[00:00:00] Speaker A: Welcome to Within WordPress. With me today is Matt Cromwell. Welcome, Matt.
[00:00:11] Speaker B: Hey, thanks for having me. Glad to be here.
[00:00:13] Speaker A: Well, thanks for having me. You invited yourself, so, I mean, I know.
[00:00:19] Speaker B: Thanks for saying yes to my. My bossiness, I guess.
[00:00:22] Speaker A: Yeah, no, no, you came with a. With a good and interesting question, so. And. And you were long overdue anyway, so for those of you who don't know who Matt is, please introduce yourself.
[00:00:35] Speaker B: Yeah, I'm Matt Cromwell. I'm a co founder of GiveWP, the number one donation plugin for WordPress still.
And I'm senior director of Customer experience at Stellar WP currently and also founder of WP Product Talk, which is a fun podcast that Remkas got a mention on just recently, actually.
So, yeah, among other things, I like to blog and write stuff and I like to give Remkus a hard time, so.
[00:01:05] Speaker A: Well, I think that's pretty synonymous for a lot of people.
For some reason, people think they can get away with it, but, you know, I'm keeping. I'm keeping track.
[00:01:19] Speaker B: Nice.
[00:01:23] Speaker A: There's a few things already you've mentioned that are interesting. I specifically would like to know more about WP Product Talk. Specifically.
I've been on it. Apparently I've also been mentioned, so, you know, there's that.
But I saw on the socials you mentioning, as well as Katie, Katie Keith, that you are going to present at WordCamp US, and I don't normally bring up when somebody's going to talk or whatever, but this is a very interesting case and I'd love for you to share a little bit more about that.
[00:01:58] Speaker B: Yeah, awesome, thanks. Yeah. Well, just for your info, the episode we mentioned you in was when we had James lepage on the show and he had just been on your show like a couple weeks before and we were talking about the core AI stuff, and I had told folks on WP Product Talk that if they wanted to hear about James and the AI Core team and. And things like that, that they should go listen to your show.
Because I was like, I don't want to go into all of that because he's been doing that in a lot of different places. We talked more about, like, the impact of AI in core for the future.
So it's like Remkus is a must listen and then come and listen to our stuff.
[00:02:40] Speaker A: So cool. Thanks for the shout.
[00:02:43] Speaker B: And yeah, it was a good interview. You did a good job.
But the. The word WordCamp us talk.
Katie and I had actually done a joint presentation at WordCamp Europe in 2023 and.
And it was, it was a WP product talk presentation basically where we highlighted a bunch of guests who had said different things about best practices for building a product in WordPress.
And we kind of wanted to do something similar for WordCamp US.
And so we pitched this in one of the, in the speaker application they gave you lots of different options of what kind of presentation it would be. And one of them was like other like pitch what you think would be a cool format. And we said honestly, we would love to do like a masterclass format of how to do product product in WordPress. Because the opportunity that Katie noticed in particular was she had gone to a Shopify conference and all of the presentations there, well, not all, but like a large majority of the presentations at that Shopify conference were all targeted at product businesses specifically.
And she contrasted that to your typical WordCamp that's really focused on the end user website builder a lot more and more in general. And she was like, there's just a shortage of really good product focused talks at WordCamps so let's, you know, go full throated and get a big bang and try to have a really significant presentation at workemp us. And we actually went back and forth with the programming team before we submitted to kind of talk through like different options and ideas and whatnot that we had to beef up the, the, the product oriented talks in general, not just for our talk, but in general.
And we landed on this idea of pitching a masterclass idea which they seem to like and we're excited that they accepted it.
[00:04:56] Speaker A: So yeah, I think this is one of the things where. So first of all, I'm super happy and stoked to hear that the, the content team over at workcamp us actually engaged into.
Let's try and find the version of this, this concept, this talk that just nails it on both ends.
I think that's highly recommendable because we don't see that too much at work camps.
For the vast majority that's just send in what you want to do and we'll just kind of go from there on what we like and what we don't like. This is actively engaging and recognizing because if you're engaging, you're recognizing the topic is something we need to have discussed more.
Yeah, what I, what I particular particularly like about it is the, like Katie hinted at the, the switch of who we're talking to.
I feel for me it's a bit of it. So first of all this depends on which word camp, but I feel wordcamps have started to realize it's smart to think about who we're talking to.
As a longtime organizer for wordcamp Europe, as well as wordcamp Netherlands, we've always been very careful of who are we picking as our target audience. But I've noticed many other Word Camps just are happy to serve. Right. They're come with whatever talk you have and whatever angle you like, and we're happy to facilitate that because, you know, that that also makes a Word Camp. But I like that we're going into a direction of more purpose.
[00:06:32] Speaker B: Yeah, absolutely. Yeah. There's, I mean, a lot of folks make big distinctions between kind of like your professional trade show type of conferences where they're paying, like, top dollar for big speakers and things like that. And wordcamps are the opposite of that and always have been. And it's been the opposite of that on purpose for real reasons, you know. But when it comes to what they have started to call our flagship conferences, which is the big continental conferences, let's say WordCamp Europe and WordCamp US and WordCamp Asia, it feels like there should be a little bit of a higher standard for the types of presentations that show up there and the way in which they get a little bit more choosy about the topics and the speakers themselves and things like that. So I agree. And I, I, I do think it's a, a good direction, and I hope to see more of it, honestly.
[00:07:30] Speaker A: Yeah. So most people don't even know what the origin of let's start a Word Camp of or let's organize a WordCamp. What that was. The intention was never to be just one version of a WordCamp. The intention has always been however small the crowd, however large the crowd, or however dedicated on one type of crowd, whatever you like, it's fine. Just start organizing.
I know the there's been a Word Camp in.
I'm gonna say this probably wrong, but I think it's Ventura, something like that. But it was essentially somebody's room. There's 20, 30, maybe 40 people, whatever. But that, that was it. That was a Word Camp, and that works. And there is what we now have had in Italy, the core days.
It's a good example of what a Word Camp just should have been it now we're calling it differently, but it's just way more targeting the actual group that we want to talk to. And it's not the goal to have two and a half, three and a half thousand people necessarily. It's the goal to educate. It's the goal to elevate.
We get caught up in the. Let's organize for the sake of organizing.
[00:08:41] Speaker B: Yeah, yeah, it's definitely evolved and changed. I mean, I do think the days of local meetups being really significant are kind of passed at this point, unfortunately.
I agree and I think that's a shame, honestly.
I feel like, especially as we continue to put so much focus on these big flagship events, that there's a huge need for local meetups to come in and kind of fill that gap. Because I noticed this and kind of highlighted it earlier this calendar year on X that if you
[email protected] for how many Word camps are scheduled or planned in the continental US in this calendar year, it's like two and preco it was, you know, a couple dozen or you know, up to sometimes some years up to 50 different word camps in a year just in like North America alone.
So it's, it's a very different environment that we're in now.
And I don't, I, I do like the continental Word Camps a ton. I just kind of miss the local experience a lot too. So.
[00:09:58] Speaker A: I, I hear you.
I, I also think it kind of so the reason for meetups not picking up and therefore WordCamp's not picking up as much as they potentially could have. I think it also has to do with the topic that we just touched on. The more specific or geared towards a particular crowd you're going, the less difficult is going to be to get a substantial amount of those coming to your meetups. I think it kind of goes hand in hand as we all specialize in our. And WordPress is large enough now for us to specialize like hyper specific previously what you had. So I started going to meetups in from 2007 or 8 and what you saw was either developers or heavy content people like bloggers and stuff like that. That was roughly it.
So the first WordCamp Netherlands organized in 2009 and 2010 was heavily focused on those two groups because those were the people showing up the meetups. And we knew if we, if we presented our. Our idea, that's the type of people, the type of people that would show up now. Yeah, we have way smaller niches and we need to do way more to get them to that meetup. And certainly in Europe that works differently because distance is. Is perceived as distance, not how it is in the US like if you need to drive for an hour to go to a meetup, that's a non.
It's not even comes across as, as a distance in the US In Europe very much. It does. I don't know how that works in Asia, but.
Or Africa for that matter.
But yeah, so I, I think we're kind of here because that's ultimately the balance where we end up.
If you specialize.
[00:11:52] Speaker B: Yeah, yeah, for sure.
Yeah. It's. It's a. It's a strange, different time that we're in by far. And I do wish that there was a way to, to influe.
Have.
I don't know if it's like big meetup weekends or, or if it's smaller word camps or if it's.
I, I feel like we tried to experiment. As soon as co hit, everybody tried to experiment with like, well, let's get this kind of content and quality of conversation online now. And, And I don't think that that was ever particularly successful.
There's just not enough of us who want to continue to sit behind our screen because we've been working behind our screen all day long and not listen to more things. You know, we'd much rather get out and see people. Like there just the. Seeing people and being together just changes the whole environment so significantly.
[00:12:51] Speaker A: I, I agree. For me, an online event is an hour, maybe an hour and a half, and then I'm done. Like, I'm done done. Because from that moment on, it starts costing me way more energy than I'm. I'm getting out of it.
100 in real life, I don't have that. I'll be dead tired after work, I'm sure. But, you know, 50 plus, it's gonna happen at some point. So, you know, I, I accept that.
But. No, but, yeah, the online, the move, the movement to online, I don't think ever was.
I mean, I think it was probably as best as it was ever going to get, but it's just not the same. It's just not even remotely the same.
[00:13:31] Speaker B: So, yeah, I also miss that.
There was a time when you would go to any WordCamp and almost everyone, especially in the States, they would say, raise your hand if you are. If this is your first time at a WordCamp ever. And almost half of the whole audience would raise their hand.
It was such a great funnel for bringing new people into the WordPress community and into the WordPress project.
And I don't get the feeling that that's the case with the big FL ship ones. I get the feeling overall, I could be wrong. Organizers might tell me I'm wrong, but I get the feeling that it's a lot of repeat folks.
[00:14:08] Speaker A: So for a very long time, this was the case for War Camp Europe. I know that for a Fact.
So bringing in fresh people.
I don't know of the last three, four work Camp Europe's with what the actual stats was or if they even asked to raise their hands. I, I, I don't know, but it's. Yeah, yeah. Again, very different.
[00:14:36] Speaker B: Yeah.
[00:14:36] Speaker A: Different times.
On the, on the talk you're going to do, can you share a little bit more on, like, what are the things we have to be expecting in terms of, you know, WP Product Talk, first of all.
[00:14:56] Speaker B: Yeah.
[00:14:56] Speaker A: Maybe even explain a little bit more what that podcast is specifically about and then perhaps share a little bit more on the content of the talk.
[00:15:06] Speaker B: Yeah, I mean, I started WP Product Talk now about two years ago now because I wanted to have what I have felt like is like one of the big advantages of wordcamps is talking with people in the hallway track, as we call it, about your business, about what's going well about what's not going well, about things you've learned. And in sharing insights and things like that, I got most of the value in those hallway track conversations.
And I thought I would love to have those conversations a lot more often in public so that others can learn from these things.
So I thought a podcast was probably the most effective way to, to do that.
I started it out as a.
What was it called again? Twitter space. It was a Twitter space.
[00:15:59] Speaker A: Okay.
[00:16:00] Speaker B: Pre, Pre. Elon Twitter space situation where you just jump on your phone, because Twitter spaces only worked on phones, not on your desktop or anything, and have a conversation live and people just joined and you could even add people into the conversation if you wanted, and things like that.
It was great, it was fun, but a lot of folks were like, oh, I'm not on X. I'd love to hear it, but I'm not there.
And can you do it somewhere else? And so we decided to do it as a live podcast, streamed live to YouTube and then distributed to all the podcast platforms through a WordPress website through the seriously simple podcasting plugin.
And our target audience has always been that we are WordPress product owners, having a show for WordPress product owners all about how to do a good business, how to, you know, do best practices and learn with each other and grow.
And it's, it's been really great. We just crossed our 100th episode just recently.
That was the episode that James LePage was on. It was awesome. It was great. I gave it a bunch of hats. So that was a fun time.
And yeah, it's been quite a ride. So with that in mind, we were like, we can bring a lot of the insights that folks are bringing into our episodes. We can bring that to a WordCamp talk.
And now that we've interviewed, you know, I think across a hundred episodes, we've interviewed at least, like, 85 people over that whole period of time. All product owners or product employees in one way or another, all doing what we feel like are great things in the WordPress product space. So it's a ton of insight that we can provide in a short amount of time, hopefully.
[00:18:05] Speaker A: Yeah. Can you give an example of the type of insight, like, very concrete example of. So this is the type of stuff you'll be learning.
[00:18:14] Speaker B: Yeah. Like, one thing that we joke about every once in a while is it's ironic how many times over all of this, all these years, that whether we're talking about development or marketing or customer experience or business, that somebody will say during the course of the conversation how you need to listen to your customers.
And it's funny because it feels very, very, like, obvious or like, oh, yeah, of course I listen to my customers. But I literally, because it's part of my job to listen to customers and do something about it, I will talk to other product businesses and I'll ask them, so how do you know what your customers are actually saying about something? They say, oh, I hear it all the time. I was like, yeah, I know, but, like, how much and what percentage of your customers feel this way? And I'll lean in a little bit. And. And they very often have no idea how to answer those more detailed questions because they're not logging the feedback. They're not doing something about it. They're not being actually very actionable about it. So even though we'll say, oh, of course I listen my customers because I. I jump into the. The help scout queue or I look at my support tickets or something like that, that might be something that you do occasionally, but that. That is still a very pretty anecdotal knowledge.
[00:19:33] Speaker A: Yep.
[00:19:34] Speaker B: You can have a lot more actual data and actual knowledge about percent. Here is the percentage of people who got refunds who said that they are actually leaving the platform because they don't have a WordPress website anymore. Like, do you know that kind of thing? Or they're what percentage of your refunds are leaving because they went to one of your competitors.
So it's a very different thing to say. Anecdotally, I listen to customers. It's another thing to actually have a process for how you process listening to your customers.
[00:20:09] Speaker A: So we have a saying in Dutch that goes, maten, is weighten, meaning if you measure, you can actually know.
[00:20:20] Speaker B: I like that.
[00:20:21] Speaker A: I guess, I guess we like the rhyming part of it. But, but it's essentially the, the, what you're saying is this is one of the tips, but what you're saying specifically is if you don' looking into reasons behind the things that you see come across your board, you're not actually learning about opportunities to improve or at least know the, the possible impact you might have had.
[00:20:48] Speaker B: Yep.
I mean, sometimes folks will say I listen to customers and then they go and build a whole brand new feature and then they find out that they were only listening to the loudest customers.
And it wasn't that people really needed that. It was just that they had some really loud people in a Facebook group.
[00:21:07] Speaker A: So it, you know, we have, we, we have. When I say we is Barry and I for scanfully.
We have sort of agreed with each other that we're, we're actively listening to feedback. Yes.
And certainly when it comes to little improvements of the existing functionality, but in general, in terms of new features, we're going to prioritize whatever we have listed first and bring that to fruition before we very actively start to monitor for additional features and stuff that we haven't thought of in any way, shape or form.
And then when we do, we're going to have to find a way to take into account the loud voices.
So Barry's been selling products for WordPress for ages. So he, you know, he's, he has a lot of hands on experience. But it's. Because what he said made me think like, yeah, okay, so prioritizing loud voices sounds logical because you get rid of the loud voices, but it's not definitely, not necessarily the smartest thing for your business.
[00:22:17] Speaker B: I mean, what do you mean by that?
[00:22:19] Speaker A: Well, if you add a feature because there's loud voices and you spend a lot of time testing, building and all that stuff and you have it implemented and the adaption rate of that is just poor on every single level, then what have you really spent time on? If you spent that same time on a feature that not only solves a latent problem, those are the best, but it's a, it's a, it's a feature you've added that nobody knew they wanted, but when they saw it was there, they got really excited about it and then started sharing and started talking about it. All that sort of stuff.
That's just way more valuable. And I'm not saying you don't need, you shouldn't listen to the loud voices. And they might have good ideas, but it's not a given. So you need to validate whatever you're doing anyway 100% and then kind of disregard the loud voice. Like, okay, ask everybody. Here's three things we're thinking of implementing. What are your favorites here?
[00:23:18] Speaker B: Yeah. One of my favorite anecdotes about this came from Pippin Williamson, who, you know, it was not in the WordPress space that much anymore, but, like, built easy digital downloads.
And at one WordCamp hallway track conversation with him, he was like, you know, I think customers are great. I listen all the time. But if all of my customers told me that I need to build a shipping module for easy digital downloads, I would ignore them completely and tell them I'm never going to build a shipping module for an easy digital downloads product. It just doesn't belong. And that's because I know what my product is, and I know what it's for, and I know who it's for. And all the folks who are saying that they need shipping actually want woocommerce. They don't want EDD anymore.
And that little anecdote has always stuck with me really well because there is a sense in which customer feedback and, and opinions are really important and should be really heightened in your product roadmap a lot, but not at the expense of the scope of the core purpose of your product. You still have to be a good product owner and with a. With a good head on your shoulders.
[00:24:35] Speaker A: Yeah. Yeah. Product ownership is. Is both easy and extremely difficult at the same time.
[00:24:43] Speaker B: 100%.
[00:24:43] Speaker A: I don't. And I don't think that will ever change. I think that's just the nature of the beast.
How, how, how long is the presentation?
Is it a 44?
[00:24:52] Speaker B: Honestly, we're not totally sure yet.
We got to figure out those details.
They.
[00:24:58] Speaker A: Hour and a half.
[00:25:00] Speaker B: It's going to work. We'll see. We'll see. There's a. We've pitched a lot of different kind of ways that it could flesh out, but we're still trying to work out those details. Katie is actually away this week on a short vacation, and we're going to probably get. Get together next week and, and really try to nail down all the details, so.
[00:25:22] Speaker A: Cool.
Well, I'm, I'm looking forward to it and I will definitely be in the crowd.
[00:25:29] Speaker B: Awesome.
[00:25:29] Speaker A: Because, yes, I am going to work mpos.
[00:25:32] Speaker B: That's awesome. I'm looking forward to it, too. I'm curious how it's going to go.
Yeah.
[00:25:39] Speaker A: Hey, you. You mentioned AI A Bit from the James's recording as well as in general a bit. Your. Your question you asked me, which kind of prompted us sitting.
Well, I was gonna say across, but I have no idea where you directionally are, but although that is south. So, you know.
Anyway, you had an interesting viewpoint with regards to the implementation of AI and the questions you had around that.
Could you summarize that for.
[00:26:20] Speaker B: Yeah, yeah, I just. So there's like several different little inklings of trends and things that I've noticed about AI usage in code or in WordPress, let's say, that have made me really curious about where this is all leading.
And it's everything to do with the way that I'm using AI myself on my personal website and on the WP Product Talk website, how I'm, you know, coding stuff that help me solve problems from a customer experience perspective at Stellar, all these things.
And you have a unique focus on performance and security. And so those are two aspects of all of this that I feel like are a little, I don't want to say in jeopardy. I don't want to get like drama queen on this stuff. But like, I think there's a big.
Yeah, I think there's a big opportunity here for everything that goes south badly to be like, oh my gosh, we can AI the best things in the world right now. And.
And now all of a sudden we've AI'd all the great stuff and our websites all come crashing to a halt and there's a billion security vulnerabilities going on and everybody's hacked and slow.
I mean, it can go that direction. But me, myself right now, I've been like, well, I don't want to always deal with like, I'm trying to, you know, my personal website and WP Product Talk, these are things I do outside of my business and I need them to not take a ton of time. So I will AI a snippet and I'll throw it into like the code snippets plugin. Right. All of a sudden on WP product doc, I have 25 code snippets in the Code Snippets plugin, where I had four before.
And like the lines of code altogether is in the hundreds. Probably at this point, maybe all of that should be one functionality plugin. I don't know. I'm not sure. I'm not sure which one. If, if I did all of those different snippets that I'll do a hundred different things, put them in one functionality plugin versus having them in the Code, snippets, plugin. I'm not sure which one would perform better or which would be more maintainable or things like that.
The other part is that, you know, WordPress has this, has this perception from the outside of security being a problem in WordPress. And we all know as insiders that it's not WordPress itself that is the security issue as much as it is the plugin ecosystem that jeopardizes the security of WordPress.
And that said, there is a huge benefit to the way security is managed and done in the plugin ecosystem compared to a whole bunch of individuals AI ing their snippets on their website.
From a security perspective, I kind of feel like I trust the WordPress plugin ecosystem a lot more than folks like me who are just AI in snippets into their websites all the time.
So on both of those fronts, I'm like, AI is awesome. It's democratizing the ability to code the stuff that we want to.
But maybe that's the greatest thing and the worst thing all at the same time when it comes to a performance and security perspective.
I didn't keep that very short, but that's basically the problem as I'm seeing it right now.
[00:29:57] Speaker A: Yeah, I don't think the goal was for you to tell that in a short fashion. I think the goal was that for you to, you know, convey all angles of what is an interesting problem. Because it is a problem, but it's an interesting problem that we're having.
There's a lot of stuff to unpack, but I'll start with the most obvious one.
As good as snippet plugins are, in my purest point of view, they should be banned.
For the simple reason is that they allow you to store PHP execution code on a database level.
Now, not even considering the possibilities of if someone gets access, they'll have an easy way already, but specifically just from a architectural standpoint, that should never ever be on your favorite list of how to solve a problem.
So the first step would would need to be you thinking about site functionality plugins. Yes.
When you're going site functionality plugin, you immediately have the opportunity to structure whatever you're doing, meaning code that belongs together. So everything you're doing, for instance, inside of Woo, or inside of a seriously simple podcasting plugin and filters and overrides and all that sort of stuff, you, you, you, you sort of combine that together into one file so stuff stays organized.
Just as important as that move is, the move to document is because I'm guessing You don't have a whole lot of documentation as to why you have implemented this, nor have you asked the AI specifically to walk you through the code in the code. So the internal documentation sort of stuff, most people just forget that because they're just looking for a solution and let me just copy and paste.
But if, if you're looking for an answer in terms of should I even be doing this? Well, yeah, this is how you extend WordPress, right?
If you're not using an existing plugin to extend, then that's the thing you do.
Moving it to a well documented, well structured site plugin is the smart move because that introduces separation of concerns. So there's stuff in your file system, not in your database.
And just for the record, if you have a snippets plugin active, it's by pure definition your site is less secure. That's because just the pure nature of what that does.
And I mentioned that before. But the second part, that is from my performance standpoint, an interesting thing to keep in mind. You're loading an extra plugin just for the sake of it retrieving extra functionality code out of the database into active memory, roughly on all page loads.
[00:33:06] Speaker B: Yeah.
[00:33:08] Speaker A: So that means if you solve something specifically for a WooCommerce checkout page, you will load that snippet and I'm going to have to assume you probably didn't escape that snippet from the perspective of only do this on the WooCommerce checkout page.
So it's going to load on every single page, which makes no sense. There is nothing optimized about that.
And depending on what you do, that can lead into serious memory hogs and resource usage if the site gets a little bit more than average traffic, especially if that site is having uncached traffic.
So there's a whole bunch of security and performance concerns just by this mere fact of you using a snippets plugin.
So my advice would be to think more in regardless of what, irrespective of what AI solution you're using.
But I would do all the things per topic in one chat and then ask it to look at all the code it's generated, explain it the way it should be generated and assume it's the same specific class you class PHP file. So for instance class hyphen woocommerce overrides PHP and inside that files will be everything you override in WooCommerce, if it understands that sort of context, it will present you a file that is in namespaces and all of that is perfectly done. If you don't ask it to adhere to WordPress coding standards, it's not going to do it. If you don't ask it to consider not loading when it doesn't need to load, it's not going to do that. If you don't ask it to optimize or ask feedback from you. I have two of three different ways of implementing this. How should I implement this? If you're not doing that, you're not guaranteeing you're getting the most optimized version. If you're not asking it to look at escaping, you know, security escaping and nonsense and all of that. So your inputs, if you have them, are just whatever I paste in there, it has direct access to whatever I do on the database, which is the vast majority of how you get cross site scripting type of problems.
Then there's this. There's a long layer of things you are introducing into your solution that are just problematic from a purist standpoint and problematic from a. Oh, I've been compromised.
Yeah, a few times over.
So it, it helps to have a purist state of mind when looking at these flyby type AI implemented solutions.
There is, this is.
I'm very much aware that this is an uncomfortable truth for those who want to just quickly solve a problem.
But the underlying problem that needs to be fixed is people understand that if they don't implement these types of things, they'll end up with a situation where your site is just never fully scalable, performant and secure. And those are base things you want because I promise you there will be situations where they are beyond their control of how people have manipulated the poorly implemented solution. Now if it's private code, it's roughly.
There's one argument to be made against if it's private code and nobody knows what you're doing on that site, how unsafe can it be really if nobody knows how to target you or nobody has extra account on your site? Sure.
But that doesn't mean it's impossible because there are certain ways you can start hammering your site and just try to penetrate it in whatever way.
And that's happening. That's happening 24 7. Because we've all seen the bots, right?
[00:37:25] Speaker B: Yep, 100%.
So no, I super appreciate that. I think, I think the like, as somebody who built websites in the past, like I know that a site functionality plugin is best practice and I made choices to use snippets because I was like, I just need like two snippets. I'm just going to Drop them in there, that's one thing. But I think the big reason why, in my mind it's such a concern is that the folks who want to use AI to code are not going to take the time to figure out how to FTP a site functionality plugin up to their website.
They're just not going to do it. They're going to lean on a code snippet plugin for sure, potentially because that's also what AI will tell them to do.
It'll be like, okay, now you need to put this into FTP and blah, blah, blah. And they're like, no, I don't want to do that. Is there, what else can I do? And it's like, oh, you can use a code snippet plugin.
Like the, I think the usage of code snippet plugins in WordPress is going to exponentially grow over the next couple years.
[00:38:28] Speaker A: I think that's fair because, you know, the, the AI will just feed you with the, the, the, the information that matches the request that you've made.
And at what point do we, so if you don't ask it to, to adhere to coding standards, it's, it's not going to automatically do that if you, if you don't ask it to, to think in performant ways of security ways and you know, specifically, or harden this as much as you can.
If you don't ask it to it, it's not going to do the job at, at its best of its abilities.
Yep, yeah, it's, it's a real problem. And, and one of the things that I've said is what I like about us now, having the ability to use AI to create functions and snippets and codes and all that to make WordPress do what it, what you want it to do.
A huge upside of that is we're introducing WordPress in its fullest flexibility to a group of people who have never known it to be this flexible.
So there's a certain layer of.
[00:39:41] Speaker B: I.
[00:39:41] Speaker A: Think calling them software developers is too much, but there's a certain level of WordPress professionals that are now understanding that there's more to their site than just installing that plugin. They can indeed have flexibility and, you know, really play with functionality and all of that. And there's, there's fancy stuff you can do with even just snippets.
So what I like about it is that we are now seeing a new group of people joining WordPress in what I think is a more interesting way than just, you know, consuming the end products.
But yeah, it's, it's still not without risk. So there is, there will be a point in time where they will get hacked and they will have slow ass sites coming to a halt and they will not understand why and they will need to look for external advice when that's the case and hopefully they're not too deterred and understand, okay, I need to be slightly smarter and thus become better at what they're doing. And then, you know, I'm pretty sure in two, three years we can say here's this, here's a defined group of people who are, you know, smart in what they build but don't necessarily understand the basics yet.
Which is fair because unless you've studied computer science and you started from a background where you understand the importance of certain things happening in programming or not happening in programming, you know you're, you're going to be limited in what you understand and how things should work 100%.
[00:41:17] Speaker B: I mean you can end every prompt by saying and make sure to make it secure, performant and accessible but that doesn't mean that it's actually going to do any of those things.
You have to be able to have enough knowledge to validate that the output.
[00:41:35] Speaker A: Yeah, one of the things I'd like, I like using is whenever it comes with a solution. So for instance, here's a few things that I love using AI for the same sort of snippets you use because I don't, you know, if I don't need to scour through the documentation or the files of a particular plugin to find a filter or a hook or whatever, AI will do that faster anyway.
So I'll use the code base, look at this. Find how I could extend xyz.
I'll always ask it to and this is probably one of the best tips I can give you. I'll always ask it to find what you can find. Report back to me and ask me three questions you need validation on before you proceed.
Super simple line, but it builds in. I'm checking what you're coming up with and you know these things you can script, you can use default automated ways of, of whatever building or sorry whatever thing you build. You're adding this to the mix of questions. So the mix, so the, the same thing in the same way you can add.
And I should probably blog about this as I'm realizing it, but there's a few snippets where you can say include your best practices knowledges from and then name your field. So security, performance, accessibility, whatever it has access to the knowledge, it doesn't necessarily know. You prioritize it. So it's just going to prioritize the basic request you ask it.
But it's super interesting see that group become more active with WordPress because I, I do see that.
[00:43:26] Speaker B: Yeah, yeah, it, it's, it for me, it's, it's helped me a ton to code again. I literally, in my day to day, I was not coding really anything.
And I did start, I'd say about six to seven months ago, I started using Cursor for, for things overall. I have a article coming out next week. I have an article on my site about building a WordPress plugin with AI. And I did, I wrote that over a year ago now and it was with like a custom GPT and I was just copy pasting from ChatGPT into a file and I was like, look, you could totally do it, it totally works. But you do cursor, it gives you all the context of your whole entire WordPress website.
And I literally built a pretty fun little plugin that I didn't code any of it. I don't think I touched a single line of code.
I validated it.
I'm able to see all the code and validate the output and tell it to do it differently and all those kinds of things.
But Cursor did all of it and I'm going to be shipping it to my website as a plugin, not a functionality snippet.
And it's, it's, it's a fun little thing, but it, to me, it's amazing. I can't believe that we've gotten here so quickly.
It. The next, that iteration of where these things go, I think will probably shock us again, most likely.
[00:44:58] Speaker A: Oh, I think the.
[00:44:59] Speaker B: Which is good and bad, honestly.
[00:45:03] Speaker A: Yeah. And I haven't decided which direction I, I'm leaning towards most, if I'm really honest.
And that's more to do with.
There's a level of unchartered territory here that is far exceeding what we possibly can imagine because. And, and so first of all, let's, let's at least acknowledge that you and I most likely are still in the 1% of 100 people.
There's just a really small percentage of people actually working with AI in a, in a meaningful way.
Yeah, let's, let's at least start there.
But from that 1%, turn that into 100 people.
I think there's probably one or two who have a proper full grasp of where this thing is going because I know for a fact that the ones Developing the AI, building on it don't have a clue.
They don't necessarily think in terms of problems, of things that need to be solved. I read an article, I shared it in, in my Within WordPress newsletter a few weeks ago.
It, it highlights that they just, they're building the thing and they understand the thing you want to solve with it, but don't, they don't have all the problems in the head, you know, in their head as we do. Right. So as soon as we see, oh, here's this magical thing, let me throw something at and see what sticks, we're going to have different ideas than they do.
But if the assumption can be set to true that the vast majority of people doing meaningful stuff with AI seriously have no clue what the capabilities be like in a year, three years, five years, I mean, I can't, I myself can't begin to comprehend what the potential is going to be. All I can see is that it's going to become increasingly more easy to tell a story to a machine which, giving a particular context I've given, whether that's the context of I'm doing it in cursor, which is then software related, or I'm doing this in VO type of environment where I'm creating movies or whatever.
But we will increasingly be able to wield the magic wand into whatever direction we want. And that's wild because that means I.
A friend of mine, he's retired now, but he's a cobalt developer and he made a fortune in being one of the very few people who understood COBOL enough to maintain and fix COBOL applications and stuff for large financial institutions such as insurance companies and banks and whatnot. So he's made a killing in that.
He made a killing in it because he could.
But at some point AI is going to be. I don't care what language you throw at me, give me the documentation and I'll figure it out.
Yeah, and that then becomes scary. And I genuinely mean that when we start making the switch from generative, because that's where we still are to actual version of coming up with something entirely new, which is we're miles, miles away from that.
That's where it becomes interesting.
So the word AI is a misnomer in my book because it's yes, it's artificial, but it's for sure not intelligence.
Yeah, because intelligence comes up with something nobody else has.
[00:49:02] Speaker B: Yep. It's not creative. It's still not creative.
[00:49:05] Speaker A: I'm computer, it's generative still.
It can only create based on something. It knows.
Yeah.
But that level of uncertainty of where this thing is going is slightly concerning in my head.
[00:49:21] Speaker B: Yeah, It's a classic Jurassic park kind of situation.
[00:49:25] Speaker A: It has that potential.
[00:49:26] Speaker B: I'm so excited that we can. That we're not thinking about whether we should. Yeah.
[00:49:31] Speaker A: I'm sure you've seen the meme where somebody says, I'm saying thank you to every single prompt and, and receive that I get from the AI just in case this.
[00:49:40] Speaker B: Just in case.
Yeah, exactly. Let's be polite to our new overlords for sure.
[00:49:47] Speaker A: Yeah, exactly. So, yeah, I see those memes and I go like, funny.
But also, if it has memory, if we.
And you know, even if you look at where and how we're storing memory, there's interesting projects going on where DNA is being used to store data, where crystals are being used to store data, and there's, you know, gigabyte terabyte, petabyte, however high you want to go.
These solutions, they have not found a way to stabilize them yet. But the opportunity that those methods hold for the storage of data and combine that with computing power, we are already exceeding beyond or getting close to atomic level.
At a certain point, they, those things are going to converge into a scary machine for sure.
[00:50:53] Speaker B: Yeah.
Yeah. I mean, that's a big part of the whole thing, is that all of this doesn't happen in a vacuum.
It all has to be powered by something.
And pushing the limits of, of how we use AI today means we're going to be pushing the limits of hardware in general and how things are computed, how much power we can generate, how much bandwidth we can save. All those things have to innovate along with all of this or else, or else to get. It stagnates.
And it's, you know, it's a similar problem with like, electric cars.
We would love to see electric cars everywhere, but there's no infrastructure to support them at all.
[00:51:34] Speaker A: We would not.
[00:51:35] Speaker B: You can't. You're going to really like cart before the horse on that one.
In many ways, I feel like AI is similar. It's like now we have this ability to, like, process amazing things with AI if only we can have the bandwidth to catch up to it, you know.
[00:51:51] Speaker A: So I've in fact seen, I think two, two, two companies of late probably this week where I saw something where they said, we have stopped investing in AI because we cannot keep up.
And I go, and, and those were companies who could benefit from AI.
And I go, like, we, we, we can't keep up.
So we're just gonna Keep it basic and just do a few things but just kind of look at it for now because just we just simply cannot keep up.
[00:52:26] Speaker B: Yep. And I got that. It's a lot for sure. I mean even in stellar right now we have been trying to use it and incorporate it into our workflows in different ways, but it feels like every three months we're like, oh wait, now there's a whole nother way that's way better for us to do all of these things.
And honestly a lot of that is disruptive to day to day operations.
So I can relate to folks who are like we're trying to keep up and all it's doing is making us slow down and have to reinvent the wheel every three months and that's not healthy for our business in the long term.
[00:53:08] Speaker A: One of the things I like playing with from the very early days from ChatGPT, well the, the 3.0, the, the meaningful version to play with.
There was stuff that I was just testing out. I, I played with JavaScript, I played with PHP, I played with Bash, you know, all these types of things.
And the fun thing of being what is it two years now later that the original question that I asked it to produce, I'll just find that thread again, change the agent that I'm using and I'm asking the same question again and it's wildly how, how much that's been improved and how much extra, I don't know context it gives and it's just wild what the difference is. And if, if I couldn't perceive where that went in two years, then I'm for sure as hell have no clue what this is going to be in five years.
[00:54:09] Speaker B: Yeah.
[00:54:10] Speaker A: Except for understanding that it's going to be monumental.
[00:54:14] Speaker B: Absolutely. I think the other big factor in all of it is the amount of private information that we're sharing with these AIs. Honestly, like I wish that OpenAI was a lot more aggressive about its user security, I'll have to say because you know, a lot of folks will, will joke about like oh, I logged into my grandparents laptop and I saw their Google search history and it's so embarrassing.
Like the types of questions that we ask AI nowadays are probably a lot more embarrassing than what we would do in a normal typical Google search.
And that's like amazing on one sense, but it's also frightening like in terms of privacy laws and all those types of things. Like your, your AI account starts to be like a much bigger trove of personal information than your email is and that's really got to be a big consideration folks take in as well.
[00:55:22] Speaker A: And the fact that OpenAI and probably all the other ones for the same reason are not allowed to fully delete whatever chat you have deleted, because it's still there. It's just not accessible for you anymore.
Yeah, that's a concern. And you know, there's many of these types of concerns because for the most part these are relatively young companies.
So they're there are, you know, they're just trying to figure it out as they go along. Anyway, short legal is on their side and they have good legal teams and it starts with the best intentions and whatnot, but you end up with a bucket of data on a incredibly price. I know, I know people who do whole psych evals sort of. I'm feeling this. What could it be?
I'm worried about this. How could I solve this? Or this, Is this stressing me out? How should I handle this? You know, I will never ask a computer any of those for a whole wide array of reasons, but privacy for sure is one of them.
[00:56:23] Speaker B: Yeah, potentially. But the thing is the AI encourages you. They're like, oh, yeah, yeah. Should I give you a whole like two week plan of how you can combat your depression? Like it, like it'll, it does those follow up questions to, to encourage you to, to ask it more and to dig in even more and to share even more.
It's really fascinating.
[00:56:46] Speaker A: It is.
I think I was gonna, I was gonna say a second. It is super fascinating as it is scary.
[00:56:54] Speaker B: Yeah.
[00:56:57] Speaker A: Well, all right.
I would say on that bombshell, thank you so much for being on the podcast.
I think this is a fun, fun conversation for anybody who wants to hear a different approach to AI and the friendliness of it. And at least it should. I hope it, this makes you think 100% appreciate it.
[00:57:23] Speaker B: Glad I was here.
[00:57:25] Speaker A: Glad you were here. Thank you so much.
