AI Made Developers Faster. So Why Does Work Feel Harder?

[00:00] Dave Pengelley: Start of the hot intro, like because you get the countdown on the restream, the 54321, and then I start talking, or we're just talking while it's happening anyway. And then when you watch YouTube, there's that slight buffer delay where it actually loses the first sentence or two anyway. So it's always interesting when I get the download the the transcript for um watching it. Where did we actually come in? Like what half hot sentence do [00:21] Matt: On [00:21] Dave Pengelley: our viewers [00:22] Matt: my [00:22] Dave Pengelley: get? [00:22] Matt: like there's a little live uh box that pops up on my end and uh that always does take a little bit of time to pop up, so it could be then. [00:30] Dave Pengelley: Yeah. Yeah. It's lots like the the bumpers at the end that always used to cut off before the bumper finished. That's why now I have this long hold where we just sit here waiting for like ten seconds where I'm like, okay, and now I'll cut it and then we'll come back and we'll just do a little debrief at the end. So Annie, just as a note, as a guest, when we finish, just if you want to stay on camera for a few minutes and we'll just we'll wrap up afterwards. Don't hang up. Some guests go. And oh the show's finished. Hang up. And it's like, Where [00:52] Annie Vella: And [00:52] Dave Pengelley: do they [00:52] Annie Vella: they got [00:52] Dave Pengelley: go? [00:55] Annie Vella: sure. [00:56] Dave Pengelley: Um But great, great to have you on the on the show, Annie. Um welcome. [01:00] Annie Vella: Thank you. Thanks for inviting me. It's fun. [01:03] Dave Pengelley: You're uh you're our first international guest. [01:07] Annie Vella: Ah. Well, [01:08] Dave Pengelley: You are [01:08] Annie Vella: privileged. [01:08] Dave Pengelley: not yeah, I mean, technically, I mean, my I'm from Perth and and my friends and family say that Perth's basically another country. And we've had we've had guests from Perth, and we've got Mano and Kalgooli. So people say WA needs a passport, but you're you're genuinely international. You're New Zealand-based, right? [01:27] Annie Vella: I am, yes, I'm in Auckland. And I'm grateful that you think of New Zealand as actually another country, not just, you know, a part of Australia, because uh happens as well. Um yes, I'm in Auckland. I've been here for about five and a half years. I am I am a Kiwi. My dad's a Kiwi, but I've moved around the world an awful lot. Um, grew up moving around because my dad worked for the New Zealand Embassy as a [01:49] Dave Pengelley: No [01:49] Annie Vella: child. I [01:50] Dave Pengelley: all [01:50] Annie Vella: had no option but to, you know, pick up and move every three or four years. And then I got used to that, so I kind of carried it on into adulthood. [01:57] Dave Pengelley: enough. Nice, nice. And um What what I mean you're you're in between gigs at the moment, right? But [02:04] Annie Vella: I [02:04] Dave Pengelley: but what [02:04] Annie Vella: am. [02:04] Dave Pengelley: were you doing most recently as far as, you know, work and AI [02:08] Annie Vella: Yeah. [02:08] Dave Pengelley: and so on? [02:09] Annie Vella: Right. So about three weeks ago, I finished up at Westpac, New Zealand, where I I was there for the whole five and a half years that I've been back in New Zealand for. But for the first two years, I was a tech area lead, which is kind of like a senior engineering manager, I suppose, so big team and And then I moved into a distinguished engineer role, which was a lot of fun. Um, I was the first distinguished engineer at the bank. I helped them realize that the staff engineering career path is is a thing. Um, [02:41] Dave Pengelley: Yeah. [02:42] Annie Vella: I read uh Will Larson's book, Staff Engineering Path or Staff Engineer or can't remember the title now, but brilliant book, really short, easy to read. And I felt like I was reading, like, where has this book been all my life? And you know, this this is what I should be doing. Um, so yeah. Distinguished engineer. And um, you know, that I guess I started off in 2003. Well, by which time ChatGPT was already out, but early days. And um at the end of 2003, I started thinking I really needed to go back to university and study the impact of AI on software engineering. Because I've been a software engineer for an awfully long time, since I was about six years old, and I got my first computer, Commodore 64, and I fell in love with it, and [03:27] Dave Pengelley: Wow. [03:28] Annie Vella: basically don't think I could have done anything else with my life. So yeah, I saw that the potential for this technology to really shift how things work in I and well, particularly in software engineering. And I thought, oh, always wanted to do a master's, been promising myself that I'd go back and do one for 20 years. So this is it. So I did that part-time over the top of the the job. [03:47] Dave Pengelley: Well, [03:47] Annie Vella: And um it was a lot, honestly, it was it was quite a lot to do both of those things. And at the same time decide maybe this is the time to learn how to do some public speaking because [03:57] Dave Pengelley: that's [03:57] Annie Vella: I've [03:58] Dave Pengelley: like [03:58] Annie Vella: got interesting things to say now and that Felt like a third job over the top. So yeah. Um and yeah, we can go into more detail on the sort of things I was doing, but that's that's me basically. [04:09] Dave Pengelley: And I think I think I've seen in when I was at Salesforce as well, organizations realizing they need a career path for individual contributors that don't want to manage, that probably shouldn't manage people. Um, but [04:21] Annie Vella: Yeah. [04:22] Dave Pengelley: how do you give them that same level of progression and promotion? And uh they [04:25] Annie Vella: Exactly. [04:25] Dave Pengelley: had like I was in pre-sales, so the solution engineers or whatever people call them. But they um it originally it was, you know, you went principal and then the last one was master. And then there was like, oh no, you can't talk about masters because they've got all these slave connotations. Like [04:39] Annie Vella: Oh [04:40] Dave Pengelley: so [04:40] Annie Vella: yeah. Oh yeah, [04:41] Dave Pengelley: I think [04:42] Annie Vella: true. [04:42] Dave Pengelley: they ended up going distinguished or something similar as well, because of like, oh we can't use the word master. I'm like, oh everything's a loaded topic. [04:49] Annie Vella: I actually [04:50] Dave Pengelley: But [04:50] Annie Vella: met uh at my first fellow recently, three of them actually, three women, one working for IBM, another one for Northrop Grumman, and another one for LIDOS or something, American company. Um anyway, yeah, so that's the step beyond distinguished. You become a fellow, it's uh [05:07] Dave Pengelley: Jol Jolly Goodfellow. [05:08] Annie Vella: yes, indeed. Yeah. [05:10] Matt: I've never heard any of these terms. [05:14] Dave Pengelley: So that's what corporate world does to you, everything stacked [05:17] Annie Vella: Yeah. [05:17] Dave Pengelley: and ranked and hierarchied and [05:19] Annie Vella: That's [05:19] Matt: That's crazy. [05:19] Annie Vella: right. [05:20] Matt: Like I was doing engineering like in another life. Uh I was at uni doing mechatronics, which is you know like the control systems systems like in general. Kind of funny where I've ended up. And I quit it after six months. Like I loved the content, but I was like, I can't do this, man. Like it's just too much. Like the the whole even in uni, like the corporate kind of stuff that comes out of it. I just wanted to get my hands dirty and just do things. Um, so I don't know. I think I sometimes think if I did go, if I had have stuck with it and gotten that full bachelor's, would I be doing anything different? And I honestly think no. I just would have more stuff in here. [06:01] Annie Vella: Well, [06:01] Dave Pengelley: Yeah. [06:01] Annie Vella: you know, [06:02] Dave Pengelley: Sorry. [06:02] Annie Vella: oh um the the corporate thing. So um I've been working for 23, 24 years now. It feels like forever and still forever to go. But um it's only the last five and a half years that I've worked at a like a more corporate y sort of environment, right? So a bank is one larger in size, so way more people that you'll never meet. Uh and um second, yeah, a regulated environment. So very different to where I had spent my time previously, whereas I've generally worked at startups and scale-ups before that. And the difference is quite massive. There's there's almost a bit of culture shock between the two. I had to see what it was like. I never tried it, so why not? And um it's different. It's very different. I think some people really thrive in one or the other. Switching between the two takes a bit of effort. [06:54] Matt: For sure. What what was your actual masters that you did? [06:58] Annie Vella: I didn't masters of engineering. It was a research-only project and it was really just looking at the impact of AI on software engineering. I had four broad, very broad research questions. My poor supervisor was constantly like, you need to narrow it down. I'm like, yeah, I I get that, but it's such a broad topic. I really, really, I mean, I'm doing it because I want to understand the the broad impact that it might have and it really is having such a broad impact. But yeah, I really wanted to understand how it was Shifting like where we might spend our time, because I feel like where we spend our time and effort and focus is that you know, that's where we build skill. So if we're now not writing code by hand so much anymore, does that mean we're gonna lose the ability to do that? And depending on where we shift our time to, maybe we are going to build more skill in that area. Um, then also the impacts on productivity and developer experience, on self-efficacy. And I had a question around whether prompt engine. Engineering might become the new core competency. And that didn't age very well because by the time I was publishing anything, it's like, hmm, no one's talking about prompt engineering anymore. It's uh it's gone through so many phases. It went moved on to context engineering, harness engineering, loop engineering, graph engineering. Like, yeah, gosh, prompt engineering was so 2023. Yeah. Um, so yeah, it was a very perception-based um, you know, professional software engineers. do they feel about it? [08:31] Dave Pengelley: I th I think that's where I found your content on LinkedIn. It was around that um engineering experience stuff, and and we even brought up one of your articles and and talked about that whole in the short term they have some feelings, but in the long term, like the breaking them out of the flow and all that kind of stuff [08:45] Annie Vella: Yeah. [08:45] Dave Pengelley: that you were talking about. [08:46] Annie Vella: Yeah, and it's it it's it's not a it's not an answered question. I I I think there's still many open questions in this area. I mean just today I have literally had three other conversations. Telling Dave, I got okay, so I've I've quit my job so that I could have a bit more free time, and instead I'm just spending all my time having calls that are really interesting, so I'm yeah, doing more of it. But um, yeah, in all three cases, we talked about that developer experience. And I mean, does developer experience even matter anymore? Is what does it change? Are we are we measuring the wrong things now? You know, no one really knows. [09:25] Matt: We we talk about this stuff all the time and it's it's so much fun. Like I think you're a perfect guest for This. Um, because like something that happens with my brain as we talk is I get so many questions piling up and we never get through them. So it'll probably sound like jarring when I just come out with this random stuff that you've said like 20 minutes ago. Um yeah, the the thing that I am like stuck on it when when you talked about prompt engineering, that's something that we've we've discussed at length here in the sense that prompt engineering. Engineering basically is like it's just communication skills. You know, it's [10:04] Dave Pengelley: And [10:04] Matt: that [10:04] Dave Pengelley: and [10:04] Matt: like. [10:06] Dave Pengelley: I I I was always a big fan of prompt inception as well, going these LLMs are pretty good at writing their own prompts. And so I would just, you know, do the brain dump and say all the things I'm thinking about, and then I would finish it with now turn that into a prompt to get the job done. And then it would give me a prompt that I go open a new thread and then drop that in. Like you said, like that was back in the old chat GPT days. Nowadays, with the the the new agents and the context and the skills and everything. Is way less relevant. [10:31] Annie Vella: Yeah. [10:32] Matt: Hmm. And then the other [10:37] Matt: thing about like the developer experience, you know, do you still need to write code in order to be good at the things that we're now focusing on? You know, I kind of go like deeper than that. Like, what fundamentals do you actually need to be a good engineer while you're writing code? And then let's say you're not writing code anymore, but you still have those fundamentals that then try translates really well into this next era where you can actually speak properly, prompt engineer properly, and use the right terminology as far as debugging and asking questions and going deeper, as opposed to the people that just get a chord subscription and all of a sudden start building things with no engineering fundamentals. You know, like there's a term, another term we use which is technical literacy. You know, like there's a there's like a few bars on the rung that you kind of need to meet. Um But yeah, it's fascinating, and I I I'm looking forward to diving deeper into more questions with you. [11:35] Annie Vella: Yeah, well, you know, I think what when I had a bit of a realisation that because I've got I had one foot in industry, I feel like I probably still do and and will have, but um what's new to me is having one foot in academia. 'Cause in doing when you sign up to university again in your forties, uh after having had quite a career behind you, they don't really treat you as A student, they treat you more as a peer. At least that's how it has felt for me. So um I have [12:08] Annie Vella: two uh one last year and one this year, two events that I traveled to o internationally with a whole bunch of academics who came from all around the world. Got one more later this year in Japan. And um so I've been in rooms with like full of academics, um, professors, PhDs, and me, and like huge imposter syndrome. Like, who am I to be in this room? Like you guys are all such experts. [12:33] Dave Pengelley: Because you've actually [12:34] Annie Vella: Um [12:34] Dave Pengelley: done stuff like to [12:35] Annie Vella: But [12:35] Dave Pengelley: be [12:36] Annie Vella: yeah, exactly. They that's you know, I after I get over that feeling of not fitting in, I realize that I I do add some value in bringing like this is what industry's actually doing, not maybe what they think they should be doing, but what's actually [12:52] Dave Pengelley: a good one. [12:52] Annie Vella: happening. Um so in those rooms, and then likewise earlier this year I was really privileged to go and attend um a future. A software development retreat in Utah in the same mountains as the Agile Manifesto was signed in 25 years ago. It was hosted by ThoughtWorks and Martin Fowler in particular. So that that room was filled with like technical leaders from around the world, people like Martin Fowler or Patrick Dubois or Kent Beck and these sorts of folks. So um, no matter what room I'm in, nobody has all the answers. Like everyone's got more questions than they do have answers. That like, what are this guy? [13:30] Dave Pengelley: So [13:30] Annie Vella: Skills that we [13:30] Dave Pengelley: I [13:30] Annie Vella: need [13:30] Dave Pengelley: think [13:30] Annie Vella: to [13:30] Dave Pengelley: it's [13:30] Annie Vella: be [13:31] Dave Pengelley: going [13:31] Annie Vella: teaching [13:31] Dave Pengelley: to [13:31] Annie Vella: students [13:31] Dave Pengelley: be [13:32] Annie Vella: these days? What what [13:33] Dave Pengelley: one [13:33] Annie Vella: skills uh will erode? Does that matter? What, you know, what is the [13:37] Dave Pengelley: of the [13:38] Annie Vella: role of a software engineer? Is it dead or is it just evolving into something else? What happens to all the adjacent sort of, you know, like the architects and product owners and testers and [13:48] Dave Pengelley: same [13:48] Annie Vella: BAs? Like it, [13:49] Dave Pengelley: things. [13:49] Annie Vella: you know, these are not vacant spaces for software engineers to suddenly claim more of. So and do software engineers even want that? And why did people get into software engineering in the first place? Some Some because they really enjoyed the craft. Is that now disappearing? Oh gosh, so many questions and yeah, no answers. [14:06] Dave Pengelley: Yep. That's great to know because we we we challenge and and ask all those things on here all the time. Like, oh, you see the articles, is SAS dead and and this and build versus buy and all the different things and and creativity and taste and and where does the AI live and where where do you need to still have the human in the loop and all these different things? And I think it's just such an evolving space. I mean, you mentioned your masters having like being a bit vague and broad, but that's because This is such a fast moving tiger, right? How can [14:32] Annie Vella: Yeah. [14:32] Dave Pengelley: you lock it down? Like you said, prompt engineering was huge two years ago. [14:36] Annie Vella: Yeah. [14:36] Dave Pengelley: And now you just wouldn't even say that out [14:40] Annie Vella: Yeah. [14:40] Dave Pengelley: loud, probably. [14:41] Annie Vella: That's right, yeah. It's all about or at least the context, if not the harness that manages the context that generates the prompts, and it's like Russian dolls in a [14:50] Dave Pengelley: Yeah, [14:50] Annie Vella: way. [14:50] Dave Pengelley: because I mean I I I w can't even remember where I heard this. I was watching something or listening to something and they talked about ideas and thought. And how does humans How does humans do that? Because if you know things, you know them. But then you get that little idea, that little voice that connects the dots for you. But like I already knew all the things. So what was the mechanism that made my brain decide right then to pop that out as an idea or to link those two dots? Because then the knowledge was all there. What is happening in the human brain that the AIs will arguably never do, depending on where we where we end up with synaptic networks and so on, that you know, makes us human, that we have knowledge, but how do we make the knowledge into ideas and thoughts um that gives us that humanness that that and and I I I talk about all the time on this this show [15:37] Annie Vella: Okay, [15:37] Dave Pengelley: that we have ambient context because we're alive, we're 24 seven, we're in the world, we've lived 10, 20, 30, 40, 50, 60 years. Um we just know stuff. And our brain has this giant network of knowledge that at random times we'll just connect the ideas together and hear something go, oh bang, bang, bang, bang, bang. New idea, new thought. And that's humans. [15:59] Annie Vella: I'm not sure. [16:00] Matt: You you had me at random times. [16:05] Matt: That that sort of thing is is so true. And like I think it was last week or the week before, like I started talking about like what are decisions? You know, like what even is a decision if you look at it chemically, mechanically, like what actually is a decision? It's the same thing that you're coming, you're just saying. So like how that how we Yeah. Well, like how we assume that these these agents that show some sort of level of intelligence. you know, using these better and better LLMs, different harnesses, different controlling of context layers and stuff, we sort of expect them to have that same intuitive grasp on decision making. And then we just get angry at them when they can't do it. Like it's such a funny thing. And I'm curious, Annie, with your experience and like your sort of research into how it affects the developer, have you looked at that aspect of it? Like the increase of, I don't know what you call it, like Stress or anger towards your tooling. [17:04] Annie Vella: I have felt it myself. And I've also had conversations with people around, you know, are we all just becoming managers of agents now? So that is it like an equivalent type of role to being a people manager? And some people have some pretty strong reactions to that, like absolutely not, or yes, very much so. And okay, you know, there's no right or wrong answer here. It's it's how people might feel about it. But one thing that I think does differ massively is you better not be talking to your human, you know, team the way that you end up talking to your agent because I know when I get frustrated with it and it used to happen more. And it you know, it to be fair, the models as they've gotten better and the harnesses as they've gotten better, you know, like the they they are better at doing more or less what you want them to do. It feels more and more like they're reading your mind. But yeah, there was a while there was like capital letters, I told you already to You know, and you would not want to be learning those traits and then applying it to your human um [18:05] Dave Pengelley: And [18:05] Annie Vella: team members. [18:05] Dave Pengelley: I I think I think my experience on that rage stuff um was and welcome Richard to the show. Good to have you here. Um mate. Better late than never. Um [18:16] Dave Pengelley: But [18:16] Annie Vella: Good, [18:17] Dave Pengelley: I [18:17] Annie Vella: thank you. [18:17] Dave Pengelley: I think um I think sorry, Richard, we're just running hot in on topic. We're we're g we're great. Um my my thing was the the context compression was the [18:27] Richard Webbe: Hey, [18:27] Dave Pengelley: issue [18:27] Richard Webbe: hi [18:28] Dave Pengelley: where [18:28] Richard Webbe: guys. [18:28] Dave Pengelley: that wasn't [18:28] Richard Webbe: Hi [18:28] Dave Pengelley: particularly [18:28] Richard Webbe: Matt. Hi [18:29] Dave Pengelley: clear in [18:29] Richard Webbe: Annie, [18:29] Dave Pengelley: the early [18:29] Richard Webbe: how are you? [18:30] Dave Pengelley: Days of your chat GPTs and stuff that you were getting this context compression. And so you'd hit your window, and then it was you know the photocopy of a photocopy syndrome, where all of a sudden it's like just running in loops because it didn't know it wasn't didn't even know how to intelligently compress context the way them the newer models and agents work. And right, so you'd just be like, This is the best thing ever, and we're progressing forward and writing some code and we're making ideas, and now you just can't do the most basic thing. And you tell me you're doing it and we're running in circles, and you say, I've done it, and you're like, No, you haven't, [18:59] Annie Vella: Yeah. [18:59] Dave Pengelley: you've changed. Nothing. [19:01] Matt: Matt Pocock used to call it the dumb zone. And uh [19:06] Richard Webbe: Anyway, [19:06] Matt: like [19:06] Richard Webbe: all [19:06] Matt: you [19:06] Richard Webbe: right. [19:07] Matt: hit a certain level of context window, and all of a sudden you're in the dumb zone, and you basically need to handoff. Where he had this full mechanical process of like, okay, I'm hit that. I'll I would use a skill called handoff, it would generate this summary. I would then start a new thread with that handoff summary, and we're going again. And he would just do that constantly. [19:27] Dave Pengelley: I was doing the same thing back before I I Discovered um, you know, but before Claude Code became popular and those sort of desktop harnesses were as much. I was deep into chat GPT, and I'd have the same thing. I'd get that feeling. I'd be like, we've been going for a while at this. I'm like, how is your window going? Should we be uh summarizing this and moving to a new thread? And then we go, no, we've got a little bit more. I was like, no, now's a good time. Now's a good time. And I'm like, cool. Give me the prompt with a summary of where we're up to for starting a new thread and give me the page, and I'd go paste it in and not lose. [19:56] Annie Vella: The thing is, like m maybe it's it's a good nudge for us to think about how that also happens with humans, right? Like if you overload someone with too much information, especially somebody who's maybe recently joined your team or company, and you're like, okay, we're going to do some onboarding. And then you just, you know, dump a whole bunch of information on them. There's only so much they're going to take in and retain. And you know, like somehow we're we're more accommodating of the agent and we we understand that there's maybe techniques to do exactly as you were just saying, Dave, like, you know, summarize it off and start up and you Yeah, maybe maybe in that sense it should be forcing us to stop and think about how we how we take the same concepts and apply it to to our human agents as well. [20:39] Matt: Same concept exists with like SOPs. You know, [20:42] Annie Vella: Yeah. [20:43] Matt: if you have a you're onboarding someone into the team and you're like, okay, you need to read through this whole file stack manual thing of all of the business SOPs and sign off because you can't start anything until you've signed it off. You know, that that kind of thing is crazy for people. But like the the benefit. Of that is that it's tracked, right? Like it's a thing, this is the repeatable process that everyone needs to do, and this is how we know that things are getting done right. You can apply that same logic to your agentic systems where you can actually have your SOPs, you have your run books, and you have your checks in between. And that's just another way of like managing that context. So each one can have a lower context window each [21:23] Annie Vella: Yeah. [21:23] Matt: time. [21:24] Dave Pengelley: To just [21:24] Matt: Um [21:24] Dave Pengelley: to just clarify, mate, you're saying SOPs, not SOBs, right? When we're talking about yelling at agents. [21:30] Matt: Basically, [21:33] Dave Pengelley: Thank you, Sean, friend of the show. Uh I can't yell at my team, my agents are more forgiving. But yeah, you're talking about standard operating procedures. [21:38] Matt: it's a good idea. [21:39] Dave Pengelley: But um yeah, I [21:40] Annie Vella: Yeah. [21:40] Dave Pengelley: mean, even even and Annie, you talk about uh and research into the whole developer experience thing. I know when I'm, you know, running three, four, five, six different VS Code windows on different projects in running parallel tasks, you can only do that for so long [21:56] Annie Vella: Yeah. [21:56] Dave Pengelley: before you go, I'm overwhelmed and I can't make the decisions, I need to shut half. And just focus on one or two of them because it's just too much. Like I think um [22:04] Annie Vella: Yeah. [22:04] Dave Pengelley: humanly we have capacity and we can we can push through a a a certain capacity level for a little while, but it's not sustainable. [22:13] Annie Vella: And [22:14] Matt: I wonder [22:14] Annie Vella: it's [22:14] Matt: if [22:14] Annie Vella: a [22:14] Matt: that's [22:14] Annie Vella: different [22:14] Matt: how [22:16] Annie Vella: Now [22:16] Matt: sorry. [22:16] Annie Vella: you go. [22:19] Annie Vella: Oh, well, I was gonna say, um so developer experience is has been defined by some people. So the the DevX framework that I used in my research. Was a somewhat simple one with just three dimensions, you know, feedback loops, cognitive load, and flow state. And I figured those were three things that an average person, developer or otherwise, can kind of understand just as base concepts, right? Um, and and and I know you've already had a quick look at this, but it flow state was the one that was most negatively affected. [22:53] Dave Pengelley: Yeah, [22:53] Annie Vella: And I kind of [22:54] Dave Pengelley: you have [22:54] Annie Vella: you [22:54] Dave Pengelley: to make [22:55] Annie Vella: sort [22:55] Dave Pengelley: a second. [22:55] Annie Vella: of think it's kind of the opposite of the feedback loop, right? Because The the quicker you're getting feedback, the the more interrupted you might feel like you're getting when the agent's responding with this giant wall of text that you now have to stop and read. And because we can't help ourselves, I'm waiting now for maybe a few seconds or a minute, I'm gonna spin up another loop and ask it another question while this one's working. And now this one's also kind of stalled. So I'm gonna spin up another one. And now they're all coming back with these giant walls of text and markdown isn't really the easiest way to read. And you know, I've experimented with having it generate. HTML instead, [23:30] Dave Pengelley: Mm-hmm. [23:31] Annie Vella: which is a little nicer to read, but then like that that didn't really stick. It's not fast enough. And like so the very nature of it having this faster feedback loop might be what's causing us to feel like the uh flow states I'm gonna hear myself. That's weird. [23:50] Dave Pengelley: I I don't [23:51] Annie Vella: Why is that happening? Are we good? Okay. Um yeah, flow states deteriorating and cognitive load, everyone felt like Well, this this tooling should help with cognitive load. It should simplify things for us. But instead, what we've ended up with is feeling like actually, I'm now having to make so many decisions. There's no downtime. There's no that toil actually maybe gave us a little bit of a rest in between harder tasks, more, you know, consequential decisions we might need to make. And now that's been automated away, and we're having to spend all of our time making more difficult decisions. And yeah, you can last like three or four hours. Like Steve Yeeggy talks a lot about. This in his um AI vampire blog post, which um if you haven't read it, was is a really good one because I think he went pretty hard pretty early and felt the the consequences of maybe spending too much time with um the bots. Um and it's like a slot machine, right? It's like you you're pulling the the the lever and sometimes a really good outcome uh happens and sometimes it's not, so you just keep pulling on it. And yeah, people say that that's a bit more like dark flow rather than The good sort of flow that you want. So yeah, this you can go down so many rabbit holes with this. I can see how you guys have a lot to talk about. [25:08] Matt: It's true. I love that concept of dark flow. That that really hits home because that's that's that state where if you don't realize you've been doing it and you kind of slip into that zone, all of a sudden you get this weird fog in your brain. Like you sort of start to regress. I think you regress in intelligence. Like I my ability to make decisions all of a sudden is just so swamped and I'm but if you get it back into that good zone then all of a sudden you're you're firing and everything's going back on all cylinders again. [25:42] Annie Vella: And I I think part of that is um other people talk about comprehension debt or cognitive debt or cognitive surrender, you know, because if you're not stopping and really thinking about like engaging the the brain and those neurons we're talking about firing before if If you're kind of delegating that off to the bots as well, and you're not utilizing it yourself, you're not really understanding things the way that you used to when you were doing it a little slower by hand, making those decisions, you know, with a bit of time in between and all of that. Um, but the consequences of that cognitive debt won't show up for a while. So we're we're kind of we are taking out a loan on something. We don't know when we may need to repay. It may never happen, we may not ever have to. have you know be able to fix something ourselves without the help of an agent. But but you know, you're still accountable for every line of code that you write. And this is what businesses are telling their engineers. Go faster, use the tools, but you're still accountable for every single line of code that you produce and you commit. And that's putting a lot of pressure on engineers to go, well, do you want me to go faster? Or because that means I'm losing control. Like I I can't drive the sports car twice as fast and still see the scenery the way that I used to when I would drive slower, right? Like it You [27:00] Matt: Yeah. [27:00] Annie Vella: there's a trade-off there, and yeah, you you you feel it in you, like it it's a it's an emotion, I think. It it's yeah, [27:08] Dave Pengelley: I talked about that when I first went to Hermes and I tried spinning out everything into individual robot profiles and had like 30 agents and they were all making decisions and they were asking me for feedback [27:19] Annie Vella: I a good it's think very thing. [27:19] Dave Pengelley: on the decisions. And I'm like, I don't even know what you're doing. I can I don't have the cognitive context to even understand where you're at to give you an informed decision. So this is your idea, you're working on it, just work it out. Like just keep going until you give me a result. Stop asking me for decisions on things that I never actually fully owned or asked you to do. Like it's it is this weird thing of this handoff and [27:56] Richard Webbe: I think we spoke about that last week, didn't we? I mean, you you learn [27:59] Dave Pengelley: a it's not [27:59] Richard Webbe: when [27:59] Dave Pengelley: a bit little of little [27:59] Richard Webbe: you [27:59] Dave Pengelley: bit [27:59] Richard Webbe: write things out by pen, you learn a certain way. When you see them on a television or by video, you learn a certain way. And it can be the same thing you're learning, but there's multiple media ways of of learning and covering that off. I find um that we are going as a human race, we're going to lose some. We've lost our skill to write properly. Uh, I can't hardly read my own writing anymore because it's always a keyboard. And that's now getting worse because I'm doing speech to text. I mean, I I told someone the other day I haven't opened my laptop up except for today for about a week because I'm doing everything on speech to text on my phone with some LLM off-site processing uh and media handling at you know at a cloud somewhere. And so you you do lose. You lose. Certain skills, but you move quicker. But as we said, as you leave those uh cognitive contexts behind, like I don't need to think about how I turn the key on my car to start it, I just push a button now. Whereas in the old days, you know, on a cold morning you turn the key a certain way, so your car started, that all goes. And now we step into [29:07] Annie Vella: Yeah. [29:07] Richard Webbe: a whole [29:08] Matt: Richard, [29:08] Richard Webbe: new realm. [29:08] Matt: yeah, you just [29:09] Richard Webbe: Uh, [29:09] Matt: said you [29:09] Richard Webbe: and [29:09] Matt: just said [29:09] Richard Webbe: uh, [29:10] Matt: something [29:10] Richard Webbe: you know, [29:10] Matt: about [29:10] Richard Webbe: we [29:10] Matt: like [29:10] Richard Webbe: talk [29:10] Matt: you [29:10] Richard Webbe: about [29:11] Matt: can [29:11] Richard Webbe: sovereign [29:11] Matt: move quicker. [29:11] Richard Webbe: independence [29:12] Matt: And I want [29:12] Richard Webbe: of [29:13] Matt: to know, [29:13] Richard Webbe: intelligence, [29:13] Matt: Annie, did your ears [29:14] Richard Webbe: and [29:14] Matt: perk [29:14] Richard Webbe: that [29:14] Matt: up [29:14] Richard Webbe: is, [29:15] Matt: to that? [29:15] Richard Webbe: do I [29:15] Matt: Because [29:15] Richard Webbe: know enough [29:16] Matt: like [29:16] Richard Webbe: about what's [29:16] Matt: is [29:17] Richard Webbe: under [29:17] Matt: there [29:17] Richard Webbe: the hood [29:17] Matt: there's a difference [29:17] Richard Webbe: to recover [29:18] Matt: between moving [29:18] Richard Webbe: if the [29:18] Matt: quicker [29:19] Richard Webbe: System [29:19] Matt: and [29:19] Richard Webbe: I'm using is [29:20] Matt: like [29:20] Richard Webbe: failing, and [29:20] Matt: the [29:20] Richard Webbe: I think [29:20] Matt: illusion, [29:20] Richard Webbe: that's what we're talking [29:21] Matt: the illusion [29:21] Richard Webbe: about. [29:22] Matt: of [29:22] Richard Webbe: But [29:22] Annie Vella: Yeah. [29:22] Matt: speed? [29:22] Richard Webbe: yeah. [29:23] Matt: Like yeah, [29:24] Annie Vella: Yeah. [29:25] Matt: yeah. [29:25] Annie Vella: Yeah, [29:25] Matt: Where [29:26] Annie Vella: absolutely. Cause I mean, this is I think on on the minds of every organizational leader on the planet that you know they've read the headlines, the promises of the productivity [29:37] Richard Webbe: Yeah. [29:38] Annie Vella: gains. And then you know you can go down rabbit holes of what do we mean by productivity? Like define productivity. That it's it's not even a well-defined term and then try and measure it. Often we come up with terrible proxies for measuring productivity. So then we get stuck in that corner of the universe whereas um I think you know what organizations Really need to stop and think about is that it's a socio-technical system. At the end of the day, everything's a system. I love systems thinking now. It just makes the world a much easier place to reason about systems within systems within larger systems. And so [30:16] Matt: So [30:16] Annie Vella: with a [30:16] Matt: we've got [30:16] Annie Vella: system, you've got [30:17] Matt: technical. [30:17] Annie Vella: many moving parts that interact with one another. What we've done is apply AI into some, not just writing of the code, but maybe also the generating of requirements or generating of tests or something like that. Little bits and pieces. So we're amplifying those areas and increasing in a sense the variety of those areas. But then it just creates the next bottleneck if you don't consider the system as a whole and the interactions between those components. And so when you zoom out and look at the overall efficiency or productivity gains of this technology on a company as a whole, you might not be seeing that much because yeah, you've you've improved the productivity of one small Area, but you've just now created a new bottleneck somewhere else, and there's a lot of pain being felt there right now and [31:11] Annie Vella: that's right. [31:20] Richard Webbe: It's the any, you're right, so right. It's sequential thinking versus parallel process thinking. I used to sell once upon a time supercomputers many years ago, and they came out with vector processing versus parallel processing. And at the time my tiny brain couldn't work it out, and I had one of the scientists explain to me, Richard, what we do is this. And you know, we set up a task and then we run it very quickly because we know what that task is, as opposed to do a bit here, hand the answer back here, process [31:48] Annie Vella: That's [31:48] Richard Webbe: that. [31:48] Annie Vella: right. [31:49] Richard Webbe: best metaphor [31:49] Annie Vella: Doesn't [31:50] Richard Webbe: for describing [31:50] Annie Vella: work. That's [31:50] Richard Webbe: the confusion [31:51] Annie Vella: right. [31:51] Richard Webbe: that [31:51] Annie Vella: Yes, [31:51] Richard Webbe: comes [31:52] Annie Vella: I [31:52] Richard Webbe: with the [31:52] Annie Vella: heard [31:52] Richard Webbe: way we [31:52] Annie Vella: that [31:52] Richard Webbe: approach [31:52] Annie Vella: one before. [31:53] Richard Webbe: things is [31:53] Annie Vella: It's [31:53] Richard Webbe: you can [31:54] Annie Vella: a [31:54] Richard Webbe: make [31:54] Annie Vella: good [31:54] Richard Webbe: a baby [31:54] Annie Vella: one. [31:55] Richard Webbe: with [31:55] Annie Vella: Um [31:55] Richard Webbe: a woman in [31:56] Annie Vella: I've [31:56] Richard Webbe: nine [31:56] Annie Vella: had [31:56] Richard Webbe: months. [31:56] Annie Vella: this conversation multiple [31:57] Richard Webbe: You can [31:57] Annie Vella: times [31:58] Richard Webbe: make [31:58] Annie Vella: just [31:58] Richard Webbe: one, [31:58] Annie Vella: in the [31:58] Richard Webbe: but [31:58] Annie Vella: last [31:58] Richard Webbe: you [31:58] Annie Vella: few [31:58] Richard Webbe: can't [31:59] Annie Vella: days. [31:59] Richard Webbe: make [31:59] Annie Vella: Uh [31:59] Richard Webbe: a [32:00] Annie Vella: And [32:00] Richard Webbe: baby [32:00] Annie Vella: I was reminded [32:00] Richard Webbe: in [32:00] Annie Vella: of [32:00] Richard Webbe: one month [32:01] Annie Vella: I don't know [32:01] Richard Webbe: with [32:01] Annie Vella: if [32:01] Richard Webbe: nine [32:01] Annie Vella: you guys have [32:01] Richard Webbe: women. [32:02] Annie Vella: read the book, [32:02] Richard Webbe: It [32:02] Annie Vella: Wiring [32:02] Richard Webbe: just doesn't [32:03] Annie Vella: the Winning [32:03] Richard Webbe: work. [32:04] Annie Vella: Organization, written by Gene Kim, Stephen Spear. So it's a good book. Um, I'll see if I can find it. Wiring the Winning Organization. So it came out probably three years ago, and it didn't talk about AI at all. Maybe it was four years ago at this point. Um, but what [32:21] Richard Webbe: No, [32:22] Annie Vella: they talk about is how you know organizations are [32:24] Richard Webbe: I'm [32:24] Annie Vella: socio [32:24] Richard Webbe: not [32:24] Annie Vella: technical [32:24] Richard Webbe: gonna [32:25] Annie Vella: systems, and they go further to explain that there's sort of layers within an organization. Organization. So three layers. Layer one is around the technical [32:34] Richard Webbe: go. [32:34] Annie Vella: bits that you that your company produces. So it might be software, it might be physical parts, whatever that's a the purpose that the thing that you produce. Layer two is the tooling that you use to produce those things. And layer three is the social circuitry of the organization. So the processes, the procedures, the norms, the routines, the collaboration, the communication. And when you Break it down that way, you start to think about how if you go and uh put a new tool into the tooling layer and don't change anything else around it, well, you're only going to get limited benefits from that, right? And if anything, it's going to put additional pressure on upstream and downstream on those two layers. I think that's [33:22] Annie Vella: exactly. Yeah. So it really forces you to think about if you want to see these outcomes, these better outcomes, and I think It also makes people have to think [33:32] Richard Webbe: It's [33:32] Annie Vella: about [33:32] Richard Webbe: like putting [33:32] Annie Vella: what [33:32] Richard Webbe: in [33:32] Annie Vella: is [33:32] Richard Webbe: a [33:32] Annie Vella: the outcome [33:33] Richard Webbe: it's like putting [33:33] Annie Vella: that [33:33] Richard Webbe: in [33:33] Annie Vella: we're [33:33] Richard Webbe: a freeway [33:33] Annie Vella: after, [33:34] Richard Webbe: with [33:34] Annie Vella: like [33:34] Richard Webbe: that [33:34] Annie Vella: how [33:34] Richard Webbe: enough on ramps [33:35] Annie Vella: that's [33:35] Richard Webbe: and off [33:35] Annie Vella: what [33:35] Richard Webbe: ramps. [33:36] Annie Vella: we need to be trying to measure, but that's a harder conversation to have, right? Like, okay, so what's the purpose of our company? Is everybody on the same page about that? And what are the outcomes we're chasing right now? And how do we measure those outcomes and then start working back from that? What do we need to change about how we do work our processes and procedures so that the amplification by AI and the tooling layer can help, etc. etc. It's a really Big transformational conversation. [34:01] Dave Pengelley: It's it's the Jurassic Park thing. It's um everyone was got so so excited um thinking about whether we we could create dinosaurs. No one started to think maybe whether we should. Um because like like AI enables you to do all these things. So people are, I need to chase all these things and do all these different things. And it's like, is that your core business? Is that actually going to drive the most value in your business? I I've been talking to a few businesses recently where they've got these ideas for things they want to run off and do, and they've been inspired by things I've seen online. And I look at it and go, yeah, that's a different country and everything's a little bit different here. And that's is that core to what you're doing? And what about your existing systems and things you're doing? Like, is there an argument you should focus on accelerating and improving the things you're already running versus trying to build all these side things that you've been inspired that you can do because AI is there, but should you do them? It's that um, you know, is it performative or is it performant? And they they sound like similar words, but they're completely different. [34:58] Annie Vella: That's a good point. [35:00] Dave Pengelley: Um yeah, it's it's it's pretty wild. I I there's uh but I think the general, you know, um uh zeitgeist people are getting there. I think people are understanding that all this AI like like people like us have been saying this for for for a long time around, you know, make sure you you're just doing things that drive value, not chasing the shiny objects and doing things. But you know, there's the category of people like us that get that, and then there's decision makers that don't get that as much. I saw this little video, not yours, Richard, but we'll we'll play your video next. But I I I saw this little funny clip that I thought illustrates what we're talking about pretty well. [35:49] Richard Webbe: Yeah, so [36:06] Richard Webbe: I'm not sure. [36:33] Richard Webbe: Great, great, good. [36:56] Matt: And all the people there with noteboys going, Hm, yeah, this is [36:58] Dave Pengelley: Leveraging [36:59] Matt: good. [36:59] Dave Pengelley: AI. Be leverageable, that's [37:00] Richard Webbe: Oh, [37:01] Dave Pengelley: uh [37:01] Richard Webbe: I love that. [37:01] Dave Pengelley: insightful. [37:01] Richard Webbe: I'm I'm I'm wondering how many boardrooms with [37:05] Annie Vella: Come [37:05] Richard Webbe: uneducated [37:05] Annie Vella: back to you on [37:05] Richard Webbe: senior [37:06] Annie Vella: that. [37:06] Richard Webbe: leadership people are talking [37:07] Dave Pengelley: Yeah, [37:07] Richard Webbe: like [37:07] Dave Pengelley: yeah, [37:07] Richard Webbe: that without [37:08] Dave Pengelley: like [37:08] Richard Webbe: having [37:08] Dave Pengelley: so [37:08] Richard Webbe: a [37:08] Dave Pengelley: I think [37:08] Richard Webbe: friggin' [37:09] Dave Pengelley: you know, there's the [37:09] Richard Webbe: clue. [37:09] Dave Pengelley: the the sat the satirists get this, um, but you know, there's still that executive leadership layer that probably [37:15] Richard Webbe: Exactly. [37:16] Dave Pengelley: are not getting it. [37:17] Annie Vella: And [37:17] Dave Pengelley: Um, [37:18] Annie Vella: you know, uh at at that level, so in the the role that I had at Westpac, I was um focused on AI governance [37:26] Dave Pengelley: I a good that's thing. think [37:26] Annie Vella: for the last year, year and a half, and It was actually quite a different topic to AI and software engineering, which is what I was studying, right? So I was it they [37:37] Dave Pengelley: I [37:38] Annie Vella: they [37:38] Dave Pengelley: think [37:38] Annie Vella: do [37:38] Dave Pengelley: we've [37:39] Annie Vella: interact [37:39] Dave Pengelley: tried to end [37:39] Annie Vella: and there is [37:40] Dave Pengelley: up [37:40] Annie Vella: governance [37:41] Dave Pengelley: that's [37:41] Annie Vella: on [37:41] Dave Pengelley: a [37:41] Annie Vella: the software [37:41] Dave Pengelley: lot more [37:41] Annie Vella: engineering [37:42] Dave Pengelley: than [37:42] Annie Vella: that you [37:42] Dave Pengelley: anymore. [37:42] Annie Vella: you want to [37:43] Dave Pengelley: So [37:43] Annie Vella: make sure [37:43] Dave Pengelley: make sure [37:43] Annie Vella: that, [37:43] Dave Pengelley: you're [37:44] Annie Vella: you know, [37:44] Dave Pengelley: waiting for [37:44] Annie Vella: but it's [37:44] Dave Pengelley: it. Yes. [37:45] Annie Vella: more specific perhaps, whereas the more generic conversation around how you do AI in a safe and sustainable way across an enterprise is a far broader conversation. And I'm convinced that it requires significant enterprise architecture thinking. And for finally, I think I understand what an enterprise architect is meant to do. Because for a long time, like when I worked at smaller companies, startup scale ups, we barely had architects full stop, let alone enterprise. What is that? I've met some. [38:31] Richard Webbe: You're so right. You're so right, Annie. [38:33] Annie Vella: Yeah. [38:33] Richard Webbe: And then above that, where's the business analyst conversation? That's the one that's going to pay for it. That's the one that wants the benefits. That's the [38:42] Dave Pengelley: I mean [38:42] Richard Webbe: customer. [38:42] Dave Pengelley: there there is a that [38:43] Richard Webbe: What [38:43] Dave Pengelley: there is [38:43] Richard Webbe: are [38:43] Dave Pengelley: that [38:43] Richard Webbe: we going [38:44] Dave Pengelley: problem, [38:44] Richard Webbe: to do [38:44] Dave Pengelley: right? [38:44] Richard Webbe: for the [38:44] Dave Pengelley: Where [38:44] Richard Webbe: business? [38:45] Dave Pengelley: um [38:45] Richard Webbe: Forget [38:45] Dave Pengelley: a [38:45] Richard Webbe: what [38:45] Dave Pengelley: lot of [38:45] Richard Webbe: tools [38:46] Dave Pengelley: these tools [38:46] Richard Webbe: we use. [38:46] Dave Pengelley: are personal [38:46] Richard Webbe: Just let's [38:47] Dave Pengelley: productivity [38:47] Richard Webbe: find [38:47] Dave Pengelley: based. [38:47] Richard Webbe: out what [38:47] Dave Pengelley: It's like you [38:48] Richard Webbe: the [38:48] Dave Pengelley: as [38:48] Richard Webbe: business. [38:48] Dave Pengelley: an engineer, you as [38:49] Richard Webbe: Need. [38:49] Dave Pengelley: a as a [38:50] Richard Webbe: Um, [38:50] Dave Pengelley: finance [38:50] Richard Webbe: you know, [38:50] Dave Pengelley: person, [38:50] Richard Webbe: it's the [38:51] Dave Pengelley: you [38:51] Richard Webbe: old [38:51] Dave Pengelley: as [38:51] Richard Webbe: story [38:51] Dave Pengelley: a [38:51] Richard Webbe: of running around [38:52] Dave Pengelley: legal [38:52] Richard Webbe: town [38:52] Dave Pengelley: representative [38:52] Richard Webbe: with a square peg [38:53] Dave Pengelley: checking, [38:53] Richard Webbe: looking for a [38:53] Dave Pengelley: you [38:53] Richard Webbe: square [38:54] Dave Pengelley: know, case [38:54] Richard Webbe: hole, [38:54] Dave Pengelley: law [38:54] Richard Webbe: it [38:54] Dave Pengelley: through [38:54] Richard Webbe: can be [38:54] Dave Pengelley: Harvey [38:54] Richard Webbe: a real waste [38:55] Dave Pengelley: or whatever [38:55] Richard Webbe: of [38:55] Dave Pengelley: it is. [38:55] Richard Webbe: time. [38:56] Dave Pengelley: Like they're all individual productivity things, but that shared enterprise knowledge and Shared context, brain across things. And that's the next battle that I think, you know, a lot of these frontier, frontier, whatever that means, companies that are building these sort of bot platforms, whether it's OpenAI or Anthropic or now Grok with their bots, um, and you know, the third parties and homies, et cetera. They're just trying to get, I think, some of these concepts established in people's brains that you've now got these semi-autonomous AI robot agent people that you can pass work to. And I think the next step is okay. Okay, once people are used to that using that familiar, then they're going to be going, Oh, well, how do I do that across my business? And you know, Slack is becoming a bit of a hub for those things. Jack Dorsey's just launched Buzz to try and be a thing for that. Um, different companies are taking different approaches to it, but it is that next battle from that enterprise layer. How do we give people the same reliable shared knowledge access for their productivity boosters? [40:08] Richard Webbe: Yeah, we don't we don't know what we don't know. We don't know how the next version of the world's gonna go. So we have to be on our toes. I I remember in the old days talking to people about online marketing. go to market and the web and how that was going to change things. And I was doing a uh strategy session with a company and someone in the corner kept saying, but yes, where are we going to get the brochures printed? And I kept saying, no, we probably won't [40:34] Annie Vella: Yeah. [40:34] Richard Webbe: need them in the future, right? No, but where are we going to get the brochures printed? It was a bit like you know that scene from Notting Hill, the guy said, where can I find these books? And he said, no, we're a travel bookstore. And the guy kept asking him the same question. People don't know where we're going, right? And uh and their context of thinking is down the traditional path [40:54] Matt: I [40:54] Richard Webbe: where [40:54] Matt: always [40:54] Richard Webbe: literally [40:54] Matt: think of um [40:55] Richard Webbe: what we're [40:55] Matt: previous [40:55] Richard Webbe: talking about is [40:56] Matt: technologies [40:56] Richard Webbe: multiple [40:56] Matt: and [40:57] Richard Webbe: contextual [40:57] Matt: previous [40:57] Dave Pengelley: Oh no, [40:57] Matt: little [40:57] Dave Pengelley: I was gonna [40:58] Matt: like [40:58] Dave Pengelley: tell [40:58] Matt: revolutions [40:58] Richard Webbe: uh [40:58] Dave Pengelley: Alex. [40:58] Richard Webbe: automations [40:58] Matt: that changed things. [40:59] Richard Webbe: that break [41:00] Matt: Like Annie, [41:00] Richard Webbe: the mold [41:00] Matt: you mentioned [41:01] Richard Webbe: of [41:01] Matt: Steve [41:01] Richard Webbe: traditional [41:01] Matt: Yeeggy before. [41:02] Richard Webbe: synchronistic [41:02] Matt: Um, one [41:03] Richard Webbe: thinking, [41:03] Matt: of [41:03] Richard Webbe: if that [41:03] Matt: his [41:04] Richard Webbe: makes any sense. [41:04] Matt: previous [41:04] Richard Webbe: And you've got to think [41:05] Matt: colleagues [41:05] Richard Webbe: outside the box almost [41:07] Matt: or [41:07] Richard Webbe: regularly. [41:07] Matt: uh in the space, Jeffrey Huntley. I always mention [41:09] Annie Vella: All [41:10] Matt: Jeffrey, [41:10] Annie Vella: right. [41:10] Matt: and he's just somebody that comes up to me a lot. Um, I don't know why. I kind of resonate with him. But he always talks about the loom, you know, the weaving loom, and how it changed everything. You know, another example could be like a calculator. You know, like how like what is this version of that? You know, is it Chat GPT is it claud, you know, I I think it's [41:33] Richard Webbe: Yeah. [41:33] Matt: more of like a fundamental mechanism, like you're saying, the the enterprise architecture and doing that safely. And what does that actually look like? And when you distill everything down and carve everything away, like what what is that mechanism? Like what is the loom that we're looking at? And yeah, I don't know. I like listening to people like Jeff because he he talks really conceptually and philosophically like that a lot while being still like a really deep technical nerd at the same time. [42:06] Dave Pengelley: Steve Yeege's cartoons that he generates [42:08] Annie Vella: His [42:08] Dave Pengelley: to uh illustrate [42:08] Annie Vella: politics. [42:08] Dave Pengelley: his ideas for for Gas Town. And and you know, he's uh this was when he was publishing and talking about Gas Town, which is a [42:14] Richard Webbe: And [42:14] Dave Pengelley: loop-based [42:14] Richard Webbe: tapping [42:15] Dave Pengelley: concept, [42:15] Richard Webbe: into [42:15] Dave Pengelley: right? [42:15] Richard Webbe: your [42:15] Dave Pengelley: Where he's [42:15] Richard Webbe: David, [42:16] Dave Pengelley: just like, [42:16] Richard Webbe: you've [42:16] Dave Pengelley: you know, [42:16] Richard Webbe: got [42:16] Dave Pengelley: well, [42:16] Richard Webbe: oh [42:16] Dave Pengelley: we [42:16] Richard Webbe: Vogue. [42:16] Dave Pengelley: don't have precision, we're just gonna throw volume. If you like throw [42:19] Richard Webbe: Like [42:19] Dave Pengelley: a [42:19] Richard Webbe: this. [42:19] Dave Pengelley: million monkeys at typewriters, eventually you'll get Shakespeare, right? He's just like, let's just spam as much code as possible until we get the outcome we're we're we're going for. [42:42] Richard Webbe: Yes, that is cute. I yeah, I always remember the um I did some research for uh a talk I did a while ago. And um when uh newspapers uh sorry when radio came out, everyone said newspapers were dead, but newspaper sales actually went up. And then when color magazines came out, everyone said, Oh, newspapers are dead, but again, newspaper sales and radio listening went up. And then television came out, everyone's saying, Oh, this is going to replace all of that, but no, it actually accelerated the use of those other communications. And we need to get used to the concept. Of augment rather than rip and replace. And as we augment using AI and automation, the rip and replace will naturally happen because we'll go, I don't need to do that step anymore because it's done that. So you don't have to go, let's rip this out. You just go, go along with it. Although it's a bit like uh, I think Dave, [43:37] Dave Pengelley: Talking to [43:37] Richard Webbe: you brought [43:37] Dave Pengelley: uh [43:37] Richard Webbe: up [43:37] Dave Pengelley: someone [43:38] Richard Webbe: last [43:38] Dave Pengelley: yesterday [43:38] Richard Webbe: week [43:38] Dave Pengelley: who runs [43:38] Richard Webbe: people [43:38] Dave Pengelley: another [43:39] Richard Webbe: running [43:39] Dave Pengelley: podcast, and [43:39] Richard Webbe: systems [43:39] Dave Pengelley: he was saying, and [43:40] Richard Webbe: that [43:40] Dave Pengelley: he's got [43:40] Richard Webbe: are [43:40] Dave Pengelley: he's got [43:41] Richard Webbe: still [43:41] Dave Pengelley: a few [43:41] Richard Webbe: running [43:41] Dave Pengelley: staff [43:41] Richard Webbe: in [43:41] Dave Pengelley: that [43:41] Richard Webbe: the background [43:42] Dave Pengelley: that help [43:42] Richard Webbe: that no [43:42] Dave Pengelley: do a [43:42] Richard Webbe: one [43:42] Dave Pengelley: lot [43:42] Richard Webbe: still [43:42] Dave Pengelley: of the [43:43] Richard Webbe: needs, [43:43] Dave Pengelley: editing [43:43] Richard Webbe: and [43:43] Dave Pengelley: work, [43:43] Richard Webbe: someone's [43:43] Dave Pengelley: and they [43:44] Richard Webbe: replaced [43:44] Dave Pengelley: had [43:44] Richard Webbe: it, [43:44] Dave Pengelley: this [43:44] Richard Webbe: but it's [43:44] Dave Pengelley: one [43:44] Richard Webbe: just being [43:45] Dave Pengelley: person [43:45] Richard Webbe: maintenance [43:45] Dave Pengelley: who [43:45] Richard Webbe: is [43:46] Dave Pengelley: decided [43:46] Richard Webbe: being paid, it's running in [43:47] Dave Pengelley: she [43:47] Richard Webbe: the corner with [43:47] Dave Pengelley: didn't [43:47] Richard Webbe: a tape on [43:47] Dave Pengelley: want [43:47] Richard Webbe: it [43:48] Dave Pengelley: to [43:48] Richard Webbe: saying [43:48] Dave Pengelley: be part [43:48] Richard Webbe: we don't [43:48] Dave Pengelley: of [43:48] Richard Webbe: know [43:48] Dave Pengelley: it [43:48] Richard Webbe: what [43:48] Dave Pengelley: anymore, [43:48] Richard Webbe: it does. [43:49] Dave Pengelley: but she had all the systems and procedures in her head. Like when she left, all of a sudden there was a huge gap in their whole publishing pipeline and workflows and everything. So um they spent a bunch time and they just use the he said primarily they're using you know custom GPTs or now Gemini gems or whatever they're just using the web-based things but they've set up the rules and the system prompts to get the results they want they put the data in they get the result out they use that um and you know sometime later this person said I I actually would like to come back and he's like all right you can come back but here's all the new systems we use now uh and that person you mentioned earlier like Annie like they've they've called themselves oh so I'm like a manager of agents like I'm a manager of robots uh and they're like that's a whole Whole different style of interaction and workflow for them. You've reduced that sort of single point of failure, but it's still that manual. They're still manually working through things. They've augmented. They haven't necessarily replaced and gone, like I'm trying to build with all my Hermie stuff. When I download this, I run a script and then it, you know, transcribes and pushes updates to YouTube and generates comics. And you still need the human in the loop there because you know it's taste of like going, oh, I'm gonna take this sentence fragment and put that in a speech bubble. I'm like, that is out of context. That was like a side joke that we made about something. Don't use that as a core thing for our conversational topics, but um building those automated systems is the next step. And that's we're working through those things of uh of not just replacing but augmenting and and systematizing. But again, for anyone, I say to people, even this guy I was talking to yesterday, like if you're gonna make changes, what value is it driving for you? Is it actually driving value to do anything different? Or is what you're doing there good? Where else should you be thinking to drive? Yeah. [45:31] Annie Vella: That reminds me of. So another one of the contributions that I've made as part of the research that I did was so one of the questions, one of my research questions was around that where we spend our time and effort, right? So when you add AI into the equation as a software engineer, do you find yourself now spending more or less time on these six pretty standard core tasks that or activities that most software engineers might feel they spend their time on? Probably wish they spent less time in meetings and more time on these things, but you know, things like writing code, refactoring, debugging, testing, reviewing, designing. [46:08] Annie Vella: So looking at the quantitative data there, uh most participants felt that they're spending less time writing code, which makes sense because they're generating it now and it's faster and you're generating more. So you're spending maybe less time on that. But what was more interesting was actually of those six [46:25] Richard Webbe: Oh, [46:25] Annie Vella: that I asked about, the only one that was slightly above median or Neutral at both time points was reviewing code. So [46:33] Richard Webbe: I'm I'm not sure what saying. [46:33] Annie Vella: that was the only task of the six that I asked about that most participants felt maybe they were starting to spend a little bit more time on. And this is all perception-based, right? Nobody's tracking anything. It's just I feel like I'm spending a bit more time. Not even testing was it. So testing did go up between the two surveys that I did because I had two years to do it part-time, I could do this longitudinally, right? Um, so it was uh reviewing and testing that shifted sort of towards. I'm now spending maybe a little more time on this by the second questionnaire. But it didn't seem to make up for feeling like we're spending quite a lot less time on everything else. So then it sort of begs the question: like, oh, are we all like just spending less time at work now? Like we're we're getting a lot more done with less time, less effort. We're all going home early and you know, putting our feet up, going to the beach, slipping latte. [47:24] Annie Vella: No, if anything. [47:34] Annie Vella: Yeah. [47:37] Richard Webbe: We're never going home early. One thing I learned no Hannah, you know, got look how modern a [47:43] Annie Vella: With [47:43] Richard Webbe: day we [47:43] Annie Vella: something [47:43] Richard Webbe: have today. [47:44] Annie Vella: else. So [47:44] Richard Webbe: Going back a, you know, a few [47:46] Annie Vella: what [47:46] Richard Webbe: decades, [47:46] Matt: Parkinson's [47:46] Annie Vella: I [47:46] Richard Webbe: we're [47:46] Annie Vella: f [47:47] Richard Webbe: growing our [47:47] Matt: law. [47:47] Richard Webbe: own bread, we're mining our own salt. [47:49] Matt: Parkinson's [47:49] Richard Webbe: or killing our [47:50] Matt: Law. [47:50] Richard Webbe: own cows and we still can't get much done. I think humans, like you were saying, tend to fill [47:55] Annie Vella: like [47:55] Richard Webbe: the [47:55] Annie Vella: it [47:55] Richard Webbe: available [47:55] Annie Vella: feels [47:55] Richard Webbe: space [47:56] Annie Vella: like a [47:56] Richard Webbe: with [47:56] Annie Vella: gas, [47:56] Richard Webbe: what attracts [47:56] Annie Vella: right? [47:57] Richard Webbe: them [47:57] Annie Vella: Um [47:58] Matt: Yes. [47:58] Richard Webbe: in between [47:58] Annie Vella: the [47:58] Richard Webbe: what [47:59] Annie Vella: quantitative [47:59] Richard Webbe: they're doing. Yeah. [48:00] Annie Vella: Data showed me that the qualitative data [48:02] Richard Webbe: Sorry? [48:02] Annie Vella: showed me that, or maybe there was this new type of work emerging, [48:05] Richard Webbe: That is ex- [48:05] Annie Vella: a new [48:05] Richard Webbe: sorry, [48:06] Annie Vella: category [48:06] Richard Webbe: it's [48:06] Annie Vella: of [48:06] Richard Webbe: exactly [48:06] Annie Vella: work [48:07] Richard Webbe: what [48:07] Annie Vella: that [48:07] Richard Webbe: it [48:07] Annie Vella: I [48:07] Richard Webbe: is. [48:07] Annie Vella: propose a name for, supervisory engineering work, name it, well, another engineering. Yeah. I've contributed to that problem. But what's interesting is it like as much as I had to narrow down to four research questions, this was not one, right? I wasn't looking to see if there was a new type of work emerging. That is actually something that seemed to show itself and emerge. From the data itself. But now that I know that I have this concept of supervisory work, go back in time and look for research that has been done on supervisory control in the past, and you find some really interesting material. So in the 70s, there was a really very comprehensive and quite hard to digest paper written by some folks at MIT about teleoperators of undersea equipment. And how they were going through stages of automation. Like now then they're not down in the water moving the machines around and making sure that they're doing what they're supposed to be doing. They're using software that controls the undersea equipment, which I mean, that's better, right? Like it's it should be better. But but now there's this disconnect between is the machine really doing what I think it's doing? And when how do I know that it's not? And if it's off by a bit, is that good enough? And and so So [49:30] Dave Pengelley: I [49:30] Annie Vella: they [49:30] Dave Pengelley: don't [49:30] Annie Vella: talk [49:30] Dave Pengelley: think [49:30] Annie Vella: about different types of delegation [49:32] Dave Pengelley: what might be [49:32] Annie Vella: and automation and and then of course there's [49:35] Dave Pengelley: a a bit bit. little little more than [49:35] Annie Vella: been research that has been cited that research since then throughout the decades, um, the ironies of automation. That's a good paper from the 80s. And I think that there's quite a lot of material we can lean on now to to see like people have been through this before. It's just new to us as software engineers because we're automating away the thing that gave us a real sense of what we were doing. Like if you're writing the lines of code yourself, you know what that code's doing. I mean, unless you were somehow sleepwalk sleep coding, I don't know, but but you should really know what you've built. [50:14] Annie Vella: Oh [50:19] Annie Vella: yeah. [50:24] Richard Webbe: You're talking to some people who I know I go back to REM statements uh in my basic program. So I know exactly what you're saying. You know, you own it, you live it, you're all over it. And you know, you see people in many industries, uh, physical engineering, car manufacturers, people who build houses, and when any of that process is automated in a way they don't understand, [50:43] Dave Pengelley: But I [50:44] Richard Webbe: they [50:44] Dave Pengelley: think we're [50:44] Richard Webbe: really [50:44] Dave Pengelley: all [50:44] Richard Webbe: get [50:44] Dave Pengelley: getting [50:44] Richard Webbe: a bit [50:44] Dave Pengelley: stuck [50:45] Richard Webbe: lost [50:45] Dave Pengelley: in the [50:45] Richard Webbe: and [50:45] Dave Pengelley: these [50:45] Richard Webbe: don't [50:45] Dave Pengelley: sort [50:45] Richard Webbe: know [50:45] Dave Pengelley: of [50:45] Richard Webbe: where [50:46] Dave Pengelley: shorter [50:46] Richard Webbe: to start. [50:46] Dave Pengelley: dopamine [50:46] Richard Webbe: What [50:47] Dave Pengelley: loops [50:47] Richard Webbe: do you mean [50:47] Dave Pengelley: as well, [50:47] Richard Webbe: you've [50:47] Dave Pengelley: because of [50:47] Richard Webbe: poured [50:48] Dave Pengelley: social [50:48] Richard Webbe: the slab [50:48] Dave Pengelley: media, [50:48] Richard Webbe: on a [50:48] Dave Pengelley: because [50:48] Richard Webbe: bit of [50:48] Dave Pengelley: of [50:48] Richard Webbe: foam? [50:48] Dave Pengelley: Doom [50:49] Richard Webbe: How [50:49] Dave Pengelley: Scrolling, [50:49] Richard Webbe: does that work? You know, [50:50] Dave Pengelley: because [50:50] Richard Webbe: it's [50:50] Dave Pengelley: of [50:50] Richard Webbe: it's [50:50] Dave Pengelley: all these [50:50] Richard Webbe: all [50:50] Dave Pengelley: things, [50:50] Richard Webbe: very interesting. [50:51] Dave Pengelley: shorter attention [50:51] Richard Webbe: And [50:51] Dave Pengelley: spans. [50:51] Richard Webbe: and that's [50:52] Dave Pengelley: And so [50:52] Richard Webbe: I was [50:52] Dave Pengelley: we [50:52] Richard Webbe: saying, [50:52] Dave Pengelley: want the [50:52] Richard Webbe: um, [50:53] Dave Pengelley: wins [50:53] Richard Webbe: new [50:53] Dave Pengelley: faster, [50:53] Richard Webbe: ways of [50:54] Dave Pengelley: we [50:54] Richard Webbe: working [50:54] Dave Pengelley: want the results [50:54] Richard Webbe: and a new [50:55] Dave Pengelley: quicker. [50:55] Richard Webbe: ways of thinking, I think goes [50:56] Dave Pengelley: Right? [50:56] Richard Webbe: with the new [50:56] Dave Pengelley: That [50:56] Richard Webbe: ways [50:57] Dave Pengelley: that's [50:57] Richard Webbe: of [50:57] Dave Pengelley: what [50:57] Richard Webbe: working. [50:57] Dave Pengelley: we want. And so, you know, whereas before you had to physically spend an hour writing out your code before you could, you know, test it and compile it or whatever it was and go, yes, it worked. Or no can break break breakpoints, step, step, step, step, watch. Um go through. I I I'm most of my programming experience was pre web stuff. So where you could actually set breakpoints and things versus [51:20] Annie Vella: Yeah. [51:20] Dave Pengelley: all this modern web stuff. But yeah, no, you you get that that that that win like because you did the work. Uh and now you sort of just want to know just give me Win, give me the win, give me the win. I don't wanna just just I'm gonna go do something else. Come back, give me the win. Um, without necessarily having done it. And so it does change also that actual sense of accomplishment. I heard someone talking about with kids. Don't just tell your kids to have good self-esteem or to be proud of themselves. Make sure if they've done something, you praise the thing they did and make them actually feel good about actually having done something. So that way they've got something to lean on, and it's like not this empty praise, but it's earned praise. And I think All this fast development agent robotic stuff, it does take away some of that earned satisfaction [52:07] Annie Vella: Yeah. [52:08] Dave Pengelley: of having done it and yes, okay, I agentically supervised this thing and I had the idea and I made it happen, but it's not quite the same as you know carving it out yourself and and doing it by hand. [52:20] Matt: I wonder if there was elements of that to that that term supervisory engineering. You know, like is there satisfaction in supervisory engineering? [52:29] Annie Vella: Yeah. Probably not for everyone. [52:34] Richard Webbe: It's the climbing the mountain. [52:44] Richard Webbe: Well, it's like climbing a mountain, you know, someone, you [52:46] Annie Vella: Yeah. [52:46] Richard Webbe: know, it used to be oh I climbed Everest. [52:48] Annie Vella: I um [52:49] Richard Webbe: I climbed [52:49] Annie Vella: I [52:49] Richard Webbe: a mountain [52:49] Annie Vella: might [52:49] Richard Webbe: or I [52:50] Annie Vella: have [52:50] Richard Webbe: went [52:50] Annie Vella: written [52:50] Richard Webbe: to this place. [52:50] Annie Vella: about it [52:51] Richard Webbe: Now [52:51] Annie Vella: in [52:51] Richard Webbe: people [52:51] Annie Vella: my [52:51] Richard Webbe: say, Oh, I've been there. [52:52] Annie Vella: blog [52:52] Richard Webbe: They [52:53] Annie Vella: post [52:53] Richard Webbe: don't mention [52:53] Annie Vella: about [52:53] Richard Webbe: that [52:53] Annie Vella: the [52:53] Richard Webbe: they were [52:54] Annie Vella: productivity [52:54] Richard Webbe: driven there or they [52:54] Annie Vella: experience [52:54] Richard Webbe: caught a [52:55] Annie Vella: paradox, [52:55] Richard Webbe: helicopter [52:55] Annie Vella: or [52:55] Richard Webbe: there. [52:56] Annie Vella: if [52:56] Richard Webbe: And [52:56] Annie Vella: I didn't, [52:56] Richard Webbe: that whole [52:57] Annie Vella: um [52:57] Richard Webbe: I earned the right to talk [52:58] Annie Vella: there's [52:58] Richard Webbe: about it [52:59] Annie Vella: a [52:59] Richard Webbe: thing [52:59] Annie Vella: really great [52:59] Richard Webbe: is is gone. [53:00] Annie Vella: Blog post by an ex-colleague of mine, David Kerlewis. He works at Canva as an engineering manager. And he was writing about tracking activities versus capabilities. And you should go more for the capability side, like just monitoring, you know, measuring activities is not the right way to look at the impact of AI or or the how your um your team might be performing. And in that blog post, uh there was this just this little section on where engineers find. joy and how some of them find their joy uh they might be a little bit more uh results oriented. The you know get a real kick out of seeing the outcome done. Doesn't really matter how, it just but I completed a thing. And some get their joy from the journey and the experience along the way, what you learn on the way to, you know, um so I've taken that concept and thought, well it's not very I I don't think it's very binary. I think it's more of a spectrum and you can sit somewhere along that spectrum, right? And you can And so for me, I get a lot of joy out of the journey. I I really enjoy the earning part of what we've just been discussing, right? I [54:07] Dave Pengelley: Yeah. [54:07] Annie Vella: I like being able to put in the hard work. I've I've done some long distance running. There's a lot of earning the right to be able to do that, right? Spend a lot of time pounding the pavement and wondering why I'm doing this to myself. But I'm sort of earning the right to be able to run that half marathon and say, I did it, right? I didn't just show up at the end in a taxi. Um but a lot of people you know I also I won't do something by halves either and then not have a finished product at the end of it. So I I definitely can see how you can be a bit of both. But depending on where you sit on that spectrum might help explain why you are feeling a certain way about this technology, right? Because if you like seeing the finished output, then this technology is like a superpower now. You oh my God, I pumped out 50 different prototypes [54:54] Richard Webbe: What [54:54] Annie Vella: for [54:54] Richard Webbe: is that? [54:54] Annie Vella: new startups and I've launched this and I've launched that and but you you're not doing it yourself. So much you're not learning along the way, and but if you're more on the other end of the scale, you like that journey, then you might feel a bit cheated out of it now because this thing can do it for you, and there's an expectation on behalf of businesses that leverage the tool, right? And so you're you're not having much of a journey anymore, you're just sitting there like yes, accept, yes, auto accept, just do the thing. So [55:26] Annie Vella: well, thank you. [55:31] Annie Vella: Yes, I am [55:34] Richard Webbe: Annie, you are uh you know I had a look at your profile before we got on. You're very learned, very qualified, multi-talented individual. And when you say you're in New Zealand still, aren't you? Just so you know, [55:47] Annie Vella: a good if not sure very you're [55:47] Richard Webbe: so [55:47] Annie Vella: idea. [55:47] Richard Webbe: when I went to New Zealand, I was doing Some work over there for Telstra and Telecom New Zealand. And uh I uh I was running back then and I was staying in uh uh where's the place with all the guys again um [56:02] Richard Webbe: yeah I was staying in Rotorua and I thought I hadn't really been to New Zealand unless I've seen those, you know, the guys that think and and I was into my running then, so I thought I'd go for a run as well. And I went for a run and I'm running around that lake there and I got stuck in some really heavy mud and I'm lifting up my feet. You know, when your feet get So heavy, you can't the shoes are falling off because of the mud. So I retraced my steps back and washed my shoes off. Went to a gex morning. I got up. I thought I walked down and look at the geysers. I couldn't find them at night time. And I got down, then there's a guy giving a talk, and he said, How you lose a person [56:33] Annie Vella: Oh, [56:33] Richard Webbe: every [56:33] Annie Vella: you [56:34] Richard Webbe: year [56:34] Annie Vella: almost [56:34] Richard Webbe: or [56:34] Annie Vella: wouldn't [56:34] Richard Webbe: so in the [56:35] Annie Vella: have [56:35] Richard Webbe: geysers [56:35] Annie Vella: been able to tell [56:35] Richard Webbe: down [56:35] Annie Vella: that [56:35] Richard Webbe: there. [56:36] Dave Pengelley: Oh [56:36] Richard Webbe: And [56:36] Annie Vella: story. [56:36] Richard Webbe: he [56:36] Dave Pengelley: my [56:36] Richard Webbe: goes, [56:36] Dave Pengelley: goodness. [56:36] Richard Webbe: Look, there's an example. [56:37] Dave Pengelley: That's [56:37] Richard Webbe: Look [56:37] Dave Pengelley: that's [56:37] Richard Webbe: at those footsteps [56:38] Dave Pengelley: that's an earned story [56:39] Richard Webbe: there. [56:39] Dave Pengelley: right there, Richard. Um [56:41] Richard Webbe: Never [56:41] Dave Pengelley: yeah, [56:41] Richard Webbe: run round [56:42] Dave Pengelley: it's [56:42] Richard Webbe: a [56:42] Dave Pengelley: it's [56:42] Richard Webbe: geyser [56:42] Dave Pengelley: interesting that [56:42] Richard Webbe: in [56:43] Dave Pengelley: whole [56:43] Richard Webbe: the dark [56:43] Dave Pengelley: where [56:43] Richard Webbe: in a [56:43] Dave Pengelley: where [56:43] Richard Webbe: national [56:43] Dave Pengelley: are you at? [56:44] Richard Webbe: park. [56:44] Dave Pengelley: What do [56:44] Richard Webbe: It's [56:44] Dave Pengelley: you enjoy? [56:44] Richard Webbe: not [56:44] Dave Pengelley: And [56:44] Richard Webbe: very [56:45] Dave Pengelley: I and [56:45] Richard Webbe: safe. [56:45] Dave Pengelley: I don't think it's a static thing, it moves around, right? Sometimes you just want the result depending on the task and and however it's done. And [56:50] Richard Webbe: Yeah, [56:51] Dave Pengelley: sometimes [56:51] Richard Webbe: yeah, [56:51] Dave Pengelley: like [56:52] Richard Webbe: it's [56:52] Dave Pengelley: I [56:52] Richard Webbe: okay. [56:52] Dave Pengelley: think back to like Lego kits, the for me it was always more fun building the thing than oh cool, now I've got a thing, and you play with it a little bit, but you break it down and you want rebuild it again because it was a fun following the instructions and building the plan. But then I spent you know nearly a decade in management managing a team and delegating workout and getting results through delegation. And so for me translating that to the agents, the agents aren't as good as the humans. Like I would prefer to be managing a a team of developers and getting them to do things and just developer agents. But um like I'm happy to sort of get delegated results back as well if it's if it's sort of the plan and and the vision and moving the the the bigger picture forward so it's horses for courses and and like all like all these things it's contextual it's all about context. [57:50] Richard Webbe: It's like, you know, David, you've managed salespeople before, [57:53] Annie Vella: Yeah. [57:53] Richard Webbe: and uh I'm an engineering background, [57:55] Annie Vella: Yeah. [57:55] Richard Webbe: but I've been in the sales and marketing game for many years. [57:58] Dave Pengelley: Well [57:58] Richard Webbe: And you [57:58] Dave Pengelley: that [57:58] Richard Webbe: talk [57:58] Dave Pengelley: and then [57:58] Richard Webbe: to [57:58] Dave Pengelley: that's [57:59] Richard Webbe: young [57:59] Dave Pengelley: always a [57:59] Richard Webbe: marketing [57:59] Dave Pengelley: thing with it [57:59] Richard Webbe: and salespeople, [58:00] Dave Pengelley: With any any [58:00] Richard Webbe: and [58:00] Dave Pengelley: management [58:00] Richard Webbe: how was [58:01] Dave Pengelley: role, [58:01] Richard Webbe: your day? [58:01] Dave Pengelley: any KPIs, [58:01] Richard Webbe: Oh, I've got [58:02] Dave Pengelley: any [58:02] Richard Webbe: a hundred [58:02] Dave Pengelley: targets, [58:02] Richard Webbe: proposals [58:02] Dave Pengelley: it's like [58:03] Richard Webbe: out [58:03] Dave Pengelley: which [58:03] Richard Webbe: today. [58:03] Dave Pengelley: are the measures [58:03] Richard Webbe: Yeah, that's [58:04] Dave Pengelley: that [58:04] Richard Webbe: not [58:04] Dave Pengelley: actually [58:04] Richard Webbe: a success. [58:05] Dave Pengelley: help [58:05] Richard Webbe: But to [58:05] Dave Pengelley: you [58:05] Richard Webbe: them, they [58:06] Dave Pengelley: predict [58:06] Richard Webbe: felt really [58:06] Dave Pengelley: perfor [58:06] Richard Webbe: proud [58:07] Dave Pengelley: like [58:07] Richard Webbe: because [58:07] Dave Pengelley: uh [58:07] Richard Webbe: they [58:07] Dave Pengelley: the [58:07] Richard Webbe: did [58:07] Dave Pengelley: outcomes [58:07] Richard Webbe: so much [58:08] Dave Pengelley: you [58:08] Richard Webbe: work, [58:08] Dave Pengelley: want, right? [58:08] Richard Webbe: right? [58:09] Dave Pengelley: You can measure like if [58:10] Richard Webbe: It [58:10] Dave Pengelley: you just measure [58:10] Richard Webbe: doesn't [58:10] Dave Pengelley: raw [58:10] Richard Webbe: make sense [58:11] Dave Pengelley: activities, [58:11] Richard Webbe: and it feeds your [58:12] Dave Pengelley: that's [58:12] Richard Webbe: ego. [58:12] Dave Pengelley: not necessarily the best measure. Sometimes it's the only measure you have, that's better than no measure, but like how do you actually [58:30] Richard Webbe: But that's true, because with with AI, [58:33] Dave Pengelley: Yeah. [58:33] Richard Webbe: your progress is much [58:36] Dave Pengelley: This [58:36] Richard Webbe: as as [58:36] Dave Pengelley: the [58:36] Richard Webbe: you [58:37] Dave Pengelley: this [58:37] Richard Webbe: and Annie [58:37] Dave Pengelley: week on LinkedIn, [58:37] Richard Webbe: were uh alluding [58:38] Dave Pengelley: I've actually gone [58:38] Richard Webbe: to, [58:38] Dave Pengelley: back old school [58:38] Richard Webbe: your [58:39] Dave Pengelley: for [58:39] Richard Webbe: progress [58:39] Dave Pengelley: a few of the [58:39] Richard Webbe: is [58:39] Dave Pengelley: images [58:39] Richard Webbe: accelerated. [58:39] Dave Pengelley: I've been posting. I've [58:40] Richard Webbe: So [58:40] Dave Pengelley: been dropping [58:40] Richard Webbe: you don't [58:40] Dave Pengelley: in [58:40] Richard Webbe: feel [58:41] Dave Pengelley: Canva [58:41] Richard Webbe: like you've earned [58:41] Dave Pengelley: and just like [58:42] Richard Webbe: the [58:42] Dave Pengelley: grabbing [58:42] Richard Webbe: success [58:42] Dave Pengelley: stock [58:42] Richard Webbe: that you've [58:43] Dave Pengelley: photography [58:43] Richard Webbe: had because all [58:43] Dave Pengelley: images [58:43] Richard Webbe: of a [58:44] Dave Pengelley: and [58:44] Richard Webbe: sudden [58:44] Dave Pengelley: then [58:44] Richard Webbe: you're there. [58:44] Dave Pengelley: compositing [58:44] Richard Webbe: I mean, AI [58:45] Dave Pengelley: my own [58:45] Richard Webbe: was developed [58:45] Dave Pengelley: little [58:45] Richard Webbe: to [58:46] Dave Pengelley: layered [58:46] Richard Webbe: write software, [58:46] Dave Pengelley: things rather than going, [58:47] Richard Webbe: uh, [58:47] Dave Pengelley: AI [58:47] Richard Webbe: and then [58:47] Dave Pengelley: Jen. [58:47] Richard Webbe: all of a sudden [58:48] Dave Pengelley: Like I [58:48] Richard Webbe: you're [58:48] Dave Pengelley: did [58:48] Richard Webbe: there. [58:48] Dave Pengelley: one where I have these little LinkedIn [58:49] Richard Webbe: Sorry, [58:49] Dave Pengelley: logos [58:50] Richard Webbe: Dave. [58:50] Dave Pengelley: on my eyes, and I was like, I could give a photo of myself to AI and say do this and do that. But I was like, you know what? No, background remove, stick it on here, drag and drop. And I'm like, I made that. That's That's that's mine now. I I I I made a picture, I did a thing. Um I mean I've done plenty of AI gen too, but uh it's you know they're little things sometimes. Richard speaking of generating things, uh it would uh behooves me to uh present your work of art. [59:30] Richard Webbe: I was [59:32] Richard Webbe: don't laugh at me, Annie. I was messing around. Play it. I was messing around with AI. [01:00:57] Richard Webbe: Thank you for playing that and embarrassing me. Dave. Here's the funny story behind that. A few years ago I was a multimedia producer. I actually produced some work with Max Walker. We got a a a multimedia award. I did some work for the Carlton Football Club and a few other things. That would have cost co that would take me, I reckon, a month to produce. I produced that. In bed in the morning at 5 a.m. when I was bored a few nights ago, and it didn't take much effort at all. And so I kind of feel ripped off that. But the key thing that I say about [01:01:30] Matt: I think [01:01:30] Richard Webbe: all [01:01:30] Matt: it's [01:01:30] Richard Webbe: this [01:01:30] Dave Pengelley: It's [01:01:30] Matt: really [01:01:30] Richard Webbe: automation, [01:01:30] Dave Pengelley: it's [01:01:30] Matt: good. [01:01:31] Dave Pengelley: the eyes [01:01:31] Richard Webbe: whether [01:01:31] Dave Pengelley: that get [01:01:31] Richard Webbe: it's [01:01:31] Dave Pengelley: me. I [01:01:31] Richard Webbe: a [01:01:32] Dave Pengelley: was watching [01:01:32] Richard Webbe: program [01:01:32] Dave Pengelley: it again [01:01:32] Richard Webbe: or a [01:01:32] Dave Pengelley: and [01:01:32] Richard Webbe: video [01:01:32] Dave Pengelley: and like [01:01:33] Richard Webbe: message [01:01:33] Dave Pengelley: all the [01:01:33] Richard Webbe: or [01:01:33] Dave Pengelley: all [01:01:33] Richard Webbe: something, [01:01:33] Dave Pengelley: everything all the other [01:01:34] Richard Webbe: it's [01:01:34] Dave Pengelley: expression [01:01:34] Richard Webbe: all [01:01:34] Dave Pengelley: moves, [01:01:35] Richard Webbe: in [01:01:35] Dave Pengelley: but [01:01:35] Richard Webbe: the [01:01:35] Dave Pengelley: the [01:01:35] Richard Webbe: message [01:01:35] Dave Pengelley: eyeballs [01:01:35] Richard Webbe: and [01:01:36] Dave Pengelley: don't [01:01:36] Richard Webbe: the [01:01:36] Dave Pengelley: move, [01:01:36] Richard Webbe: creativity. [01:01:36] Dave Pengelley: the irises just stay [01:01:37] Richard Webbe: The [01:01:37] Dave Pengelley: like [01:01:37] Richard Webbe: production [01:01:37] Dave Pengelley: like the pupils [01:01:38] Richard Webbe: is getting [01:01:38] Dave Pengelley: don't move, [01:01:38] Richard Webbe: easier. [01:01:38] Dave Pengelley: you do [01:01:38] Richard Webbe: And that's [01:01:39] Dave Pengelley: these [01:01:39] Richard Webbe: why [01:01:39] Dave Pengelley: like [01:01:39] Richard Webbe: I [01:01:39] Dave Pengelley: static [01:01:39] Richard Webbe: was playing [01:01:39] Dave Pengelley: marbles [01:01:39] Richard Webbe: around with that. [01:01:40] Dave Pengelley: and [01:01:40] Richard Webbe: So [01:01:40] Dave Pengelley: it's [01:01:40] Richard Webbe: thanks [01:01:40] Dave Pengelley: like [01:01:40] Richard Webbe: for [01:01:40] Dave Pengelley: the [01:01:40] Richard Webbe: playing [01:01:40] Dave Pengelley: old the [01:01:41] Richard Webbe: that, [01:01:41] Dave Pengelley: old [01:01:41] Richard Webbe: Dave. [01:01:41] Dave Pengelley: like [01:01:41] Richard Webbe: I'm [01:01:41] Dave Pengelley: chicken [01:01:41] Richard Webbe: a bit embarrassed [01:01:41] Dave Pengelley: head. [01:01:42] Richard Webbe: by that, but [01:01:42] Dave Pengelley: They don't [01:01:42] Richard Webbe: it was [01:01:42] Dave Pengelley: move [01:01:42] Richard Webbe: my [01:01:42] Dave Pengelley: while [01:01:42] Richard Webbe: first attempt [01:01:43] Dave Pengelley: everything else [01:01:43] Richard Webbe: in [01:01:43] Dave Pengelley: does. [01:01:44] Richard Webbe: bed at 5 a.m. I'll make this up. [01:01:46] Dave Pengelley: Yeah, the little [01:01:48] Matt: I love all the dramatic causes, like the [01:01:50] Annie Vella: Yeah. [01:01:51] Matt: real like all the [01:01:53] Dave Pengelley: and [01:01:53] Matt: emphasis, [01:01:53] Dave Pengelley: the and [01:01:53] Matt: like [01:01:54] Dave Pengelley: the [01:01:54] Matt: is [01:01:54] Dave Pengelley: jump [01:01:54] Matt: really [01:01:54] Dave Pengelley: cuts. [01:01:54] Matt: good. Yeah. [01:01:56] Richard Webbe: The sa [01:01:56] Matt: And all [01:01:57] Richard Webbe: I didn't [01:01:57] Matt: the different [01:01:57] Richard Webbe: put [01:01:57] Matt: rooms [01:01:57] Richard Webbe: the sign [01:01:58] Matt: that you [01:01:58] Richard Webbe: in. [01:01:58] Matt: kept changing [01:01:58] Richard Webbe: The program [01:01:58] Matt: studios. [01:01:59] Richard Webbe: put the sign in for me automatically. I didn't put it. [01:02:05] Richard Webbe: Yes, it I didn't [01:02:06] Matt: Awesome. [01:02:06] Richard Webbe: do that. [01:02:07] Dave Pengelley: I [01:02:08] Matt: No, [01:02:08] Dave Pengelley: reserve [01:02:08] Matt: I actually [01:02:08] Dave Pengelley: the right [01:02:08] Matt: genuinely [01:02:08] Dave Pengelley: to veto. [01:02:09] Matt: like it. I I watched it at two times speed though, so like that was a very different experience. [01:02:13] Richard Webbe: Oh [01:02:13] Dave Pengelley: Oh [01:02:14] Richard Webbe: it did. [01:02:14] Dave Pengelley: I [01:02:14] Richard Webbe: It did. It did. But don't [01:02:16] Dave Pengelley: I [01:02:16] Richard Webbe: laugh, because I'm [01:02:16] Dave Pengelley: like [01:02:16] Richard Webbe: working [01:02:16] Dave Pengelley: our [01:02:17] Richard Webbe: on [01:02:17] Dave Pengelley: podcast [01:02:17] Richard Webbe: one for you, [01:02:17] Dave Pengelley: comics. [01:02:17] Richard Webbe: Matt, and I'm [01:02:18] Dave Pengelley: Um [01:02:18] Richard Webbe: working [01:02:18] Dave Pengelley: that [01:02:18] Richard Webbe: on one [01:02:18] Dave Pengelley: that [01:02:18] Richard Webbe: for [01:02:19] Dave Pengelley: that's my [01:02:19] Richard Webbe: Dave [01:02:19] Dave Pengelley: speed. [01:02:19] Richard Webbe: right now, so stay tuned. [01:02:23] Dave Pengelley: Yeah. Yeah, well um have to make sure you send me a photo like for a modelling for your [01:02:28] Richard Webbe: Much [01:02:28] Dave Pengelley: uh [01:02:28] Richard Webbe: better [01:02:28] Dave Pengelley: comic [01:02:28] Richard Webbe: experience, [01:02:28] Dave Pengelley: version of yourself, [01:02:29] Richard Webbe: much better. [01:02:29] Dave Pengelley: Annie. [01:02:29] Richard Webbe: I'm gonna [01:02:30] Dave Pengelley: for [01:02:30] Richard Webbe: speed [01:02:30] Dave Pengelley: when I [01:02:30] Richard Webbe: it [01:02:30] Dave Pengelley: generate [01:02:30] Richard Webbe: up. [01:02:31] Dave Pengelley: that. Um [01:02:33] Richard Webbe: They are good. And and Tony was most impressed with those days. Well done from last [01:02:37] Annie Vella: Sure. [01:02:38] Richard Webbe: week. [01:02:40] Dave Pengelley: we we we we're at the hour. We're gonna wrap it up there. But I mean not everyone's AI generating stuff. [01:02:44] Richard Webbe: Yeah, [01:02:45] Dave Pengelley: I I [01:02:45] Richard Webbe: Annie, [01:02:45] Dave Pengelley: saw this [01:02:45] Richard Webbe: yeah, [01:02:45] Dave Pengelley: this [01:02:45] Richard Webbe: Annie, [01:02:46] Dave Pengelley: post. [01:02:46] Richard Webbe: give [01:02:46] Dave Pengelley: Uh [01:02:46] Richard Webbe: me some of [01:02:46] Dave Pengelley: Canadian [01:02:46] Richard Webbe: your really [01:02:47] Dave Pengelley: theater [01:02:47] Richard Webbe: smart, [01:02:47] Dave Pengelley: employee [01:02:47] Richard Webbe: famous [01:02:48] Dave Pengelley: drew [01:02:48] Richard Webbe: quotes [01:02:48] Dave Pengelley: their own [01:02:48] Richard Webbe: and [01:02:48] Dave Pengelley: post [01:02:48] Richard Webbe: I'll [01:02:48] Dave Pengelley: of [01:02:48] Richard Webbe: create. [01:02:48] Dave Pengelley: a Spider [01:02:49] Richard Webbe: We [01:02:49] Dave Pengelley: Man [01:02:49] Richard Webbe: had [01:02:49] Dave Pengelley: brand [01:02:49] Richard Webbe: a little, [01:02:49] Dave Pengelley: new [01:02:49] Richard Webbe: you know, [01:02:49] Dave Pengelley: day after [01:02:50] Richard Webbe: Pixar [01:02:50] Dave Pengelley: the cinema [01:02:50] Richard Webbe: style [01:02:50] Dave Pengelley: failed [01:02:50] Richard Webbe: cartoon [01:02:50] Dave Pengelley: to receive an [01:02:51] Richard Webbe: of [01:02:51] Dave Pengelley: official [01:02:51] Richard Webbe: that for you. [01:02:52] Dave Pengelley: one. [01:02:57] Dave Pengelley: Sorry. Uh No. [01:03:03] Dave Pengelley: That person earned every piece of praise they get for that. [01:03:06] Matt: Do you do you know that do you know that shape? That like that [01:03:10] Dave Pengelley: Well this [01:03:10] Matt: spoda [01:03:10] Dave Pengelley: is the [01:03:10] Richard Webbe: I [01:03:10] Dave Pengelley: the the [01:03:11] Richard Webbe: think [01:03:11] Matt: man? [01:03:11] Dave Pengelley: the Spooter [01:03:11] Richard Webbe: that's [01:03:11] Dave Pengelley: Man, [01:03:11] Richard Webbe: fabulous. [01:03:12] Dave Pengelley: right? It's [01:03:12] Matt: Yeah. [01:03:12] Dave Pengelley: that that that's sort of yeah yeah. [01:03:14] Matt: Yeah. [01:03:14] Dave Pengelley: But [01:03:15] Matt: That's [01:03:15] Dave Pengelley: points [01:03:15] Matt: very [01:03:16] Dave Pengelley: for [01:03:16] Matt: old. [01:03:16] Dave Pengelley: effort. I like it. I like the fact that they drew one by hand. Even if it's [01:03:20] Richard Webbe: Absolutely. [01:03:21] Dave Pengelley: modelled off a meme. Yes, anyway. Um thank you, Annie. It's been a pleasure having you join join the show today and and de di dive into all things. developers and careers. [01:03:32] Annie Vella: Oh gosh, thank you so much, Richard. Look, it's been a real pleasure. It's been a really fun conversation. And um yeah, keep doing this and and get more people on and get me back one day when [01:03:43] Richard Webbe: Annie, [01:03:43] Annie Vella: I have something new [01:03:44] Richard Webbe: you are [01:03:44] Annie Vella: to [01:03:44] Richard Webbe: an [01:03:45] Annie Vella: say. [01:03:45] Richard Webbe: outstanding [01:03:45] Dave Pengelley: Brilliant. [01:03:45] Richard Webbe: talent. [01:03:46] Dave Pengelley: You're all definitely welcome. [01:03:48] Annie Vella: I'd [01:03:48] Dave Pengelley: All right. [01:03:49] Annie Vella: love [01:03:49] Dave Pengelley: Well, [01:03:49] Annie Vella: to [01:03:49] Dave Pengelley: um, thank you everyone for for joining the show. Uh if you have questions, comments, please drop them in and we'd love to get back to them later. So make sure you do that. Do the liking and the subscribing. [01:03:59] Richard Webbe: We'd love that. [01:04:00] Dave Pengelley: And all [01:04:00] Richard Webbe: Come [01:04:00] Dave Pengelley: the things [01:04:00] Richard Webbe: come [01:04:00] Dave Pengelley: you're [01:04:00] Richard Webbe: back [01:04:00] Dave Pengelley: meant [01:04:00] Richard Webbe: again [01:04:01] Dave Pengelley: to do, [01:04:01] Richard Webbe: soon. [01:04:01] Dave Pengelley: whether you're on uh if you're on LinkedIn, find the AI operators pod page and like us on there so we can sort of keep growing on the business side of things, but also on YouTube. It's 2026. You guys know what to do. Just just please do it, because no one does, [01:04:14] Matt: Share [01:04:14] Dave Pengelley: but [01:04:14] Matt: it. [01:04:14] Dave Pengelley: do it. [01:04:15] Matt: Share it's the best algorithmic interaction. [01:04:17] Dave Pengelley: Yeah. [01:04:21] Dave Pengelley: All right. [01:04:24] Annie Vella: Hi George. [01:04:25] Dave Pengelley: Thanks, Richard. All right. I'm gonna play us out and uh thank you everyone for joining. We'll see you next week. [01:04:31] Richard Webbe: I have another podcast to go to, so I'm I must away, but I will talk to you guys later. And lovely to meet you, Annie. What a wonderful time. [01:04:49] Richard Webbe: Yeah.