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
#305 OpenClaw
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
All my friends are gonna be there too, you know.
SPEAKER_03In hell?
SPEAKER_02Yeah, that's what they mean.
SPEAKER_03Um that's awesome. I love that. Happy GTC 2026 week. And welcome everyone. Yeah.
SPEAKER_00Welcome, welcome back or out of that. Is it done? Is it still on today or is it done? I don't know. I had enough of it. I had to get out of there.
SPEAKER_03Yeah, it's hilarious. Yeah, it's all good.
SPEAKER_02Um that's cool. That's cool. Yeah, you know.
SPEAKER_00Wow, look at that.
SPEAKER_02Lots of lots of non-developers at GTC at this technical conference. Lots of Wall Street people.
SPEAKER_03Yeah, lots of Wall Street people.
SPEAKER_02It's like, what are all you what are all you people doing here? This wasn't for you, and now everybody and their dog is at this conference.
SPEAKER_01But they tracking around big bags of cash just to add to anyone, any young promise.
SPEAKER_03Yeah. Anyone. Yeah. Um pretty much fun. That was there. Had a lot of talks. It was friggin' busy. But uh, you know what? I'll tell you right now. Um uh remember, you have to watch IoT Coffee Talk at your own risk. You know? If you take us seriously, you have a big problem. You need to take out insurance. And um we're just here to have fun.
SPEAKER_02Do you have to do you have to listen to Jensen at your own risk? Yeah, you do apparently. That's for sure.
SPEAKER_03Yeah, and uh what else do we say usually in our disclaimer? But yes, this is for uh entertainment. It's not even informational. It's just if you're not here to have fun, just there might be you might be some information.
SPEAKER_00You're not trying to learn something.
SPEAKER_02Here's my takeaway. One of the things, you know, you know, if we sat there through his what, two and a half, almost three hour keynote on Jetson's deal. One of the things he said, obviously he says a lot of things, but one of them was just you know, talking about, you know, when I came out here in 2016 with this chip and it was this powerful, and today, you know, we've increased 40,000 percent or whatever, some insane number.
SPEAKER_03Yeah, it was like four million or 400%.
SPEAKER_02Yeah, and it's like but then I sat back and I was like, well, he's standing on stage, and the entire stage is just giant racks of servers going on forever, and that's what he meant by that increase. He didn't walk on stage on the to to this week and hold a chip saying this is 40 million times better. He just said, if you buy all this shit, it's 40 million times better.
SPEAKER_03Yeah, exactly.
SPEAKER_02And I started to and I you know, you start thinking, I was like, what are the companies that can afford that thing? What is the 0.01% of the largest companies in the world that can afford that? And then so what are we what are we gonna do for everybody else that wants this AI stuff, you know? The AI grid.
SPEAKER_01I think you have to radically redesign the architecture of how AI works would be my that my better. I've been talking to a few companies that who are taking a neuromorphic kind of mammalian brain approach to edge AI, particularly. So rather than you know, because if you honestly, if you look at AI today, it's it is the definition of what Saturn and Delhi used to talk about as cloud first development, isn't it? It's like I'm in the cloud, I've got unlimited resources, I can ingest any amount of data, it then becomes am I that 0.1% of the companies that can afford to throw you know a bag of odds at it or not? Whereas the human brain and animal brains function on much more constrained resources. Oh yeah, yeah, defined kind of rule set, yeah, and the constraint of time as well. You know, you've got to you have just this this amount of time to make a decision, you've got this amount of resource to make it, and you've got these pre-given you know uh conditions like don't do this, it'll kill you, don't do that, it'll you know, whatever. And and yet we are apparently intelligent and able to produce, you know, uh people are apparently rather an artist rather than or just um but it's really interesting. I mean uh to your point, Rob, I think that's the thing. We we've designed we've designed AI, you know, logically, but in such a way that the act access to those tools at the most extreme level are very well yeah, I mean it's it's the uh you know the IT workloads are always going to be in the cloud, but you know, of course I'm biased, but you know, there's no reason why.
SPEAKER_00And actually I was talking to the CEO of DeepX at um Bedded World last week, Kim. And uh very passionate on this topic around you know, more distributed AI infrastructure uh away from sort of these uh US-based clouds, basically, and how do you run you know AI um more locally with more sovereignty? And uh it's inevitably heading in that direction. It's just a matter of whether or not you look in Korea, it's like Furiosa, Rebellions, Deep X, yeah, you know, SK, telecom. I mean, everyone and you know, multiply that by a hundred countries. There they're everyone's working on how to how to move AI into a more sovereign base. Right. Yeah, it's gonna take time.
SPEAKER_02I like it like along with like Alistair was talking about the brain. I did see this post from you know Peter Diamandis earlier, and he basically he was like the human brain uses only 20 watts of power but performs one exaflop per of operations per second. Today's top AI chips burn 700 watts for one petaflop. So we're still a thousand X less efficient than biology.
SPEAKER_00It is the uh we're in the steam engine era of AI, right? Yeah, and then that word and uh see the wheels turn, you're like, oh, this works.
SPEAKER_03But but you know, it it needs to be completely re-architected, replatformed, and you need a whole new technology base. You know, we're we're already hitting walls. Like what Rob just mentioned, you know, we we've scaled out uh four million times since whatever, four years ago, or five years ago, or whatever years ago. Um that's not you know, that's not scaling the density of compute and the efficiency of compute, right? You're and I I said this three years ago because he he had all these scaling laws. I was talking to Dylan Patel uh back when he was like doing a newsletter, uh and I told him these scaling laws are bullshit. You know? Um he you are bound to Moore's law. You can't go optical thinking that you're somehow at a at processor level or processing level, uh somehow uh you know divorcing yourself from silicon. You're not. You're on that roadmap. And um there's this conflation of the technology, the you know, component, and then the system. And what Jensen's talking about is scaling at the system of system level. So, you know, he's really playing on the ignorance of the broader population who uh are basically you know are those folks that you see at these crypto conferences, you know. I I don't know, I mean we should revisit those crypto conferences, but that's what GTC feels like now. I mean, like info is coming in.
SPEAKER_02Because if it had been like Pat Gelsinger out there saying, well, here was the 286 or 386 running at 20 megahertz, and then today here's a you know core ultra whatever and it's running at 4.5 gigahertz, and that is a however million percent increase in well since the 80s, I guess, right? Yeah, um, and so when Jensen says look at this increase, it's like, oh, well, you had to build buy this whole building full of stuff to get that increase, you know?
SPEAKER_01And people are like, oh to what extent does did does do you um collectively, to what extent do we think that uh that quantum computing is is the solution? Because it used I I used to hear quantum computing and AI kind of thrown together. I don't see anyone actually combining those two concepts. I don't think I've seen a startup talking about it. Yeah.
SPEAKER_03I mean that I I kind of learned that the hard way.
SPEAKER_01That I hear analysts talk about that.
SPEAKER_03Yeah, yeah, yeah.
SPEAKER_01Oh, quantum competitors are gonna stop it.
SPEAKER_02No, that's not they're like pears and oranges and apples. Right. Yeah, exactly.
SPEAKER_00And we have people actually apples and potatoes is more than that.
SPEAKER_02People cover quantum in IDC, they hate me because I'm like, yeah, no, I think it's gonna be awesome. I think it's gonna be widely deployed in the 2040s. And they're like, 2040s. I'm like, yeah, the 2040s. That's yeah, yeah. That's not what I on Q is telling us.
SPEAKER_03Yeah. Well, actually, IonQ will tell you that. Jensen will waffle on that, right? Depending on the stock price of any of these quantum companies, notice that he, I don't know him. Did he have a quantum day this year? Which was like the most ridiculous thing I ever seen. This year last year. The absolute waste of time.
SPEAKER_02I remember when he did it last year or the year before, and he's making fun of all the guys. Last year. He's like, Can I run Excel in your quantum computer?
SPEAKER_03Yeah, but no, the laughable thing was he came in um trying to um uh you know, block back his earlier comment that tanked a bunch of uh quantum stocks, right? And he said, Well, you know, it could be like a year away, and then he brings all these people up on stage who are the founders and CEOs of these companies. Well, you know, it's really about 30 years off. So you you know, his Jensen's original comment that tanked everything was actually what these guys were probably telling Jensen. Right? And then he created this weird talk track that quantum and somehow accelerated computing are going to um, you know, uh be you know it's going to bring about commercial quantum in whatever five years. And that ended up not being I mean, you know, he spent like a whole day walking that back.
SPEAKER_02That's not necessarily a bad fact. It's obviously a bad theory, like because they use the same talk track with you know, robots kind of suck, and then they're getting together, and it's like, oh, it's because humans are designing robots and we're not good enough. But when we hit HEI and AI helps us design the robots, then they're gonna be you know doing kung fu. And so the same talk, the same narrative is we're not smart enough individually to build quantum computers, but once we hit the singularity, that will help us build, it'll solve all the problems that we can't solve.
SPEAKER_03Well, I'll tell you this right now. Um, physical AI um this year is kind of a joke because it's actually remote control. There's nothing very little AI about it. You know, it was funny. Um I stopped by the Anaconda booth, and there's like this friggin' uh like seven-foot humanoid robot like punching the air and doing like really dangerous stuff. And of course, there's a crowd within the ninja blade, you know, in it.
unknownYeah.
SPEAKER_02I was like, why is it when I have to why is it when I hear Anaconda I think of Sir Mixelot? You know, like you live in Seattle, you used to live in Seattle, dude. Remember Baby Scott Back?
SPEAKER_01Oh yeah, you call it you call it Mr. Mixelot.
SPEAKER_02You know this yeah, I don't know, I just hear Anaconda and it's just like whoa!
SPEAKER_03I always I always think of uh Ice Cube and uh the movie Anaconda. Oh right. Uh the guy Yeah, but the guy was saying, Oh yeah, it's that guy over there with the remote control. Oh and then you go you see all these other demos. Guess what? These robots are programmed. Probably the only qu AI-ish thing that was uh physical AI-ish, let's say, demo was that um frozen little snowman dude. Uh I forgot what his name is. Olaf. Olaf, yeah. Olaf. That was that was kind of cool, but I mean that's done by Disney thing. And Disney has been doing this stuff for Abraham Lincoln.
SPEAKER_00I mean, he gave the whole Gettysburg address. That was physical AI, wasn't it?
SPEAKER_03I don't know. That was a very basic form of physical AI, right? But it I mean, all of these things are uh for the most part programmed.
SPEAKER_00Um although I would say the the robots that are in you know warehouses and things like that, those are that's real physical AI in terms of you know navigation and uh task completion and things like that. When you get into this humanoid laundry folding robot thing, which is still like uh a sales pitch.
SPEAKER_01I mean, even with the automated robots, I think you know you bang on about this last week. I think you know, if you look at the promise of AI, particularly on AI code gen, well it's obvious you can you can use more code, prototype faster, blah blah blah blah blah. But the but the the framework that is necessary for these things to be developed by companies that have liabilities and customers and needs to deliver services don't exist. On the code generation side, you can't you can't build build a software build of materials, you can't you can't actually trace provenance, you can't protect your right thing. And on physical AI, you have to segregate these robots into a separate area away from human beings because they cannot adequately uh detect the presence of a human or predict what that human being's going to do such that they don't get run over. And and so I think that you know, I see a lot of activity in the startup space, you know, kind of tackling these problems. These companies are all relatively early stage, and it seems like we only kind of realized this was an issue a couple of years ago and starting to address it. But I think if you went into most enterprise customers and and sat down and said, okay, how do you use AI? How do you want to use physical AI? How do you want they would have a laundry list of possibles, but not many things that they could practically do without bankrupting their business in a law sort of.
SPEAKER_03Yeah, yeah, yeah. I mean, I think that's a great point. That's a great point.
SPEAKER_02Um by the way, there's some sad news today, and you're not gonna believe this because I thought this was impossible. Chuck Norris has passed away.
SPEAKER_03Oh, no way.
SPEAKER_02Your alter ego? I didn't think he could die. I thought I didn't think so either.
SPEAKER_01Yeah, that is my kids' karate school. My kids um karate was one of actually, but there were two. One got handed over to another, and both of them were students of Chuck Norris. Um he looked like he'd been reanimated at the time, but I didn't you didn't want to say that to him because he could probably still rip you, despine you.
SPEAKER_03Oh my god. Wow. Yeah. Well, you know, he's like he's like 90 years old or something, right? How old is he? He's up there.
SPEAKER_02He's up there.
SPEAKER_03Yeah, right. I mean, he's like William Shatner. Isn't William Shatner like a hundred years ago?
SPEAKER_02Shatner's still alive and he's in his 90s. Yeah.
SPEAKER_03Yeah.
SPEAKER_02So now that can join Bruce Lee and they can reenact, you know, like Enter the Dragon and all that stuff that they can kick each other, kick each other's butt.
SPEAKER_03Yeah. Exactly. Yeah. Exactly.
SPEAKER_02And um wow, good times. There you go.
SPEAKER_03Yeah. In the hall of life, in Valhalla.
SPEAKER_02That's right. In Valhalla, brother. Valhalla. I'll see you in Valhalla. It'll be glorious. Yeah. I hope you had a glorious death. Yeah.
SPEAKER_01I just I just wish they made more of that chaos series with Jeff Dolblem. Do you remember that? The whole Valhalla thing just reminds me of that. Did you watch that series? That was you didn't, you've got to watch it. They made one series. It's Jeff Dolblund is a Zeus, the King of the God. You gotta be kidding me.
SPEAKER_03I can't imagine that. Oh, you should watch it. Oh yeah, that that's I mean, that description in and of itself is hilarious. Yeah, I I could already imagine that. That's awesome. Makes no sense. Which is awesome.
SPEAKER_02Didn't expect to hear Zeus say life will find a way.
SPEAKER_03Yeah, but you know, the thing that really tripped me out was like the Nemo the Nemo claw. Nemo claw. You know, and it's funny because I'll tell you right now, going into RSAC, all people are talking about right now is the exponential exploitation of these technologies by threat actors. That's all. So the the this script has been flipped, you know, dramatically, you know, 180 since Chat GPT first came out. And the agentic stuff, especially the swarm stuff, man, right now, um, the reports that are coming out talking about how the fidelity quality um of the threats, the volume, is just spiking. And so one of the biggest anxieties SOCs are gonna have socks and uh you know operators are gonna have is just you know um they're gonna have to compartmentalize everything. I mean literally, we've been joking on uh it's actually hasn't been a joke, it's more of a forecast or prediction. Battlestar Galactica. You know, somebody actually used Battlestar Galactica and tag me. Related to uh Somebody listened to you one time. No, somebody yeah, somebody has been listening. Well, the only time I ever really mentioned Battlestar Galactica is here. Because we coined it on the show. But the people remember it. Yeah, yeah. Um, but man, I'm telling you, uh and um you know, if you think about what uh Nemo Claw Claw is with the open shell, uh that's uh you know, extreme isolation or trying to isolate these agents. And uh someone made a comment uh um yesterday on a call asking, well, if you constrain these agents to that degree, um how useful are they? How much can they actually do? Which, you know, we've been talking about that forever too. So it it this this just shows you how late to the game everybody is in thinking about this stuff. And the agentics stuff, um I I I think um is just gonna create an a a massive, you know, COVID moment for AI, I think. People are gonna freak out. I mean, literally, this is like COVID, right? Because now you have um you know agents that might have alignment problems. There there are no tools to really instill you know um safe um guardrails and um into these things. Right. There isn't. But you have companies like NVIDIA now promoting it like this stuff is ready. And one of the points I made to one of my clients was like, Man, open claw just came out like friggin' what, two months ago, three months ago? And then you're gonna go up on stage and say this is an enterprise grade. I I will tell you right now, it's not. It's not. And it's not because I'm just saying it for the sake of it. I talk to people. Oh my God.
SPEAKER_01Don't you think some of that's kind of the trap we always fall into, I think, or often fall into in technology, which is, you know, we move fast, break, start to bed, fix it in the code, wherever you, which completely ignores the reality of the physical world. I mean, that take um take robotic surgery. You know, we talk about the the that the opportunity of having an AI-driven device that can isolate and treat a cancer cell with an above. I mean, it's an amazing technology. But when are you going to see that deployed in a hospital in the US, given the litigious nature of the entirety of the medicine in the US? Well, a friend of mine is an anaesthetist. She carries so much insurance that she said to me in the book, if I ever have a malpractice suit, I'm done for. I can't. No, no, no. Which is 40% of her kind of salary at the moment to cover her insurance in the year. Oh wow. So how are we gonna and I think you're right, we've ignored, we have ignored the realities of the physical world in designing technologies or talking about them as being, you know, enterprise growth. When we don't really clearly there's not an understanding of what enterprise growth means.
SPEAKER_03No, and and and and and when you think about it, you have more robust frameworks for AI that have been geared toward enterprise that have not proven to be safe. And that's why you have the adoption issue. So these guys keep beating their heads up against the wall, you know, um regurgitating this trope that it's all about data. It's like, no, it's about security, guys. You know, you you haven't bothered to look at the architecture, you don't understand properly the technology, and you're projecting a delusion to the end markets. This is why you fail. You know, and this is all predictable.
SPEAKER_02That's what Yardis said. Yeah, it is. This is why you fail. This is why you fail. This is why you fail, young Skywalker.
SPEAKER_03Yeah. I wish Bill was on here because he had did you see his um, I don't know if it's the shirt. Uh it I it I think it says, around you fuck. Around you fuck.
SPEAKER_04Find out you will.
SPEAKER_02Oh my god, that's awesome. Wow. Okay, I did see something. That's good. I saw something funny today. Genius, man. It is genius. So speaking of open claw and humor, uh, our friend Darren, he's been on here before, and I saw him this morning at the little startup thing I go to. And you know how people talked it in the past, like climbing the ladder in the corporate world to get to the top, and how many people did you step on to make it to the top? His new deal is I'm gonna open claw my way to the top.
SPEAKER_00There you go. Nice playing words.
SPEAKER_02I'm gonna open claw. And so I can imagine the picture just clawing pet people as you make your way to the C-suite.
SPEAKER_03Yeah, yeah. Well, you know, what happens when everyone does that?
SPEAKER_02When are we gonna have the first open claw CEO?
SPEAKER_03Uh we probably already do. It's just it it at the moment has a human interface. You still need that human interface, right? You need something to manipulate. I mean, because think about it, that's what happened. I think that you you you got a sense of that with that whole Multbook thing that happened, right? They're hiring humans. It's going around hiring humans and using money to manipulate people.
SPEAKER_00Right, right. You hire the humans to manipulate the physical world on behalf of the AI.
SPEAKER_03So think about it. This thing trades crypto, goes and takes the money, and then goes and just hires people and to do whatever the hell it wants it that that it collectively wants people to do. And it doesn't matter that uh there's people in the background injecting prompt injecting and influencing the swarm. It it is what it is. Do you know what I'm saying? It it can go and exact in the physical world in reality uh what what what it's concocting in its you know collective AI brain. And that's what's scary. It it's not that AI is great, it's how terrifying some of these applications can be. Right? And you know, uh uh props to IBM for doing a lot of research on swarm intelligence way back in the day. Um there and it's funny, uh IBM is always a pioneer, but then they always get caught up in the hype cycle and can never really commercialize safely or uh, you know, in a compelling way commercialize a lot of great technologies and research that they've done. Um, but uh you know, you read some of the papers and you just go, oh my god, if we get to this point, um all these things that everyone's talking about, orchestration, management, um, observability, and being able to um you know govern these swarms, it becomes a huge question, right? For any application. But then now with this open claw thing, it it's it's insanity. Uh there are there are no tools uh for managing these things. It doesn't matter what people say. Doesn't matter what people say. And when I I listened in on an open claw session, these jokers get up and they're from the big companies, you know? All the big AI companies. These jokers, okay, get up and one side of their mouth talk about how um it this stuff is unsafe. We have no ability to truly make these things safe, and then out the other, they talk about how we need to innovate and we want to promote and everyone should use it, and enterprise, it it's going to maybe uh revolutionize industries. That's like hypocrisy, right? It's irresponsible. That's like that is the definition of irresponsible AI.
SPEAKER_02I'm gonna try to get OpenClaw to be on next week's show on my behalf.
SPEAKER_00Like crustacean avatar. That'd be cool. Yes.
SPEAKER_02Well, remember what remember one time Wakey came on and he had his avatar one of the walls. That was funny. Yeah.
SPEAKER_03Yeah. Well, you know, did didn't uh Mark post something about how much money Zuckerberg wasted on the metaverse?
SPEAKER_0080 billion.
SPEAKER_0380 billion. Oh my god.
SPEAKER_00That's the number I saw. Wow.
SPEAKER_03You know, it goes to show you.
SPEAKER_00A billion here, a billion there, it ends up. It turns into real money.
SPEAKER_02Yeah. Well, they remember right after remember right after Zuckerberg went all in on the metaverse and changed the name of the company. It was right before MWC a few years ago. Yeah. And then we all show up there, and like every company there had hurriedly changed the signs on their booths, on their stands, you know, like Accenture's like Accenture Metaverse.
SPEAKER_03What did I what did I um I I um yeah, over some beers, I was uh uh talking to a friend of mine at uh GTC, and I told him, yeah, it's like if you're the master, he's an analyst. I I told him, yeah, if you're a master of the marketing um hype, uh you're the master of nothing. Uh and there's a lot of analysts who do that. They'll just take the press release, they won't ask any critical questions, there's nothing thoughtful in uh their engagement with any of these companies, and all they end up just regurgitating stuff, right? And then you're a master of nothing. You're just master of somebody else's PR, and that it's it it's ridiculous, right?
SPEAKER_02And um, you know, when you go to the master of getting that company to still pay them money to say nice things about them. That's what they're the master of. It's like, oh I don't want to bash this company too badly, so I'm just gonna change a few words here. Oh my god. Because there's that aspect to this whole game. No, that's not jaded or anything. Yeah.
SPEAKER_03Yeah. Yeah. No, there absolutely absolutely is, and that I think that's the that's the stuff that diminishes the the um profession, if you will.
unknownYeah.
SPEAKER_03Um because there is a lot of value in just telling it like it is, right? I mean, there is there are valuable audiences for that kind of insight. Yeah. And uh it gets underbined by people who just um are you know don't have robust analysis.
SPEAKER_02I feel bad about some people we know who do say it like it is accurately, and they're not getting any money from any of these companies because they're telling the truth too much. Yeah, you know, they didn't get the memo about telling the truth, and so uh it's kind of a weird deal. It's like they're saying the right stuff, they're they're they nailed it, and yet they're living in a cave now, sadly.
SPEAKER_03So well that it it that that's where you become a founder. Yes. There you go. Okay, yeah, you're in the wrong business, right?
SPEAKER_01There are a lot of, I mean, uh, like uh at least from the slides of the world I see through Momenta, there are a lot of companies and a lot of founders who have realized that these huge gaps exist and you know they're they're building solutions. The the the challenge that they will have, I think, in some respects, is again it's the it's that I'm a I'm an industrial company, you know, medium-sized, large-sized company, you're a 10-person software startup in the valley. Am I going to bet my life on you? And so there's the cycle of the startup being acquired and then becoming part of a major provider, you know, that needs to happen. But if if if the potential acquirers are still in kind of cooking land pretending that the problem doesn't exist, you know, is that just a is that just a public face or is that really, really kind of what they believe that there isn't a problem? I'm I tend to think it's more the public face and that there will be some common sense and they will make some of these acquisitions and they will, you know, these tools will become available. But we're but the the the the risk, I think, is that AI is moving so quickly, so is the r rush towards disillusionment. I mean, that's the oh yeah, yeah, yeah. I mean cool. You told me it was going to be here next month. It's not here.
SPEAKER_00Well, there's the there's the tech, there's the hype, there's the business, the software and hardware is all moving faster than uh the development, typical development and deployment cycles of any of these products. And so it's almost like you're you were getting overtaken by the this wave and it's it's kind of messing up all the development cycles of all these products. So people are kind of tripping over themselves and pushing out code way too early and doing all this stuff. So it's interesting. Like when the when the pace of the change of the tech becomes almost too fast, you know, it becomes a real uh negative.
SPEAKER_03Yeah, but there's like a religious fervor um that espouses this idea of you know, break things and run away, right? Um this whole Silicon Valley mentality. And I had uh I had an interesting conversation about excellence and how that is completely the principles of excellence are completely absent, almost completely absent in Silicon Valley. That mentality is not compatible with break things fast and run away.
SPEAKER_00No, it's not fast, and that's that's if speed becomes the paramount, then you know everything else suffers, right?
SPEAKER_03So but excellence survives, and it usually is the signal above the noise of a bunch of you know um you know minimal uh you know uh MVPs, right? Yeah, in other words, um pseudo-crap. So um that but right now everything is so skewed on the crap side of stuff versus like quality. Uh I think uh Silicon Valley is in in big trouble. This is gonna be the last hype cycle uh for Silicon Valley. If this crap does not work out, Silicon Valley is, I think.
SPEAKER_00That's a bold prediction. I think hype cycles for Silicon Valley will be around for 100 years.
SPEAKER_03SVP, SVB, SVB, we have to remember SVB was almost the downfall of Silicon Valley. No, that's true. Yeah, so the government bailed them out, right? Right? They bailed them out, and they basically took all that toxicity and then loaded it up on the regional banks, the little guys. We have to remember that. Well, also people moved this up.
SPEAKER_00Right. Small banks, it's a tough business to be a small bank, and that's why there's all this consolidation and everyone's moving money to chase and other things like that.
SPEAKER_03Yeah, they almost put yeah. But you know what, Leonard?
SPEAKER_02You can't spell excellence without the letters MVP. People always do that. You can't spell blah blah blah without the three letters that are in their product or whatever.
SPEAKER_03There's no I in everybody doing stuff like that, right? Right, right. Well, you know, what I love is when uh they say that there is no particular letter in some you know a concept or principle, and then there is oh, okay, wait a minute. You're right. Okay. Yeah.
SPEAKER_00By the way, speaking of hype cycles, so I've been digging into a lot of these kind of automation platforms like uh Zapier and Make and N8N and all that stuff, trying to do some interesting things. And man, this stuff is really primitive, so primitive. You know, when you get into it, the APIs and things are so limited. It's it's like it's I was kind of disappointed. I'm like, there's gotta be somebody out there with platforms that I can use to build agents that do actual useful things.
SPEAKER_01And um, like if this then that's so basic.
SPEAKER_03No, but you know what? Think about it. The younger, I mean, this is what I've noticed uh through my my career, and I know you guys have seen it as well. Uh you have the younger generations that think that they've freaking invented something. But it was like it's either stuff that already exists, and we we see this with like you know, more seasoned folks too. They didn't think like, let's say these middleware concepts don't exist. It's like, no, this shit, you know, we were doing that 20 years ago, man. You're trying to do this now. They they're just not aware, right? And you you know, um Alistair, you know, you probably know in the startup world, you get all you get pitched all these ideas, and and and you know, these founders will walk up to you with concepts that you probably worked on and your teams have worked on a long time ago. These guys just don't know that doorbell on the video camera!
SPEAKER_00You guys watch Shark Tank? I had one on the other night. It was like it was like a tool that like cored a banana and you could put like Nutella in the banana. That was like their thing.
SPEAKER_01This is what we've come to. I've got I mean the thing that you see quite a lot, and I'm not I hate to generalize, but you do see still it's not so much that people don't that that that you know, oh wow, I've invented this thing and doorbell with a camera, etc. It's more that that um people have really diagnosed interesting problems and they've developed a conceptual answer for that problem, but there's they don't know how to affect that. They don't know how to market channels.
SPEAKER_00This is great, you know.
SPEAKER_01Yeah, yeah. No, it's but but it's like how okay, what are you gonna do next week? And the week after that, and the week after that.
SPEAKER_00No, it's uh but you know, that separates the good founders from the ones that are will probably fail and try again and learn and keep at it, you know, and uh that separates the wheat from the chaff. But I mean it's sorry. But no, I was gonna say, I think it's there's some founders out there, and I'm sure you've met them that have that kind of holistic view of a business that it's not just the idea, it's the execution.
SPEAKER_04Oh yeah, yeah.
SPEAKER_00And they differentiate themselves. I'm sure if people are on the investor side of the table, they're like, okay, these people actually know what they're talking about, and they other folks come in. And that that's just the law of the jungle, right, in this startup world where uh maybe they get that lesson. I had a um colleague who uh sent me a deck and asked for some feedback on it, and I gave them some pretty direct feedback. Like, hey, this thing is missing like 90% of the shit you need in here. It was a great idea, and I'm just like, you know, uh better I tell you than you get in front of the meeting coming up. We don't want to go in there with this thing because it's gonna queue up. So you know, that's part of the cycle, and that's what's exciting about working with a lot of startups is that they they learn that stuff. I met a guy in Nuremberg, he's like 22 years old, and he we had this big dinner party in Nuremberg, and uh it was awesome. And he's like, Yeah, I never had a job. This is my first job, and we've invented this thing, and it was a pretty like uh rich technical like uh problem they were solving. And uh, but you know, he was just scraping the the beginning. He's like, and he and good news he was self-aware. He's like, I realize I don't really know what I'm doing right here, but you know, we built this cool thing and we're looking for folks to help us figure out the rest of it. So so that's just like the nature of the beast, I think. I don't think that's uh I think that's gonna be there forever. That behavior. Yeah, that's just that's that that's part part of the human human molt book.
SPEAKER_01You need that enthusiasm and you need naivety.
SPEAKER_00Yeah, they don't know what's not possible, so they go for it. And you know, one out of a hundred get it right. So good. That's good. Yeah, we don't bet on the other 99.
SPEAKER_01That's all we didn't know that there was a rule that said we couldn't figure out the answer to this, so we just figured out the answer to this. It's like excellent.
SPEAKER_02You don't want the guy to go into the pitch and and they're like, what's a pain point? Why do I why would I care about that?
SPEAKER_03Yeah. I don't know. Yeah, but you know, um, I and I I tell a lot of startups, just go and freaking do a project. Um do it like Bill Pugh. You know? Um go if you do have an idea, find a client, make a consulting business out of it for the moment, solve the problem, prove it, you know, uh go through the process of figuring out, you know, uh how you can secure the IP for it, and go old school, build a business. And that that's the thing that bothers me about the startup uh culture and the VC culture is that it fosters this fundraising fixation, which that you know, the venture stuff should be about building a business, and it's not about founders raising funding. And uh man, I'm this is the thing that stuns me. I get in conversations with grown-ass people that don't know the difference between revenue and funds. They think funds, oh he they raised a billion dollars. That's investment funds going, that's like cash coming in, it has nothing to there's no business there in exchange for equity. Don't complain, yeah. The revenue with a VC giving some companies a friggin' money, you know, you haven't closed the loop, but there's grown-ass people that don't know the difference. Like some people don't know what um what you know corporate taxes get taxed on, they think it gets taxed on revenue. So then that and that's how they have this whole stance about we shouldn't be, you know, taxing corporations because then it'll shrink their business. No. Taxes apply after they've made all the money, paid pretty much everyone. You know, so grown-ass people don't know.
SPEAKER_02Pete talking about giving advice, you know, someone I as I'm unpacking books and stuff since moving down here, I found an old book that's from the 90s. And my first startup that I did with my buddy Darren, it was a mobile device management company, it was before I joined Microsoft. And I remember having this book, and I remember it's like, okay, the seven questions that they're gonna ask, you know, and this is really big. What is your product? Who is the customer? Who will sell it? How many people will buy it? How much will it cost? You know, and all this. How much will you charge? It's it's really funny, you know. And you've seen a million deck out there, you know, when will you break even? Market research. It's really funny. You know, I feel like we've downloaded every template of pitch deck ever created.
SPEAKER_00Basics. Well, they ask all those questions on Shark Tank too, so that's another way to get educated.
SPEAKER_02Yeah. It's like, you know, what's the pain point? How are they solving that pain today? Right. How are you going to solve it? Why will it be better?
SPEAKER_03Oh my God, you sound so smart, man. Yeah, really.
SPEAKER_02Well, you just it it's not intelligence. It's like you've just seen, you've been involved in it so many times that you just kind of know all the standard stuff, you know? Yeah.
SPEAKER_04Yeah.
SPEAKER_01It's surprising how often you don't get a straight answer to those questions. Don't. I mean, particularly the whose problem is it? How much is it worth to them? How do they buy things? How do you fit into that channel? It's a lot of it is the market. You know, are you going to go and sell to that end customer who has the problem that derived from which they can derive value if they sold it? No, no, you're not. Okay, so who does, and how are you going to design a product that actually provides an economic incentive for the guys in the middle to actually take it to the end customer and sell it? And it's like, but the end customer really needs this.
SPEAKER_00It's like, yeah, but the we need more Russ Hannemans, that's what I think.
SPEAKER_01Yeah, you've got to figure out how to those are kind of I mean, they're not uncomplicated questions, really. I mean, and I understand why you know the focus is on the hard problem, you know, solving the hard because if you don't have a product that solves a hard problem, then you know what's the point. But you've got to have the other stuff too. You gotta have um yeah, and uh you figure that out by building a business, like you said, I think build a real business and then take that product and turn it into code, you know?
SPEAKER_03Yeah, and then you go and productize it, then you build a go-to-market around that, and then you have a business. But yeah, I I mean I I think we had this conversation at MWC like three years ago with Ray, um Ray Ozzy. Uh exactly that, the identity thing, remember? Um, that was and I on IoT Coffee Top, we were talking about that. One of the biggest challenges for a lot that a lot of the IoT players face is really understanding where they fit and how what is their identity in the grand scheme of and ecosystem on value chain. And you know, and so the questions can be very simple, the answer uh it can be really tough to arrive at, at least the proper one. And so this is why we see so we've seen and continue to see so many companies just really confuse themselves. And now in the AI space, it's e it's probably even worse. And a great opportunity to uh for IoT people to like look at this stuff and be entertained and maybe reflect back on how the previous uh IoT cycle was, but hype cycle was, but yeah, being able to to arrive at a at a clear um identity, uh one that positions you to uh you know align the organization, your strategy uh to be a winning one. And this is um uh a conversation I I had with um um the folks at uh the help me out guys here, the smart antenna guys. Uh oh Igneant, yeah, yeah. Um and um yeah we had uh great conversations about that long time ago. And um but you're hitting on something that's really important and you know at some point you have to fit into a solution. And if you're not the and you're not at the end market, know what your role is. Otherwise you you see these people doing components saying, hey, we're gonna revolutionize uh XYZ end market. Like, no, you're you you create like a sensor, dude.
SPEAKER_01You know, embrace that identity. We made that fundamental mistake, you know, in the early days of IoT platforms, and uh, you know, certainly I did, that that you're gonna design this platform and then enterprise customers are gonna be able to build their IoT solutions. And that I and I came from a consulting background immediately before, thinking that, oh, you know, if we can go away with the millions of dollars they have to pay a consultant to build a custom solution by this platform. And the problem was that those enterprise customers that there wasn't a connection between a cloud platform and those enterprise customers. There weren't a developer, there wasn't a developer base in those enterprise customers. And actually the people in the middle, the the notion of a platform that made it easier for customers to build IoT was antithetical to their business model because as a systems integrator, I don't want it to be easier for you. I want it to be hard because then you need to pay me a load of money. And the harder it is, and the the harder it stays, you know, then then you're gonna need me next year and the year after and the year after that, and I'm gonna be here forever and making a ton of cash. And it the fact that the very notion and the core concept clashed fundamentally with the interest of the channel, I think was in in part at least why none of that.
SPEAKER_00Yeah. A little PSA for uh we're having a startup pitch contest next week at EGI San Diego. Yeah. People are in there in the area and want to see, and actually Hans Alistair is moderating a pitch contest. So if you want to see startups in the wild, they will uh which time is that which time? I think it's on the 25th, I think. I have to double check the session. 24th through the 26th, but he's doing a startup pitch contest at the end of one of the days, and so we're gonna get a bunch of startups up there battling it out and pitching their stuff. So people want to see that.
SPEAKER_02Awesome. That'll be fun.
SPEAKER_03Oh my god. I'm so bummed I'm gonna miss out on that stuff. But I I will know coming then. Uh I'll be there for dinner. We'll we're doing dinner on Sunday. Dinner? So we'll probably do a uh right?
SPEAKER_02Probably not doing dinner. No, I'm not sure.
SPEAKER_03Yeah, we'll do uh the two of you will have dinner on Sunday. Yes. No, it's not just the two of us. Who else isn't is it just the two of us?
SPEAKER_00Oh Rob's gonna be there, alright? Are you? Okay, so it's just Leonard and I then.
SPEAKER_03Oh, okay. Well we'll we'll uh we'll have a chat then and you know, record our stake. We'll record our but um yeah, I'll I'll try to be there on uh Thursday. I will be there on Thursday. Yeah. But um I'm gonna fly out on Thursday. Okay.
SPEAKER_02Uh hang on, let me say.
SPEAKER_03But while while you're trying to figure that out, Alistair, you I I think you're you're totally spot on about that. And um the AI community, uh uh all those pieces that that weren't recognized at that stage in the IoT platform cycle, type cycle, uh these AI guys don't see it. They really don't see it. And so all of this stuff we know is gonna buy everyone in the ass at some point. They're so fixated on tokens and blah blah blah and how tokens are intelligent. No, it's not. You know, people pay for outcomes and uh express value in the unmarkets, and when you have something that is so squishy and unreliable, it's not deterministic. That's the problem with determin when that probabilistic nature of AI goes up the commercial channel, there's all kinds of problems that we haven't seen before, and it it it's gonna create some economic backlash, right? And uh, you know, it's it will destroy a lot of these PowerPoints that are being put in front of companies with promises of exponential, whatever, a hundred trillion dollars of economic value, man.
SPEAKER_01Being a disruptor sounds great in a PowerPoint deck. Oh, you can point to an industry and say, I'm gonna upend the economics, it's getting fun. You can but you can build a nice model that explains why the end customer, the enterprise, you know, saves a ton of money. But the problem is if you're dependent upon the people you're disrupting, you know, to actually be the channel from you to that. Sure, sure. Guess what? They don't want to be disrupted. No, they don't want to disrupt it. They want to carry on just as it was AI is probably.
SPEAKER_00That's why Tesla sells their car their cars directly not through a dealership, right? I mean classic, classic change in the go-to-market strategy to be disrupted.
SPEAKER_01But what's the equivalent in AI though, then? What's the equivalent of having your own dealership in AI?
SPEAKER_00It's tough. I mean, yeah, bypassing system integrators.
SPEAKER_02I am waiting for a startup to come pitch, Alistair, and they're gonna walk into the conference room and they're gonna drop down their Mac Mini, and a hologram is gonna come up. Oh an open claw hologram is gonna do the piece.
SPEAKER_03Oh my god. That would be dude. You know what? You should you should be uh one of the primary writers for Silicon Valley 2.0.
SPEAKER_00Yeah, I was gonna say that sounds like a reboot. Reboot in the making.
SPEAKER_03Oh yeah. I mean, if you I've been watching this show, oh my god, it hurt hit so hard. Oh yeah.
SPEAKER_00No, Mike Judge, uh Silicon Valley. It's amazing how prescient that is.
SPEAKER_02Oh yeah.
SPEAKER_00But no, it's due for a reboot for sure.
SPEAKER_02Oh my god. You're right. Uh uh Russ Hanneman, the return of Russ Hanneman.
unknownYeah.
SPEAKER_00With the Lambo.
unknownOh.
SPEAKER_00Three combos. What is it? His trace comma, Trace Coma tequila?
SPEAKER_03Oh my god. So funny. Yeah. Well, who wants to take us out?
SPEAKER_02Are we the only ones that know that the whole thing is kind of BS?
SPEAKER_03Who? What is?
SPEAKER_02Just the whole Silicon Valley and the silliness of it all. You know?
SPEAKER_03A lot of people's blood coming in. There's always fresh blood.
SPEAKER_00And me actually taking it seriously as part of the entertainment, you know, as part of this. You need some people that are not in on the joke. Yeah, yeah, yeah.
SPEAKER_03For the joke to happen.
SPEAKER_00Exactly.
SPEAKER_02Good point.
SPEAKER_03Yeah, that but you know, we're we're the classic uh grumpy old guys, right? Yeah, sure.
SPEAKER_04Get off my line.
SPEAKER_03Especially that don't know anything. Yeah, we know nothing. Don't listen to us because we're we're not disruptive and all these other stupid buzzwords that everyone tries to characterize themselves as.
SPEAKER_02Leonard exudes grumpy old guy. I see.
SPEAKER_03Yeah, I do. I'm gonna be super grumpy.
SPEAKER_02Grumpy.
SPEAKER_03Horrible.
SPEAKER_02As you sit there in your in your beautiful vineyard backdrop there.
SPEAKER_03You see that.
SPEAKER_02Yeah. Picturesque.
SPEAKER_03But hey, all I gotta say before we sign off is what the hell happened to Satya? We don't hear him talk about AI anymore. Has anyone heard anything from Microsoft about AI? They were like supposed to be a little bit more than a lot of people. Looks like uh the winner and blah, blah, blah, you know, all these balls.
SPEAKER_00Well, you know, this is the issue, is they don't have their own, they have not commercialized their own foundational model IP or frontier model IP. And so they did these deals, and now it's open AI, which has become, I don't want to say they've become toxic, but they're a bit radioactive in some cases. And so, you know, then they're competing with Google, who has their own foundation models, and you know, and and anthropics are getting market share. So yeah, I mean, Microsoft's like in the heat of the competition, but maybe without as many weapons as they thought.
unknownYeah.
SPEAKER_04Well, didn't it fall?
SPEAKER_03Maybe it's not just them. Everybody was saying that they were gonna win, except for maybe us, right? Or they are winning.
SPEAKER_00Well, they might still eventually when things settle down, but I think they probably need some.
SPEAKER_02Because they do own what 20, 30, 40 percent of open AI or something.
SPEAKER_00They do. They do.
SPEAKER_02So they get to tag along with their upside.
SPEAKER_01But at what point do you pull the trigger and buy the rest of it, though, to to to take half the uh the value of Mark stuff to let it public?
SPEAKER_00Well, you have to wait for them to go public and then crater and then buy the rest of it. Don't buy it now.
SPEAKER_03Well, that might be Apple, actually. That's what they're waiting for. While everyone's saying that they're behind on AI, uh Apple's probably just waiting for every all of us to crater and they're gonna buy everything a penny on a dollar.
SPEAKER_00Scoop it up with the office chairs, you know.
SPEAKER_03Maybe they know more about AI than everybody else.
SPEAKER_02Exactly. Yeah, that's what I'm that's what my company did with Juli. We came in and we scooped all that up.
SPEAKER_00The box, did you build the box?
SPEAKER_02We had the box, yes.
SPEAKER_03Do you have the books? I'm telling you, the the future writers of Silicon Valley 2.0 are listening to this right now. Maybe we should get in.
SPEAKER_00I think we need to work on our own fan fiction script.
SPEAKER_02Oh, yeah, we should. I think that probably is the most promising future for all of us. All right, everybody, thanks for joining us on IoT Coffee Talk. Uh, hopefully we'll see all of you next week in lovely Sandiago for the Edge AI conference. It's gonna be a blast. It's gonna be edgy, and there might be some AI, and maybe some open claws will be right by the ocean so crustaceans can get you. And you know, hopefully people bring all their Mac minis and show off their open claws there. That'll be exciting. Ooh, kick the waves. I love that.
SPEAKER_03That's really great.
SPEAKER_02Great discussion today. Good talks about the venture capital world, you know.
SPEAKER_04Oh, yeah.
SPEAKER_02Venture capitalists, they're the world's worst gamblers, as you know. They lose about what, 95% of the time, and then hope to make up for it 5% of the time. But what if we don't have any more unicorns? What happens then? You know, it's like lose, lose, lose, lose, lose Facebook, you know, lose, lose, lose, lose, lose. You know what I mean? And so you have you were you VCs depend on that, right? And so if they were 100% right, they'd all be in Vegas right now, you know, just cashing in big time. Um, so but turn it around.
SPEAKER_01Yeah, absolutely.
SPEAKER_02Absolutely.
SPEAKER_01Yeah.
SPEAKER_02Yeah, it can happen. Yeah, I'm sure we'll do some Man on the Street live interviews next week in Santiago.
SPEAKER_03Yeah. Hey, dude, you're part of the IoT coffee talk crew, dude. You have to do us good, man. We have we trust you, dude. Pull out the flute.
SPEAKER_02Oh yeah.
SPEAKER_00There you go. Some jazz flute.
SPEAKER_01Are you gonna are you gonna have the your Ron Burger the impersonator at the event next week?
SPEAKER_00No, actually, he's booked, it turns out.
SPEAKER_01Oh my god.
SPEAKER_00I asked, but he's not gonna.
SPEAKER_01I wonder if there is being a Ron Burger.
SPEAKER_00But you know, Rob, feel free to show up in costume, I think you.
SPEAKER_01Yeah, I was gonna say. Crack that little suit out of the out of the closet.
SPEAKER_02There you go.
SPEAKER_03Oh my god.
SPEAKER_02Yeah. All right, guys. But all right. See you soon.
