I would normally, typically random, ramble much more than that, but should we just, shall I just launch into what we're going to do
Let's do it. We can, we can launch together where it's the multiverse. The timelines are both crossing.
Okay. In which case, I'm going to do it like this. Okay, Matt, who's going first, me or you?
you know what you can go
Okay, I'll go first. The reason I'm going to have to ask that question will become obvious in about three seconds time. Matt and I are doing a podcast together, but it's going to end up in two different podcast clients because it's going to go into his podcast and my podcast, so that's kind of weird. Fairly new.
Yes, this is new. And, let me just ask you before we dive into the WordPress and AI stuff and the life changing technology. let's just talk about this old dog of podcasting. You've been doing video for as probably as long as I have, and audio. Have you been following the recent trend of the industry at large saying, Podcasting is now video and not audio. And what's your gut reaction?
My gut reaction is to firstly make a massive confession. And that is that the podcast platform that we're using to record this is currently not recording the video. Have I made a giant error?
but you do live streaming, right? So do you, you consider the live stream, the podcast, a podcast as
Yeah, I do. And my, my thing is all audio, I'm all in with audio and I'll tell you why. It's a personal preference and it's based upon my childhood really. When I was a, when I was a child, we had, we had audio on in the kitchen all the time.
And it occurred to me at an early age, I wasn't told this, but it occurred to me that I could do things at the same time as listening to the audio do things, which I. Wanted to get done, you know, and it might be chores around the house or DIY or whatever it may be. And I've always preferred the radio to the TV. I watch very little TV. But I consume a massive amount of, well, I don't consume much radio anymore. It's all podcasts. So for me, audio all the way, very conscious decision.
And the only reason that I record my show on a Monday as a video is because it's the only real way to do that. but it gets recycled as a, as a sort of audio podcast episode. And that's where, that's where my audience is. It's not really on YouTube or anything. What about you?
tons of podcasts, audio podcasts. And, and I do jump between. YouTube and, you know, my favorite podcast app, of the week. And, it's, it's, it's great, but I will always defend audio. Because, listen, when it comes to a, just production level, right, as you know, it's just so much easier to, do audio than it is video. Every second counts, excuse me, the train is moving above my head right now.
I wondered what that
counts, every second counts, in, in editing and production and setting up. So I am, all in with audio. I recently launched another project called Our Beloved Medium, which is an audio documentary slash, fictional Entertainment series. And if we did video with that, it would have been a motion picture. It would have, we would have to hire actors, actors. I mean, it was impossible to produce what we produce with audio, which is story, voice, acting, sound effects, all this stuff.
If we did that in video, we'd have to go to Hollywood and ain't got the budget for that.
did that go? Because when we sat down, because Matt and I sat down at WordCamp US in Portland, and you were talking about that becoming a thing. I think you'd even cast like voice actors and things like that at the time. So that's now shipped, is it? And you've done it.
yeah, we, we have the first episode out and this is the episode that we're sort of using as the, the carrot on the stick to get, some sponsorship dollars out there. RSS. com, podcast hosting company is the, is the first one to, to help us out with a, a nice chunk of, of cash. Helped us get our first, you know, pay for our first voice actor and get this off the ground. I have a small part in it. you know, it's the, it's the last radio station on Earth. It's, there's this company called Opus 3000.
It's taking over all of the, radio stations. And, we're the last radio station on Earth broadcasting stories, of radio importance throughout humanity. And I'm the general manager of WFYZ. Yes, it's a, it's a, it's a bit of a drama. It's a true documentary as we interview folks who have covered the space over the last, you know, 20, 30 years that we can reach out to. And it's
Do you, like, okay, sorry audience, we're just going to totally digress here, but this is fun. would you, would you drop the, the bits and pieces that you've done for the last sort of 20 years? Like the, you know, the tech based podcasting and all of that. Is, have you enjoyed it enough to think, gosh, if I could pivot to that, that would be where I'd end up?
because I, you know, I think it's still the act, you know, of creating art. It's an art form. for me, right? Like, the WP Minute is an art form for me. it's a way of storytelling, through, you know, highlighting stuff, that folks have done in the WordPress space. And I see it, you know, very similar. if all of a sudden this took off, and somebody was like, Hey, you have a book deal or a movie deal out of this, out of this, audio story that you put together?
I mean, yes, I could see myself being like, Hey, you know what, WordPress? That was fun, but now I'm gonna
Yeah, because it, although I know it's like striking a vein of gold, if you're a, you know, a gold miner, the few people that do do that, it's retireable, it's retireable amounts of money, isn't it? You know, you can have one hit and it'd be the, the end of your working life if you choose to go that way. Whereas the enterprise that we're involved with, tech news, particularly WordPress tech news, there's a, there's a small audience for that. It's a much more of a slog and you'll never, ever.
Hit that, gold sort of thread of gold that enabled you to retire. Gosh, that's really interesting. I pray for you that that works out. That'd be great.
it's not the intent. it's not the intent by any stretch of the imagination, right? The, yeah, you know, the particular challenge for me with this project, Our Beloved Medium, is to work, with a, a former colleague of mine, Stuart Barefoot. He's the, he is the, the creator behind this, and he's the editor. I'm the executive producer. The challenge for me was I want to wear the executive producer title.
Like I want to, I want to learn that space, which is like largely what I do already for the WP minute, which is awareness, finance, like getting sponsorship dollars and, you know, having some creative angle on, on the stuff that we're going after.
But the WordPress stuff, you know, I think maybe another phrase for it is like, this stuff is a labor of love, the stuff that you do, the stuff that I do, this is a very small, you know, this is a very small me air quotes media opportunity for folks like you and I, and I'm sure we, we can get into it, but, you know, we've seen other folks come into this space and come and go because they're approaching it from, Like the typical, like, I want to be a YouTuber or I want to be this content creator
that sells courses and do other stuff. It's that's fine. It's a fine goal. but the attention in the audience for WordPress feels big, but it's very, very tiny. I've always said 10, 000 English speaking humans in the world, maximum care about the stuff that you and I. Well, at least the stuff that I talk about. I don't want to, I don't want to cast the 10,
No, I think, I think it's a smaller number for me.
yeah, yeah, so like that, like, that's the way that, that I see it. so you really, like the stuff that, that you and I do tirelessly doing this and folks like Ray, folks at WP, Mayor, you know, these, these folks who haven't given up on this space, you know, we're doing it because we love WordPress and, and that's the, that's the motivation.
fascinating though. Honestly, seriously, I really wish you the best of luck with that. When, when you've got it all, you know, when we finish recording this, can you send me the links for that? I want to listen.
we'll do.
Okay. So. It's Matt Madeiros, it's me. We're going to be talking about, AI because everybody's talking about AI, but particularly in the WordPress space. So I'm going to, I'm going to lay my cards on the table, Matt. I'm a bit of a Luddite when it comes to AI, I haven't really explored it very much personally. So more or less everything that I have to say comes third hand, but the tiny forays that I've made into it have really frustrated me. And I found more or less nothing works.
So I was at a meetup in London the other day, and I, I was sitting at the front and I said to the audience, I tried four things the other day, on AI, how many of them do you think worked? And the response came back, all four. And I was like, no, it was none, none of the things that I tried worked, absolutely nothing that I tried worked. And so that tells me, A, that I don't ask it the right things, because I'm sure that the basic stuff that I was asking is in the realms of AI.
but B, I haven't really played with it enough to, to learn that I was making those mistakes. You, on the other hand, you've like taken all your You've rolled up your trousers, you've jumped in the lake, and it seems that you're like a real adept now. You're like pro, pro, super pro AI person.
Well, Nathan, what you've forgotten is I sold cars for a lot of my life and I'm pretty good at selling the story.
Okay, so many, many, many, many hours of banging your head against the wall, eh?
I'm just a, what I would say just a few paces ahead of you, you know, on, on this, on this learning curve. but yes, I, I recently, put out a, an app called an app. Let's call it what it is. It's an RSS reader slash aggregator. You can find it at pulsewp. cc. And, it was just a way for me to exercise the idea of building something with AI. And, largely solving a problem that I have every day, which is, How do I keep up with all the WordPress news? you know, but my way, right?
You know, I have, I'm gonna forget the name of the app that I have on my phone. net something that I use as an RSS reader. And it scoops up everything and I have my newsletters and I have obviously my slack channels. and That keeps a, that keeps a pulse on things for me and I was like, there's some things I want to do. Like I want to be able to summarize these things quickly.
I want to have like little social media messages automatically made for me for every new article so I can share these out. And then I started building it. And we can break down how I did that, but that's how it was born. It took Forty ish hours up until this point to do what I did. lots of failing. lots of experimenting.
the 40 hours include the failing?
Oh, yeah.
Oh, that's not as bad. I would have thought it would have been a lot more than that. But have it, but I know that you just said a minute ago that you're a good storyteller and sold cars and you can put a spin on all of that. It does seem though, from everything that I've seen and the tenor of your language, it does seem that this has captured you. It's not like, you know, you wouldn't have got to hour 40 if it was pure frustration.
I'm presuming that, you know, at hour 2 and hour 6 and hour 10 or whatever, there were moments of, whoa, look what I just did!
Yeah. so here's the biggest thing. For me and, and I hope, and I'm sure you can relate to this too with your stint with Drupal, way back in the day, CCK and views and all this stuff. those moments were very powerful for me, right? When I installed my first PHP Nuke bulletin board, I was like, Whoa, this is awesome, right? Like I put a lamp, like before I even knew the terminology, I put a lamp stack together. I installed PHP bulletin board. I access it through a browser on another computer.
Like, this is amazing, right? Poking holes in the firewall and accessing it, from your, you know, home server. Like, this was amazing stuff. And then you learn Drupal and, like, the web and the technology and CCK and Vues makes you feel like a PHP developer and you're like, wow, this is awesome. Fast forward, WordPress. You have those same shining moments, 2008, 2009, where you're like, I'm solving problems for clients, I'm charging money, I'm building a business.
This AI stuff, it's like that same moment in time where it's like, Wow, I just built this app and it becomes this thing that you just constantly want to do. I will say, a little bit of, like, I'm not a gambler, but it feels like a little bit of gambling. Cause every time you send it the prompt, you're like, is it gonna get it? Is it
yeah, yeah,
get it? And then it gets it, and you're like, yes!
or at least it looks like it got it,
Right, it looks like I got it. So it feels like it's, it's fun right now. and you know, that's, that's what I'm, that's why I'm liking it. But, let me just ask you this question. With your failed attempts were you doing those in like chat GPT or like a vanilla sort? I'll call a vanilla AI prompt
I was, I was doing it in chat GPT. So a client on the Mac for chat GPT and I was asking it discreet questions, but my discreet questions were so confined and constrained that I presumed that it was to do with getting a loop and connecting two different custom post types and, you know, comparing one against. The other, it was the kind of thing which any decent developer I think would knock off with their eyes closed. And so I thought that this would be fairly possible. No, it really wasn't.
It took many, many, many attempts. I've subsequently learned from somebody in the WordPress space, Tim Nash, actually, that if you ask it and you literally append at the end, can you make the output, use WordPress coding standards that immediately will make it more or less better. and I didn't know that. But so everything that I got was wrong, but I wasn't using any kind of, environment or setup. I was literally asking the pure, chat GPT. And so maybe there's a mistake that I made there.
Yeah, were you were you was your intent to build it as like a proper plug
Yes. But first of all, I just wanted to, to see what the output would be if the simple things, and I've tried it at various different points and it has succeeded. But in those cases, it was something so very, very minor. but yeah, I haven't set up the, the development environment and everybody, everybody at the minute seems to be talking about this thing called cursor.
I don't really even know what that means, but, the principle being that there's, there's an environment, there's a setup, there's tools that you can get that are better and superior and advancing at a faster pace. I wasn't using any of that. Can I just ask you a question? This is a weird question, but I'm going to ask it anyway. Do you feel in relationship with this thing?
Do you feel like when you sit down and ask it these questions, do you feel disappointment in it when it doesn't give you something back? And do you feel like happiness for it and for you? And, you know, is there some sort of, and I, that's weird. I know I don't mean relationship, but do you, are you drawn to it? When you sit down and you sort of like, right, some more AI smashing, let's crack on. This is exciting.
is, like back to that gambling thing, or like, when I was younger, like video games, like You know, when I played, I don't know, whatever, your favorite video game, you know, back in the day, where you were like, you have to clear the map, you have to hit all the points, and like, get the trophy or whatever, that's what it feels like.
I don't feel, I think what maybe you're saying is like, do I feel like I'm, cause I've heard this before, where people sit down and it's like, oh, I talk to ChatGBD all day, it's like my, it's like my buddy, I don't feel that, but it's, it's wanting that output, to make me feel like, alright, this is gonna get done. And I'll tell you why, because as somebody who, so the app that I built, PulseWP, it is React, right?
This is not WordPress, this is not PHP, that's not, you know, it's largely React and JavaScript, which I know. 0 percent
Yeah. Yeah. We'll join the club.
I mean, I know a little bit more now, which, you know, is great. This is a great side effect of playing with this stuff is you do get to learn APIs and routes and how react is structured and pulling in components and libraries. And you really start to learn like the fundamentals, just like when you were hacking away at a WordPress site 15 years ago, you were learning what this stuff does. Now I'm doing the same thing at a faster pace. But because I know nothing, Okay.
I rely 100 percent on the agent to build the thing, right? So what that means is and why this feels kind of like a video game slash like a slot machine that you pull the handle down on is because it take it actually takes time. So if you're asking it to build something complex, you ask it and then you wait and you stare at the screen and you go and I'm just waiting for it to finish and I've found myself.
Now, when I was really like deep into building that app, launching a prompt over here and one tab, sending another prompt in another tab, asking it to like, think ahead about the next feature that I was working on. So it's this thing where you're just waiting for the output. It's time consuming and it feels like. I guess I'll call it work because you're just waiting for this thing to do the task and you have no say.
And if it fails, which it does, very often you have to then go back and say, do it again. But now think about it this way. And then wait for that output to to, to finish. And you have to wait to see if you're getting the goal that you're after And it does feel like little levels in a video game, like, can I beat this level today? And like, that's the connection that I have with it, because, again, it's making me feel powerful.
People are seeing these features, I'm shipping, air quotes, features, that makes me feel like a proper developer.
So I was going to ask that, that was going to be another question and it was, and I think you've just answered it because when you said that you did air quotes, your fingers went up and you did air quotes, who, who built this? Basically is my question. Did you build it or did something else build it?
Something else built it.
And, and, and
had the idea.
how like the, the people like me with no coding capabilities whatsoever, I would feel that compulsion as well, I would fall over myself to, to kind of explain, look, I didn't do this, it was chat GPT, or whatever it may be the latest thing, but it, but it's curious how, well, you, you did do it. But you did it with the assistance of some other thing. But it definitely couldn't have been done without you. Or I suppose without So it's like a team effort almost. It's kind of weird.
Yeah, yeah. And you know, I would fast track the, like to fast track the concept of. You know, if you, if you open up your favorite social media platform, if you have one and you look at, you know, the, the discourse around AI right now, it's either this thing is replacing all of us, or this thing is replacing none of us. you know, like everything else in my life anyway, I kind of see it somewhere in the middle and. You know,
us. Ha ha ha ha
of us. See you later. One of, one out of two of us on this
Either that or half, half literally of us. One half of me and one half of you. Ha ha ha ha ha!
and, you know, these, these tools are great. They do fast track it, but you, I, I think what it'll, it'll do is eventually, you know, will we need an app store? Right. You know, when you look at it from, I guess maybe either a dystopian point of view is like, will you need an app store if you can just go on your iPhone, which is now somehow powered by Apple intelligence? Cause the, the, the new colorful rainbow ring around my phone turns on instead of like the little Siri circle anymore.
And will I just turn to my iPhone and say, give me an RSS app that follows just the WordPress news and it makes. Basically, what I made with Pulse and just installs it on my phone. And I pay Apple, like, nine bucks a month for Apple Intelligence to maintain that app. What does that do to the landscape? But then also, like, as a developer, What does it do for humanity, right? You know, like, do the developers go away and the AI?
So, there's all of these thoughts that are going through my mind as I'm building this stuff. I think it'll get better, but I don't think it's replacing all of
I think it's interesting because I think there's always, I really don't know. I can't distill it, but I have an equal amount of trepidation as I do excitement. And so the rainbows are very, very close to the dark black clouds. and the bit of me, which worries is that the industry that we're in is full of very, very nice people who have spent an age getting good at, let's say, plug in development or what have you. And it does seem to me that at the rate that.
The development of the AI is moving that a lot of that might be possible with a voice command. Like, obviously if you build something complex, like a complex, I don't know, an LMS or something, I'm sure that's probably out of reach at the moment. You'd have to have all the different parts, but if you've got a plugin, which is very successful in just doing one or two simple things, that concerns me a little bit. So that's a bit of a dark cloud. However.
The, the sort of rainbow side of that is, well, we're here to democratize publishing and if anybody can speak into their phone and achieve the thing that they've asked, that's fabulous. And, and the boundaries just get pushed further of what's possible, which is great. Here's the other weird thing, which never really gets into my head when I'm thinking about it is, I don't know if it's sustainable, in terms of, I don't know if the economic model is keeping these LLMs going.
is something that can be sustained. It seems like there are some dark clouds gathering there, where the amount of energy that's being used and the amount of dollars being spent just to is kind of being kept to get kept alive on a venture capital venture capital promise that something around the corner is going to flip the economic argument and all of a sudden we're going to start shelling out great amounts of cash for, for this.
I wrote about this, on, my resume site, Crafted by Matt, and it is, One of the things I think that's going to hold this whole industry back are these industry leaders themselves. Because how many times have we seen and heard this story before, right? You must jump on this bandwagon. This is going to change everything. This is the technology of the future. And everything looked like AI every time that thing was happening, whatever that was. NFTs, cryptocurrency, like all these different things.
The internet, books are going to go away. Self driving cars, you know, cell phone, like all of these technology points have always said that kind of thing and it's a little true but then it's like you scale back 60 percent of that and then you scale out the other 40 percent over two decades because that's how long it takes for people to catch up. Like people still don't have websites, right? Like we thought that was going to be done and all over with 20 years ago.
so I think the biggest thing holding these, these That will hold us back, you know, I'll say it is like the corporate greed, right? We need to be the biggest, the best. We need all of it and we need trillions of dollars to do it, right? Oh, how much are you making? Zero. In fact, we're losing trillions of dollars while we're asking you for more trillions of
and we're pumping a load of carbon into the atmosphere and, you know, all of that kind of stuff. The interesting thing here though, let, let's, let's just use, because I said it, self driving cars. The, what you get, if that technology comes to fruition, you get self driving cars, you get cars that can drive themselves. However, if the AI, if the AGI, the artificial general intelligence comes about, then you basically solve more or less everything.
I mean, you invent a load of other problems, but there isn't just like one solution. It's like you just give a computer or a device or something. can you invent me a new game that nobody's ever played before that'd be really interesting for like under five year olds? yeah, sure. Try this one. Oh, yeah, that's really good. or do you fancy just getting rid of cancer for me? Yeah, all right, that's fine.
You know, so the promise, the promise here that they, again, the rainbows and the unicorns and all that is so vast that I think this one maybe differs slightly and it wouldn't surprise me if the spending cycle keeps going for a little while longer, but I can't see it going on forever. That promise of AGI seems to be no closer today. Then it was two years ago, but the output of the LLMs is to my mind. Anyway, it makes you think that there's intelligence there, but it's not really.
There is, I don't know if they're testing fire alarms in my building right now.
Yeah, there was something, I can hear something, but it's okay, keep going, I
are going wrong. the, You're wearing a pair of glasses. I'm going to take you down one of my, crazy conspiracy theories, with AI. You're wearing a pair of glasses, and I wrote about this. I said, well, look, at some point, the hardware, the technology, the software is going to get so optimized that you'll have an LLM on your pair of glasses.
You'll walk into a car dealership one day, if you're still buying cars, and the salesperson is going to look at you and go, The car is 35, 000 and that's my best price. Your glasses will analyze that and prompt you in your lens to say, That's not the best price. And here's how you should reply, right? And it, what it will move to is like this zero sum game for a world where, you know, lack of knowledge, lack of education, lack of information.
Allowed people to capitalize off of swaths of, of humans, right? I'm not saying it's a bad thing, but what it is, is it's gonna empower people, I hope. To look at that sales deal, buying a house, looking at a legal contract, talking to your doctor, and feeding you instant information that betters your situation.
So that, that competing negotiation or that lawyer who's like, sign the contract, it's the best thing for you, and you look at the contract and your glasses say, this is not the best thing for me, it looks like I get screwed in year two! Like, how does that happen? You know, and it's going to allow people, hopefully, to have that slight advantage that they didn't have before.
And, you know, the dystopian side, is going to be, well, if you don't have the best and the fastest LLM in your glasses, you lose. And by the way, the best and the fastest LLM in your glasses, it costs 5, 000 a month.
I think the dystopian runs really parallel to the, like the, the rainbows and unicorns. I think the two things, it's so difficult for me to differentiate the two at the moment. A lot of the positive seems like the negative and vice versa. One thing that is interesting though is that I came to the same intuition, and I don't know if you meant glasses as a As that's where you think the technology will land. But that is where I'm coming down now.
I really do think we, it will be in glasses and I know we had that Google glass debacle a few years ago where the, like, the promise was so great and the delivery was so small. I think that was held back by the, the kind of technology available at the time. But it does seem like the, the thing on your head. That nobody really cares about, you know, it doesn't have to have forward facing cameras, but it's got a screen and you can push, push things onto that screen.
I think augmented reality will become hot again once the glasses come, but I think the AI mixing in with the glasses is the perfect, the perfect vehicle. You know, imagine, imagine the things that you could do, if the, if instead of the computer and the screen, if it was just the pair of glasses that you take on and, and put off again. I think that is the, that
Yeah, we're, and we're not that far away, right? We all have phones in our pockets, right? So we can achieve that. let's not scare off the listener too much. Let's talk about the tools, right? How did I achieve some of this stuff? And let's talk about maybe some of the points where, where, you were having some challenges. What did your workflow look like? You were testing it out in ChatGPT and just like copy and pasting into like a snippet plugin tool
I mean, basically something like that or into a sort of functions. php file. And like I said, I was trying to do something really simple, but my goal was just to see if I could get things done in under like two minutes. In other words, how would a, how would an inexperienced WordPress user manage with this? And the reason I said it's in. The whole thing collapsed is because for an inexperienced WordPress user, it was an endgame. It was like, this should work, but no, it's destroyed my site.
And that's what it did. It took the site to the white screen of death kind of thing. And I got the email notification. So I thought, okay, that's interesting. If I was a novice. Coming into the WordPress space, it would, it would ruin my expectations of what WordPress could do. But, but let's flip that round again and ask you why, why this particular topic? Why PulseWP? I'll just give the URL again. PulseWP. cc. I mean, you, you said you were sort of scratching your own itch, but tell us why
Yeah, yeah, so I mean, I, yeah, I wanted a way to feed, yeah, that news and information for WordPress, to the way that I need to consume it, which is You know, cut, like, watch everything, summarize it, so that I can quickly scan it, and then go into the full article or, you know, whatever it might be that, that I'm watching for, and then quickly generate social media because I, social media, posts because I also share other people's content, so I wanted a tool that quickly allowed me to Watch
all the stuff. Summarize it so I can quickly jump into the bullet points that I need to know about. And then, help me share it on social media, because it's important to me. like you, I started, I had pneumonia. So, maybe not like you, but I had pneumonia. A few months after you and I met in Portland. And I was just laying on the couch, couldn't really do anything. And I was like, let me get into this AI coding thing. And I started to hit the limits of, Claude.
At the time, that's what I was using. And it would just tell me like, Hey, you're trying to build this piece of software or whatever. Come back tomorrow. literally it would say, come
Oh,
tomorrow because you've ran, you've run out of tokens. And I'm like, like, and I go back to like that gamification, that gambling thing. Like, wait a minute, I'm like halfway through this idea. I need to continue to build it. And that's what, and then I started, what I started doing is.
I would have half the idea with ChatGPT and use the credits over there, and then bring, and I would say, Hey, take this little idea of this little feature, write the code for it, but then give me a prompt that explains how to build that code, because I wanted Claude to build it, so then I would, I would do a lot of ideation on ChatGPT to save credits, and then make it write me a prompt. And then I would paste that into, into Claude and that helped optimize the stack a little bit. Right.
And eventually, and we can get into this, eventually I started to uncover the cursor, app, bolt. new, which is a fantastic, AI app. And then what I eventually built. PulseWPON is an app called Repl. it, which I like to say is like the Elementor of AI coding. Like, think about everything encompassed that you need. Because Pulse is, well, it's JavaScript. It's, you know, the React app framework. It's a database. It's a user authentication system. it is the hosting provider. it does a lot.
And once you start getting into this space, you realize, Oh, it's not just the website that you need or that front end component. You need the database, you need user authentication. User authentication is huge. and I should say that all of this built me a better, a better recognition for WordPress, right? Like, I look at WordPress now and you might go, Oh, that thing's like 20 years old, it's old and monolithic. Yeah, but it's done. Right? It's done.
I can pick this up and put it on a website and thousands of humans care for it. Versus me. On my couch and chat GPT and it's just me and him building this thing and nobody else is caring for it. Nobody else is giving me ideas. It's a stark difference from building code by yourself without any knowledge versus you look at WordPress now and you go, wow, 20 years and thousands of people have touched this. And this is why it's
can I ask you a question? So you mentioned a variety of tools there. So you chatGPT, Claude, you mentioned, Repl. it and you mentioned Bolt. What, what, so I think it's fair to
How did I make that path?
And also what is the path? What, why go to Bolt? What does Bolt do that's different from Repl. it and what does Repl. it do that's different from chatGPT? So yeah, talk us through the journey.
Yeah, I'm gonna, put a video together on this as well, comparing the, the two, platforms, but here's my, my take on it, right? I think you can, if you're somebody who's a developer and you're very, and you're, you even consider yourself, I'm halfway there. I think I'm a decent developer. Even if you consider yourself, I'm not the best, but I'm just a beginner. I think a great way to leverage AI now is to use something like cursor. Like if code looking at code doesn't scare you.
I would say use Cursor, because it's, it's a coding app, it's a fork of VS Code, which is a, an app to help you code. and it has all the things you need to accelerate your coding abilities. If you're somebody like me There's these two other apps, and I'm sure there's others. These are just the two that I've explored. Bolt and Repl. it. And you go to bolt. new is the website for the AI stuff. And I really haven't even had the time to explore these guys too deep. Guys and gals, I should say.
But they actually have an open source component of Bolt, which you can run locally. I haven't even touched that yet, but that's just like In my mind, like, this is kind of why I like these, this team, because they have that, and I do want to explore it more. But bold. new, you just go to the, you go to the site, you throw it the idea, and it builds it beautifully.
okay, so wait, wait, hang on, wait, wait, wait. You say you throw out the idea, like, how complex do you think you could actually make that idea?
And this is where the things kind of fall, fall down a little bit. So when I started, once I got off of, using vanilla, Claude to build my idea, I went to cursor because I heard that was a thing and what cursor gave me was vs code, which I've never touched in my life. No idea. I mean, the last time I use a code editor was notepad plus plus when I was running my wordpress agency, you know, going back to 2006.
That was the last time I've looked at code like this and it was very powerful, but it wasn't guiding me at all like it. You have to be a developer to use it. And if I was a developer, Cursor looks amazing. Because you already know the code and it's just going to accelerate that. And I was Fumbling around with it. I was, you know, I was at the time, when I first started, I was like, give myself a new challenge. Learn the CloudFlare infrastructure hosting, right?
Like I heard CloudFlare is better than AWS for some of this stuff. I used to sell managed WordPress sites for Pagely and we used AWS so I didn't know my way around AWS, but I was like, let me try this CloudFlare thing. So I was trying to like pipe all these things together, run CloudFlare. Pages, Cloudflare functions, Cloudflare databases, and I'm trying to tie all this stuff in together. Impossible.
Ha ha
And then I found Bolt by just hunting around on YouTube and Reddit and stuff like that. And Bolt will do really good at straightforward requests. Like you can ask it to make a user login system, but at the time when I started this, it would create it. But it would be just like a form, right? There was no security logic behind it. It would show you a login page. You could log in, but you could.
There was no like WordPress authentication, like it wasn't checking your password, was that strong enough, is there a reset, you know, I need to change my password, didn't have any of that function. what it did integrate with at the time was a hosting company, called Netlify. Those of you, those of you listening probably know about that. So what you could do, is ask Bolt, build my app. And then you could click deploy and send it to Netlify, and it would host the static pages that you built.
And it worked, but there was no database. Which then led me down the path of, I need to figure out how to connect up a database. And I found another platform called Supabase, S
Oh, that rings a bell. Yeah.
yeah. And it is a hosted database platform, And it does a ton of things besides just databases, does file storage, but it also does user authentication, because that's a huge thing, user authentication. Another thing we take advantage of, or take advantage for in WordPress is the whole, like, stack of user authentication is already there. So Supabase allows you to do that.
So I started stringing things together, and I'm debugging, like, build scripts and deployments, with Bolt and Netlify and now Superbase. And then I had to learn to tie in GitHub to all of this stuff so that I could version control it. It was very, very difficult to like keep it all going. And eventually you hit this limit of Bolt. Where it's just like, it's not working. And here's a pro tip, for any of these AI platforms, If it can't get it, in three or four tries, It's not gonna get
just give up.
Just give up, on like that feature set. So what that led me to do was, I would, I would build, I would ask all these complex things, Do this, do this, do this, do this, all in one big prompt. Build it for me, make it do this, and then move this button over there, and it would just crumble.
Right. Right.
So, I had to learn, and then I would delete the whole app, and start again. So then I would learn to take it slower. Here's the goal at the start, build it out, and then I knew the features in my head, but I wasn't telling Bolt to do it. And I would do it step by step. Now build this feature. Let's test that feature. Okay, great, that worked. Let's build this feature. Let's test it. Okay, do the work. All of the things you normally do. In like real development.
because a real developer isn't taking all those ideas and slamming it all out into his or her code editor right there. It's all done methodically. So I had to do that. And I'll fast track it here. Bolt now has a super base integration,
Oh, neat. Okay.
So now it'll create the database for you. You can host it on Netlify. You'll still have to connect it up to GitHub, to do like deployments and stuff like that. but it works. And then I found Repl. it. And you can think, like I said before, you can think of Repl. it as like Elementor. It's all in one. So it has all the database, the hosting, the user authentication. It's all built in.
And you're not having to hunt around and go, okay, now I need to set up Netlify to look over here at GitHub and Superbase. It has to know about Superbase. It just has all of its database and hosting all built into
So is, is replet, replet, is that your current Nirvana thing that that's where you're at at the
That's what PulseWP. That's what PulseWP is built on, but what I've found is what I will do, cause you're still faced with the same, no matter what, if you're, if you're running VanillaChatGPT, your own account, if you're running Bolt, or you're running Repl. it, you're still faced with the same, issue, which is, conversations get really complex, like, as you extend it and keep working it, and it just starts to fail and give you a bunch of slop, or, it gets costly, and you run out of tokens.
Right? So you, you start to, Replate has a different billing system, but we can talk about that later. But, you, it still costs you money. And you still run out of that. So if you're not, If you're, you're not, strategic about how you're building your app, then it can start to cost you money. And, or run out of credits altogether. but my favorite way of doing this is ideate in a vanilla, chat GPT, then I bring it to Bolt, and I say Hey, build me this idea.
And I just let it, what Bolt is really good at is that MVP. Like, what does it look like? Do I like the buttons over here? Do I like this feature function? And then if I'm real serious about it, I will ask Bolt to, once I get it to a kind of working level, I'll say, Hey Bolt, everything, you know, about this app. Can you make me a, prompt that I'm going to place into another AI system? Feels a little bit like cheating.
but, but it's like, hey, you did a really good job over here, but you know what? I'm not going to stick with you, but tell me everything you can do. And I'm going to bring it over into Repl. it and then I'll start to build in a
It's kind of like when you go into a store and look at an item and then in the store, buy it from Amazon on your
Correct. That's a hundred percent. That's a hundred percent. but for whatever reason, and this is stuff that I still need to explore. Bolt does a fantastic job with UI and the design and Repl. it is just Terrible. Like, like Pulse is not a very good looking website to look at. It does the job. It's very stark black and white, you know, and even when I say add more color or do more like Repl. It doesn't, it doesn't do any of that creative stuff, but Bolt does a really, really good job of it.
And their apps look a way, look way better. And for whatever reason, maybe their LLMs are just trained better. Like when you say improve design. Bolt does a really good job of improving design, whereas Repl. it just looks at you. Yeah, it's
in the LLM, if you're using the same language, but it, you know, your ex, well, or, or your expectations are sort of differently met, you know, you prefer the way that looks, but, but, but, okay. So given all of this in this journey and the fact that the three years ago, if I suggested. No, maybe three years. Yeah, three, let's go three years. If I said any of this to you, your, your three years ago, self would say my, my future self is lying. This cannot be done. That will never be possible.
And yet here we are, you, you've achieved something. You've built a, you've built an app that there's no way that you could have pulled off. And so the technology has gone from that's a lie that cannot be done to I've now done it. So that's three years. If we, like three years from now, surely that's going to be a bit exponential.
Let's assume that the, the venture capitalists don't run out of money and they're still very excited and all of that, then surely this is just going to get easier and easier. Your expectations will go up. Your demands of the prompts will go up. You, the output will get better and better. God, it's, it sounds like utopia. Coding utopia.
Three years ago, if you would have said that to me, I would have called the cops on you a year ago. I would have said, I don't think this. Yeah, not me. But now it's just like, okay, kind of me. And this is, you know, for all of the naysayers out there. Like, I think the reality is that the missing links right now are speed because this stuff takes time.
Like, it's not just time of like testing, but if you're somebody like me asking it to build robust apps, You literally have to wait for the code to finish, right? Like this is a waiting game. So I'm totally dead serious. Like when you start getting into something that's, you know, a little bit more robust, I'll, I'll send the prompt and then switch to tabs and like go back to work and let it finish.
And then if you have to, if it, if you go back to that tab and it's still not working, You gotta run it again. You know, so it's like, okay, let me go back to work. You know, did you finish over here? You know, but if you're somebody who's a developer, you know what you have to solve. I don't. So I just have to constantly keep prompting this thing. But yeah, three years from now what'll happen, and I think we're starting to see it now, are these agents.
Which are, for lack of a better word, these little robots that are supposed to be actively or passively working on this stuff for you. And that's the next sort of shoe to drop where I can build the app and then tell the agent, Hey, every day, just make sure I don't have any security flaws in pulse. Or if there's an update to react, or JavaScript, make sure you update it to the latest packages for me. And you know what? Every day, think of a new idea.
And send it to me, message me, and see if I like that idea. And if I like it, we'll add it to Pulse. Repl. it kind of does that. As you move along with Repl. it, it'll prompt you and say, Hey, did, do you like, okay, did that work? The thing that you just asked me to do, does it work? If so, I've got these three ideas that I think we can do next. And you're just looking at it going, oh my god. Like this thing is thinking ahead. you know,
yeah. The, the, the, the exchange of finances, it's, it's, it's cogs are ticking there. It's thinking, how can I get you to spend more tokens tomorrow?
Correct.
but, but it's interesting because I foresee a future then. But, and I didn't know this until we had this conversation, but it feels to me. That we're going to have a future in which certain AIs are going to meet the needs of a certain niche of people. So in the example that you gave, come to, I think you said, Bolt, if you want it to look nice. So we're the, we're the nice looking AI, you know, we'll, we'll give you a UI that looks beautiful. Oh, there might be a WordPress.
version or a Drupal version or a whatever, you know, we're really good at putting web pages together. Or we're really good. If you're, if you're into travel, we've got the database of all the travel agents and flights and blah, blah, blah. All of that. Maybe that's the way it's going to go. and you'll go to a different AI for a different thing. And then, you know, the And then the sun will never stop shining. And we'll all be incredibly happy forever and ever after.
And, and of course we'll never ever, none of us will ever need to work again.
Yeah. So, so how do you think this all competes or how does WordPress compete with this in your eyes? Like I look at it as we're going to, we could see 10, 15 percent of that market share drop. Where people go, you know, I'm just gonna build it. I'm just gonna build it over here and repl it, forget WordPress. How do you see WordPress sustaining through
I think my gut reaction at the moment is that it's going to be a really tough time for developers. I think people are going to experiment more and more with AI. And given that the WordPress code base is completely visible to all AI, it can consume it. So I'm imagining a scenario where a lot of simple tasks that might have been sent to your developer. You know, now you'll just ask you a 16 year old son, have you got, you know, this AI stuff? I've got a website and I've got this problem with it.
Can you just have a look at this for me? Just try it out on your AI, because you know, you're doing computers at school, that kind of thing. I think it's going to potentially eat in. To that market, but equally, maybe the expectations will go higher. So we'll all expect more from our WordPress website because the AI agents are helping us. So it's, it's a time of change for sure. And I think it probably is a time to, to reevaluate what you're doing.
I was speaking to somebody yesterday, very intelligent person who basically was saying, and they're in the WordPress space, so it's really, you know, on message, they were saying that really it's time to get into AI. If you don't start using it now, there will be people a year from now who are so good at it that they've kind of made themselves like that top tier and they will be the elite developers.
and even though they might not have a development background, if they do have a development background, they'll be, you know, elite plus. so I think, I think it is definitely a, a time of change. yeah.
really cool thing, It's not really an aside, but it's sort of an aside. Like when you're thinking use case, and even thinking outside of our little tech bubble, my, so another great thing about Bolt is it is pretty quick, and like I said, the UI is really nice. And what I found it to be really fun, with kids, is doing homework, right? With my kids, right?
So my, first grade son, my middle son, you know, they come home, they have their homework, there's whatever 12 words that they're working on, you know, for the week or whatever the homework is. And, you know, it's just, we're still writing on paper, right? Like the homework comes home, we're still using a pencil or a crayon, right? To do this stuff. So, you know, trying to get him in the, in a different mindset to learn these words, I was like, Hey man, how about we build a little game?
And if you can guess the right words, we'll have a dinosaur pop up with a little sound effect. And he was like, yeah, let's build it. So I just. Literally, take a picture, I'm grabbing my phone, take a picture of his homework, and I say, you know, and I paste it into, or I upload it to Bolt, and I say, grab these 12 words, I mean, I was so lazy, Nathan, I didn't even write the words, I took a picture.
And I said, look at these words and make me a game, a guessing game, for, my, six year old son to, guess these words. If he gets it right, it makes a little ding sound. If he gets it wrong, it makes a little wrong sound. and he has to guess the missing letter. And if he gets whatever I said, 80 percent of it right, we'll have. confetti and a dinosaur show up at the end, right? Just a little fun game for him to play with. And it built it in, whatever, two minutes.
And in two minutes, I had a little video game for him to play. And then we added to it. So I said, okay, this video game is great. Let's make a flash cards. And. It'll show the word for one second, and then go away, and then you'll have four words that are random, and the right word that my son has to pick from. A, B, C, or D. And then it built that, and we added to the video game. A little menu system at the beginning, and he could pick which game we could play.
And then I started thinking to myself, like, Oh, I could extend this, like, you know, I could make this a weekly thing, I could have a login, a top score chart, I could do this with all of his brothers, because we've got two other boys, and I was just like, my mind is just like racing with the, like the good opportunities.
what? It's really interesting. Of everything that we've talked about over the last 50 minutes or so, the most profound bit is that
hmm,
okay. That was impossibly complicated until really recently. But also, you've, there is no barrier to entry for a child to do that. So, a game that would have been the purview of a fairly seasoned developer, and that card game that you've just described, I imagine would have consumed a decent amount of a day. And you can imagine the cost of the equipment required, and the cost of the, you know, the amount of time spent learning all of that.
And now, You can have a child with a phone and a camera and a microphone and they can build something like that. And the reason it's profound is, is not so much that it's possible, it's that it's in the hand of a generation of people who are going to grow up. And that will be normal. And the idea that you would do anything other than ask an AI to assist you through this problem will be Like crazy, you know, why wouldn't I do that?
You know in the same way that I now rely on things that my parents It's just, it's beyond them. They've reached that point where certain technologies can't be used. Like the mobile phone is a real enigma to them, but for me, it's just so straightforward. And that's going to be the case, you know, old curmudgeons like me, I don't have that relationship with it and I probably never will, but the generation coming up, if they can just get a handy thing for every problem that arises in life.
You know, whatever it is, I don't know, build me an, I don't know, I've just come out of a tube station in an unfamiliar city. Can you, can you design me an app just to take me on the most interesting route possible around this city, take in some sights, and then give me a description of what I should be eating and how many calories it's got in it, and then send me to a decent hotel at the end, and whilst you're at it, will you just book it for me?
You know, I imagine all of that's possible with Google Maps, actually, now that I say it. But the point being, you can just speak. The interface is your voice. And that's remarkable.
There, I want to add in a couple other tools here, that I've, I've used and, and I think, another great way that instead of us looking at it like, Oh my God, WordPress is doomed. I think there's, I think there's some extensions that, that AI can help us with.
So another tool that I use, especially with my, my children, That is this tool called WhisperFlow, I don't know if you've heard of it, it's called W I, it's Whisper, W I S P R, Flow, WhisperFlow, and it's, it, it, you turn it on, I just turned it on, there's a little button you click and you just speak, and it takes your words, and puts it into text, it's very simple, but the, the key factor of Whisper, and I'm sure there's other tools that do this, this is the one that I found, is it will clean
up Your garbage vocabulary and all your filler words just like we do with Descript so if I turn it on and I go into bolt or my at my my chat GPT I Can just talk to it and be like hey, so here's what I want to build I want to build this app that does like a video game for my seven year old to learn I can just talk ums and ahs and spacing and pauses and then it just cleans it up and condenses it into the acute information that is needed Right?
So it doesn't have all the ums and the ahs, and then you just paste that into the prompt and you hit enter. And then it just takes your idea. So I was letting my son, my son do that, you know, and it's just because he's seven. So he's just talking and he's, well, dad, should I do this? And yep, tell it what you want. And then it just cleans it up into like bullet point text, all like the two minutes of him just like kind of like babbling and rambling.
And it just creates a prompt for them that's accurate. And then we just hit enter and it builds it. and I think, you know, tools like that, just like you said, these are going to be so common. Especially for, you know, for our kids as they grow up to like do this kind of thing. But here's the extension I think that's super valuable. that you can do with, that you WordPress.
Or any system really, but think about databases of important information, tax information, real estate information, land Plots, forgetting the word for it right now You know stuff that's impactful in your local community Information that's hard to get out from like just like your local government like who has access to what water filtration all this stuff That's like stuck in millions of databases around the world I'm building something now with Gravity Forms, API, REST API.
So, Gravity Forms has REST API. If you're using Gravity Forms to store information on your site, I can turn to Bolt and say, build me an app that analyzes the data in this Gravity Forms form. Here's the API keys. Go build me a table of that, that data. And by the way, analyze it with ChatGPT. And give me whatever, you know, whatever you think is the, the valuable information, draw a graph, do all this stuff, hit enter, boom, like it accesses data and WordPress is a great data store.
Like that's my future vision of WordPress. When Matt many years ago said operating system of the web, I can see WordPress as like a future, just database of content, open, accessible, build it your way. And then these AI tools can then go in and get this data and help us build robust solutions for whatever it is
Gosh, it sounds like, okay, so I can't, can't allow us to finish on a positive note.
You're english
that's right, we have to bring it down a little bit. And, so I, I do wonder, you know, what is going to be our capacity to cope with boredom in the future and to, you know, just to cope with not being able to achieve something like that, you know, if we're out of, if we're out of contact with our AI, if we go on holiday and the, I don't know, the, the internet breaks where we're living, you just have this like, but wait, no, I can't cope.
I don't have AI plugged in to my head kind of thing that, that, that, that. Bit of life does worry me. I do, I do see that, you know, there's something quite nice about being disconnected and the more that we connect, the more that we require to be connected and so on. So I do think there are some guardrails that we need to build in. and I don't know how busy we want our heads to be in the future. I mean, it all sounds so great, doesn't it?
All this amazing stuff that you can do and access to all this information all the time and the ability to wrangle it and manipulate it. Don't get me wrong. It really, really sounds amazing. But, maybe, maybe at this point I will go and sit on my desert island and. smash coconuts and drink the, drink the liquid that comes out of them and that'll be me.
Yeah, well, my lap. My last point is, like you said before, it have to get stuck in to AI and start learning it now. And I think that's all that I think that's all that means is to learn it and understand it. Like, I don't think anyone needs to be a superhero with AI or you have to go out and build apps. But understand the fundamentals are like when you used to crack open WordPress, you looked at HTML, CSS and PHP.
And you just looked at it, and you're like, I kind of get what this does, this functions. php file. Like, I think that's the same mindset you have to take with this, like, understand Vanilla, ChatGPT, and Claude. Then understand, like, what these other, like, Repl. it and Bolt environments do, and what the differences are, how to connect them up. You don't have to be building something super robust, and I think even that is a step ahead for yourself, you know, in a couple of
Yeah. Well, we'll see. that was a fascinating conversation, Matt. I really enjoyed that. Thank you. We'll come back in a couple of years time or we'll see if, you know, see if there's, there's anything left of humanity.
I'm either gonna be in Hollywood or like connected up to 16 machines. I don't know.
Well, good luck. Good luck in the year 2025. Thanks, Matt.
Yes. Thank you, Nathan.