From Doer to Director, Getting Value From AI

Paul Boag: Hello and welcome to Boag World show, the longest running web design podcast where we're looking at user experience, design, conversion optimization and working in the web. On this month's show, we ask just how good is Claude Design? We explore the changing role of UX professionals and how we're not going to get the most out of AI if we're organized.

Marcus Lillington: Um, Ungorganized.

Paul Boag: Paul.

Marcus Lillington: That's a new word. Is that because you're so busy? Ah.

Paul Boag: Uh, bless. Are we. We seriously not going to redo that introduction? You're just gonna. You're just gonna keep going on? Anyway, so my name is Paul Boek, and joining me, as always, as you've already gathered, is Marcus. I'm bloody hating Billington. Hello.

Marcus Lillington: Oh, rude. Even I know. Hello.

Paul Boag: That's the mood I'm in.

Marcus Lillington: Yeah.

Paul Boag: Get it done today.

Marcus Lillington: God, then. All right. Okay. I did another. I did another article. We'll save that one for next month, so. Because we're out of time. Nice to see you, mate. I'm sorry to hear that you're, um, being a bit stressed, but being busy is no bad thing. I may.

Paul Boag: No. Why can't the world just cooperate to give you everything on a nice, even spread out over the year, ah. Kind of thing? No.

Marcus Lillington: Yeah.

Paul Boag: It all just comes suddenly and of course I'm away again, so that's what I tell you.

Marcus Lillington: Don't look like you're at home. Are you in. Are you. What? You're in a caravan, aren't you?

Paul Boag: Uh, I hate saying that. Can we call it a trailer like the Americans do? It doesn't sound.

Marcus Lillington: You're in a caravan.

Paul Boag: It's bad.

Marcus Lillington: Are you at a caravan park?

Paul Boag: I mean. Yeah, I am. I'm in a caravan, um, park. It's not even a very nice one. In the middle of Exmoor while it's pissing with rain. Yeah, Exmoor's lovely. No, Exmoor's lovely. We're going to be here next week, which is when the weather's going to be good.

Marcus Lillington: Yes. Yay. Yes. I've taken an extra day off next week because.

Paul Boag: Why not?

Marcus Lillington: Yes.

Paul Boag: Well, so have I. I'm doing exactly the same.

Marcus Lillington: Yeah.

Paul Boag: Um, and it'll probably end up pissing down because that's the weather in Britain. I bet there are Americans right now listening to this, just rolling their eyes. Oh, is that all the British ever talk about? But you don't know what it's like, you see, because if you live on a little. Is that's, uh, at the bottom of the Gulf Stream, the Weather turns on a dime. You watch your weather forecast and they're generally pretty accurate in the uk. Could be anything. It changes from one minute to the next in the day.

Marcus Lillington: So they're better than they used to be at it. But I've learned that you can only really trust 48 hours.

Paul Boag: Yeah.

Marcus Lillington: Ahead. Any more than that, it's like.

Paul Boag: Nah. Yeah, absolutely.

Marcus Lillington: Yeah.

Paul Boag: So. So this glorious experience is made possible because of my Starlink Mini, which is lovely.

Marcus Lillington: Okay.

Paul Boag: I guess.

Marcus Lillington: Well the thing is cutting in out and making you look go all juddery. I'm joking.

Paul Boag: No, it's perfectly fine. What are you talking about? They're so cute mine. Because you can just shove them in a backpack. So you can even take them when you go walking if you're so inclined. Not that I ever am, but they're really tiny. They're great. I love them.

Marcus Lillington: A Starlink Mini, that's like a little aerial I take it?

Paul Boag: Oh, do you know? No, no, no, no. It's um, it's imagine.

Marcus Lillington: I know what Starlink is, but.

Paul Boag: Yeah, a star, A Starlink mini looks exactly like a normal Starlink except it's about. It's between the size of an A4 piece of paper and an A3 piece of paper in terms of dimensions. Um, and it's got its own little flip out stand, it's got the router built into it. Everything is a self contained unit. Can run off of a 12 volt. Um, and it, you know, you can use it anywhere.

Marcus Lillington: Okay.

Paul Boag: Um. And it seems to be really good at ah, picking up a satellite even if it's not pointing in quite the right direction. And so literally I can put it in um, on the dashboard of my car and work driving along. Not obviously when I'm driving. Well, I could, I could give it a go. Don't do that kids.

Marcus Lillington: Exactly. You'd probably end up getting arrested. Paul.

Paul Boag: Well, yes, I was gonna say again

Marcus Lillington: then but I can't think you have been arrested before. Maybe you have.

Paul Boag: I've been done for speeding a couple of times so it kind of counts, doesn't it?

Marcus Lillington: I don't know.

Paul Boag: Yeah, one ah, time I'm one of those times. I, I will fully own up to the other time. It was a trap. It was designed to get you. Did they make faster? First up it was a slip road going on to a dual carriageway that was in theory still 30 miles an hour. Dual carriageway was 70. Of course you have to speed up to merge with the traffic.

Marcus Lillington: Bastards that you haven't let that one go. Have You.

Paul Boag: When was that? Do you know when that.

Marcus Lillington: 1995 it was, wasn't it?

Paul Boag: It was, it was. It would have been before town pages. So, yes, probably about 1995.

Marcus Lillington: That's, yeah, pretty bang on then. And it's still. Oh, it's still in there, isn't it?

Paul Boag: It still annoys me. Yeah. So let's talk about, um, uh, the app for this month which is going to be Claude Design. Um, so, so Claude Design is come out and everybody's going ooh and ah, and you know, getting really excited about it. So this is a, ah, this is an app basically that will build out, um, designs for you. Um, and uh, it's pretty damn impressive. Uh, so it'll, it'll give it a good damn go. I've created a, um. Well, I mean I, I run, well, not run, but fundraise, um, for a tiny little charity in the hit here in the uk, um, uh, called uh, Bethesda Project and they basically raise money for children's education out in India. And I built them a website four or five years ago on Squarespace, and it cost me a fortune in, in like hosting and stuff for a really, really simple site. And I thought, screw it, I'm gonna just build it as an HTML site and m. You do it like that. And, and I used Claude Design to design it and then Claude code to build it. And um, for a small little website, I did it in, I would have thought, probably about eight hours worth of work, entire website, all new content, all new everything. Just bloody brilliant for that.

Marcus Lillington: Yeah.

Paul Boag: Um, the other. Sorry, go on.

Marcus Lillington: No, I was gonna say I, I had a look at it. I was, I was very impressed.

Paul Boag: What would be of interest specifically to you? Mine, which I discovered today, is that it doesn't just do webcam, uh, web design, it can do, um, uh, presentations too. Oh yeah. And it can export those into a normal PowerPoint presentation that you can edit offline. And I've been working on that today because I've got, uh.

Marcus Lillington: I'm glad you said that, Paul, because I was going to ask you. So can you export it into PowerPoint or whatever? I thought I'll save that until we record the podcast. I would have forgotten otherwise. That's a sign of age.

Paul Boag: So the answer is yes, you absolutely can. And it also produces a lovely PDF or, um, you can have it online as well. So I've just written, um, I've been working with the National Trust and I'm putting together a presentation for them. Um, and normally that would have taken me I would have thought probably about four days to go from nothing, um, uh, to all the way through to you know, it's a 53 slide deck.

Marcus Lillington: Yeah, you've got to come up with all that content, haven't you?

Paul Boag: So yeah, so, but I'd already written like a draft report, um, but it didn't have quite the, quite emphasis so I using notion to write the content and then Claude designed to produce it. I was done in two hours, I'm not kidding you, it was incredible. Um, and obviously I had to go back through and check it and blah blah blah blah. But the first pass of it was done in two hours and it was bloody incredible for that. So you can use it for presentations, you can use it for small websites. Um and it ticks a lot of the boxes for me of what I want from a, a kind of AI web design tool. So it can work with design systems. So you can get it to build out a design system and then it's capable of working within that. Um, or you can import 1m from Figma if you prefer. You can make manual edits, which is a big one for me. You can go in and go I'll just add a bit more padding there or whatever and you can manually do that without having to ask the AI and burn through a load of tokens doing it. Um, you can select elements visually, um, and you know, say that thing there that I want to be updated but it's got a couple of things that I don't like so much. Um, one is you can't export out of it to back into figma. You can import from Figma but you can't export to Figma which is annoying. The other thing is it doesn't give you an easy way of publishing what you, you produce for a website. So um, my, my um, the other tool I use, which is called figure.design which I've mentioned before, that allows me to hit a publish button and then I can send that directly to the con, uh client. This one I just download the project and then I have to upload it to my own server. Not that big a deal, um, but a bit fiddler and a bit more annoying. So um, but the big problem with it is one of usage. So if you sit down to do a design project you will burn through your Claude usage in like 5 minutes flat or it feels that way.

Marcus Lillington: Oh, so you have to go at it, what every day for five minutes?

Paul Boag: Well this is, yeah, this is almost kind of the, the thing that they're doing they go, oh, you've used up your usage. You could wait until Sunday is what it told me today. Or you can just pay with credits. Pay as you go. Yeah, yeah. Uh, it's not surprising at the moment. AI is a massive land grab and so everything has been underpriced for that from now and we'll see a big jump in the price of all these AI tools. But uh, that said to go from days worth of work to hours worth of work, it's worth it for me because especially at the moment when I'm really busy, I could just fill that up with other work that pays me so you know, I, I end up in profit. But you know, it does get a bit annoying when it says, oh, you've reached your limit and it's like, you can try again on Sunday. I can't. If I wait till Sunday I'll have forgotten what I'm doing. And they damn well know it.

Marcus Lillington: So how much do you think it cost you in, in Claude Design to build the Bethesda site?

Paul Boag: Beyond Beyond? So I got my 20amonth, uh, 20 pounds a month or whatever it is subscription. I reckon it cost me another 50 quid maybe. I know, right, yeah.

Marcus Lillington: I mean if, if it was the local club website, you wouldn't mind that, you know. No, if you'd said 200 you'd be sort of thinking okay, right, I think twice. But 50 quid.

Paul Boag: Yeah, yeah, exactly. It's a no brainer in it.

Marcus Lillington: Yeah.

Paul Boag: If you want to see the, the um, the site that it produced, I'll, I'll put um, a link in the, the show notes and of course feel free to donate if you've appreciated the content that's been produced from this podcast, blah, blah, blah. But what's interesting about going and seeing it is not the cause, although the cause is amazing. But that Claude design has a very specific style.

Marcus Lillington: Mhm.

Paul Boag: Right. If you're not heavily guiding it, which I wasn't with that particular one, it has certain tells that you can go, oh, the Claude design built that. And I've seen reoccurring ones like for example, it's got this weird habit of creating underneath the hero. It creates this bar that has no real use, but it seems to always put it in and put something in there. And it's done it on my site as well. So it's little things like that.

Marcus Lillington: Yeah.

Paul Boag: But you could say everybody goes, oh, uh, yes. So that makes it hell. I can look at designers, I can see uh, see a design by an Ed designed by Lee the design by Andy Clark or designed by me, and I could tell you whose was what. We've all got our own style, so, you know, I don't think it's that big a deal.

Marcus Lillington: No.

Paul Boag: So let's talk about the main topic for today. This is a bit of a weird one. It's a bit kind of not directly related to, uh, you know, web design, kind of. Not really, you know, and. Well, I guess, yeah, maybe it is. Yeah, it kind of is. Yeah, it's a weird one.

Marcus Lillington: It does apply to people who do what I do, for example, but it's much wider than that. Yeah.

Paul Boag: So, I mean, basically it. I. I've seen a lot of people go, um, oh, AI isn't m. Saving me time. It's actually taking me more time. Sure, I can do more stuff, but it's taking me more time to get the value out of it. Um, and I've been coaching a few people like that. And that was really confusing me because I'm like going, hang on, I produced a slide deck in two hours where it took me four days before. How can it not be saving you time? But I think what I'm beginning to discover is that, um, a lot of it comes down to how organized or otherwise you already are in your work. Right. And how you structure and approach your work. Because if you're not fundamentally efficient in the way you operate, and particularly if you're not experienced in working with other people in an efficient way, then you're not going to get the most out of AI, Right? So the way that I've kind of brought all of that together, that kind of thinking together, is that a lot of our problems with AI and those people that are struggling it are still inherently wanting to be doers, right? They want to be in there playing around with the code or the, the pushing things around in figma or, or writing the words or whatever else. And that really, to get the most from AI, you've got to move from that doer mindset to a director mindset. Um, and it kind of reminded me of. Of, um, the. The moment in the biopic of Steve Jobs where there's this, this clip. I'll probably put it in the show notes. Um, but there's this clip where, um, you know, Steve Wolfsniac turns around to Steve Jobs and says, what. What is it that you do? I know what I do. I write code, you know, I produce hardware. I. I do everything and you do nothing. What is it that you do? And Jobs replied that he conducts, you know, he he is the conductor of an orchestra. And that, uh, Steve Wozniak is an amazing lead violinist. Right. Is best in the world at that. But Steve's role is to bring all of those disparate players together. And I think that is very much the mind shift that we need to make now. Why, why do I say that? Well, it's because increasingly those people that are getting the most time savings out of AI are directing a, a collection of agents that are doing different things, right? So let me give you an example going back to that National Trust document. Um, I was writing right, in that this is a really simple but real example that I've been doing today. So in one window I had notion open, right? And I had an AI client in notion and I was saying, okay, we're going to work on the outline for this slide deck. Um, here's all the relevant information. I want you to kind of plan out. You know, the slides go, right? So it's off there. Then I flipping over to Claude Design and I'm saying, okay, we need to work out what these, um, slides are going to look like. So I want you to go away and look at the National Trust website and you know, get their branding and their look and m feel and create a design system for this slide deck, right? So they're now both off doing their things. And in meanwhile I pick up another thing that I've got to do. I think I started another, um, uh, uh, agent running on a different project that I had at the time to write tran, uh, um, uh, scripts for videos that we needed to be recorded. So I had all of these different agents all going off and doing the things simultaneously. Up until this point, you, you would, you would always work on one thing at a time, basically. Okay, now that is a difficult skill to develop, right? Especially the people that are best prep m. Uh, the best place to do that kind of thing is people that have management skills, right? So people that are used to delegating to other people. But if you're not used to that, suddenly you've got to hold all of these multiple things in, in your head at the same time. Um, and you're switching between these tasks and feel quite distressing and quite difficult to do. And actually there is something that's come along now which is called AI burnout. And it is actually a real thing where agents are moving fast, you know, doing all of their different things and you're context switching relentlessly between these different, um, things that are going on. And the different feedback from the AI is Is, is kind of piling up. And one AI said, oh, I finished my task, and the other one said, I finished my task and there's this desire. Well, I need them going on the next thing. Um, and so we, we need to kind of pace ourselves and go, okay, how much of this can we deal with and transition into that slowly? But the kind of worst part goes back to this not being organized things. Right. So one of the big parts of all of this that I'm seeing is that if you, if you're not a fundamentally organized and structured person, then you're going to find it much harder to get good value out of AI. Um, and so people are failing because their workflows and their processes are fundamentally disorganized or they don't have processes at all. So if you come to say me, who is, I've become very anal and uptight and process driven, um, you will find that, um, you know, if you ask me, how do you do X, I'll be able to say, I do this, this, this and this. Right. That's my process for doing that thing. Whether it be a proposal writing, whether it be a design, whatever. I have processes for everything. Not necessarily always written down, although increasingly I'm doing that these days, writing them down. Um, and so it's very easy for me to explain to an AI what I want it to do. Okay. Now, I'm nothing special because of that. The reason that I'm like that is because I didn't used to be like that and I suffered from huge anxiety and trying to hold all these things in my head and work out what I was going to do next. And it overwhelmed me. So my coping mechanism was to structure everything. Right. And I think if you're probably a more laid back person like you, Marcus, I might be putting words in your mouth, but you probably don't feel that same need to prototype, you know, turn everything into a process like I would. Is that. Yeah, fair enough.

Marcus Lillington: It's totally fair enough. Um, I talk a lot about, um, logging processes. We had, we think we talked about this on the last, yeah, before, before that. And I'm still. Yeah, yeah, I'll do that one day.

Paul Boag: Yeah, exactly.

Marcus Lillington: And so therefore I, uh, you know, I, I tend to use AI linearly. I don't, I don't, Yeah, I don't conduct an orchestra. I actually quite like the doing as well writing in my case. Um, yeah, so.

Paul Boag: And that's the other big problem. Yeah, is that, you know, it's that thing of having to give up. I Mean, you see this amongst, um, design leaders all the time, right? Yeah.

Marcus Lillington: Well, design is good. It's an enjoyable thing to do, isn't it? Yeah.

Paul Boag: So, and so what happens is you get promoted into a design leadership position and suddenly you're not doing that. You're sitting in M meetings and telling other people to do the design. And, you know, a lot of design leaders find that really hard. And so they end up micromanaging because they really want to be in doing the design, you know, um, and a lot of the problem, even with the way designers are using AI, is that they're kind of the AI produces something and they go, well, that quite what I had in my head. And then they start fiddling with it and fiddling with it and fiddling with it even though it's perfectly good enough because they really want to be doing it themselves. And, um, that's a difficult battle as well. And I sympathize for people like that. I'm again, I've been quite fortunate that I'm a big, big picture person, so I'm very much of the attitude, oh, that's good enough, that'll do the job. Let's keep moving forward. And if you're not that kind of person, that creates real problems here. I've got really. Any answers to all of this. It's more, I think it helps to understand what's going on here. But going back to the being organized thing, I, I think it's going to increasingly valuable to start organizing yourself in a more efficient way to have. You know, I've been banging on about digital Playbooks for years and having those SOPs and those policies because then AI can learn from those and can go, all right, I already know how Paul does things. I can just follow those. And so it's much more likely to get it right the first time and also to have that information in an easily accessible way for AI. Right. Whether that's putting it in a system like Notion or even just markdown documents. Right. Markdown documents are great for AI, you know, and if you don't know how to write markdown, go learn Markdown. It. It like takes 10 minutes to learn how to do it. And it's so, so useful. Yeah. Um, so, you know, for me it's about having a task management system so the AI knows what I'm doing, what I'm working on, that it can access and see all of that. It's about having a repository of all of my knowledge in one place so it can go and reference that. And I'M not having to repeat myself endlessly. It's about building out those skills libraries I've talked about before on, um, the podcast, about having that playbook, all of that kind of stuff. So, yeah, it's an interesting time really, to kind of work out what it is that we're doing here and, um, and kind of work out the, the how. What we want to be and how we want to fit into this new world because it's, it's really difficult. It's going to be an uncomfortable transition for some people because if you do like being a doer, I think that's going to be taken away from you to some degree, whether you like it or not, in order to be. And I'm not saying that's a good thing. Um, I'm just saying I think it's the way things are going to go.

Marcus Lillington: Yeah.

Paul Boag: Yeah. So I think really that if you could, if there's one takeaway is really we need to start building our management habits, um, and our management skills before agents fully take over. Right. And, and take away that doing stuff. So it's going to be about getting organized, practicing delegation, learning to hold multiple threads in your head and all of that kind of stuff, which is a little bit depressing. Um, I know. I mean, what do you. Yes.

Marcus Lillington: Yeah. Okay. As I've always said, I want AI to hoover the living room. Really.

Paul Boag: Yeah. That kind of thing would be useful, wouldn't it?

Marcus Lillington: But machines have been taking over what humans do for the last 200 years, 250 years, probably. Uh, it's just a furtherance of that. We become more.

Paul Boag: We.

Marcus Lillington: We are users of those machines. Um, and less. And we understand less of how they work. Um, yeah. Cars is a great example. I get in my car with no idea how to fix it because no one expects me to, whereas 100 years ago I would have had to have fixed it on the side of the road. It's probably not a great analogy, but, um, not really. We are doing. Yeah, but I'm enjoying myself, Paul. You know, it's.

Paul Boag: I think the fun. The, the problem here. Let's. You. Let's use your analogy and keep going with it, is that if you enjoyed fixing the car at the side of the road.

Marcus Lillington: Yep.

Paul Boag: You don't get to do that anymore.

Marcus Lillington: No.

Paul Boag: Because you open the bonnet and you've no idea what's going on under the hood.

Marcus Lillington: Yeah.

Paul Boag: Um, and, and that it's sad on one hand, but it is the kind of way that life is. And I think it's coming to terms with that and finding joy in what you still can do. So uh, let's take design for example. Right? How much design, uh, there is. Okay, so you say you, as a designer you enjoy the process of making design. What do you actually mean by that? Or, or you and writing a, uh, you know, actually writing. Is it the typing on a keyboard, Is it using Figma or is it crafting words or crafting a design? Because if it's the latter, I don't think that necessarily needs to be taken away from you. At least not entirely. So let's take a look. The, the, the writing example. Right. So I love writing. You know, I do. I've written books, I blog weekly. I don't do the right kind of writing. You do as much. I don't enjoy that. But I have spent a lot of time and got a lot of joy of teaching AI to write like me, to analyzing and understanding how m. I like to word things, how I like to explain things, that kind of stuff. And I've also taken joy when it comes to where I actually get the joy is from communicating my subject with passion and enthusiasm. And so now I can still do that. I can brain dump and enthuse and excite, but I'm not having to type all the shit into the keyboard. Yeah, right. I'm, um, not having to worry about grammar or spelling or that kind of thing. And I think with design as well, the same applies that, uh, I've actually, for me and um, uh, everybody's different, but for me I've almost rediscovered my design roots a little bit that um, I don't have the time to sit around to push pixels in. In Figma. And so that was robbed from me in a way as I, you know, moved up and started doing other things. Um, but now I can come back and explain my vision, explain what I'm after to an AI. It can go away and work on it for a bit, come back to me and then I go, no, that. And I know I'm not going to offend them like I would a junior designer. And I can, you know, I can say, oh, I want this and that and the other. And uh, it goes away and does it. And actually that's quite liberating because it's freeing me from that time constraint. It's freeing me from the fiddling around with, oh, uh, how'd you do that in Figma again? And you know, so that there is still joy in it. I think the problem becomes where your joy comes from. Not the output, but the tool. Right. So I know some people love to be a master of a particular tool. Right. The master of figma, the master of the guitar.

Marcus Lillington: Uh, right, yeah, yeah.

Paul Boag: And, and I can understand that because a lot of bloody work goes into doing that. But if you can get your joy out of the design or the music rather than the tool, then you're going to be okay because it's just a different tool you're using.

Marcus Lillington: M. I wonder that. Well, we've had this conversation before and it's when, when you start dipping over into the art side of things. But it's kind of connected as well is will, how will the next you come along that knows about design? Um, or how will the next thousand of you come along who knows about design? Um, will humans stop pushing what good design is? Um, that is a concern. I know it's a bit deeper than we're talking about here.

Paul Boag: Yeah, yeah. I've got a lot of concerns at that level about AI, at ah, societal level.

Marcus Lillington: Mhm.

Paul Boag: But I can't do anything about those.

Marcus Lillington: Music is a really good example. I mean the good thing about music is it's still kind of taught a little bit in schools and people associate, you know, playing the guitar with being cool or you know. Yeah, it's a bit rock and roll or whatever. So I think people are always going to be, they're always going to kind of go to that or migrate towards that. But it's like good design that kind of came out of doing being a junior designer, didn't it? And

Paul Boag: no. Yes it did. But it doesn't mean it has to be that way necessarily. Yeah, I'm just, I mean, I don't know. I don't know because it hasn't been done yet. But I think people, you know, if I go back to where my interest in design comes from comes from being at school and you know, and when I was at school I uh, used to love doing art but I actually wasn't very good at ah, painting or drawing. Right. Yeah. I still got an enjoyment out of it. And that was when one of my art teachers said, oh, maybe you should be a graphic designer. Because I had a. I, I used to get hugely frustrated because I couldn't produce what was in my head.

Marcus Lillington: Yeah.

Paul Boag: Because I didn't have good enough hand eye coordination. It's probably the same in music as well. When you start out, you've got this really great tune in your head and

Marcus Lillington: it's still like that.

Paul Boag: Paul never goes so, so really you could argue that these AI tools are potentially uh, making it easier to express those artistic tendencies and those desires in us just the same way as, you know, creating a design back in the day would involve having draftsmanship shift skills and, and, um, being able to deal with, you know, print mechanisms and all of these kind of complexities that, that have gone now and the barrier to entry is much lower. So, so maybe, and I don't know, but maybe AI is just another step in that. That actually will see an explosion of the creativity because people that. That struggle with the tools again, um, are able to express themselves in a way that they weren't previously able to do. But I'm, I'm. I'm ridiculously optimistic and I know I probably need to be more cynical over this stuff because, you know, unless there are people shouting warning bells, you go down dystopian paths. But I'm a positive person.

Marcus Lillington: Yeah. I mean, I won't go down this, this route, but from a musical perspective, taking away the craft is. Can't be a good thing. I mean, maybe it could, but I can't see that how that's good. Um, maybe not.

Paul Boag: By the cr. What do you mean by the craft? The. The actual ability to play an instrument, for example, is that.

Marcus Lillington: Yeah. Um, I mean, why you can go in and. Using AI music tools, create music. Um, and it's. It feels wrong.

Paul Boag: Yeah, I know, but this is the weird thing, right? There's a. There was a great story doing the rounds recently about somebody who posted an, um, AI generated picture of Monet.

Marcus Lillington: Right?

Paul Boag: Monet, you know, a painting in the style of money, and said, can you guys help me to understand why this is not as good as the real money? Right. Um, and. And people wrote entire essays talking about the flow and the this and that and, you know, and, and trying to pin down what it was that was wrong about it. But it was a real Monet picture. So. So a lot of it is in. It's a very hard thing to quantify. And I wonder whether it's a degree of just a natural resistance in ourselves to change and how much of it is real. And that's, that's where. That's where I struggle. You know, I think it comes from.

Marcus Lillington: I was going to talk about the fact that AI's AI being creative can only come from what it's learned from human inputs over the last however long. Um, although I was, I was hope. I was kind of stopping there because I don't know if that's actually true.

Paul Boag: I. It's something that's bandied around a lot. I would Say it's more true if the AI is working in isolation by itself. But I think, and that is, and to some degree it's true of humans as well, right. We're building on the shoulders of giants and all the rest of it, but we, we can make creative leaps that I don't think I can do at the moment. But if you put an AI and a human together now it begins to get interesting because the, the human is able to express desires and ideas and the AI can help you realize those that to me, upsetting aside all of the other bad things about AI of which there are many, um, but you, that is a, that is a good thing in my mind that, you know, I, I, I, let's take music, right? I have no sense of rhythm whatsoever. I've got appalling hand eye coordination, right. Which means I'm never going to be able to make music. But I could with AI, right?

Marcus Lillington: Mhm.

Paul Boag: And, and I've got no desire to now because it's been ripped out of me, right. Over my life of constantly failing because of the tools. Right. And um, because of my own physical limitations. But the young me, if I could sit down and describe what I had in my mind and play around with different ideas and remix them, not just press a button and get something churned out, but, but interact and get a sense of satisfaction of going, no, I wanted, I wanted this. And isn't that a good thing?

Marcus Lillington: Yeah, I mean you can do that as kind of a dj and you could have done that, you know.

Paul Boag: Yeah. That wasn't so much of a thing

Marcus Lillington: 40 years ago, you know, Maybe M.

Paul Boag: Yeah, I guess so. Yes. You can do it with it. Well, is a DJ then a musician?

Marcus Lillington: Uh, yeah, yeah. An artist?

Paul Boag: Yeah. All right, cool. Yeah, you can, you can work. And that's what I see, uh, see this is beating, right. I mean the other example is photography and we've talked about this one before, haven't we, that, you know, my dad started doing photography just as single lens reflex cameras came along where, you know, and he took pictures of diving birds where, uh, you had to press the button at the right bloody moment.

Marcus Lillington: Yeah, right.

Paul Boag: There was effort and skill and blood and sweat and tears went into that. These days you can hold the shutter down and it takes a series of photographs and you pick the best one and then manipulate the out of it in Photoshop. Right. And for that generation, that current generation, that's photography, right. Well, my dad's generation is going, well, that is a proper photography. You're cheating. Right? What the current generation are looking at people that manipulate it with AI and go, that is a proper photography. They're cheating. So it all depends on the baseline you were introduced with.

Marcus Lillington: Yeah. Uh, well, yeah. Yes, that is an element. I think there's, There is something about the fact that AI has, uh, based its decision making off other people's hard work. Yeah. There's an issue there that makes me feel.

Paul Boag: That's a squirmy. That's a. Yeah. And that. That's the problem. I think that. And rightly so. A lot of people are unhappy with AI for lots of other reasons. You know, I said that AI has had a lot of problems. That is one of them. Just a blatant disregard of anybody's intellectual property, let alone the environmental considerations. And that there are loads of issues with AI, um, and I'm not trying to defend any of those. I'm just saying that, you know, it's already ripped off everybody's work. Right. Do we collectively refuse to use it because of that? I don't know. Can we. Is that actually a feasible thing to happen? Can you put the genie back in the bottle?

Marcus Lillington: Uh, well, I don't think we would just. We wouldn't collectively decide to do that unless literally the forest started burning. Yeah. And even then, probably not.

Paul Boag: Yeah.

Marcus Lillington: Which is a scary thought.

Paul Boag: We've never gone back on the technology that we've introduced, I think. I suppose nuclear weapons would be the closest. Wasn't it where. Where, you know, we used them a few times. They exist, but in theory are not being used. But anyway, we're getting very.

Marcus Lillington: But it's interesting, man.

Paul Boag: Yeah, exactly. But, yeah, I, it's the attitude of, do you resist the future or do you try and pick out the best bits of it? Yeah, and, and, and I respect both points of view, but I'm very much of the latter. Um, and you're probably somewhere between, aren't you?

Marcus Lillington: Yeah, I think my, my concern here is that this feels a bit different. And I can't quite put. It's because of the. This is based on all of human knowledge. It feels like you have suddenly have this access to that. It just.

Paul Boag: What, a bit like the Internet then?

Marcus Lillington: Yeah, but you could ask. You could basically get something to review. I'm exaggerating. But all of human knowledge very quickly. Um, and for a human to have done that using Google, you just give up. It's never gonna happen. So it just feels. I don't know, it feels like, yeah, more concern should be paid here. But maybe. Oh, yeah, maybe I'M just being a Luddite. I don't know.

Paul Boag: No, no, no. I think even, you know, you could argue, make the argument you've just made about when the Internet came along, when the Industrial revolution came along, etc. But that doesn't mean that you don't want people questioning it and challenging it. You absolutely do to, to curb the, the worst impulses of these things. But on which, on which subject, I do want to move on because obviously you have to balance my relative enthusiasm with, with a negative. Um, interesting read of the month that bashes AI so you go for it. Marcus, you do tell us.

Marcus Lillington: I've got a couple here. But because we're, we're rushing through today, I'll save the second one for another day. Um, this we mentioned last month. Oh, um, I think you mentioned it earlier about. I can't. What did you call it? The AI burnout. Burnout. That's what it was.

Paul Boag: Yeah, that's what I was saying today. Yes, I've just said that. It shows how much attention you give me.

Marcus Lillington: I just couldn't remember the words because again, age, book. What was it he said?

Paul Boag: What you need is an AI to remind you of what's just been in my ear.

Marcus Lillington: I need some of those, those meta glasses, don't I?

Paul Boag: Yeah.

Marcus Lillington: Um, anyway, so I found the article. Uh, it's called Life with AI Causing Human Brain Fry, so. Brain fry. That sounds quite, you know, like a, an old sci fi film, doesn't it?

Paul Boag: Anyway, the Boston sounds like a click baity headline. There's probably good stuff under it. Yeah. Um, so mental.

Marcus Lillington: Yeah, basically this, this was the article that I was referring to at the time. The Boston Consulting Group, um, has coined the term, um, AI brain fry, uh, which is they say is mental exhaustion caused by heavy use of. Use or supervision. Sorry, supervision of AI tools that pushes people past their cognitive limits. So we've just had Paul Boeg telling, telling people that you should be supervising your AI tools. But not so sure.

Paul Boag: But I did it. I acknowledged this. I acknowledged it. And I suffer. I seriously, I suffer from this. I agree with this 100%. I'm just saying it's go. We're going to need to learn to adapt because I think it's inevitable. I don't think it's good.

Marcus Lillington: Yeah. All right, let me carry on. So, um, it continues. Heavy AI users report feeling overwhelmed by tasks like reviewing endless AI generated code, managing multiple AI assistants and drafting lengthy prompts. Drafting lengthy prompts. Yeah, I guess if you have to keep doing it again Change it because it's not right. Um, the problem is hitting software developers hardest since AI agents have rapidly become proficient at writing code. But that code still needs careful human review for security flaws and overall coherence with the code base. Um, this article actually came from one of Headscape's developers, um, who was basically saying, look, look, this is, this is why. It's why AI is bad.

Paul Boag: Um, and they're, they're right. I entirely agree with this. Everything you've said so far, it is, it is a real problem.

Marcus Lillington: Um, but they go on to say so. They recommend that company leaders set clear limits on how much employees use and oversee AI to prevent burnout. And that was their conclusion, really. Um, quite what that means, though, I'm not sure, because company leaders quite often don't think. Uh,

Paul Boag: just don't think. Stop. Yeah, yeah, yeah, yeah.

Marcus Lillington: But they are more interested in output and, to be fair, things like that.

Paul Boag: So how. How are they supposed to know that this is a problem unless we tell them? Yeah, right. You know, it's easy in my situation because I am both the employee and the boss. So, you know, when. When I begin to struggle with AI and I begin to get over the pub, I go down the pub, you know.

Marcus Lillington: Yeah.

Paul Boag: But if you're an employee, you don't get to do that in the same kind of way. And so that's a. I. I think we need to get better at highlighting articles like this and sharing it with management and explaining what the problem is. And, you know, it might even be that we need to be recording just snippets or telling stories, like I did a minute ago about how I had, you know, Claude design doing one thing, I had this doing another, and I was kind of back, um, uh, jumping between those. The, the only problem that you face, to be honest, um, if I'm being a cynical senior manager, is I'll turn around and go, well, that's what my life is like every day anyway. Because as a senior manager, it is. It is constant going from one meeting and one discussion into another. You know, that's what I do.

Marcus Lillington: I always have done.

Paul Boag: Yeah.

Marcus Lillington: And even though I'm not particularly organized, I am quite good at keeping things in my head. Yeah. Uh, um. But not everybody is, I guess. Um.

Paul Boag: No.

Marcus Lillington: We don't have any answers, do we, Paul?

Paul Boag: No, none whatsoever. M. So you've just brought a big gray cloud onto proceedings with no positivity associated with.

Marcus Lillington: I think we had some gray clouds before we went to that bit.

Paul Boag: Yeah, that's true, but it's you and your curmudgeonly, um, uh, Luddite ways. That's what it is.

Marcus Lillington: Yeah. I just want to go and grow tomatoes, Paul. That's. I don't know.

Paul Boag: Yeah, well, I think that's probably going to be the best way to, to um, you know, to make money in the future. Be a plumber or electrician or something like that.

Marcus Lillington: If, if I was 25 and thinking, oh, this music business larks quite tricky, I'm going to go and do something else. I would train as a plumber or an electrician. No question, no question.

Paul Boag: I, I was talking to my daughter last night and who's doing computer science at university and she can't get, she can't get a year's placement. There's just nothing at the moment and I do worry about her. Fortunately. She's one of those quite laid back, flexible kind of people that will just, yeah, take punches as they come. But it, you know, it's not, not a great way to start, is it? At the moment?

Marcus Lillington: No. But I mean, but there are, uh, if you do have, ah, you know, if you're an engineer of any sort, which might be a, you know, a domestic plumber, but it could be somebody building, I don't know, pipelines across the Atlantic, you're going to have endless work. Um, but if you're.

Paul Boag: Until, until the robots get, well, you know, to it. So maybe another 10, 15 years away from that, I would have thought maybe.

Marcus Lillington: Yeah.

Paul Boag: Don't know, pulling that out of my ass.

Marcus Lillington: Robots doing plumbing, maybe. Not sure on that one. That would be really hard.

Paul Boag: Um, that's kind of. Yeah, yeah.

Marcus Lillington: Um, but yes, do something. Hands.

Paul Boag: Yeah. So, so basically we're all stuffed and thank you for listening to today's podcast,

Marcus Lillington: but there is one shining light of happiness. Paul, before we go.

Paul Boag: Oh, your joke. Yeah, yeah, I forgot your joke. Yeah. After that it's like basically we've just walked out of a funeral where now you're going to tell a joke that's effectively what is happening.

Marcus Lillington: And uh, yeah, you're gonna crack a joke because they would go down so well. Right, well, this joke, not only that, this joke needs an intro because I think it will. It's a UK focused, I think, and of people of a certain age. But it made me giggle so I'm gonna say it.

Paul Boag: That's fine. I'm from the uk. I'm, um, of a certain age. I'm, um, I'm all in.

Marcus Lillington: Right, here we go. Um, I guessed orange, but it was chocolate. I guess toffee but it was peanut, I guess. Strawberry, but it was coffee. I was wrong on so many Revels.

Paul Boag: Oh, that. That really is quite a niche joke. I was wondering where it was going.

Marcus Lillington: Anyone at our age. Paul. Knows what a Revel is.

Paul Boag: Yeah. And that, uh. Didn't they get renamed something?

Marcus Lillington: Oh, I don't know. I think they're still around.

Paul Boag: Kath of Revel's still around. Oh, they are still around. So there you go. So it's still relevant even though nobody eats them anymore.

Marcus Lillington: Yes.

Paul Boag: Yes. Oh, that's terrible. Right, There we go. That's. That's that one. Done. Another podcast out of the way. Hope people appreciate it.

Marcus Lillington: All the misery this is. This isn't even us. It's. It's AI Marcus.

Paul Boag: And, yeah, we replaced. We replaced ourselves. So you never. I can never imagine two AIs being able to just talk utter as well as us. I just. I can't imagine it.

Marcus Lillington: No, they'd be. Have you ever.

Paul Boag: Have you ever heard. Have you ever heard any of the AI podcasts that are created by, like, Google's, um, notebooks? They're so slick and so together. No AI could be as incompetent as us. So that's what our niche is, basically just, you know, incompetence. That's kind of true. I could work with that, really.

Marcus Lillington: It's the rough edges that make us human.

Paul Boag: Yeah. Oh, that's profound. Let's leave on that. All right. And on that. That profound revelation from Marcus.

Marcus Lillington: Bye.

Paul Boag: Goodbye.

Marcus Lillington: Hm.

Paul Boag: Sam m.