Steve Jobs' 30-Year-Old Tip for Getting ROI from AI

[00:00] Matt: Just repeat [00:00] Dave Pengelley: Richard, [00:00] Matt: all [00:00] Dave Pengelley: you [00:00] Matt: of it again. [00:02] Dave Pengelley: ran too hot. I'm um gonna see how we go today. I'm pretty tired. I've I've I'm gone for the Y Monster instead of the coffee today because I am I am exhausted after this week. Uh [00:14] Matt: Dangerous. [00:14] Richard (Syllogism) Webbe: All [00:14] Dave Pengelley: yesterday [00:14] Richard (Syllogism) Webbe: right, oh yeah. [00:15] Dave Pengelley: Yesterday I did my normal 5k walk and I did upper body weights and then did an hour of intense karate class last night where we were doing karter practice and my my my BPM, my heart was getting over 160 doing these katas, and I'm just like so this morning I'm I woke up and I'm and I'm and I mowed my lawn front and back yesterday as well, did all the edging and the whippersnipping. So today I was like, oh, I'm dead. I [00:37] John Ansley: That's [00:38] Dave Pengelley: need [00:38] John Ansley: a [00:38] Matt: Okay. [00:38] Dave Pengelley: to [00:38] John Ansley: that's [00:38] Dave Pengelley: spark [00:38] John Ansley: a [00:38] Dave Pengelley: up. [00:38] John Ansley: classic [00:39] Dave Pengelley: I need [00:39] John Ansley: what [00:40] Dave Pengelley: got to spark [00:40] John Ansley: my [00:41] Dave Pengelley: up [00:41] John Ansley: wife [00:41] Dave Pengelley: for the podcast. [00:41] John Ansley: my wife did the same sort of thing yesterday. So she goes out and she goes, Oh, I'm gonna I'm gonna do things. So she ends up, you know, we go to the gym and then she goes for a big long walk and then she goes and does a whole pile of golf practice and then she goes and plays pickleball and then she comes home and she says, I'm really tired and I'm like, Really? What [00:58] John Ansley: so I I [01:06] Matt: Very good. [01:08] Dave Pengelley: Yeah, no, I normally crash out around midnight, but I was in uh bed in sleep with lights out before 11 last night, and I woke up this morning, I was just still so tired. I was like, uh [01:18] Dave Pengelley: So good old, good old white monster. Um I had a coffee this morning and then I was gonna make another one. I was like, no, I need the hard stuff for the podcast. So [01:28] Matt: I feel bad [01:29] Dave Pengelley: um [01:29] Matt: for your heart. I know you said uh you said you're doing upper body weights or something, but [01:34] Dave Pengelley: Yeah. [01:34] Matt: I heard Apple I heard Apple body weights [01:36] Dave Pengelley: Apple body [01:37] Matt: and [01:37] Dave Pengelley: weights. [01:38] Matt: I just was picturing apples. [01:39] Dave Pengelley: No, no, well, I wasn't using Apple Fitness. I I I subscribed to um the American group Beach Body that made like Power90, P90X, and all those sort of kind of things. [01:46] Matt: Uh [01:47] Dave Pengelley: I do [01:47] Matt: yes. [01:47] Dave Pengelley: I do their programs, the uh the old video workouts in the home, the home gym. [01:52] Matt: Oh you don't just pull like a chat GPT workout [01:55] Dave Pengelley: No, no. I [01:57] Matt: on [01:57] Dave Pengelley: mean, [01:57] Matt: voice [01:57] Dave Pengelley: I I [01:58] Matt: mode. [01:58] Dave Pengelley: actually find the video. And having the instructor keep you on pace is really useful. It's like going to a gym class, but you know, I can do it in my own time at home. But just having that pick up the next way, pick up the next way. Because I'm just going off a list, you end up dawdling a little bit between things, and you have an extra drink, and all of a sudden, this half-hour workout goes for 90 minutes, and you're like, that's that's inefficient. [02:18] Matt: Yeah, no, [02:19] Richard (Syllogism) Webbe: I [02:19] Matt: definitely [02:20] Richard (Syllogism) Webbe: I have a [02:20] Matt: great. [02:20] Richard (Syllogism) Webbe: complaint about iPhones and workout because when I go to the gym, people use their iPhones to keep track of their workouts and their heart rate and that, and you're waiting to get onto the machine and And sure enough, they drift off into social media and you're standing there waiting for someone sitting there on the bloody machine playing with their phone. [02:36] Matt: Or if it was me, I'd be there like messaging my agents and keeping work moving. [02:41] Richard (Syllogism) Webbe: Exactly. No, your agents would be messaging Dave's agents, who'd be messaging my agents and then going back to Dave again. [02:50] Dave Pengelley: Yes, well, split seeing Matt Matt has used the word AI and agent, so that tells me it's time for this. [03:10] Dave Pengelley: We are the AI Operators Show, the casual pub style chat where we talk about what's going on in the world of AI, how we're using it in our lives, and occasionally having interesting guests on to talk about what they're doing with AI and to get some fresh and new perspectives. And this week, John, we'd like to welcome you to the show. Welcome to AI Operators. [03:28] Richard (Syllogism) Webbe: Yeah, [03:30] John Ansley: Great to be here. [03:31] Dave Pengelley: Brilliant. Now um what what what what uh what did you do during the week to keep you busy? Uh and how does that intersect with AI? [03:41] John Ansley: Oh God. So this this week's been a uh a busy one in terms of workouts. So uh the gym and so there's an intersection with AI because I I do use uh occasionally um use Gemini for helping me with my programme and make sure that I don't uh that I don't overtire any particular part of my body. So That's been quite a useful thing. I'm I've only really been a gym junkie for about two and a half, three years, and uh and so it was a place I was quite uncomfortable going. So having having access to um to an LLM like Gemini was a really nice way to ease me into my workouts. My my background is mostly martial arts, and so you know, going to the gym was sort of weird for me. I got most of my exercise out of my martial arts. Um, so I'm trying to augment that. [04:28] Richard (Syllogism) Webbe: I forgot to mention that you're both a bit dying. When you're not prepared or when you are prepared. [04:34] Dave Pengelley: I mean [04:34] John Ansley: Um [04:34] Dave Pengelley: my martial [04:34] John Ansley: and a [04:35] Dave Pengelley: arts is a pretty solid workout, but the but the resistance training and stuff at the gym is something you don't it has have as much focus on. [04:42] John Ansley: I think particularly too as you know as as you get older it becomes more and more important. So having a uh a habit of getting to the gym and pushing some weights around is particularly important to you. And I see a lot of people in my age group who um you can see their bodies almost start collapsing in on themselves. And so they're not Doing anything with our upper body. They go for walks or whatever, but they don't actually do anything strength wise for their upper body. So that to me is a really important aspect of it. Um, the legs I haven't had to worry about because I've been doing so many squats and so many kicks and so many stances for so long. But uh but I am doing legwork as well. Um uh don't forget leg day. So yeah, so it's uh it's um yeah, so that's that's one aspect of AI this week. The other aspect for me has been I've been doing a whole pile of strategy work on on our um on our business and uh and also So, you know, the creation of LinkedIn posts and things like that where I go into AI to make sure that what I'm talking about is well founded and that I can back it up. So a lot of a lot of the things that I put out on LinkedIn would otherwise just be an opinion rather than something that I can back with a with a solid argument. And as I said, the the strategy stuff around the business is amazing. There's sort of access to different frameworks that you can get very easily and throw questions in there and say, how does this framework? apply to our business or whatever it might be, [06:02] Richard (Syllogism) Webbe: Let [06:02] John Ansley: um, [06:02] Richard (Syllogism) Webbe: me do [06:02] John Ansley: really [06:03] Richard (Syllogism) Webbe: a [06:03] John Ansley: help [06:03] Richard (Syllogism) Webbe: quick uh [06:03] John Ansley: smooth the [06:04] Richard (Syllogism) Webbe: let [06:04] John Ansley: thinking. [06:04] Richard (Syllogism) Webbe: me do a quick pseudo intro of you, John, for our our listeners. [06:07] Dave Pengelley: Yeah, because uh yeah, we can we keep to know more about the business itself, [06:10] Richard (Syllogism) Webbe: Yeah, yeah, [06:11] Dave Pengelley: so [06:11] Richard (Syllogism) Webbe: yeah. [06:11] Dave Pengelley: uh, [06:11] Richard (Syllogism) Webbe: So so John and I met at BHP AT many years ago. And John was very big into uh you know technology and improving business when we're both at BHPIT. And then John was very big with the SAP and the supply chain and the marriage between business and efficiency. And I can tell you now BHP did [06:29] Dave Pengelley: yeah. [06:30] Richard (Syllogism) Webbe: groundbreaking work and John introduced groundbreaking strategies and architectures and ways of filling supply chain for some of the largest companies in the world. Still a good friend with SAP, I think, at times. And then John and I caught up recently and he he he he John's been a CIO at a number of major Tier One Australian companies. So he's no slats when it comes to owning the budget, owning the need, owning the business and delivering results, which he's very, very good at. And then John and I caught up recently and we're at dinner and uh Uh somebody up to John said, I've started this company called Nobi and it uses AI and it's groundbreaking. And when I had a look at it, I went, oh my goodness, that is unbelievable. So uh I'll let John explain what Nobi is, but I'm excited for him and I think it's really impressive. [07:17] John Ansley: Yeah, thank thanks, Richard. Um, you know, the the the number one thing that I've seen time and time again across almost 40 years in in technology is that as as uh new technologies come out, they they all All suffer from a very, very similar problem. And if you if you think about, you know, uh whether it was ERPs or CRMs or SaaS or client server or any of these massive the with the internet, uh any of these massive changes as they've come at businesses, um, everybody says get your get your people and your process and your and your data sorted out. And process is the one that's the worst done. And and I've seen this in like I say, decades of doing large transformation. Transformation programs, you never capture your process well enough. And you know, you're encouraged by the vendors of all these different systems to adopt their best practice. But nobody has an actual how do I do this? You know, so so um nobody is really uh um at that sort of intersection between um work and the intent that your business has for how that work should be executed. So it's capturing your best practice and making it available. to everyone across across your organization. And when I say everyone, that means any worker type including AI and automation. So and robotics if you want. So if you think, okay, we have a particular way of doing a transaction in a system, capturing it really quickly and then making it available to everybody everywhere in the world in your business, in their own languages, nice and easy to follow, and you can track that they're actually doing it. And no one's really solved that problem before the way that we've gone After this. And we've based this all around a learning concept called cognitive load theory, which basically says people learn better if you give them bite-sized, engaging pieces of information. So you can take a long, complex process and chunk it down into bite-sized steps with video or images and text, and then present that to people. And they learn about 50% faster, and they retain that knowledge about 30% better when they learn this way. And we've aimed it to be very much that just in time information. So it's not necessarily like a learning management system where you do it once a year, but it's when you want that information, it's accessible. And I sort of think of a you know Jiminy cricket on your on your shoulder telling you what how you should behave, right? It sort of gives you that guidance right when you need it, when you're about to go off to the pub with that sly looking fox, right? Um you know that's it that's uh you know that's the the the analogy. So um we spent um the last sort of We've been in market now just on three years. And we spent the first two years of that just trialing it in the small to medium enterprise market. And in the last year, we've started to focus on large enterprise and we've started to land some very large businesses. And we're now trying to expand inside those accounts as we land them. So typically they're landing with a, you know, a smallish number of users. And then we want to expand across the business because we believe that we've got a really solid benefit pattern. Package for every area of a business. There's not, when you look at what we do, there's logically not a single business in the world that couldn't use us. And every part of those businesses could also use us, whether it's in the supply chain area, whether it's in customer management, whether it's in finance, HR, wherever it happens to be, you can see this works across the entire organization. [10:55] Dave Pengelley: This this morning I was watching an old um lecture from Steve Jobs. From back in 1992 to MIT MIT Sloan School of Management students. And he referenced a book called The Business Value of Computers, written by Paul Strassman, who used to be CEO of the Pentagon and do other things back in the day. And he did a big study and [11:18] Dave Pengelley: found out that successful enterprises that were successfully deploying IT and ones that were less successful or not seeing the same level of results from IT both had about the same spend. It wasn't that some spent a lot more on IT. And some spend a lot less. They were both spending about 2%. The difference was the more successful ones were focusing on operational productivity versus management productivity. So if all your 2% spends going into making managers more efficient, you don't [11:43] John Ansley: Yeah. [11:43] Dave Pengelley: get the same returns as if you can make your operational processes and staff more [11:47] John Ansley: Yeah, [11:48] Dave Pengelley: efficient. [11:48] John Ansley: yeah. I hundred a hundred percent. There's a really there's a really cool sort of picture here too, and that is that uh something like 95% of IT spend is spent on on 20% of the people. Right. So so if you look at that, you've got 80% of your workforce who are getting bugger all spent on them. And that's that's definitely true for people who are doing any of the manual tasks in a business. So you're you're you're dead right. Sort of a lot of the systems are built to yell at the workers. Right. So they're a management tool to yell at the workers and they just keep screaming, Do something, do something, do something. They never say, Here's what I want you to do and how I want you to do it. Right? Here's how you can improve your capability. Here's how you can move up that sort of um that knowledge matrix from unaware to being fully independent and able to do the job. Right. And and that's that that's exactly the problem that I wanted to solve. We we did an enormous it was a I think it was a two-year exercise at at BHPIT where we sent two of the brainiest people I've ever met out across the entire organization. So I think we had something like 34 mining businesses at the time. And they went out and they mapped the entire Entire business processes of all of those businesses and then crunched it down into what the standard could look like. And at the end of that, we had this beautiful business process map, but it didn't still tell you how to do anything. It just showed you the things that we did. It didn't show you how we did them. And um, and I see that with a lot of customers. They come to us after they've done a business process mapping exercise. They've got beautiful detailed business process map, but it means nothing to the worker on the shop floor. Right. It means nothing to the person trying to use the system. And so when you and they and they you know, one particular customer, Storco, who's just brilliant at going after efficiency and operational gain and and employee engagement and retention, they had done that business process mapping exercise and then they went, Oh damn, we haven't shown anyone how to do stuff. And so they came to us and said, You're the how to do stuff picture, right? And [13:55] Dave Pengelley: Yeah. [13:55] John Ansley: so when when you change that, so we we we went in there and um They put QR codes onto all of their equipment in their factory. And now the startup operate and shutdown processes are followed absolutely perfectly. And they've had a 70 75% reduction in breakdown of their machinery. Right. It's [14:15] Dave Pengelley: Yeah. [14:15] John Ansley: just like, you know, and just so simple, so simple to do. And uh and you look at those sorts of results, you go, this is this it to your point, Dave, this is exactly what it is when you operationalize properly, uh, you know, the the the benefits. [14:31] Dave Pengelley: Yeah. [14:31] Richard (Syllogism) Webbe: These are hidden costs, aren't they, John, that people just don't see from above and below and around. You need to have this holistic view and then know [14:40] John Ansley: Yeah. [14:40] Richard (Syllogism) Webbe: how to fix it. [14:41] Dave Pengelley: The the business process map is is a flow chart, but they're all black boxes, and within each of those black black boxes, there's entire workflows and steps within them that you don't necessarily [14:49] John Ansley: Yeah. [14:50] Dave Pengelley: capture at that level. [14:51] John Ansley: Yeah, yeah. And it'll say it'll say do these things, but it won't say how to do these things, right? So that's that then that's the differentiation. And and and the The problem actually gets worse with AI. So if if you haven't defined the how-to properly, and then you add in automation or AI on top of it, all you do is accelerate the mess. So, you know, that this um where where where um where I'm landing in terms of the way I think about what we've done is we've built the governance infrastructure for businesses. Right? So it and it applies equally across any type of worker, whether it's an AI worker, a human worker, a robot, whatever it happens to be. these instructions become the your absolute gold, right? And then they they also are um Novi as a platform becomes your efficiency optimization platform where you can see you can get feedback either through reports or through people you know saying, hey, this doesn't work well doing it this way, I suggest we do this. And then you can make that change to the to the Novi instructions. Bang. And you you know you've got best practice improved. So it's um it's really interesting. And I've spent a lot of time over the years doing Kaizen and Six Sigma and you know all those are lean and all those sorts of things. And they're great as long as the experts in the room with the people showing them how to do it better. As soon as the expert moves out of the room, the skills atrophy and sort of 12 or 18 months later you need to go and do that exercise again. So you know there's you know it's really a missing piece. And when when I looked at the the market, there are some people playing in parts of this. They do they do it for manufacturing companies only or they do it for um they do it for desktop workers only or they do it for you know that they've really focused in on little bits and which is startups will tell you they should you find a niche and go deep. We've we've done it the other way. We've gone this is a problem for everyone we're going to go horizontal. And uh and that's that seems to be really different. And the way we're doing it using some of our own intelligence that we've built programmatically and also um doing things like We've built an MCP connector to uh that allows you to use, say, Claude to uh search for surface, create, or execute uh the Nobi instructions. So you could just say to Claude, run month end for me or do my end-of-shift report for me. And it'll go and find the Nobi that does that and follow your exact instructions. So, like I say, that governance infrastructure, so you don't just get you know random results, you get actually guided results. [17:30] Matt: So how does it how does it work? Is it just like another another tool in the toolbox that the team has to sort of use, or does it integrate with their existing tool set? You know, like if they you mentioned like ERPs and SAS and CRMs that they they may already be using, that's already yelling at the team, you know, [17:48] John Ansley: Yep. [17:48] Matt: do they then just alt tab across to Nobi and then get their little instruction, then come back to their tool set? [17:54] John Ansley: Yeah, it depends. It depends on the task. So like like my favorite my wife says her favorite thing to say is it depends. But it it depends on it depends on the task. So if it's say a physical task, take an ERP physically sort of generated task. Here's a work order coming out of SAP that tells you you need to do a maintenance task on a vehicle. Right? So it could be that you've embedded a uh Anobi URL link in the work order and they just click on that, it opens up in Nobi and it shows them exactly what to do. And as they do that maintenance task, they fill in I've done this, I've done this, I've used this much oil, I've replaced the brain. Breaks, whatever it is, bang, done. And then all that data goes back into your ERP system and makes sure you've, you know, you've reordered any parts that you need to reorder or whatever else happens. So that's that's an example. Another example might be you're executing a system task inside your browser and you're not quite sure how to do it, and you just open the Nobi next to it. And so you have a side-by-side example. Or you don't want to do it yourself, intel Claude to do it for [18:58] Dave Pengelley: That's [18:58] John Ansley: you. [18:58] Dave Pengelley: it, that's interesting. I didn't realize it was a good one. Data capture and and sort of work management as well. I I knew it was sort of the how tos and things. I didn't realize like if [19:05] John Ansley: Yeah. [19:05] Dave Pengelley: you're running through a task that actually, you know, captured, record the steps, record the data, and and you're [19:11] John Ansley: Yeah. [19:11] Dave Pengelley: doing that sort of job tracking piece as well. That's that's that's great. [19:14] John Ansley: Yeah, [19:14] Richard (Syllogism) Webbe: You've got [19:14] John Ansley: so [19:14] Richard (Syllogism) Webbe: the whole [19:15] John Ansley: we [19:15] Richard (Syllogism) Webbe: MCP [19:15] John Ansley: we're [19:16] Richard (Syllogism) Webbe: interface happening too now, haven't you, [19:17] John Ansley: Yeah. [19:17] Richard (Syllogism) Webbe: John? [19:18] John Ansley: Yeah, so we we do. We built so if you think about creating an instruction, we've tried to make it as easy as possible. Um and uh and so we've got you know a browser capture uh plugin which can capture what you've done inside a system and then automatically create a nobi from that. You go in and make sure you're happy with it, publish it, and then you can share it with everybody else. So it takes as long to make a Novi as it does to do the transaction. With [19:42] Richard (Syllogism) Webbe: Yeah. [19:43] John Ansley: with physical tasks, we we uh encourage our customers to video a person while they're doing the physical task and talk about the tasks they're doing while they're buying bootio, upload that into Novi and Novi will then look at that, apply cognitive load theory to it and slice it up into The steps that it needs to be sliced into, strip the audio and turn it into text, and then publish that and it's available as soon as they've approved and published it. It's available in I think 140 plus languages now. Um, which you can then, if you want to hear it, you can just hit the read aloud button and it'll read to you in whatever voice you've selected. So um the the speed with which you can make these things now compares to what it would have taken previous to to Nobi being around is is just Amazing. Like you think to create and and I've been working with customers who are doing this right now with ERP rollouts and they've done you know screenshots and then they've put it into PowerPoint and then they've put in text for there and they've taken all this time and it's taking them, you know, two or three hours to build one transaction. [20:48] Dave Pengelley: Yeah, [20:48] John Ansley: Right. [20:49] Dave Pengelley: I I think about when the guy did the handover for when we got a pool and he was showing me all the like, you know, the filter restart procedures and the cleaning [20:56] John Ansley: Yep. [20:56] Dave Pengelley: and everything, and I'm recording it all, and then you know, six months later when something didn't work. Properly and I couldn't get it to prime. I'm like, oh, where's that video? And I'm [21:03] John Ansley: Yeah. [21:03] Dave Pengelley: back on my desktop and I'm scrubbing through trying to find it. Whereas, [21:05] John Ansley: Yeah. [21:06] Dave Pengelley: you know, sounds like nobody would have taken that kind of exact process with [21:09] John Ansley: Yep. [21:10] Dave Pengelley: him talking about it and [21:11] John Ansley: Yep. [21:11] Dave Pengelley: broken it down to make it searchable, indexed, find exactly [21:14] John Ansley: Yep. [21:14] Dave Pengelley: what I want, see the piece, go, oh, okay, I need to flip that switch and hold that thing and then do that. [21:19] John Ansley: Exactly, exactly. It um and and I've done exactly that for my own house. You know, we've got a solar battery system, I've got a um a triple water filter system, you know, we've [21:28] Dave Pengelley: Mm-hmm. [21:28] John Ansley: got the pool, we've got the spawn. We've got to, you know, all the all the normal sort [21:32] Dave Pengelley: Yeah. [21:32] John Ansley: of crap. And um and all of those things we've got nobies for, which just [21:35] Dave Pengelley: Yeah. [21:35] John Ansley: makes it super easy. So if I'm not here and my brother's staying here or you know, surf surfing in a local area or whatever, or you know, my sister's down from up the [21:43] Dave Pengelley: Yeah. [21:43] John Ansley: coast, they can wander in here and they understand immediately how to run everything in the house. [21:47] Dave Pengelley: Yeah, my wife's got no idea to do anything to do with the pump in the pool, because that's my job, and I d [21:51] John Ansley: Yeah. [21:52] Dave Pengelley: I know that. But if [21:52] John Ansley: Yeah. [21:52] Dave Pengelley: if I'm not here for some reason, [21:54] John Ansley: Yeah. [21:55] Dave Pengelley: um, yeah, that's great. [21:56] Richard (Syllogism) Webbe: You need to ask John how he does it because what the most exciting things about I liked about Nobi when John explained to me the concept and where it's going is time to market from doing it. I mean that this is the groundbreaking bit and you just produce a QR code, don't you John, and stick [22:13] John Ansley: Yeah. [22:13] Richard (Syllogism) Webbe: it somewhere. [22:14] John Ansley: Yeah, [22:14] Richard (Syllogism) Webbe: I [22:14] John Ansley: you can [22:14] Richard (Syllogism) Webbe: think [22:14] John Ansley: do QR codes, you can do URL links. The the the I I think the the two most fascinating things about Novi, and the MCP connector's a really cool piece, right? But the two most fascinating things are the speed with which you can create Novi's, right? Which which is minutes, right? Or seconds even, depending on the the type of nobi you're creating. Um say a five minute video takes about thirty seconds to convert into Anobi, right? So um and a ten second a ten cent ten second transaction in your ERP or your C R M or whatever is converted into Anobi in those ten seconds. It doesn't take any longer than that. But the the the really amazing thing is it's a two day implementation. Right. To go from not using Nobi to using Nobi is two-day implementation, which we we're sort of like, I think, I think with Salesforce who [23:05] Dave Pengelley: Yeah [23:05] John Ansley: advertised [23:05] Dave Pengelley: that's [23:05] John Ansley: themselves this way years and years ago, we're the anti-software software. Right. So we [23:10] Dave Pengelley: that'd [23:10] John Ansley: have we [23:11] Dave Pengelley: have been [23:11] John Ansley: have [23:11] Dave Pengelley: that'd be no software campaign. Um [23:13] John Ansley: yeah. [23:13] Dave Pengelley: because I would try [23:14] John Ansley: Yeah. [23:14] Dave Pengelley: to push people in the cloud and stuff. I I X [23:16] John Ansley: Yeah. [23:16] Dave Pengelley: South Source, [23:17] John Ansley: Okay. [23:17] Dave Pengelley: so [23:18] John Ansley: Yeah. [23:18] Dave Pengelley: yeah. [23:18] John Ansley: So so my thing is I hate the cost and the effort around implementation. Right. So that's the thing that really stings you, apart from the cost of some people's licensing, which is a real pain in the And [23:30] Dave Pengelley: Yeah. [23:30] John Ansley: the hidden secrets in the licensing deals. But this, you know, the idea that if you wanted to go with with Novi for an organization, regardless of the size of the organization, it's a two-day implementation to get you up and going. And then you might do waves where you bring more departments on, but each of those departments is going to be no more than a day or so. So [23:51] John Ansley: we've got things like you determine who your users are going to be, what categories you want to put your Nobis into, what sort of teams you want your use. To be in and uh and then you know give me a whole pile of your content that you want turned into your initial nobies. So give me your existing videos, documents, whatever you happen to have. And using the MCP, I'll get Claude to create all of those nobies for you, and it'll be done in hours, and you'll have a library of nobies ready for you to start using. [24:20] Richard (Syllogism) Webbe: And [24:20] John Ansley: Right. [24:20] Richard (Syllogism) Webbe: what [24:20] John Ansley: So [24:21] Richard (Syllogism) Webbe: what I [24:21] John Ansley: there's [24:21] Richard (Syllogism) Webbe: like about what you said earlier, John, just to counter one of the things you said, if you don't have to classify who does what, they're all available. To everyone anyway, for context, [24:31] John Ansley: Yeah. [24:32] Richard (Syllogism) Webbe: so people [24:32] John Ansley: Yeah. [24:32] Richard (Syllogism) Webbe: on the floor know why they're doing something. [24:35] John Ansley: Yeah. Yeah. Yeah. You've got um sort of three levels of um of access. You've got you can publish stuff to everybody, right? In inside your organization, you can publish stuff to teams. So maybe your payroll team is the only one who knows who needs to know how to run payroll, right? So you don't publish that more broadly. And you can publish stuff publicly. So if you want to put stuff out and you're not worried about protecting IP or anything like that or you want [25:01] Dave Pengelley: Mm-hmm. [25:01] John Ansley: to help your customers understand your products or your suppliers understand how to get into your loading docks or whatever it happens to be, you can push that stuff out publicly. Right. So um is it's really interesting to see some of the businesses are doing all of that. So they've gone absolutely wall to wall, floor to ceiling, and outside their enterprise, which is really cool. [25:20] Dave Pengelley: And then [25:21] John Ansley: And [25:21] Dave Pengelley: is [25:21] John Ansley: others. [25:22] Richard (Syllogism) Webbe: Sorry, [25:23] Dave Pengelley: are [25:23] Richard (Syllogism) Webbe: girl. [25:23] Dave Pengelley: all these nobies indexed and searchable from the point of view of being able to go, I want to run an LLM across them to look for potential workflow. efficiencies [25:30] John Ansley: Yeah. [25:31] Dave Pengelley: and and optimization gains. [25:32] John Ansley: Yeah, so do it again, it depends. If if you've if you've used the category sort of approach, then you can very easily you've got a classification if you like. So some people use categories to be um by job type, some use it to be by equipment type, some by location. You know, so you can use categories. It's a it's it's a tagging system, which you can choose to use any way you want. [25:56] Richard (Syllogism) Webbe: So [25:56] John Ansley: Um [25:56] Richard (Syllogism) Webbe: to create a [25:56] John Ansley: and [25:57] Richard (Syllogism) Webbe: to create a sovereign reg with [26:00] Dave Pengelley: I'm [26:00] Richard (Syllogism) Webbe: I [26:00] Dave Pengelley: just [26:00] Richard (Syllogism) Webbe: think the [26:00] Dave Pengelley: thinking [26:00] Richard (Syllogism) Webbe: problem [26:00] Dave Pengelley: like [26:01] Richard (Syllogism) Webbe: is [26:01] Dave Pengelley: can I [26:01] Richard (Syllogism) Webbe: great. [26:01] Dave Pengelley: can I point my agent at Novi and say I'm looking for areas where we're repeating tasks or doing overly manual things that potentially we could automate. And is that capability available via the MCP or some of some sort? [26:15] John Ansley: Uh uh yeah you could definitely uh you could definitely ask Claude that right now because Claude [26:19] Dave Pengelley: Okay. [26:19] John Ansley: can search through all of your Novi's. [26:21] Dave Pengelley: Yeah, [26:21] John Ansley: Right. [26:21] Dave Pengelley: nice. [26:22] John Ansley: So um so right now um it could you the the the MCP connector allows you to search back into Novi and in fact the way I've set it But whenever I ask Claude any question, it looks to Nobi to see if there's an answer in Nobi first. All right. So it's it's pretty cool. And and you know that that but there's this ability, I I was working with a a very large fashion brand recently and took a 170-page PDF document about um some of their systems areas and turned it into more than 60 different nobies. And it took about an hour. Right. And uh and and then [26:55] Matt: Oh, [26:55] John Ansley: you put those into the library for them in draft and say they're ready for you to go through and review and [26:59] Matt: I'm not sure. [27:00] John Ansley: And publish as you approve them. But just to go, okay, you know, previously we used to give people a 170-page PDF to deal with. Now I've got a fully searchable system in Nobi, which I can start with that as this is the basis for how you do the transaction. We can make improvements, we can do some screen videos, we can do browser capture, we can do whatever you want to make these even better. [27:22] Richard (Syllogism) Webbe: I've got a new slogan for you after hearing that, John. Business at the speed of need. [27:28] John Ansley: Yeah, so it's better than our founders. One, which is um teaching done how to get done. [27:34] Richard (Syllogism) Webbe: Sorry, Matt, you were going to say something, but you've always got the big questions. [27:39] Matt: I've got a few sort of like yeah interesting heavy hitting things. Um so similar to what Dave asked, like can this work in reverse? Where like let me let me paint you a bit of a picture. So like imagine I'm a business like this fashion label, and they have their massive stack of SOPs, you know, they've got all their their documentation, you dump it into Nobi, it's now digitized, it's it's applied cognitive load theory, you know, it's the same thing. It's just digitized and in bite-sized pieces. How do we know we were doing the right processes to begin with? Like how does that company know that their their existing processes are not terrible? You know, it's just as what we've always done. You know, [28:23] Dave Pengelley: Yeah, [28:23] Matt: this [28:23] John Ansley: Yeah, [28:23] Matt: is [28:23] Dave Pengelley: it's [28:23] Matt: just how [28:24] John Ansley: yeah, [28:24] Matt: the business [28:24] John Ansley: honest. [28:24] Matt: works. [28:24] Dave Pengelley: back to that some of the six sigma thinking and and and application of that into the documented processes. [28:30] John Ansley: No, they're a great that's a great question. I mean that we we've done um so one of the things I've built out over the last uh four or five months is um what I'm calling uh organizational memory maturity model, which actually looks at um how you do things and how you share and know-how. So A, how how how do you do it? So how well do you do it? And and then secondly, you know, how well do you share it? How accessible is it? Which which sort of leads to the logical answer that nobody is an answer to your question. But understanding how well you do things is super, super important. And [29:07] John Ansley: the companies who use Nobi the best have always got a business improvement person focused on using Nobi to make those leaps. Right. So a great company in New Zealand who work here in Australia as well, called Ute Master, who do um Ute trays and canopies and all those sorts, toolboxes and all those sorts of things for. Ute's beautiful work that they do and they're they're another company that's super focused on doing things really, really well. And uh and Aaron, who who leads their business improvement side of things there, when he saw nobies like, this is fantastic. I want to do this, right? His biggest his biggest hurdle was how do I get people to start right? How do I get people to start capturing nobies? Um, because we know we want to improve. But once they started to capture those, those nobies become an improvement vehicle where they keep tweaking them and making them better and better. And I do think to to your point Matt I think you do need to have someone who is a you know uh an expert um in business improvement or has a focus on business improvement to really really drive it we we're seeing the same in the in the in the wine industry with customers in the wine industry you know those ones that have that focus they go gangbusters [30:21] Matt: Cause a lot of people ask stuff like, you know, which which roles or which jobs or which things are AI gonna replace. And you know, you always get the the responses about like healthcare, you know, you always need humans in healthcare, and um it sounds like you're saying that that kind of process improvement role within within a business like AI can't help with, and nobody doesn't necessarily help with. You still need that person to to say, you know, we're not doing this right yet. [30:51] John Ansley: Yeah, I I I think I think that there's a couple of answers there. One is most people aren't tracking enough to know how well or badly. they're doing something. And also most companies don't know this there are platforms like Novi available. So they're [31:07] Matt: Yeah. [31:08] John Ansley: not even like they don't even they haven't focused on the problem because they don't know that there is a potential solution. [31:14] Matt: Yeah, [31:14] John Ansley: It's [31:14] Matt: unconsciously [31:14] John Ansley: a really [31:14] Dave Pengelley: Yeah. [31:15] John Ansley: it's a [31:15] Matt: complex. [31:16] John Ansley: Yeah, [31:16] Dave Pengelley: I'm [31:16] John Ansley: yeah, [31:16] Dave Pengelley: I'm [31:16] John Ansley: yeah. So you know you think you're doing it well, so why would you bother looking? And [31:20] Matt: Yeah. [31:20] John Ansley: then when you see somebody doing it better and you go, oh hang on, how are they doing that? And then you find out they're using a tool like Novi, [31:25] Dave Pengelley: not [31:25] John Ansley: then you [31:25] Dave Pengelley: sure. [31:25] John Ansley: go, oh okay, maybe this is part of the answer. But the rest of the the rest of that answer is like when you when you do an organizational uh memory maturity assessment and you see where people score and then you say okay to move from one out of five to three out of five is a massive jump for you in terms of benefit. So how do we how do we go about doing that? And inevitably, you know you're going to have something like we need we need someone focused on this and and it might be somebody's a in a in a role of focusing on improvements, but it you have key people who tend to be with the the Nobi ecosystem, they tend to be an approver for their uh for their department or area of the business. And so they've they've also got a a desire to improve, right? So you've you've just got to get the right, the right structure in place and then you, you know, you you see the improvements come. But it's I I ran one of the groups I ran at at um when I was uh at Linfox, I had a the longest job title I've ever had. I was president of supply chain solutions implementations and CIO. Which meant supply chain optimization, new site development, and all of tech. And in the supply chain optimization area, we had an amazing group, and I think there's 35-ish in that group under a good friend of mine, Chris Hemstrom. And Chris's focus was the value stream mapping part of lean. So go in and find where the wastage was in processes and remove that wastage. So I'll quote you some numbers that I that I that I recall from it that each individual could run 2.3 projects at the same time. The projects were of six weeks, elapsed time. And on average, each one of those projects made a return of about $250,000 to the bottom line annually. [33:21] John Ansley: So this was an incredible piece of work. Some of the smartest stuff I've ever seen. Twelve months later, most of those projects had to be run again if you wanted to keep that saving because they didn't embed they couldn't embed the new way of working because of employee turnover and knowledge atrophy right so when you look at that and you go that's that's where I wanted to plug nobi in right because if you can go in and do that sort of work so if you have that business improvement focus and you have nobi you make the change you improve the Nobi you embed the improvement Right. Um [34:00] Richard (Syllogism) Webbe: Nobody's [34:01] John Ansley: nobody [34:01] Richard (Syllogism) Webbe: like [34:01] John Ansley: doesn't [34:01] Richard (Syllogism) Webbe: a [34:02] John Ansley: have the smells to do that without without the humans, though it needs the humans [34:04] Richard (Syllogism) Webbe: no [34:05] John Ansley: in that [34:05] Richard (Syllogism) Webbe: nobody's [34:05] John Ansley: piece. [34:05] Matt: Yeah. [34:05] Richard (Syllogism) Webbe: like a a Crm for uh for domain expertise, isn't it? [34:11] John Ansley: Yeah, I don't I don't I don't know. I don't know the it's [34:14] Richard (Syllogism) Webbe: Where you are. [34:15] John Ansley: yeah, it's [34:16] Dave Pengelley: Generation knowledge base of of sorts. [34:18] John Ansley: yeah, it is. It's it's the knowledge that people actually want rather than the knowledge that some technical person decided they create. Right. [34:26] Richard (Syllogism) Webbe: That's a good comment. [34:28] John Ansley: Um [34:30] Matt: Actually, does it help track? You said that the the biggest problem is that the people aren't tracking and saving these metrics that they can observe. Like, [34:38] John Ansley: Yeah. [34:38] Matt: does it actually help set those things up too? And it's not just this knowledge base that we're [34:42] John Ansley: Yeah. [34:42] Matt: drawing analogies to [34:45] John Ansley: Yeah, it does. Um so we've we've got a forms feature in there where you can ask any for any input that you you want. So you could be tracking the time they're taking, you could be tracking, you know, the the the um uh results into terms if you can get them to upload an image or a video of what they've done. You could get you could be asking them, um, you know, uh say say it's a weekly safety check and you're checking things on a on a vehicle for for ease, say the a forklift in a warehouse, you'll then have data across that forklift for the entire year if you do a weekly check and you're gonna have the data that says, here's the condition of the seat, the seat belt, the the oil pressure, the engine hours, um The tire where you're gonna have all this information and then you can align that up against all of the maintenance activities you've done. And you can then see how how you're doing in terms of what's working, what's not working, right? [35:41] Dave Pengelley: So [35:41] John Ansley: But [35:42] Dave Pengelley: so [35:42] John Ansley: most people aren't tracking any of that data, so they don't actually have anything they can, you know, that they can compare against. [35:48] Dave Pengelley: in in that sense you're overlapping with some of the other tools coming out, like the safety cultures and stuff, right? Where [35:52] John Ansley: Yeah, yeah, [35:53] Dave Pengelley: whereas [35:53] John Ansley: yeah. [35:54] Dave Pengelley: you've got the that that integrated knowledge breakdown and and that tutorial piece integrated with that knowledge and data cap. Capture and drop tracking pace. [36:01] John Ansley: Yeah. Um yeah, and and with with, you know, I think safety culture is probably a a good one to to look at. They don't go quite as wide or as deep as we do, but fairly, you know, they they're and they've been more checklist-based. So we do checklist if you want checklist. But we can do um we can also do, for instance, uh one of our customers in Quebec and Canada is, you know, one of the most advanced laser laser manufacturers in the world. And when they're building these lasers, there's certain points you get to in the Nobi. Where you're following the detail of how to do something and it's gonna say, okay, now um what what strength laser do you need? You'll go down path A, B, C, or D and click out there and go down to a more detailed um sort of workflow. So, you know, really we're seeing the usage go from lasers and space to um we're on Broadway, right? So we're on Broadway a theater and Broadway just signed up last week. So, you know, um it's it's bizarre. It's like the the variation can be absolutely enormous but the um the the product is built to cope with the complexity [37:08] Dave Pengelley: Yep. [37:08] John Ansley: and [37:09] Dave Pengelley: Well [37:09] John Ansley: again we've we've built a really broad platform as opposed to a a niche [37:14] Dave Pengelley: every [37:14] John Ansley: so [37:14] Dave Pengelley: business [37:14] John Ansley: that [37:15] Dave Pengelley: has operational processes, right? So [37:16] John Ansley: a [37:17] Dave Pengelley: you're [37:17] John Ansley: hundred [37:17] Dave Pengelley: applicable [37:17] John Ansley: percent [37:18] Dave Pengelley: in in every business. [37:19] Richard (Syllogism) Webbe: It's [37:19] John Ansley: yeah [37:19] Richard (Syllogism) Webbe: like the [37:19] John Ansley: we [37:19] Richard (Syllogism) Webbe: strava. [37:20] John Ansley: would [37:20] Richard (Syllogism) Webbe: It's like the strabo of business process with all your data. [37:24] Dave Pengelley: So I was [37:25] John Ansley: yeah [37:25] Dave Pengelley: I was um well I was watching the the Steve Jobs thing and he was talking about this back in ninety two and he With next computers, and one of their big things that he was building was this sort of object-oriented ability to build business apps very quickly. [37:36] John Ansley: Yeah. [37:36] Dave Pengelley: That's why we're talking about the operational thing. And I'll I'll play this clip because he talks around like the speed, and you talked about two-day deployment and why that speed matters. Uh, we'll play this little clip, um, and then we might jump into some news. [38:21] Dave Pengelley: Just lay laying out those maths of that that speed to market. And I think, you know, it ties in the nobi stuff, obviously, very fast deployment, but AI in general, and what we're seeing with businesses that are using AI. are to build new solutions and deploy things. There's this this that next generation gap now of building things is so much faster. And where are you in that in that spectrum? [38:40] John Ansley: Yeah, um so we we've been blessed. We've got a CTO who we've referred to as Mr. 10X for a long time. He he is now Mr. 100 or 200x with with Claude Code. Um he's he's uh you know the capacity we have in our very small technology team, it's effectively two two two people in development roles building. The most amazing platform you can imagine, right? And do you see the speed with which we can do things? It's incredible. And the speed with which we can pivot. So when we need to make a change, maybe it's a technology change we need to make, or um, you know, or maybe it's a slight product shift we need to make or something like that. The speed with which we can do those things, and and I've been in the development game for a long time. I've I've um I've I've started started life as a programmer, and uh, and you know, I've I've built um products for very large companies, I've built the mining solution. For SAP as an example. And when you look at how slow that process is versus how quick this is. And with the mining solution for SAP, it was only a little bit of code on top of an existing ERP just to make it look a little special for the mining businesses. Um when you look at what we've built in the time frame we've built it, it's mind blowing, right? Um it and you know, and the ability to track everything too. So all our real-time tracking of how the platform's running, all the AI in there is insane, right? And you it's You'd never get that. You'd never have a platform uh this this big, this good, this stable without AI. You you just couldn't do it in the time frame. [40:09] Dave Pengelley: Yeah. Brilliant. All right. Well, um, great, great to hear about Nobi, and um certainly something that I think has a lot of potential, and it's great to hear you're having some success there. Like that's that's it's awesome. Uh, we're gonna do this now. [40:23] Dave Pengelley: This news is [40:36] Dave Pengelley: This happened this week. Does anyone understand what this [40:40] Matt: Yep. [40:40] Dave Pengelley: this means? [40:41] Matt: That [40:41] Richard (Syllogism) Webbe: No. [40:41] Matt: is that is Mr. Turnus, who is the new CEO of Apple, getting a hundred million followers in less than an hour [40:48] Richard (Syllogism) Webbe: That's [40:49] Matt: from [40:49] Richard (Syllogism) Webbe: right. [40:49] Matt: one post. [40:52] Richard (Syllogism) Webbe: Yeah, Tim Tim Cook stepped down. I'm probably gonna lose a hundred million followers in less than an hour. [41:00] Dave Pengelley: Everyone's like, oh, what a doing series been so dumb, blah, blah, blah, blah. But like we've seen the whole movement with the Mac minis and AI and running your own models and stuff. They've they're sort of winning the AI war in some ways through a different like they're not winning, that's not a war, but you know, they're getting wins in this AI ecosystem by not being a model provider, by not doing inference, but but by being that hardware platform. I I saw someone saying like OpenAI and Anthropica buying hundreds and and and or thousands of Mac minis to train their model. on computer use [41:31] Matt: Yeah. [41:31] Dave Pengelley: because they can just run simulations on the these Mac hardware things. So Apple's in an interesting place when it comes to this whole AI thing. [41:38] Richard (Syllogism) Webbe: It is. It is indeed. [41:40] John Ansley: Well [41:40] Richard (Syllogism) Webbe: It [41:40] John Ansley: it's [41:40] Richard (Syllogism) Webbe: is [41:40] John Ansley: interesting [41:41] Richard (Syllogism) Webbe: interesting. [41:41] John Ansley: looking at looking at the new CEO's background as well. So that that's uh again a a real a real pivot, right, for those guys. [41:48] Dave Pengelley: To go back to that more very technical-based founder, whereas um Jobs was a visionary and um Cook was a bit more of an operations guy, right? So to go to a technical Based leader. Very interesting as they sort of are evolving into this world. It's funny. In that same um lecture, jobs was people like, How do you feel about not being at Apple anymore? And it's like, oh, they're they're trying to be the Sony, the Sony of uh personal computing, and they're focusing on these consumer things, their top-end sort of um enterprise and and sort of professional stuff is not as good. Talking about some of the things that we're doing there back at the time. Obviously, they've they've pivoted under [42:25] Richard (Syllogism) Webbe: Look. [42:26] Dave Pengelley: under back [42:26] Richard (Syllogism) Webbe: A great [42:27] Dave Pengelley: this [42:27] Richard (Syllogism) Webbe: business [42:27] Dave Pengelley: way with all [42:27] Richard (Syllogism) Webbe: case. [42:27] Dave Pengelley: the studio stuff. [42:28] Richard (Syllogism) Webbe: Yeah, a great business case. Study is actually comparing Sony and Apple and how they changed. And I used to talk about this all the time. Sony nearly went bankrupt because it kept leading with technology, but it didn't win the application or the use discussion. For example, it launched the beta video and VHS one. And VHS one in particular, because the content that people wanted was more available on VHS than on beta. I won't go into the details of that [42:55] Dave Pengelley: Thank [42:55] Richard (Syllogism) Webbe: because I can see [42:56] Dave Pengelley: you. [42:56] Richard (Syllogism) Webbe: that, but you know the reason. And none of us should know. The reason, but we all do. And that went around. And the one thing that saved um saved Sonya, I think, was the Walkman at the time because portable music was coming along. You know, I don't know if you remember, I remember sitting by the radio with the record button on my tape, ready to hit it when the song came on that I liked. I could then take that music with me portably quite easily, as often as I wanted my Walkman. It really saved him. Of course, Jobs was just, I mean, I worked for Apple for a number of years. I threw the CEO, that was the second one after Steve Jobs. Into a pool in Hawaii once, just over a bet. I didn't get fired, just so you know. And basically, the ability of Steve Jobs to marry technology with use and customer business, that's what made him such a unique individual. He straddled across the use case and the technology need and how that fitted in our world. He did it so well. [43:53] Dave Pengelley: Like [43:53] Richard (Syllogism) Webbe: That's all [43:54] Dave Pengelley: it [43:54] Richard (Syllogism) Webbe: that's [43:54] Dave Pengelley: was [43:54] Richard (Syllogism) Webbe: what I [43:54] Dave Pengelley: all [43:54] Richard (Syllogism) Webbe: mean. That's [43:54] Dave Pengelley: yeah, [43:54] Richard (Syllogism) Webbe: what I'm [43:55] Dave Pengelley: yeah, [43:55] Richard (Syllogism) Webbe: saying. [43:55] Dave Pengelley: yeah, yeah. [43:56] Matt: I [43:56] Dave Pengelley: Um, [43:57] Matt: felt [43:57] Dave Pengelley: I mean [43:57] Matt: the uh people always bring up like You know, nobody wanted an iPhone, nobody asked for a phone with no keyboard. You know, nobody wanted any of the features, and then he brought it out and they're like, I want that. [44:09] Richard (Syllogism) Webbe: Yeah. And you know, there's a great there's a great um there's a great bit of psychology that says that none of us can do anything unless we can imagine it in our mind's eye before we do it. So I think I've said this before. If you watch a child and they know they want to walk, but they don't know how to, but they're sitting on their bum, they're looking around, they're watching these adults walk, they're going. I think I want to do that. But they, as soon as they imagine themselves doing it, is when they first start, I think it's called cruising, where they get their feet up and they grab a hold and do their thing and moving along. And Steve Jobs was very good at describing it, but there's a book out. I never got a hold of it, but it was about how TV and TV shows modeled our use cases, and technologists built the use cases. So Hollywood [44:54] John Ansley: Stop. [44:54] Richard (Syllogism) Webbe: came first and the reality came second. And one of them was about Star Wars, uh Star Trek. [44:58] Dave Pengelley: So for yeah. [45:00] John Ansley: Yeah, [45:00] Richard (Syllogism) Webbe: So [45:00] John Ansley: Star [45:00] Richard (Syllogism) Webbe: that [45:01] John Ansley: Trek's [45:01] Richard (Syllogism) Webbe: the whole [45:01] John Ansley: a [45:01] Richard (Syllogism) Webbe: flip [45:01] John Ansley: great [45:01] Richard (Syllogism) Webbe: phone [45:02] John Ansley: example. [45:02] Richard (Syllogism) Webbe: that he had, you know, on the other planet beaming up, Scotty, really is where we probably got the idea of a mobile phone. Because before that, the idea of a mobile phone didn't really gel as much. [45:11] John Ansley: So [45:11] Richard (Syllogism) Webbe: So [45:12] John Ansley: same [45:12] Richard (Syllogism) Webbe: you're [45:12] John Ansley: with [45:12] Richard (Syllogism) Webbe: exactly [45:12] John Ansley: automatic [45:13] Richard (Syllogism) Webbe: right. [45:13] John Ansley: doors, right? Automatic [45:14] Richard (Syllogism) Webbe: Yes. [45:14] John Ansley: doors. [45:16] Dave Pengelley: Where [45:16] Richard (Syllogism) Webbe: I [45:17] Dave Pengelley: they [45:17] John Ansley: That's [45:17] Dave Pengelley: had [45:17] Richard (Syllogism) Webbe: still [45:17] Dave Pengelley: where [45:17] John Ansley: where [45:17] Dave Pengelley: they [45:17] Richard (Syllogism) Webbe: do [45:17] Dave Pengelley: had [45:17] John Ansley: I [45:18] Richard (Syllogism) Webbe: that. [45:18] Dave Pengelley: like [45:18] John Ansley: came [45:18] Dave Pengelley: in [45:18] John Ansley: from. [45:18] Dave Pengelley: interns behind the set pulling them out and they're playing [45:21] John Ansley: I'm [45:21] Dave Pengelley: that little [45:21] John Ansley: making [45:21] Dave Pengelley: gas [45:21] John Ansley: the [45:21] Dave Pengelley: noise. [45:22] John Ansley: noise. [45:22] Richard (Syllogism) Webbe: I still make the noise sometimes. If the lift's too quiet, I've got to make the noise. [45:30] Dave Pengelley: Automatic doors at a building I worked in and and and they um they weren't they were the sort of swing style. And so you know as as you walk up, you do the little little bit of a force push, you just sort of like pretend you're using the force and the door would just open up and you can walk through. [45:43] Richard (Syllogism) Webbe: Oh, [45:44] Dave Pengelley: Um [45:44] Richard (Syllogism) Webbe: we are very nerdy. [45:45] Dave Pengelley: it's [45:45] Richard (Syllogism) Webbe: Hey, [45:45] Dave Pengelley: but [45:45] Richard (Syllogism) Webbe: I've got [45:45] Dave Pengelley: but [45:46] Richard (Syllogism) Webbe: some [45:46] Dave Pengelley: with [45:46] Richard (Syllogism) Webbe: news. [45:46] Dave Pengelley: the with the Apple stuff, um what's what's uh one of the things that really pushed the Apple hardware into relevance in this new AI era was about eight months ago when the open claw thing went bizarre, like just crazy. Um, and you know, OpenClaw two point oh just launched this week. Um, after being quiet for a couple of months, they have since pretty much since Steinberger went to OpenAI, I think, uh, they've released a big new launch. And obviously Hermes has been eating their lunch and Grok and everyone else has been sort of getting into this autonomous agent space. It's funny, um, we talk about uh one of the uh the hype influencers, Alex Finn, sometimes, and he did an an open claw video. And so while I was a mowing the lawn yesterday, I was like, what? What's Alex got to say? Because normally he's like fully hyped about everything, and then every 24 hours he's dropping the last thing that was awesome to talk about the next thing that's awesome. And he came on and went, I'm not here for it. Open claw just breaks all the time. I've tried it and it's not better. Like there's a few nice things around the interface, but I'm sticking with Hermes and Grockbot. There's nothing here for me. And I was like, that's really interesting coming from Alex. [46:52] Matt: Yeah, the user interface as well, like two point oh, you'd think that they would, you know, just coming out of a chat about Apple [46:57] Dave Pengelley: Full [46:58] Matt: and [46:58] Dave Pengelley: refresh. [46:58] Matt: interfaces, yeah, you would think that they Would have spent a bit of time on actually like perfecting it, making it good, making it fun, uh, you know, taking a book out of nobi and actually making it user expressive. Like the is rubbish. Like nobody, I've not seen anybody outside of the people that's been working on it actually saying that it's any good. And I don't, I just think it's a waste of time. [47:18] Richard (Syllogism) Webbe: It is. Look, everything comes to user experience. I remember I was doing a lot of work for Nokia across Europe and China at the time, and uh someone said to me once, oh no one's ever going to Rid of a Nokia phone. We're used to the interface and all the buttons we have to hit to achieve certain things. It seems a million light years ago. And then Apple just came along with the uh the smartphone and just slaughtered it overnight so that Nokia is almost nowhere when it comes to phones. And they were the one that had the idea for the smartphone, but it was the user interface that was much more easier to use and [47:51] Dave Pengelley: Well [47:52] Richard (Syllogism) Webbe: capture [47:52] Dave Pengelley: eve even [47:52] Richard (Syllogism) Webbe: proclaim. [47:53] Dave Pengelley: the Windows phones, the the the the um the Windows smartphones that came out that had the slide out keyboards and I think the jazz jams and the Xperiors and things like that. They were some really great devices that were really good for techie people that liked using Windows on a tiny display. But the ones that had more success were from the HTC, et cetera. They were basically putting their entire own sort of interface layer over the top, but then it was a bit slow because it was a second layer of software over the OS. It wasn't the core OS. Whereas, you know, jobs came out with the iPhone and it was an integrated, smooth OS that just had limitations. And he's he's never apologized for limitations. But the limit, what it did, it did well. [48:32] Richard (Syllogism) Webbe: Yeah, and when 3Com invented their palm pilot, and um uh one of my friends used to be the CEO of 3COM Worldwide. Hello, Matt, if you're watching. And um yeah, not the same Matt. But you know, you could have done a better job than this, Matt. And he wasn't a very good surfer either, John. Just uh you've met him actually, John, who's made of Phil's. But anyway, so what happened is um uh when you look at that, I remember the story about the head of ThreeCom at The time when they were doing a palm pilot and they wanted to break that user experience paradigm. Apparently, either the CO or the designer had a little bit of cardboard that fitted in the top pocket of his suit and the top pocket of his uh shirt where you put those weird pen holders that nerds like us used to have. And every time the development team would come to him and say, Hey, we want to put this on it, he'd go, Great, just pull this card. Will it fit on this format? And if they couldn't answer the question, it was tossed to one side. Next, next. Next. And you know, it's it's a it is, these are tools we use. It's about user experience. It's about how easy it is. I tell people now, I hardly ever open my computer except for this podcast, because AI has changed my user experience so much. And um uh I'll I'll show it in a sec, but um recently I had a major project to work on. So I turned on the voice activation of Claude and started discussing with Claude as I drove down to my beach house down the coast near Bells Beach. Which John knows about. And so I had an hour and a half in the car and I didn't want to waste it and I'm driving and I didn't want to crash. So I just had this verbal conversation. And really, AI has fourfold benefits, we know, but one of them is the user experience and the ability to talk to and interact with video without text. My opinion is emails are going to become a thing of the past in the future. It's all going to be command controlled with our voice, with our eyes, with our movement. And you know, nobody's picked up on that as a use case. But it was I got this project. finished using Claude in an hour and a half by the time I got down. And Claude was great coming back and saying, well Richard, I don't think you can do that, or we've got to look at this format. It was a thoroughly enjoyable experience. [50:41] Dave Pengelley: Speak speaking palm, I got my wife who, like she's clever but not super technical or into techie toys and geeky gadgets and stuff. I got her to have a palm, one of the last palms. It was like a smaller, [50:52] Richard (Syllogism) Webbe: It [50:52] Dave Pengelley: more [50:52] Richard (Syllogism) Webbe: was [50:52] Dave Pengelley: felt [50:52] Richard (Syllogism) Webbe: her. [50:53] Dave Pengelley: palm that was on their new Web OS that they were promoting. This was pre-iPhone, but she loved it. She loved her little palm. She loved having a Full keyboard for messaging your friends, for doing emails and things, having that full keyboard from a user experience point of view, like jobs did a did a great job promoting the fact you don't need a real keyboard. But a real keyboard is so nice to have on a phone. Like I used to love my Blackberry keyboard. I loved my Sony Experia X1 because I loved just the full tactile keyboard I could type so fast on those phones. The touch screen on my iPhone, I still miss the space bar half the time, put extra M's in where I want spaces and stuff because it's Just not the same. [51:32] John Ansley: I I think I think this is a really interesting point though. This is where things are changing, right? And you think about wearables or you think about being able to talk to your phone rather than having to type. And uh my my dad was really techie. He passed away uh about five years ago at 91. But he he had lost most of his eyesight at at that stage. And and he had fully integrated his life with Apple and with Google Assistant, right? So he [51:55] Richard (Syllogism) Webbe: Yeah. [51:55] John Ansley: was just walking through his house chatting to whatever speaker was there, you know. Checking out what the weather was doing back home in Canada where we came from, you know, sending messages to me or my brother or sending emails to friends. He he didn't he didn't touch a device [52:09] Dave Pengelley: Yeah. [52:10] John Ansley: anymore. Right. [52:11] Dave Pengelley: Nice. [52:11] John Ansley: And and that and that, you know, that that that's like I say, that's more than five years ago. So when you look at that, this capability, and we we're playing around with this right now, but when you look at the wearables thing, so what's the next format? Right? And uh I know everyone's spooked about the meta glasses and the [52:27] Dave Pengelley: Yeah. [52:27] John Ansley: ability to have a camera and all that. I mean, you know, let's be honest. spy cameras are out there and perverts have had them for years, right? So it's not the glasses aren't going to change the way perverts are behaving. Um so so um you look at this stuff and you go, you know, these glasses could be really cool. So the the version of the meta glasses in the US that you can get now which also projects a screen in front of you would be a really really cool thing to be able to use and a lot more sensible than some of the smart glasses that have been available in the industrial sense for quite a while. They're not they're nowhere near as heavy, you know, they're [52:58] Dave Pengelley: Yeah. [52:59] John Ansley: they they they're pretty So I I'm I'm really intrigued to see what the next form factor is that really engages with us, right? [53:07] Richard (Syllogism) Webbe: Oh, I I I've got a great use case for that, John. I think you're right on the money. Um, I imagine, you know, we all go to exhibitions to meet people face to face and talk to them. So you imagine you've had a chat to Claude and said, now Claude, we need to talk to Nobi, we need to talk to Slatech Systems, we need to talk to the co-founder of Syllogism. I know they'll be somewhere at this show. I put my glasses on. The facial recognition's happening, key signs are coming up, and it'll tell me, oh, you know, John's over the back corner there, Matt's having a coffee at the and I just make a beeline to them. And you know, for stalkers, that might work a bit better, unfortunately. But I just think the value of connecting people that need to talk and sort [53:48] John Ansley: Mm-hmm. [53:48] Richard (Syllogism) Webbe: issues out and sort problems out and process things is going to accelerate the speed of how we run the world. [54:00] John Ansley: glasses and I don't use my computer anymore. You know, I I did this experiment. Um uh I think the first time I had a real lash at this was uh 2014. Uh I took over as the toll global CIO and I went in there and they uh they brought a a laptop out and gave me my brand new laptop and I said no take it away. I want everything to work from a note three. So I just um so I I I just had a note three that I would plug into an adapter on my desk and that was my computer. And um and you know [54:31] John Ansley: I think we need to continue to push for those sorts of things. Like, you know, I've still got a beautiful MacBook and I've got an iPad and I've got a phone and I've got a backup Windows PC if I need one. You know, I think it's but yeah, exactly. But imagine imagine just having a pair of glasses and you've got everything, you've got everything stored, basically everything stored in the cloud anyway. Why aren't we all moving to this new form factor? You know, how [54:55] Richard (Syllogism) Webbe: I [54:55] John Ansley: much [54:55] Richard (Syllogism) Webbe: think we're [54:56] John Ansley: grunt [54:56] Richard (Syllogism) Webbe: rushing [54:56] John Ansley: do we [54:56] Richard (Syllogism) Webbe: it. [54:56] John Ansley: the average person doesn't need, you [54:58] Richard (Syllogism) Webbe: Yeah, [54:58] John Ansley: know? [54:58] Richard (Syllogism) Webbe: I think we're actually [55:00] Dave Pengelley: The [55:00] Richard (Syllogism) Webbe: John. [55:00] Dave Pengelley: conference wants one once these ones [55:01] Richard (Syllogism) Webbe: Thank [55:01] Dave Pengelley: by the [55:02] Richard (Syllogism) Webbe: you. [55:02] Dave Pengelley: sound [55:02] Richard (Syllogism) Webbe: Thank [55:02] Dave Pengelley: of [55:02] Richard (Syllogism) Webbe: you. [55:02] Dave Pengelley: it. [55:03] Richard (Syllogism) Webbe: And the steel structure for my bones as well, please. Thank you. [55:07] Dave Pengelley: Just [55:08] Richard (Syllogism) Webbe: John [55:08] Dave Pengelley: a little person [55:08] Richard (Syllogism) Webbe: and I. [55:08] Dave Pengelley: identification and a little red text. [55:12] Richard (Syllogism) Webbe: John [55:12] Dave Pengelley: Target [55:12] Richard (Syllogism) Webbe: and I were [55:13] Dave Pengelley: identified. [55:13] Richard (Syllogism) Webbe: exactly well. That's but you know, that's what it's like, right? You imagine we're talking about AI changing sales. We're talking about getting rid of the noise, like a thousand cold calls or people reaching out or getting spam everywhere. New technology in ways are gonna um we bumped into Gus first. Ferguson the other night at the BHP IT exec dinner, John. And it was good to see all the old characters there. Some of them retired, but a lot of them, as you said, very smart characters that we worked with in the old days. I remember when I was working for um, I was working for Retix and they had helped develop the OSI layer model for communications. That's what it's about. It's all about the seven-layer model, about connecting together. And I remember Gus coming up to me. I was the custom, I was the vendor then, and John, of course, is my customer. Gus was was my customer and and he said, Richard, why can I email you down the hallway on the 32nd floor here of the BHP building? But I can't send an email to that mine that John's working in, you know, over in the US. And I had no answer for him. And then of course we've worked to break that down. And then once you could email people anywhere in the world or do a voice call, it just changed the way we work. So it's it's it is new ways of working. And as more we imagine new ways of working, the more these end devices are going to accelerate. And change the way we do things. [56:31] Dave Pengelley: The [56:31] Richard (Syllogism) Webbe: I [56:31] Dave Pengelley: the [56:31] Richard (Syllogism) Webbe: think this is really exciting. [56:32] Dave Pengelley: important thing, and and as we discussed with the optimization stuff, is keeping a human at the helm. I mean, if you go full augmented reality and you see all these jokes and memes on the internet around people in job interviews and and both both sides are asking the chat GPT, what should I ask the person? You should ask them what their day was like. Oh, well, if they asked me what my day was like, what should I ask them? Answer, you know, with this sort of [56:51] Matt: Oh, [56:52] Dave Pengelley: and if you've got the augmented reality and everyone's just reading their little scripts off their eyes and you can't see anything. And and I've seen some of these um dystopian vision things that were made sort of Of decades, like a decade or more ago, where without the augmented reality, the world is just all like checker boxes and plates and things and holding patterns for where the actual content goes. And so uh [57:10] Richard (Syllogism) Webbe: Yes. [57:10] Dave Pengelley: very, very black mirror-ish that that kind of one went. Um, so I think I think it's important that we make sure humans are still empowered to be at the helm and they can get augmentation and support, but that we're not completely sort of like making them colour by numbers all the time by by presenting this thing through interfaces that sort of force that, right? [57:30] Richard (Syllogism) Webbe: Segue because I'd like to now bring up that uh how to screw up selling, thanks to you, Dave, has now been published and you can get it on Amazon under my name. [57:38] Dave Pengelley: Do it. [57:39] Richard (Syllogism) Webbe: Thanks, [57:39] Dave Pengelley: Do it. [57:39] Richard (Syllogism) Webbe: Dave. I appreciate [57:40] Dave Pengelley: Um, [57:40] Richard (Syllogism) Webbe: all your efforts. [57:41] Dave Pengelley: wrap up quickly with a few last things on the news. You mentioned Claude, Claude Fable 5.1 and Mythos 5.1 are just out, and I'm sure none of us have really had time to play with them yet. [57:52] Richard (Syllogism) Webbe: I saw [57:52] Dave Pengelley: Um [57:52] Richard (Syllogism) Webbe: that this morning. [57:54] Dave Pengelley: I'm not gonna play with them. I don't have a Claude license, so um, I I won't be won't be doing that. Um Any have you heard anything about a Matt so far about the new [58:03] Matt: Yeah. [58:03] Dave Pengelley: models? [58:04] Matt: Yeah, it's it's an improvement. Um it's more expensive, even though it hits a higher case ratio. Um it's just a meaningful improvement. So it'd be very interesting to see if it's as arrogant as it used to be, where you know, [58:17] Richard (Syllogism) Webbe: Yeah. [58:18] Matt: the idea of telling the user that you know they're they're wrong. Um and not in the sense of opposite to Sicker Fancy, whatever that is, but um just being disobedient that's that's just the general kind of vibe that I've [58:33] Dave Pengelley: I [58:33] Matt: been [58:33] Dave Pengelley: can't [58:33] Matt: hearing [58:33] Dave Pengelley: let [58:33] Matt: from [58:33] Dave Pengelley: you [58:34] Matt: a [58:34] Dave Pengelley: do [58:34] Matt: lot [58:34] Dave Pengelley: that, [58:34] Matt: of people. [58:34] Dave Pengelley: Dave. [58:35] Matt: Yeah very very similar. [58:38] Richard (Syllogism) Webbe: Open [58:38] John Ansley: So [58:38] Richard (Syllogism) Webbe: the [58:38] John Ansley: the [58:38] Matt: Um [58:38] Richard (Syllogism) Webbe: hatch. [58:38] John Ansley: the advertising, the the little spiel that came out, uh so when I opened up claw this morning, the spiel that I I saw is go to Fable, it's cheaper. Right? [58:47] Matt: yeah [58:47] John Ansley: So it's less [58:48] Matt: so [58:48] John Ansley: expensive to run. [58:49] Matt: I I mean with your experience John you'd know that anything that any company says is all laced in marketing [58:54] Dave Pengelley: Huh. [58:54] Matt: so [58:54] John Ansley: So [58:55] Matt: um [58:55] John Ansley: hey, [58:55] Dave Pengelley: Huh. [58:55] John Ansley: that's one thing I didn't get to say earlier. So we we only have one rule and I know this rule applies in football teams and other places and doesn't apply very well apparently in some of them. But our rule at at Noby is don't be a dick. Don't be a dick to our customers. Don't be a dick to our shareholders. Don't be a dick to our [59:09] Dave Pengelley: Yeah. [59:09] John Ansley: team members. And it's second slide of every board presentation. And uh and we check in all the time and people have been moved on for being dicks, right? [59:18] Matt: Yes. [59:18] John Ansley: So you think you think about licensing. Licensing is really key to me. And I hate people who who do this and you could name any of the large Technology vendors, and they've all done it. They hide things in a pricing engine that you don't understand and you get an absolute sting in the tail. [59:37] Matt: So [59:37] Richard (Syllogism) Webbe: Oh, [59:37] John Ansley: And [59:37] Matt: there's a [59:37] Richard (Syllogism) Webbe: you [59:37] John Ansley: it [59:37] Richard (Syllogism) Webbe: should [59:37] Matt: massive [59:37] John Ansley: just [59:37] Richard (Syllogism) Webbe: come [59:38] Matt: thing [59:38] Richard (Syllogism) Webbe: and [59:38] John Ansley: pisses [59:38] Richard (Syllogism) Webbe: see [59:38] Matt: with [59:38] Richard (Syllogism) Webbe: the taxation [59:38] John Ansley: me [59:39] Richard (Syllogism) Webbe: system [59:39] John Ansley: off. [59:39] Richard (Syllogism) Webbe: in Victoria, mate. There's a lot in many of taxation. [59:42] Matt: There's [59:43] Dave Pengelley: Yeah, I mean [59:43] Matt: like huge, [59:43] Dave Pengelley: it's [59:44] Matt: huge uh like literal like class action lawsuits right now against Claude with regards to their subscriptions. You know, like they say, oh, it's it's 20x your [59:53] Dave Pengelley: a are be going lot not of people to who able do that. to [59:53] Matt: just generic plan. Well there you go. Thanks, Dave. Um where it's not uh at all. And [01:00:00] Dave Pengelley: Just the five hour sessions, you're getting twenty X on the five hour, but not the overall week. [01:00:04] John Ansley: Mm-hmm. [01:00:05] Matt: Yeah, so all of that kind of just sneakiness is just rubbish. Like it's such a shame that the the general vibe of anthropic as a company is so not [01:00:15] Dave Pengelley: Misanthropic. [01:00:15] Matt: very good. Yeah, when [01:00:17] Dave Pengelley: Huh. [01:00:17] Matt: with like you can't you can't argue that that [01:00:20] Richard (Syllogism) Webbe: So [01:00:20] Matt: favor [01:00:20] Richard (Syllogism) Webbe: you did [01:00:21] Matt: part [01:00:21] Richard (Syllogism) Webbe: then. [01:00:22] Matt: Yeah, that's you don't live on the internet, uh Richard, that's why you haven't heard that yet. [01:00:27] Dave Pengelley: No, no, [01:00:27] Matt: Um [01:00:27] Dave Pengelley: no, that's my that's my original map. What are you talking about? [01:00:30] Matt: That's [01:00:30] Dave Pengelley: Steal [01:00:30] Matt: literally [01:00:30] Dave Pengelley: that. [01:00:30] Matt: Elon. Um, [01:00:33] Dave Pengelley: Uh Elon hates everyone. He's not happy with that open AI either because they pulled out a cursor. There's a lot of bits of news we haven't taught touched on, but um, yeah, open AI have gone, you know what, we are 20x. We don't have a five-hour limit on our pro plans anymore. And so when you say 20x, that's 20x the pro plan or 5x. Like so, they've they've come out swinging against Claude. The the general market sentiment is like from a lot of the the technical people, is open AI is on a tear and Claude's going downhill. That's [01:01:00] Richard (Syllogism) Webbe: Mm-hmm. [01:01:00] Dave Pengelley: what [01:01:00] Matt: Yeah. [01:01:00] Dave Pengelley: I say. [01:01:01] Matt: It's a real shame because like Fable five point one is still like I would say still the best model in the world, like as far as like just model capabilities. Um my my exposure to Fable, I don't do an anthropic subscription, but I've got cursor and I'm probably gonna pick up Devon. Uh you guys probably forgotten about Devon. Um the their their actual their harness, [01:01:23] John Ansley: It's a sandwich [01:01:23] Matt: like [01:01:24] John Ansley: mate, isn't it? [01:01:25] Matt: sorry. [01:01:26] Dave Pengelley: That's [01:01:26] John Ansley: It's [01:01:26] Dave Pengelley: a sandwich, [01:01:26] John Ansley: a sandwich [01:01:27] Dave Pengelley: mate. [01:01:27] John Ansley: mate. [01:01:27] Matt: Yeah. Yeah, no, D E P [01:01:30] Richard (Syllogism) Webbe: The [01:01:30] Matt: I [01:01:30] Richard (Syllogism) Webbe: tan [01:01:30] Matt: am. [01:01:30] Richard (Syllogism) Webbe: in the staff of England. [01:01:33] Matt: Yeah, anyway, the that that is where I'm probably gonna go to be able to test and get exposed to all the different models still. [01:01:41] John Ansley: What if this this is a question my my wife asked me? So we're building up all of this knowledge and capability inside these models we use. So what Claude knows about my business right now, what you know um what ChatGPT and Gemini and others know about others others' businesses, how do I make that transferable? What if I decide I want to jump out of Claude and I want jump into somewhere [01:02:00] Dave Pengelley: Mm-hmm. [01:02:00] John Ansley: else and should I run a home AR um LLM, a home-based LLM where I actually dump all of my knowledge into rather than trust it in the hands of these bastards who [01:02:11] Richard (Syllogism) Webbe: Right. [01:02:11] John Ansley: will probably do something nefarious. Right? So so [01:02:14] Matt: Sounds [01:02:14] John Ansley: where where [01:02:15] Matt: like [01:02:15] John Ansley: do I, you know, what happens if if if if anthropic blows up and all my knowledge is sitting in there, what happens? [01:02:21] Matt: it kind of sounds like Dave that um it sounds like they need a like a centralized like place for all of the non [01:02:30] Dave Pengelley: Like a knowledge core. [01:02:31] Matt: Like a like a [01:02:31] Dave Pengelley: Umledge [01:02:32] Matt: cool a cool location. [01:02:34] Dave Pengelley: core. [01:02:35] Richard (Syllogism) Webbe: Call [01:02:35] Dave Pengelley: If [01:02:35] Richard (Syllogism) Webbe: the Brian. [01:02:35] Dave Pengelley: only they they [01:02:36] Richard (Syllogism) Webbe: Go [01:02:36] Dave Pengelley: they had [01:02:36] Richard (Syllogism) Webbe: to [01:02:36] Dave Pengelley: like a [01:02:37] Richard (Syllogism) Webbe: Brian. [01:02:37] Dave Pengelley: uh a uh something that knows knows bees things about their business and stuff. Um no so there's there's there's two aspects that uh uh and we we we could do a whole nother episode just on this, but I think depends on which surface you're using when it comes to your Claude and your chat GPT. If you're just on Claude.com or chatgpt.com and it's all in chat sessions on the left frame of the left rail, your knowledge is all. Captured and trapped in that platform. If you're using Claude [01:03:02] John Ansley: Exactly. [01:03:02] Dave Pengelley: Code and all your knowledge is in markdown files and text files and code on a folder on your desktop that clawed code's interacting with, [01:03:09] John Ansley: Yeah. [01:03:10] Dave Pengelley: all that's on your desktop. You can throw [01:03:11] John Ansley: Yeah. [01:03:11] Dave Pengelley: any other agent and model of that at any point in time, and that's all yours. Um [01:03:16] John Ansley: I [01:03:16] Dave Pengelley: because [01:03:16] John Ansley: I I think [01:03:16] Dave Pengelley: that's [01:03:17] John Ansley: there's that there's a huge percentage of people who aren't using Claude Code, they're using Claude.com, right? And they are pushing all of this information and it's they've got a great history in their chats, an amazing history in their chats. If they lost that Tomorrow that'd be devastated. [01:03:31] Dave Pengelley: I've [01:03:32] Richard (Syllogism) Webbe: I [01:03:32] Dave Pengelley: I went [01:03:32] Richard (Syllogism) Webbe: had [01:03:32] Dave Pengelley: through [01:03:32] Richard (Syllogism) Webbe: trouble. [01:03:32] Dave Pengelley: that with chatgpt.com myself, especially when I deleted one of my project folders and actually just manually burned a lot of conversation history. But that's why, you know, I'm I'm primarily running Hermes harnesses now, which means I can plug any model in at any point in time to assess across it, and all the data is living in files and folders on my machine. So the context is always mine and it's always there, and I can throw any LLM at it that I like. [01:03:57] Richard (Syllogism) Webbe: Dave, you and I spoke a while ago about Exporting your knowledge base out of your LLM. And I was having trouble getting it out of Claude, but I got some of it out and I dumped it into Chat GPT as an update. So then my idea to answer what sorry, you go on a sec, mate. My answer to you, John, is a cheap one, but it's a retail one, and that is I'll run two LLM accounts and I'll share the data across both of them as a backup, kind of like a company has two telecommunications providers. What we're going to say, Matt, you've always got A better answer for these things. [01:04:31] Matt: I have the literal answer here. And it's treat everything as ephemeral. If it needs to be saved, save it somewhere deliberately. Don't [01:04:40] Richard (Syllogism) Webbe: Yeah. [01:04:41] Matt: just treat like a web app or a desktop app as something that is, you know, never gonna go away. Treat [01:04:47] Richard (Syllogism) Webbe: Yeah, [01:04:47] Matt: it, [01:04:47] Richard (Syllogism) Webbe: yeah. [01:04:47] Matt: treat it as if they're about to ban you and they're about to delete your account. You know, how do you keep yourself protected in that sense? So, [01:04:54] Richard (Syllogism) Webbe: Yeah, [01:04:54] Dave Pengelley: Yeah. [01:04:55] Matt: me, [01:04:55] Richard (Syllogism) Webbe: like [01:04:55] Matt: like I personally use something that I built and I have everything that I have inside. I kept using the word core because I've built this thing called the core, and all of my [01:05:04] John Ansley: Not [01:05:04] Matt: stuff [01:05:04] John Ansley: the [01:05:04] Matt: is [01:05:04] John Ansley: goal. [01:05:05] Matt: in there. Yeah, [01:05:06] Richard (Syllogism) Webbe: No, [01:05:06] Matt: C-O-R-E. [01:05:06] Richard (Syllogism) Webbe: the core, the core. [01:05:08] Dave Pengelley: Yeah. [01:05:09] Matt: Yeah. And um it it you need to think about that because like you see people all the time uh you know saying that their account got suspended and deleted and they've like gotten it refreshed, but all the stuff's gone. You know, all that kind of stuff is very relevant. What about if it's not, you know, something that's excusable in that sense? What if it's actually Nefarious? What if you now have a bit of a ransomware situation where you know the the speed of development, as you were saying earlier, John, is analogous, is parallel to the speed of malicious attacks as well. You know, the actual speed of people finding exploits in people's systems. So you need to be able to treat your system as if you can take to it with an axe and then nothing is lost because you have it saved in a certain way. [01:05:58] Dave Pengelley: Yeah, it's all Business continuity disaster recovery stuff, which in this new age people haven't fully thought about. And when I was back using Chat GPT 12 months ago as my exclusive tool, I started realizing this and started taking, oh, that's a really important artifact that I don't want to lose. And it's so hard to find in the chat bar on the left. I'm going to start putting stuff out of Notion. And then I started building like a little indexing of I got this from Notion from this chat thread with a URL. So I can go back to keep working on it. And then over the last 12 months, I've obviously evolved and moved systems back and forth and And so on. But yeah, if there's things you care about, you should have a backup of some capacity. Um, and not just especially if it is a web based system that you log into and pay for, because who knows when, you know, some antitrust lawsuit will come and and [01:06:43] John Ansley: Exactly. [01:06:44] Dave Pengelley: all of a sudden [01:06:44] John Ansley: Exactly. [01:06:44] Dave Pengelley: they go away. [01:06:45] John Ansley: Yeah, I don't I think the biggest issue is that everyone's lazy. Right. So you gotta find an easy way to do it. [01:06:51] Dave Pengelley: Well, [01:06:51] John Ansley: You could [01:06:52] Dave Pengelley: you'd [01:06:52] John Ansley: you know, [01:06:52] Dave Pengelley: you'd [01:06:52] John Ansley: and [01:06:52] Dave Pengelley: know [01:06:52] John Ansley: that's [01:06:52] Dave Pengelley: from better [01:06:52] John Ansley: that's [01:06:53] Dave Pengelley: CIO [01:06:53] John Ansley: the [01:06:53] Dave Pengelley: background, like disaster recovery, [01:06:54] John Ansley: yeah. [01:06:54] Dave Pengelley: who actually wants to spend time working on your on on your business continuity plan? Like no one. Because ah, [01:06:58] John Ansley: Yeah, [01:06:59] Dave Pengelley: you'll [01:06:59] John Ansley: exactly. [01:06:59] Dave Pengelley: be right. Probably won't happen, but [01:07:01] John Ansley: Yeah. [01:07:02] Dave Pengelley: then it does. Um yeah, we might we might uh wrap it up there. I was gonna finish with wins of the week, um, briefly, but we're out of time. But you know, on the open AI side, they keep it in the reset button, which is crazy. Um, to the point they're saying, should we should we do a merch line calling ourselves the reset company? Um so on the other side of the fence, they're not necessarily perfect at all. But like I on Monday woke up to find out that they were gonna do a reset in three hours. And I was like, had 80% of my week's quota left. And I managed to burn that down to 20 something by the time they hit reset. And then 24 hours later, I'm back down to 30. So in like 24 hours, I did like 150%, like a whole week and a half's worth of work through uh through codex, which is just wild, which is why yes, yesterday I was I was like mentally smoked. I went, I'm gonna go do some like walking and mowing and stuff because I can't think about all my projects because I blitzed it so hard on Monday. Um but [01:08:00] Richard (Syllogism) Webbe: I'm [01:08:00] Dave Pengelley: Yes, [01:08:00] Richard (Syllogism) Webbe: getting [01:08:00] Dave Pengelley: which [01:08:00] Richard (Syllogism) Webbe: I'm getting sad for all your notes on a Sunday night or a Friday night saying I've run out of tokens. I'll have to mow the lawn this weekend. [01:08:10] Dave Pengelley: just uh it's just you know to to tokenomics, just getting the maximum out of my subscriptions while they last, right? [01:08:17] Dave Pengelley: Uh well thank you, John. Been a pleasure having you. Good to good to hear about Nobi and and how that's sort of doing things and how that interacts with the whole what we all what we're on about here. So uh very good. Thank you, Matt. Thank you, Richard, as always. Great to have you here. And um remember, I always forget to do this earlier in the show, but uh like and subscribe, drop comments, um, help the algorithm see us. Keep an eye out on the channel, not just for the live show, but for the weekly field notes segments that we're putting up as well. So keep an eye out for those. We are slowly getting there with the content. And if you want to see what these episodes look like as a comic book, make sure you're following the AR Operators Podcast page on LinkedIn where we post the Weekly comic book summary of the episode as well. So uh look out for that one, John. We'll tag you when we get the comic up once my AIs process and transcribe and draw some pictures for us. [01:09:10] Richard (Syllogism) Webbe: Let's see how John looks like a comic. [01:09:14] Dave Pengelley: Indeed. All right. Well, thanks everyone. Thank you, gentlemen. We'll see you next week.