Expert in the Loop: AI File Recovery, $1.50 Personal Agents

[00:00] Richard Webbe: Hello, [00:46] Dave Pengelley: hello. Busted, Richard. [00:50] Richard Webbe: Sneaking a quick sip before we start. [00:52] Dave Pengelley: How dare you have a drink on a podcast. [00:54] Richard Webbe: Very rude of me, sorry. [00:55] Dave Pengelley: As long as it's, uh, alcohol and not just water. [00:59] Richard Webbe: [00:01:00] No. Yes. [01:00] Dave Pengelley: That's all that matters. [01:01] Matt Slager: She's [01:02] Richard Webbe: a full [01:02] Matt Slager: strength coke, that one. [01:04] Richard Webbe: It's a coke zero and it is my most outrageous drink because I don't really drink alcohol or coffee or tea. [01:11] Matt Slager: That's pretty outrageous. [01:13] Dave Pengelley: Coffee number two for the day. So, um, I'll drink it. Um, but it's midday here, Mano. It's not like you where it's like 9am. So two coffees in this four days, not, not horrendous. [01:27] Richard Webbe: Text I sent you this morning, Mano, was probably at 3am. [01:31] Marno Brits: That's not a problem. That's normally, um, I try to wake up earlier, but I went a bit, went to bed a bit late last night. [01:36] Marno Brits: We've got the sauna sessions on Tuesdays. Which resulted in me not getting home till about 10 o because it was, we were trying to do a smoker, like a last minute, Hey boys, everyone bring a piece of meat, so we had like some fish that was just caught, we had some beautiful chicken wings, and then we spent about an hour getting the smoker ready before we go into the sauna. [02:00] Marno Brits: So the whole thing, this went till 10pm, just [02:00] Marno Brits: eating sauna, cold lunch, eating again. I would not recommend, um, but I'm glad I experienced it. [02:05] Dave Pengelley: If you don't [02:06] Marno Brits: just take me [02:06] Dave Pengelley: into the sauna with you and just steam it in the sauna. [02:09] Marno Brits: It's probably cool, but it's too hot like in New York, yeah, it's too hot. [02:13] Matt Slager: Why, why not recommend? [02:14] Matt Slager: That sounds great. [02:15] Marno Brits: Um, it's great and it was the atmosphere is nice until you get into the sauna and your body is allocating all its energy to digesting the food and you're sitting there and you're like my head is getting a massive rush it's all getting spinny it's 110 degrees i can't see where i'm going Like you start, your body starts fighting against you because it's allocating the energy to your stomach, not to you surviving this heat. [02:38] Matt Slager: Wow. Sounds like a smart lunch. [02:42] Marno Brits: Your favorite place now. [02:45] Matt Slager: That [02:45] Marno Brits: was a good time. I'm glad I did it. Like probably next time we normally eat at the end of the sauna sessions, like we do three sauna sessions and then at the end we eat. But there was so much food we just had to stagger it. Just need to [02:58] Dave Pengelley: sleep. [02:59] Dave Pengelley: Sleep [02:59] Marno Brits: [00:03:00] faster. Good sleep. I had a brilliant sleep. So let's check the sleep. [03:04] Richard Webbe: Manu, just to remind us quickly as we're talking about weird cultural experiences. [03:07] Marno Brits: Um, [03:07] Richard Webbe: where's your background from? Where did you learn to do sauna and eat smoked meat? [03:11] Marno Brits: Oh, um, learned that in Australia, but my background, I'm from South Africa, but I moved here when I was 10. [03:17] Marno Brits: This is just... The, the health kick that I've been on for the past three years and hopefully for the future. Very big on just longevity and endurance. So anything that can help me perform a bit better or feel better throughout the day. And your cold plunges and saunas have been extraordinary in that. [03:35] Dave Pengelley: I [03:36] Richard Webbe: eat a lot of meat and I surf, so I suppose it's the same thing. [03:39] Marno Brits: Yeah. [03:40] Dave Pengelley: What kind of smoker? [03:42] Marno Brits: Oh, it was literally just like a standard little Weber. Two tiers, so at the bottom we had... [03:47] Dave Pengelley: Like a bullet smoker? [03:49] Marno Brits: Yeah, there we go. Um, that was good. It was fun. [03:53] Dave Pengelley: I was hoping I could lean into AI and IoT with connected grills and... Some of the cod grills were there, the keg drafts and [00:04:00] stuff. [04:00] Dave Pengelley: No, it was all charcoal. Uh, my first ever smoker was a little bullet smoker. It was like a $45 one from BCF that my wife bought me. Because I said, I'd like to get into smoking. You know, as you hit 40 as a, as a man, you either get into smoking meats, trains, or like, um, world war history stuff. And I chose smoking meats. [04:19] Dave Pengelley: And, uh, it was like, oh, we don't want to over -invest in this smoking thing, just get a, just get a cheap one. It was so cheap, it was almost impossible to use, it leaked smoke, wouldn't hold its heat, it was a nightmare. We got this tiny little mini, like, one kilo piece of brisket, but it, we, we managed to cook it enough, we were like, this is, this is pretty good, we should spend some more money on this, so. [04:38] Richard Webbe: Let me ask the million dollar question, or the billion dollar, how can AI help us cook better meat? [04:44] Dave Pengelley: Well, on the connect the grills thing, I don't know about, about, about, uh, necessarily AI. I mean, these are good for recipes. We've all gone chat to your PT and said, I've got this in the fridge. What do I make with it? [04:53] Dave Pengelley: Uh, and it's come through with the goods, right? I'm not the only one. Yeah. [04:58] Matt Slager: Similar concept, like Richard, [00:05:00] just, you know, if you've never done it before, how do I do it? What do I need? You know, that kind of stuff just get you kicked off. But yeah, where, where Dave's at now, well, I know even, I know, I don't know how like AI could come into it to help them. [05:12] Dave Pengelley: For me, it would be research. Like, so my, my grill that I've got, I've got like this, um, $500 acorn grill from Bunnings cause it was. What I needed it was a good prize point. I've been great, I had it for like five or six years now. But it is a steel built 1, that's why it was so cheap and it started to rust and because I think to the future and I'm looking the paid girls because they are connected, they got apps that are easier to fire up, use, turn them on, like I've done the charcoal saying I've done the sleepless night up and down adjusting my events to do all that sort of heavy listing I've had that adventure. [06:00] Dave Pengelley: I'm getting older, I just want to smoke me quickly and easily and tastefully. So I'd be looking at pellet grills and so you go, okay, I will get AI to research across different pellet grills in different price ranges and like water layers, as far as, you know, emptying the hopper [06:00] Dave Pengelley: and maintenance and any potential issues. [06:02] Dave Pengelley: So for me, it becomes a research project. [06:04] Matt Slager: Yeah, that's true. [06:05] Dave Pengelley: Necessarily using the thing itself, um, because I'd probably rather go and watch Aaron Franklin on YouTube and trust his advice on how to cook a brisket more than board codes. [06:14] Marno Brits: But. [06:16] Dave Pengelley: As far as finding which grill to buy, that's where I would do the AI. [06:19] Marno Brits: Have you seen the blog that used AI, this is probably last year now, um, when voice agents started getting a bit more traction? He built a voice agent, some custom prompting, and it would call every single Rolex. dealer to see where you can buy the best priced Rolex and it called I think like 100 or 200 shops and then found the best ones and collated them into a list and then he went and bought the cheapest Rolex but I'm like you could use that to just arbitrage it and then see okay cool where can I get the best grill and then resell it or make a grill gumtree type of setup yeah yeah [06:53] Dave Pengelley: that's so on LinkedIn today someone actually um published a Guinness index where they set up voice [00:07:00] robots and ring 700 pubs to get the price of a pint of Guinness And apparently there used to be a thing that there used to be a published Guinness, like a pint index in Ireland. [07:09] Dave Pengelley: Uh, and this, this guy resurrected it by getting a robot to automatically call all the pubs and find pints anywhere, like, you know, three year or up to 10, uh, listed out there. So, um, and then you get a, people go, Oh, yeah. I don't think that would be a contravention of the GDPR or the EU rules for declaring that you're an AI robot to do that. [07:27] Dave Pengelley: And I'm like, oh, Tim, all this red tape always gets in the way, right? [07:32] Richard Webbe: I always have to think about, because what we're doing on AI, in my opinion, is just grabbing, when we grab mass data off, public data. We're just grabbing the sum net average of everyone's opinion, collating that and then deciding. So, you know, there can be corrupted data in there. [08:00] Richard Webbe: And I always think about, you know, the early crashes of the share prices in the old days when they ran those sun systems that automatically told people when to buy and sell. I'm just wondering how that affects our cooking because [08:00] Richard Webbe: tastes are different. Power costs are different. It'll be very interesting. [08:05] Richard Webbe: It'll be more nuanced. [08:07] Dave Pengelley: I heard, I heard a term this week, so I, I, um, we, we should, we should go around and talk about what we've learned this week. Uh, specifically on the topic of AI, not just, just cold plungers and, and grilling meats. But, uh, um, I went to a Claude Code meetup on Monday night, and one of the speakers there, Mark Monfort, who's been involved in, you know, um, decentralized finance and other bits and pieces, he gave a little talk around what he's been building with, uh, So it's like Claude Code. [08:33] Dave Pengelley: And he talked about this term expert in the loop versus human in the loop. So human in the loop, we've most people have heard of. If you are joining us for the first time, you haven't heard that. That's the concept that you give the AI agent or the chat prompt permission to go do a bunch of stuff. But when it gets to critical points, it comes back and brings a human into the decision making loop in order to confirm whether you go forward or not. [09:00] Dave Pengelley: But he talked about it being an expert in the loop. Because having a level of domain expertise or subject knowledge [09:00] Dave Pengelley: will affect the questions you ask. The agents, what you ask them to do, how you ask them to operate, and then bringing your expertise into that loop gives you much better results. And so, you know, Richard, you're talking about your personal taste and what you like to cook with and all that kind of stuff. [09:15] Dave Pengelley: That's your personal expertise. If someone like Nigella Lawson goes and talks to Claude about recipes, it's going to be a very different conversation than someone like me or Heston Blumenthal or whoever, right? I [09:28] Richard Webbe: have one of those phone -connected sous -vides. Are you guys familiar with a sous -vide? [09:32] Dave Pengelley: Yeah, the little water boilie thing and you put the shrink wrapped meat in [09:37] Dave Pengelley: it. [09:37] Richard Webbe: I don't know if you're going in the keto protein way, sous -vide's a must have. So most restaurants use them on a large scale and they pre -cook the meat vacuum sealed in a plastic bag before it's, uh, you know, um, cooked. Braised or caramelized, right? And so it almost gets the meat to the perfect temperature. [10:00] Richard Webbe: And once it's at the right temperature, you may leave it there for half an hour or 20 [10:00] Richard Webbe: hours, depending on your choice. And then you just drop it on the grill and sear it either side and the steak's done. And that's how restaurants get you a perfectly cooked steak. to your table a lot quicker than when we're at home and we toss it on. [10:11] Richard Webbe: So sous -vide, superheated water, and you can buy these great ones that just sit on the edge of a pot. Yeah, we've got one of those from Malta. Yeah, yeah, the sous -vide, but they're connected to your phone, so you could be at work and you could start it up. Cooking before you get home. So it's at the right temperature. [10:28] Richard Webbe: And so I was considering an agent approach to using that along with my electronically measured barbecue could be a very accurately cooked piece of meat. [10:41] Dave Pengelley: I'm sure people are lining up their open claws and getting them doing all that kind of. Connected, right. Um, anyway, I've, I've, I've had a few wins this week that I'm, I'm keen to talk about where, where, where's everyone else's, uh, week's been, uh, in the AI sphere, gone to any interesting events, done any interesting things, had any, any good wins.[00:11:00] [11:01] Matt Slager: Yeah, I've had a fair few. So without going into like hectic detail, basically Um, I've had a couple of really good fresh client chats, really good communications with people, you know, even to the point where like they may not necessarily go anywhere, but just having the conversations. [11:21] Richard Webbe: And [11:21] Matt Slager: I think that is something that you're going to see more and more of, um, because at the beginning it was always like almost taboo. [11:30] Matt Slager: You know, oh yeah, I do that. I do that AI thing and um, and if people are like, oh yeah, I don't really, I don't really get around that too much. Um, but yeah, now like I literally was speaking to business owners over the weekend. Um, I think I told you guys I went on this, this huge walk. It was for men's mental health and. [11:47] Matt Slager: It was the longest walk I've ever done. [11:50] Richard Webbe: How far? How far, man? [11:52] Matt Slager: It's 25km, so I know it doesn't sound like a lot. No, that's a [11:55] Dave Pengelley: big walk. That's like 5 -6 hours. It's a big walk. [11:58] Matt Slager: Yeah, okay, cool. [00:12:00] Um, but yeah, it was pretty intense. And, um, but just during it... Just being able to mingle amongst these other, these other gentlemen, um, and talk about stuff and like what do you do and like, you know, when it comes to that time where I had to explain it, you know, I'm not sure how you guys answer that question now, but it's still really hard, but I went down the path trying to explain it and the guys were like, whoa, that's really interesting. [12:22] Matt Slager: We need to get you in and talk to our team about this. Yeah, I had that chat like a good few times and I don't know if anything's going to come out of it. I'm probably too busy to take anything on right now anyway, but just having the conversations was my biggest win. [12:38] Richard Webbe: That's a great one. [12:39] Dave Pengelley: My, my, my, my experience with some of those chats was they're like, yeah, they're really keen when they're face to face with you. [12:43] Dave Pengelley: And then as soon as you try to follow them up, you just get crickets because it's not actually a big enough priority for them and that they are interested. Not that they're not interested. They weren't, they weren't spinning you a line, but when they get back to the, the coal face, It's like, ah, yeah, we'll, we'll do that AI thing later. [12:58] Marno Brits: Yeah. [12:59] Richard Webbe: Maino, what about [00:13:00] you mate? [13:01] Marno Brits: I feel like I've had a similar response, but more the opposite when I was in town. Anyone that's international or national, if I don't get to see them face to face, then that's like, Hey, this sounds awesome. I'll get back to you. And a bloke just responded, what, maybe a week ago. [13:14] Marno Brits: And it's been like five, six months since I last spoke to him. It's just, it's not a priority because he doesn't see the immediate impact. Like businesses that are in Kalgoorlie, I'll email them maybe twice. And then I'll go to the office, just catch up, just be in their face, see how it's going, provide value there. [13:32] Marno Brits: Then like, oh wait, there's an out. I don't have to be in this turmoil of constantly just chasing my tail. It's been a month now. Let's get this done. Um, and because I'm physically there, I can just show up, do a little changes or show them Microsoft shortcuts to cut about two hours off their day. Like it's so much easier when I'm getting to go see them. [14:00] Marno Brits: My talk like the last week has been awesome. Um, too many opportunities coming up. So now it's deciding what do I [14:00] Marno Brits: say no to? I've really built out my anti -gravity bot and now Matt's knowledge core in there as well. So I've essentially built this personal assistant, this PA for me. That checks up every morning. [14:11] Marno Brits: These are the emails that you've missed overnight. These are the calendar events that you have to be aware of Um, these are cold leads that you haven't followed up with and then also what are my Action plans for today in order for me to achieve my go -to -market strategy in order for me to achieve my my quarterly financial goal I need to follow this 12 -week step Um, so keeping me accountable on that and then at night time, again, checking in, did I do the things that it asked me to do? [14:37] Marno Brits: And if not, what are we going to do tomorrow to adjust the plan? Um, so that's been tremendous and it's so cheap, like I had one day, I ran it the whole day. I did 72 tasks that I handed for it to do. And it cost me $1 .50 for the whole day. [14:53] Richard Webbe: That's pretty good. [14:54] Marno Brits: Unreal. [14:55] Richard Webbe: Mate, that sounds like a wife without the nagging. [14:57] Marno Brits: Ha ha ha. Um, yeah. I [00:15:00] mean, it's pretty, pretty, um, blunt, because last night I was talking to her before going to bed in my review saying that I'm hoping to spend time today. doing my 5 and 25 year plan and then it just bluntly said no that seems like procrastination what you should be doing is focusing on the plan that you've already made that's focused for the next 12 weeks so don't even worry about that you're just trying to find movement like you're confusing movement of progress um so it's really picking up on me as it's learning more and more about me and how I operate That [15:30] Richard Webbe: is a great, actually all three of those are great stories. [15:32] Richard Webbe: I think for me, um, and David and I have been speaking about this offline a couple of times as well, and I know you and I are now going to have a chat. I remember as a person in go -to -market and sales, no matter what the new technology was, people get overwhelmed. Then they hear what it might do and they go, that's exciting. [16:00] Richard Webbe: But like the ostrich, and by the way, of all the ostriches that were ever surveyed, none of them ever dug a hole and stuck their head in the sand. That's a misnomer, but anyway, a lot of them take that metaphor really and they [16:00] Richard Webbe: put it down. And I remember I was trying to reach the head of an automotive company in Victoria when I was at IBM. [16:05] Richard Webbe: And the same thing happened. He was overwhelmed by the concept of what I was talking about. I think I was at IBM at the time. We were talking about image processing, which is the early start for AI and recognizing things and scanning pages and processing things quickly. And there used to be this great cartoon that floated around in marketing and sales circles. [16:25] Richard Webbe: And it had a picture of General Custer with the hordes of Indians coming over the hill. Of course, we know the story that General Custer died and didn't survive because he was a really bad military strategist. And there's a taint there. And then one of his sergeants is standing there with the hordes of Indians coming over the hill going, there's a salesman at the door. [17:00] Richard Webbe: I think you should talk to him. And General Custer says, no, look at all this crap I've got. Why would I talk to a salesman now? Because when you look to the right of the cartoon, there's a salesman there with a machine gun. And that's what I see AI in [17:00] Richard Webbe: metaphoric terms. Cartoon terms recently. So much so I was going to get AI to recreate that cartoon for me because I couldn't find a copy and start sending that with a machine gun with AI written on it. [17:09] Richard Webbe: Anyone who takes that idea, well done, go for it. And I actually did this with that particular person and sent them a note. And no one did, we were beyond faxes in those days, but I wrote, this is you, this is me. That's your competition. We need to talk. It's okay, right? Um, and sending that to a CEO of a large company, uh, as a young salesperson is not what's up. [18:00] Richard Webbe: But it got me the meeting. We eventually toured the world, world together on a, you know, one of those IBM Discovery trips. And I think that's where we're at, right? People really want, right, what you guys have got, right, I won't say what I've got, I'm not the technical nerdy, but what you, people really want, yeah, I really, Mano, I want you to send me that, there's a man with ADHD, I really want that, that agent that you just wrote, so please email it to me, and um, I just think that's the challenge, I'm, I'm reaching out to a lot of [18:00] Richard Webbe: people at very senior levels and talking to them and they're going, yeah, that's what I want, I spoke to a friend of mine who's a very successful marketing executive, and I think that's what's going to help me, I think that's what's going to help me, and I think that's what's going to help my team. [18:07] Richard Webbe: Um, and she's covering off the, uh, AI and, uh, and, uh, cloud and data center space for a very large company in Australia that is representative of a large vendor. And she told me that they, she went and did one of those females in tech chats, kind of like what you guys do all the time. And they came out with, um, The go -to -market message has to be more specific around what we do, so people go, It's not that scary. [18:33] Richard Webbe: And of all the things they worked through, this came up as number one, and no one adopted it. Because everyone was too scared. [18:40] Dave Pengelley: I don't know what stories you've been hearing, but I've never been to a female in tech conference, Richard. No, [18:46] Marno Brits: that's a pretty good initiative. [18:47] Dave Pengelley: Not the [18:48] Richard Webbe: right thing to say. I'll get into any gender politics today. [18:52] Marno Brits: Is it called She Codes? Was that the organization? I know they're quite large. [18:55] Richard Webbe: I don't know. I don't know. And, um, I won't out her in terms of, you know, uh, [00:19:00] male, female profiling. So a very successful executive. One I respect very much and could very well be listening right now. But I got it, right? That, that, and the reason it was a female in tech thing was she just went there and I wasn't. [19:12] Richard Webbe: So it's nothing to do with male or female. It's everything to do with. Everyone knows what the problem is. [19:17] Marno Brits: Yeah. [19:18] Richard Webbe: But no one's doing it. [19:19] Dave Pengelley: Yeah, a friend of mine published a report, um, from a HR firm and was saying like people, uh, they all know they should be doing it, but no people are doing it and the people that are doing it are not seeing and recognizing value from it because they're just doing random things with it, not understanding the value, whether that's, you know, creative outputs, research outputs, time -saving outputs, whatever it is, people are not Knowing how to use it and so productivity is actually going down as people using AI for the wrong things on boarding times are going up. [19:49] Dave Pengelley: Because now people are trying to work out what, what is even my job now? I don't even know. [19:53] Richard Webbe: And everyone thinks it's an LLM or something like that. I mean, I've just, as the end of my little anecdote, and I'll shut up, [00:20:00] is I, uh, I got introduced to Claude by you guys, of course, after playing with him. And everyone thinks it's all about LLMs or these command line, you know, uh, type, um, AI supports. [20:10] Richard Webbe: And, uh, David knows, and I've written it a few times. I wrote it once and it was stolen on a laptop. But writing my definitive book of how you screw up selling just. kind of a laugh, I think, make sure I tell everyone all the mistakes I made rather than how successful I am. And, uh, and, um, and I think that's a more valuable book. [20:26] Richard Webbe: And so I'm using it as my sub -editor and my, you know, punctuation and phrasing and formatting, right? Because when you're writing creatively, formatting is a real pain in the butt because you're trying to think, should that go there? Should it go there? And, uh, I was using Claude and I'm up to 20 chapters already. [20:42] Richard Webbe: Whereas I think the first chapter took me about two years, and then over the weekend I get another 17 chapters out. It's quite amazing, but life isn't about LLM. Because everyone's using them, so you're not ahead of anyone. You're just keeping up. [20:57] Dave Pengelley: Well, I mean, that comes back to what you were saying around the, the, [00:21:00] the, some net average is what everyone gets, unless they're putting their own expertise into it. [21:04] Dave Pengelley: Unless you're putting your own knowledge and experiences in. Uh, there is becoming a growing pushback against all the AI slop, and I think creatively humans are gonna re -engage in creative pursuit somewhat. Uh, take that, take that space back, because people don't want to see it. That's funny, like on the Instagram reels and stuff, you used to see the people getting randomly, like, um, savaged by animals. [21:25] Dave Pengelley: Like, not terminally, but you'd just see, like, goats headbutting people and stuff, and you'd think that's hilarious. But now it's like over the top and you're like, I don't, I don't, is that AI? I don't care. Like, there's no, no, no one actually got hurt. So I'm not interested in it. Like, it just, it, uh, I mean, that sounds really bad when you say it like that, but you know, that concept of like the, the whole lack of reality and not knowing you can trust stuff just takes out like, [21:49] Richard Webbe: That's very true. [22:00] Richard Webbe: That's very true. I remember, I don't know if I've said this before, but my ex -wife and I were at a, you know, a child was at a nearby, um, uh, private school [22:00] Richard Webbe: and, uh, half of the mums there had become influencers, really big influences, right? And it was kind of hilarious, right? And, uh, my ex -wife at the time was doing, uh, charity work for really important charities and asked a couple of these influences. [22:13] Richard Webbe: Some of them are on TV and you'd know them or I won't use their name. Thank you. Could you please mention this charity on, in your Instagram or your podcast or whatever you're doing? And they said, yeah, give me $40 ,000. And we went, what? But what I find funny is that Wimbledon girl appeared and, uh, I don't know if I've mentioned it on this podcast before, but someone created a fake, uh, Female, uh, rather gorgeous, I might add, blonde, perfect mini skirt, perfectly positioned and they placed her at scenes in Wimbledon and said, she kept saying, I'm enjoying, I think she got like 40 million followers and she wasn't a real person. [22:47] Dave Pengelley: Wow, yeah. [22:48] Richard Webbe: And was saying all the right things and had, Probably because of the sum net average of people who are interviewed of what sort of a girl would you like to see at Wimbledon? There it is, right? Instantly created using our [00:23:00] data and stats. And as you said, David, and as you guys are leading to now, all of a sudden, the value of someone who really does that goes through the floor. [23:07] Richard Webbe: And I'm expecting all those influences to disappear because no one will know if it's AI or real and no one will care. [23:14] Matt Slager: I don't, I'm going to, I've got an opinion. Um, I'm not going to talk about miniskirts, but, um, with, with the, the idea of the influences going away, I think it's the exact, exact same landscape and framework that a lot of these other. [24:00] Matt Slager: Um, people that we learn from in the AI, you know, automation and coding ish space where it's the, it's the reshuffling of the wealth. So the reshuffling of the market, you know, the reshuffling of the, the monopoly. That's probably the best way to put it. So you've got all these big influencers that, yeah, might start to get less traffic or be swallowed by the, the slop fluences. [24:00] Matt Slager: Um, but [24:01] Richard Webbe: it's [24:02] Matt Slager: also gonna, it's gonna, like what Dave said, it's going to cause people to seek out the real ones, you know, [24:10] Dave Pengelley: That's why we're here. That's why this is a live show. It's real and authentic. We can't fake the level of awesome authenticity we have here. And [24:18] Matt Slager: to bring all of this back to Smoke Meats. [24:21] Richard Webbe: I think, you know what, I think, Matt, I think what I was saying, Matt, I think you're 100 percent right. Remember having a webpage was all R and all about quality and I want to deal with them because they've got a webpage and I can touch them and reach out to them. We're so far past that now, aren't we? And AI is taking us into that very dynamic space. [24:38] Richard Webbe: So I think, um, it's going to be all about content. and all about reality and you guys are the cornerstone of that you're creating the real content and the real reality and we want to hear more about it [24:48] Dave Pengelley: as long as we don't go back to that whole writer's strike era where everything's reality tv like i actually like scripted storytelling can we have some of that please [24:56] Richard Webbe: you're right david it's just the color around it right it's [00:25:00] just the color around it [25:00] Dave Pengelley: yeah but but shows like this and i'm sure we'll hit our groove and one day I was talking to the boys the other day just around, you know, that whole hard work versus luck, uh, and you know, comes into which and things like the, uh, the, the YouTube algorithm are one of those things where you've got to do the work and you've got to put the videos out there, but then there's also that luck where one video, for whatever reason, you just get the title, right. [25:23] Dave Pengelley: It attracts a few clicks, you get the extra views and all of a sudden you'll have like. Videos with a hundred views, a hundred views, a hundred views, a hundred views, a hundred views, 10 ,000 views. And you, there's no rhyme or reason to it. It's the same kind of content you've been doing the whole time, but you just get that little stroke of luck where oneness takes off. [25:40] Dave Pengelley: And I [25:40] Richard Webbe: did a video during COVID just for fun because I was practicing video to market through LinkedIn and it was about being ghosted, which is probably the most common thing. That executives and people, everyone finds nowadays. And so I just did this thing on ghosting, which was, I didn't even plan it. I just put the camera up and my phone hung it on the wall and spoke to her and said, [00:26:00] well, don't feel bad. [26:00] Richard Webbe: We all get ghosted. 7 ,000 views in less than an hour. [26:04] Marno Brits: Wow. Yeah. [26:05] Richard Webbe: It just broke the bank. And I went, how'd that happen? [26:09] Matt Slager: What was that posting that on too? Is that where you had like a big following already too? [26:13] Richard Webbe: Look, it's part of, it was part of me fragmentally putting together my go -to -market strategy around how I screwed up selling and my book and all of that. [26:20] Richard Webbe: And I, I set up a video thing and I think it was, the content was okay, but I think people love the content cause I was including them. It's not just you. It's me as well. And everyone loves that. So the vulnerability way to get it. And I think the name, right, what about ghosting or I forget what I called it was all ghosting and it just click, click, click. [26:38] Richard Webbe: And away it went, went very, very well. [26:40] Matt Slager: And that was on your TikTok account? [26:43] Richard Webbe: It was on my LinkedIn account. And if you look up on LinkedIn and there's two very embarrassing videos there, but I'm vulnerable. Go and have a look at them. I'm badly done about I've screwed up selling. But the book's coming out soon, available at all good distributors, including Amazon, in a little short while.[00:27:00] [27:00] Marno Brits: Beautiful. I'm curious, was that your first video that you posted to LinkedIn? Because I've seen that as quite a big, um, good tick for the algorithm. When it's your first video posted, or first one in a very long time, then it's almost kinder to you and then it'll promote that more. [27:15] Richard Webbe: Yes, you're exactly right. [27:16] Richard Webbe: And it was just at the start of COVID when we were all locked in. So I had three key market things. People were sick of being locked in and I wanted to see or talk or listen to someone. Number one, it was probably the first video I ever posted. And, uh, number two, I'm a pretty funny guy. So, uh, I think it went well. [27:33] Richard Webbe: Beautiful. [27:34] Dave Pengelley: I think that's why they're saying Tiger King went so well, right? Because it was at the start of the lockdown, everyone was like, uh, That was a great meme where, like, someone was talking to Trump because he was still the president of pre -Biden at the start of COVID. And, uh, someone's like, yeah, the people are getting restless, and he's like, release that tiger. [27:49] Dave Pengelley: That [27:50] Richard Webbe: is so true. [27:53] Dave Pengelley: Uh, sometimes, and again, it's just that perfect storm of events. You didn't necessarily plan that, but you saw something, and it is that whole mix of [00:28:00] taking advantage of opportunities, doing things, not just waiting for things, but also that just spark of the universe. Just right time, right place, that happens. [28:09] Dave Pengelley: Uh, which we can't control all those external factors that govern success, but we can, we can just keep doing the [28:15] Richard Webbe: internet. Look, I was asked to go and do some talks in London a number of years ago, uh, and it was for British Telecom and Cisco. I don't mind saying that. And they said, Richard, would you come over? [28:26] Richard Webbe: And it's the same story. There was the new technology of the internet and online business flow models and how it was going to help people and no one knew. It's back in 97. So they said, could you come over and do some talks? We think that might spawn some more conversation, by the way, which by the way, guys, we are going to do a online talk and then our face to face one in the next few months, David and I just planning that. [29:00] Richard Webbe: And so I said, okay, and I'm sitting in front of the second in charge of BT's office that I was shuffled in there. I bought a three -piece pinstripe suit in London on the way because I didn't know what to wear. And I'm a bit [29:00] Richard Webbe: keen on jackets as you know, and so I turn up and this guy, funny enough, he was American, but running most of British telecom, he swore at me and said, Why am I talking to an Australian about solving this problem? [29:09] Richard Webbe: I said, I don't know. Anyway, he said, What do you want to do? I said, Well, basically people don't get the concept. They think it's here, here or here and your sales team, he goes, Okay, I'll let you loose on them for a few months and we'll see how it goes. It went so well that we increased sales, wait for it by 600 million pounds. [29:25] Richard Webbe: In the first six months. [29:28] Marno Brits: Beautiful. [29:29] Richard Webbe: I ran out, I ran a program across, I think it was 2000 sales reps in the first six months and I dumbed it down like we do here. [29:38] Dave Pengelley: Not, not that our audience are dumb. No, there's a lot of confusion. A lot of, a lot of people misunderstand not knowing what's going on. So I went to this Claude Code meet up on Monday night and, and, and, and, and, and, and, and, and, and, and, and, and, and, Um, it was focused on more of an engineering audience, like it was a coders meetup. [30:00] Dave Pengelley: That was the intention. They're doing lots of different things and they're the community ambassadors and... Uh, and Dom, the front of the channel, was organizing this one. [30:00] Dave Pengelley: And it was intended to be more technical audience. But one of the, one of the speakers got up there, um, and asked how many like engineers were in the room. [30:09] Dave Pengelley: And not even a quarter of the audience put their hands up. Now, I don't think that was a shy. I think it's like only a quarter of them consider themselves actually developers, engineers, coming from a coding background. More of them might be coding and using things other than tools to do it. But they don't identify as engineers. [30:23] Dave Pengelley: Uh, and so, some really interesting demos of some, some interesting ways people are using things, uh, and some tools that can, can help people. So, that was, that was good, but there were a few times where, uh, and you know, the, they were showing kind of like the open core type things, kind of what you're talking about, Mano, how I've got this agent and it goes and gets all this data and gives me a briefing and does this and does that. [31:00] Dave Pengelley: And I'm thinking in there. This whole audience, most of these people, they want that kind of thing, they hear these stories, but where do they even get started? How do these technologies become approachable for them? Uh, we've talked about OpenClaw on this channel, [31:00] Dave Pengelley: and Matt, you've gone deep in trying to lock down and secure OpenClaw and started experimenting with it. [31:04] Dave Pengelley: I'm keen to know if you've had any successful wins or if it's actually changed your life yet. [31:09] Matt Slager: I did it before Jensen did it. [31:11] Marno Brits: Yeah, actually. [31:12] Matt Slager: Yeah, when the emote clock came out, I was like, Wait a second. Is he been reading my notes? No, I can't be the only person. It was, yeah, I just wanted to have safe defaults. [31:22] Matt Slager: You know, like the, the whole idea of installing it and running it and how good and autonomous it is, sure. But then, you know, obviously that increases, introduces security concerns. So for those that don't want to read the, the chapters and chapters of security notes that do exist out there from the official team, I wanted to just make a secure default, you know, like [31:42] Dave Pengelley: you [31:43] Matt Slager: install it and the whole thing's locked down. [31:45] Matt Slager: And as you learn more and want to do stuff with it, you actually upgrade it. So [31:51] Dave Pengelley: you've, you've installed it. Are you running it? Are you getting value from your open core? Is it doing meaningful things? [31:56] Matt Slager: Yeah, yeah, yeah, yeah. I actually, it's kind of embarrassing that I've spent [00:32:00] more time on infrastructure and benchmarking and testing and, you know, just drilling it than I have actually to build my own sort of results sequence. [32:08] Matt Slager: But I'm doing something very similar to Mono right now, where I have my daily briefing and Richard, it's the, uh, it's the, the ADHD sort of. focusing where i've specifically said it i'm like i only need to get told one thing a day you know if you can tell me one thing and it's the thing i have to focus on tell me when to do it in my calendar so yeah it basically it checks my my knowledge core checks my active linear issues it looks at my calendar looks at my recent stuff that we've done together and is like right you haven't finished this yet do that do it in this time block and that's that's all i have it doing yeah we [32:43] Richard Webbe: either have hyper focus or no focus [32:45] Matt Slager: Well, yeah, because I got the friggin like blunt end of the stick. [32:48] Matt Slager: I've got the whole AUDHD, so it's, you know, the autism slide mixed in as well, which just makes it chaos. [32:54] Richard Webbe: My daughter and myself as well, mate. It is fun chaos. [32:58] Matt Slager: Yeah, it's either infinite [00:33:00] distraction or hyperfocus. [33:01] Richard Webbe: That's it, that's [33:02] Matt Slager: it. Yeah. [33:03] Dave Pengelley: So, so you're a pretty technical guy, Matt, and you, you did the hard work and you made a secure version of OpenCLR and you're getting results. [33:09] Dave Pengelley: Mano, you, you laid more technical, you, you didn't do much for a while. You were like, I'm too busy, I'm too busy, I'm too busy. You didn't go open core, right? You've built your own custom solution. [33:18] Marno Brits: Yeah, so I did, I did flawed code and antigravity. So I started with just antigravity for some projects as I was helping to automate the website development. [33:26] Marno Brits: And in that there was one day where I had a Tuesday open and I felt incredibly anxious that there was nothing my calendar because it just felt like unusual. So I was like making sure I don't miss anything. Hey, how can we do this to have a look for all my emails, read every single keyword. Is there any reference to Tuesday to this date where I sent, I'll send an email, but I never sent an email or calendar invite. [33:50] Dave Pengelley: Demonstrated [33:51] Marno Brits: for me took about 30 minutes to build up and then from that basic email calendar checkup I've now expanded it into this I did play with nemo claw when [00:34:00] it came out just because I was curious to see what the secure version looked like I didn't like it then I went to open claw to see what's the difference and I was like this is all done I'll just my claw code code setup is significantly better and it's already customized and I have full control and I guess a better understanding of what it has access to and what it doesn't. [34:21] Marno Brits: So I'll just stay for code for now. And now I've desk dispatch available in channels. It completely replaces the need for, um, open floor at all. [34:30] Dave Pengelley: Yeah, but you're, you're running your own custom program somehow. It's not just sitting in cloud code and stuff, right? Like your thing that's checking, like what, what's your interface that you chat with it to using it through discord, right? [34:40] Marno Brits: So that's just for channels. [34:42] Dave Pengelley: And then you said that dollar 75 thing. So you've actually set up API tokens. You've built like a, kind of a bespoke. Custom Mono's Claw that is running. I [34:50] Marno Brits: would like to say that, but at the same time, I don't think it's that at all. All it is, is I'm running Claude Code with a bunch of skills and I've set up a channel so I can talk to [00:35:00] it via Discord. [35:01] Matt Slager: Where is it running? Is it on your computer? [35:03] Marno Brits: Yeah, it's all local. [35:05] Matt Slager: Interesting. [35:05] Dave Pengelley: But is it using Claude Code, like your Claude Code plan or like? [35:10] Marno Brits: Oh, so I've got API tokens for certain tasks and then I've got Claude Code for the actual computing. Interesting. But when there's like when I have to build a proposal plan or I am further developing something out on my computer, then I'll use cloud code. [35:23] Marno Brits: But for most of it, it's probably going to be ABI tokens. [35:27] Richard Webbe: Mano, can I ask a, can I ask everyone a question as the ignorant nerd here? Um, because of the tools we have available to us at AI, if we pick a different architecture, uh, as an orchestration platform, like, [35:40] Matt Slager: uh, [35:41] Richard Webbe: you know, open floor or something like that, it wouldn't be too hard for us to migrate, would it? [35:47] Matt Slager: OpenCore itself is an amalgamation of different things and wrappers. It's actually running off PyAgents, P -I. P -Y? [35:57] Dave Pengelley: Isn't it P -Y? Python? [35:58] Matt Slager: No, P -I, P [00:36:00] -I .dev. [36:00] Dave Pengelley: That's different than Pydantic Agents? [36:02] Matt Slager: Correct, that's just a Python framework. Yeah, so like PyAgents, Um, it's not for everyone. It's, it's super techie, it's super like developer, but if you want to build like Richard, like for yourself, if you want to like find an application layer where you can wrangle in agentic layers to it, something like Pi is exactly what you could use, or you do what Dave and like what myself have done, where you take the OpenCore public repo. [36:31] Matt Slager: So the actual source code of something like OpenCore, and then you give that to your local agent. You know, in the case of code code, for example, or codec CLI or whatever. And you say, let's just take the framework here. Let's just take the foundational fundamentals and adapt it to this and just use this as a reference, learn from this, you know, what you want. [36:55] Dave Pengelley: So, so when I was building my, what I called anti -claw where I was doing manual tasks, a bit like, [00:37:00] um, like running skills and stuff in anti -gravity. Uh, then I, [37:05] Matt Slager: is Richard frozen? Yeah, [37:06] Dave Pengelley: I pointed at OpenCore. Can you still hear me? Yeah, we can still hear you. Are you frozen yet? They've got the much better, the much better freeze frame than last time. [37:15] Dave Pengelley: Yeah, he looks [37:15] Matt Slager: really excited to listen. [37:20] Dave Pengelley: But, uh, I, I gave it the OpenCore repo and said, learn from this. And that's where I started getting some of my memory structures and some, some of those sort of ideas from that are integrated in pretty much all the stuff I do with any LLM. Now I've got this sort of memory. memory structure they use across all my projects, which is awesome. [38:00] Dave Pengelley: Uh, so I did that, but over the weekend I did move from my Mac to a PC. So I'm running on a Windows based install now because I, you know, I had a PC that had better RAM and CPU stuff than my Mac mini and I was sick of hitting limits on my Mac. So replatforming all your AI stuff from a Mac. Windows -based thing to a Windows -based thing is a whole journey in itself because Macs are a little bit more friendly because of their Unix [38:00] Dave Pengelley: -based nature Versus Windows where I actually had to put in like Windows WSL like a Linux wrapper in like Windows Server, like, I can't, I don't even know what WSL stands for. [38:09] Dave Pengelley: But I basically had to install Linux as like a sub -operating system integrated with my Windows. It's been a massive journey, but I put Docker on, so I wanted to play with NemoClaw as well. So for those that don't know, NVIDIA released NemoClaw, which is, they've got this thing called OpenShell, which is just a security layer, and OpenClaw lives in that. [38:28] Dave Pengelley: And so when OpenCore tries to do things, OpenShell says, Are you sure you want to give her this access? So it's just kind of a safety firewall around your OpenCore. But try to do that within a virtualized Docker environment on my PC. I was just hitting all these errors, I couldn't even get OpenClaw to boot up. [39:00] Dave Pengelley: And so I was like, okay, too hard. And I had ClawedCode looking at the folder. And I had my ClawedCode saying, um, trying to fix it all for me and rewrite config files. And I just could not, it was too much. I was like, you know what? I don't even know if this is a NemoClaw issue or an OpenClaw issue or a Docker [39:00] Dave Pengelley: issue. [39:00] Dave Pengelley: Just strip it back. So we ditched NemoClaw, went with straight OpenClaw. I still needed ClawedCode to actually configure it and get it working. So, so that's what we did. This OpenCLAW thing is not super simple. It's not a point -and -click install. It's a hassle. And then I finally got it installed and I've got it on Telegram and I've given it some tasks and said, just go away, build stuff. [39:16] Dave Pengelley: And then I don't hear from it for days and I'm going to go, are you still working? It's like, yeah, I'm still working. I'm like, sure you are. So I, I don't really know what's happening with my open core. I'm trying to get it to pay for its own batteries. That's what I told it I had to do. Uh, for you people, Richard, and this is what I said to the, the people at the cloud code meetup where they're seeing people using these things integrated into Slack and discord and, and open claws. [39:39] Dave Pengelley: And cause I asked like, who's tried to do open core in the whole room? Basically nowhere. Just a few of the presenters are the ones doing it. Like in this room full of techie people, almost no one has actually tried setting up OpenCLAW. [39:48] Richard Webbe: It's [39:49] Dave Pengelley: such a [39:50] Richard Webbe: great framework. [39:53] Dave Pengelley: But, but you see all the hype online and you think everyone's using OpenCLAW. [40:00] Dave Pengelley: Like I had massive FOMO. That's what I'm like. I need to, I'm, I'm, I'm an AI [40:00] Dave Pengelley: operator and I haven't even tried yet. I need to, I need to, I need to, I need to, I need to, I need to, I need to, I need to, I need to work this out. I need to have a foot in the, [40:05] Richard Webbe: I think that's a, I think what, what you're saying, that's a message to everyone, our audience and all our friends and all our customers and clients. [40:13] Richard Webbe: And it was very similar back again in the internet and the online days, even the computing dates, but what computer should I buy? Or what webpage? What, what programming? Who should I get to do it? Don't ask yourself those questions. Be like Nike, just do it. And then once you get your hands dirty, the concept becomes real in your mind. [40:33] Richard Webbe: I, you know, um, some of you know I did a lot of work for Mark Andreessen and Ben Horowitz. Of course, Mark started the first Netscape browser being, uh, first browser being Netscape. And I'm lucky to have those guys in my, uh, in my, uh, uh, my contact list. And I went to their, I think it was the Ray Summit, which is the, uh, cloud, uh, Python, whatever thing. [41:00] Richard Webbe: And what I thought was amazing is they got everyone in the audience. They're all, they're all nerdy. Smart people [41:00] Richard Webbe: like you, not that smart, but they had us recompiling some Python code and just playing with it. And I remember looking at my business partner, Rob, uh, Robby Perrin, who I'd gone over with, and we're looking at each other going, what the hell are we doing? [41:11] Richard Webbe: But you know what, just that hour of recompiling and adjusting a code, because I haven't programmed since I was like five, um, it just opened up my eyes to what was actually happening back then, what's possible. So my say is what the, the summary of what you guys are saying is just do it. Download it, play with it, but more importantly, if you're that scared, give us a call. [41:35] Richard Webbe: We will walk you through it, no charge, and help you get started. And then, away we go. [41:41] Dave Pengelley: The winning move is, tools like Claw and Anthropic, they're on to this. And they know, they know that Peter Steinberger, the guy that created OpenClaw, has been acquired by OpenAI. They know OpenAI is going to be, Building in this kind of claw style functionality, this sort of personal assistant 24 -7 agent stuff. [42:00] Dave Pengelley: They know that OpenAI are going to be making a play on that [42:00] Dave Pengelley: because they bought Peter Steinberger. And so they're just jumping in straight away. And so one of the things I shared, um, with them without trying to share too much, um, behind the scenes stuff here. But [42:11] Richard Webbe: this [42:12] Dave Pengelley: is my, uh, this is my co -worker. [42:17] Richard Webbe: And, [42:18] Dave Pengelley: uh, this is, in my co -work, I have my dawn rhythm that runs at 7 .04 AM, uh, and, uh, apologies for the scroll thing, it's a bit awkward because I'm remote terminated in my Mac where I've got this set up from my PC. [42:31] Dave Pengelley: You're [42:31] Richard Webbe: such an, [42:31] Dave Pengelley: uh, camera [42:32] Richard Webbe: buddy. [42:32] Dave Pengelley: But, but, here's, here's my instructions. We've talked about my, my, my Marvel people before, uh, my DC people. You are Perry White, Editor -in -Chief of Ballad Agenda Operations. Execute the dawn warning briefing. So go check all your memory files, then go fetch the calendar from an 8n where I've set up the link, then go query linear MCP for potential issues that are going on, go scan the headlines for sovereign AI, agentic workflows, uh, go check for brand references, for valid agenda, for David, [00:43:00] not the US math professor version of me, but me for syllogism. [43:03] Dave Pengelley: And so this is just, uh, just a skill. This is just text that you can get the AI to write for you. And I'll put this. Running on a schedule at 7 or 5 a .m. It'll use all your connectors and your MCPs and things that you sort of set up here in Claude co -work. It's all point and click to set this sort of stuff up. [43:20] Dave Pengelley: And then you can see all the different runs for your different rhythms and things. So it's actually pretty crazy that you can do a lot of this sort of morning briefing. Check all my details stuff, uh, without needing to... How do I... Close the screen share because now you can see all my Claude code thingies. [43:37] Dave Pengelley: It doesn't matter. [43:43] Dave Pengelley: All right, there we go. Does that stop it? [43:47] Richard Webbe: No, not yet. Reminds me of the CEO at a conference I went to who accidentally shared his account profile at the hotel, which did highlight several movies he'd watched the night before. [43:57] Dave Pengelley: Oh, great. [00:44:00] Uh, no, what did you see on that one? You saw that I'm doing something with 20 CRM. [44:04] Matt Slager: 20 is great. I'm surprised more people don't know about 20. Like, as far as just open source CRM stuff. It's really good. [44:11] Dave Pengelley: Yeah, just the fact that you can run it all. Like, I've got it up on my hosting at KVM. And, uh, the... Sort of open data model reminds me very similar the way the data model works to Salesforce circa 2011. [44:24] Dave Pengelley: Um, but in a free open source CRM, which is crazy. [44:28] Marno Brits: There's so many like CRMs that I remember this vividly when I first started getting into it. Bolt .new was still like the leading one before lovable came in and just splashed out in the water. And there was a bloke that was on one of the very first dev calls of bolt .new where they have like this weekly catch up what the community built. [45:00] Marno Brits: This guy built, I think it's called Panda CRM or something like that. It's just a basic, basic, basic CRM platform. It's making like 20k a month. And it's based, it could probably just be a 20 [45:00] Marno Brits: like, repo copy and then a new UI or pretty UI and that's it. Um, CRMs are so underrated for just a service because a lot of people just, They don't know what exists, and again, coming back to the scare thing, like, they don't know what they can trust, because there's no one that facilitates the safe space for them to feel comfortable enough to ask the questions. [45:21] Marno Brits: And see, okay, why is clickup .com better than monday .com? Or why is, like, Candle CRM better than 20 CRM? [45:30] Dave Pengelley: So, so, so is SAS dead? Everyone's going to build their own software [45:33] Marno Brits: now? No, definitely not. People are still using fax machines. So I'm sure that software will be fine for a long way to come. [45:39] Dave Pengelley: And I think from my point of view, there's always the build versus buy conversation. [46:00] Dave Pengelley: And the thing about buying versus building is there's a product team ostensibly behind that. They're thinking about the feature upgrades, their features to it. You run your business, unless your business is building CRMs, do you really want to be building a bespoke CRM? That's why something like 20 is great, because if you just want [46:00] Dave Pengelley: something basic. [46:00] Dave Pengelley: Uh, that's good enough that you want to run on your own metal, then just use 20 or, you know, pay their hosting fees and buy it by their cloud version, right? There's so many CRMs and so many of them are overbaked. So many of them over promise and under deliver or they're too heavy or they're too, sometimes for a good reason, they're very opinionated on how you should do things and how you should manage your data. [46:21] Dave Pengelley: And some of the old school ones like HubSpot are very opinionated because they came out of email marketing. And so they're very. heavy on some of that sort of stuff versus the others and the free editions give you a lot of stuff but there was a lot of overhead you don't need as well with some of these things but that little sneak peek you saw there to what I'm building in cloud code is um and I think I mentioned I'm building a little side hustle for website upgrades and flipping websites and I'm trying to make it as automated as possible so that I can I've got a n8n crawler that runs every night to look for businesses with old crappy websites That loads them as opportunities into my CRM. [47:00] Dave Pengelley: I [47:00] Dave Pengelley: can review them and if I look at the website and go, Yep, that's a, that's a possibility. I'll update the stage. That'll stick it in a build queue. Where then I'll have an automated job. Pick up new, like check the CRM periodically. And then go, okay, we're going to build these preview sites. It'll go and pull down and use my local cloud code to rebuild a new preview site, at which point then it will update the CRM. [47:21] Dave Pengelley: And the next step is then to integrate things like Twilio and SMS. It'll SMS the customer and say, Hey, we noticed you're on your website, could do with a refresh. Here's our offer. Here's a link to see what we've built for you. Would you like it? Yes or no? [47:33] Richard Webbe: I'm [47:33] Dave Pengelley: just trying to automate that whole process from discovery to preview build to, to sale and do it as, as hands off as possible. [47:42] Richard Webbe: Um, [47:43] Dave Pengelley: using [47:43] Richard Webbe: tools. I met with a, a very successful Australian entrepreneur. uh, late last week, uh, I've made this conversation with David, but, and, uh, I was talking to them about this and they have seen all the evolutions. There are, there are a venture capitalist, a private equity, [00:48:00] they're running a very large, um, uh, service provider for the power industry. [48:05] Richard Webbe: Uh, and, uh, he paraphrased what he thought was going to happen. And I said the same thing. Um, you know, people are paying their cloud application providers 10, 20, 100, 1 ,000, a million bucks a month in very large ins. He envisages that people are going to go back because the data is the critical component, not the application. [48:26] Richard Webbe: Because now, the application and the plug -ins and the handling, like you guys are doing. So, you know, it's, uh, If you can imagine it, maybe you can build it, but there's a line where you can't build it because you don't have the architecture capability in your hands, but it needs someone like you guys, but you guys are a lot more effective than me paying maybe a large multinational. [49:00] Richard Webbe: software sass vendor. So that whole power, as Matt said, is going to shift that, that cost retention. And he envisions that in some areas, they're going to say to the cloud application provider, we've been paying you a hundred grand a [49:00] Richard Webbe: month. He's 500 dollars a month. Otherwise get lost because we've done the, and that's probably an extreme view, but he said, we've done the buyer on computers again, you know, the whole bill shock with cloud. [49:11] Richard Webbe: We all know our customers and clients have suffered that. Okay. I need. Uh, as you're saying, I need that data. I need to know when my clients, um, you know, subscription fails. I need to know when they have faults. I need, you know, there's service now in there and all that. So I gather all that data and I put my own agentic agents, right? [49:29] Richard Webbe: I don't have to pay for those extra modules to those big companies. And it's all about hosting my data. So my data is reliable and available, merging that in a sovereign AI and merging that with my own sovereign data so that I have my unique data. Competitive advantage or value propositions. Tell me to shut up, I'm waffling on. [49:48] Richard Webbe: But it just really dawned on me that, uh, you know, we spoke a couple of weeks ago about the typing pool disappearing and then the fax machine disappearing, but not completely, Marno, because I've still got one. [49:58] Marno Brits: Yeah, a lot of people still have it. I think that's [00:50:00] the big argument the gentleman is making in the video. [50:02] Marno Brits: He's talking about everyone's, um, all, there's all this hype about AI. But very few people are considering the infrastructure requirements. And even then, you know, much talked about the same thing with TerraLab or TerraFab. Where we're just, we're running out of space, we're running out of, well not space, but we're running out of energy and we're running out of infrastructure and we just have to get, there's so many constraints that are going to start lining up, like with chips, we need the energy to make the chips, we need the water to produce the energy to make the chips, and then we need the chips to create the infrastructure that we can run the AI on, like there's so much that's gonna, hopefully not, but all at the same time just blow up in our faces. [50:42] Richard Webbe: The [50:42] Marno Brits: future is going to be, for me, I'm optimistic about it. [50:46] Richard Webbe: I am, but it's creating a brand new geopolitical situation worldwide, isn't it? And when I say that I don't mean religion or politics, I mean supply chain geopolitics. Who's got what there? Where are we going to do it? [00:51:00] I [51:00] Dave Pengelley: saw an article saying that apparently, supposedly, allegedly China's I'm throwing in a bolyry and I'm like, uh, trying to send some chemical shipment through to Iran which would help them make their rocket fuels and other things. [51:11] Dave Pengelley: Um, And it's like going through the same place where the u s sub sunk and Iranian ship, and it's that because like China doesn't sort of sort of hope IRAN, they're gonna get their oil cut off. But if the U .S. stops China, then China's going to cut off rare minerals and rare earth materials that the U .S. [51:30] Dave Pengelley: need for their targeting systems and their missiles. Like, there's this whole, like, stalemate deadlock happening with who's got which resources and who needs them and how this one little strait in the Middle East is completely jamming up global supply chains. And, and over, over the last 50 years, a lot of economies have de -sovereigntized their supply chains, and they've outsourced everything, but under free trade agreements, and, and moved it out, and we don't own anything. [52:00] Dave Pengelley: Um, but when it comes back to the, the, the top of AI and [52:00] Dave Pengelley: data, NVIDIA is clearly in that space. And so [52:02] Richard Webbe: props [52:03] Dave Pengelley: to Jensen Huang from, from NVIDIA. Yes, he's sort of proclaiming the, the great open core Nemo core thing as the new operating system for AI because it's got IO and blah, blah, blah. But he also has a very clear commercial reason to pitch this because NVIDIA will sell the hardware to run your own local models. [52:22] Dave Pengelley: Thank you. So that way you can keep all the AI in -house behind your firewall, keep your secrets protected. And so if people buy into this claw thing where they're going to need a lot of tokens. then they're going to need cheaper tokens than they want to buy from Anthropic. They'll want to run their own models, so they need to buy the hardware from NVIDIA. [52:42] Dave Pengelley: NVIDIA wins. [52:45] Richard Webbe: It's, it's interesting. I know you're going to say something, Matt, so I'll shut up, but in the networking game, when networking took, yes, Matt, I've just, I've just put you in it, so think of something to say. But what I'm saying is, In the networking game, everyone wasn't [00:53:00] sure what protocol to go for, what connection to go for. [53:03] Richard Webbe: IBM tried to lock in token ring, Xerox brought Ethernet, there was IP networking. What I learned from those days right through to now, anyone who tries to create a proprietary environment loses. So whether it's Entropic or it's, you know, OpenAI or NVIDIA, if they try and, Lock it in and say, well, you have to do that with it. [53:26] Richard Webbe: And I think even more with AI, that's not going to happen. And so rather than do that, we're looking, I'm seriously, we're looking at to the people like yourselves, if you can think it, how are we going to build it? And if we build it one way and we have to change platforms, AI makes it, it's not that hard. I mean, [53:43] Dave Pengelley: that's what OpenCore took off because it was open source and this is the future of software. [54:00] Dave Pengelley: Not necessarily being able to build something and sell it. Not everything, not everyone wants to build a full SaaS platform and manage those customers. But can you build the ideas? Can you get the ideas out there to grow and expand? And [54:00] Dave Pengelley: if you want, do you monetize that with services and advisory and consultation to help people get up and running? [54:07] Richard Webbe: Yeah, my friend Ben Horowitz said, he wrote that book, Software is Eating the World. With inputs from Mark Andreessen, I think he was a bestseller worldwide. And my entrepreneur mate who I caught up with last week, he said, No longer is software eating the world, AI is eating software. And I thought that was a very unique comment. [54:24] Richard Webbe: That is, again, back to these tools. If you can imagine, you can build it. You don't have to pay a fortune for it. [54:31] Dave Pengelley: You saw, you saw in my briefing that it uses linear and it goes and like logs issues and basically build a task list using linear. Matt, you use linear, but you never log into the linear, linear web interface, right? [54:41] Dave Pengelley: It's all done by your end. [54:42] Matt Slager: Yeah, I'm part of a specific workspace that I built for one client and that. That actually hilariously is a CRM and complete software suite that I've built for them. But yeah, in there, I've got just this infinite list of issues because I've got agents that will take my ranting and feedback, [00:55:00] take all of the tickets that I received from that team, and then turn those into things and then structure them and schedule them for the next releases. [55:07] Matt Slager: And that's just for one client. So then like I've got my own Um, linear workspace as well with all my project separation for my own personal stuff, professional stuff, um, leads, you know, like, you know, potential proposals coming up and that sort of thing. And it's literally all triggered by my local agents that I run. [55:25] Matt Slager: That [55:25] Richard Webbe: is so cool. That is exactly where we're headed. And it is connecting the ideas people like you guys to actual business output much quicker and much faster. And that's where we all want to be. I think that's just so cool. [55:37] Dave Pengelley: Yeah. I'm building, I think I've mentioned my other SAS that I've got in beta at the moment, chartreporter .com. [56:00] Dave Pengelley: For people that are doing forex trading using MetaTrader, they can, they can journal and analyze their trades on there. And some people are going to just want to click in and use my interface and do that. Others might want to use their agents. And so one of the things that I haven't finalized in the beta yet is [56:00] Dave Pengelley: MCP. [56:00] Dave Pengelley: Now think what you will of MCP. It is a standard to allow agents to use their agents. to talk to the data. Um, and you know, that that's an avenue that I think some people will prefer to use. I'll just want their agents to go review their journal and their trades. I don't want to go look at the data on the mat. [56:14] Dave Pengelley: They want the data accessible. And so what my tool is doing for them is taking the data out of this, this restricted black box of MetaTrader and putting it in a super based database and then giving them access for their agents to go and talk to it. Uh, and that comes down to the reason I could build any of that and make any of it make sense is I am in the, the, the, Frame, but that particular product I've spent over a year in the Forex trading space using MetaTrader, seeing the limitations, understanding that space. [57:00] Dave Pengelley: And so just to close us out as we, as we top the hour, that expert in the loop thing becomes really, really critical. If you just see AI as a generic tool, that's going to do things for you and it's a magic black box that you just expect the world from. Increasingly the magic fairy dust has worn off and oh good, it can write an email that sounds like [57:00] Dave Pengelley: Benjamin Franklin wrote it. [57:01] Dave Pengelley: Excellent. Thank you. Um, that's not that interesting anymore, because... Who cares? Uh, I don't need Lin -Manuel Miranda's version of my, of my, uh, my idea written in the style of Hamilton anymore. Um, but taking your own creative inputs and your own knowledge and domain expertise and using the AI to solve problems and develop that. [57:22] Dave Pengelley: But the guy who, who got his dog cured from cancer because he had ideas, he asked questions, he worked with the right researchers to then progress those things on and made it a, made an mRNA. custom genome sequence that, that fixed his dog's cancer. It's applying your knowledge and your expertise. And so to close out, I just posted this on, on LinkedIn overnight. [58:00] Dave Pengelley: But my big win, was uh, someone came to me and they had a voice recorder, just one of those little Philips manual dinky voice recorder things for recording meetings, with a corrupted file. They're like, we, we've, we've got this recording, but it, but it's not there. The file doesn't work. And I plugged it in, had a look and went, that's because [58:00] Dave Pengelley: it's zero bytes big. [58:00] Dave Pengelley: The file is not there. There's no timestamp on it. There's clearly an issue where the recorder didn't save the file properly and it's, it's gone. And they're like, Oh, it's gone. So just, just leave it with me. And so my first instinct, like, like, you know, cause I'm old was Google. I'll go to Google and look for a file recovery software. [58:16] Dave Pengelley: That's what I'm going to do. I'm going to, there must be a tool out there to do it. But before I, before I did that, I was like, hang on. I have Cloud Code and now it's all running on my Windows environment and it's all set up with all the Linux things that I talked about and it's all ready to go. So I just said to Cloud Code, first of all, I copied the zero -byte file across and said, can you do anything? [58:34] Dave Pengelley: It's like, this is a zero -byte file. There's nothing to fix. What are you talking about? I was like, okay, well, that's from this USB recorder. It's plugged in right now. Go take a look and it had a look and it then helped me build an image file of that recorder so that when we weren't touching the original data, we were just pulling a disk image of it and then it knew that I was like, I've got this Linux thing. [59:00] Dave Pengelley: Is that better than using PowerShell? It's like, actually, that's much better. We can do things in there. So then it went and found the right tools that could [59:00] Dave Pengelley: analyze the underlying disk partitions. and it found the hexadecimal fragments of file data. I went, I found a file that is not in the file system. [59:09] Dave Pengelley: Let me check. Yep, this is a 96 meg file. But it's a, it's a web file. It's recording. Let me recover it. And it recovered it. And put a recovered web file on my hard disk. And I showed the client. They're like, that's, that's exactly it. How did you do that? That's amazing. And this took under an hour. Now, my dad used to do the hard disk recovery from crashed hard disks in the 90s. [59:28] Dave Pengelley: Take him days. Days staring at sort of these, these, these ASCII screens of hexadecimal codes and, and bytes and bits and trying to map the data and recover certain files. Claude did this for me as a background task while I was multitasking and building my other projects in less than an hour. Like it's just crazy, but I'm that expert in the loop thing. [01:00:00] Dave Pengelley: I knew! That sort of stuff was possible because I knew my dad had done back -end partition file disk recovery. I knew that these recorders, when they're recording, they're writing the data onto their little [01:00:00] Dave Pengelley: storage, but there was just some issue at the end where it didn't wrap it up and top and tail it and register it. [01:00:06] Dave Pengelley: But the data was there, the thing was recording, so I thought there's a chance. The information is there, we just have to find it, and sure enough, Cloud Code was able to do it. So it comes down to, you know, knowing the right questions to ask. [01:00:19] Richard Webbe: It's all about the, that's what I always say, it's all about the question. [01:00:22] Richard Webbe: And it's at this stage, I'd like to promote David's book, Haiku, Personal Responsibility, available on Amazon. Sorry. [01:00:31] Dave Pengelley: Very good. Um, yeah, so anyway, that was probably my biggest win. I mean, like, aside from all the other things, I'm loving being in the Cloud ecosystem now. I can't say me. Going back, anti -gravity is dead to me. [01:01:00] Dave Pengelley: And... Between the things that Cowork's doing now with those schedule tasks and actions, what Claude Code's able to do, it's just... Crazy, just crazy what's going on. And you just got to ask the questions and experiment. And that's, that's all why we're here and talking about our stories. We're not, I'm not, [01:01:00] Dave Pengelley: not a high bro. [01:01:00] Dave Pengelley: Like go, go can do anything. I can do anything. I'm I'm, I'm happy. Did you like [01:01:09] Matt Slager: achieved some sort of sponsorship from lanthropic or what? [01:01:14] Dave Pengelley: No, I wish. [01:01:15] Matt Slager: Because like for anyone who like listened to Dave's amazing story there, which for starters, like massive pat on the back, like that's huge. Um, you can, you can cut the word cloud code there and insert any particular coding agent. [01:01:28] Dave Pengelley: Codex or something might've been able to do it. It's [01:01:31] Matt Slager: not a might. It comes down to if you can ask the right questions again. [01:01:34] Richard Webbe: But here's the trick. It's like people don't know they've got a spanner and we know how to use the spanner. So if anyone out there is confused, give us a call. And it's fine. We want to help. [01:01:44] Richard Webbe: And [01:01:45] Dave Pengelley: the difference is though, Matt, like, like, if people think they can just go to ChatGPT on the web and do it, that's not going to work. Like, we're talking about, you know, having the, the code, the local running execution tools, like a codex. Like Gemini CLI or Claude code, but you've actually [01:02:00] got to go a level further down and then just go into your web browser and go, Hey chat GPT, help me recover this file. [01:02:05] Matt Slager: Well, you'd start there and say, how do I do this? And then it would say, well, you probably need to run the codec CLI and go from there. Yeah. It's a crazy, crazy position to be in, to just be able to do anything. Like you said before, Richard, if you can imagine it. You can do it. [01:02:22] Richard Webbe: Absolutely true. And I froze again. [01:02:25] Richard Webbe: I do that. [01:02:26] Dave Pengelley: Well, I'm on that. [01:02:28] Richard Webbe: Uh, [01:02:30] Dave Pengelley: he's gone. All right. Good to see you back to say goodbye. Uh, thanks. Thanks guys. Good chat. Good luck with the tech this week and all the things and keep, keep experimenting, keep operating. Uh, awesome to hear different wins and stories around how you're personalizing stuff and You know, everything from actually personally talking to people, Matt, and the impact that's having, and just hearing those stories around where business owners are at, to what you're doing, Marno, with your whole custom dev AI agent, agentic operating system thing, [01:03:00] um. [01:03:00] Dave Pengelley: Yeah, and, and Richard is good, like you're not as hands -on yet, but that gives us that great point of view on, on my next job is to help you get set up with co -work. [01:03:11] Richard Webbe: Yeah, let's go, please. [01:03:14] Dave Pengelley: And I think once you start playing with it, you'll just ask your questions and you'll probably do stuff that I would never think of. [01:03:19] Dave Pengelley: I'm a sub -representative [01:03:20] Richard Webbe: of the ignorant business of Australia, so please help. [01:03:24] Dave Pengelley: I'm excited to set you up and let you run loose on co -work. Alright, any last comments from anyone else? Thank you. [01:03:32] Marno Brits: Thank you. Have a good week. [01:03:34] Matt Slager: Yeah, thanks to anyone who's listened, watched, whether it's live or recording. Um, like Richard keeps saying, you can reach out to us, like, we each have our own places. [01:03:43] Matt Slager: Um, we're not going to give you our mobile numbers, but, um, yeah. If you, if you need to get help or if you ever get stuck, um, the worst we can say is no. [01:03:54] Dave Pengelley: links links are in the description for our websites and where to find us [01:03:58] Richard Webbe: if you ring us we'll [01:04:00] drop we'll do it or we'll point in the right direction drop [01:04:03] Dave Pengelley: a dm uh be quiet on the chat today no no uh chat crowd interaction everyone's a bit shy today but uh if you've ever got questions jump on live drop them in the chat we'd be we'd love to interact with the crowd and see if we can help you in real time but thanks boys [01:04:18] Matt Slager: thank you [01:04:19] Dave Pengelley: see you all on the next one [01:04:20] Matt Slager: catch you later