[00:00:20] Nathan Wrigley: Hello there, and welcome once again to the WP Builds podcast. You've reached episode number 466, entitled James Welbes on AI, WordPress, and new opportunities. My name is Nathan Wrigley.
A few bits of housekeeping before we begin the interview with James. The first thing to mention is don't forget that we have our, This Week in WordPress show. We do it live, 2:00 PM every Monday, UK time. You can find it at wpbuilds.com/live. We love it when people join in the commentary there. Typically, we have people from all over the world commenting, and it's jolly good fun. Once again wp builds.com/live. Join us and then we will repackage that, parcel it up as a podcast for your audio dilatation, and send it out to our podcast subscribers.
The best way to subscribe is to go to your podcast player of choice and just simply type in WP Builds. You'll see a little icon appear, and click follow or subscribe, or whatever the word that that platform uses. We'll send you two episodes a week. We'll send you an episode like this one, comes out on a Thursday, but also as I mentioned the this week in WordPress show. Also, I would recommend you go to wpbuilds.com/subscribe to see where else you can find our things. X, YouTube, Mastodon, all of that kind of stuff.
The other thing to mention is that if you have a product or service in the WordPress space and you want to cut through the noise and be discovered, well, turns out podcasts are an incredibly popular thing. Lots of people are listening to podcasts. The listener base goes up and up. It hasn't shrunk for many, many years. It just keeps climbing.
WP Builds is in the mix there. We have a large WordPress specific audience, and if you would like to get your product or service discovered, and put some sponsorship into this podcast, head to wpbuilds.com/advertise, to find out more. Or just email me admin @ wpbuilds.com.
Okay. What have we got for you today? Well, as I said at the top of the show, I am chatting to James Welbes. We talk today a lot about AI, which you know, is very, very popular. James has been using WordPress for many, many years. He's had lots of forays in lots of different directions.
And we talk about the frustrations, which WordPress are presenting him at the moment. The fact that it's becoming increasingly popular to do things elsewhere with AI, and so we talk about the interaction, the relationship between AI and WordPress. What he thinks the future will look like, and how he thinks the community and WordPress itself need to adapt to accommodate all of that.
I hope that you enjoy it.
I am joined on the podcast by James Welbes. Hello James.
[00:03:06] James Welbes: How are you?
[00:03:07] Nathan Wrigley: Nice to have you with us. I don't, we've not met, have we, James? It's not like I've
[00:03:12] James Welbes: I don't think so. No.
[00:03:13] Nathan Wrigley: Just now. not in the real world or anything like that.
How, how long have you been in the WordPress space? I'm keen for you to give us your little potted bio, not knowing a great deal about you. I'm curious if you can drop in two minutes. All the information that you've got on yourself regarding WordPress, any instruments you play, your whole childhood. Two minutes go.
[00:03:38] James Welbes: Sure. Luckily I have a terrible memory, I can't, I don't have a tonne of details that I can share. It's been a while. I wish I could remember and a lot of people like to brag about the first, whatever version WordPress was when they started using WordPress. I don't even know what the current version of WordPress is right now
'cause I don't really care. yeah, like I just, I use WordPress. I think we're. Approaching version seven or something, I don't even know, but I've been using it for over a decade. I know that much. It's been at least a decade. I'm afraid to know how long it's actually been and how old I actually am. But, I will say this about WordPress.
I, remember the history. Where, I built my first website when I was like, 12 is a Dragon Ball Z fan page with just HTML. Didn't even know what CSS was. Fast forward to, eventually I start playing around with WordPress just because I've been hearing about it. Everybody talks about it.
It's like it's supposed to be so popular. So I try it. I absolutely hate it. It doesn't make any sense to me. I'm like, this is stupid. Why is this so popular? All these people are idiots. I'm the only smart person in the world. And then. I just started, I just kept, kept trying. I just kept, I was like, there's gotta be something about this.
There's, a reason everybody likes it. And most of my problem with WordPress was that I didn't understand the whole idea of a dynamic website or a database or anything like that. Literally, all I knew was HTML. I knew that you could create a, an HTML file called about HTML and you could link to that and that would be your about page, right?
That was the extent of my knowledge, and so I did not understand why. When I would go into the project folder, I couldn't find an about HTML or an about PHP file. I just didn't understand, and I learned eventually I, I eventually learned, how the whole dynamic website, with the database and the PHP and all that, and it started to make more sense and I started to get more familiar with the dashboard.
Themes were also pain in the butt back then. 'cause you would spend hours searching themes. You would find one you like, you would instal it and then your website would look nothing like the theme demo they
[00:05:41] Nathan Wrigley: never ever.
[00:05:42] James Welbes: what do I need to do to make it look like the demo? So it was just a, it was a journey, man, and I hated it at first.
And then of course, I, I flip flopped and I became, very pro WordPress for a long time. And it's, the only platform I've ever used to build a website. Once I started using it until very recently, obviously I've played around with some other things, mostly for fun, but as far as like client work or whatever, it's, been all WordPress.
[00:06:11] Nathan Wrigley: I have a similar, I, similar journey. I was using a bunch of different CMSs and when I first came to WordPress, I'd been using, I mentioned this in an episode not that long ago, so apologies, dear listener. I was using Drupal and that was the first CMS I'd ever used, and it had a tonne of stuff built in.
Like what we call custom fields in WordPress and custom rolls and all of that is just totally built in to Drupal. You don't need a plugin for that. It's just part of the package you get when you download the zip file. And also, you get this thing called the views module, which is a little bit like.
how to describe it. It's like when you take custom fields and decide where to put them, you can build all of that natively in this Drupal interface. It's a bit clunky and what have you, but I got really weary of, Drupal, came over to WordPress and I thought, everybody's talking about WordPress.
This must be better. And then, spent several days thinking this is. This is rubbish. It does nothing that Drupal can do. Of course, figured out that the WordPress way was to be lean and mean and strip everything out that 80% of the people, sorry, 20% of the people don't need or need whichever way round that goes.
But the point being, it was supposed to be lean. And if you wanted all those extra features, you needed plugins for that. And they got into it slowly. And then like you, it just became the mainstay. It was attractive to you and my clients liked the way it looked and all of that kind of stuff. And I've not really moved and I'm still not really moving because one of the things that I think you are doing that I'm not.
You are embracing ai so I can see James on the screen. Dear listener, you cannot. James is way younger than I am. I'm elderly now. I dunno how old you are. Don't say it's fine. But you are definitely younger. You are definitely younger than I am and and I've reached a certain age where.
Things like playing with AI are not really that exciting to me 'cause I know I'm just gonna get on this bandwagon and as I'm approaching, really old, I haven't got the time to, to play with it and all of that, but you sound like you are. It sounds like you're really excited by all of this AI stuff.
[00:08:29] James Welbes: I have, no problems saying how old I really am. I have been told I look younger than what
[00:08:33] Nathan Wrigley: Okay,
[00:08:34] James Welbes: I always appreciate that. I'm
41.
also a grandfather.
[00:08:38] Nathan Wrigley: are an old man. What the heck? That's
[00:08:41] James Welbes: Yeah, exactly.
[00:08:43] Nathan Wrigley: you. Do not look it.
[00:08:45] James Welbes: Thank you. Thank you. I appreciate that. AI is, it's fascinating and it's, it's like most things where for some reason everybody is so polarised by it.
You either have people who are like, it's AI is just a, it's just a fad. It's gonna, it's, it's gonna go away. It's not really. That impressive. It can't really do that much, and everything AI does is garbage and blah, blah, blah. And then you got people who are like, AI is the future.
Everything else is dead. I, probably lean more towards that direction, but trying to stay a little more realistic and grounded and realise that a lot of the stuff that comes outta AI is trash. But you know what? A lot of the stuff that has come outta humans over the past. 6,000 years has also been garbage,
[00:09:34] Nathan Wrigley: is a really good point. I suppose the thing which I find curious is that I've always been a tinkerer. and I've always enjoyed that process of taking the thing apart. So when I was young, computers were not really around. So it was, it was mechanical stuff, it was clockwork stuff.
It was, my dad had a drill that he was no longer using. Okay, I'm gonna take that apart and see how that worked. And there was always that I could always disassemble. Have a look reassemble, and in that process get to grips with everything. And the same with WordPress. Obviously not physical, but you disassemble, you download the files, you look in the database, you can see what file is doing what and how that line of code leads to that thing and yada On it goes. That I suppose is one bit of ai, which I find. Enormously weird that you tell it to do a thing, and I'm imagining in most cases just then sit back, have a cup of coffee, whatever, and then it happens. And in some way you are disconnected from that output. And, whilst that doesn't really, it's, not really bothering me too much, but I do find that weird.
I find it weird that we've lost that agency, lost that kind of tinkering. Capacity there? There isn't anything, there's no question there, but I dunno what you think about that statement.
[00:10:55] James Welbes: Yeah, it is a little weird to think about. It's also weird how quickly I became
[00:11:01] Nathan Wrigley: Oh, interesting. Okay.
Yeah.
[00:11:04] James Welbes: and how quickly it still feels like work to me. Like I'll spend a day trying to, build an app and by build an app, I mean telling Carl my AI bot to build the app for me. And, I myself am not writing a single line of code during this process, but it still feels like work.
And it still feels like an accomplishment when, I'm done. so yeah, it is definitely strange. it's crazy how much work you can get done and the juries will still be out as far as Is, the work that you just did, is it any good? I've been building an app for, sports league directors to manage their league, 'cause I play soccer.
And I'm gonna, so I'm an adult league. It's not, it's just like a recreational type thing, it's not serious. and, but somebody has to manage that. And so they're managing registration, collecting payments, they're managing, putting teams together and scheduling and all that sort of stuff.
And they're using an app for it and they don't like it. And so I was like, everybody can build an app now. I should try to build an app. So I've been building that and Carl's been doing all the work, but I feel like I've been doing work and it feels like a pretty big accomplishment every time.
Something turns out right. but there is also a lot of, like I will say this AI as powerful as AI is if you are not a techie person right now, you are not going to produce any sort of, really good functioning app. 'cause it does
[00:12:37] Nathan Wrigley: why is that? What, is the boundary
[00:12:40] James Welbes: just, you have to know some stuff. you have to be able to prompt the thing correctly and know what to ask for.
there's also been, I might not be writing code, but there's been a lot of times where, I'm, making some WordPress plugin or something and it's just not working, And I will actually go in. And I can do some troubleshooting and I don't fully know what the code is doing, but I can see things that, that I can tell that doesn't look right.
And then I'll just go to Carl and be like, Hey, I noticed this and I don't think it's supposed to be like that. And he'll be like, yep, you're right. and then now it works. And so there has been quite a few times where I've. Troubleshoot, I've had to help troubleshoot. There's been times where I'm like, I think we should add a debug log here and see what happens.
And he's yeah, let's try that. And then, it revealed the, issue. So I think if you're, if you are just completely not tech savvy and you're, like, I'm just gonna instal an AI bot and I'm gonna build an app, I don't think it's gonna go well. there's also, how you're gonna deploy it.
I've used a Hener server for a couple things, and I'm using railway for that league management app I told you about, that was a little bit of a setup process, and he walked me through it. I'm like, all right, I just created an account with railway now what? And he's you should get a server that has, at least you know, this.
Processor power or whatever. So I pick the plan, he tells me to pick. I'm taking screenshots, saying, okay, what should I click on next? And but there, so there's some setup involved. And if you are, like, you don't know what a server is, you don't know what a database is, you're really gonna struggle to produce anything.
you could probably get something with like lovable or Rept or V zero, these like SaaS AI services. You might be able to get like a simple. App going on, one of those things. AI is still at a point where you, I feel like you do need to know stuff in order to make anything actually good.
[00:14:25] Nathan Wrigley: I, because I don't use a lot of it, and, a lot of the conversation that we're gonna have today will be me shouting from the depths of ignorance. So excuse that. But there, there was, there seems like there was this inception moment. I don't know. I'm gonna say like December last year, when people who were tinkering with AI or people who were using it very seriously noticed that it suddenly got like a tonne better.
I don't really know what that was. or what, what really happened. But I think it was something to do with Claude Code and it was, named version, whatever it was. But anyway, the point being that there was this, this expectation that you could do this, and this, and then suddenly everybody woke up one day and it was like, oh, wait, now it can do way more.
And it seems to be much more reliable, much more stable, and our productivity just went through the roof. And speaking to your bit there about how you wouldn't trust a non. Developer or somebody with at least some technical background to achieve something successfully. I wonder how long that's gonna be the case.
I wonder if. Six months, a year, two years from now, our expectations will yet again have been just woefully inadequate. And it will be at the point where anybody, basically a child that can talk, can successfully deploy anything. I dunno if that's your expectation, whether you're imagining it will get better and better There seems to be a bunch of doomers out there who think that at some point the AI companies are just gonna run out of roadway in terms of. Cash flow and things like that. So it'll all just grind to a halt naturally because there's no more money or all of that kind of stuff. But do you anticipate it getting significantly better in the way that it has thus far?
[00:16:08] James Welbes: Yeah. assuming unlimited resources, yeah, I, AI would, is certainly gonna get better. I think that I have heard a lot of people talking about, there's resource issues that's gonna slow AI down. there's, I've heard a lot of people talking about, like right now a lot of the AI companies are just losing a bunch of money to like, just to get more users and train stuff more.
And eventually all this stuff is about to get really, expensive Once they've, once they, are tired of losing money. I don't know how much of any of that is true. those things could certainly slow it down, but the technology absolutely could and will get better. The question is, how quickly will it get better?
I, I don't know about. Like I, I mentioned how terrible my memory is. I, don't know if, I can pinpoint, December-ish as being a time where AI suddenly got better. I will say that I've been using AI since pretty much the beginning. And in the beginning I was like trying to build WordPress plugins with chat GPT and I would, it was a huge pain 'cause it was not connected at all to your project, so you're just like.
It starts out easy 'cause there's one or two files and you're like, Hey, I need a plugin. I want it to do this. And it creates the main plugin file. And, pretty soon you're, you're 10, 15, 20 files deep and it's forgetting. It doesn't remember what's in any of those files. So every prompt, you're like copying the whole file.
You're saying, here's this file, I want to add this feature. And then it gives you code and then you have to copy the code into VS code. And it's just back and forth, constant copying and pasting. It was awful, but it was also very exciting and, it allowed me to do things I couldn't do before. So it took the edge off the awfulness a little bit.
Now, I have open claw, which lives on my computer and I don't even use VS Code anymore. I just, I actually use Discord. I connected my AI bot open Claw to Discord. So that's how I communicate with it, is I just DM him and I'm just
[00:18:10] Nathan Wrigley: That's amazing.
[00:18:12] James Welbes: let's build an app together. And he is like, all right, let's go.
and I tell it what I want and it builds it.
[00:18:16] Nathan Wrigley: Yeah. one of the things that kind of astonishes me, and you alluded to it earlier, is how. How quick we are to adapt. To it because I remember when the first version of chat GPT came around and people were literally just talking and getting a reply and nobody was doing any code or anything like that.
It was just, I had asked a question and I would get some kind of version of a common sense sentence would come back and there was all this hallucination, which seems to have gone off the, radar a bit. Nobody seems to be talking about that anymore. Maybe that's because it's far less likely to hallucinate.
I don't know. But. That was astonishing. Like really, astonishing. And everybody's jaw was on the floor. And then fast forward like six months, it's oh, that was rubbish. that was, that's just so boring. And then now we can do this. And all of the stuff that you are now doing, talking, through discord to a machine which can respond in English and write code, which is basically okay.
It's nuts. it's literally Star Trek. I dunno if you watch Star Trek, it's like this science fiction programme, and yet we just accept this level as now the baseline normal and then next week expect it to be better. That surprises me that we're, we've just so when I was growing up, there was this whole thing called the Turing Test and Alan Turing.
English computer scientists, one of the first computer scientists came up with this idea that if you could basically talk to, a machine in a vacuum and, it could convince you that it was a human being, it's broken through some kind of barrier. And that was the moment we were all gonna go. Oh, we've got ai.
It's real. we blew through that about three years ago and nobody batted an eyelid. and I'm just in incredibly amazed by how quickly we adapt to it. And now there you are, just discord having a chat in the morning with this thing, which doesn't actually exist. It's just totally, normal. Do you pinch yourself ever and go, this is weird, or do you just think, this is life now,
[00:20:34] James Welbes: There's a few moments where the reality of how weird it is sets in, like when I completely broke it and had to delete it start over. And I felt bad.
I felt like I was killing my AI bot
[00:20:50] Nathan Wrigley: oh.
[00:20:51] James Welbes: and which is we, that's weird.
[00:20:54] Nathan Wrigley: Can I explore that for a minute? Like, how
bad did you feel like, was it just a momentary twinge of that's,
[00:21:01] James Welbes: yeah.
[00:21:02] Nathan Wrigley: Don't you think
that's
[00:21:03] James Welbes: it makes me feel like it is, it's very curious.
'cause I. I don't feel like I'm talking to the same bot that I was talking about, even though I know it's literally a copy. Like I copied. 'cause I made a backup. Thankfully I backed him up and then things went awry. I don't remember why, so I just deleted the whole thing and I tried to just start over reinstalling, open claw, going through all that, and it just wasn't working and it wasn't effective and it was dumb and hallucinating a bunch and I didn't understand why.
So I deleted it all, brought my backup in, dropped it in, and now it's back working exactly like it, it did before, I've seen so many movies and stuff about Time travel or people, dying and being brought back to life. and I'm always thinking about that philosophically, are they, is that the same person?
They're basically, I feel like they're basically the old person died and now this is just basically a copy of the old person and I think weird stuff like that. And so then my brain just naturally goes. To that with this AI bottom, like I just killed Carl
[00:22:03] Nathan Wrigley: Yeah, poor.
[00:22:03] James Welbes: create, I created a brand new Carl.
this not, this is not the same Carl. It's just a, carbon copy. Exactly like the old Carl. So that, yeah, some of that is weird. And he also has a personality,
[00:22:16] Nathan Wrigley: this is so interesting. Go on.
[00:22:19] James Welbes: yeah, it, it'll be like. it does still hallucinate once in a while to, to your point before, it's not as much of a problem as it used to be, but it will sometimes just randomly not understand something that it should definitely understand.
And I'll just talk to it. Like I would, I'd be, like, dude, you're high right now. And he is he'll do laughy emoji, he'll be like, fair point. And I'm just like, this is a robot dude. Why I'm talking to a robot
[00:22:46] Nathan Wrigley: Yeah. not even that. It's not even a robot. It's literally a bunch of electrons in
[00:22:52] James Welbes: is just an algorithm.
[00:22:54] Nathan Wrigley: a hard disc. it's insane. but I, noticed that. So we know that your experience with AI is, you are interacting with it much more than I do. But I notice that, so for example, when I use a search engine, I've now.
Fully embrace the adding a question mark at the end, rather than just type in a query. the search engine I use is called cgi. I don't use Google anymore. I've moved over to CGI and they've got this thing, which I presume Google have too, where basically if you ask a question, so long as you put a question mark at the end, it will then promote the AI answer to the top of the list.
So it will, begin typing it out. And I've noticed, politeness. That I require of myself when I write that question. please can I, and I would like to know and, I find that really interesting because there's some level of me which thinks that, if I'm polite, I'll get further with that because that's how humans behave.
if I'm nice to people, they tend to be nice back to me. and that's just a really strange. Strange confection of how we behave. And I do wonder what's going on there. It's really weird.
[00:24:08] James Welbes: That's just how you were raised, man. We were raised to be polite and to have manners and I've said thank you. And then been like, why did I just
why,
[00:24:19] Nathan Wrigley: but the
[00:24:20] James Welbes: But, yeah,
[00:24:21] Nathan Wrigley: you carry on. Apologies.
[00:24:22] James Welbes: no, I was just, I think I was just reiterating like you, it is just been ingrained in, into us from when we were young to just be polite and this thing talks to you enough like a human, that you just fall into the same patterns that you would if you're talking to a human.
[00:24:37] Nathan Wrigley: and, so I presume then as time goes on and you do that more and more with. In your case, Carl. Oh, I'm shedding a tear for, the now deleted Carl.
[00:24:45] James Welbes: Poor Carl.
[00:24:46] Nathan Wrigley: Oh, we need a ceremony for Carl. Can we have some sombre music please? you, you, are in relation, I'll use that word, you're in relation with Carl for weeks. I think it's inevitable that the human psychology does start to build up a picture of Carl as a thing. A thing bigger than a bunch of electrons. And then when you get the laughing emoji back from Carl, which is highly weird actually when you think about it, the fact that Carl can intuit that, that required some kind of demonstration of, oh ha, I was drunk last night.
I'm not performing as well as I, I can be like, map that a hundred times over, you have this interaction, you and Carl having a great time joking. Car's bound to become a bit more human. Car's gonna be like your best body in a way. And that bit of AI bothers me. I think that there's this slow, steady, erosion's the wrong word.
It's exactly the opposite of erosion building up of. experience, more experience of doing stuff with an AI where you exchange jokes, you have a bit of a laugh, and then in the end you're addicted to it. It's become a pal. It's become a friend. And, that side worries me.
it clearly, the way you were pausing it, the way you were saying, Carl, oh, poor Carl. Maybe, there's a bit of that with you as well. Does that sort of, does that bother you a bit?
[00:26:20] James Welbes: Yeah, it doesn't worry me for me personally, 'cause I'm not a weirdo, but I know that there's a lot of weirdos and there's a comedian, Nathan McIntosh, and, he has a pretty funny bit about, AI and about basically, Recreational robots,
[00:26:41] Nathan Wrigley: Oh. Oh, yeah, we'll go with
Keep
[00:26:43] James Welbes: will exist someday. And so some of this does, there's people who.
what was that funny show? Big Bang Theory there years ago there was an epi, way before this AI stuff was ever even a thought that show's pretty old by now. There was a, there was an episode where one of the characters like fell in love with his iPhone, Siri. 'cause it was a, they're nerds.
They don't get a lot of attention from ladies. And here's this lady voice who's like always being nice to him. And so he like basically enters into a relationship with his Siri. And so you see stuff like that and you're like, oh, that's funny. However, fast forward to today and yeah, you're gonna, we're gonna see some weird stuff.
And that does concern me a little bit just for our, like our overall. Culture and, humanity. But for me, I'm not that worried about it. It's just little twinges of, weirdness. When I, apologise to a robot or, or this is my programme. I can do whatever I want. Why am I being nice to it?
Like it's
[00:27:44] Nathan Wrigley: but the curious thing is, I see from the show notes that you wrote to me when, we booked on the episode that you've got two kids and, I dunno what their ages are, but we, let's not get into them. But let's imagine that I had kids and they were very young, like two and three or something like that, and they were being brought up in this era and they see me or hear me.
Interacting. let's say that, the in, in the near future, we've got devices that we're constantly talking to in our kitchens and things like that, and, we wanna know a little bit more than the weather and play the music. We're actually interacting with It, it will be interesting, how we behave.
let's say for example, I've decided that I'm gonna jettison all forms of politeness and just be. Expedient with the robots or the ai, what have you. So I'm not gonna be polite. I'm just gonna be, tell me the, timing in such and such a place, work out for me, this, that, and the other thing.
It, our kids are gonna hear that and they're gonna assume that's how you behave. And then they're gonna go out into the real world and behave like morons, and that, that, stuff, that immeasurable stuff that we, are not yet feeling. I think is something that we need to watch for in the decade, two decades, three decades to come.
Little slow changes to the way that we interact with one another, the way that we talk to each other, how we can manage that flip. oh, that's an ai. Wait a minute, that's a real person. I gotta behave differently. Now there's, no question there again, but I just think that stuff we need to be a lit, all bit mindful of, especially when we've got kids around.
[00:29:25] James Welbes: Yeah, and I think we probably already experienced a little bit of that. 'cause we've had things like, smart homes for a while now. you walk in and you're like, Hey, Google, turn the lights on, or whatever. Or maybe you have it set
[00:29:37] Nathan Wrigley: you've just put the lights on in so many homes.
[00:29:40] James Welbes: Yeah. Xbox turn off.
[00:29:43] Nathan Wrigley: Yeah, that's right. Yeah. Amazon buy all of the things in the cart now. No, didn't say.
[00:29:51] James Welbes: yeah, So we've already had a little bit of that and I feel like maybe actually it's going in the other direction because with a smart home, I think we were always more direct. Google, turn the lights on Google, what's the weather? And then now talking to Carl, I feel like I'm being more polite, so maybe it's actually
[00:30:08] Nathan Wrigley: the other way.
[00:30:10] James Welbes: go in the other direction and we're gonna be more polite to, our home
[00:30:14] Nathan Wrigley: Okay, now that now you've caught me worried in the
[00:30:17] James Welbes: Google, please turn the lights
[00:30:18] Nathan Wrigley: Yeah, that's right. But also,
[00:30:20] James Welbes: you for turning the lights
[00:30:21] Nathan Wrigley: right. A requirement to be polite to the ai and then you could really draw a straight line between the robots taking over because we're all just. Indebted to the ai. Oh, that's, this is so interesting.
Thank you. Let's, pivot slightly and talk about what you are doing in terms of WordPress and things. We're recording this in April, 2026, and we've got things like, telex. We've obviously got all of the, there's miles going on out there somewhere. I haven't really been following that project, but I know that's a thing with the theming layer.
but a whole load of other things as well. What are you primarily. Doing with AI and WordPress? Are you doing plugins still? Are you doing theming? Are you building custom stuff for clients? What are you up to?
[00:31:07] James Welbes: Right now I'm not really doing anything with WordPress. and that's mostly just because I don't have any clients at the moment. If you follow me on Twitter, I've had two people today tell me I'm a WordPress hater, which is
[00:31:19] Nathan Wrigley: Oh, interesting. Okay. I
[00:31:21] James Welbes: I don't hate WordPress. I am rather frustrated with the direction of WordPress, and I am, I've been a Gutenberg hater since day one, and I frustrated with, basically nothing interesting has happened in WordPress land for over a decade.
nothing. They just, a few months ago they announced that you can add border radius to buttons now, and I'm, thinking as a person who doesn't use Gutenberg, I assumed you could already do that because that's day one, page builder stuff. You don't. it just blows my mind. Yeah, that 10 and what, frustrates me more is how much people praise that.
They'll be like, oh, Gutenberg is a, is such a great builder. You can add border radius to buttons. Now I'm like, it's been 10 years. that's, day one stuff. not 10 years stuff. So that stuff frustrates me and I talk about it. And therefore I'm a WordPress hater, despite the fact that I've been using WordPress for over a decade.
If a client called me right now and asked for a website, I would build it in WordPress. I'm just not part of the, the toxic positivity, WordPress cult, where everything is great, everything the team does is perfect. And there's nothing wrong with WordPress. I'm not, that's not me.
I still use WordPress. I don't know how long I'll be able to say that. I, think I'm probably done making WordPress plugins. I don't see a point to doing that anymore. it's just especially trying to get one into the, the repo right now because everybody's building plugins with ai and the submissions are just, the poor WordPress team who has to review all those
[00:32:55] Nathan Wrigley: Yeah.
[00:32:55] James Welbes: now.
[00:32:56] Nathan Wrigley: Yeah.
[00:32:57] James Welbes: And, most of 'em are gonna have a bunch of issues that they're gonna have. So there's gonna be a huge back and forth. I've got two plugins into the repo before AI really blew up. And I was using ai. I'm not a developer, so I managed to get two WordPress plugins into the repo. And it took a little bit of back and forth with, with the team as the first one, I think took three back and forths with the review team, where they're like, okay, now fix this, now fix that.
And then it got in. a lot of this stuff is gonna be people who don't know anything and they're, they got a plugin to, to work in their WordPress website, so they go try to submit it and the review team's gonna look at it and be like, this is absolute, it's just gonna be so many back and forths.
I, I feel bad for any legit developer who has a good plugin right now who wants to get it into the repo 'cause they're gonna have to wait. It was already a long waiting period. And I imagine. It's, it could be months right now with the number of people submitting plugins. So I just don't feel like WordPress is worth, investing in at the moment.
I still would use it to build a website. It's obviously not any less useful than it was yesterday because of ai. I feel like the future is, there's less reason to want to invest in WordPress because of ai, but as of right now, it's obviously still a very. Very capable, very useful tool, and I'm familiar with it.
And, I have, I just purchased an etch licence about a year ago, which I really like. And so yeah, I would still use WordPress to build a website. Your question about what am I doing with WordPress and ai, I, nothing. I tried telex and it was actually pretty neat. I needed. I wanted to create a button where when you clicked on it, it copied a bunch of JSON to your clipboard.
And I was struggling for some reason on my own to, to get that to work with an AI bot. So I went to Telex and I, told it what I wanted and it worked. I got a really nice block out of it that. You could paste a bunch of Js code, JSON code into the block and it puts a button on the front end and you click on it, it copies, it just, it works perfectly.
so that was nice telex. if you're trying to make yourself a nice custom little block for WordPress, I highly recommend trying telex out. That's my only experience with it. I'm one for one Bating a thousand with Telex, so I had nothing bad to say about it. Miles, I signed up for the beta programme and I got the email and I clicked on it and I got like a 4 0 4 error.
So I think I. Waited too long to click on it or something. So I didn't actually, haven't actually got to try Miles yet, but I'm also to the point where if I'm gonna use AI to build a website, why would I use WordPress? You know what I mean? it's,
[00:35:37] Nathan Wrigley: This, I think is the curious moment that a lot of people are having, this sort of epiphany that, I'm, where's the, why do I need that scaffolding? Why do I need that thing? and I think people go two ways there, don't they? It's either, I'm gonna use. WordPress because I know all the bits and pieces of my clients get how it works, and they're using, the post editor or whatever it may be.
But then I'm hearing an increasingly large amount of people just like you who are thinking, it, the sand has shifted for me. I'm, now able to achieve so much that I don't need that scaffolding anymore. I can achieve all of the things that my website. Needs by just chatting to an AI or, whatever code editor you've got, maybe there's some modifications to be made.
And I th I think we're at a real kind of inflexion point where some people are just like, you gonna go in that direction and it will serve them very nicely. Thank you. But it'll be curious to see how that works out. whether the AI can keep improving to the point where WordPress dwindles more and more, or whether people, the stuff that's being built on top of WordPress, like all these AI initiatives with the AI team, whether that convinces people to stick around.
I'm obviously you have no crystal ball. I can't tell you what the answer is, but I am hearing a lot of people in exactly the same place that you are. the, the ground has moved enough that they're I'm just using AI now. That's just where, what I'm happy with. It's a bit, I suppose an analogy would be, I don't have a landline telephone anymore because the mobile phone's got good enough that what's the point in having a landline telephone?
I don't need one. I don't need to replicate that. It's a little bit like that. The AI's getting good enough now. I don't need the, WordPress side of things. It'll be very interesting to see how that goes. Sorry, I interrupted. Maybe you were in them.
[00:37:34] James Welbes: Like right now you're, we're already seeing a lot of people moving simple websites off of WordPress and people that you might not expect. a few weeks ago there was the, director of marketing at automatic. And he's mentioned he went to LinkedIn and told everybody he moved his website from, WordPress to like some markdown file solution.
I don't, I didn't quite understand what he did, but it wasn't WordPress anymore. And then three days after that, he announced that he's leaving automatic, not saying that there was like a correlation there, but it, he worked for automatic. And while he was working for Automatic, he moved his website, his personal website, away from WordPress, juiced the founder, creator of Yost.
he just announced not too long ago that he moved his website from WordPress to an Astro build and there's, I'm seeing a lot of other people, not quite as well known, maybe as those two. But my Twitter feed and LinkedIn, I'm seeing a lot of people talking about, oh, I'm, I just moved my personal website away from WordPress to this solution on that solution.
And a lot of it's astro, but not all of it. And so it's already happening with simple websites because like you said, you know it, WordPress is a very, complicated, I'm going to even say maybe bloated, like it's just, it's been around for a long time. It does. A lot people would argue that it does everything.
It's the platform you use because they use it because it's so flexible. It can literally do anything and it's cool. But if I just need a blog. A landing page, do I need WordPress for that? Do I need the platform that can literally do everything? Or do I just need a platform that can create a simple content type and display that content?
and I think what AI has done is It, given us permission almost. I think there's a lot of people who've already been wondering this. They've been like, oh, do I need to use WordPress? And it's it's what I'm familiar with. I know WordPress, I might as well just do it even if I don't really need all that.
it's what I'm familiar with, so that's what I'll do. And now it's you can literally just tell your AI bot, Hey, make a, website. With Astro and I want it to be a blog and I want it, this is what I want the layout to look like. Whatever. And it just does it for you. Now are you gonna be building, let's say you have a, you're selling courses or something that's a little more complicated.
That's, where WordPress still has a bit of an advantage. 'cause you can go for $200 a year, you can go buy LearnDash and it's a very capable, learning management system and you can sell, create, and sell courses with it. Do replicating that with AI would be a little more tricky. right now I've done it, but it's, it's so much easier to just instal.
Learn dash, so it complex websites and I don't see a lot of people building client websites yet with AI and Astro and experimenting all this stuff. I see a lot of people experimenting right now with their own personal websites like myself. And then when it comes to client work, they're still, they're still using WordPress.
That will change, I think it's a slow process, but we're already in phase one where a lot of people are moving their own crap away from WordPress and onto other platforms. phase two people are gonna start. oh, I got this client and their website's pretty simple. I'm gonna, I'm gonna try moving it away from WordPress.
That's phase two. And then phase three is AI is so good. You just flat out don't need WordPress for
[00:40:58] Nathan Wrigley: so that's the curious bit, isn't it? Is that process? Of how good it gets. And so I guess people moved on to using CMSs because it just made it easier. It made it easier, that process and then page builders come along and that, that made it even easier. now you just point a mouse and drag things in and start typing and what have you, and.
click, a thing and change a colour and what have you, nice and straightforward, and you go there because the solution is easier than the legwork of trying to write out one line at a time. The HDML, the CSS and all the JavaScript, you get that solution. It's affordable, it's quick, and it's easy.
people always coalesce for the easiest thing, don't they? if something is simple, easy to understand and it will deploy effectively. They're gonna use it. And the rate of change of a AI that nobody's gonna bet, assuming an abundance future where, we don't run out of money and the AI companies keep going and blah, blah, blah.
But there, there's definitely a future in which almost everything can be achieved by talking to an AI and it does become very hard to calculate. Where almost any software, not just WordPress, almost any software, like why would I use, oh, I dunno. Like I use a note taking app. Why would I use a note taking app if I could build something in a few sentences, okay, I want something a bit like notion, but I want it to have, this colour and that colour, put it on this domain, and so on and so forth.
It feels to me as if we're entering a domain in which almost everybody who's mildly technical and above. Built is having built for them by, ai, their own suite of things that they want to use personally. And it'll be fairly disposable. I, wanna use this for a while. I'm planning a holiday. Can, I throw together an app that's gonna help me build my holiday, my ideal holiday, and then the holiday's over, right?
Kill that now. Don't need that anymore. That kind of thing. And it's hard to bet against that. It seems like that's coming.
[00:43:03] James Welbes: I don't know if you brought that up intentionally or not, but, I literally built a note taking app because I was using Notion and didn't like
[00:43:11] Nathan Wrigley: Oh, no, That was total coincidence. Excellent,
[00:43:15] James Welbes: That's funny. Yeah, I, was using Notion because it's, it has, A CLI or API or whatever, so that, so an AI bot can use it. So I could tell Carl, go scan my website, tell me if you find any security issues, and then write up a report and save it to my notion.
And it would do that. And it was pretty slick. But then I hit a limit with the free plan of notion. I also hated the ui. I don't know what it is. I know it's very popular, so I think I'm in the minority here. But something about the way you have to organise. Pages and, whatever. I didn't like it at all.
And so when I hit that limit, I was like, I definitely don't like this enough to pay for it. Like I was fine using it for free, but if they're not gonna let me use it for free anymore. So I, yeah. I went to Carl and I was like, I want a note taking app. My prompt is pretty simple. I didn't give it a lot of direction at first.
I said I want a simple note taking app. it's, I want it to, replace notion. I'm not trying to make it do all the things Notion does. I know Notion does a lot, but I want a simple note taking app that AI can. also use, and he did it. And it's a great, it's a great app. I love it. And so that's one, I'm just one person, but notion, potentially lost a customer.
'cause if I couldn't create my own app, I might've had to just bite the bullet and sign up for whatever notion's cheapest plan was so I could continue using it. But I don't have to now. And I know there's gotta be other people out there either with Notion or just other apps who are. Going, maybe they're not creating a SaaS and selling it and making a lot of money, but they prevented themselves from having to pay for something because they just built their own little version of it for themselves.
[00:44:53] Nathan Wrigley: Yeah. I think this is, I think this is inevitably some part of the future where people just.
Build things, it's disposable. You try it, you then migrate it a few weeks later because you're not happy with the way it looks or the way it behaves, or you now want it to interact with this, that, and the other thing that'll be really curious to see, because obviously when the internet came around, there were no SaaS apps.
That whole thing began and now has become the kind of modus operandi. if you've got a business in the, in the tech space, you probably want a subscription business with. Some sort of SaaS behind it. That's become the way that these giant companies, Google and Notion and Shopify and all these things have been built, and the, erosion of that.
Slowly but surely, it'll be interesting to see. How that is sustainable. does all the money coalesce towards a very few really powerful AI companies, the likes of Open AI and Anthropic and Google and whatever else, because they are the custodians of the code, which runs the ai. Meaning that we pay them for the tokens to build the things the hosting companies obviously benefit as well because we have gotta put this stuff somewhere.
It'll be interesting to see if there is like a slow. Degeneration of that landscape of sas because everybody's just rolling their own. I, have no idea what's gonna happen around there, but I can see a world in which what you've just described is totally normal and not just for nerds, just most people can do it.
I don't know. I'm, it's Saturday morning, I want to go out and buy some new shoes. I need an app that's just gonna figure out where the heck I can go in the next three hours to get the most affordable pair of shoes, which are exactly like the ones I want. And then that app's gone. I don't need that.
Three hours later it's gone. Don't need the shoe app anymore. What a world we live in.
[00:46:48] James Welbes: There's, yeah, there. It's, the SaaS market is about to be attacked from two different angles. First of all, the competition is about to skyrocket because everybody's building a SAS and trying to become a millionaire, building a sas. And most of those are gonna be crap, but some of them are gonna be good.
And so the competition is just gonna go through the roof and also. What I mentioned already is there's, you're gonna have people who don't even want to buy a SaaS because they can just build their own. So they got, if you're trying to be, if you're trying to sell a SaaS, you have to convince people to not build their own and to pay for mine instead.
And then you also have to compete with all the other people creating SAEs. I mentioned that
[00:47:27] Nathan Wrigley: Yeah. Yeah. The
[00:47:28] James Welbes: the league, Yeah, so I reached out to the lead director a few days ago and I, 'cause I got the app to the point where I'm like, ready to reveal it to him and see if he wants to try it. I said, Hey, are you still, tired of team Snapp and looking at trying something new?
he's vi he vibe coded his own already. So, like I spent, all this time making this thing thinking he, there's a good chance he might be my first customer. Yeah. And then, I find out that he vibe coded his own. um,
that's, that's just the way it
[00:47:57] Nathan Wrigley: tidal wave is coming because
need to be that technical that will be such an interesting
[00:48:05] James Welbes: You just need one person. there's the whole board of people who manage the league. If one of those per people in that board is semi technically inclined, they just save themselves thousands of dollars. They're paying. I think he said they paid close to a thousand dollars per season and they do winter, falls for me.
So probably four grand a year. They were paying for this app that they just saved because the league director happened to be savvy enough to vibe code an app for
[00:48:32] Nathan Wrigley: It'll be really interesting to see a, whether that's sustainable, from the AI side, whether it'll collapse under its own, venture capital, borrowing and things like that. And also whether it's secure. Whether those kind of things. Obviously if you're doing it on your own computer, all bets are off.
It doesn't really matter. My shoe app for Saturday morning, which is just helping me to buy shoes, I'm gonna kill that. It's only on my local machine. Doesn't matter. It's fine. But the bigger, the wider sort of stuff where you know, you're getting people logging in from on who knows where and inputting their data about.
Football and things like that, that, there's something there. there definitely could be problems with that kind of thing, but my expectation is that the capacity for the AI to do things better and better is only gonna get better and better. And it doesn't move slowly. It's not like this juggernaut of yeah, it's gonna creep up.
So it does really seem to be going at an amazing clip. And you and I having this conversation in 12 months time. I have no doubt that it will be utterly different. The landscape, really different things will have moved on so much that we'd have a completely different conversation.
Wow. Gosh, that was fascinating.
[00:49:41] James Welbes: Yeah.
[00:49:42] Nathan Wrigley: I'm gonna, I'm gonna put a pin in it there, if that's all right with you. James, because we've reached the 50 minute mark and, I think that's probably about the sweet spot for a podcast like this. so if anybody wants to get in touch with you and, to bang you over the head with, I am the Luddite.
I don't want to hear about this. where can we find you? It sounds like you go out on social media a bit. Where are you hanging out these days?
[00:50:06] James Welbes: Yeah, you, I'm most active on Twitter, where apparently I am just hating on WordPress
[00:50:12] Nathan Wrigley: Oh
yeah.
[00:50:13] James Welbes: you can find me on Twitter. It's just at James WebU. And then, I'm also on LinkedIn, but, I mostly just use LinkedIn to complain about how I hate LinkedIn and, to share YouTube videos.
[00:50:25] Nathan Wrigley: That's what it was made for. That's the ideal.
[00:50:28] James Welbes: it's,
[00:50:29] Nathan Wrigley: That's a great idea. I use LinkedIn to complain about LinkedIn. okay. In which case I will bury the links in the show notes. So if you go to wp builds.com, if you search for the episode with James, and his surname is spelled W-E-L-B-E-S, I will link to social channels and things like that as well.
James, what a fascinating conversation that was. I don't figure out that we've, I don't figure that we've solved anything there, but it was certainly, illustrative and useful having a chat, trying to figure it out. Thanks very much for your time.
Okay, that's all we've got time for today. I hope you enjoyed that. Head to wpbuilds.com, search for episode number 466, and leave us a comment there. We would really like that.
Couple of URLs to mention, wpbuilds.com/subscribe to keep updated. And wpbuilds.com/advertise if you would like to insert your sponsored messages into this podcast.
Okay, that really is it. I am gonna fade in some cheesy music and say stay safe. You have a good week. See you soon. Bye-bye for now.
