Welcome to another episode of the Osom to know‬ podcast! In this episode we explore the changes in the WordPress ecosystem with our guest, Hendrik Luehrsen. Hendrik, a veteran WordPress developer, brings invaluable insights from his extensive experience. We discuss the evolution of WordPress, the rise and fall of WordPress themes, and explore the impact of new technologies like AI on the ecosystem. Hendrik shares his thoughts on full site editing, Gutenberg, and the strategic shifts his agency has made in WordPress development. Whether you’re an agency owner, a developer, or just curious about the future of WordPress, this conversation is for you.

Hendrik Luehrsen [00:00:00]:
And he’s in our agency, an inside joke how bad ThemeForest, code quality is. The standardization of content of the web in markdown, that is something I don’t think Matt would enjoy. We were ourselves every time we opened the block editor.

Maciej Nowak [00:00:17]:
Hello, everyone. My name is Maciej Nowak, and welcome to the Osom to know podcast where we discuss all things related to great websites. Today, we are diving into evolution of WordPress with our guest, Hendrik Luehrsen, a veteran WordPress developer and the CEO of Luehrsen-Heinrich web development agency. Hendrik has been part of WordPress ecosystem since 2006 and brings valuable insights from his years of experience. In this episode, we will explore how WordPress has transformed over the years, the rise and fall of WordPress themes, and new technologies like AI, and how AI are reshaping the entire ecosystem. Hendrik will share his honest perspective on full site editing, Gutenberg, and why his agency has made strict strategic shifts in how they approach WordPress development. Whether you are a development agency owner or just curious about WordPress, this conversation offers a candid look at where WordPress has been and where is it heading. If you don’t want to to miss new episodes and keep learning more about WordPress, please subscribe to our newsletter at osomstudio.com/newsletter.

Maciej Nowak [00:01:26]:
This is 0s0mstudio.com/newsletter. Our if you are watching this on YouTube, please give us a thumb and subscribe to our channel. This means a word to us. Without further ado, please enjoy my conversation with Hendrik Luehrsen.

Lector [00:01:48]:
Hey, everyone. It’s good to have you here. We’re glad you decided to tune in for this episode of the Osom to Know podcast.

Maciej Nowak [00:01:56]:
Hi, Hendrik. How are you?

Hendrik Luehrsen [00:01:58]:
Hi. I’m good.

Maciej Nowak [00:01:59]:
That’s great to have you here today. And maybe just, like, for starters, can you introduce yourself to our listeners maybe?

Hendrik Luehrsen [00:02:08]:
Sure. My name is Hendrik. I’m from Munich, Germany. I’m the CEO of a small advertising agency. And, we had the fortune of, jumping onto the WordPress bandwagon very early on, which, kinda led to me being in the community, for a long time now, and I’m always there and, always have an opinion on everything. But, well, here I am.

Maciej Nowak [00:02:34]:
Very good. That that’s why you’re here. So would you say you’re a developer?

Hendrik Luehrsen [00:02:38]:
Yes. So I started my first, WordPress blog back in 2006 for my own personal site. I was just moving to Munich from my hometown, close to Bremen, and, I wanted to blog a bit. And I first used, like, a niche blogging tool called, Serendipity, because I already knew PHP, so I wanted a PHP system. And, Serendipity was not very good. So I, jumped on board with WordPress, and that’s that was probably my first contact with WordPress.

Maciej Nowak [00:03:23]:
Alright. Yeah. So so how did it start in terms of, like, I’m curious now because we jumped the bandwagon a little bit, later when it was already you know, the train was rolling. And I’m curious to know how back in the days, how did it differ from what we have now?

Hendrik Luehrsen [00:03:41]:
From the term of contexts, if you look back at 2017, not that much. The first big evolution I visibly remember was the introduction of, the menu, meaning the patch where you could control the menu different for separate from the structure of your pages, which was like the advent of, true proper content management where you could, on one hand, write content, organize content, on the other hand, organize your menu structure. And that was that was, in the beginning, something big. But if you are so close, to the thing, the evolution feels very natural. But I believe if you would jump back into, WordPress system from even 2016 or or 2006, they would feel very different. It’s like watching your grow kids grow up. You don’t really see it.

Maciej Nowak [00:04:55]:
And then they’re 18.

Hendrik Luehrsen [00:04:56]:
You don’t really see a difference day by day, but suddenly, they are grown up.

Maciej Nowak [00:05:01]:
Exactly. Yeah. So now I like, you are providing me too many, like, options to continue this conversation. Like, is WordPress grown up? Is it a grown up? I certainly, it’s, you know, of came of age, as they say. But is it a grown up technology?

Hendrik Luehrsen [00:05:19]:
It is over 20 years old now. And whenever a 10 a technology can, not only have a foothold for over twenty years, but dominate the space for such a long time. I would say it’s it’s appropriate to call it grown up. Additionally, to the fact that there is, a boatload of people earning proper money with WordPress and in the WordPress ecosystem. Thankfully, me included. So, yeah WordPress is very grown up.

Maciej Nowak [00:05:53]:
Mhmm. Now if you continue growing up, you start to get, you know, elderly. And with all with all of the big technologies, there is that moment in time on the evolution of of the technology when you start to joke around joke about the technology that it looks like. This piece of software looks like the piece of software that is for, you know, forty, thirty, twenty years in the on the market, and this is no longer a grown up technology. It’s, you know, the dinosaur technology. Right? And with this, would you say we are approaching being an you know, in the an elderly space or still, you know, very fit and in shape were WordPress, you know, athlete.

Hendrik Luehrsen [00:06:42]:
And this is a question where I have to give my agency CEO answer. It depends. It depends on where you look at. We certainly have parts in the WordPress system that are ages old. I mean, one of, the first big, functions Matt wrote in back in 2003 are the actions and filter system, the hooks, that enables so much growth, that pretty much enabled the plugin ecosystem and the theme ecosystem. And much of the code that you see there is probably over a decade old, that hasn’t been touched in in over a decade. But, that in itself is not a problem because the ecosystem around WordPress and within WordPress is evolving very quick. We have, so many awesome contributors that spent time, on on things that are not very shiny, PHP compatibility.

Hendrik Luehrsen [00:07:47]:
The new block editor is certainly or school site editing is certainly a big, big, big, progress in term of staying up to date. And, well, that’s my it depends answer. Depends on where you look.

Maciej Nowak [00:08:03]:
Mhmm. How about the incumbents? How about the new technologies that, how to put it? Like, maybe they are, like, chipping away part of the the the market. The market, is huge. The market share for WordPress is huge. But, like, there are companies, like, you know, for the certain groups, like Wix or Squarespace or or Webflow, for example, which are growing so fast. You know, they are growing for a for a reason. They are growing so fast. Obviously, it doesn’t hard work as yet, you know, if you look at the big numbers.

Maciej Nowak [00:08:46]:
But there is a reason for their, like, absurd growth, especially for Webflow, for example. But there are also many other more niche technologies. So, like, the pie is getting bigger and they are growing because the pie is getting bigger, or maybe there is some, you know, migration of of customers? What like, what’s your take?

Hendrik Luehrsen [00:09:09]:
Certainly both. The pie of, people publishing content on the Internet is certainly bigger, And WordPress has not been able, to capitalize on a lot of these trends, especially if you look at, video content. Like, look at how big TikTok is, for example. It’s it’s humongously large, the amount of creators on there. Or Instagram image content, serialized image content. WordPress hasn’t been able to cut capitalize on that, not, because of lack of trying. I mean, Google had, like, a format back in 2021/22 called Web Stories, I think. Pascal Bircher, worked on that, and we had a working prototype for that.

Hendrik Luehrsen [00:09:59]:
But that sadly never took off. Like, creating TikTok or Instagram stories right in your WordPress and being able to have them shown nicely on Google, it was a nice experiment. Never went anywhere, sadly. But that doesn’t mean that those, creators would have come, to WordPress if TikTok didn’t exist. So the question is, albeit WordPress not being able to capitalize on that, is that the loss for WordPress? And I would say no. You would say, maybe Wix or Squarespace or Webflow are fishing in the same space, and I would answer to that or I say to that yes, for a to a certain degree, because Wix and Squarespace and also Webflow offer certain things that are very hard in WordPress right at the moment at this moment, it’s the time to publish content from nothing to published content in WordPress is fairly long. It’s a fairly long complicated process and having to navigate that alone is certainly easier in one of these platforms. But if you are up and scaling and suddenly have to worry about things like accessibility, search engine optimization, performance, traffic scale, these closed or, data protection, GDPR, these closed source plat platforms suddenly give you limitations you do not have with WordPress.

Hendrik Luehrsen [00:11:40]:
So for every problem, there’s a proper tool. And at least for now, most problems can be solved by WordPress fairly well.

Maciej Nowak [00:11:52]:
That’s interesting what you said about TikTok because, you know, it’s, like, totally different story where…. Mhmm. The the main benefit of TikTok is that it’s, like, video only. It’s all glued together by the the legendary algorithm that propose you the proper content that will, that you want to consume. And I really can’t see that transition or any anything similar to what you explained regarding the video format, like, tapping into video craze, let’s say, with WordPress. Because, like, by definition, this is not a so far, not a tool for any kind of, like, video editing. It’s like different continents. You know? Like, I totally does don’t get it. You know?

Hendrik Luehrsen [00:12:47]:
So I would, let’s stick with the TikTok video idea for a bit. I would challenge you to abstract the concept of TikTok one level further. Do not think about video. I think the appeal of of TikTok and Vine beforehand, was the speed from I have an idea in my head and that idea is published published on the Internet. That speed is incredibly quick in on TikTok and on Vine, on an Instagram and in Facebook stories. That idea to published content, that time is so quick because, in a picture, in a single picture, you only can transport so much information in an entertaining way. Video adds a whole new dimension. So video can allow you, to transport more information, than you can in a picture.

Hendrik Luehrsen [00:13:53]:
So, that is the, in my mind, the appeal of that. If we could somehow, find a way where within WordPress, we can have an idea and publish content within minutes, like entertainment, highly polished, beautifully designed content within minutes, I think that, would allow for another wave of growth. And I think AI can enable that.

Maciej Nowak [00:14:30]:
Yeah. I for sure will ask about this AI part. But I think for TikTok and any other platform, the difference is huge. I can’t agree with you on this because there are a couple of alerts like lights that are, you know, what people want. Like, they want the, the validation. You know? There’s the dopamine hit lights. Then there is discoverability where the algorithm pushes the content. So you are you not looking for new like, you are not looking for new content.

Maciej Nowak [00:15:03]:
All of the content is there within your one window versus in on WordPress. It’s like, you know, you will be on a dominion of one creator, one URL. Yeah. Another creator, another the the the like, it will require, like, something like a federation of websites to start to look like something very mostly similar to those closed platforms. And the native tools where you can record with your, mobile application. So, like, I think that maybe that would be easier if you have something is like, faster from creation to from idea to creation. But all of those missing components are, like, separating your content from the rest of the world, in my opinion.

Hendrik Luehrsen [00:15:53]:
True. It’s it’s certainly a problem because, discoverability is a huge thing. The TikTok algorithm is very good in serving me exactly the right video to keep me engaged. And we don’t have that in the traditional web. We don’t have that in traditional publishing. The Feediverse has the concept to allow that. I know that Automattic is actively working on integrating the Filiverse, tighter into, WordPress. And I think at least theoretically, replicating a discoverability with WordPress generated content is possible.

Hendrik Luehrsen [00:16:43]:
But that would require an amount of work, focused work in one direction that is simply not available for an open source project. That’s that’s the sad thing by design. What we do is designed by Comite and, being fast with a project this size is very hard. I really, do not want to trade places with Matt because, Matt has the job of steering this, huge behemoth of a ship that the WordPress project is, and it’s not easy. Certainly, it isn’t.

Maciej Nowak [00:17:26]:
I reached out, to record this conversation because, this started around the topic of themes. And back in the day, themes were super popular. Now, not so much. Like, I don’t hear so many clients, like, even mentioning the word theme. It’s, like, no longer a word in the vocabulary, you know, in the in the dictionary of client language. And I’m curious about your opinion on it. Like, what what has changed that this is no longer a a theme is not a thing anymore.

Hendrik Luehrsen [00:18:02]:
That is certainly, a case in a big area of the WordPress ecosystems. Themes have certainly, taken a step back from the limelight, since 2017, I would say. Before that, I mean, let’s let’s go back to 2006, how I, found the theme of my blog. I went through the WordPress theme repository, and I probably found I’m a bit inter interpolating here because I don’t really remember. And but I think I found a few quirky themes that probably had a lot of character and probably were in in one or the other way, broken, but did what I want them to do and had, like, purpose and character. And I wanted a blogging theme and not a photography theme. And I did not want to spend a lot of time on it. So there was that.

Hendrik Luehrsen [00:19:07]:
And there are places on the Internet where that is still true. I mean, look at Envato, like ThemeForest. ThemeForest is still earning a lot of money. There are still some successful theme vendors on there selling WordPress themes for, like, architects, photographers, whatever. And if you look at at what Squarespace is doing or what Wix is doing, they offer purpose built, utility first themes, what they call it, templates for specific needs. Me as a if I would be an architect that just came out of university and is now a freelance architect, I don’t need, a full site editing. I need a theme, for architects where I spend, like, an afternoon hacking my own, texts in there and be done with it. And that’s something, full site editing is severely lacking, and the other members of the WordPress editor ecosystem are not doing very good.

Hendrik Luehrsen [00:20:20]:
The closest, competitor, to, like, Wix and Squarespace that does this halfway decent is Elementor. They offer, like, also purpose built themes for whatever need, doctors, for the sports club, the regional sports club. You can find that there, but it is fairly hidden. And I think that’s where a lot of value was lost because, as I said, the time from idea to published content is very long. And choosing a theme or a template or how your website looks, whatever you want to call that, is currently in core WordPress way too long, which leads to, one of two effects. The first effect is all pages suddenly seem to look alike because everyone is using, the the default theme with some minor tweaks, or they are using Squarespace or Wix or workflow.

Maciej Nowak [00:21:27]:
But don’t you think that maybe the reason the themes were abandoned is because they were they came with a lot of problems, like, you know, sustainability of the theme, a lot of, like, plugins that that were required by the theme and so on. And at the end of the day, you would break the theme if you are that, you know, person that would, you know if you don’t have certain knowledge about what you are doing, like, with everything, you would, like, at some point break the website, then error 500 wouldn’t you wouldn’t have any any website because of that error, because of something you you you did.

Hendrik Luehrsen [00:22:09]:
At least knowledge and see, an insight joke how bad ThemeForest, code quality is. I mean, like, if we, happens rarely these days. We get a client website that is based upon a ThemeForest theme. One of our first demands is to switch out the theme to either something purpose built or something more standard like Elementor or something. So that was a problem we had, like, in 2019/20, just before the big advent of full site editing where in this theme space, you only had a chance of success if you had, like, a gazillion of features. Because we arrived at, a state of the theme market where you could not sell with style alone. You had to bring some tangible upside with your theme, through, like, sliders, customizable, whatever, the customizer, your own templating system or whatever, and it was a mess. I mean, vendor lock in being locked into, like, Visual Composer or Divi is a problem that many people had.

Hendrik Luehrsen [00:23:39]:
So the intent of what full site editing and the Gutenberg project, tried to solve is certainly valid. It was a mess back in the day, and, I mean, it certainly led to the loss of quirkiness in in the WordPress ecosystem. I mean, I remember we, tried our luck in the public WordPress theme ecosystem in 2019, and, we published, very unsuccessfully, an own, public, free WordPress theme. And it was hard. Building that thing was hard. Getting that thing through WordPress theme review was even harder. They are relentless with with their rules. And, seeing the state of the WordPress, theme repo back in 2019 really broke my heart.

Hendrik Luehrsen [00:24:38]:
It it was really a bad site because it was gamed left and right to the cost of of discoverability for proper good themes. And if you’re a new user coming in seeing the onslaught of bad themes, of course, you’re disheartened by what you see. I think

Maciej Nowak [00:24:59]:
How is it how is it the game?

Hendrik Luehrsen [00:25:01]:
For example, a usual strategy to get a theme, into, like, the the popular, area of themes is to let the theme have a lot of child themes. So, a theme author wanted to, game the, WordPress theme repository, would publish a fairly decent but bland core WordPress theme. I put it out there. And then they would start uploading child themes that did minuscule, like, minimal changes. Change the primary color, change the font, a sans serif font with red, a serif font with blue. So that one core theme had, like, thirty, forty child themes, and child themes had their own review queue. So, not only, took it way longer, to to publish a new core theme, you would also, as as a competitor, see all those child themes going past you and and, like, helping the original theme author, getting attention and, like like, broadening their visibility in somewhat I would consider unethical ways, but it was within the rules of the WordPress theme repo. So, and it certainly did not help, the users.

Maciej Nowak [00:26:39]:
And and if the review was so relentless, why, you know, there were so many bad themes?

Hendrik Luehrsen [00:26:47]:
Because while the theme review was relentless, once you knew the rules, you had your ways to circumvent them. So, basically, themes were only checked on the first submission, I believe. And all subs subsequent submissions were mostly just checked by, the automated tests that back then were not very good, and rarely did or very little time was spent on subsequent reviews. So you could do a lot of shenanigans once your core themes, core theme was, submitted.

Maciej Nowak [00:27:37]:
Okay. So we release a pilot and then, do the actual upgrade with, you know, cutting corners and extended functionalities and child themes. Okay?

Hendrik Luehrsen [00:27:47]:
And we have to it certainly flipped on its head. Being an ugly theme, not being pretty, not being a good theme was not a reason to be denied from WordPress theme ecosystem, from the the WordPress theme repository. So the the the theme review team was in no position, and that that’s nothing they could do, to to, like, judge the visual quality of the theme. They had a set standard of rules. They, had to check boxes for the theme. And if the theme followed that that box, it would go through. And so there was no, like, gut feeling quality control that this is a good theme or not a good theme because it is very hard to define, like, within rules what a good theme means.

Maciej Nowak [00:28:43]:
Mhmm. Okay. And how about now that we have, you know, the times when the time right now when we have Gutenberg and full site editing? Aren’t there themes that go that that would be full site, like, full site editing compatible that, you know, non developers could buy and use and launch their websites in no time?

Hendrik Luehrsen [00:29:12]:
We certainly have a handful of very good, very good themes in the FC ecosystem. Ollie jumps to mind. But, and that that is a big but, we have basically scared away, a lot of the old theme development companies. Back in 2019, so June ago now, we calculated that building a proper, not multipurpose, just the standards of a WordPress theme, building a proper theme with good styling, with good design costs about €25,000, with with development time in Germany. So, we are a little bit higher on the wage side, etcetera, etcetera. So €25,000, back in 2019. When you invest that amount of money, you need a way to recoup that in some way. ThemeForest and back in 2015/14, and pro themes and upsells and, all the nasty things we had, in terms of banners and pop ups back in the day.

Hendrik Luehrsen [00:30:32]:
There was a way to do that. Today, with full site editing and how close the system is and how little, upselling is possible and with no clear, path to revenue, that investment of creating a theme is very hard to justify because it we we scared away the c theme developers with uncertainty of, how themes are being developed because, basically, up until today, we do not have a complete best practice system how to develop themes with full site editing. There there are a million ways to solve a million problems, and they are constantly changing. And there is no clear way to revenue. And I do not, judge anyone for the decision not to participate in that market because, well, it’s economic insanity to participate in that market.

Maciej Nowak [00:31:32]:
Mhmm. So what’s the way? Is it always custom development or using Elementor? Like, because, for example, we are doing only custom design, custom development. Like, it was always from the start. Like, this was because we had those bad, you know, bad experience with, you know, clients coming with existing teams. It was always a mess. There was a reason they were looking for a change, so we were always doing custom development. Now discussion with you about the themes is something like no. Not the the themes were not is are not my cup of tea.

Maciej Nowak [00:32:11]:
Right? And that’s why I want to learn, you know, about the ecosystem and what is available now for customers, at least, for example. So I’m curious to know What are what are the options for clients if you don’t want to go with custom design and custom, you know, Gutenberg development, for example?

Hendrik Luehrsen [00:32:32]:
It’s hard to give an answer for that because, it very much depends on on the position of, you as a freelancer or as agency, and the clients you’re on board. I can answer that question from my point of view. For us, we have been doing custom Gutenberg development since 2018. We were very early on board with blocks. And and, by all criticism, I do about the block editor and how much I, tend to, vocalize my anger about it, I think it is the best content editor we have. It is a pain to develop for it, from time to time. Whenever you see me on on certain Slack instances rambling about it, you must certainly think, that I hate that by a passion, and that is, to some extent, true. But I also think it is the best editor we have, for content editing.

Hendrik Luehrsen [00:33:35]:
That said, in 2022, things changed a bit. We, through our ten year agency lifespan, always had the fortunate position that we were able to choose clients who had the budget to do custom development and custom design. So we could always do some pretty amazing things. We all always had all the freedoms that Colt was able to give us. That kind of went away after, like, the COVID high. There came a COVID low. I think everyone in the industry to some extent felt that. We felt that by bigger agencies suddenly pressuring into our market space from above, we were certainly competing, with agencies for clients we have never competed with because usually our budgets are way too low for them to even wake up in the morning.

Hendrik Luehrsen [00:34:36]:
That said, that, led us to be bold and jump over our own ego, and take a hard look at Elementor. Elementor, for custom development people like me, for the longest time, had a fairly bad reputation. Sites were slow. You were limited in what you were able to do, and it’s very easy to make a mess in Elementor. But, from a standpoint of 2023 and onwards, they have come a long way. The ability of Elementor, to fix issues they have, fairly quickly, is astonishing and certainly fueled by their success, in the market, and and, the size of their community. I mean, when I look at, our, WordPress meetup, I think 80% of the people are Elementor users. So, that’s a new avenue of how we build sites that opened up in the last two years.

Hendrik Luehrsen [00:35:53]:
And we are experimenting with that, and we are getting better with that in the moment. So, basically, to explain, that in short, for the lower end clients, like for the lower third, we’re using Elementor to build out sites and try to do it in a very conscious way in terms of accessibility, performance, and and, because we know how the system works behind the curtains. We know because we program these systems, not in that extent, to us for ourselves. And for the upper two thirds, we do custom development, solely in in in Gutenberg, but not for full site editing. We completely ignore full site editing for now.

Maciej Nowak [00:36:39]:
Mhmm. Why?

Hendrik Luehrsen [00:36:42]:
And because, we had the pain of supporting Gutenberg and block sites for the first few years, which meant constant deprecations. We fixed this data. I mean, for the first year, we had to go into some sites every four weeks to to to every release to, fix some, deprecation or something breaking or, some control is suddenly popping up that was not there before. And that has stabilized since more attention was was, being put on full site editing. And for me, I do not think that full site editing is in a place right now where we have the amount of documentation and and, stability where for a mid market agency like us, where we could risk the investments going into full site editing, because things are still changing, documentation is not there yet, and the the investment is too steep for the the results we’re getting.

Maciej Nowak [00:37:59]:
Mhmm. So so just for our listeners who are listening to this, so they can visualize the problems you had. Like, full site editing versus, you know, building on Gutenberg and those new controls popping up. Right? Like, how how did manifest itself? Like, these problems were like, you you are supporting the rendering on the editor. Right? So that the user could, arrange the content in the editor fully visually. Right? This is what you are referring to.

Hendrik Luehrsen [00:38:32]:
Right? We had, like, a thing, I think it was in 2022, where suddenly content widths were, changing. Just changed because, back in the day, the way content, and and wide align, especially within group blocks was calculated changed. It just changed at some point, leading to all the themes we built beforehand just breaking, until we, I think it was a fix. It was a fairly simple fix, within, theme JSON where we had to opt out of a certain, thing. And there are, issues on on the GitHub repository, the Gutenberg repository, which are, like, dozens and dozens and dozens comments long, about these issues where suddenly things just break. And that is unacceptable if you, are an agency that spends, like, three to six months on a project and then has to move on. We need a stable product, and WordPress was always that. It was a stable product.

Hendrik Luehrsen [00:39:42]:
I mean, for example, we, two years ago, redid a site that was that, we did in 2013, meaning we were able to come back to our own site ten years later. They they kept it running for ten years, and it worked. It worked like a charm. It wasn’t the best UX in the world anymore. It certainly had its edges, but it did what what it was built to do. And, we built we built that in Gutenberg, and was fairly stable because Gutenberg has come a long way now or the block in the house has come a long way now. But I do not think that you can, even go back to an FSE site, like two years ago, and it would look the same now. Just virtually impossible because of the decision that the project is doing for a variety of reason, being the most options, being opt out rather than opt in.

Hendrik Luehrsen [00:40:44]:
So maybe, for example, I can talk about our way how we develop Gutenberg blocks. That might, clear things up. In the beginning, we religiously followed, the documentation, meaning we learned React or at least I had to learn React. And, we built our blocks back in the day before even patterns existed, by using JavaScript, doing, the editing and then the saving, to the database, meaning you had to save function and you basically generated HTML code that you saved to the database. And that was fine until we had, like, a few instances where we had to update, the DOM structure, the HTML structure after that. And the client had the audacity to actually publish content, on on his website. So we had, like, for in one instance, three, four thousand posts we had to update manually because if you update the block, if you update the JavaScript, the generated HTML doesn’t update its itself. You have to write a deprecation.

Hendrik Luehrsen [00:42:05]:
You have to go into the post. You have to hit save, and then it’s done. But they are still, to this day, not an automated way to do that. I know that the big agencies I’m I’m I’m I forgot which one it was, TenUp or Human Made. They tried to to publish, like, a method of doing that, but I think it went nowhere. I I didn’t look at it in the past few years. So then we, solved that problem by, rather unceremoniously going to advanced custom fields because advanced custom fields, solve the problem in a way. We knew you had to write the PHP code once.

Hendrik Luehrsen [00:42:48]:
The PHP dictated, how the site is rendered and if you had to update your DOM because style change, whatever, you could do that. And for everything, every content that has ever been published with an ACF block got, updated instantly. But, the UX was garbage. That’s that was certainly the case. The UX in the back end was suboptimal. We had one case where we had a lot of ACF blocks, on on a post or a certain landing page. We had, like, more than a hundred blocks, ACF blocks. And, back in the day, server side rendering, worked a bit differently where each server side rendered block was one ping, to the, back end server.

Hendrik Luehrsen [00:43:38]:
We were ourselves every time we opened the block, block editor. It was it was a challenge, and that made us rethink. And I think it was discussions with Brian Kortz, where we were talking about how to solve that. And, basically, our current method is, to, write the editor component, like edit JS and everything like that, how it’s supposed to be written, but we save null. We save nothing, or the just the inner blocks. And, we basically take the PHP render and, render everything in PHP, which has the downside of having to write, the the block markup twice, once in React and the second time in PHP. But it gives us the flexibility to change everything whenever we we want without having to worry about applications, without having to worry about existing content in the database. And that is certainly something that lets me sleep very well.

Maciej Nowak [00:44:50]:
Okay. And I don’t get it. Why does it solve the deprecations problem?

Hendrik Luehrsen [00:44:55]:
Because, the block editor has introduced something called the block validator or the the code validators. Meaning, if you, for example, updated to just the edit function that a certain, HTML code is suddenly different or, a class is different or something like that, that block would suddenly be, considered invalid because the rendered, instance of the block in the database is different from how the React application or the editor application expects it to be, and that leads to a conflict that is usually resolved through a concept called deprecations.

Maciej Nowak [00:45:48]:
Alright. I thought I was my mental picture was, deprecated functions that you shouldn’t be using for you know, because they are

Hendrik Luehrsen [00:45:56]:
We also have that plenty, but we are actually talking about a standing concept within the block editor called block deprecations where you basically have an old version of a block that is deprecated. Your new version of the block, recognizes that old version of the block and gracefully upgrades that. That is…

Maciej Nowak [00:46:21]:
Alright.

Hendrik Luehrsen [00:46:23]:
A concept within the block editor.

Maciej Nowak [00:46:25]:
Early in the conversation, we started with the, you started with the with that concept of getting, like, let’s say, time to market of your content. You know? Getting the content to the audience as fast as possible. Now there’s the the very young concept also of vibe coding, and there are certain tools that helps you do that even, you know, free GROG three, which is tends to be, you know, exceptional at at code generation. Where is this going to take us? You know? Speaking broadly about the developer community, agency, you know, space, client space. Obviously, you know, everyone can do this, which means everyone can generate a lot of problems, technical debt, and so on. But at the end of the day, you will have something you’ll be, like, 95% happy within two hours. So

Hendrik Luehrsen [00:47:23]:
Honest answer, I don’t know.

Maciej Nowak [00:47:25]:
Very good answer.

Hendrik Luehrsen [00:47:26]:
We can only we can only speculate because there are some massive things happening in in the broader ecosystem at the moment. And and it is still uncertain how that impacts, us in the WordPress space. First, we’ve seen, we are seeing, solutions already existing in the market, with the projects trying to solve the time to market, or time to content thing. I mean, look at InstaWP, what we Vikas Singhal has has built. I hope I didn’t butcher your name, Vikas. What they have built with with Insta, WP work, you can, I think, go to WPnew and and, you suddenly own a WordPress website at least for two days, until you pay? Or I think Elementor has a similar product. And that is something traditional hosting companies cannot solve because they are in a fundamentally different mind space of selling hosting. They do not sell you a place to publish content, but they sell you a place to put your web application.

Hendrik Luehrsen [00:48:43]:
And that is fundamentally a different headspace, and I have not yet seen a single hoster that approach the things from, like, an InstaWP angle. That’s the first thing of of solving time to market. I think if we can get something like InstaWP, that is even more refined. Well, even InstaWP itself as they refine the process and, become more popular. Maybe that is what Squarespace is for the WordPress system, just a lower end, safe serving solution for WordPress. And once you’ve grown out of InstaWP, they offer, like, migrations to CodeEdit, whatever. Yeah. One of the myriad of hosts we have in the ecosystem.

Hendrik Luehrsen [00:49:36]:
The second thing is vibe coding. I think vibe coding works when you have a general understanding of the basic concepts of code, how code works. I mean, there was a very popular tweet, recently on X, or post on X. I don’t know what to call that.

Maciej Nowak [00:49:57]:
Joel something. Joel something. Joel forty seven.

Hendrik Luehrsen [00:50:01]:
Yeah. Yeah. Where he, like, where, where where his code base was, deleted.

Maciej Nowak [00:50:06]:
Deleted. Sorry. Deleted. Okay. Yeah. Four months of work deleted because he’d never committed to the code repository and cursor wiped out his whole call base.

Hendrik Luehrsen [00:50:18]:
Yeah. Because he had no idea that such a thing existed. Because when you are coming in without general knowledge or overview of the concept, even AI can only help you as much. AI is brilliant in helping you if you know how to phrase the question. But if you don’t know how to phrase the question, you’re on your own. That’s the thing about vibe calling. I think, especially, what we see with Levels.io, like the airplane game and stuff like that, it’s brilliant. But, obviously, Levels is a fairly decent program himself.

Hendrik Luehrsen [00:50:58]:
So from my point of view, I do not think that with the current methods that we have at this point in time, everyone can be a programmer. But I fundamentally believe that a programmer using AI basically has superpowers. It’s it’s like that. And the third thing is where this is going within the WordPress space. A few days ago, we had the CloudFest hackathon where a team I forgot the the name of the concept, developed, like, a WCLI plug in of MCPs, basically.

Maciej Nowak [00:51:41]:
I’m not sure.

Hendrik Luehrsen [00:51:43]:
For the listeners, WCLI is a method of interacting with your WordPress installations through the command line. So you do not have to visit within your browser, but you have the command line, very structured, machine readable, machine writable, and stuff like that. And MCP is a protocol where you can allow large language models like ChatGPT or Claude or, Grok, to talk to an application. And they, on the CloudCheck stack a thon, spent, I think, two days on, large language models being able to talk to WordPress. And I think that has huge upside potential. Being able to, like, do InstaWP on steroids where you have, like, a prompt, give me a WordPress page for architects in Munich, Germany. And the large language models instantly have the proper tools at hand to generate, the WordPress to give you a properly configured WordPress for the German ecosystem with, like, cookie banners, GDPR stuff. Good placeholder content for you because you told them who you are and, like a decent theme.

Hendrik Luehrsen [00:53:02]:
And that that has huge potential.

Maciej Nowak [00:53:07]:
Yeah. I read about this on X as well. I think, Joost de Valk was involved in in in creation. Like, there were more more more authors The usual suspects. The usual suspects. Yeah. The the WordPress mafia or whatever. Yeah.

Maciej Nowak [00:53:25]:
But for also for our listeners that MCP is, like, in broad terms, serves as a layer of communication between LLMs or large language models and all sorts of different applications so that the LLM can consume and interact with different applications. Otherwise, you would have to write on your own code. And if you use ready made MCP, it’s like a plug in method, plug in tool, something like this. Yeah. I think this was missing, Kent. This would mean, in my opinion, that you would be able to, steer your WordPress website, like, manage your WordPress website within the LLM chat window through MCP and WP-CLI. So, finally, the missing component because, from what I know, it’s rather difficult to work with WordPress and LLMs. Like, that like, they don’t talk well to each other yet.

Maciej Nowak [00:54:30]:
Pretty much.

Hendrik Luehrsen [00:54:31]:
I always use, like, I’m I’m a JetJPG user, on on the desktop app on one screen and my WordPress on the other screen, and I’m copy pasting. I have good methods. I I I strongly believe that, LLMs will spark the standardization of content of the web in markdown. That is something I don’t think MET would enjoy because MET’s automatic or the WordPress ecosystem would have much preferred to be the block protocol, to be the standard of the on the Internet, but I think it’ll be marked down. So, yeah, that’s that.

Maciej Nowak [00:55:10]:
But since you are copying from one screen to another, what’s your ID for development? Don’t you use Cursor or or Copilot maybe? What do you use?

Hendrik Luehrsen [00:55:23]:
I use Copilot in Versus Code, but usually for smaller task, rewrite that function, give me documentation for this task, whatever. I started a few weeks ago, using a Python application called Git ingest, which basically summarizes your whole code base in one TXT file and copy that, copy that whole file into, an LLM with an in large enough context window. You need an LLM with a large enough context window window like, ChatGTP3 or high, I think.

Maciej Nowak [00:56:04]:
And Or or Gemini. There there is, like, 2,000,000 tokens context to like, apps like, crazy big.

Hendrik Luehrsen [00:56:13]:
Then you can start asking these LLMs about your code base and suggest changes. And but that’s the way I at least preach what the LLM gives me. It is a suggestion. It is a junior developer handing me code, because that code might not be the most secure one. It might have some flaws, different strategies than what you are using. It is pretty good. I use it often, but you have to treat it with a certain respect.

Maciej Nowak [00:56:52]:
Obviously. Yeah. Trust and verify. Yeah.Alright.

Maciej Nowak [00:57:01]:
Okay. Hendrik, this was a pleasure to to have a chat with you.

Hendrik Luehrsen [00:57:04]:
Thank you for having me.

Maciej Nowak [00:57:05]:
Thank you very much. This is yeah. And maybe maybe we’ll have a, another round one one point in time, and, yeah, thank you once again.

Hendrik Luehrsen [00:57:16]:
Sure. If you want to have it have me again or if someone wants to listen again, my absolute pleasure.

Lector [00:57:22]:
If you like what you’ve just heard, don’t forget to subscribe for more episodes. On the other hand, if you’ve got a question we haven’t answered yet, feel free to reach out to us directly. Just go to osomstudio.com/contact. Thanks for listening, and see you in the next episode of the Osom to know podcast.