KI & WordPress

Music

Jakob: Hello, everybody, and welcome to Greyd Conversations. Today I have I would say two of the smartest people in WordPress, definitely two of the most experienced people in WordPress here. And I feel honored that the two of them take the time and talk to me. We’re going to dive into a topic that’s shaping how we build, how we manage, how we scale websites, and that’s AI in the WordPress ecosystem. We are going to take a look at what is happening as of today, how agencies, customers are adapting or can adapt, and where this might be heading in the next few years. But before we actually get started, there are two awesome people, Jake and James. Actually, just a quick: Jake, James and Jakob – triple J podcast today. Yeah, maybe introduce yourself. So starting with James, maybe you can start telling people who you are, what you’re working on in WordPress, and what your role might be in terms of AI in the WordPress ecosystem.

James: Sure. It’s a pleasure to be here. Thank you for having me. And it’s also awesome to be alongside Jake and Jacob. Three Js. My role in the ecosystem right now is: I’m the engineering director for AI at Automattic, leading the company and building multiple products for WordPress, and also thinking about where the future of our industry could go from that lens. I’m also a co-lead of the core AI team with the WordPress project, thinking a lot more about the enablement of folks building in WordPress and the enablement of the core project.

In terms of what can happen in the future, both technically the project itself, the direction of WordPress, the open web, things like that. Before that, I ran a startup called WP AI – very creative name – and it did AI for WordPress. We did a lot of research around code generation and genetic, systems for WordPress, and that’s kind of where I cut my teeth with the WordPress and AI-specific implementations. I studied at university. I ran an agency in the past for WordPress as well. Also did another startup outside the industry. So I think that’s the 30 second to one minute background about me, and I can dive deeper in the future.

Jakob: Thank you very much. And, Jake, maybe you can introduce yourself, tell the audience who you are, where you come from, what their background is.

Jake: Yeah, sure. So I think, I’m a professional services guy that’s been or I’d say dealing with multiple areas of great change in what digital means to consumers. I like to say I’ve been working on the internet since the internet was built. That’s gone through the mobile revolution. There was a time when we thought animation flash was a revolution. The responsive revolution, I think, now, still thrilled to be in this industry. At the time, we are probably going for our biggest step change in what it means to do digital and be on the internet in the last 25 years, which is, of course, AI and its implications of the next 5, 10, 15 years.

I probably am best known to people in this WordPress ecosystem as the guy that started 10up, the agency back in 2011, bootstrapped that agency to around roughly 300 people in 2023, at which point be merged in with a company called Fueled. That was, I would say what 10up was to WordPress, Fueled sort of to app development, some of like the app product and design space to really round out our skill set. I think a very fortuitously time integration in many ways, because I think that sort of cross-functional capability moving beyond single platforms, to a larger sense of how you serve customers‘ digital needs, even with you still use many of those same platforms.

Like WordPress is quite timely. Since taking on that role I’ve probably worn a bunch of different hat in the last two years. I now just I think for simplicity, so I don’t have to to keep changing titles every six months, I just go by partner, really work closely with our current CEO, the board of directors, and our platform executive team taking on special projects, special interest areas in the company. For me, for the last roughly six months now, I guess, that’s been two very closely related subjects at the moment. One of those subjects is, probably for you to formally give it a function, it’s really been like the storytelling marketing, talk about what we’re doing as a company, talking about it publicly, helping write white papers with our team, who write blog posts, writing case studies about how we’re serving the industry.

I think, insofar as our agency’s value proposition, it’s on top of our new website, fueled.com, has a lot to do with how we use AI to still create very human experiences and drive human-centered design and do truly high caliber work. A lot of the storytelling, a lot of thinking, a lot of my collaboration with the larger engineering strategy and marketing team has to do with how we’re using, applying AI. And I guess the other thing I should say, at least have a shred of competing with my colleague in the industry here in terms of expertise in the space, it’s, I like to say I’ve been kind of into AI since before it was cool.

And back when we just called it machine learning. So we’ve been doing work in that space and a ton of work that I was very personally curious about and tend to be very involved in this, I would say like roughly 2018, as a company going back to the first version of our plugin ClassifAI, which at that time use like IBM Watson to do, you know, language processing and tag recommendations and everything else, and have stayed very close to projects that we do in-house around applications of AI and machine intelligence.

James: And, Jake, an interesting overlap is that my agency used ClassifAI back in the day, and I found it really interesting. And that might have been my first use case with AI and WordPress. And I remember the alt text stuff through the Microsoft services, and I found that really interesting and kind of started piecing things together of, hey, like, there’s a lot that we could be doing here, even back then. So thanks for ClassifAI, I appreciate it.

Jake: That’s really inspiring to hear today. Thank you.

Jakob: Yeah. Wow. That’s really inspiring. Speaking of ClassifAI, can you explain a little bit what you do, what ClassifAI does, where you come from?

Jake: What was the base idea starting like back in 2018 and what it is today maybe. I mean, I think it’s fair to say maybe one of the like since day one guiding principles that defined 10up, that defined how we look at the world, is the idea that the process of creating content on the web, articles and stories, pages, marketing campaigns, whatever you do, publishing content on the web, is that we always wanted, we used to say we like to be easy, maybe even fun.

The idea about being able to tell a story, to be able to publish something, to create, shouldn’t feel like you… I mean, this might sound silly today, but certainly back in 2011, should not feel like you need a computer science degree to do so. It should not require five hours of training. And looking at the manual anymore, that like a word document was like exercise or creating an InDesign layout or something would be back in 2011. That’s why I think on that continue that’s always been a real personal passion of mine, I think it was been a guiding principle, for 10up. And now for all this, how do we make it very easy for our customers over time, make it easier and easier and easier.

So maybe even joyful to go on, be able to create content. And so I think what we started to see for customers, you know, over and over again with huge archives of content tags or mass categories or mass accessibility audits in 2016, 2017, over and over. Nobody remembers the alt text today. It generates itself like the guy with the file name of the image that you uploaded to all that kind of stuff. So when we see those patterns coming up of like, this is a pain point, how can we make this easier and make this more seamless?

We’re always on the lookout for solutions for creative efficiencies to innovate. how we have this problem. So, I think it was that spirit very specifically for ClassifAI, the first feature being that like IBM Watson driven NLP auto categories and tags came from a project for the customer. It was the Saturday Evening Press. It’s been a publisher for decades that had huge archives and had this problem of like, how do we intelligently or how to we go manually or make some stupid like logic decisions, clean up and re categorize all this, and just doing a little bit of research.

At that time, sounds goofy today to think of this player, at that time, IBM was actually doing really interesting work with Watson and again machine learning. So I think that was the initial impetus that we kept looking at, okay, what else does our platform do? And James mentioned it was like the Azure Vision services. Do Amazon have like Polly. But this is before we started rebranding everything AI, we just sort of got more and more into this machine intelligence models and, you know, just kept building incrementally from there.

Jakob: Yeah. What I find really interesting about ClassifAI specifically is, it actually targets an audience that is really professional about that workforce and already knows a lot of stuff that they’re doing, but they’re making it easier for them. And that is something that, at least in my opinion, I think not all of the AI products out there are getting necessarily right at the moment, because currently and I think maybe, James, you can agree on that, most AI products or solutions are mostly targeting that more of a beginner type user, I would say. And even if you take a look at the projects Automattic and WordPress is releasing, is mostly focused on giving you a head start and getting you easily there, so mostly unexperienced users could get somewhere with AI, with this really easy interface of having a conversation. And I think that is really something that ClassifAI does different. Do you have, how did you, how has the feedback been with agencies? Because you, as far as I’m understanding, ClassifAI is not necessarily built for each user or you’re not necessarily building it for the public market to target like this beginner type user I was talking about. How did you incorporate it into the workflow of an actual agency or yourself or your client? How how did that feedback process go?

Jake: I am trying to make sure I follow the question, but maybe I just quickly build in on our previous points. I think you made a good one. I think in some cases AI tends to target very, I don’t even know beginner is the right word, but people that are not, people that don’t have a lot of sophistication in the space, help them become more sophisticated about the space, or they go the other extreme end of the spectrum, which is, to say this maybe in a imprecise and semi cruel way, they’re really targeting the nerdy super early adopters that are extremely technical. People that can go in and even grok what it means to like set up an agentic flow to do like automatic updates to plugins or site changes.

And I think, one of the things I actually kind of say this, one of the things that it’s easy and understandable, we do some of this too, with early technologies, is to confuse what’s sexy with what’s useful and practical in the space. Like so the earliest, like when ChatGPT exploded, we clearly needed to pivot, ClassifAI, deal with that horrible name it has now, that it inherited. We did some early experiments that allowed people publicly do another interesting experiment. It was kind of like, it’s interesting to see, well, what if I could generate an entire Gutenberg page layout from a prompt?

That’s interesting, it’s sexy, it gets attention because it’s harder to kind of like to imagine how do you get from just a prompt to a full page created? But it’s also, especially in those early days, highly impractical in terms of like, you know, use cases where somebody would actually in a flow of trying to getting back to a point of trying to create content or trying to publish something on a magazine where real people that are creating things in WordPress, which again, often not engineers or very technical people, would be ready to make that leap.

So like for us, I think what we see in like enterprise being biased in that direction is in like large publishers and large media. I mean, it’s a couple of things. One is, it’s a bit of work before we run, before we start saying, I just trust this thing to write pages for me, and I can just copy, clean up some copy here and there, it can give me some better or smarter SEO or title suggestions, it can help assist me and guide me and support me. Certainly knock off the plate things that are like the most routine, almost busywork kind of tasks.

Yeah. I see virtually zero interest at this point in that space. Why that is? Stodgy, slow enterprise, people are afraid of their jobs, human nature, whatever the reasons are, I see I would say virtually zero interest at this point in that part of the market. Again, enterprise, larger buyers, larger publishers. And I just wanted to completely do it for me. Where people are, I think doing something closer to that, where a writer for a publication wants to get a first draft co-created, they’re not, let’s be honest, they’re not, and I think we’re pretty far away from this, not doing that in the CMS.

Right. Where they’re doing that, is they are going to a tool specifically dedicated to the writing process, which is a whole other tangent. Like I don’t actually think for like article creation, the creative prompt, the first step in the creative process. The draft happens in the CMS, happens in Google Docs, it happens now in ChatGPT… I mean, from my experience it does not. It happens in Figma.

Sometimes I do think people should get earlier into the CMS like start doing the layouts, start doing the creative. But by that point they’re not looking for „create a page for me“ in most cases. They’re looking for, to go back to your example, make sure the content is accessible and there’s an audio version, and there’s a searchable transcript, and there’s alt text in all the places that there should be and help me make sure the title is optimized. So in that last going, that last line, I also, I don’t know, I’m rambling a little bit here, but I guess I would say, like all of our experience working with customers, it was more about assisting, supporting, providing tooling, providing assistance, providing automation around things that are just routine work in the process.

And I think not, we’ll probably get there at some point, but I think not fronend centering from a practical use case standpoint. Things that are cool, things that are like, kind of mind blowing, but things that are a little impractical in terms of whether somebody is going to adopt them. Like, I just want you to do the whole thing for me. It can be done. Or things that are just so hard to overcome in terms of their technical complexity, in terms of like setting up agents and servers and workflows that, you know, already the barrier is high and ClassifAI to some degree that says you have to go in and opening an AI account or a user account.

James: Jake, I also think, to build on some of what you just said, there’s also the discovery part in enterprise. So with Automattic and through the products that we’re building, we have WordPress VIP, which serves a very similar group to, I think, who you were just mentioning. And from what I’ve seen with these guys, everybody knows that AI is here, everybody knows that it’s very disruptive. But they’re not really looking for this crazy all in one solution because they don’t even know what that could be or how it would work or how it would work within their organization.

What they do love are these little pieces of the puzzle, of the tags, of the whole text, of the stuff that would be annoying and is annoying to have to handle manually, and that’s an enterprise. But I see that in an agency, I see that with individual users, the power users as well, where it’s like, yeah, like I have a good process and I know what I’m doing here, and I can see how a system could eventually evolve to be really beneficial. But I need space to experiment and explore more.

And I think those types of lower complexity implementations of AI through ClassifAI, VIP has a product called Content Helper, similar things, that gives that space in enterprise to do the exploration. And then I think the individual user and the person who you mentioned before is somebody who’s kind of less sophisticated related to the specific software. And I love that group, because they know what they want, and they’re experts in what they want, but they’re not experts in how to do it. That’s kind of a different bucket of folks, where maybe you can play around with more baked end-to-end systems.

But even then, there’s still the requirement to experiment as the producers of these systems and figure out what works, what doesn’t work, what works for the user and the people, and then also like, how does it work within WordPress and how do we actually build and maintain and create it?

Jakob: Very, very interesting. So, James, you probably tend to see a lot of different solutions in and out Automattic in terms of AI. So as I understand, you kind of target the different user groups specifically. Are you building different AI tools for different groups? For the more enterprise experience users, are you targeting those specifically in a different way or with different tools than, for example, like a month ago, WordPress and Automattic releasee a big statement and a big announcement where I think the eternal name was Big Sky, which I if I’m not mistaken essentially it’s a WordPress AI site builder. This, at least from what I’m seeing and what I’m playing around with, targets this user group of being more beginner friendly, more I’m talking to an AI and getting a website out of it. That’s essentially what it’s doing. But you’re talking a lot of these specific tools that we see as well. Are you actually targeting those and building different tools for both of these audiences?

James: Yeah. So Automattic is fun because there are so many different places that it targets through multiple businesses, multiple products and streams of thought. Even so, for the Big Sky, internal name which went external, but the AI website builder for WordPress.com, that is something that is very squarely aimed at the beginner to get people into WordPress and solve the blank page problem that I think everybody sees and addresses in their own ways.

But you end up in WordPress without AI assistance and without a theme that has great onboarding, and the user is kind of left to their own devices to do that discovery. So the thesis there is we can probably alleviate some of that by giving them a website and putting them into this experience, where they can then see how the Gutenberg site editor, post editor GW works with real content related to what they actually want. And that’s one approach for WordPress.com. But WordPress.com, that approach works very well, because that’s who we target there.

For WooCommerce, it’s a completely different group of people. It’s merchants. And those merchants may want to better understand, what’s driving their sales and understand how they can configure their plugins to convert more, for example. And then for VIP, it’s enterprise. So I love my position because I get to see all of it. I get to see the entire spectrum of these solutions and then also work to figure out, well, these are individual solutions and they are targeted at individual people right now.

But what does it look like in one year or two years, five years within the Automattic ecosystem and the WordPress ecosystem in general? And how can we ideally bring these things together? So there’s a pretty cohesive end to end experience, even though it is multiple products, even though it is multiple user groups. So I think there is a question of like, what’s Automattic’s strategy around here? Like kind of pointing at that. For us, it’s this end to end AI experience. We want somebody to end up on one product, assisted with AI through the entire experience of it.

So if you look at WordPress.com, the user gets onboarded, they have the onboarding, but then the site is created and they can work through the site editor with the assistance of AI. They make it to the admin. They can use assistance within the admin as well. And that’s really fun to try and design and think about from the technical perspective, but also the product perspective and even going back to like, what are we doing in terms of these bigger kind of, more cohesive but more opinionated approaches like the AI website builder versus maybe some of these single shop AI features or base quality of life features that solve these problems that have existed in WordPress for a while, like alt text and the media library and things like that. So it’s fun.

Jakob: Yeah, yeah. Thank you very much. Jake, I can imagine, in a huge company like Fueled, there are a lot of specific or different users like you talked about publishers, enterprise companies that you’re obviously working with in certain projects. Is there a similar thing where you see different users interacting at the end of the day with your company and what your company offers and solves, and are you targeting those in a different way, or do you see different types of experience that we need different solutions for? What what is your experience on that? Because I can only imagine in such a large company, such a large user group.

Jake: Yeah, I could speak to that a little bit of, you know, maybe in working my way to a more direct answer and building on a lot of what James just shared with us. I think that, part of what you’re getting at is that the problem of like, who is the audience for WordPress and what do they expect from AI is a much… It’s not a problem unique to Fueld, it’s not a problem unique to even Automattic, which let’s be clear, there serves a much wider range of use cases than Fueled does as an agency and professional services business. WordPress itself has been both, how do I express this, has been both incredibly inspiring as a platform, but also sometimes infuriating to build as a platform because it is uniquely wide in its audience base.

I don’t think you can point to, I challenge myself on this, there are some that are emerging, but there are not many CMS website, I oversimplify, I know it’s more than this, website building platforms on the internet that go from someone’s little blog that they pay $0 a month for on WordPress, powering WordPress.com to whitehouse.gov, and huge e-commerce, thousands of news stores. It’s really incredible in a lot of respect to the incredible testament, to open source, to the openness of the platform, to the, you know, all the things we can get into that could ever even reach 40% market share is a one size fits all solution.

It also is a challenge for us, both those of us providing companies like Automattic and companies like Google coming at it from different directions, but also for the entire platform, the open source project, because I think, you know, if you are Sanity, if you are even Webflow, which kind of broad in what it does, if you are these other platforms, I think you ahve a pretty crisp identity of who ends up, who your audience is, what part of, what slice of the market you’re really trying to target and do well with it.

I think it provides a certain clarity for the kind of solutions you really need to focus on building. Where I think WordPress has this challenge, has almost this, you know, continuing identity crisis would be melodramatic, but identity challenge of like, is the priority, Joe for a site builder who never wants to go into code and build an entire site, is the priority of the people that are building the biggest sites in the world for NASA, for other companies that don’t want to have a LLM build their entire site.

They want assistance. They want automation. They want workflow. Is it for high end developers that want to plug into an API in services to be able to customize features and do really unique things in WordPress? I’m rambling a little bit, but I think, yeah, you’re reinforcing in a different way, but it is a unique challenge in our industry. Fueled specifically is mostly focused on what we would call, it’s an oversimplified term, but the enterprise buyer. Yeah. You know, now, to be clear, there was a lot of good technology we can develop for the enterprise buyer that has huge appeal as we can make it more simple, as we can make it more approachable for a much larger audience.

Then we do have some solutions. We serve some smaller customers to our site watch program. We do some small quick get in and get out engagements. We have an Elastic press product, which is doing interesting things with AI, which is more small to medium. But let’s be clear that the dominant focus we have here are larger buyers. People that are not looking for AI to just do it all and be done, people that are willing to pay for it. They have the budget to pay for professional services. There’s a couple of things there.

One is their expectation across the board is, as we’ve discussed that already in this conversation, how does AI bring some efficiency within their workflow in business as the customer, which is to say, you know, I think it’s a stupid example to keep using, but like, can I get better titles faster or can I make sure I’m not breaking accessibility, not to be manual chore or to bring efficiency to their process? They want to know how to use these technologies to better their products. So not the efficiency play, but the do new things, interesting things, we just built a whole like improved search experience for a part of Starbucks that uses AI, right?

So that’s the innovate, do more creative things, offer more options, which is exciting for us. It’s new territory to trap. You can’t just plug in to vibe coding at this point. And then there’s a third category, which is, I think maybe the biggest challenges for a business like ours, which is frankly, not the how do customer directly take advantage of it for their products in their workflows, but frankly, how does the art of how you how we efficiently build a website and stay competitive change, which is to say, probably the most, it doesn’t almost doesn’t feel exaggerated to call it existential AI question for us.

It’s not what we’re building in ClassifAI. It’s what does it cost? And how do you build a website in 3 or 4 or 5 years? And how do you take a 300 person team to get there and get them from here to there? It’s just building engineering, tooling, designing our processes, that are the most disruptive.

Jakob: Yeah, that’s actually a great topic and great introduction to… We talking about a lot of solutions that are targeting customers of your companies, for example. But what about internal things? So as far as I understand, everybody of you uses AI on a day to day basis, I would guess at least. What are you doing Jake, internally to make things more efficient at the end of the day, to deliver such projects? Are there tools that your are using, are there workflows that you’re developing for those teams, or how do you approach that?

Jake: Yeah. It’s multi-pronged. And frankly, I just got back from a two and a half week vacation, so I feel like I might be out of date again back with the team, three weeks later. But I think there’s a few prompts to it. So one is absolutely building tools, building processes, building workflows for things that are repeating, challenging problems. A good example of this would be, we just published about this right before I went on break, I feel almost a little bit embarrassed, we vibe coded a great tool that we never got around to before because it felt like it was arduous and tedious to build, which is, you can read about it on fueled.com on our blog, which is, automating going from Figma styles and tokens to a WordPress theme.json It’s such an example.

It’s like it hit so many checkboxes because it’s in a not circular, but like in a perfect example, it was both an example of with the right culture, with the right push for our team, with the right go do this, the reward, the recognition, the encouragement to go build these kind of tools is an example of how we could build things that help build things faster and more effectively in some cases. And it is an example of where we now have a tool because we’re able to do so that lets us streamline our processes more quickly.

That’s part of it. Part of it is choosing, like trying to rally around what collectively with 150+ engineers, what are the platforms that we think make the most sense? So if we sort of start, you know, roll back the clock nine months, there’s a bit of a let a thousand flowers bloom. Go try whatever AI platforms you want. Be careful about PII going in and things that are truly confidential, but otherwise just try different platforms. Now I think we’re going back a couple of months, we are really at the use one of these. Right, maybe a little bit more room, but these two, we’re pretty sure based on our experiments, use these tools.

Here’s some examples of prompts, there’s education narrowing the tool set, building new tools to help the team work more efficiently. And I think, the way that I look at it, the other, the last prong is and I alluded to this, is cultural and systems and incentives within the business to get people excited about it. And some of that is easier there. I’m, you know, I’m curious, I imagine probably similar things in both of your businesses. Some of that is creating a culture of celebration and recognition of and reward for those that embrace it. Just even, bluntly, maybe, you know, a bit of like friendly inter-engineer competition that you kind of want to build within the company to show that you’re getting ahead on this platform.

So as cheesy as this example might sound, we have a channel in the company where, you know, what we ask is, hey, if you save a little time on a project, first of all, go enjoy that afternoon and second of all, all we’re asking is go in there, tell us what you did. Point to the repo. If it’s something anybody in the company can look at and our leadership will rally around that, celebrate that. And it really starts to create not only a culture of like, we’re not afraid of this, we’re recognizing this, rewarding this.

We’re good to take risks, good to try some new things within the business. But it also, again, it creates a little bit of like an atmosphere and a culture of like, maybe I need to be right on top of this as well. The other thing is, we need to have systems that like, how do we get very, kind of psychological and like economic about the way I think about these issues. I think there can easily be a pattern that inside of a company, like if you’re not a little boutique, everybody has a share of the businesses.

When you’re more skilled business, which is sort of a, what to call bluntly, what’s in it for the individual engineers to resist? It is a sense of like, great, if I really embrace this, who am I? Cool way of saying, who am I enriching in doing so? Is it we, am I getting better in this process? Or is it just my management? Yeah. There’s some fear involved as well. And I’m going to be rewarded by saying great, you’re able to do that thing. And half the time now years of the work right to get done in that process. So I think there’s a challenge in business, especially of scale. I mean, this speaks to the larger culture of the universe at the moment of like getting a large team on board to unlock its efficiency and creativity rather than afraid to do so, rather than doing like, in a sense, like in the no good deed goes unpunished kind of way, that they be punished, right, for being able to do so.

So I think there’s a bunch of like systems and incentives you create that range from Jake’s going to go write a story and help me to talk about all the work that I did publicly on LinkedIn and on the blog and everything to literally just things like, hey, if we got this budget for it, we got this time allocation for the project. If you can do it faster, fine by us. We’re not going to just throw more work on your plate. We’re going to get some of that time back. So I’m again, I’m droning on, but I think I can’t ask enough for scale businesses like the cultural business economic incentives is being key to helping make that change. I have no doubt that like most of our engineers, certainly our best ones, are capable of making this transition or capable about embracing this technology.It’s not the skillset that’s the problem. It’s the incentive structure.

Jakob: Yeah. Leaning into that transition and really embracing it, I would say is the key factor. Actually, shout out to the theme.json converter. I already tried it out and I really liked it. Theme.json converter is probably the wrong wording, but I saw the post and I tried the tool and really liked it. And that’s actually similar to a couple of the trends we’re seeing as well with our clients, who are agencies as well. They’re quietly building internal AI tools.

So specifically do a thing that they usually did manually. Now AI is kind of automating that instead of, and that’s maybe a shift that I’m seeing, or we are at least experiencing, instead of buying little SaaS tools to help you, they’re just quietly building something that specifically does that for them, which is then really individual but really embraces their workflow. Which obviously works great for them and their tooling. What I find curious, getting back to you, James, is, all these internal tool things and I think, both of you can agree on in general, there are a lot of little tools that people write and a lot of different workflows they’re creating based on AI now. And they’re mostly maybe in Ireland and they’re having their own workflows. How do you get all these little workflows to prompts and solutions people build into? Kind of like a general, as you were speaking about this vision of AI in WordPress, getting one streamlined AI in there, how do you tackle this in these different projects? And like companies, I would say with different solutions for different audience and get that into one core idea? How do you do that?

James: It’s fascinating. It is very interesting. And I’ll answer it from an internal perspective first, and then focusing on kind of what we would like to present to customers in 1 to 3 years from Automattic. So first of all the internal: like Jake mentioned, there are so many opportunities and in terms of the overarching business, if a company is able to adopt AI effectively, we’re in a really good spot. So this is something that when I came into Automattic, I saw a lot of opportunity for this type of internal tooling and enablement of the organization and celebrating the usage of AI, as opposed to trying to like, move away from it and stay away from it. Just because that’s where everything’s going.

And the three major things that Jake mentioned, the tools, the culture and like celebrating enablement, are things that we’re also doing at Automattic. But I would also assume that many organizations in a large scale or even agencies are doing really similar things where, it’s like, how do we use AI? How do we make sure that we’re the company adopting this stuff so we’re not the ones left behind when the other companies go and adopt this stuff? Two additional things, just internal that we’ve done as well is, create guidelines so people can trust that they are able to do something and able to explore and create these little one off prompts and solutions and look at a tool and say, yeah, I can use this.

And here’s the guideline that kind of explains, oh, like, we’re not going to use personal information. And maybe for this client, everything is excluded. All of that. And from that they can trust that they can go and build and experiment and explore. Then if they create really interesting tooling and, this happens kind of throughout Automattic as the organization, it can be surfaced and it can be more visible and it can be showcased. And if that makes sense, to go and bring to the wider organization. Automattic specifically has an internal AI team, to go and do that.

But again, if you’re able to surface this, then especially in technical organizations, you don’t even need the team. You can just go and kind of get people to continually adopt and enhance these solutions, especially if you have those guidelines and the trust that people need for that. So you’re giving them guidelines and then you are giving them the platform to present it internally into your company. And then you are deciding whether you want to create something? Yes.

And the final bit of things, and this isn’t as baked as I would like it to be, and it’s going to be more baked over time, is we’re biasing, as Automattic, the company, towards a similar approach to what we’re doing in WordPress core with the core AI, where me as the core AI lead or one of the leads in Automattic. We are creating building blocks for our developers and even our non developers who now have access to ChatGPT and can vibe code a bunch of different things.

We’re giving them building blocks to go and build these solutions and kind of the roadmap of, okay, this is an internal framework. This allows you to go and do a genetic loop. This allows you to go and register tools. You can go and expose it into WordPress. In this way you can go and pull information in this unified way. So now we have people who not only are celebrated because they build these tools, not only do they understand what they can build and what data they can use. And, a way to review the solutions that they want to and have created. But they also now are having a increasingly baked and growing toolkit and tool box to go and build these things with. And I’ve found that anywhere.

And you don’t need to be building your own toolkits, because these things do exist and there are many open source libraries. And in some organizations it makes sense to do it yourself and others, in the majority of others, I’d say it makes sense to go and point at something. I think that’s kind of like the perfect storm to enable companies to move more agile, to move in a more agile way. And again, like, keep up with the other companies that are doing the exact same thing. So that’s the internal bit. I can transition to external. But I saw Jake, you shake your head on something, and I wanted to see if there was anything you wanted to…

Jake: No, I’m 100% with you. Okay. I was going to comment on its applications to external, but, your role.

James: Okay. I’ll mention, I think there’s an interesting segway and overlap with that internal stuff within Automattic kind of flowing into our products within Automattic. So like I mentioned before, there’s kind of this overarching vision of, we would like everything to work together. We would like this to be an end to end experience. It may be multiple products, it may be multiple services, but it can feel the same if all installed in the same environment. That’s something that is a very big vision and it’s something that takes a lot of work to get to.

It’s also interesting because given how wide of a demographic Automattic serves with all of the products, I think if we were to build this unified thing outside of AI, that might not be a good vision, because of the individual demographics of like, this is enterprise, this is not enterprise. Which I can kind of scale up and down in not the most completely adaptive way, but it can scale up and down to say, I’m going to service this user who is a beginner, who doesn’t know how to code.

And I’m going to service this user who is a developer who does know how to code. And you can kind of build systems that can be deployed into different contexts and tweaked a little bit to go and be more effective for the demographic that they serve, sharing the same underlying stuff. So with this general vision of AI, I think it is possible to go and create a system that adapts to the users and spans across multiple products in a cohesive way. And the way we want to build that within Automattic is, have these multiple product groups target these individual demographics, build it all using the same fundamental building blocks and pieces, some of it being within Automattic, some of it being hopefully what comes to WordPress core, doing so in a unified way, and eventually result in this really, really impressive end to end system.

So again, it’s like have the building blocks, we do have multiple teams focusing on multiple products building in the same way. And that allows us to kind of think in this overarching direction as opposed to, oh, this is one product for this one group, and we’re going to build it in this way, and it’s not reusable anywhere else. And that flows into core AI, which I can get into now, but I can also yield back to you, Jakob to see where you want to go.

Jakob: Yeah. No, that’s very, very interesting. I could go on for hours or I would go on listening to you for hours, to be honest. But I want to shake things up a little bit right now, and I want to jump into a section I just quickly called, I came up with „hot or not“, where I am gonna share a trend or a buzz word. And I want you to say if you think that’s hot or not. And whatever hot or not means for you. So essentially, if it’s here to stay, if you think that’s good, if you think that’s bad, that’s solely up to you. So just simply say hot or not, and then just give maybe a quick reason why you’re thinking that way.

James: Yeah.

Jakob: And I’m just going to go for it and I’m going to target one of you specifically. And then if the other one wants to share a thought on it, feel free to, but you don’t necessarily have to. So, James, first off, we talked about it. Vibe coding – hot or not?

James: Very hot. I think it’s awesome if you do it in the right way, in the right context. So I see a lot of folks who are developers and really good and tenured and have been doing this for a long time, using this new way of creating, to experiment and explore and create MVPs. And that is very interesting from a personal perspective in terms of being expressive and creating things that you couldn’t really have the time to ever do in the past. And it’s also very interesting from like an exploration perspective from an organization. We see product managers creating not in Figma anymore, but in code.

So I think it’s very hot. But you have to do it right. You can’t be just willy nilly using this device to go and build things that end up on hundreds of millions of sites. So you need to be smart.

Jakob: Yeah, Jake, I see you nod your head.

Jake: I agree. Very good.

Jakob: Okay, Jake, one thing for you, AI written articles.

Jake: Hot.

Jakob: Hot? Why?

Jake: I mean, I think they’re just not very different. I mean, despite coding for writing, right? It’s effectively the same answer, which is, people that know how to write, people that know the story they want to tell, is a tool to skip some of the, like fast forward, accelerate some of like the mechanics of the syntax and the prose shaping to be able to more efficiently express their ideas, maybe even express their ideas better, I think is only good. I would say, you know, this focus on content that I’ve had, I’m able to do this now, in about 24, 30 hours a week on average now, with one person helping me.

We’re able to do that. We don’t, to be clear, like just shove it in with a short prompt and publish on the web. You’re going to get, you know, garbage or slop, I guess, is the term of the year. But as a tool to, like rapidly accelerate, help more effectively communicate, should the LLM that the voice need more naturally without such manual effort feel consistent? It’s fantastic. We’re not, it’s not the storyteller. It’s not the person that knows what the story is, what the idea is, not the final cut, but as an accelerant.

Jakob: Yeah for sure. Hot. All right. Thank you very much, James. AI generated logos.

James: Not. No. I actually, so this is a very personal opinion, but I don’t really like AI generated art that much. And I find that the simpler you try to make these generation specifically around logos, they all have the same look and feel. And I really like a logo being an expression of a business. And I don’t think AI gets you there specifically on logos. That’s a funny question because we didn’t coordinate this, and I was just having a very detailed discussion with some of the designers in Automattic in the New York office about this. And yeah, I love the human thought and creativity and process that goes into logo design specifically, but art in general. And that’s not like anti all AI art, but with logos it’s anti AI generated logos. And that’s a again a personal opinion.

Jakob: Yeah, that’s what I’m asking for. Yeah. All right. Thank you. Jake, AI generated client proposals. Have you done that? What do you think about it?

Jake: In my binary choice here, I might have to say hot again. Again, I probably keep giving the same answer, like trying to say, you know, AI autonomously responding to them right now – no. Again, for all of these things as a tool to help accelerate the craft, to help shortcut things and or just if we, if we’re being honest with ourselves, are just grunt labor tasks and not actually creative tasks. Yeah. I mean, I know it’s happening across some agencies as an assistive tool. I think every agency that’s not doing it is thinking about how can I take some repetitive tasks out of this process.

I know I did this thing or some permutation of this thing over and over and over again for the last ten proposals. How do I not spend my day doing what that, you know, how can I take some repetitive tasks out of this process. the rhetorical point, like the blue collar worker just typing it again right to the screen, yeah.

Jakob: But would you go ahead and brief a project and say, give me an offer and give me a proposal for it?

Jake: As of the 0.5, to say like pull in the right case studies… Right. You know, the section on how we do QA, I don’t need to write that for the 700th time. As a permutation that we edit, clean up, augment, bring creativity in. Yeah. But I mean, to be honest with you, I don’t think any of these AI things, with maybe the exception of something like stock art or something like that, are ready to be, you know, are ready to be, press a button, then you’re done.

Jakob: Yeah. Actually, perfect overlap to the next one. James, I’m giving you AI memes.

James: AI memes. I do like AI memes, and I like, I, I mean, especially when you can tell that there was, like, a lot of intention that went into creating this thing. And it’s like, you can kind of extract the humor coming from the person that has created this thing. And I’ve seen a couple that are just absolutely hilarious. And it’s like either a remix of one of these classic memes and you can kind of like get the wit that somebody infused into it. I think it’s very funny.

Jakob: Interesting. You just said you don’t like AI art, but AI memes, it’s different.

James: Well, I think my common thread is that, and Jake has kind of enunciated this in some of his answers. It’s like when you can use AI to kind of express yourself or be more effective and quicker, but you still have the human element and the human connection. So it goes back to the art, it goes back to the memes. Like, I don’t like memes that are just like a remake, a mash of a bunch of random things that aren’t funny.

But like when you can tell that somebody put thought into this thing. There’s also like funny video generations now, and it’s just like, oh, like somebody wrote that script and this AI character has said it, but he’s like, say it in this way. It’s hilarious. It’s very fun, and it’s something that you would never unlock without AI.

Jake: I think the way that I think about it, shut me up if I am breaking your hot or not.

Jakob: No. No, that’s the intention of this.

Jake: The way that I think that it is like, I see this as a good way to express it. I always loved that metaphor of, like, I know, like, cliche tech nerd, the Steve Jobs quote, but the Steve Jobs quote of the computer is the bicycle for the mind. And I have sort of taken to my own head on this very publicly, very often that I think of AI is like the motorcycle for the mind from there, where it’s what I think, what is powerful about AI, where it is most, at least today, transformative, is not replacing human creativity.

It’s the places where it can accelerate human activity because it takes the manual work. You have an idea, to the meme thing, you have a really creative idea or meme in your head. You may not have the tools or it would take you four hours to actually go open up Photoshop, right and craft it, but if it helps you tilt it, accelerate how fast you can express that idea. Which of the computer and Photoshop, you know, cut the time down by 70%, you know, bring out your stencil set right, tracing things out with an eraser.

And if I can take that down even further, I think when we think about something like a logo, I really share all of James’s perspective. I could, maybe some designers would slap me around for this, I think the reason that feels like that falls into the not camps, because for a good creative designer, getting from idea to the expression of an idea doesn’t feel like a herculean task when you’re talking about a logo in the same way that like, you know, how to say it, like crafting all the panels of a meme like obnoxiously tedious compared to what you’re trying to express.

James: I think also Jake, my thing with the logo is that, a lot of the work around logo design specifically is, and this is also with web design and actually a lot of creation, it’s not the actual act of creating these things. It’s really interfacing with the client and figuring out who they are and what they want and how they want it, and going back and forth and back and forth and infusing real artistic direction and talent into this. So that’s a funny question, because the logo stuff is stuff I have deeply thought about, which is just an interesting thing.

Jake: Yeah, I think that one we’re also brushing up against, like it’s still, it’s incredibly impressive. But I also think if we’re comparing image generation and video generation to where like, frankly, even code or text generation is right now, it’s not quite there in the same way. Yeah. Yeah, it still feels very… I think of those like, you know, embracing the if the photograph becomes a Picasso print when you run it through it. There’s something about the way these language models work that just doesn’t quite feel as mature, but you can sort of see a version of the logo creation, where again, I think where we would agree is like, here is my company name, make a logo. That’s not, that’s robbing us of human creativity and exploration. There are some, I will say, if you saw me like sort of smirking, there is some joke in here that humans also managed to make every logo at the same, without AI.

James: Yes. Well, I also bucket the Fiverr logos versus the, major ad agency logos and things like that.

Jake: Sure. There I would be a little…

James: Yes, yes.

Jake: But could I see a process where if, like, these AIs got a little bit better about, like, really being able to refine and engage with, visual images in the same way that they can with copy and code right now, where you could say, here’s what I’ve learned about the company. Here are five inspirational logos. Here’s kind of the direction I’m thinking of. And I’m not a designer, but it’s colorful or it’s bubbly, or here are ideas I feed into it and be able to bring back to you ten different things to react to and keep shaping.

Again, AI assisting the creative process, not replacing the creative process as a person. I’d still be excited about that. I would be surprised. I would get a lot of like, very good designers are, whether they’re telling us or not, are maybe even starting today, that process of how do I get to a logo with some interactivity, back and forth collaboration or acceleration. So I think the point… I’m rambling. The point that I’m trying to make is accelerant of human creativity as opposed to replacement for human creativity is, I think, the way that I…

James: Me too.

Jake: …at least today, the way that I process it.

Jakob: Yeah. You guys made the perfect transition from this hot or not to the next topic, because that is, I think, you are tackling one of the most important topics, because there’s a big fears out there that AI might make certain services, certain agency services or agencies entirely obsolete at some point. But I don’t think that’s necessarily true. Or I think the answer is more nuanced. As you said it already, I like the metaphor of having this as a motorcycle of the brain. So obviously, one of the most quoted quotes out there is AI won’t replace humans, but humans with AI will replace humans without AI. And I think that tends to be exactly what you’re getting to. If we’re existing, if we’re making things more efficient, if we’re shortcutting things, that is actually something that is really good. But if we’re trying to replace creativity or replace understanding the client, understanding the goals, I think that then it’s not really feasible. Or is that the point of what we are trying to do? Is that how you see it as well, in terms of will AI replace agencies or agency services? Is that how you see it?

Jake: That’s a tough question. In the short term, yes. That’s how I see it. I think that’s where we are today. I think that’s reality in terms of what these LLMs actually do. I think that’s, yeah, I’m sorry, I’m rambling. I think what’s hard for this is like, it’s like I’m trying to parse ideology from practical application. Like, ideologically, it’s very difficult for me to say I can be excited about a future regime where AI replaces you. I don’t know if that is because it feels like we’re basically just in…

Jakob: Do you think that AI would replace humans in some parts?

Jake: Yeah, I think so. Right now I don’t think… It’s a two-way station that question. Personally, ideologically, I hope we don’t get to a place where for the most creative pursuits and endeavors, AI replaces the human spark of creativity as opposed to rapidly accelerating and helping us bring creativity to life much faster. That’s ideological. As a statement. Practically as a statement, fortunately perhaps, I don’t think AIs today do that effectively. To the logo point. I don’t really think… I’ve might be a little too simplistic. I don’t think if you’re looking for something truly original or creative that they really do that. Yet to the extent that you want to, like, really rule something out for your business or a final product, right, or something like that. It’s probably a debate that what’s harder is when you look out 3, 4 or 5 years, right, in this ecosystem, if you asked me, like for, not for everyone, for 70% of people that need to throw up a marketing site or, a creative looking story about their restaurant with a menu, whether AIs can do that. And I mean, yeah, they probably can. I feel like I’m rambling here, because I’m thinking about your question.

Yeah, but ideology, we could then go on into the discussion of what actual true creation is and what makes things truly creative. Is t the result of all the things you experience during your life, or is it something truly unique? Most art is derivative. Most art. Yeah. Because that is that is what I’m thinking is really interesting. If you would have asked, most people in the streets a couple of years ago, or actually maybe even today, what are the jobs that AI is going to replace or like differentiate what shape the fastest and the easiest, you’re going to say, okay, these low skilled jobs. But the reality actually is that jobs, like developing jobs, creative jobs, are tackled a lot by the services of AI.

Because, for example, as you said already, code generation is kind of already there in a lot of places and already usable while image generation and really like for example, tasks that you need to have a thumb to do that task, the AI cannot really do all the… Well, the problem, and again, you can go really deep on this and go around circles, the problem for most agencies, certainly agencies that look like a Fueled, 10up kind of agency is that, how to say this, like creativity, your innovation, your creativity, your business problem solving is the 10% of the work you do that differentiates you to sell the rest of it. Which is to say, like most of the work that agencies and our service, it’s not that it’s low skill, whatever the hell that means anymore, low skill labor, it’s that it’s manual labor that you’re doing that’s being replaced 90%.

And don’t take 90% too literally. But to me, like 90% of the work that we actually do and bill hours right for, have our time compensated for, is production. It’s moving the pixels on the page. If you’re a designer, it’s typing the code into the editor and deploying it. Yeah. So I think there’s no question at this point that that model is already being radically disrupted, will be more radically disrupted. I think there’s, you know, going off the rails here, there’s a question for agencies like ours of, okay, that 10% that is unique and not just go do the thing or build it, go code it, go move the pixels, go move the code. The creativity, the what’s the strategy, the how do you approach this problem?

Do we just do more of that and fill up the same budgets with that? Does that just cost more? Beginning with the business model side of it. But as a buyer say, again, picking a random number, a buyer that would spend $100,000 on a website. Is the future in five years, that I now only am going to spend $5,000 on my website or $1,000?

Jakob: Or do I get more value out of it, right?

Jake: Right, or do they say no? That’s still what it costs. But this is going to be 20 x.

Jakob: Yeah.

Jake: In terms of like features, I could have creative application art that will create on the website a better logo, right to torture that example that I would have gotten before. I think that’s the harder question that, you know, I could argue now, like half glass full half glass empty. And it’s probably somewhere in between. It’s probably not going to go down to I’ll spend 20 bucks on ChatGPT to build a website, but it’s also probably not cost what it did five months ago.

Jakob: To make this a bit more concrete. James, maybe you can refer on that from your past days, but even from your present: Are there parts of agency work that or what parts of agency work, in your view, most likely is to be automated? And parts are most likely not to be automated as of right now, maybe also in the future?

James: Yeah. Well, something I was thinking about as you and Jake were talking was, if I took my agency isotropic from 2018 and dropped it, even into today, but definitely like two years down the road, it would not succeed. It wouldn’t work. And it was a very successful agency. And I think that’s a good example to kind of underscore: If you as an agency or professional in this industry, don’t understand that there is disruption that comes from AI, and don’t best position yourself to make sure that you’re on the right side of it. Then there is definite risk that there will be negative impacts to your business and the way you do things in the process.

And I don’t think you can really sugarcoat it, and you have to be realistic about that. So if I dropped my agency into today’s day and age, I think the thing that would be most impacted would be the actual process of running this company. And there’s a lot of efficiencies to unlock from AI in terms of maybe, for example, drafting the client proposals. If we’re spending, I don’t know, five hours on this one, we could be spending one hour on this and then using the rest of the four to go and prospect and sell. That’s not good.

And other agencies now are doing that, and we’re not going to be able to compete that. A big thesis of mine when I founded WP AI, which the first product was called Code WP and it made code, was, hey, when I ran this agency, we built large B2C WooCommerce sites. We built a lot of them. While they might have served different industries, I always found us developing 80% similar plugins to glue together a lot of WooCommerce extensions and functionality and features. So the thesis there with Code WP, and it seemed to be right from the usage that we had, was hey, this repetitive coding work, this almost busy coding work is something that I can disrupt with AI and I personally would want this to be disrupted.

So then I can go and focus on the architectural design of these WooCommerce stores, make the better decisions there, maybe build out more impressive solutions that are more boutique for that client, if that makes sense. So that’s what got me back into WordPress and that’s what the thesis of of WP AI was. Let’s go and solve some of these low lift things that will be disrupted if I do it or if somebody else does it, it will be disrupted. Let’s do something specific for the WordPress community, where I know there’s probably not going to be many other startups doing it in WordPress. And create a tool that is better than generalist things like ChatGPT.

So that was the idea there, and the usage and the adoption and the exponential curve of growth that we had definitely proved that that was at least somewhat right, and that many agencies were definitely automating away some of this coding busywork. And eventually we started working on an agent where it was like, this is not going to go and run your entire WordPress site, but it’s going to help with the automatic plugin updates, for example, or like basic scheduled maintenance tasks. So all of that to say, the products got a lot of use.

This is what is most disrupted right now. If you zoom out and think about the future as an agency owner, anybody in this industry, you need to be realistic that things will change and you just need to kind of go at it with an open mind. And also a glass half full mentality of things will change. How can I adopt these tools and systems and ideas to be the most effective version of myself? And then finally I’ll close on how do I also contribute to the future that I want. So I would assume that we all want human creativity value to go up as the price to create code goes down. How do we go and make that thing happen? This is something that I also think about with WordPress itself through the core AI position.

I want WordPress to exist in ten years. And how do we make it exist in ten years? We create the systems that other developers need to build the experiences that people expect in the coming century, and enable people to do that and steer in the right way. So like, not pretend this isn’t happening, not take action and just be like, it’s going to be fine. Understand that there is disruption and then steer towards that disruption. Embrace it and and try to create what we want for the future.

Jakob: Yeah. I think that’s a very, very valuable lesson or like take away. Jake, do you want to add anything there as we’re kind of at the end of this episode? Is there a takeaway or a tip for agencies like going into the future that they should be embracing a certain thing, that you might recommend to them?

Jake: No. I mean, think on the agency specific point. I think we’ve touched on that earlier. It’s a don’t just focus on the technologies, focus on building the culture. Focus on making sure that your engineers are unlocked to feel like they can get ahead of that. We have a unique and precious moment for hell knows how long, a year, maybe two years, where it’s not so dramatically changing, so your business model is broken overnight. So you have a moment to start changing the way you work and adapt and adjust. But you have to move swiftly. You have to move purposefully. And focus on culture or focus on providing tools and resources. Focus on efficiency. Focus on coaching, creativity.

Disabuse yourself of any notion that you’re value is going to be the just mechanical work of job as an agency in a few years. There’s one point about where I think core is going and where WordPress is going as a platform, maybe building on the last point here that James made, which is, two things in terms of rallying around WordPress core, in terms of where it goes. I think it speaks to, in my opinion, what are the two, facets of WordPress that unlocked its growth over its entire trajectory? Those are two things. One is for the baseline, what 60 to 70% of all people who use CMS software need to do. Let’s rally around shared built-in functionality rather than having the thousand people trying to solve the same problem. So very baseline stuff, probably many things that are in ClassifAI or other solutions today becoming just basic hygiene for a product that’s about easily creating and publishing content. Alt text, text for, you know, smarter proofreading, proofing collaboration on a document, image generation, those kind of tools.

And then I think James hinted at this, we maybe said this in a way that I know what he’s saying. Maybe we can make it even more direct, which is if the first thing WordPress does is be very good for the baseline things, for just easily creating content. The second thing I think unlocked WordPress‘ growth was being a very flexible, agile, open development enabled platform that left thousands, if not tens of thousands of developers create integrations, build a stack on top of it to solve all kinds of different problems. So creating the tooling infrastructure inside the platform to be able to build out new AI features, be able to plug more advanced functionality into the system.

This is a very crude analogy, but I kind of think of it as like if in some ways look at the introduction of custom post types into WordPress. Not that WordPress needs to do this on its own, create immediately a shopping cart of product like shopping content type or a calendar event. What it needed to do was create a framework that would let developers who wanted to create all kinds of content to be able to very easily plug in without having to, like, you know, frameworks, APIs, without having to torture the software. We need the equivalent of that for AI in WordPress, a framework to make it very easy for developers to open up all the different niche last 20%, last 10%, last 1% market kind of use cases. And so, I think that your earlier question rallying around adapting core, table stakes, hygiene for publishing and creating content standard way versus everybody having to go out and figure out how to add tires to their car in a few years.

And second, all the development frameworks, the tooling, the under the hood stuff that will let a rich ecosystem of developers safely, efficiently build out on top of WordPress for tons of other applications that we won’t we can’t even imagine. Certainly not just the core teams going to imagine.

Jakob: Yeah. All right, guys, thank you very, very much. I could honestly go on for hours and hours listening to you two and getting your opinion on certain things. I have hundreds of questions that I didn’t even ask here, but to close this out, that was a blast. Thank you very, very much. I really, really appreciate it. You taking the time and sharing your thoughts and your experiences on AI, on that very much evolving, topic and reality, but future as well. Yeah. Thank you very much for being here. I just quickly wanna mention to the audience that you can find us on YouTube, on our website and check it out. Next episodes are coming in probably a couple of weeks. So stick around and check those next episodes out. Yeah. And Jake and James. Jake. James, Jakob, we are closing this. Thank you very, very much for taking the time.

It was a pleasure. And hope you have a have a great day. You have a great week, and I hope to see you around. Thanks. Bye.

Music