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
#313 AI Adoption
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Hello! Welcome to Lincoln Team Talk, and that was Terrible Karaoke. Um you might not know the song. It's Mean Street.
SPEAKER_02Yeah.
SPEAKER_01And uh oh my gosh, if you didn't show up, man, we would have nobody. What's up?
SPEAKER_03Yeah, well maybe I don't know, maybe I scared everybody, but yeah, I've been a bit of a my been a bit of a my. So uh as I would tell you in the in the prep call, I was pretty busy all Friday mornings for a couple of months, but yeah. I've I've marked my calendar slot locked, non-negotiable.
SPEAKER_01Yeah, well, um that's what you have to do. You have to uh block out the time to jump onto IoT coffee topics exactly, exactly.
SPEAKER_03So, what have you guys been discussing in the past weeks? I haven't been able to catch up, actually.
SPEAKER_01You know, what everybody's talking about all the time, which is like AI, AI, AI, AI, A, AI, AI, I, I, I, I, I, so uh, but, anyways, hey everyone, um, thanks for joining. Remember uh to like, share, and comment. Um, you know, we're doing this for fun, so if you take us seriously, you're doing absolutely the wrong thing. Number one rule in this house is have fun, and we're just gonna rant and rave. This is how we decompress, right? After a uh whole week of AI fatigue. I mean, um it's just it's tiring. But uh this is how we vent, and out of this venting, believe it or not, comes out some very important and um timely uh insights. You won't know that it's timely until you watch it six, seven, eight months from now. Actually, you know, it's funny because you know, this whole Multbook thing, you talked about this long time ago. It was like um almost seven or eight months before Multbook happened, right? I mean, you it you stated it as kind of a joke. Well, a potential ridiculous scenario that ridiculous people would do. And it happened, right? It it's crazy. So kudos to you, man. Yeah, but that's a great example, right?
SPEAKER_03It's it's a good example, and uh and don't don't the audience take it wrong. I think that there's I'm actually running one of these things, and but uh I'm doing it the proper way in your three different isolated boxes. But if you give the keys to your life to an artificial thing that can actually use it to get more things out of your life, what do you expect?
SPEAKER_01Right, right. And and and that's all we've been saying. We're not saying that AI is worthless or anything like that. Um the key lesson that we're trying to convey here, or learnings that we have uh that we've accumulated over the years that we're trying to share with everyone who gets overly excited about all this stuff is this stuff is typically really hard. The path to good is very difficult and very expensive, and the potential for exponential damage more than an exponential benefit is much higher, right? It's an asymmetrical thing. And we're trying to help, you know, uh technical communities, business communities, end users, even vendors understand this because if you don't align with reality and you invest in selling hype, I guarantee you will not sell anything. And we say this because many of us on IoT Coffee Talk, we have sales experience. We've had to go out and we had to sell things, we have experience delivering, solutioning, um, having to be accountable to customers, right? We're not just pontificating based on you know something we read in business school. This is like field experience that we've had in the past, or like in Bill's case, I mean, dude, you have to deliver all the time in the gap of the ugly, right? So that's where we're coming from with this stuff. And people who get that twisted, yeah, you know, you're gonna get hurt listening to us. It's gonna be super uncomfortable for you. But you know what? At the end of the day, what's gonna be even more uncomfortable is when you fail your investors, when you don't make your um sales quotas, right? And then you lose your credibility because you you're trying to sell some ridiculous nonsense to a customer who will remember at some point that you're ridiculous. You don't have discernment, you don't know how to see beyond the hype and what and meet that customer where they are, you know. And it's funny because nowadays you hear a lot of companies, well, we're gonna meet the customer where you are. No, you're not. If you're if you're good still trying to sell ridiculous hype-driven nonsense, you are not meeting the customer where they are because most customers aren't even ready for any of that aspirational stuff. So, anyways. Obviously, I had a really, really special AI nonsense week. So, yeah.
SPEAKER_02You've been on the wrench, man. In the trenches, agents all around you.
SPEAKER_01Well, you know, I I'm I'm trying, I'm working with a lot of clients to figure out uh how to help them you know wade through this stuff.
SPEAKER_03What is the biggest what is the biggest nonsense you heard this week?
SPEAKER_01So I posted about it. AI adoption. Why is that a problem for anyone? Whose problem is that? AI adoption?
SPEAKER_04We have a what what does it mean though? Exactly. I mean, that that's a that's a it's a broad term, man. I mean, AI adopt. Um, okay, so you use Chat GPT, is that adopting AI?
SPEAKER_01Um, you know, yeah, it's sort of it sort of is. But the thing is, is that if it adopting it for what? What purpose? If you're adopting it like what we we were talking about for the uh past few weeks, um uh Dimitri, is uh if you're using it as a drug, right? Um then there's nothing really great that comes out of it. I mean, you know, you falling in love with uh AI chatbot. I'm sorry, that I I mean, how is that healthy? Seriously, how's that gonna help you with your social life? I mean, it's just I don't know for me, man.
SPEAKER_04It's just part of it though purpose, purpose is I feel like I feel like part of it is is that we're looking at it from a practical perspective in terms of what problem are you trying to solve. All right, yeah, that's the first thing. And by doing so, you take the fun out of it. Yeah, yeah, yeah. I mean, I'm I'm positive that there are a ton of people that that listen to us, and you know, and those that are literally just having fun with it, not thinking about the repercussions, not thinking about infiltration, not thinking about any of those types of things, are literally just having fun and let's see what I can create havoc. Fuck around and find out. Yeah. But there we go again. We took the fun away. And I think that there's I think that there's a there's not even a fine line, there's a literal chasm between using AI for a for solving whether it be a business problem, a technical problem, or whatever it is, um, to the creatives that are like, what can I do with this that's going to one tell you know allow people to see how creative I am by leveraging these tools, um, and and things of that nature. So part of it, part of me, I'm I'm kind of you know, if I step back and I look at it objectively, then I'll I'll be like, Yeah, you guys are taking the fun out of it, man. It's supposed to be fun. But when you hear the business side of it and you hear what people are talking about and what they're saying, you gotta call bullshit.
SPEAKER_01Yeah, yeah, yeah, yeah. And and and you know, and that's the thing, is we we've had to we had to we've had to deliver, you know, whether it's from the sales perspective, whether it's from the solution delivery perspective or product development standpoint, we've we've had to commit to being and we've we've been accountable for being successful. Uh you know, um delivering value to the customer is how you match you measure success. It's not tokens. How many tokens are you consuming? That's total nonsense, right? It's a distraction. Um and and that's really where I'm coming from. But you know, um, I mean, we have a lot of people who watch our show, right? For one reason or the another. And I'd rather have people watching our show because they want to figure out where ground truth is rather than get um, you know, a diet of exponential nonsense. There's plenty of that stuff. I mean, the all the you know, um all these responses that you get from these these chat, you know, these AI tools is nonsense average crap, you know. Um I I I saw some folks putting out some content. Do you see one of my posts that kind of made fun of all these people making uh these diagrams and shit? You know how easy that is? It and how thoughtless there's no thinking, and then the output is like, oh my god, this is like all just marketing material, right? Well, I mean, but but okay, so again, it's fair.
SPEAKER_04Marketing is a creative exercise. Yeah, technical marketing combines the two, and you should get something out of that. I've always thought that um what what what are these um infographs and stuff like that? I've always thought they were garbage. Yeah, I've always thought that, no matter where it came from. I'm like, that's garbage. Yeah, you know, you're you're you got all the garbage on one page, and that's all it is, you know. Um you know, and when people want to do these taxonomies about, you know, who's here and who's there, I'm like, pay attention to that if you want. You know, I'm too busy with my head down trying to deliver on these things and trying to pick up the next customer.
SPEAKER_05Yeah.
SPEAKER_01I mean, or you know, or even just figure out who the the best partners are to work with, and you know, in your case, like you know, suppliers, which what what technologies to actually use?
SPEAKER_04Well, you know, I mean, it there there is absolutely something to be said about that, but that to me is a general part of business anyway, because if you believe that you and your solution are going to be the one throat to choke, and you're a startup, again, you're living in the valley of disillusionment and not facing reality. You're not, you're not going to be that one throat to choke.
SPEAKER_01Yeah, yeah.
SPEAKER_04Google isn't even the one throat to choke because they have partners. Yeah, you know, I mean, so it I I think I think it's it's it I think there's a lot of a lot of failures in general business practice across the board, irrespective of the tech that you're using.
SPEAKER_01Yeah, yeah. What do you think, Dimitri? What have you been up to, man? By the way. Yeah, well, you sound the most talkative.
SPEAKER_03Yeah, it's just uh I I need to rewarm up a little bit.
SPEAKER_00So you gotta give you a chance.
SPEAKER_03I mean, I think I I thought some of you are starting your position at the end of March, so it's the beginning of the establishing your position and those kind of things. Uh uh, but yeah, I mean the the the the point is you know I think that you know companies and it's also an overabused term, but we we truly have to think it as an ecosystem nowadays. You're not the one that's gonna do everything. You you play a part in the value chain. We're actually learning that at the at a global level in all our economies. Uh value chains right now, and how can you influence the value chain is the most important thing.
SPEAKER_01Yeah, that's true. That's that's very true.
SPEAKER_03Whether it is using AI or automation, all those kind of things. Yeah, it doesn't matter. The the problem, I believe, with AI that to add to your earlier point, uh Leonard, is uh is uh is the scale, because now it can it can have a much more important dramatic exponential impact. Uh that's been the problem. You know, before the first time you get a stupid analogy, the first time you get a you get a TRM, or you can call more customers and you more on their phone and more on the app. But you're still limited by humans. Now you can uh I basically receive a dozen pieces of crap email. So I'm stopping accepting by connecting it. Because you you it's harassment level now. And it's all AI generated. It's harassment level. And actually I've been saying that with some of some of you will recognize maybe you probably not, you're not our audience. But I'm saying, guys, you know, you're second in my connection. This is the tensie made this week. Who do you think you are? And I know you're not a person you're an AI agent, but stop.
SPEAKER_01Yeah, I mean, you know, uh now these people uh now I'm getting like even longer emails than before.
unknownYeah.
SPEAKER_01Where people were a little bit more to the point. I mean, that was like one of the things that I learned in consulting. It was a huge coaching point by one of my managers. I I used to write these really long ass emails, and she told me one day, took me in the room and and and told me, look, you know what? Try to be more concise because nobody reads the entire email. I went, really? And she goes, No, um, you write too much. And just get to the point and synthesize your thoughts. But now what's happening is you put intent in and this thing freaking goes bonkers and then creates a two-page email that no one reads. Um and you know, I lo like I was telling you, um, I think there's some companies now in the AR programs, when analysts ask questions, what they do is they have uh a chatbot in the background, their chatbot providing answers, right? I I think that's what's happening because a lot of the responses just look the same and they're really friggin' long. And I'm scratching my head wondering, how did you write all this stuff? And oh my god, I just asked a simple question, and you guys are giving me you know all kinds of irrelevant crap on top of it, right? So um there's much more noise, and that we're increasing the requirement for filters, right? That's that's nonsense.
SPEAKER_03Do you know what I'm saying? And and I think the the answer to that is that people are gonna start putting other bots to read those bots. By the way, on on email, I always been applying three good the three golden rules of email for me. The first one is no more than ten lines in the core. Now you can have references. The second one is spend 25% of the time crafting the subject line. This is the most important part. And the third rule is be very cautious on the T and on the BTC.
SPEAKER_00I would put that number one.
SPEAKER_03Yeah, well, no, it's a very well, that's the political aspect, that's why I put it at the end. But uh, but yeah, but that's the most important thing to do. But yeah, but yeah, yeah, yeah, yeah, yeah. And I usually have been saying that for decades. If you send me an email whose subject area is urgent, this is the last one I'm reading. Oh man.
SPEAKER_04Urgent why? I I guess, I guess I'm I mean, I don't know. I uh again, I'm one of those that's look, I I don't have a whole lot of time to be sitting there writing emails and shit. So I will email will just the here's what I want to talk to you about. Two lines, give me a give me a time, a date, and time, just jump on a call. Yeah, best BP. Done. I'm out. And and I mean, that's it. You know what I'm talking about, you know the context that I'm putting it in, uh how I'm framing it, and you know that we need to have a conversation. So, what else does there need to be said? And if things go have gone wrong, like I mean, we deliver and we deliver and we deliver. And in certain instances, like in you know, one of the programs that we're that we're working on right now, we should have started back in February. And and so my guys are sitting there going, Are we still doing this? I'm like, Yeah, we got the PO, but you know, the way everything is structured, I'm not invoicing until we start. We told them how long it's gonna take, and now because of their lack of execution, they're saying, Can you compress your time? I was like, No. No. In email, lessons learned. This date, this time, we're gonna talk. Yeah. Because, you know, we can't we can't continue to operate this way. So I mean, there's there's I think I think email TLDR is a real thing. Um, so I'm like, I'd rather get to the point and have you respond back and go, what are you talking about?
SPEAKER_01Yeah, yeah, yeah. Yeah, exactly. Just uh be straight up. And you know, the thing is, is the the less, I mean, I think this is the longer email, the the the less trust between parties. Is sort of like the correlation I see and yeah, you know what I'm saying? Um if I trust a party, I I just need to be all I need to do is be concise. I don't have to do a huge sales job, you know, adding a bunch of nonsense to sell a point. You know what I'm saying? If there's mutual trust, they just they they'll know what you're talking about. But the people that go around that are trying to build trust and this is the irony, they will send you a long ass email and um uh and and then the recipient will typically not trust that person for sending that long ass email. They'll just think you're a spammer. You know what I'm saying? And so um yeah, man, I don't know.
SPEAKER_03It's it it's just crazy. It's crazy. We we don't we live in societies where you know being being a man of few words is something that everybody has forgotten. Probably the Gene AI world, it makes sense because again, those things are so easy can so easily spit some stuff automatically. You give them a sentence, you get two points. It's it's crazy.
SPEAKER_01Well, you know, think about it. Isn't it crazy? Like you have like the uh the younger generation that have diminishing attention spans. Why are you going to use gen AI to create a longer uh email than you would be capable of writing or would even bother writing? Expecting that a younger generation person is actually gonna read the friggin' thing.
SPEAKER_04You you okay, so now it's it's so it's so fucking ironic that you're saying that. And you recall last week when I said, Hey, come here, let me tell you something. Don't fucking do it. I just get this email. Do you see the cops?
SPEAKER_03Oh my god. Yeah, yeah, yeah, yeah, yeah, yeah. I'm not surprised. Yeah. It's your name, yeah. Oh, yeah.
SPEAKER_04The cup has my name on one and his on the other. Oh, and by the way, I found True North Software Solutions on LinkedIn. I really like your product. I'm a marketer and product manager who has helped several people. I'd love to learn more about True North Software Solutions. Wait a minute. If you use us and you understand us, then you know what the fuck we do. Yeah. People stop doing this shit.
SPEAKER_00Yeah, yeah.
SPEAKER_03You know, you know it's not gonna happen. It's yeah, it's crazy.
SPEAKER_04By the way, this is I mean, I mean, but how how do you think that this is a good idea? I can help you secure more calls. How? I work with local city and state, yeah, yeah. I work with utilities. Are you gonna go and introduce me to CEOs of these utilities and and and CIOs of these cities? And and I mean, how are you going to do that for me? Because if you can, damn it. Let me let me bring you on.
SPEAKER_01Yeah, yeah, yeah. I mean, it dissuades people from investing in building trusted relationships, right? Um, you know, uh it's like what you're doing is just trying to build a pipeline, but when everyone's using these tools to create a pipeline, it generates all this junk. You you you know, you're not gonna have a huge conversion rate, you know, because the reason why you're using these tools is to improve your conversion rate, right? Yeah, yeah. And then if everyone else starts using it, guess what?
SPEAKER_04But you know, I think you will nothing changes. Can we? I mean, but here's the thing. It this is, and I mean, I've done this, I've gotten calls like this, and I'm I'm like, where is my fucking information out that people are calling me saying, hey, uh we'd like to, we like so so I get one, and I sometimes I'll entertain them just to rip them the fuck up. Yeah, okay. I'll entertain the call. Hey, this is Bill. What's up? Hey, you know, I saw that you got this kind of business, blah, blah, blah, blah, blah. And we were wondering, I mean, what were you using to track your what? Are you using a CRM? Yeah, we're using HubSpot. Oh, really? Well, we got this product that can do this, and I'm like, really? We built our own. Well, well, how does it work? I mean, it takes everything from the moment I'm having a meeting, and I and I that meeting is recorded and it goes into HubSpot, it lines up everything, it spits out a uh time, it ages the deal, and then at the time we close the deal, it tells us how much cost it cost us to close that deal and subtracts that from the total margin. So I understand what my net profit is.
SPEAKER_00Yeah.
SPEAKER_04Oh, you do all that? You're talking about.
SPEAKER_01Yeah, I mean, and think about it, man. Um, the conversion rates on um social marketing spend, right? Uh is still very low. I think it's still in the low single digits. And then so when you hear these guys say, well, you know, we used AI to improve the conversion rate by 5%, which it's like, okay, prove it. Um because I I don't know if there anybody actually audits what these platforms do in the back end, right? All I can tell you is they it there is a huge cottage industry, global cottage industry, to inflate the uh impression numbers. I think it's very easily in fact, I know that it's easily provable. Uh you have click farms that are basically creating, you know, they're like minting coins. Do you know what I'm saying?
SPEAKER_03And um we did bot farms now.
SPEAKER_01Yeah, now we have bot farms that gin up like synthetic engagement, fake engagement. This isn't real stuff, but what are the social platforms doing with their ad engines, right? And working with advertisers to make the sure that that what what the advertiser is paying for is uh you know for legit stuff, right? Real impressions. And then what is an impression? What is engagement?
SPEAKER_04Uh how do those I mean, those are all people are people are it are are enamored by it by that, right? You know, clout chasing and all of that good stuff, you know. But here's here's something. I just want to throw this out. I mean, you guys tell me what you think because I'm I'm kind of looking at this and I'm like, we talk and and and we kind of we disagree with each other from time to time, you know, and but you know, everything is kind of like you you got your point, and you know, I don't know if it resonated or not, but we should open up the lines to calls.
SPEAKER_01Yeah, we should. We should. But uh we'll have to figure that out. But you know, it's the uh open invitation to people who I mean not necessarily joining, somebody might just want to call and be like y'all bullshit.
SPEAKER_04Click. I don't do that, yeah.
SPEAKER_00You know, hey, look, we're sports that we're not we're not Jerry Springer, man. Hey, but I'm gonna go. That would just turn to Jerry Springer.
SPEAKER_04I'm like, I I like that idea because you know, a lot of people don't want to get up first, they don't want to jump on and and and and be face to face with you. They kind of want to, you know, be behind the keyboard and let me just make sure that what I'm saying is gonna land. If it does land, I might join the call. You know, that kind of thing.
SPEAKER_03Yeah, the the the problem the problem is the filtering for that because you know we we could do a Twitch style where you get you know the audience ask questions, but it's a big it's a big feed on the on the left. But the problem is I've seen with all those Twitch things is well you get the noise, the noise. But but actually we have we have a more targeted audience, so it might work for us. Maybe we want to explore that. You know, we have hey, open forum, log in and you know, drop a question, and we'll select one or two or three and we'll we'll talk about. And if you have a a good point, maybe we'll invite you. We say, hey, uh XYZ asking this, you want to jump in and have a chat?
SPEAKER_01Uh I think we're I'd rather have a conference, dude. Just like a real in real time.
SPEAKER_04I mean I would I would love to do that, but I I mean as a stepping stone, that might be the first catalyst to it. Yeah, yeah. You know, I mean, I feel like I just feel like I want to I definitely want to hear from from people.
SPEAKER_01I mean, you know, uh I I can I mean people hear it think of it this as like a reaction video, you know how like you have these people that react to music videos and movies or whatever. You know, this is a reaction video. It was so we have a couple of things that happened in the week, and then we react to it and then get all ridiculous about it, and that's sort of what it is, you know.
SPEAKER_04But I would bet that there are people out there. Um I mean, I've I've gone I've gone into speaking engagements and whatnot, and people have been like, hey, uh you know, you're on IoT coffee. Yeah, I am. You know, man, you you you probably have the worst mouth on there. Yeah, I probably do. I probably do. Yeah, yeah, kinda, you do. And then I'll I'll tell them right then, you know, I've done it before. It's like it's no disrespect or anything. I'm like, look, I don't give a fuck. I will say what I want to say.
SPEAKER_05Yeah.
SPEAKER_01And that's that's it. That's what, but you know, that that's the thing. Going back to what I mentioned before, um, you know, at the end of the day, people won't be listening to people who just gin up hype. Seriously, they lose credibility over time. Um, this whole show is about ground truth. We we uh uh debate each other, we react to the nonsense that's out there, and then we try to help our audience uh figure out where they need to uh go to to get things that will actually work and help them in um value creation, right? I mean, think about what the IoT stuff. We had the same thing about adoption. Well, you know, they're not adopting IoT fast enough. Well, what the hell do you mean? And what is that? Wait, you so you're gonna you want to drive adoption and to satisfy some ridiculous forecast for 20, whatever, 2026 or 2025, that it's supposed to be 1.5 trillion dollars or some ridiculous number of economic benefit. What about uh how you know what about considering the fact that that number is nonsense, right? And think, you know, again, all the battle scars uh the the IoT community suffered. I mean, they're still suffering from it. I you know, when I talk to people in the community who watch the show who know me, uh they give me testimonials, man. None of them tell me, oh dude, you know, um this was like an astronomical opportunity. They go, man, that was brutal, right? I this AI stuff is gonna be brutal because most people out there talking about stuff haven't an effing clue what they're talking about, you know? And that's that's scary, especially with this kind of technology that can have huge, huge impact. It it it in an in that regard, yes, it is a transformative technology, but with the degree of misuse and misguided use, that is what's scary.
SPEAKER_03Yeah, that's that's what I was mentioning earlier, because you know the reality is our society, our structure that you know investors are only responsible for the money they invest. They're not responsible to understand the thing and and be cautious about it. Now, in the different ways we had in the past, whether it was the internet bubble or the the whatever mortgage crisis or whatever you know the IoT thing, uh the technology had some impact, but it was not as transformational as this whole AI thing is, because I keep on saying that over and over again in the show. I think what people understand is that it is the first time that we have now technology that can use something profoundly human, which is language, at a scale. That is crazy. And this is going to be transformational in society and can be extremely positive, for example, for education, or extremely dangerous in many other ways.
SPEAKER_04Yeah. And uh so wait a minute, let me ask you a quick question. Was was was uh the the this latest hack, right, with Canvas? Is that AI?
SPEAKER_03Uh I don't know, but the thing is right now the the the the thing that you the the the the the the the the the problem which you the thing is the i I have some experience with the hacker communities, you guys know that. What is interesting is that the in the early days of the hacker community it was actually a very select group of people and it was knowledge based. Uh so you would go to a BBS button service or to a forum and you would not ask, hey, how do I hack this? The first question you would get is what do you know? And say, Oh, what have you done? And if there are nothing, okay, go find something, and once you've done something, come back. Okay. So that was the early days. Then came the internet and the forums, and people started to share knowledge. And then hacking become hacking as a service, so you can actually buy some hacking tools, some something. The problem now with those DLLMs is that the the the the the entry level for hackers is the same level of the defenders. Very low. And that's dramatic.
SPEAKER_01And it depends on what you mean by AI. Uh, it might not have been entirely AI, because you know, things like um mythos. What the the reason why it's a powerful tool, it's not exclusive to mythos, is its ability to uh to enable um you know discovery. Uh it gives you visibility to stuff that wasn't visible before. And and that's what hacker hackers are gonna be using. You know, the point is that the point is that does that mean that it's AI, it's AI enabled? Uh, but then we're gonna go to the next level where now these guys are gonna be using agentich. They're already doing it, using agentic uh to uh basically uh you know up their game in phishing and everything else.
SPEAKER_03And I give you uh I give you another interesting example observation in the hacker community. Are you guys familiar with what is a CTF? Capture the flag? These are those competitions that you and part of the occasionally. The problem now in the community, in the hacker community, is that you go to the CTF, and because usually you will go to a CTF, you have the teams, and you will see the the points over time will be a ramp. Now you see the top 15 ones, they all have the same number of points. Because they're all using AI, and then they and they're gonna have so the CTF scene is being destroyed by LLMs. Totally destroyed.
SPEAKER_04Well, and I mean I mean, but there's there's also one point that we that we didn't uh address, and that is the majority, I would say better than 80% of the attacks are no different than they were 10-15 years ago.
SPEAKER_02No, right?
SPEAKER_04The foundation of them is still the same. So you have social engineering three things, yeah.
SPEAKER_03Social engineering, buffer overflow, previous escalation. Yeah, three things that's that's hacking, period.
SPEAKER_04You know, I mean, but that's that's the you know I mean, th those things now you can just kind of do it faster. And every anybody can do it.
SPEAKER_03Yeah, and anybody can do it. That's the problem. The bar is low, the buy is just the cost, and actually, that is in the CTF communities now. It's how much money do you have to spend on token. We're back to your question earlier, Leonard. Is how much money you can the one with the most money in tokens will win.
SPEAKER_04Tokenomics is a business.
SPEAKER_01Yeah, yeah, but you know, the funny thing is uh you you've you've created it it's a complete waste if you think about it. Because um you have to spend tokens to defend yourself, and then at the same time the attackers are using tokens to take you out, and so same with military, my friend. Yeah, it it literally is a war going on. Um and the sad thing is the cyber criminal economy will grow faster than the regular economy. And the cyber uh criminal economy doesn't produce anything, it just steals.
SPEAKER_03Well, and and it's actually it's it's it's actually what I'm saying is that it's different than that. It's not the cyber criminal, it's the criminal who now have access to cyber because it's I know because now they can vibe it.
SPEAKER_00They can vibe criminality.
SPEAKER_03Yeah, no, it's true. That's the problem. Again, that's the idea. The the barrier to hacking has been lower and lower and lower. So anybody can do it. Now it's a tool for criminals. You know, before you had like half of the hackers were just doing it for the fun, and the other for the money. And in the beginning, it was actually the opposite. It was one person doing it, doing it for the money. 99% of us were doing it for the fun. Now it's probably the opposite. 99% is just pure robbery, it's just you know, professional criminals now can use an LLM instead of a gun. Yeah, and I stay in my basement with a computer, and maybe I have these kids that know a little bit more like me, but hey, I have access to it before. I had to learn.
SPEAKER_04But the other end of it is when people in this, I've never understood this. You know, oh, our security is XYZ. Our security, we are so secure, and it's like you are just dishing out an invitation. Exactly. Come and prove me wrong because I just told you that we use layers of security across these areas. It's like, oh, really? A layered attack. All right, here you come.
SPEAKER_01Thanks for laying out the blueprint for how we take you out. But you see, that's the thing. Eventually, um uh you know, security through obscurity, how can you hide? Uh, because these things will be scanning and and can micro-target things, right? And you know, so everything that's happening in advertising, right, uh, is being flipped and used in cyber secure uh cyber um criminality. You know, those uh those they're literally the same uh tool sets and espionage, right?
SPEAKER_03So now on the on the defense side, they're there are actually area of research and I I had conversation with a few friends recently. You can actually one of the good things is because of the ability to generate code so fast, you can actually go back to pretty much imagine like generating a very specific new uh encryption algorithm just for you. Before that, you couldn't do that. You have to rely on standards, and standards would be beat up and people would find the problem. But see, but now it's easier to generate your it's G it's easier to generate on a regular basis your own obscurity in defense.
SPEAKER_01Yeah, yeah. But this is the thing that I think is really I don't know. I get frustrated hearing people just talk about code. Well, you can code this and code that. No, at the end of the day, you need to build an application that works. Yeah, but just because you generate code doesn't mean that that equates to the custom application or uh bespoke capability. And that that's no, and you know, I'm not uh I'm just reacting uh to stuff that I've heard all week because you know a lot of these guys are pumping out code generator, agentic code generation tools or SDLC automation tools, and but the thing is, is there's always a person in the middle, you need to know what the hell you're doing, yet you have all these people conflating vibe coding with AI or agentic enabled SDLC or software development. They're two different discussions, right? And and a vibe coding is just only one thing that might be for one step or one phase of your SDLC. And lay people, people who don't code, who are not um you know software engineers, don't know this stuff, right?
SPEAKER_03Yeah, I mean it's the same, it's the same thing we had before the beginning of the show, Lenard, when we're talking about you know prompt crafting. If you think that two lines are gonna get you great code, no. No. But no, but what what I was mentioning is that if you have if you spend the right amount of energy and focus in actually specifying in natural language something very sound, then the code generation automation can be achieved much more easily than before. That's the part yeah, it is not again. I think I said to all people, oh, here I I look at a result of a work. What I asked them is what was the length of your prompt? Yeah. And if the prompt was two sentences, I know it's shit work.
SPEAKER_04I mean yeah, and then they say, you know, it it's it's wild, it's wild to me. Like, like I said, there's two there's there's you know, the whole vibe coding thing, my team, fine. You want to do that on the side, I don't care. If I ever see you commit that code, we got a problem. Um flat out. I'm I'm just there there is not a there is not there is not just a human in the loop, there is a human in the lead, there is a human intervention, there's a human at the end, we still qA our shit. Yeah, yeah, of course. Yeah, yeah. And you know, we've had co code generations Claude writes better code than any, you know, well, you know what? I ooh has so many things I want to say about that shit.
SPEAKER_01I mean, um it does the average, it does like I I mean, but the thing is, um you know, it it the it we have to remember it's not the AI, it's actual the the harness, right? The the harness that actually guides a series of you know like quote unquote agents, uh call it disaggregated intelligence to uh that work together in a chain to to come up with the outcome. But yeah, you're right. I mean you know the thing is the there we're are we're already seeing this. So if people are saying, well, you're full of shit, my experience is that no, you this is what's happening uh to uh what we're seeing, or this is what we're seeing with some of the big guys that have been talking agentic for more than a year, pumping out a bunch of frameworks and tools, and finding out you still need to have the person in the middle. When three years ago they said this shit is magic. You know, they go all in. Guess what? Now they're like stepping back and saying, Well, you have to use all this code now to make sure that your AI and your agents don't go bonkers on you. Well, I mean, look the same people three years ago who went overboard. I mean, you know, in my mind, they don't have credibility because you FA'd and FO'd, and you should have known better from the get go.
SPEAKER_04Well, definitely. No credibility, man. No credibility. I mean, it's it's going to happen. It's there there are there are I had a uh a lengthy debate with uh someone from another company that was looking at the twins that we build, and they said, Oh, we do that. I was like, No, you don't. I mean, it I'm sorry, you don't. Because, oh, well, you know, uh, can you ask your twin questions? I was like, you don't talk to a production twin that's on the operational side working on critical infrastructure, it's busy doing that. Yeah. Well, you should be able to know. Well, if you just sent the question and the data sits in the graph database over here, then you can query that out. So what you want to do is create a twin of the twin and unidirectionally feed that insight data over there to that one. Sure, you can do that. You're not talking to the production twin, you fucking idiot.
SPEAKER_03I hate it. Well, well, well, hang on a second. You know, I mean I I work for the Union of Digital Twins. They need breaks when you can talk to them. So you need to put them off the production lines.
SPEAKER_04It's modern it's modern slavery.
SPEAKER_03We have digital twins now.
SPEAKER_04Hey, you know what? These things, these things get fed by data, so you don't need a break, nor do you need a rest.
SPEAKER_03By the way, I was thinking on the funny, we are on the funny, on the funny strope strope now. Uh your your question about your falling in love with a digital twin or an AI agent. Then maybe there's a new work, there's a new job, which is uh uh a divorce attorney to uh divorce with your AI agent.
SPEAKER_00Oh my gosh. Oh, no, that's rehab talk. Yeah, yeah, drug rage AI rehab.
SPEAKER_03Yes, but yes, but it's much more sellable if it's called AI lawyer, AI marriage lawyer.
SPEAKER_04Well, I mean, if you if you do that, then somebody's gonna be talking about custody, custody of the data, custody.
SPEAKER_03Imagine how many tokens per hour can charge for that.
SPEAKER_01Yeah. Yeah, you know, you know, um, I think if you really think about it, the people who are gonna I I you know, they they um uh you have all these people that say, well, you know, um AI won't replace you, uh the person who knows how to use AI will replace you. I don't necessarily think that's true. Actually, the the person who knows how to who's amassed experience and um actually can operate um and deliver value without AI is gonna continue to outpay.
SPEAKER_03Yeah, I I would actually rephrase that in a different way. I would say the person who knows when not to use AI will replace you.
SPEAKER_01Yeah, exactly. Actually, that's that's actually that's also a very interesting point because you know uh I coined this uh concept called zero trust execution, and I've had a few people say, well, how's you know that's just uh zero trust uh you know security. And I I would I told them no, it's not. Because usually what you know when you look at agents and what they're now people are trying to do with open claw with Nemo Claw, and you know, you have like uh they introduced uh what's that open open script or something like that? Yeah. All these uh isolation mechanisms, sandboxing, blah blah blah, you know, zero access uh upon instanti instantiation. All this is is completely opposite thinking. Um and it's all about not not um you know you know uh not trusting something that you're actually uh um you know is there, right? And um that's the approach that you you know you need to take, even as a a user of AI, be aware of its limitations, otherwise you're gonna really screw yourself, right? But then you know don't be a Luddite, just know what it's good for and what safe use is, and then uh adopt it in a way that delivers value. But these people that went all in, I'm telling you, man, all of them have to backtrack, dude, because in the beginning it was all about we have to remember this. These guys were telling you to train your own foundation models. Is that happening? No, no one's freaking doing that, you know, and no one's doing that, and that was only like two years ago. Where are we now finding out? Oh, geez, wow, um, it's not secure. How do we do this uh in a safe way? And then uh it it's just been uh it's been uh a circus act, quite frankly, and that's why we're uh at this adoption discussion now. How do we drive you know, people aren't driving adoption fast, you know, adopting AI fast enough? Who gives a shit? If it's valuable, it'll get used, it'll be adopted.
SPEAKER_03I mean, you know who gives a shit, it's the is the billion of dollars of investment that you've got.
SPEAKER_01Well, too snapping bad. You know, you should have thought of this before you took your FOMO gamble. That that's the thing. They took a gamble and it looks like they're losing. Because if AI adoption is a problem, that's their problem. Yeah, not a not a not a business's problem, not a CEO's problem, even though CEOs are making it their problem. It's not the board's problem, even though they're making it their problem. You don't need to you don't need to adopt AI for the sake of adopting AI satisfying any a vendor or anyone else's or an investor's um AI agenda, right? Just like that was their gamble. You don't need to gamble on it.
SPEAKER_03So that's yeah, well, I'm not sure the boards and investors think that way, but yes, it should be.
SPEAKER_00No, they don't. They don't. They that's right.
SPEAKER_03And and we we and actually to some extent in tech we've we've lived in this in this kind of uh even though there were some bubbles that burst, but in general, if you look at the long term, the return on investment in tech in the past you know 50 years has been generally very positive.
SPEAKER_01Well, you know, think about that for uh for a moment though. Why is it why is it positive? And what is tech anyway?
SPEAKER_03What is it? Well, I mean everything related to digital, but I think that now we're reaching a point that the scale, again, because the scale of what it can do in impact is kind of also the scale of our the amount of money we need to because it's a very inefficient way to produce intelligence. The amount of electricity it consumed, the amount of CPU, the amount of chips you need is gigantic. So it's I mean, even the building of the internet infrastructure. I haven't done the math, you probably know that much better than I. But you look at the how what was the cost of building the internet infrastructure? We've probably passed that with all this AI crap today.
SPEAKER_01Yeah. Well, um apparently what uh AI and all this infrastructure is good at at the end of the day is generating code, which is then software running on CPUs. Yeah, you see my post? I actually have a next curve graphic for that. But no, at the end of the day, the tech is about tools. Yeah, business tools. It doesn't actually create real economic value. Because remember back in the day when there was an argument about well, software, software industry, they don't really they create they don't really create a product, right? So you might deem software a product, right? But in terms of driving gross national uh product, right? Creating, you know, like stuff that people buy that are essential. It doesn't do that. Um business processes improve um it you know, it can be the basis for um services, but it it and it can improve business operations and uh all the things that it's used for, but uh it's a tool.
SPEAKER_03Yeah, but yeah, but the thing is, and we might have to close soon, but maybe that's a closing thought, is because we have this kind of at least at the moment illusion and maybe over time reality of thinking entities which are digital, you can probably derive money out of that directly, because you can talk to your I mean the same way that you you talk to your AI lover, this AI lover might ask you for money and you might want to give him gifts. I mean uh it's not easy to imagine that, and I'm sure that might be a fear again, because those things being language-based, uh I can see a future where you have these very small companies that create armies of digital agents that become friends with people and start extracting money out of them, or or influencing them to buy some goods where they make a profit. I've told you already on the show this is the future of advertising for me. I think advertising is gonna totally disappear, uh, and it's just gonna be entities that talk to you and persuade you to buy stuff. Yeah. The big boards will go away. They will be replaced by somebody popping up in your corner and saying, Hey, how are you today, lovely? Oh hey, I got these new things yesterday. You should buy one too.
SPEAKER_01No, but yeah, yeah. Anyway. No, I I it's it's um. No, Proxy. There's a company called Proxy uh that I I uh wrote about almost seven years ago. Um and yes, hyper hyperpersonalization and basically um mapping your your psyche. Exactly. Um yeah, it it's and you know, metaverse was sort of uh a staging ground for like what we're I mean Cambridge Cambridge Analytica did was doing those kind of things as well in terms of influencing behavior, that's the same logic.
SPEAKER_03And again, now we have the scale because of the compute power we have. So yeah, yeah.
SPEAKER_01Anyway, yeah, all right. So hey, thanks everyone for listening. Yeah, thanks for listening to our weekend um or end of the week rant. Um, remember, we're all about communities, so check out Elevate Our Communities at www.elevateurcommunities.org and also follow IoT Coffee Talk at www.iotcoffee talk.com. We are not quite yet the Jerry Springer of tech, but what we are definitely a unfiltered, uh, as honest as possible form for having discussions and debates on um, you know, wide range of tech related topics. But most importantly for us, it's an opportunity to decompress. It is a tough week. But thank you so much, and we'll see you next week.
