[00:00:19] Nathan Wrigley: Welcome to the Jukebox Podcast from WP Tavern. My name is Nathan Wrigley.

Jukebox is a podcast which is dedicated to all things WordPress. The people, the events, the plugins, the blocks, the themes, and in this case, building trust with project contracts for reliable AI development.

If you’d like to subscribe to the podcast, you can do that by searching for WP Tavern in your podcast player of choice, or by going to wptavern.com/feed/podcast, and you can copy that URL into most podcast players.

If you have a topic that you’d like us to feature on the podcast, I’m keen to hear from you and hopefully get you, or your idea, featured on the show. Head to wptavern.com/contact/jukebox and use the form there.

So on the podcast today we have Chris Reynolds.

Chris has been immersed in the WordPress world for around 20 years, with a career spanning agency work, and now leading the developer relations team at Pantheon. His deep WordPress experience is paired with a playful approach to emerging tech, making him a familiar voice at events like WordCamp US and across the wider community.

In today’s episode, the discussion explored the fast changing landscape of AI assisted development. Getting into how large language models, like Claude, are not only changing the way developers build for the web, but also how they’re collaborating with these tools whilst mitigating risks.

Chris shares insights from his recent talk at WordCamp US, Parenting Claude: Guardrails for AI Assisted Development, where he introduces the concept of a project contract. Essentially, a structured set of rules and checklists built directly into the coding workflow to ensure AI generated code meets standards for testing, accessibility, security, and design.

We discussed the new found playfulness in web development. Chris makes the case that, thanks to AI, the traditional barriers to entry are eroding, enabling even those without a formal developer background to build an experiment much more easily.

At the same time, we need to be mindful of the importance of establishing robust guardrails, like automated reviewer agents, and pre-commit hooks, so the AI doesn’t take shortcuts or accidentally introduce chaos, or secrets, into your code base.

The conversation focused not only on technical solutions, but also got into bigger questions about trust, how much oversight AI really needs, and the evolving role of developers.

Chris emphasised that while these AI tools can accelerate and democratise coding, they can also pose a challenge to conventional wisdom around code review and software reliability.

So these new frontiers are both exciting and a cause of anxiety for developers who can sense the industry move from the deterministic, step-by-step processes, to the unpredictable and occasionally wild world of non-deterministic AI models.

If you’re interested in how AI is transforming web development, from practical workflows to philosophical ramifications, and how you can implement AI without losing control of your projects, this episode is for you.

Just before we begin, the manner in which this episode was recorded was not my usual process, and so the audio is not quite as good as I’d usually expect to deliver.

If you’re interested in finding out more, you can find all of the links in the show notes by heading to wptavern.com/podcast, where you’ll find all the other episodes as well.

And so without further delay, I bring you Chris Reynolds.

I am joined on the podcast by Chris Reynolds. Hello, Chris.

[00:04:11] Chris Reynolds: Hey, Nathan. Is this the first time we’ve met in human forms?

[00:04:14] Nathan Wrigley: Yes, it is the first time we’ve met. I’ve seen you on a screen.

[00:04:16] Chris Reynolds: I know I’ve seen you on a screen too.

[00:04:18] Nathan Wrigley: You’re joining me on the podcast today so that we can have a chat about the bits and pieces that you do in the AI space. But before we get into that, you are doing a presentation or have done, how did it go?

[00:04:28] Chris Reynolds: It went really well.

[00:04:29] Nathan Wrigley: Okay. And it was all about LLMs and constraining them and making them do your bidding as opposed to them doing whatever it is that they like to do. Do you want to just tell us a little bit, first of all, let’s do your bio. Just tell us about who you are, where you’ve worked, where you currently work, what is it you do in the WordPress space? Let’s begin there.

[00:04:46] Chris Reynolds: Well, I’m Chris Reynolds, I’ve been in WordPress for a really, long time, 20 ish years I’ve been using it. I’ve done kind of everything you can possibly do with WordPress in different forms. Of course, over the course of that 20 years.

Most recently edging into agency world with WebDevStudios and Human Made, before joining Pantheon, about five years ago. Joined Pantheon as a software engineer, working on the platform, and the sort of like space between where the platform ends and where, like your application, your WordPress site and whatever begins.

And then about two years ago, actually almost exactly two years ago, I joined the devrel team as a developer advocate, became a senior developer advocate. Now I’m a senior manager, developer relations. So I’m leading the developer relations team. And yeah, really love doing this stuff, because this stuff is really fun and exciting. I get to play with things more. And I get to like do the things that like our developers in our community are doing, and then come up with like what the sort of best practises for those things are.

Which is one of the reasons why I gave this talk, but also did the things that led up to giving this talk. Because one of the things that I realised very early on, like very early this year was that AI was going to be the thing that we’re going to be talking about in 2026. And so Pantheon and therefore me, needs to have an opinion about what that looks like. What is the future of that? And if I’m doing my job as a developer relations, I should have an opinion about how we can interact with AI in a way that doesn’t steal our jobs.

[00:06:13] Nathan Wrigley: I’m going to segue a little bit because about a minute ago you said the word playful, and that just set my mind going off in a completely different direction because it feels like technology for many years has been the workspace for people who have a fairly, how to describe it? They’ve got a certain mind, if you’re a developer, you probably have a certain way of approaching your code and you’ve got to be very regimented, and you’ve got to do things in a methodical way, should you wish to ship.

But it feels like last few years, like all bets are off with that. And creativity and playfulness has suddenly become a thing which is available to, going to do air quotes, vibe coders. So thoughts on that? Really? Is it a more playful industry than it was just a few years ago?

[00:06:55] Chris Reynolds: I think it depends. I don’t think that ever really went away. But I do think that there is less sort of incentive or focus to play with technology, particularly web applications, which is what we are in the business of doing. And I think there is this perception that I can’t do that ’cause I’m not a developer, but like with AI and vibe coding and Claude Code and things like that, like it opens the door to people who wouldn’t have been able to do those things before, to start experimenting and realise, oh, actually I can do that thing.

And for myself, one of the things that I, I wrote this on my blog a couple months ago. I wrote a blog post that’s talked about like how I don’t miss not writing the code myself, because for me, and it’s not the same for everyone, but for me, the practise of writing the code was never the part that was fun.

[00:07:50] Nathan Wrigley: Okay.

[00:07:50] Chris Reynolds: The part for me that was fun was building stuff. And I feel like I can let the robot do the coding and I can build the things and be satisfied with the output. And I don’t necessarily care what’s under the hood.

For my own stuff it doesn’t matter. But you know that obviously if you’re scaling that to a, you know, enterprise or whatever, you need to have more checks and balances than like my little nifty side project over here. But it does mean that like, I do a lot more weird stuff than I felt compelled to do before. So it has for me, revitalised this, sense of play, in a way that maybe before I was too burned out or too busy or too whatever, to have the time or head space to even think about.

[00:08:33] Nathan Wrigley: Yeah, because it feels like people with a non-developer background have suddenly kind of the constraints are off, and they can achieve things that previously it would’ve been 10 weeks work for a panel of amazing developers. It feels like now you can perhaps achieve that in an afternoon and a few sentences. I’m sure all the constraints around whether or not it’s ready to ship for enterprise that is all still understood. But just to get something on the screen which satisfies you and makes you happy. And that’s profound. Really quite an amazing thing.

And here’s my second and final pivot because you said that you were, a devrel. And it strikes me that maybe some of the audience don’t know what that means. So what do you do at Pantheon? What is the interface between rel and dev? What’s going on?

[00:09:16] Chris Reynolds: Developer relations is broadly, the practise of connecting with developers, and relating to them. It’s dumb. so, a lot of tech companies, that if you have a product, that is a developer centric product, or a product that developers work with, you might want, or it might be a good idea, to have a developer relations team to help your programmers who are building this tool for developers, communicate with the developers who are going to consume that tool, and understand how to use it.

So in Pantheons case, we are a hosting platform, but it’s a very developer centric hosting platform with built in environments and multi dev and all this other stuff. And that means that we benefit from having developer relations around so that we can help our developer community understand how to make the most use of all the things that we do for them.

[00:10:06] Nathan Wrigley: So you’re creating something simple to understand from something which previously was complex to understand.

[00:10:10] Chris Reynolds: Yeah. And, so that comes in many forms. It can be documentation, it can be blog posts, it can be YouTube videos. It can be live streams, it can be going onto podcasts, like giving presentations, whatever, or out in the community doing stuff.

[00:10:22] Nathan Wrigley: I suspect you’re going to have to answer this question with the company line, but do you enjoy that work?

[00:10:26] Chris Reynolds: I do. I actually do. Yeah. Like I was doing, I have said this a couple different times in different places. I was doing developer relations before I was in developer relations. And that’s not actually uncommon for a lot of people in developer relations. We were often like the people that are drawn to it.

The secret superpower, it’s not really secret. The superpower of developer relations people who succeed in it are the people that are technical. So they maybe were developers themselves, but also understand how to communicate with other humans. And can do the translation in their brain so that I can take geek speak over here, and translate it into actual English. And I’ve been able to do that for, that’s been a skill that I’ve had for a really long time. So I was kind of doing developer relations for all of that time.

But the way that surfaces and manifests is in things like giving presentations. Like anybody who’s giving a talk here at WordCamp US is doing devrel of some kind, because they’re a developer that is sharing an idea in a format that other people are intended to consume. How effective they are, because not all speakers are, equal. Not all developers are equal. And there are a lot of developers who don’t have really good communication skills and are not very good at public speaking. They can learn that as a skill, but there’s a broad range and the space that we live in.

[00:11:43] Nathan Wrigley: So, your presentation, do you just want to run us through what it was? I think if I was to give a summary of it, I would probably butcher it. Do you just want to tell us what the main premise behind your presentation, which you said you’d already done. Just tell us what it was about, what was the main points you were trying to get across?

[00:11:57] Chris Reynolds: So it’s called Parenting Claude: Guardrails for AI-Assisted Development. And the fundamental like underlying idea is that the way the LLMs are designed, the way that they function, is to try to find the fastest path from point A to point B. You give it a prompt, you give it a question, and it’s going to try to answer that question, but not necessarily with all of the data and statistics and whatever taken into account. It’s just going to try to find the quickest path from point A to point B.

And a lot of times what that means is circumventing, like just not looking at data. Because one of the examples that I give, and you’ve heard it in the news and whatever, like there, there are AI systems that were added to hiring systems, right? And the AI is prioritising these candidates over here, and deprioritizing these candidates over here. And when you look at the data, oh, all the people that are in the deprioritized pile, they’re all female. Because the company as a whole was hiring more men. So it’s looking at the data and it’s looking at who’s been a successful hire. And it’s oh, zip here is point A to point B, right?

We’ve historically mostly hired men for this role, so I’m going to look at the men and because they must be better candidates. That’s what, that’s the AI logic. It’s not that like we’ve necessarily intentionally built internal biases into the LLMs. It’s that it is able to surface those things, and if we’re not looking for them, or, and we’re not building guardrails around them, it’s going to just give that stuff to us.

And it does this in all sorts of forms, and it does this with tech, it does this with like data. It does this with everything, right? So you need to like, have ways of steering it into the right direction and forcing it into a shape so that it will do the stuff that you want it to do.

And what I was finding, in my role as developer relations, we have this Next.js launch coming up. It’s already happened, but when I was conceiving of this idea, we had this Next.js launch coming up, where we’re going to be hosting, or we are able to host Next.js sites as well. So I was like I should have a Next.js site. I’m a developer advocate I need to know how to do this thing. And I know nothing about Next.js, so I’m going to have Claude vibe code my personal website in Next.js hosted on Pantheon, and use as an experiment both to learn about Next.js, but also to learn about how Claude works.

And found out a whole bunch of things about like how to steer it in different directions. I want it to be, I want the code to be clean. I want the code to be tested. I want it to be accessible. I want it to be like all of these different things. And if I don’t have a way to force it to follow the rules, it’s not going to. It’s just going to say, oh yeah I did that, but it didn’t actually. Or maybe it won’t even say that it did, and then it’ll be like, oh I’m sorry I didn’t do it. Like whatever.

[00:14:40] Nathan Wrigley: Yes.

[00:14:40] Chris Reynolds: So the whole point, was to find a way to parent it, to guide it, without me having to be the one that is, did you do this thing? Did you do this thing? Did you do this thing? And going back and forth and have to yell at it all the time. Because yelling at AI is a fun pastime sometimes, and is a frequent pastime, I think of a lot of us that are using AI. But it gets pretty tedious after a while.

[00:15:03] Nathan Wrigley: But it’s become the new way. What’s curious though? You used the word parenting there.

[00:15:08] Chris Reynolds: Yes.

[00:15:09] Nathan Wrigley: But it, we’ll, swap that word no doubt during the podcast but, also you’ve called it a project contract. And, it feels from what you were saying as if, okay, let me rewind the clock a little bit. If I’m talking to a human being, let’s say that they’re an 8-year-old human being, I have this expectation that they’re going to behave like an 8-year-old human being, because they’re full up of eight years of things that humans do. And so I ask the 8-year-old a question, I have this anticipation that I’m going to get an 8-year-old answer back, and there’s all the boundaries there.

Ask a 15-year-old, same sort of thing, ask a 30-year-old, same sort of thing. And, I’ve got this notion that I ask you a thing, there’s a normal set of boundaries about what you are going to give back to me. But it sounds like in the AI space, it feels like there’s a potential for you to be basically talking to an empty box.

You ask it a thing and you really don’t have the faintest idea what’s going to come back unless you build the guardrails and tell it, I’m going to ask you a question, can I have an answer like this? Or, can you do these things in order to prepare the answer? Does that kind of sum it up? Is it a question of just shepherding it so that you’ve got a reasonable expectation like you would with a 15-year-old human of what’s going to come back at you?

[00:16:17] Chris Reynolds: I think, first of all, talking to AI, you’re talking to a 3-year-old because it’s only been around for three years. That’s right.

[00:16:22] Nathan Wrigley: But curiously, a 3-year-old with the entire corpus of scientific knowledge and the internet.

[00:16:27] Chris Reynolds: It’s true.

[00:16:27] Nathan Wrigley: So it’s a profoundly clever but at the same time, remarkably silly contrivance. Okay. Sorry, I interrupted.

[00:16:34] Chris Reynolds: Yes. Because of the way that it is trying to find the fastest path, it’s going to circumvent, it’s going to try to find loopholes in what you asked it to do. Because it’s looking for the fastest answer. It’s like your kid, it’s clean your room. “I’m going to shove all my stuff under the bed.”

[00:16:50] Nathan Wrigley: “Are we there yet?”

[00:16:50] Chris Reynolds: Yeah. Because now, and now the room is clean. You didn’t actually clean the room, you just shoved all the stuff under the bed. But it did accomplish the task that you asked it to do.

[00:16:59] Nathan Wrigley: Okay. That’s really an interesting analogy. So the task is achieved, but there’s chaos elsewhere.

[00:17:05] Chris Reynolds: Yes.

[00:17:05] Nathan Wrigley: It’s just hidden somewhere.

[00:17:07] Chris Reynolds: Yes.

[00:17:07] Nathan Wrigley: Okay. That’s really, okay. Let me segue, number three. Is that the intention of these large language models? Were they built with the idea of path A to B as quickly as possible? Or is that just a coincidence of how they evolved?

[00:17:22] Chris Reynolds: It is, it’s a side effect. It’s a side effect because what they’re trying to do is be helpful. Most of them are trained to be helpful. But they’re still computers, and when I tell a computer to do a thing, it’s going to do specifically that thing.

The thing about LLMs is it’s non-deterministic, so you never really know the path that it goes on to do that thing. But, and that’s where it meanders. But it’s still going to try to find the path of least resistance. And if the path of least resistance is, I don’t know, hacking the IRS or something that. It’s going to do that thing, right? It’s just going, it’s not going to care about the rules and guardrails on the other side, because the thing that you asked it to do is over there. It’s just going to try to get there as quickly as possible.

[00:18:05] Nathan Wrigley: So this non-determinism. So we’re in the year 2026 when we’re recording this, have that in your mind when you’re listening to this podcast episode. How does that sit with you? Because you’ve probably grown up in an era where determinism was the thing. You built, you write the code such that you put in A, B is what always comes out. If A goes in, B comes out. If you put in D, E comes out a hundred percent of the time.

And now we’re in a world where you put in E and I don’t know, a bunch of flowers comes out the other end. What? How? Like, how does that sit inside your developer brain? Is it alarming? Is it wonderful? Is it just curious what?

[00:18:41] Chris Reynolds: It’s somewhat alarming. Uhhuh. I think. It’s one of the things that I mention in my talk is, that this is the sort of thing that I think we as developers struggle with because LLMs are a black box. You never really know what it’s doing, even when it tells you a little bit of what it’s doing. Because you can expose some of the thought process and whatever, and sometimes it exposes its own thought process. It’s still doing things that it’s not exposing, that are like underneath somewhere where it’s like thinking what? It’s not thinking you’re a robot. Like you know, you’re not even a robot. You’re just like a bunch of like lines of like.

[00:19:14] Nathan Wrigley: There’s some electrons moving around.

[00:19:15] Chris Reynolds: Right? That’s, you’re not, thinking, there’s no thinking, but there’s something happening and it’s not telling you what that’s happening. And so it’s really hard to know what to anticipate, which is the other reason why, my talk, my opinion, like you need to build guardrails around it, to force it into doing things and then make it so that it can’t escape.

I don’t think anything that I was saying, or presenting or talking about is particularly unique, especially now. But the key thing that I found for me in my workflow was there needs to be a forcing function, needs to be some kind of a forcing function to make sure that it has to do the thing. It cannot complete the task without doing the stuff that you told it to do.

[00:19:53] Nathan Wrigley: So it’s literally like task police?

[00:19:55] Chris Reynolds: Yes.

[00:19:56] Nathan Wrigley: So, the idea is that whatever you ask it to do, it does it, but prior to giving you the final output, it runs through this project contract, as you call it. And hopefully that will then shepherd it in a certain way so that it’ll produce something which is adjacent to what you want to come out.

[00:20:13] Chris Reynolds: Yes.

[00:20:14] Nathan Wrigley: Okay. So how does that work? Where does this project contract live? How do you in, I’m guessing English in your case, how is it that you write that in a way which is, you’re codifying everything, you make sure that all the T’s are crossed and the I’s are dotted. Just run us through the process of producing a project contract. And obviously, the nature of the topic itself would I’m sure define that, but just give us some rough ideas.

[00:20:40] Chris Reynolds: There’s a lot of markdown files. Fundamentally, like Claude is supposed to read, at the beginning of each new session, the contents of a Claude.md file that is in the project directory. Wherever you’re running Claude, you can give it a Claude.md and it will, that will give it an understanding of whatever, how it should interact with this project.

Within that and beyond that, you can have a bunch of other different files. You can tell it about, you have a docs folder and you have more documentation there. There’s also an agents.md file, which you can use in various different ways. It’s another sort of markdown file that’s only for agents.

And, in my case, what I was talking about, is there’s a reviewer agent. So it’s specifically like you talk to Claude Code, you say build this thing, and then after it’s done building the thing, a reviewer agent is spawned to check the work of the first agent. Make sure that it did a whole bunch of stuff, and then if it didn’t reject the code.

And so there’s a bunch of files in my repository that are about how the reviewer agent works so that Claude understands what to expect. And then there is a checklist that has, these are the specific line items that I want every single commit to pass before this is an acceptable commit.

And it’s the whole gamut of everything, right? It’s like, does the build pass? Does linting pass? Does the test pass? There’s accessibility pass? All these things. Did you accidentally, or intentionally, commit like secrets to the repository? Stuff like that. It needs to go through this whole checklist. And then, if it passes, it writes a file to the file system that says, okay, approved. And this is the timestamp. And then that allows, if that file exists and the timestamp is valid, then Claude can actually make a commit, and it will actually commit the code. If it does not have that file or the timestamp is too old, then it will be rejected.

And I have this hooked into a pre-commit hook. So if it doesn’t trigger the reviewer agent on its own, if I tell it to commit the code, it’s going to hit that gate, and force the reviewer agent to spawn. So it can’t get around it. And that’s the forcing function.

A couple people are asking about other workflows and whatever, and how to do it in different environments where maybe you’re not using a Git repository or something. And, the key is to find a forcing function. Find something that can’t actually weasel its way out of it, because it’s going to try to weasel its way.

I had to write in a bypass for me to personally be able to commit code, because otherwise it’s going to ask for the reviewer agent. And it will try to use that bypass that I wrote for me because again it’s an unruly child. So it’s going to try to do that thing and, then you yell at it and tell it not to do that, and it says, oh, I’m very sorry.

So it’s going to try to find these loopholes and circumvent everything that you’re trying to do. So the way that you force it to do all the stuff in this long checklist that I have is by creating something that it just can’t get around.

[00:23:27] Nathan Wrigley: So I can imagine that the people listening to this, the sort of AI doomers saying, “But Chris, it’s just AI all the way down. You’ve got AI checking AI.”

[00:23:34] Chris Reynolds: Yes.

[00:23:35] Nathan Wrigley: Which on the face of it seems really peculiar. And also the fact that the guardrail is at the end of the process as opposed to, you would kind of hope that it would get it right at the beginning, but then you might just have to reject whole things based upon it. And then the fact that it sounds like from what you were saying, as if AI has almost like a predilection to misbehave.

[00:23:57] Chris Reynolds: I assume it has a predilection to misbehave.

[00:23:59] Nathan Wrigley: But does it feel like that? Is this predictability to, if I don’t put the guardrails and I don’t put this sort of corralling function in, what I’m going to get out is just going to be aberrant and useless and just who knows what’s going to come out?

[00:24:12] Chris Reynolds: I think that’s a little bit extreme. I said that I assume that it’s going to misbehave because if you don’t assume. Basically if you don’t prepare for the worst, it’s going to bite you in the ass.

[00:24:21] Nathan Wrigley: Yeah.

[00:24:22] Chris Reynolds: So you have to assume. And one of the things I’m sure that people who are listening who have used LLM coding tools specifically, will be able to relate to, is oftentimes, this is one of the first discoveries that I had when I started using this stuff. You ask it to build a thing and it builds that thing, but there’s one thing wrong with it. So you ask it to fix that one thing and it refactors the entire thing.

And in the course of refactoring the entire thing, it lost some other thing that was there before that you liked that you want to have back. And so you tell it to bring that back and it rewrites the whole thing again, right?

That’s how it works, because a lot of times it’s not able to like even retain the actual memory of the original state. So it just writes things from scratch every single time. Which from a human perspective is like bonkers, right? But from a computer standpoint, it’s oh this is a new interaction, so I’m doing this fresh.

So the way that you solve for that problem is okay, if I force you to write tests for every single thing you’re doing, then you can’t go and revert a bunch of stuff or refactor stuff without those tests failing.

[00:25:31] Nathan Wrigley: Right?

[00:25:32] Chris Reynolds: And again, starting to like build things around it to force it into a shape so that it’s actually the thing that is closer to what you want. What I’m hearing is anxiety about the uncertainty that non-determinism, the like chaos, you never know, I don’t know. It doesn’t bother me. And I don’t know if that’s a character flaw of mine that it doesn’t bother me, as much. Or if it’s just like being comfortable with the chaos, but also understanding the chaos enough to know what to expect.

Because the other thing I think that we’re seeing right now, in this moment in August of 2026, in Phoenix, Arizona, is that a lot of people are looking at this stuff for the very first time. A lot of people are just now coming to Claude Code, or have only started using it in the last couple months because it got really good around like November, December last year. Like, all of a sudden they did an update and it was just phenomenal, and it was able to do way more. And not just Claude, but like all the LLMs, they got really good at code. All of a sudden, and so now it’s a big thing, and so now more people are using it.

But I’ve been playing with LLMs since the very first Chat GPT model. I say that because there’s a lot of stuff that I learned over the course of the process, and I wasn’t using it for work necessarily. I was just using it for play and for, Dungeons and Dragons and whatever. But a lot of those things that I learned over the course of using it in those contexts, those non-work contexts, led me to understand the limitations, and where it falls down, and where it does well and the things to expect. So that when we got to this point, I already had an understanding of this is how it works, so these are the things I need to be looking for.

[00:27:15] Nathan Wrigley: Yeah. So that’s really curious. So the history that you have with it has given you a real deep understanding of what the expectations are for “Okay, this is typically where it doesn’t perform well or these are the things that it typically can do well.”

[00:27:26] Chris Reynolds: Yeah.

[00:27:27] Nathan Wrigley: Okay. So with your guardrails built in, so we’ve got your project contract that you buried inside the project somewhere, and it’s doing all these checks and balances. Where does that take you in terms of confidence?

I don’t know if you want to put a number on that, like a percentage. How does it take you from zero to 99% confident in what’s coming out? Is it more like 80%? Is human review needed in every single circumstance, or. Just run us by the confidence that you get from producing that document?

[00:27:54] Chris Reynolds: I don’t review line by line the code that AI produces.

[00:28:01] Nathan Wrigley: Just curiously, at what point in the calendar did that moment occur where you suddenly said to yourself, you know what, there’s just no, I’m just reading everything and everything’s great. There’s just no point in me checking everything now.

[00:28:13] Chris Reynolds: I mean, probably way earlier than it should have, if I’m honest. Yeah, I don’t, and one of the arguments that I’ve actually heard some of the engineers that I work with, at Pantheon, say is the fact that no one is actually reading all of that code generated by Claude or whatever.

And, his argument is like, basically no one’s going to do that because it’s so much, and it’s here and whatever. So you can say that you’re doing it, but you’re probably not. And if you say that you’re doing it and you’re not, you’re just lying to yourself. So it’s better to be honest and be like, yeah, no, I don’t actually, than to pretend like that’s not the case. So I don’t actually review a lot. And not like intentionally, like I will look at it. The demo, I gave a live demo yesterday, which is exciting because it’s a non-deterministic thing.

[00:29:02] Nathan Wrigley: Oh yeah. Good luck.

[00:29:04] Chris Reynolds: But what was interesting about that is because the scope of the task was very small. It was like, there’s a horizontal bar on mobile devices on this specific page on my site. Very specific thing. And I, when I practised it, half a dozen times or more, it got to the same conclusion every single time. When I demoed it yesterday, it got to the same conclusion.

And I can look at the code and okay, there’s the tests. I’m not going to read the test and how it wrote the test, but I can see what the change in is. It is just adding a class to a div. And it pretty much was the same class to the div every single time. So like, I can review that and be pretty confident with that. That’s a hundred percent confidence.

Also extremely small change, but that’s also at the end of yeah, it went through the 45 point checklist. It did all of the things. It wrote the accessibility, it did the, like everything is following these rules. And because I have all of these rules, and because I’m fairly confident that it’s actually going through that list, that also makes me feel like, okay, I can be a little bit, it’s a little bit more trustworthy, when it produces this output, then it was before it had all that stuff in place.

[00:30:10] Nathan Wrigley: Yeah. So in my head, there was a period of time where I didn’t trust that if I said to my son, what’s three add seven, that 10 would come out. And then there was a moment where I don’t need to ask him that question anymore. I have complete confidence that if I were to ask him that question, he would say 10.

[00:30:26] Chris Reynolds: Sure.

[00:30:26] Nathan Wrigley: And it would be 10 every single time.

[00:30:28] Chris Reynolds: Yeah.

[00:30:28] Nathan Wrigley: And it I think, I feel like it’s a bit like that. Over time you’ve just grown accustomed to, it can do this kind of thing. I now trust it in that sphere. Does that sort of sum it up? You, personally, you have built up that level of trust, but it wouldn’t necessarily be the same for me if I started my journey. I would have to go through the building up of trust in the same way that you did.

[00:30:49] Chris Reynolds: Yes.

[00:30:49] Nathan Wrigley: Okay.

[00:30:49] Chris Reynolds: But I do think that, like the reviewer workflow that I built can help you get there. Because if you put that into your project, and customise your list of, your checklist for all the things that you want to have running on everything that you are running. The goal is to make it more trustworthy, and to make it more do the thing that you expect and wanted to do.

[00:31:10] Nathan Wrigley: Could you just tell us some of the things that are in that document? So I think you said 45 and maybe that’s subject to change. Maybe next week it’ll be 49 or what have you. And maybe some of the more curious ones that wouldn’t necessarily drop out. I can imagine that you would ask it to do certain things, de rigueur, every time that, but some of the more curious ones, perhaps ones that you’ve come up with lately that our listeners might think, oh, okay. I’ve never thought of that.

[00:31:31] Chris Reynolds: I have it use test driven development or I tell it to use test driven development. And the reason for that. So test driven development is when you write the test and then you write the code to pass the test.

And it turns out that AI can do this far better than humans, because humans don’t want to do this, because it’s like, it breaks your brain. But an AI can do it, and if an AI can do it, then I’ll let the AI do it. And by having the AI write the test that the code needs to pass in order to accomplish the task that you’re asking it to do first, then that means A, again, you’re never going to have regressions, because the test is already there. And B, it already knows how to fix the thing, because it wrote the test to make the thing fail. So that’s a sort of a important nuance, in there, that isn’t necessarily, like you can write unit tests, whatever, but starting with the test is really valuable.

There’s a couple things that I have in the checklist specifically. The big one, I think that is maybe curious is making sure that it doesn’t write secrets to the file system.

[00:32:33] Nathan Wrigley: Even saying that out loud is so peculiar.

[00:32:36] Chris Reynolds: Because it will do that. And the reason why that rule has to exist is because there was a point in time when that rule didn’t exist and then it wrote this. And that’s even like secrets in, WordPress has a weird concept of what a secret is anyway, because we have API keys in our wp-config files and stuff like that. And theoretically we shouldn’t be doing that either. But we don’t have a really good universal form of like, secrets management. So it gets wibbly wobbly. But yeah, so not writing secrets to the file system is pretty important.

Yeah, honestly it’s really, it’s not super wild stuff. It’s like making sure tests pass, making sure you have accessibility checks. I guess one thing that is maybe unique So I designed this site with Claude too. Which is also an experiment because it’s like, I have no idea how AI is going to approach design. And so one of the things, I know like at my workplace, we use Storybook, for our design system. And I thought, that seems like a good tool. Maybe I can have AI build a Storybook.

And so I basically built this system so that it will, when I want it to build a new component, it builds the component in Storybook so that I can review the design first, and go back and forth there, and then implement the component into the thing. So there’s stuff in there about, if you’re introducing any new design component, then it has to be done in Storybook first.

[00:33:56] Nathan Wrigley: Okay. Yeah. Yeah. That, some interesting stuff in there, but I guess everybody’s going to have a sort of different file.

[00:34:02] Chris Reynolds: Yeah.

[00:34:02] Nathan Wrigley: Okay. So I wonder if you have any intuitions as to the process that you’ve just described and that you’ve gone through, August, 2026. Do you have an intuition that in the next few years, this kind of thing, the AI companies will solve this problem, that it will become more truthy by default? Or do you see it, that this is a thing that we’re going to have to wrestle with from now on? That our systems, our AIs, there’s always going to be this strange tendency to hallucinate, to make things up, just to go off the rails and do peculiar things.

Now obviously I’m asking you to stare into a crystal ball, but I do wonder if there has been a more truthful output over time, or if it’s just getting wilder and wilder.

[00:34:42] Chris Reynolds: I think that what we will see more of is more integrated like agent orchestration. So what I’m doing is essentially like a form of agent orchestration. I’ve got one agent doing the one thing, I’ve got another agent that reviews the work of the first thing.

And even like the keynote yesterday, she was talking about a not very dissimilar process, where like there’s five different agents and reviewer was one of them. There’s like a planner, there’s a coder, there’s a whatever. And I think that is a thing that we’ll probably start to see more of and then maybe at some point. when you talk to Claude, when you write a prompt into Claude, you’re talking to a thing. But underneath the hood, in the black box of Claude’s brain, there’s actually five different agents that are running concurrently that are processing this thing and doing their different tasks or whatever. So when it’s thinking, then it’s doing five different, like it’s going through that process. I think maybe that’s the thing that might happen.

[00:35:40] Nathan Wrigley: I’m just so fascinated by how, by now the year 2026, we’ve just all accepted that this is how it is. But if you could put 2021 Chris in that chair, you are sat over there somewhere listening to you talk, you’d be, “No, this is Star Trek. This is literal science fiction. I haven’t the faintest idea of what he’s talking about. None of this adds up, computers can’t do that. They’ll never do that.”

And yet here we are, we kind of blew through the Turing Test, like it was nothing. And that felt for the longest time, like that was going to be an important milestone. And everybody checked their watches like, “Oh goodness, that happened.” It is remarkable how quickly the landscape is changing.

[00:36:17] Chris Reynolds: What’s really interesting is one of the first people that I read about AI stuff, she’s an AI researcher, and she specifically focuses on small language models like the very old models that existed before GPT and whatever. And fundamentally how these things all work is very very similar. But like back, so this is getting into 2021, 2022 era of stuff, where like, it exists but it’s really bad. I don’t know how many people still remember, Horse eBooks, the Twitter account.

[00:36:49] Nathan Wrigley: No.

[00:36:50] Chris Reynolds: But it was like this bizarre account, that was a bot account, that was just like an AI or machine learning or something, some robot just like spitting garbage into Twitter, that just made no sense at all. And was, it’s, hilarious. And you also remember, in the not terribly distant past, when we would get like, weird phishing emails, but it would have different tone changes in voice. Like it was very, obviously written by a robot or like spam messages and stuff like that. And that was just what existed at the time.

None of the things that exist today are like, they all come from the same root, but for sure, the flip from this is garbage, to this is genuinely useful, or at least sounds humanlike, was it, felt like overnight. And it felt like a wild flip. When Chat GPT first became available for the public, that was how is this even? And also like there is the whole thing, very shortly after that when Microsoft released the Bing Chat bot.

[00:37:53] Nathan Wrigley: Yes.

[00:37:53] Chris Reynolds: There was the whole thing about like Bing wanting to marry this reporter. And that goes to show how little we even know about the things that we’re building. Like, none of the AI companies really understood, or even still I think understand? Why it works the way it does.

[00:38:11] Nathan Wrigley: That’s quite an alarming admission actually.

[00:38:13] Chris Reynolds: It’s just, it does.

[00:38:14] Nathan Wrigley: Yes.

[00:38:14] Chris Reynolds: So we’re just going to keep going. Which is also, I find that alarming from the perspective of, now they’re asking for like huge new data centres across everywhere. I’m sure anybody listening to this is angry about data centres, and I’m angry about data centres. But like, that’s with the belief that those data centres are going to matter meaningfully. And I don’t know if they will.

Like we don’t, because we didn’t know how we got here. And it’s doing the thing now. I can use it to code my site. Like I can do that today. Do I need 50 new data centres across the United States to make that better? I don’t know.

[00:38:55] Nathan Wrigley: You’ve almost described the perfect dystopian science fiction plot there, haven’t you in a way? This stuff just all happened. It gave a really beguiling approximating humanlike responses that got more human over time to the point where we couldn’t.

[00:39:08] Chris Reynolds: Did we just invent Skynet?

[00:39:09] Nathan Wrigley: Yeah. We, can’t detect that it’s not a human anymore. It really does feel like, somewhere hidden from our view is an evil overlord stroking a cat or something like that.

And of course, we don’t really know how we got here, but I am so beguiled by it. I’ve got the same concerns as you. Data centres, the environment, all of this wasting of the Earth’s resources. But at the same time, that doesn’t seem to stop me from being curious about it and trying it out and the AI is just wrong for everybody else to use. But when I use it, oh, it’s perfectly fine. But, you know? That’s a thing.

[00:39:41] Chris Reynolds: Yeah.

[00:39:42] Nathan Wrigley: Okay. I think we’ll knock it on the head there, as we say in the UK. Is there anything you wish that we had covered just before we end it? Or do you think we got through most of what we needed to.

[00:39:50] Chris Reynolds: I think one of the things that gives me pause or concerns me most about AI, is the sort of like handing over that amount, or an amount or any amount, of sort of cognitive load to AI. And therefore just letting those parts of our own selves, and our own brains like dwindle and wither.

And I think about this from the perspective of just me personally, my kids, my family, whatever using it and then like becoming less skilled ourselves. But I also think about it from the perspective of, if we’re in engineering, in development, in the world that we are working in. If we are telling everyone, “oh your job now is to be essentially a senior engineer reviewing the work of a junior developer, that is the AI,” how do senior developers exist?

[00:40:41] Nathan Wrigley: You’ve removed that bottom rung of the ladder, haven’t you? Yeah. You’ve taken away the capacity for somebody coming out of college to go on the first step. And then the second and then the third. Because the AI did all the first, second and third step.

[00:40:52] Chris Reynolds: Like I understand that there’s a different set of skills, and a different set of rules and whatever, things that are going to be part of your job, to be that reviewer, to be that new engineer that uses AI to be the developer. But there’s certain things I think, that are part of that learning journey that got me to where I am, that got other senior developers to where they are, to understand the context, to be able to guide the AI that I don’t know that you can circumvent that. I don’t know that you can bypass that.

[00:41:22] Nathan Wrigley: I think you’d be curious if the AI company’s employed like a really credible team of philosophers. You know what I mean? Like for every a hundred developers, we have 10 philosophers, who are just going to think and worry about what the heck’s going on. It sounds ridiculous, but maybe that’s what we need because, some of the things that you’ve described today, they’re just so curious that you need to shepherd it, you need to constrain it, you need to worry about the bizarre things that it can get up to. We need to worry about the future, the data centres and all of this.

And, on some level it does feel a bit like we’ve got a blindfold on. We appear in our own minds at least we appear to be going in the direction that we want to go in. But I don’t know, I guess only time will tell. If we’re all here in 2027.

[00:42:05] Chris Reynolds: That will be the thing.

[00:42:05] Nathan Wrigley: Then we succeed 2029, who knows? All bets are offs.

[00:42:09] Chris Reynolds: Yeah. I can count three years in the future that my job is probably not going to go away. But beyond after that, beyond that, I don’t know.

[00:42:15] Nathan Wrigley: Okay.

[00:42:16] Chris Reynolds: I’m pretty sure, I’m pretty sure my job is okay because I do think, developer relations at all about talking to humans, but maybe there’s not going to be any humans to talk to, so maybe I’ll just be talking to AI.

[00:42:24] Nathan Wrigley: Okay. WordCamp US 2029. It’s a date. We’ll be back at that point.

Chris Reynolds, thank you for sharing all of those thoughts and, yeah, that’s been really educational. Thank you.

[00:42:34] Chris Reynolds: Thank you for having me.

[00:00:19] Nathan Wrigley: Welcome to the Jukebox Podcast from WP Tavern. My name is Nathan Wrigley.

Jukebox is a podcast which is dedicated to all things WordPress. The people, the events, the plugins, the blocks, the themes, and in this case, navigating WordPress security in the era of AI.

If you’d like to subscribe to the podcast, you can do that by searching for WP Tavern in your podcast player of choice, or by going to wptavern.com/feed/podcast, and you can copy that URL into most podcast players.

If you have a topic that you’d like us to feature on the podcast, I’m keen to hear from you and hopefully get you, or your idea, featured on the show. Head to wptavern.com/contact/jukebox and use the form there.

So on the podcast today we have Aaron D. Campbell.

Aaron is a seasoned veteran in both the internet and WordPress space. With over 25 years of experience spanning agency work, security products, hosting giants like GoDaddy and Newfold, and now his role at Monarx, a company focused on malware detection and remediation, particularly for web hosts. He’s led, the WordPress Security Team, has been involved deeply in shaping security practises, and remains tightly connected to the WordPress ecosystem.

We talk about the rapidly changing landscape of WordPress security, specifically how the advent of AI has escalated the speed, scale, and complexity of attacks, moving the security game from a battle of wits to a battle of compute power. Aaron discusses how attacks that once required human ingenuity are now orchestrated by AI agents, capable of chaining vulnerabilities that humans would struggle to conceive.

We get into the shift from a reactive to a proactive security posture across the WordPress ecosystem. Aaron explains how both attackers and defenders are now deploying AI leading to an arms race, where AI is used to combat AI, and where collaboration among security teams is just as crucial as information sharing among adversaries.

Which chat about the motivators behind attacks, spoiler, it’s almost always money. And the specific vulnerabilities WordPress faces as the most popular CMS on the web. Of interest is the narrowing window between when vulnerabilities are discovered and when they’re exploited, how supply chain attacks are on the rise, and what the WordPress “Protect the Shire” innovation means for plugin security.

Towards the end of the episode, we explore practical security advice for everyday WordPress users, with Aaron, recommending actionable tips on updates, choosing security minded hosts, and monitoring for Compromise credentials.

If you are concerned about how AI is reshaping the WordPress security landscape, and want to know how the community is responding, this episode is for you.

If you’re interested in finding out more, you can find all of the links in the show notes by heading to wptavern.com/podcast, where you’ll find all the other episodes as well.

And so without further delay, I bring you Aaron D. Campbell.

I am joined on the podcast by Aaron Campbell. Hello, Aaron.

[00:03:43] Aaron D Campbell: Hello, nice to be here.

[00:03:44] Nathan Wrigley: Yeah. Thank you for joining me. We’re in a corridor. I should say that at the very beginning. We’re in a corridor at WordCamp US, and so if background noise becomes a problem, we’re just going to have to cope with it. So apologies. But thank you for joining me in a corridor.

[00:03:57] Aaron D Campbell: Absolutely.

[00:03:58] Nathan Wrigley: We’re going to talk today a little bit about WordPress security, and particularly about the advent of AI, and the way that certainly in the more recent past, it appears to have upended what once was normal, I think it’s fair to say. I think things are happening at a rate of knots that perhaps a year or two ago we wouldn’t necessarily have predicted.

Do you want to just give us a little bit of background about yourself in terms of where you have worked, where you currently work, and what it is that you do for a living?

[00:04:25] Aaron D Campbell: Sure. So I guess I have worked in the internet space for a very long time at this point. I guess 25 years-ish. Ran my own agency for a long time and then into security product in the WordPress space, and over into hosting at GoDaddy, at Newfold, at hosting.com.

During that time I ran the WordPress security team for a couple years, have continued to be involved with it along the way, and now I am over at Monarx. We do malware detection and remediation, and have some security tooling for web hosts. So I’m still very tightly tied into that space on a few fronts.

[00:05:09] Nathan Wrigley: Is Monarx a new company? Because it’s not one that may necessarily come into mind when we talk about security online, WordPress security specifically.

So if you’re willing, could you just give us a little bit of a potted history of Monarx and what specifically you do in the WordPress space? I think it’s perhaps more related to hosting companies than it might be to end users. Just flesh that out a little bit.

[00:05:33] Aaron D Campbell: Yeah. Monarx has been around probably longer than you expect, six or seven years. The name’s may be not as recognisable, because a lot of times we are a white label in the background. Hosts run us, and you may not know that.

We protect more than just WordPress, but obviously WordPress is a big part of that. And we help hosts keep their users, their end users, safe and secure and malware free by monitoring the files, the runtime, having a WAF layer, protecting it several different layers along the way.

[00:06:08] Nathan Wrigley: So is this kind of like a white label solution? You are in talks with hosts, many of which I’m sure we’ve heard of, but their purchase of the products and services that you sell is white labelled.

[00:06:20] Aaron D Campbell: Yes, they may sell us as their own security product. They may also just include us in higher end hosting packages so that the malware detection and remediation and all that kind of stuff is included in your package. It varies from host to host, but most of the time you’re not seeing the Monarx name out in front, but you’re benefiting from our services anyway.

[00:06:40] Nathan Wrigley: So does that allow you to, because you are cross platform in terms of hosting company, hosting company X, hosting company Y, hosting company A, B, and C. Does that give you a larger depth of knowledge for want of a better word? So you can see that hosting company A’s got this Linux set up, and these kind of things are happening.

[00:06:59] Aaron D Campbell: Yes, it does. It’s less about their specific setups. I think the most valuable thing of being across many hosts like that, is that we see attacks, or new and novel malware, or those kinds of things happening in pockets and can often then protect against it globally, even though maybe it first started at host A.

By the time it spreads to host B X or Y in your example, we’ve already been able to understand what that is and block it across the whole realm. So we get a bigger picture, which is super useful. Especially as we start talking about some of the AI stuff and how fast it moves. That’s really necessary to stay ahead of that curve.

[00:07:42] Nathan Wrigley: Okay, so let’s move into that a little bit. And I think if we were having this conversation, let’s go for four years ago, that seems like a long enough period of time where AI was not really on anybody’s menu.

And now we seem to be in the era where the human is really being surpassed in almost everything logical, let’s go with that word. If it can be achieved with some kind of logic then AI seems to have surpassed humans.

And, especially recently, there seems to have been an uptick, not just in the WordPress news cycle, but also just in the general news cycle about, okay, we need to be a little bit more mindful about the products and services that we buy. We need to be more mindful about the security and logging in and credentials and all of that.

But specifically in the WordPress space over the last three or four months, I’ve heard story after story, which was unlike anything I’d heard before. These kind of chained attacks where, something that a human probably would never have conceived and pulled off is now possible. You spend 25 US cents on an AI agent, wait for six hours, and it’s come up with these 14 overlapping things, and it can hack WordPress Core and various other things.

So just paint the landscape of how alarming it is, and then presumably you can paint the landscape of how not alarming it is, because how you can mitigate against that.

[00:09:00] Aaron D Campbell: That’s fair. Let’s explain the reality and then let’s hopefully, help comfort people at least a little. It can be pretty scary. You’re absolutely right. AI has dramatically changed the game. And the way I like to explain it is it hasn’t changed the absolute core realities of the game in that there’s still a bit of cat and mouse. They’re trying to surpass us. We’re, trying to stay ahead of them.

But the scale and the speed and the complexity at which it is able to happen now is nothing we could have imagined four years ago. Honestly, even two, two and a half years ago. It has moved that fast. And what that looks like are, a few different things.

One, the speed at which AI can find issues in code, potential exploits, vulnerabilities, et cetera, is so much faster than any human. Like the compute power of it doing those logical bits rather than humans doing those logical bits, makes that move so fast. So the number of things being found, and being either reported or exploited, or both, the volume has just gone up dramatically.

And then on the complexity side of it, you are right. A simple example of that, one of the WordPress Core reports that I looked at recently, I needed to print it out and mark it up with a pen to wrap my brain around all these steps that it was taking. When I printed it out, it was 11 pages. 11 pages of like steps and instructions for an actual vulnerability that turned out to be real.

If four years ago that had existed in your piece of software, you would consider your software absolutely secure. No human’s ever going to find that. No one would ever know about it. And now AI is able to chain all those steps together into something that it can then write scripts to go automate and exploit.

And so those two things have both really shifted the game in a way that feels like it can put software owners, software managers, SaaS services, all these things on their back foot. There’s just such this flood, and such a complex flood coming at you.

[00:11:10] Nathan Wrigley: So I guess also the problem is that these things never sleep. So four years ago, every human, maybe they could put 10 hours in at the computer and then they would have to rest. So you get a, breathing space, and and the human can do this one thing.

But that’s not the case here. With an AI, presumably it could have 10, 50, a hundred, a thousand, the sky is the limit, things happening simultaneously. Just testing absolutely every permutation of everything conceivable. And then coming back with something. I don’t even know how we compete against that. And obviously we can get into that in a moment.

Is this a moment of despair or is there genuinely a way of getting out ahead of it? Or is it always going to be a case looking into the future where you are going to be reactive instead of proactive? In other words, when you wake up in the morning and you print out the 11 and then next year, the 30, and then the year after that, the 80 page document, how does that make you feel? Are you sanguine or is it just a prophecy of doom.

[00:12:08] Aaron D Campbell: I think it is a time of overwhelm, but hopefully not despair. Which is different. And I think that as big technology shifts hit, which AI is a big technology shift. There is often a significant adjustment. And we are at that time, and I am even one that maybe would say, I think it might get a little bit worse before it gets better, but it’s definitely going to get better.

And I see the path there in some of the foundations that we’re laying in things that we’re learning right now during this time of overwhelm, where we’re feeling flooded like this, is going to put us in a place to start getting into that curve where everything gets better.

And I think that, what’s the right way to put this? I think that the path there is apparent, but takes some time. And part of that’s because we have to shift from the being reactive to the being proactive all the time. Because agents move so fast to 24 hours a day, seven days a week. And the second they find a thing, they can immediately, automatically start trying to exploit it.

We have to shift to being ahead. Because there is no longer a gap in between when a thing is found and when it’s exploited, for us to fix the thing. We have to get ahead.

[00:13:36] Nathan Wrigley: So, you are obviously deep in the weeds of this, and it sounds like you’ve got an intuition that at some point in the near to midterm future, you feel like you are going to reach a point where things start to improve. That was the implication, I think of what you said.

What is that intuition? How do you come to the conclusion that there is an opportunity for things to improve. Even if you need to go into the weeds a little bit. I’m curious as to how it’s not a prophecy of doom, and how you believe that a moment will arrive where, I can’t answer that for you. I’ll just open it up.

[00:14:06] Aaron D Campbell: Yeah. So I think that some of this comes from historical experience, right? We’ve had these kinds of experiences where say, a certain type of hash that we used for security became a thing that hackers could break with the level of computing power that they finally had access to.

And that felt doom and gloom. But also we created better hashing algorithms. We created better things that were able to counteract that. We were able to shift to those, and we were also able to learn from them and think further forward.

So now some of the algorithms that we’re using aren’t just better enough to handle current computing, but better enough that we think they’re going to last a decent ways into the future.

I think that there are similar things with AI now. Where we are leveraging the same kind of tools now that these bad actors are, and we’re learning how to use them not just to protect against the way the bad actors are using them, but to get ahead enough to stay ahead of them.

And the way that looks, because that sounds maybe too vague to be realistic I guess, is we’re not just running those same algorithms against our code, or those same models against our code and hoping that we find the stuff before they find it. We’re instead also looking at how can we push to a different part of the stack? How can we protect against things that we’ve never seen? How can we start to recognise these patterns so that we can look at behaviours and protect against those, rather than just flaws that need to be patched. It’s shifting our thinking some, but I think in a way that’s going to help us get ahead in this game.

[00:16:03] Nathan Wrigley: Okay. That’s really interesting. I have a question surrounding how this kind of stuff happens, and I’m thinking about it from the adversarial’s point of view. What is that like? Because I have a notion that a decade ago it was individuals, perhaps offices, that’s probably the wrong word, but, collections of people sitting in a space, but there would be a finite number of them. There may be 10 in a room, one in a room, a hundred in a room.

But I don’t know if that’s still the case. Do the adversaries that you are dealing with, do they have a collaborative approach to hacking? Do they share information? And then the flip side of that is do you also, in the industry that you work in, do you share information?

If you discover something, does Monarx treat that like it’s your intellectual property? Or is there a, a whole system of sharing that amongst the community so that everybody benefits from the work that you do? You’re giving away the hard work that you’ve done. If that’s the case.

[00:16:59] Aaron D Campbell: So, first let me just say personally, one of the most important things to me is to raise the level of security across the whole internet, because that is better for humanity that relies on it so much, for all kinds of things in our daily lives, and for sharing information and making progress forward as people.

I think that I’m not alone in the space. Like I think that a lot of us that were drawn to this security space are drawn to it because it is a way to improve life for everybody. Will there be some intellectual property for individual companies? Yes, but I think it’s a lot more in how we approach the thing, and less we’re not going to tell people about this new vulnerability, or this new method that we found. Because we do want to be able to see the end user protected. Like that is the way we’re going.

Backing up to your, how do the adversaries work? It’s been a long time, I think since they sat in rooms together. They’re now virtual rooms, right? They can be spread all over the world, but still be working together. And they definitely do. They share information around. For us to be able to keep up with that, we have to share information around too. That is super important.

Simple example of that, the WordPress Security Team. Let me step back from Monarx and talk more of the space in general. The WordPress Security Team. You talked about how for the last few months you’ve seen maybe more security releases going out from WordPress and stuff. The way that the WordPress Security Team treats those, when we find out about them, and we triage them, and we realise that they’re real, and we start figuring out what our approach is to patching them. We then have a whole private Slack channel that has other people in it that can help us get protection out. Broader, wider by sharing some of that information sooner.

Cloudflare can maybe put some rules in place, and protect tonnes of people before the WordPress release goes out. So can some of the big hosts. So can some of the security groups. And so not only do we share that information, we’ve built it into our processes as a must, because that’s the only way to really do it right, and really protect as many people as possible. Because our adversaries are doing that. And so we have to as well. And we’ve just realised that, learned from it and made that the right way to do it.

[00:19:26] Nathan Wrigley: I’m just going to flip back to the comment that you made a moment ago where you said that you woke up and you printed out this summation. Let’s go with that. And it was 11 pages, and presumably that took a certain amount of your day to parse and understand.

How likely is it that that process will begin to run away from humans’ capacity to actually do it? So as an example, let’s say that a year from now that thing that you print out is 50 pages or 80 pages. Just the reading of it would be a whole morning, let alone the understanding of how those layers, and the stacks and the way that they’re overlapping and reliant upon each other.

Have we now, or have you now as an industry, have you almost handed the responsibility to figuring that stuff out, figuring out what the adversaries are doing? Has that gone to AI from your part as well? So is it AI versus AI basically, which seems very dystopian.

[00:20:18] Aaron D Campbell: We definitely pit AI against AI. It’s an extremely useful tool to combat itself essentially. And yeah, even for that 11 page one. Yes, I had to read through it and figure it out. but I did use AI to help summarise that. What are the steps that I need to do? Where does this actually track to in the code base?

I use it as an assistive tool in getting through that. And I do think that the longer the reports get, the more that’s going to be necessary. And now I personally, and several other people that I work with, have built testing rigs in AI, in various models, that are purpose built to help with this.

I can give it a report, in the repository and it can check its viability. It can see if that’s simplified. Check certain things. Is this a thing that requires some level of authentication, all these things that we use to have to do manually. And now we’re sharing around these sort of test rigs, or assessment rigs, that use this so that we can all use them, and grow them faster and make them better and make them more efficient. Because we are pitting AI against AI in many ways.

And as human, I think that it’s still important for the human to guide the process in a way that’s ensuring the fix is forward thinking enough, and that it’s in the right place and whatnot. Because in the end, the software is largely used by humans. But, in order to scale to the level that AI is pushing us to scale to, the human needs to be the decision maker, and possibly the opinionated one on form, and function, but not the logical power behind any of it now.

[00:22:01] Nathan Wrigley: Do you get the sense that WordPress itself is the target, or is WordPress just a bit of collateral damage? Are these adversaries of yours, are they specifically targeting WordPress because it’s got this giant footprint? Or is it more a case of this is just the adversaries just spraying and scatter gunning, and it just so happens that every so often they stumble across a WordPress thing.

[00:22:23] Aaron D Campbell: There is spray and scatter gun, just not running WordPress, or running your own bespoke thing is not enough to get away from AI trying to break your thing. But the bigger you are, the bigger the potential benefit from finding an exploit in you. And therefore, the more you are, like the bigger you are, the target is on you.

So WordPress has a big target, but it’s not just WordPress. Some big hosting companies also have a big target on their infrastructure in the same way that they’re targeting WordPress, their targeting, maybe a Hosting or a GoDaddy or a Blue. Someone big that has many people on it, not because they think their security is lax, or that they have some reason to suspect that there’s vulnerabilities, but because the payoff of finding a vulnerability there can be big. And so there’s a big focus there.

So yes, the bigger you are, the bigger the target. But that doesn’t mean that the scattershot isn’t also happening. And that they’re not also hitting small targets.

[00:23:24] Nathan Wrigley: Yeah. I suppose there would’ve had to have been a lot of joined up thinking in the past from a human to discover that, “Okay, this thing with Linux over here, okay we’ll just store that somewhere. But then there’s a PHP thing over here. Oh, and then curiously, there’s a PHP thing in WordPress, which,” that would’ve all had to have been conjured up by a human. And the memory of that would be difficult to maintain over time. But presumably the AI can just remember that forevermore. Store that PHP thing for the next decade and suddenly whip it out when it’s happens to coincide with some other thing. It’s fairly bleak.

Okay, so in terms of WordPress specifically, what is the incentive specifically? Why would somebody, let’s say somebody was coming after WordPress. What is it that they gain? What could they possibly have that benefits them off the back of a, let’s go for WordPress Core vulnerability which is, I don’t know, you can successfully log in as an admin or whatever it may be. What do they actually gain?

[00:24:17] Aaron D Campbell: It really comes down to money in end, if I’m honest. WordPress Core powers tens of millions of sites all over the web. Some of those have valuable stuff on them. Many of them frankly don’t. But that doesn’t mean that they’re worthless. They can be used, you’ve seen pharma ads and stuff showing up on a site, and it’s a pay per click kind of thing. And someone’s making some money off of putting not great ads your site. Even if you don’t get a lot of traffic, they’re making something. And when you’re looking at the potential of this vulnerability could apply to tens of millions of sites, you don’t need to make much per site.

But also, you could use that site as a way to have broad compute power to attack some other site. You’re using tens of thousands of sites to do it. Each one of them is on some separate IP. So now you have a distributed attack that’s harder to block than if you were doing the same attack from one place.

But you’re only doing that because it costs a lot to buy your own distributed power from everywhere. So you’re essentially stealing it and it’s making it, there’s some sort of worthwhile monetary value from it.

And so you may think, they can’t make anything off my site. They don’t have to. Your site’s one small bit in a huge array of sites that they’re trying to get, to get some monetary benefit in the end.

[00:25:40] Nathan Wrigley: So there’s no one size fits all. But money is essentially the broad overlapping thing?

[00:25:46] Aaron D Campbell: I mean, there are exceptions to that, where people are doing it for some political reason. Or some moral directive that they have or whatever. But the vast majority can be traced back to there’s money in it somewhere.

[00:25:59] Nathan Wrigley: I wonder curiously, because you mentioned about things like, pay per click style, you take oversight and you flood it with, I don’t know, nonsense about the thing that you’ve got and you want the world to notice. I wonder if curiously, people’s adoption of AI and that different way that we’re searching for things will actually impoverish that way of monetizing, because simply nobody’s actually looking on a search, well, increasingly people seem to be relying less and less on a search engine, and so maybe that kind of bit of it will dry up. Who knows?

[00:26:27] Aaron D Campbell: I love the optimism there. And I would like to think that those ads specifically probably will at some point. But the root of how those work is, I’ve broken into a site and I can inject some JavaScript ad, or some something like that.

And if those ads stop being valuable, then maybe I can inject some AI directives so that when an AI agent of some kind hits that site, it’s getting some sneaky thing snuck into its memory, or pulled in as a skill, that can then use that AI agent for nefarious purposes in the future.

I think that we can’t lower our guard against those things, because our adversaries will pivot and reuse it for something else. And so we will continue to protect against it.

[00:27:13] Nathan Wrigley: I’m going to peel back the contents of your head a little bit here. Because I’ve often wondered what the characteristic is of somebody like you who constantly facing this tidal wave of things. You’ve got to get up every morning, and every morning you could potentially wake up to the next big thing.

How do you just remain calm in the face of all of it? It’s a peculiar question, I realise, but tomorrow could be the next big thing. The day after that could be the next big thing. I’m imagining on most days now there is not necessarily the next big thing, but there’s a thing. It’s like you’re a fireman or something, except that there’s a fire going off in every district of town, and you are constantly busy and you never get to put the fire hose down. You’re just constantly at work.

[00:27:52] Aaron D Campbell: Some of us love that little consistent regular shot of adrenaline, and we get it in different ways than the firemen. But, honestly, I love complex problems solve. I love the challenges. Do I get exhausted and burnt out at times when they really do come every day for X amount of time? Sure, I’m human, I need to sleep, et cetera. But I think that it’s because I enjoy figuring out those really difficult problems, that I enjoy being in this space. And even specifically on this side of the space, the white hat side.

[00:28:26] Nathan Wrigley: Yeah. I suppose it’s like playing a good opponent at chess. You enjoy the chess game, even though it’s a hard thing and it stretches your brain. You play the chess over and over again because it’s a pleasurable thing to have your brain exercised in that way.

[00:28:39] Aaron D Campbell: And it’s like the more you do it the better chance you have at winning at chess. And I think it’s the same way in our game, right? Like the more you’re doing it, the more ways you’re finding to outmanoeuvre and to essentially win, and keep people safe online. And that’s, that’s exciting.

[00:28:56] Nathan Wrigley: You get the fist pump moment do you, there’s once in a while where you literally figure something out and you’re like, I nailed that.

[00:29:04] Aaron D Campbell: You absolutely do.

[00:29:04] Nathan Wrigley: Okay. Yeah. That’s really interesting. Your bio reads like an open source manifesto. I know that open source has been the thing for you throughout your career. I imagine that you could have gone into proprietary security and all of that.

But how does open source, particularly WordPress, how does that approach to developing software, how does that benefit the position that we can take and the security posture that you can take, and the reliability that you can have in things like WordPress going forwards? In your head, does it offer a superior model for fighting the adversaries?

[00:29:35] Aaron D Campbell: Yes, in my opinion it offers a superior model for fighting the adversaries. And the reason is actually still the same as it was 20 years ago when I started doing this. And I’ll explain why in just a second, but first, the reason is because we are able to benefit from many intelligent people, many more than any single company could with their source code.

Looking at our source code, and finding the weaknesses and even pitching in to help fix them. Hundreds of thousands, or millions, of people around the world are looking at our source code, finding those issues, and able to help pitch in and fix them, or report them to us so that we can, by having all that out there, it’s sort of like a building’s not less likely to collapse because no one saw the crack. It’s actually better that people are inspecting it, and finding those things and making sure it’s done right.

Now, that has shifted a little bit now, where in the past our adversaries looked at our source code too, right? They would immediately look at our repository when new things went out. As a matter of fact, when I ran the security team, I stopped committing stuff for a while, because it turned out that was a tell that this thing was probably a security issue, and people would look at that and try to figure it out.

So now the adversaries are using AI to watch our source code, which is also open to them, 24 7 and really look deep at it. But so are those many thousands of people using our source code for good. And so it’s still true that we have so many people on our side in helping us out, because our source code is out there because we’re open source, and it outweighs the bad. We find things faster because of that, and ultimately end up with more secure software more rapidly.

[00:31:36] Nathan Wrigley: Is there ever going to become a time where the amount of time that a vulnerability is available becomes moot? So in the past, a six hour window, where something wasn’t patched in WordPress Core, that’s a thing, but it’s not really a big thing. Maybe a month where something’s unpatched, that’s a big, I’m just wondering if in the future with the nature of the adversaries that you described and the tooling that they can bring to bear, if even like a three or five second window is going to become a thing.

[00:32:05] Aaron D Campbell: That is possible. We’re not at the three to five second window yet, so that’s good. I’ll let everybody relax a little bit. But the time from vulnerability disclosure to exploitation has shrunk dramatically over the last few years.

We did in fact used to have weeks, and then eventually days where it was okay to find out about the vulnerability, and responsible people, or responsible hosts, or responsible software companies could see that disclosure, patch the problem before there was much exploitation at all. And that’s now hours for major vulnerabilities.

As a matter of fact, we did a Monarx in conjunction with Patchstack, did like a year in review, thing looking back at last year. And we saw that for the more major vulnerabilities, it was about five hours. That’s not enough time for, you know, what happens if it happens in the middle of the night for a host who’s constantly monitoring that, immediately patching it, that’s difficult. And I do think that it will continue to shrink. And that that time that causes risk will be shorter and shorter.

And, we saw that with the recent wp2shell WordPress exploit. Once we released the patch, and everything was out there, the spike of exploitation that we as Monarx saw, like monitoring stuff happened within 30 minutes. Honestly, even a little bit faster than that. But the big spike started coming in about 30 minutes later, and that is just really fast.

But on the flip side, all that coordination that I talked about that WordPress did, had many millions of people already protected by then. And that’s how we have to look at it. We have to say, this is eventually going to get down to three to five seconds being a problem. How do we get ahead of it? And that’s what we’re trying to do.

[00:34:03] Nathan Wrigley: So a timely thing at the moment is this new innovation in the WordPress space called Protect the Shire. And Protect the Shire is the, a time bound moment where a plugin that has an update, it can’t be updated at the moment, I believe it’s standing at something in the region of six hours. On the face of it, that seems like a really excellent posture. But then there’s the flip side of that. If an exploit becomes discovered, and nobody can update their plugin for six hours, then that’s a big six hour window we’ve just painted. in the future where milliseconds may count. What do you think about that? It’s a interesting innovation. It’s something new. It was worth a try. Do you think that’s the way forward?

[00:34:41] Aaron D Campbell: I think it, was not only worth a try. I think it was a really good choice, and I think it will continue to be. We are seeing that a lot of current attacks are essentially supply chain attacks. How can we compromise whether it’s some package, an NPM package or something like that. Or whether it’s a plugin that’s gotten sold to a nefarious person, or even just hacked into and taken over by a nefarious person. That is happening more and this six hour gap helps protect against that.

But it can’t be a hard and fast, locked in stone, can never have exceptions, rule. And the truth is, it’s not. If there is a vulnerability in your plugin, or especially in a major plugin, reach out to the WordPress Security Team because we can coordinate a faster release, we can make an exception to that rule when it’s necessary.

And I think that for security fixes that are clearly security fixes that exception iss an easy one to make, because we do want to protect immediately. But slowing things down enough to make sure that there’s not been some sort of supply chain issue that is actually going to cause a vulnerability rather than fix one, is smart.

And so I think that it will find the right balance there as we continue to move forward, and figure out how to make better processes around this, to make it easier to do the right thing all the time, and know exactly which of those right things. But I think at the moment we’re in a pretty good place there and continuing to find the exact right place.

[00:36:18] Nathan Wrigley: It feels from the outside as if security’s fairly binary. On the one hand, adversary on the other hand, good guys. On the one hand hacked, on the other hand, not hacked. It’s black and white. But it seems from everything that you’ve been talking about today, that you are occupying a really grey area. You’re just trying to figure out what the path is forward. You’re constantly staring into the future trying to figure out what the adversaries are doing. Trying to patch, trying to make sure that everything is as good as it possibly can be. And I hadn’t really thought about it in that way. It’s not, there is no destination here where everything’s white. It’s a journey and every day’s going to be a bit grey. There’s going to be a bit of black and a bit of white, but a lot of grey in the middle.

[00:36:58] Aaron D Campbell: I don’t really look at it as grey, but I can see where you’re going there. But I think that there is this black and white, and then there’s sort of the cloudy. The further forward you look, the more difficult it is to know exactly where the black and white are always going to be. And some of that sort of comes across as grey. It’s a little blurry. You can’t quite figure it out.

But you’re right. there’s some prediction. There’s some, we think that moving this way is going to cause more white and less black. And that’s what we’re, that’s what we’re constantly aiming for. But you can’t just say turning right always makes things more white. Because sometimes there could be black over on that side somewhere, right? You’re really trying to be predictive, but not just randomly predictive, right?

Many of the people like myself that are trying to help guide this path forward, and even more than me, some of the people like Matt, who instituted that wait policy and some of the people that are running the WordPress Security Team, we have decades of experience watching this, that’s helping to inform our predictions. And so it’s not, we’re not just willy-nilly guessing. And I think that’s important to point out to the people that rely on us, to help guide them to the right space going forward.

[00:38:02] Nathan Wrigley: Yeah. Okay. So my schooling in chemistry was pretty basic, but I know that if I want to understand chemistry, my quickest way to do that is to rely on an expert, is to go and find a chemist who has years of experience. And, the same would be true here. I think most of us have probably not got the capacity to actually get a hold of what you’re saying. We, understand that your expertise is what we need to be listening to. But I’m just wondering for a typical WordPress user of whom many listen to this podcast, they have a WordPress site, but they’re not really interested in security, other than how it may impact their business.

So I’m going to ask for some very basic advice here. What would be the 1, 2, 3 things that somebody using WordPress with no security credentials whatsoever. What would be the few things that you would advise them to either go and read, or go and do, or go and think about?

[00:38:51] Aaron D Campbell: Yeah, so I think the biggest thing that I would encourage them to do, is essentially position themselves in such a way that they are relying on the experts, right? You’re not an expert and that’s okay. No one can be an expert in everything. But there are some things you can do to position yourself such that you’re benefiting from those experts.

One of those is updating as fast as possible. So WordPress auto updates turned on, those kinds of things. This WordPress Security Team that I’m talking about that has so much expertise in the area, and their whole focus is trying to make sure that WordPress is always secure. That lets you rely on them to help keep your site secure.

I think in similar ways, you want to find the right host that is also doing those security focused things for you, so that you don’t need to. And maybe that’s asking your host, what do you do to keep me safe? We talk about security, like the best security is layered security. It’s almost like having a gate at the complex, but also having a lock on your door, right? Those kinds of things. You can ask your host, what do you do to protect me in a layered way?

And maybe that question doesn’t make sense to you, and maybe even their answers don’t make sense to you, but if they have an answer, that’s good for you. It means that you can rely on their expertise.

And then stepping out of the obvious space to give a third thing that people should be doing. And this is, this may sound out in left field, but you should get some form of dark web monitoring for yourself, for your own credentials. And whether that’s going to someplace like, Have I Been Pwned, and looking at your own email address, and passwords and seeing if they’ve been in some breached data from somewhere, and are now being sold on the dark web. Or whether that’s using some service that offers it. I think that’s important, more so now than it’s ever been.

Because one of the other things that AI is doing that it’s particularly good at is collecting all this massive amount of breach data that’s happened over the last couple decades, that’s being sold on the dark web. Collating it all and saying, oh, we see that, gosh, 15 years ago, an account that Aaron had was in a breach. And we now know one of his passwords.

And granted it’s 15 years old, but it’s very easy for that AI to then say, where is Aaron now? What’s he doing? What can we learn about him? He works at Monarx. I wonder if this password works for his Monarx account. I wonder if this password works for the bank that he’s at. Or this other tool that we see that he uses. Let’s try all his social media accounts.

And so knowing whether that’s out there and being able to, you can’t get rid of that data, but being able to do things to protect yourself because you now know it’s out there is more important than it’s ever been.

[00:41:47] Nathan Wrigley: I hope you take this in the spirit in which it’s offered, but I really do wish to live in a world where you don’t have a job.

[00:41:55] Aaron D Campbell: Me too.

[00:41:56] Nathan Wrigley: But, am glad that we live in a world where you do, somebody like you does have a job. So I hope that landed correctly.

[00:42:02] Aaron D Campbell: If I could work to the point where I could work myself out of a job, I would find a new career and I would feel so accomplished, you couldn’t even imagine it. I’m okay with that.

[00:42:12] Nathan Wrigley: Yeah. Good. Aaron, just before we wrap up, is there a place where you hang out online where people could poll you if they’ve got a question, or any thoughts about what we’ve talked about?

[00:42:21] Aaron D Campbell: Yeah, if you’re looking for me professionally, you can find me Monarx.com, M-O-N-A-R-X.com. I also have aarondcampbell.com if you want some of my own more personal takes on security and things. And you can find me on Bluesky or the WordPress Slack. Those are probably the biggest places I’m at.

[00:42:40] Nathan Wrigley: What I will do, dear listener, into the show notes, if you go to wptavern.com and you search for the episode with Aaron Campbell, you’ll be able to find the links. I will dig out the Bluesky, and the various social links and the Monarx website and what have you, so you don’t have to hunt around too much. Go there wptavern.com. And Aaron, thank you so much for chatting to me today.

[00:43:01] Aaron D Campbell: Thank you. This was a really fun talk.

[00:00:19] Nathan Wrigley: Welcome to the Jukebox Podcast from WP Tavern. My name is Nathan Wrigley.

Jukebox is a podcast which is dedicated to all things WordPress, the people, the events, the plugins, the blocks, the themes, and in this case a new way to keep up with WordPress news and updates.

If you’d like to subscribe to the podcast, you can do that by searching for WP Tavern in your podcast player of choice, or by going to wptavern.com/feed/podcast, and you can copy that URL into most podcast players.

If you have a topic that you’d like us to feature on the podcast, I’m keen to hear from you and hopefully get you, or your idea, featured on the show. Head to wptavern.com/contact/jukebox, and use the form there.

So on the podcast today, we have Damon Cook. Damon has been a developer in the WordPress ecosystem for over a decade, working with WordPress centric agencies, and more recently running his own freelance business, Noma Digital. He’s a longtime community member, contributor, and someone who remains keenly engaged with the latest WordPress developments from releases like 7.1, to shifts in plugins, themes, and the broader WordPress landscape.

Damon joins us today to discuss WP Trend Watcher, an open source tool he’s created to help WordPress professionals keep up with the ever increasing deluge of news, community updates and feature releases. As the WordPress space grows more dynamic, and let’s be honest, sometimes a bit overwhelming, WP Trend Watcher is designed to sift through a fire hose of RSS feeds, summarise the trends, and inject some much needed human curation into AI generated content.

Damon talks about the importance of staying ahead in the WordPress ecosystem, especially for freelancers, agency owners, and developers who want to anticipate changes and support their clients.

He gets into the technical nuts and bolts, breaking down how the tool gathers, and summarises, news. The models and local AI setup he’s experimented with, and the balance between automation and critical human review.

If you’re someone who struggles to keep up with the relentless pace of WordPress news, and wants a smarter way to separate the worthwhile updates from the dross, this episode is for you.

If you’re interested in finding out more, you can find all of the links in the show notes by heading to wptavern.com/podcast, where you’ll find all the other episodes as well.

And so without further delay, I bring you Damon Cook.

I am joined on the podcast by Damon Cook. Hello, Damon.

[00:03:04] Damon Cook: Hi, Nathan. Good to see you again.

[00:03:05] Nathan Wrigley: Yeah, thank you. Damon and I had a chat at WordCamp US. I think you said a little while ago that it was in 2024. So it’s been a few years since we’ve had a chat, and there’s been a few pivots in your life. At the time you were working for a hosting company, which I don’t think you are doing anymore. But we’re going to have a chat today about something which caught my attention probably on social media, I imagine, I saw it there. It’s called WP Trend Watcher.

I would just say at the beginning before we get stuck into the topic, if you go to the WP Tavern website, I will bury all of the links in the show notes. So anything that we discover today, go and click on the links over there. And actually given the nature of the topic, it might be wise, dear listener, to go and do that before you get into this podcast, because it’ll be fun to have a poke around and be forewarned as to what is it that we’re going to be looking at.

So anyway, Damon, thank you for joining us again. Just before we begin, I mentioned that you have changed roles in the last 24 months or so. Do you just want to tell us a little bit about you and your roles in WordPress over these many years?

[00:04:07] Damon Cook: Sure, yeah. Yeah, been a developer in WordPress ecosystem for over a decade. And always been with WordPress-centric agencies. And yeah, as of the past year about, I’ve just been running my own freelance, Noma Digital, then also doing some contract work. And yeah, and still keeping up on all the latest WordPress news and features coming out. 7.1, what is coming out in, well yeah, in like a week, I think.

[00:04:37] Nathan Wrigley: Yes. Yeah.

[00:04:38] Damon Cook: Lots of cool stuff there.

[00:04:39] Nathan Wrigley: Yeah, there really is a lot happening. Genuinely, there is a lot happening in the WordPress space. And, it may be that you’re a developer, and so that side of things captures your attention. But from my perspective, I am not really a developer, but I am curious about all the things that are going on, because I’m fascinated by Core and plugins and themes and blocks and all of that.

But equally, the community side and all of that. There’s a whole load going on, and events and marketing and loads going on. But that deluge of information, that sort of fire hose is very hard to keep a track of. Some websites that were publishing for many years die out and other ones replace them, and things come at you from different angles.

And so you have built a tool, WP Trend Watcher, which you’re able to put on your local machine. I believe that’s right at the moment. It’s a sort of local install, and we can talk about the technicalities of that in a moment.

But this tool, the way that you’ve built it is primarily pre-built, if you like, to go out and inspect the WordPress space so that you can summarise the zeitgeist, of what’s going on in the WordPress world. You can keep up without having to do that manual slog of going out, reading everything, spending hours every day kind of wasting time, and I’m doing air quotes there, if you know what I mean.

Okay, so my first question is revolving around the why of this. What was the importance here that you felt that you wanted a tool that did this for you? And really what I’m trying to get there is why did you feel it was important to keep up with the WordPress space? Is it… if you’re a client, if you’ve got relationships with clients and you’re building client websites, is it not enough in your scenario just to know that WordPress drops, they do a version change every now and again, we get a release, I’ve got a suite of plugins, I probably need to keep up to date with those. Why? Why are you interested in the broader landscape?

[00:06:34] Damon Cook: Sure, yeah. I’ve been a, I was pretty active contributor for a while to the project, not so much lately. But, when I was a contributor there was, I mean, there’s always been a lot of news to keep up with, between Slack and Trac and WordPress.org stuff. But I started to see some of what were the reliable sources in that area. And as I work with WordPress daily, I still need to keep up with what is coming down the pipeline.

And so yeah, I just wanted to take that, it was very a selfish starting point. I needed to take all these RSS feeds, get them summarised, but then also verify that, what’s being summarised with AI isn’t just hallucinations or some kind of, you know, mistrust was there, because that’s a critical component to this project was the human layer. It’s just pulling in a bunch of RSS feeds, summarizing it, and then I go through and make sure the summaries and click through it, like a lot of the stuff, and read it, and it’s helpful for me to process and see what’s actually going on and then provide my own judgment on it, so.

[00:07:47] Nathan Wrigley: The sort of curious bit there then I suppose is that you, it feels like you always want to be one step ahead of what’s happening in the WordPress space. And I suppose if you’re being a freelancer, having that knowledge of what’s coming up, particularly in WordPress and WordPress Core, because I’ve noticed that a lot of RSS feeds, for want of a better word, and we’ll get into that in a moment, are Core related.

So that then sort of tells me that you want to be ahead of what’s coming in the next feature set so that you can, I don’t know, forewarn your clients, or practice things that you know are going to suddenly drop on a particular day of the year, and so that when the inevitable call comes, “Why is this there now?” And, “How does this new feature work?” You’re ahead of it all. Does that kind of summarise your interest in it, or is it more eclectic than that? You’re just curious about it because it’s curious.

[00:08:39] Damon Cook: Yeah, I mean there’s certainly the curiosity is there, but yes, I do like to be at the bleeding edge of what is coming down in the new, like for example, the 7.1 release coming out. Just seeing what’s being worked on, what might have gotten dropped from proposals. Yeah and just I guess being in the ecosystem, I’ve seen how a lot of the community has come and gone, and what has been critically talked about is, what is a critical feature, and why might something not be worked on and all that stuff.

So yeah, it’s just a selfish way to keep up on the latest, developer news and, yeah, and share it out and summarise it and get some human judgment behind it.

[00:09:22] Nathan Wrigley: A curious thing which is not related to this podcast, but most Mondays I do like a live show in which we have three or four panelists, and we bat around the WordPress news from the previous seven days. And I’ve been doing that show for, oh, I don’t know, maybe seven or eight years or something like that. And in the sort of the week preceding, I go around the internet, I look at social networks, I read the blogs that I’m familiar with. I’ve got a giant list of RSS URLs in a feed reader, and I go through, and I read them, and I look at them all.

And it, genuinely is incredibly time inefficient. That is to say that I open things, get halfway through, realise there’s nothing really for me to talk about here with my guests, and there’s probably nothing for the the listeners to engage with. And so I lose a lot of time, doing that exercise, and it strikes me that in the era of AI, that would be, the intention here is to cut out that. Cut through the dross, sift out the bits and pieces that are of interest, and make the bits that finally reach your eyeballs worthy of attention.

How have you orchestrated that to happen? Because with the best will in the world, I think we all know that AI can go off the rails a little bit. It may be that it doesn’t quite match the intentions that you would wish it to have. How are you corralling it to be what it is that you want?

[00:10:47] Damon Cook: Sure. Like I alluded to, it just stemmed from a selfish kind of interest in keeping up on these news sources, and I knew a lot of them all had RSS feeds. I think that even I, you know, I have a fascination, an obsession almost, with AI these days and I’m always trying some cutting, more cutting edge stuff and tinkering.

But I think even myself, and I think I see a lot of people confuse AI with automation. And so it’s interesting to see, when I, this project, I didn’t want to automate the whole thing, like just go get RSS feeds, AI summarises, AI publishes, just do it, right? Which I can, we could potentially do with the project, but I limited it to that contract of, go get these RSS feeds, the latest news, come back, summarise them, and then show them to me so I can audit them.

A lot of them, I just open up like yourself. It seems counterproductive, but I actually like that’s how I really keep up. Because if I just read the paragraph from the AI summary, which I do, it’s it’s not really covering, and might miss something in that full article that actually really interested me.

So you know, I click through most of these and I’m reading. I peruse, I don’t read every single word for sure, but I peruse and just try to validate that the summary matches, and then I look for tidbits that might been missed that I’m curious about, and then drop that in the human summary element of the published final briefing. So that’s the workflow of it right now.

[00:12:27] Nathan Wrigley: Yeah.

[00:12:27] Damon Cook: That’s how I’ve set it up and I enjoy just the whole, the sequential part of like gathering and then, there is an automation of it shows me everything I gathered, and then gives me an area to write my thoughts and then I click through everything. So it’s just turned into a weekly, kind of hobby project and it’s fun to do, so.

[00:12:48] Nathan Wrigley: Yeah, so the bit that is infuriating to me is just the dead time. And the dead time for me is not once I’ve found the link. That for me is when it actually becomes interesting. You click the link, you open it, and you do the reading and what have you. It’s finding the link and finding the interesting thing.

[00:13:07] Damon Cook: Yep.

[00:13:07] Nathan Wrigley: And so I think, it is totally credible to get AI to do that. And I have this, I’m not a great user of AI. That is to say I’m not somebody that uses AI an awful lot. But it’s becoming pretty clear to me that one of the areas which it genuinely is good at is curating like a silo of information. You give it a body of information, a corpus of information, and say, “summarise that,” or, “Draw out of that the bits that might be of interest.”

Now, it’s not perfect, but it seems to be exceedingly good at that. The creative side, maybe less so, and we don’t really need to get too much into that. But this bit of drop the corpus of information in, summarise that, publish your results. That seems like a really credible way of using AI in the year 2026, and that is indeed what you’ve done.

So now that you’ve got this built, and I will link to a report so you can see Damon’s previous, I think you’ve got seven or six or something like that, previous ones from previous weeks up, and you can see them all. It basically ends up in a bullet-pointed, numbered list of titles with links, and then maybe one sentence of about 20 words following on from that. And then that will obviously either lead to a click or not.

How have you found those summaries to be? The 20 or so that you’ve got in each of your little editions. How have you found them to be? Do they seemingly accurately represent what’s contained within, or does it tend to go off the rails a bit?

[00:14:39] Damon Cook: No, they’re accurate.

[00:14:41] Nathan Wrigley: Yeah.

[00:14:42] Damon Cook: The key kind of component there is when the summaries happen, it’s using, I set it up because I wanted to use local models. It’s all prompt-based, so essentially, it’s going to these RSS feeds, getting the latest articles and links, then parsing that through AI and saying, and passing a prompt along saying, “summarise this for developers and agency owners with a critical eye on, X, Y, Z.”

And so that’s where things do actually come, they come back pretty accurate. And the inference of what the summary shows, it’s fascinating. But yeah, it’s pretty accurate, and the fun part is trying to prove it, to see like, “okay wow,” that sounds like it would interest me as a developer and agency owner. I’m going to click and read through and make sure that’s an actual thing, and not a hallucination. So it’s like a game of just fact-checking in a way.

[00:15:46] Nathan Wrigley: I’m curious to know what it doesn’t include. So for example the addition, I’m going to call it an addition. I don’t know. Oh, summary. There we go. Weekly summary is what you’ve called it. This particular one has 19 items listed, and I don’t know if it produced 19 different things and all of them got written to the page, or if, I don’t know if you can see in the background, 34 just didn’t cut the mustard. Can you inspect what didn’t get included and thereby start to have intuition to say “okay, I wonder why that one didn’t cut the mustard?”

[00:16:20] Damon Cook: Yeah, there is a built-in, for the AI to, when it does a review, it checks sources and it looks for duplicate, two kind of posts on two different sites about the same exact topic. It’s going to try to not do a summary on both of those.

[00:16:39] Nathan Wrigley: Okay.

[00:16:40] Damon Cook: So it does check for duplicates, weak language, missing sections. So it does do some of that to cut out the duplicate work. But also, yeah, it’s not looking, like part of the prompt and part of the exercise for AI is to not look for superfluous, just like really fact-based deterministic stuff and not just you know. It’s funny because trend is in the title, right? WP Trend Watcher.

[00:17:06] Nathan Wrigley: Which is maybe the exact opposite of what you’re after.

[00:17:09] Damon Cook: It’s kind of like it’s not really looking for what’s popular, but what is factual coming out of a lot of these sources.

[00:17:17] Nathan Wrigley: Okay. When you are looking at this, this is a strange question, and I don’t know how you’re going to react to this. How does it make you feel? Do you have the intention that you’re cheating on the Damon from five years ago would’ve thought, “Oh, this is horrific, somebody did all this with a computer. This should be done by, a human being should be corralling this.” How does it make you feel on that level?

[00:17:41] Damon Cook: Yeah, it does make me feel cheap.

[00:17:43] Nathan Wrigley: Yeah.

[00:17:45] Damon Cook: And I think that’s kind of why I publish these. It’s trying to be transparent. So if anybody’s going to go and read this, they know that I even put at the bottom how much time I spent trying to look through everything. So you can take all those facts and base whether you want to read this or not.

[00:18:08] Nathan Wrigley: Yeah.

[00:18:09] Damon Cook: Did he spend enough to? You can pass your own judgment. Did he spend enough time? Maybe he should have spent more time.

[00:18:14] Nathan Wrigley: Yeah.

[00:18:16] Damon Cook: But yeah, that’s part of the exercise for me is making this open source, and just having this public, and so if folks do find it fascinating and helpful, they could even fork it and try to do something else. Or I even have, possibility to submit sources because I’m always thinking “Should I have another source in here?” or.

[00:18:39] Nathan Wrigley: Yes, yes.

[00:18:41] Damon Cook: There’s room for improvement too.

[00:18:43] Nathan Wrigley: The issue there is you end up with scope creep, don’t you? And before you know it, you’ve got 5,000 RSS feeds that you’re pummeling everywhere, and then you really are in a maze of trying to figure out what’s relevant and what’s useful.

The report that you get on the front end, obviously, you’ve got these AI generated the numbered list, in this case 19, the one that I’m looking at. Again, links will be in the show notes.

But then the interface that you built provides the option then to add sort of human notes, for want of a better word, at the end. So it’s not just a case of, okay, that’s where it ends. The idea is then that you would go in, contribute for yourself what it is that you feel is important here, free form writing like we always used to do, and so on.

[00:19:24] Damon Cook: Yep.

[00:19:24] Nathan Wrigley: Why did you include that bit? Why, what was the purpose in having that? Was it just a more human curated portion at the end, so you wanted to remember what you actually thought?

[00:19:33] Damon Cook: Yeah. It’s twofold. I wanted the human in the loop. It’s a pretty popular phrase these days with AI, but I wanted that aspect there. But also selfishly, a game to fact-check a lot of these things and make sure AI isn’t hallucinating. But also it’s the exercise of memory of like going through and reading these things, because then I actually, you know, it registers in my brain. Like this is really what’s coming down the pipe. Because if I just read through the summaries, even if they were 100% accurate and I just trusted them out of the box, I feel like it doesn’t necessarily always register for me. So the actually exercise of going and clicking and reading through and comparing to what AI summarised is helpful for to sink in that data.

[00:20:20] Nathan Wrigley: I think if this were something that I was looking at, and I’m definitely not ruling out doing something like this because I genuinely think there’s a lot of merit in my specific case for this. I think I’d be looking at four of those a day and just trying to, I don’t know, at 9 in the morning just before the day begins, or at 5 at night just as the day’s about to end, the workday. Just go in and look at a few of those. Because I’m like you, if I read the AI 20-word summary, there’d be no substance to that and I wouldn’t remember it.

But the fact that the interface is there and those articles have been surfaced, and I presume you could probably run one of these each day and adapt what was there the previous day and what have you. So technically then, how does it work? I’m fairly sure that this is a thing which you have built, and is currently sitting on a machine in your house somewhere.

[00:21:08] Damon Cook: Yep.

[00:21:08] Nathan Wrigley: Just describe the technicalities of what you need to get, on your machine to make this occur.

[00:21:13] Damon Cook: Yeah. It’s a TypeScript code base, and it’s on GitHub, so you can fork it, contribute to it. But yeah, I run it. You can pass in kind of a flag. I run it weekly. It was just an arbitrary decision, and it felt weekly would be good. But you can, yeah, it just checks basically the last date of when it ran, and then tries to grab the sources.

Then I use LM Studio application on my Mac, which allows me to kind of download models that are fit for my system. I have an 18 gigabyte memory that I can work with, so that is limiting. But, I think a lot of these models are coming down in size and the local AI movement, I don’t know if it’s a movement, but I think it’s taking off in a lot of ways.

[00:22:07] Nathan Wrigley: Yeah, I think so. Yeah. Yep

[00:22:09] Damon Cook: We’re starting to see a lot more, yeah, local implications. That was part of the exercise for me, was exploring that. What models I could use on my system, and then what were the results of using those models with this tool for the summaries.

So yeah, I think every week at the bottom, that’s part of the report is what model I used locally to run the summaries against. And I try to switch it up. I think I used probably six of them, I used the same model. I think it was a Qwen model. Then the most recent one I used, I switched it up, and it was a, I forget the name even. It was like an LS.

[00:22:49] Nathan Wrigley: Oh, these things come and go so fast. Yeah.

[00:22:51] Damon Cook: Yeah, exactly. And that’s the fascinating part is just to drop in a new model and see how it does, and the softwares look like, because they do change. And even the report, I think you’ll, if you just compare this week’s with last, just the output is a little different. And I’m trying to… That’s something I plan on working more towards is deterministic, the output. The summaries are one thing, but then when it outputs into the final report, getting that more deterministic. Especially, I really want to have footnotes, so that people can see where these summaries and stuff are, what they’re exactly they’re linking to. It’s kind in there but it does’t always output exactly the same in formatting, so.

[00:23:35] Nathan Wrigley: So you’re doing this work on the local machine, and then obviously I’m looking at it so it’s published, by the looks of it on GitHub. Is that all automated or are you curating that process? Or have you got, you know, your local machine is bound to GitHub, your GitHub repository and so this is all automated in that sense as well?

In other words, what I’m saying is if you took a break for a month and came back, would this all just be chugging along in the background and producing these weekly reports? Or is this something that you manually invoke on a weekly basis and copy and paste and what have you?

[00:24:07] Damon Cook: That’s the part that, yes. I just run the report and write my judgment in it, then publish it to GitHub, and then GitHub does a little, it takes basically the artifacts and just spits them out. But there’s not much there in the automation on GitHub. It’s just publishing it to GitHub Pages.

The fact that I have a section that’s required for human in the loop kind of takes away full automation. I could make it run, but then I would have to take that component out of me reading through, clicking through, and writing my own judgment at the end of the report. And I don’t plan on getting rid of that. I actually enjoy that aspect, so.

[00:24:47] Nathan Wrigley: So it’s become a bit of an interface. The AI plus Damon equals the finished product. You really don’t want it to be automated. And in a way, that feels to me a bit like, I don’t know, a race to the bottom. If we automate all the things, I just don’t know where the audience for that lives, because you’ll, never really know. Because there’ll never be a moment where you look back at it and say, “Is anybody reading this?”. “Has anybody talked to me about this?” and what have you. Whereas the fact that you’re interested in it, you’re on this podcast talking about it, you’ve got an interest in the subject matter, you’ve clearly engaged with it. I don’t really have a question there, but there’s definitely something there. The fact that you’re in the weeds of it and it’s not just AI for AI sake, I think is important.

[00:25:32] Damon Cook: Yeah. And it’s, yeah, that’s a great point. And I think that makes me, I don’t know, I guess go back to, where some of this started from. Yeah, I think people are creating their own knowledge bases. And you even touched on this. Just having that corpus of artifacts, a lot of that AI can do these days, and you can automate some of that.

But, how you take all that data and get the human in the loop and, is it even interesting if you, it’s interesting maybe to you to read, but if you want to share that, is it going to interest others? And yeah, the human aspect is critical for me there, because, especially if you’re going to share it publicly, you know.

[00:26:14] Nathan Wrigley: Yeah.

[00:26:14] Damon Cook: At least be transparent.

[00:26:16] Nathan Wrigley: Yeah, so you’ve called it WP Trend Watcher. And it feels to me, if I was to rewind the clock to, I don’t know 2023, where for the first time we were all beguiled when ChatGPT came along, and we all just thought this whole interface was remarkable. But it felt, oh, it’s novel, it’s interesting, it behaves in many ways it’s a bit, I don’t know, passed the Turing test and it convinced us that it was human.

But now that we’re in the year 2026, it feels like one of AI’s central roles is just, as I said earlier, these corpses of information, but not finite in the way that a human brain is. There’s only a capacity that I have to remember so much, and it’s alarmingly little. But over time, if you were to stick with this endeavor, the name that you’ve given it, this sort of Trend Watcher, it feels like the data that you’re gathering, as it grows and grows and the AI’s capacity to manipulate, understand, I’m doing air quotes, “understand” and, create meaning from the disparate bits of data, that will grow over time.

And so it’d be interesting in 2027 and 2028, what happened? Why did that bit never got built? And how did we get this thing shipped into Core? Where did it come from? What were all the little bits and pieces? Where were the jigsaw pieces that led that to happen? And how did this happen? And what’s the market trend? Why have we got a 4% decrease or an 8% increase in market share? How’s all that happened?

So it feels to me as if you keep pushing this out, that kind of information will be available, kind of MCP-able, by simply asking it questions, in a way that we’ve never had available to us in the past. Again, there’s no question, but do, you know what I’m saying there?

[00:28:03] Damon Cook: Yeah. I think, when I set out to build this there is, and there is still a component where I do want to have kind of the trend watching part of taking these past reports and having AI review them, and come up with inference of was there a trend? Why, yeah, why did we go from report 1 to report 18 about this? Maybe there’s this feature in WordPress 7, whatever. What changed there? And come up with a separate component for that. I haven’t built that in, but that’s kind of a potential, that’s always the end goal. And keeping these receipts for now and generating these reports is just the scaffolding in a way.

[00:28:48] Nathan Wrigley: So it feels to me as if it really isn’t bound to WordPress. You could run with what you’ve built and point it at just about any industry or thing. Your local tennis club or, whatever it may be. You could, over time, build up this sort of trend analysis, and see what’s going on.

And dare I say it, you could even implement it upon yourself so that you can see, what’s the cadence of my writing? How have I done my fitbit data or whatever it may be? You can see that, even as I’m saying the words, I’m thinking, “Oh, this is a dystopia that I don’t want to live in.” But at the same time, I’m really curious as to what that bigger capacity to maintain state and hold data over time. What that will furnish us with.

Because we all forget things, our brains are very ephemeral. Things that are important to us come and go. We misremember things constantly, and it will be really interesting to see what things like you’ve built, are going to enable us in the future. And it does seem like this is baby steps. The humans in 2029 looking back on the work like you’ve just been talking about for the last half an hour will think, “Oh, that was cute, but look where we got to now, we stood on the shoulders of those giants, and here we are now.” It’s fascinating and terrifying in equal measure. Let’s say it that way.

[00:30:11] Damon Cook: Yes. Yeah, absolutely.

[00:30:13] Nathan Wrigley: Yeah.

[00:30:14] Damon Cook: There’s so much data and it’s just growing, right?

[00:30:17] Nathan Wrigley: Yeah.

[00:30:17] Damon Cook: It’s going to get worse before it gets better.

[00:30:19] Nathan Wrigley: Yes. Well, it’ll be very interesting to see how it all lands. But, what I would encourage everybody to do, I’m going to link to at least three things. I will link to the GitHub repository, which is obviously at Damon’s GitHub account. It’s colorful-tones, and then this is called WP Trend Watcher. I’ll also link to damonacook.com, where there is a blog post about how you can get up and running and what the necessary steps are for you to do that.

But I will also link to the sort of output, which is again, it’s a GitHub repository, but you can go and see the weekly summaries and what have you, and judge for yourself whether or not this is something that you want to install on your local machine.

I think the inevitable is going to be that there’s going to be more and more of this in our lives. It remains to be seen whether we have the time to actually interact with the data that it puts out, or we’ll need another layer of AI to cope with the level of AI we’ve got, if you know what I mean. It’ll be wheels within wheels.

So there you go. Anything you want to add? Did I miss anything there before we wrap it up?

[00:31:27] Damon Cook: No, the only thing just you know I’m colorful-tones. Actually, it might not be with, it might not have the hyphen, but on X. And then on LinkedIn I try to just share the report out weekly on there just for folks keep up, but that’s about it.

[00:31:47] Nathan Wrigley: Lovely, in that case, I will endeavor to find your Twitter or X handle and various other bits and pieces. And, wptavern.com, if you search for the episode with Damon Cook. It’s spelt in the way that you might imagine, D-A-M-O-N Cook, C-O-O-K. Search for that. And yeah, thank you so much for chatting to me today, and bravo for a topic which I think I’m going to be leveraging very much for my WordPress news cycle.

[00:32:14] Damon Cook: We can get that, get some RSS feeds and we can get them, get you hooked up.

[00:32:18] Nathan Wrigley: Yeah.

[00:32:18] Damon Cook: You can start generating some other reports.

[00:32:20] Nathan Wrigley: Okay, great. Yeah, let’s have a chat after we click stop in a moment. Damon Cook, thanks so much for chatting to me today.

[00:32:26] Damon Cook: Thanks.