, we've been old school, this is what it takes to build. There's a crucial thing.
SPEAKER_03No MCP, none of that nonsense, none of this abstracted crap. Uh straight up, straight to straight to your privilege access systems. Dude, look, let me tell you that.
SPEAKER_07I don't know, I don't know how many of you have ever done anything. All right, so so you go to you go to a trade show, everyone's like, oh, do we have internet connectivity? Yeah, you have internet, but you get enough people there and you're getting stepped on, right? So the noise floor is off, I mean, off the hinges. And so you go to building local, right? You know, your your local 3000, that's where you're you're gonna you know run your demo because you don't have to worry about internet. So I watch people saying, oh yeah, you know, we're using these Docker containers. Docker doesn't update overnight, your your Docker won't open. Ouch. Ouch. I'm like, and then they're like, Well, I'm trying to use the UI. I'm like, go in and command line and you can get everything back running again. Can't do it. Why? I need the dashboard, I need this, I need that. It's like we have forgotten what the hell it is.
SPEAKER_04Wow, Bill is the teacher of the future. I'm gonna see him in the classroom, and all the kids are learning everything on the command line and stuff like that, and they're like throwing shade at people who are using use UIs to do stuff. It's like you guys are a bunch of you guys, a bunch of losers.
SPEAKER_02There's there is a re-inbrace, there is a reinbase of of analog kind of lo-fi. And seriously, I mean, I went like I said, I went to this gig last night with my son, young cut band from LA audience somewhere between 16 and 22. And uh usually you used to see everyone holding up their smartphones and all the screw screen. There were plenty of those, but I'd say roughly 50% were old flip phones. I kid you not, they were flip phones. There must have been 60-70 of this wasn't a very very very big menu, and I stood right at the back. There must have been 60-70 people with with actual flip phones. Oh they look like motorrider razor flip phones, and to your point, Bill, I mean, I think the the there is a strong kind of like rejection of AI for a variety of reasons amongst that that that kind of age range, not least because it's going to screw with their lives. Um but the tools are still very you know compelling. I mean, at the moment, my 11-year-old is building multiplayer platform games using clawed code. Um, and then I'm trying I'm trying to do with him. So he's hooked, he's hooked on the ability to create stuff, is then getting to go back, reverse back through and recreate it. So he's using VS Code already as a as an adjunct to clawed code. He's taking the code, dropping it into an HTML, you know, and he's figured all this out himself over the course of the last two or three months. He's a bright kid, um, but not not hasn't had any training on coding or have you. But but my hope is that we if you can hook him on something that's fairly quick to deliver results, then you need then you need to go back and kind of build the foundation of the building. So I agree with you. We're we're losing we're losing the ability. I had a long conversation with him the other day. Why do I need to learn math, Dad? You know, I can just get the computer to do it. Um and you know, my answer boils down to if you don't understand how the machine works, then how do you fix the machine? And and what do you do when the machine you know takes over parts of of uh functions that you don't want it to? Um, and and you know, math obviously being the root of of all things. And he was like, well, okay, okay, I bought it.
SPEAKER_01I don't know whether it means it's math grade will improve, but you know it's even more fundamental than that because math is really it teaches you the thinking process. So it's it's the it's the language of science. So if you can't solve math problem, it gives you the approach to solve any other problem. That's what people don't realize. Uh I mean calculate top of your head doesn't make any sense. 125 multiplied by 693. You don't need you have tools for that. But knowing how to do that uh shapes your brain so that you have a thinking process.
SPEAKER_07There's something called the order of operations when you're solving a problem.
SPEAKER_01No, well, you mean typing you mean typing typing into ChatGPT or into Cloud?
SPEAKER_07No, no, order of operations.
SPEAKER_01There is a workflow associated with the Yeah, you start with Yeah, launch Cloud, ask the question.
SPEAKER_07Wrong answer. Wrong answer.
SPEAKER_03All right, wait a minute, guys. So let's say hi to the audience. Hey everybody, welcome to IoT Coffee Talk. And um, you know, uh remember, we do this for the communities. We're trying to elevate communities. And there was some interesting talk last week about maybe doing an event, and I think that's a really great idea. And we had, I think we have some of that event um management and development firepower in the IoT coffee talk crew here that uh we can make that happen. So keep an eye out for that because I think that might be a real thing. And um, yeah, and remember, we're here for entertainment purposes only. Right? We're completely non-informational.
SPEAKER_02Uh by saying in any business any investment decisions or sort of things, yeah, yeah, yeah.
SPEAKER_03So um yeah, take us seriously at your own risk. And uh, you know, if you take us seriously, that's pretty much your fault. And we encourage you to take out insurance if you do. So anyways. Welcome.
SPEAKER_04Yeah, it's the only part where you have to take out insurance to listen to it. Yeah, exactly. Wow. Yeah.
SPEAKER_05Love that. Love that.
SPEAKER_04I don't like YouTube.
SPEAKER_05Yeah.
SPEAKER_04I love I'm loving hearing that this small young age group has got foot phones and stuff. Yeah. Are we gonna bring back the StarTac?
SPEAKER_02I've got one. Maybe I should offer that to my living one. I mean, it's funny though, this is the first generation that is being stripped of basic abilities to understand, dissect a problem, source information, synthesize that information and reach a conclusion because they are going straight to Claude and typing in the question and getting the answer. And it'll be very interesting to see, you know, scary to see what the implications of that long-term are. And and incumbent upon us to figure out how to fit how you fill that gap. Yeah, how do how do we get how do we end up with people who understand how to code?
SPEAKER_03Yeah, you can't. Um, we've already been uh going through this um with Google. Um, so people copying and pasting. There's a copy-paste problem that we had uh we've had over the last two decades. And so, you know, I I've mentioned this before. You know, back in the day when we had uh laptops and we had to store all of our files for our projects on on our laptops as we're you know bouncing across the uh the country or around the world, um, and there wasn't this thing called the internet and cloud. Um, we had to prepare, plan, think. We had to pay attention in meetings. Uh, we didn't have all these tools that took notes for us. And um the skill sets and what we were trained on, right? Um the the sixth sense that we developed as sentient beings, uh we were forced to do a hell of a lot of thinking. Um over time, in the consulting uh industry, I saw that decline where people were just copying and pasting stuff, there's no original thought, inability to create original thought, because everyone was just copying and pasting stuff. And I I can't tell you toward the end of my consulting um career, I had to throw away so much stuff because it it was you know that people were producing because it was flat out copied.
SPEAKER_04But don't worry, don't worry, the stuff that's the stuff you've created is gonna live on in PowerPoint decks delivered by the world's largest consulting firms when they present it to their customers as their own stuff, like we've probably all seen before. Oh, yeah. I'm not gonna I'm not gonna name any names of the biggest consulting firms on the planet that I see that all the time. Oh, you guys don't have an original thought, do you?
SPEAKER_07Well, I mean they've lost it, right? I mean, at the end of the day, you know, they they make thousands of phone calls to folks like us. Your name doesn't get mentioned, but your your data is there. Everything that you've told them has been either formatted and synthesized to some level, but you know who said what. And you know, the the the way that I mean for me, the the the fun part is you know, someone wants to ask me a slick question, let's have a conversation. Let's talk about that. When you start saying that this this is wrong and this is right, let's talk about that. Yeah, and let's peel the onion back because I guarantee you, somewhere along the way, the fumes of that onion are gonna burn the fuck out of your eyes.
SPEAKER_04He was hoping you were gonna talk about pop music, but instead it was onions burning your eyes.
SPEAKER_02Yeah, yeah, that's literally what it is.
SPEAKER_04Wow, wow.
SPEAKER_02I mean, I think consultants, I mean, uh back back in the 90s, mid-90s, I spent a stint as a consultant. And I, you know, I was actually talking to a friend of mine who stayed stayed the course and his like partner and in a poem I won't name. And he said to me, no, what we're what we're now doing and what we've been paid to do is helping enterprises figure out how to adopt AI. That's it. You know, um, nothing's people don't ask us questions, there's no market study anymore, there's no strengths and weaknesses of your competitors, there's no all of that stuff out the window because people just use ChatGPT or Claude or you know, whichever tool they prefer. Um, and and so our job is really serving up the serving up the means of our own doom, is how he put it.
SPEAKER_03Yeah, I know, but but the outcome of that is fine because it actually creates uh val, you know, people can create uh opportunities for themselves because as the noise floor rises, that it is uh it creates higher value opportunities for people who can filter through the noise. So I mean you know what I'm saying? It it's it's not all doom and gloom. Uh and at some point, um the real money is going to follow uh grounded stuff. Um so uh yeah. The thing is is more concerning is um the state of society as cognition continues to decline. I mean, we have to exercise our brains constantly, especially as we get older. Otherwise, man, you know what I'm saying?
SPEAKER_04You're gonna be with a keytar. That's how it's gonna work out for you, man. You're gonna be with Howard Jones, top of the pops. Top of the pops maybe in the 80s. You know how sometimes you work really hard on a post and you think it's the most important thing you've ever done, and like five people like it. Yeah, and then I post Howard Jones, you know, from the 80s, and there's like 4,000 likes and 280 comments. This is the best thing I've ever seen on the internet.
unknownOh my god.
SPEAKER_04So like there's a message hiding in here, I think, about this.
SPEAKER_01Rob, you you you have to tell us about the other 65 Rob Tiffany, Rob.
SPEAKER_03No, no, no. And I'll tell you right now, I'll tell you right now, all this stuff is fake. All this stuff is fake. Um is that air you're breathing? Exactly. It is. And so um you you just can't take you just can't take things seriously. And but it's not all because of AI, it's because of all this synthetic content, contrived more, you know, uh this perception that impressions are everything and subscribers and followers are everything. You know, most of them are fake. There are all these click farms out there that and engagement farms that are producing synthetic engagement. And this is something that Debbie and I talked about a long time ago, as usual, right, Rob? And uh and now you have Meta buying open. Did they didn't they acquire Open Cloud? No, they bought Multbook. Yeah, Multbook.
SPEAKER_04The social network for agents to talk to each other. Yeah.
SPEAKER_03But why?
SPEAKER_04Yeah, but think about the value now in the it back in the day, you were having to pay a whole lot of people in the Philippines for your click farm, and now you can just have legions of agents doing it for you.
SPEAKER_07Well, I mean, you have those agents even responding, right? I mean, you look you do anything. If you're on Reddit, if you're on anywhere else, you got to look at the responses to see is that a bot? Yeah.
unknownYeah.
SPEAKER_01I mean there's actually that maybe there's hope in this dimension because we we might reach a point where all those things actually spend their time talking together and we can get out of it. So the the matrix became a matrix for the that's a great idea. Actually, the matrix is reversed. It's not the humans that are in the matrix, it's the machine. We put a matrix on the machine to be inside and they just burn venture capital. Yeah, that's the whole purpose is to burn venture capital.
SPEAKER_03It's not virtual capital, though. That's the problem.
SPEAKER_01That's the way you that's the way you change the world. You create that infrastructure where all those investments go and work into one area and then just talk to each other.
SPEAKER_03No, but the problem is it's not virtual stu um capital. It's real money. Advertisers dumping a ton of money into these platforms like Meta and Google search.
SPEAKER_01Yeah, but this this monetary most of it is fake.
SPEAKER_02I I think I think kids are, you know, again, going on my two, the kids are very, very attuned to AI generated content. Um Really? You think so? Very, very attuned. Oh, yeah, yeah, yeah. I mean, they sit and watch um the YouTube videos with my wife quite frequently, and and they are expert at spotting what is they and they it's a fairly instant rejection of. So I do think you you might be right to me. I mean, I there is a hope in me that there is more focus now on creative professions. Um both my kids are like we weren't what what's the point of being you know going into um computer, you know, the IT world, etc. Um, we'll we're gonna go and do art and design. There's more focus on people actually talking to each other um and community, um, which has a long way to catch up post-COVID because we all got used to not talking to each other face to face. Um, so there is hope, I think, in this generation, and that their rejection of of what's happening right now is is firm. I mean, it is firm. Um, and they can discern the difference. So they might be losing the cognitive ability to define and analyze a problem, etc., etc., but they're they're they're developing a more, let's say, critical filter. Um to be able to filter out the the drops from from all the stuff that's being thrown at them. But the volume worries me because there is so much, so much, so much, so much. LinkedIn, we talked about before. LinkedIn is is worth is kind of useless at this point because so much of the content is, you know, here's my AI generated, you know, designed to get uh grab attention from people, content and it isn't it's very difficult to actually find useful information amongst all of the noise. Um the noise.
SPEAKER_07That that part right there, you uh you are spot on with that one. I've I can't tell you how many times a week I get people sending me stuff saying, Did you see this guy? You see what he said about digital twins? I'm looking at it, I'm like, this is AI generated. I could care, I could give less than I could care less about something. You're gonna come up with something creative.
SPEAKER_04How many of those Fs are you not gonna give?
SPEAKER_07See, I I I'm I'm just saying I don't want to get, you know, we're too early into the end of the talk for me to just get crazy like that.
SPEAKER_04Oh, we'd love to be crazy. Hey, that's what builds our audience is when people go crazy and stir the product.
SPEAKER_07I mean, but for real, it's like all of a sudden you have like uh you know 2,000 people that are experts in this space. And I'm like, you know what? If you say that you're if you qualify yourself as an expert, if you are calling yourself an expert, if that's the the if that's the the name that you're giving yourself, then I mean, damn it, you better be like on you better be the Tony Robbins of whatever it is that you're talking about. Because if you're not, you know, kick rocks and don't let the doorknob hit you with a big boy. Because I I don't I don't have I don't have anything for all of these things. I mean, you know, if you have time to post three, four, five times a week deep content about a specific subject, one of two things is happening. Either you don't have anything else to do, or that that's one, right? You don't have anything else to do, or you are clamoring for attention.
SPEAKER_05Yeah. Yeah.
SPEAKER_01But you know, no, no, I mean we are talking about you know kind of society level issues, and and that's kind of the I think that's kind of the biggest point here, because we see how some of our leaders operate by saturating the media space with content and things like that. So this is just I mean, it's a mechanism. So unless there is a generation, as Arista was saying, that's starting to realize that enough is enough and we have to build another model with other values, another education, because the only the only way to solve that is education. It's just the only way to solve that.
SPEAKER_03Yeah, well, you know what really bothers me more than the um the impact that uh AI has on the younger generations what it's uh the impact it's having on older generations. Oh yeah, totally past the fakery, you know. They're I mean, like my parents are I mean, they are problematic. They will share all this fake stuff and they will think that it's true. And they are so gullible. It's unbelievable. And it's just starting. And what really scares me is are these um you know, engineered um and generative AI assisted and uh you know engineered um spam emails and phishing attacks. I get them and they're convincing. My parents get them and uh other folks in their generation Oh my god. Yeah.
SPEAKER_02But I think I think what's happening now is people are spending more time actually using I mean how many people are using Chat GPT and typing in, take this AI generated content and make it sound as though I wrote it? Because I think now that's the thing. Everybody can generate the same report in ChatGPT. Had a discussion with a kind team meeting the other day. Um, everyone can generate the same content using ChatGPT or whichever tool you like. But who actually understands what they're talking about? And as you said, Bill, who can actually figure out what the answer to the problem is? Yeah, increasingly you see people, I do, I do it, I don't make slides anymore, I draw, I write, you know, in analog, you know, here's here's the diagram of how you know it could work. Uh, but but as people start to use AI to make AI sound more human, the the problem that you point to, Leonard, will just get more, more, more pronounced. But I would guarantee, I would bet 90% of executives um are typing some some derivation of that request into an AI. Make this content sound like I wrote it, make this look as I mess it up a little bit, add a few spelling mistakes, don't use the same word. And it's um thing nonsense on nonsense.
SPEAKER_06Yeah, yeah, yeah.
SPEAKER_01Yeah, but the thing is, and you're on the other way, is why you're back to my matrix idea is that the people that receive that do the same. So there is a point where this is gonna circle back. So, and then the hope is that a new generation emerged saying this is bullshit, I want to build another society, or the whole thing collapsed. So, this is really the question to ask. This is really the question fundamentally. Because yeah, this all content is produced by AI, but I receive it, I also use AI to process it. So it's just a lot of fun.
SPEAKER_03Yeah, and here's the great thing is um uh AI is completely optional. There is no essentiality in any of this stuff, it's just convenience. It's like e-commerce, you don't really need it.
SPEAKER_04Uh, it hasn't necessarily transformed retail as we know it because are you saying that I can drive my car to an Amazon fulfillment center and go in and go shopping there?
SPEAKER_03Yes, and then you can have your own damn delivery service, dude. Or you you know, it it's there's nothing original about it. That concept is service merchandise, dude.
SPEAKER_01No, you can drive.
SPEAKER_03Do you remember service merchandise?
SPEAKER_01I do remember.
SPEAKER_03Best toy catalog ever, especially their Star Wars collection.
SPEAKER_01It's actually the alternative, Rob, is not to drive to the Amazon Center, the alternative is to drive to the producer.
SPEAKER_04That's that what? I thought it was all produced inside that building. What? Like kind of like Santa's Elves create video game consoles out of the.
SPEAKER_01Oh, you you mean the North Pole, you mean the North Pole Resanta, right? Yeah.
SPEAKER_04So you think that's I've been putting some dots together. I was thinking about the IoT conference that we're gonna have, maybe in Bandera, Texas. But then thinking of a conference, I was also thinking of, you know, I got Bill and Dimitri here, the leading digital twin people on the planet. Bill talking about people, Bill talking about full of shit people writing about digital twins. I was like, we need to have a digital twin conference. But he also mentioned Tony Robbins. So I think at our digital twin conference, people have to walk on hot coals for the final exam to prove that they actually know what they're talking about.
SPEAKER_07They don't have to just walk on hot coals, they have to walk on hot coals while writing code to build digital twins.
SPEAKER_05Yes, yes, I love that. By the end of the track, they love that.
SPEAKER_01On a terminal connected to a mini computer or a mainframe. Not one of those codebook nonsense things, okay?
SPEAKER_04We're gonna build digital twins with kick screens. Yeah. Anybody remember those? IBM? Uh never mind. Yeah, whatever.
SPEAKER_05That's old stuff.
SPEAKER_04I love, have you even seen that ad on TV? You know, I forgot who it is. They're on a factory and the factory's not working, and the girl goes, We need someone who knows COBOL. And the guy's like, what's COBOL? He's got his hard hat on, they're like, you know, working and the robot's smashing into the windshield of the cars they're making.
SPEAKER_03With um all the technical debt out there, the reason why you you still have systems that are based on COBOL.
SPEAKER_01But do you know why? Do you know why? Why there are so many so many systems sitting on COBOL? Because it was a great powerful language to process transactions.
SPEAKER_03Well, yeah, but there's that, but the reason why you can't modernize is that all the people that knew how that stuff works are dead now or they're retired and they cost $100,000 an hour to hire uh to unravel all that mess. Um and there is no there's no documentation, right? And so it makes it really difficult for all that to be unraveled. But it it's a testament to what we're looking forward to, where everybody is vibe coding and working with these abstractions into you know, Bill and Alistair's point, you don't know what how things work. Eventually you can never fix what gets broken, right? And that that's such a that's that's such a uh that thinking is such a disservice because going forward, what I'm afraid of is this is that we will find out that we actually do need computer science folks. We need developers, we need folks who can test for all the messes that everyone's gonna be making with AI, right? Because no one's really thinking through how does this stuff really work and is it safe? And that thinking is still a huge deficit in um, you know, I m MWC. I have so many conversations. I'm I'm listening to these guys going, oh my god, you just don't give a shit about safety. That's first. If you're a vendor, right? And these guys are jokers, are still popping out crap.
SPEAKER_02We talk a lot about the role AI is going to play, and the thing we don't talk about so much, and this actually goes back to something Bill, you said a few weeks ago. When you look at the use of AI within enterprises, essentially you're playing Russian roulette. You know, if you you then there is an absence, and there's a few startups I've seen, including this startup entire AI, which is uh Thomas Donkey's startup, that are aiming to provide an enterprise grade toolkit for leveraging AI agents in a safe way, because otherwise you've got intellectual property contamination. You you if you've developed something and an AI was in the mix, there have been several cases recently where the notion that there is attorney client privilege if an AI is involved, or there is is is patentable IP if it is being blown away, there are security vulnerabilities on an industrial scale. I mean, there was a thing I saw recently that 50% of the time an AI will choose an unsecure deployment model or the development model in preference to a secure one. And then there's the never mind the liability and the regulatory and governance gaps, and when when you're when your truck runs someone over, you know, being able to prove the provenance of the code, being able to show that you put it through the appropriate set. There, I think the gap will be in folks who can understand the system in depth enough to be able to go in and forensically analyze you know what happened um after the fact. And I we don't I don't see a generation of people being coming out you know having the opportunity to learn those skills. Yeah, because you learn those skills as a junior developer.
SPEAKER_07That's why we're gonna that gap that I'm talking about all the damn time that it's not pretty. But the the real the the reality the reality guys, man, the work that people really have to do to make this stuff go. I mean, when you're putting these things together, there are standards, there are best practices. The reason why NIST and IEEE and and ISO and all of them exist.
SPEAKER_01Yeah. Yeah, but the reality the reality, guys, is the drivers for that. So because you know we I think we're all quite well-informed people in this talk, and we can all realize that. But look at what happened for cybersecurity. We never did it from the ground up, and now we're in a mess. Now, with all this AI stuff, it's gonna be uh a hundred X. So the only hope is a crisis big enough, but not too bad enough, so people will start realizing. No, because the next the the next big I mean the the the next big thing, and you rob, you you might have seen some of my comments on some of your recent posts in this area. To me, the next big, big, big catastrophic event is when we're gonna start connecting those AI things to the physical world. Right now we're still safe because there's still a human that reads the crap and pushes the button. Uh the day it is over and that's going to be accelerated. You can actually argue that the uh the the war climate that exists in the world is going to be a significant incentive for that uh the day it starts happening without any of the guard uh guard rate that we discussed, uh there is going to be a crisis and it's going to be to a scale that we have no idea. But I don't see any sign that there is any uh one or anybody or any leadership. I mean, come on, look at the leadership we have in the world, not just in the US, everywhere. I mean, there's no no indication that anybody has an idea.
SPEAKER_03I think the EU has an incentive to But they have no power.
SPEAKER_01What the EU is it's they have no power.
SPEAKER_03We'll we'll see what happens this year. They do we'll see what happens this year.
SPEAKER_01How long has it been a war? Leonard, how long has it been a war in the middle of Europe? How long? Four freaking years. There's been a war in Europe four freaking years. I know, I know. No, no, no.
SPEAKER_03It's not related to the war. It has it has a lot to do with the DMA, the AI Act, and some of the stuff that I saw at MWC related to uh sovereignty. But um no no you here's the thing people still have no clue how dangerous this stuff is gonna be. I sat in a guardrail um orientation thing for um vendors. Okay? Um if you encourage your vendors to use AI with your data, be ready in in about a year uh to just be banning the use of AI amongst your vendors and service providers. It's not your data, you just go to AI. I I was just sitting there listening to this thing, going, Oh my god.
SPEAKER_04Yeah, I mean, it's that same conversation we had when we started talking about self-driving cars. Who's responsible? Who's liable? This is worse.
SPEAKER_02Well, the challenge the challenge though is it's similar to if you go back, well, uh if you go back a few years, there was a great deal of pressure on manufacturing companies to use the fastest path to the outcome. Never mind if that produced pollution, never mind if that was child labor, never mind, you know, whatever. And the countries that were prepared and the companies that were prepared to take that risk on, one, you know, they were able to build cheaper products. Look at the dominance of China and Southeast Asia in general in manufacturing. I think the equivalent is is happening is going to happen with AI. The countries where AI is treated with some skepticism, and there are guardrails put around it, are not going to move as quickly as the countries where there are no guardrails. And not and not pointing specifically at China. Korea, for example, I read something the other day. So for every X number of workers in the US, there are 200 robots, I forget what X was. The equivalent in Korea is like 15,000. There is an adoption of technique, the means to produce more for less, you know, regardless of the risk. It comes down to your appetite for risk. If you're in a litigious Western society, then maybe you're a bit more cautious. If you're in the Wild West, who gives a crap? Or you know, there, you know, all no holds barred. You can do what you want, and that's going to create some really quite concerning um risk situations. And you see this kind of regionalization of data. A lot of the enterprise customers I talk to are like the data's got to be stay in my country, in a factory. If you can if you can have federated learning on the edge and the data doesn't go anywhere, like great, excellent, that solved my problem. So there's this kind of retrenchment that's happening, but at the same time, global systems that are being originated in countries with few guardrails are going to be pervasive. And we're all going to be subject to those risks. And I hope you're wrong, Dimitri. I hope there isn't a huge shock. But you know, history would suggest that human beings require a fairly, fairly high shock to change course.
SPEAKER_01No, I agree. I mean, I I I I can't agree more with you. It's totally totally what's gonna happen. We have the kind of the Nobel Prize effect, you know. Yeah, we're just gonna do it, it's gonna blow up, and then somebody says, Oh, let's create a Nobel Prize to try to buy ourselves back.
SPEAKER_04You know how I get to see Democrats.
SPEAKER_01And I do believe your Europe will be left behind. Europe will be left behind.
SPEAKER_04I get to see Dimitri's comments and posts along with Rick Bellata's. I go into LinkedIn and type in search for negative comments. And then Dimitri and Rick Bellotta come up.
SPEAKER_01This is not negative comments, this is the reality, my friend. I'm the pessimistic, but hey, somebody has to be. If you have no idea, you're the most pessimistic person in the world. Well, I mean, I don't I don't know.
SPEAKER_03No, maybe Rick is yeah, it's not pessimistic. I think it's it's probably being c it's closer to realistic.
SPEAKER_04Um you know, you probably saw this thing floating around years ago, and I saw it popping up again. This was a quote that IBM made back in the 70s. It basically, you know. And so, you know, IBM was always clear. A computer can never be held accountable, and therefore computer must never make a management decision. Well, that was kind of set in stone at IBM back in the day.
SPEAKER_01Yeah, read Yazimov laws.
SPEAKER_04Yeah, yeah. And it's like, oh, has that changed? Because obviously it has, you know.
SPEAKER_01Oh, yeah, we forgot, we forgot to be able to do that.
SPEAKER_02Yeah, I think that the challenge at the moment is a lot of the a lot of the hyperbole you hear about AI, well, some of it's not hyperbole, some of it is. But uh, you know, the the temptation is to you know, call call attention to you know what could be a very significant risk, and do so using language that will attract people's attention. The downside of that is that people read something that sounds so extreme that they just immediately dismiss it and move on to the next thing. Or they think it won't affect me, or they think it's not gonna happen. I mean, look at climate change. It's not gonna affect me, it's not gonna happen for a bunch of years, you know, not sure about whether or not it's a risk anyway. I'm just gonna carry on driving a five-litre you know truck around the place. We're in the same situation, and and human beings are not good at reacting you know effectively to this sort of circumstance when it's a massive risk, but it's not directly impacting you, you can't really do anything about it, and it could be long-term in nature, and no one you know, the future's not certain, maybe it's gonna work out fine. Those sorts of situations we don't do well at. We do well at where you know when a dinosaur is chasing us, should we run away? Yeah, we do well with that.
SPEAKER_01This is why I was mentioning leadership earlier. The only way we've solved that at some point in humanities is when you have leaders that point us into the right directions. And the problem no, it's the only to it's the only way to solve the problem. The only way to solve the problem is it doesn't care for me now, but it will care in 10 years, is to have leadership that tells you it cares.
SPEAKER_07Who the hell trusts any kind of leadership up around this place?
SPEAKER_01No, but we used no, but we but but the point is that we the the the this is why I'm more and more of a pessimistic because we used to, guys, we went to the moon under leadership. That was 50 years ago. Yes, we did. My digital twin told me, my digital kick, my digital team told me that's how we got there, digital twin.
SPEAKER_07We Capricorn one, baby. I'm just saying, I mean, there's there's some legitimate damn questions about that. I mean, the amount of infrastructure it takes to communicate from the moon back down here, where the F was that? Like a little Motorobo.
SPEAKER_01There's an excellent you there's an excellent YouTube channel that's run by a French guy called Corius Mark that explained that in greater details. And they've actually even rebuilt in the lab the whole infrastructure to do that. So there is evidence of this.
SPEAKER_07Thank God for other at that time actually doing it. I'm sorry, the jury is out.
SPEAKER_04Thank God for other countries that are circling the moon now and taking pictures of our the equipment we left behind. Find a so-called leader.
SPEAKER_02Well, finding a leader today who fits the definition of the word leader is tricky, I think, anyway. But it find a leader who understands the first principles of AI, and I think you're doing very well. And like you were saying, Leonard, you go to you go to industry conferences, and people in the enterprise kind of you know, well, industrial automations based, etc. You know, there are there are lots of folks who haven't really thought through or understood the implications of AI. But otherwise they'll be building different products. You know, folks are building applications on top of an LLM. It's like, guys, you you are two steps ahead of the steamroller behind you, uh, and you're gonna get squished. You better focus on feedstock or governance.
SPEAKER_04That's why I'm I'm putting together my pitch deck that I'm gonna pitch to Alistair for my new company. It's gonna be called open claw manufacturing. Open clouds. Oh, we're just gonna drop that thing in the factory and say, make it better. Oh my god.
SPEAKER_01So and and then Alistair, the commercial is going to be a Pac-Man, you know, eating different things.
SPEAKER_04Exactly. Yes. Oh my gosh.
SPEAKER_03I mean, you know, uh it it's really funny that at MWC um this year there's been a come to Jesus moment finally around Copilot. And that it it sucks, it didn't work, it didn't do anything that people were promising um back three years ago. And you do you really do get a pretty solid sense that all this AI stuff uh has been kind of a joke. It's just a rolling circus act, right? It's like one circus act after the other. And it it it's it's funny to watch how everyone is just starting to get sick of this in the background. They're going, can we stop talking about this crap? Because number one, we can't sell any sell any of it. Our custom we're in these like infinite POCs, and um it's not delivering uh to expectations. And so there are vendors out there that are desperately trying to build a uh you know, basically a last minute case for the this stuff, which I think it will be because physical AI is such nonsense, right? It's just friggin' robotic intelligent robots.
SPEAKER_04Jesus is gonna be on stage next week at GTC and he is gonna be telling you that the future is physical AI. And so I don't know what you're talking about here.
SPEAKER_07Oh, I know, I know uh you'll know Jesus. I know Jesus as well. Yeah, he goes by Jensen. Where's he gonna be at?
SPEAKER_04He's gonna be in San Jose at the SAP center. Yeah.
SPEAKER_01And you too big guys. All right, we'll see.
SPEAKER_04Yeah. I mean, you know, Jensen's just pounding the jumbee. He thinks the that's the future is just what it's.
SPEAKER_03But how much of it's actually new?
SPEAKER_02Well, I think I mean if you can solve if you can solve the whole governance issue, we know how to how to generate forensic information about what an AI has done, the decisions it made, etc., so that you can protect your IP, you can't, and if there's a management structure which means you're stopping your employees feeding all of your company data into an AI and therefore making it completely public. And and you know, we figure out how to because at the moment, I mean a lot of enterprises. I've been talking to a few people who've gone from building applications on top of LLM to building um applications that that kind of synthesize original data sets and present them in a managed way to an AI, so through an API call or through the data itself, stay separate. Because I think the enterprises who have adopted AI have pretty quickly realized that, well, my AI is the same as your AI. So what how how does that give me a competitive advantage? Oh, if I use my data, if I have a proprietary data set that I can train my AI on, then then but then it and then the battle becomes as it always was to a degree, it becomes about the data. My data's better than your data.
SPEAKER_06Yeah, right.
SPEAKER_02I can use it more quickly because I've got an AI, and I can I can use that AI in a physical situation to optimize a system faster, but it's down to the data. You know, if I don't have data on the system that the AI is trying to optimize, if I don't understand the manufacturing workflow, then who gives a crap about my AI?
SPEAKER_03Well, see, that's it, that's the ironic thing, is you're we are starting to see, and you will see in the future, uh, more uh compartmentalization of these models and the data sets and the knowledge bases. Because it's too dangerous to train any single monolithic um AI with all of your corporate data. So we we're we're off that thesis already. That's my idea. Then you're you're on the wrong track, buddy. No, we're training it, all of it. No, that because uh uh because of security reasons. The thing is is you're offline air gap LLM running on your infrastructure and you're gonna fine-tune it with your but even then it's gonna be uh compartmentalized because you're going to have cybersecurity threats. Anyone who can get that graph that can get that model will have everything. And it that is way too dangerous.
SPEAKER_04It's going to be What if I put a strong password on my vector data?
SPEAKER_02Well, yeah, I mean, I I was talking to a company the other day, and what they build is a um AI-based solution for optimizing power consumption in data centers, quite a pressing you know topic. Um and they use a that uh A, it's uh completely air gap because it's a freaking data center, but they're also using a federated learning model where it's kind of like um it's kind of like terrorist cells. You know, each machine you know learns from the collective, but the information that's shared between them is really quite limited. So so that is a is a compartmentalized AI actually uh in each edge node, you know, effectively.
SPEAKER_04Is that company was that company called Sleeper Cell? Was that the name of that company?
SPEAKER_02Hold on, hold on, I just need to go and buy a URL.
SPEAKER_03That's a great name for a company that does federated learning at the edge.
SPEAKER_02No one has to be put off by the get the terrorists in your process.
SPEAKER_07Yeah, that's what you but I mean, but but when you say federated learning at the edge, isn't that kind of like what neural networks are supposed to do anyway?
SPEAKER_03Oh neural networks.
SPEAKER_07Yeah, and it is a neural network. This is a Unix system. I know this. I mean, yeah.
SPEAKER_02Yeah, I mean quite honest.
SPEAKER_07That's what that's what they're supposed to do, right? Yeah, it's neural network.
SPEAKER_02I mean, uh there's there's there's lots of uh everything is everything is kind of being slapped with the AI sticker, you know. It's you know, IoT, you know, well tell that's telematics, M2M, you know, let's chuck all that stuff in the bucket and brand it IoT. Machine learning, data analysis, neural networks, chuck it into the AI bucket. The thing that pisses me off, honestly, is that I mean we all talk about IoT and the brand IoT branding, I think, was deeply unhelpful. But why did we do such a shit job in in publicizing the term IoT? Why is AI so much bloody better? How did that how did AI get to be a global phenomenon when people still understand what IoT is?
SPEAKER_03Yeah, because they start layering all kinds of other shit on top of IoT because they found out, oh, wait a minute, these huge numbers that all these consulting firms and research firms are throwing out there aren't going to be met if we don't start double counting and layering a bunch of other shit on top, like big data. You know? And then you lose perspective. And then everyone discovers we're double counting, just like physical AI. What's gonna happen is Jesus Christ is gonna get up on stage and say, the you know, the impact, the economic impact, according to this consulting firm who is gonna get it catastrophically wrong because it's X trillions of dollars, right? And it's gonna be triple, quadruple counting of stuff that already exists, right? Uh and you know, robots don't have to be intelligent, they can be remote can remotely controlled, completely non-intelligent. That's robotics, and most of robotics is that's what robotics is. It's non-intelligent robots. Doesn't mean make them useless. In fact, they're probably a hell of a lot better than any of these humanoid, you know, this humanoid bullshit that people are trying to uh you know shill, right?
SPEAKER_04I think there's a robot behind you there to check it out, man. They're coming for you.
SPEAKER_02Was there a lot of humanoid stuff at MWC or embedded world? Because I mean, we all I think we all went to CES, and CES it was like frickin' robot hands. You could buy 50 robot hands without walking 10 paces. Yeah, true.
SPEAKER_03Yeah. Yeah, I don't know. Rob, don't don't laugh, man. Three years ago I saw one of these things like you know You got molested by a robot three years ago. I saw some some guy get molested by a robot.
SPEAKER_02It was it might have been villain. It might have been villain. You might have just you might have just re-traumatized it.
SPEAKER_03Well, yeah. Yeah, you gotta be careful, man. There you go. Yeah, absolutely.
SPEAKER_04Wow. Yeah, I know we've said this before, but everyone is everyone is regurgitating the exact same use cases and outcomes for AI that we used for IoT. Like, literally, it's like, wait a second, that was my use case. We already used that. You're I'm serious. It's back to the double counting, regurgitating, whatever. They're just like, well, yeah, now that I have AI, I can tell the sprinkler system to turn on if it, you know, if it's not raining. I'm like, no, no, no, we didn't need an LLM for you.
SPEAKER_02You know, if value X equals more than switch on.
SPEAKER_04Yeah.
SPEAKER_03Yeah. I know. I mean, literally, you're so right, dude. Yeah, and then I'm like, I saw a lot of that at MWC, and I'm looking at these demos going, wait a minute, you're this is like shit from like, you know, when we're at IoT solutions world congress, Rob.
SPEAKER_02Like almost we need is an AI-enabled vending machine that anticipates what you want and and spits it at you in or makes a drink that's I'm sure there's someone prints it somewhere.
SPEAKER_04Walking by the vending machine and it just like shoots it right out at you. You want a Snickers bar right when you want a Snickers bar whether you want whether you like it or not.
SPEAKER_03Exactly. Oh my god. Oh my god, it's too funny. Wow. I mean, you know, we're not being negative. It's just friggin' silly, right? I mean, we're it's why why does pointing out silly stuff equate to being negative or pessimistic or cynical? It doesn't, right? Well, I mean the people huh?
SPEAKER_02Sorry. So the positive aspect might be a course correction post-COVID, because if you think about the societal damage that was done, you know, during COVID, but yeah, when you look at again kids, the kids' generation, you know, they they they had a pivotal period in their lives where their social network was gone.
SPEAKER_00Yeah.
SPEAKER_02I you know, and a hopeful interpretation is that actually AI is the shock. AI is the shock that is pushing people back towards analog, face-to-face, human what can what humans good at, creativity, you know, conversation, you know, etc. Maybe I mean that every generation is smarter than the last. Maybe this generation is is figuring out you know how to go back and fix some of those problems spurred on by you know the trigger of AI. I mean that that that's the that's the positive interpretation, and there's no reason why it wouldn't be the case.
SPEAKER_03Yeah, that would be a positive. And maybe it's the AI slop that will make that that that happen. I mean, that's kind of what you're alluding to, right? It's it's because the AI slop is just gonna be tremendously unhealthy for everything.
SPEAKER_04I love the return to analog.
SPEAKER_02Yeah, it's it's my kids have bought my 15-year-old buys LPs. He goes and he goes and buys records, and you know, it's it's everything analog.
SPEAKER_04Everything analog start caring about things like cartridges and needles and things like that, and belt drive versus direct drive, and Dolby B versus Dolby C versus DBX. Like we should. That's what we should be worrying about instead of all this other nonsense. Yeah, yeah, yeah. You know, electrostatic speakers versus you know three million, three million uh uh what is it?
SPEAKER_03How many like 11 million songs that we'll never listen to? Yeah. Yeah.
SPEAKER_04That's why all the kids are going back to Yacht Rock, baby. Yeah. All right. Yeah.
SPEAKER_03Alright. So hey, um, you want to take us out, Rob?
SPEAKER_04No, I think I think the guy with a cool coffee talk background should take us out.
SPEAKER_03Yeah, I think you should too. Dude, that's freaking awesome.
SPEAKER_02This is this is what you get when you feed your resume to ChatGPT and say, I've got an INT coffee talk podcast coming up this morning. Build me a build me a background. This is this is what I came up with. I was like, let's use it. The irony being that the brain, the AI brain is is replacing mine. I think there might be a subtle message from uh from Chat GPT.
SPEAKER_03It's like I'm right behind you. You're you're being powered by AI.
SPEAKER_02I know I've been watching, I've been watching this dinosaur series with Morgan Freeman, and and you know, the analogue to what's happening today is kind of like well, all these dinosaurs didn't understand what the fuck was happening. Is that good, by the way? Is that a good dinosaur? Oh, it's really good.
SPEAKER_03Netflix, right? Netflix? Uh yes, yeah. Okay, okay, I gotta check it out. I love that.
SPEAKER_04Anything where I hear Morgan Freeman talking, oh I'm all in. It's like I love it.
SPEAKER_02It's almost as good as David Attenborough. But yeah, yeah.
SPEAKER_04I love David Attenborough, absolutely. Yeah. Yeah. But watch out, Richard Richard Attenborough, his brother, was in Jurassic Park. So watch that.
SPEAKER_03Oh no way. Wow.
SPEAKER_04No, no, the guy's name was that in the movie.
SPEAKER_03He was the the old guy.
SPEAKER_04Yeah, the old guy. We just found them in amber and then we put some DNA together, and then used the rock for the missing lake. What could go wrong? Yeah, they're all girls.
SPEAKER_03They could have just used bird. I mean, that's birds are dinosaurs, you know. So they're just flying. All right, dude, take us out, man. What does that involve? I don't know that I've overcome that.
SPEAKER_02Thank you for joining us. Thanks for listening in. Uh yes. Well, yes, indeed. Thank you for listening to it. Let's make shit off, Matt. Um, hopefully this rambling has um has contributed something to your thought process. Um, if you've witnessed the the analogue machinations of our old wrinkled brains um and you take some value from it, fantastic. If not, go put it in Chat GPT and see if that can come up with something more sensible.
SPEAKER_04Love it. Love it.
SPEAKER_03That's it. By the way, everyone's brains are wrinkled, just so they know.
SPEAKER_04Yeah, well, we have other wrinkled things now, so that's why we are old. That's why we have filters on our laptops that do smoothing and things like that. Indeed. All right. All right, guys. See you guys.
SPEAKER_06Take care. All right.
