WordPress Isn’t Dead... But the Old Way of Building Websites Is

By Dan SanchezPublished July 31, 2026Updated August 5, 2026

WordPress isn’t disappearing overnight, but AI-assisted development is changing how small teams build websites, landing pages, internal tools, and custom marketing systems.

I killed my WordPress site last week.

That sentence still feels strange to write. WordPress was a big part of my career. I defended it for years, knew how to customize it, and could work my way through themes, plugins, CSS, and page builders without much trouble.

But “without much trouble” isn’t the same as “without much time.” Building a landing page still took days. Updating a site meant navigating settings, plugins, templates, and forms. A small change could turn into an afternoon of troubleshooting.

Then AI-assisted development got good enough that I rebuilt the site another way. The result looks better, takes less time to change, and gives me more control than I had with a typical WordPress setup.

So my prediction is intentionally nuanced: WordPress probably won’t vanish, but the old way of building websites is entering a slow decline. For small and midsize teams, AI-assisted development—or what many people call vibe coding—is becoming a practical alternative.

Why WordPress started feeling like a bottleneck

WordPress was revolutionary because it made publishing accessible. You didn’t need to build a website from scratch. You could install a theme, add plugins, drag elements around, and get something live.

That model still works. The problem is that every layer adds friction.

A page builder adds another interface to learn.
A form usually requires a plugin.
Analytics requires configuration.
SEO features require more settings or another plugin.
Custom functionality often means searching for a plugin or hiring a developer.
Security and performance become ongoing maintenance tasks.

I’ve been proficient with all of this. I knew how to hand-code pages, use WordPress, work in a drag-and-drop funnel builder, and make the pieces connect. It still took too long.

That was the real problem. Our marketing bottleneck wasn’t always writing the copy. AI could already help with a first draft. The bottleneck was getting the copy, images, design, forms, integrations, and pages assembled into something we could actually test.

We’d say, “Give me a few more days,” and then two weeks would pass with one page finished. Once the first page existed, duplicating it into a funnel was easy. Getting the first page live was the hard part.

AI-assisted development changes that equation. The thing that used to take weeks can often be done in a day, sometimes less.

What vibe coding actually looks like

“Vibe coding” can sound like asking a chatbot to magically build a website. That’s not really what I mean.

For me, it means describing the outcome I want in plain language, letting an AI coding tool handle much of the implementation, and then reviewing, testing, and refining the result. I’m still responsible for the direction. I still need to recognize good design, explain what feels wrong, and make decisions about security and functionality.

The process I used looked roughly like this:

Start with the existing site. I gave the AI my website URL and explained that I wanted to move away from WordPress while preserving the important content and overall visual direction.
Pull in the assets. The tool inspected the site, gathered the pages and images, and created a local working version.
Choose a better foundation. Instead of generating every component from a blank folder, I used an open-source content management system called Payload as the base.
Build locally first. The site was developed and tested on my computer before going live.
Deploy through GitHub and Vercel. GitHub stores the code, while Vercel handles deployment and hosting.
Iterate through conversation. When I wanted dates removed from blog posts or a layout adjusted, I described the change and let the AI make it.

The important detail is that I didn’t build everything from nothing. That would be wasteful and riskier than it needs to be.

Don’t vibe-code what open source has already solved

One of the most useful ideas I’ve encountered recently is the “third way” between vendor lock-in and building everything yourself: free and open-source software.

When I started rebuilding my site, I initially thought the AI would generate a custom content management system from scratch. Then I stopped and asked a better question: Has someone already built a reliable system for this?

The answer was yes. Payload gave me a strong foundation for content management and ecommerce, with a codebase that other developers have already worked on and examined. I could customize it instead of spending thousands of AI tokens recreating basic functions.

This principle applies far beyond websites. If you want to build a CRM, internal chat tool, project tracker, or knowledge base, first ask an AI tool to search for established open-source options. Then build on top of one that fits your needs.

That approach has several advantages:

You start with tested functionality instead of a blank screen.
You reduce the chance of forgetting important security or accessibility details.
You spend fewer tokens and less time generating basic code.
You can inspect and modify the system yourself.
You reduce dependence on a single commercial vendor.

Open source doesn’t eliminate risk. You still need to evaluate licensing, maintenance, security, and the quality of the project. But it’s usually a smarter starting point than asking an AI to reinvent everything.

This is similar to the broader shift I’ve written about in automation and AI agents: the advantage comes from combining capable tools with good judgment, not from blindly handing over the whole job.

The website becomes a custom marketing system

The biggest change isn’t that I can make prettier pages faster. It’s that the website can now become a system designed around how my business actually works.

For example, I can create a custom form that connects directly to my CRM. I don’t have to settle for the form builder provided by a plugin or accept a generic workflow because that’s what the software supports.

I can describe the process:

A visitor completes an assessment.
The form sends the contact to the CRM.
The system applies the appropriate tags.
The visitor receives a custom result page.
The team gets a notification.
A follow-up sequence begins.

That can all be built around the business instead of forcing the business to adapt to the limitations of a plugin.

I’m also building an importer that can take a new YouTube upload, find the transcript, turn it into a properly formatted article, place the thumbnail correctly, and publish it to the site. The point isn’t to publish more generic content. The point is to remove the repetitive implementation work between an idea and a finished asset.

That distinction matters. AI can help with the right marketing workflows, but it shouldn’t be used as an excuse to produce more material nobody needs.

Security is the reason to move carefully

I don’t recommend that every company immediately abandon its current website.

If you’re a large enterprise with hundreds of employees, sensitive customer data, and serious compliance requirements, this approach deserves extensive review before it touches production. New technology creates new failure modes. AI tools can make a bad change quickly. A database can be misconfigured. A deployment can expose something that should have remained private.

For smaller teams, the risk may be manageable, but only if you take basic precautions.

A practical security checklist

Use an established foundation. Start with a maintained open-source system instead of generating critical infrastructure from scratch.
Keep backups. Store your code in GitHub and maintain recoverable copies of your content and data.
Separate development from production. Test changes locally or in a staging environment before pushing them live.
Limit access. Don’t give an AI tool unrestricted access to every system, credential, and database unless you understand the consequences.
Run a security audit. After the functionality works, ask the AI to research common vulnerabilities for your specific stack and audit the project against them.
Use a second reviewer. Have a different model or a knowledgeable person inspect the code and challenge the first system’s assumptions.

That last step is especially helpful for marketers who aren’t developers. I can ask one AI tool to build something and another to red-team it. I still need to understand the recommendations, but I don’t have to rely on one system grading its own homework.

There’s also a useful form of future-proofing here. My code isn’t trapped inside one AI product. It lives in files and a repository. If I stop using one tool, I can give the same project to another. If I leave Vercel, I can deploy the repository somewhere else. The tools may change, but the underlying work remains portable.

AI content needs more human input, not less

The same shift is happening in content creation.

LinkedIn has started testing a way for users to report AI-generated content. I understand why. The platform is full of posts that sound polished but empty, along with comments that simply paraphrase whatever someone just wrote.

I’ve used AI for social posts myself. It was convenient, and sometimes the writing was sharper than my first draft. But marketers have become very good at recognizing the familiar rhythm of AI-generated content. The hooks, transitions, tidy conclusions, and vague claims all start to sound the same.

My approach has changed. I’m giving AI more of my actual thinking and less freedom to invent the substance. I might ask it to organize rough notes, clean up awkward sentences, or identify a confusing section. I don’t ask it to manufacture ninety percent of a post from ten percent of my ideas.

That’s the ratio problem. If a thirty-minute interview becomes thirty days of content, the system has to invent most of the material. That might be efficient, but it isn’t necessarily useful or authentic.

AI should help us repurpose original thinking, not replace the thinking. The human contribution needs to remain visible through experience, taste, specificity, and judgment.

This is part of why I’m skeptical when people describe AI as the death of creativity. The real danger isn’t that nobody can create anymore. It’s that people will accept generic output because producing it is easy. I explored that tension in the debate over AI and the art of marketing.

Stop chasing the biggest AI story

There’s always another major AI story to worry about: geopolitics, model competition, economic disruption, job losses, open-weight models, or some prediction about what happens ten years from now.

Some of those developments matter. But most marketers don’t need to make a decision about them this week.

The more useful question is: What is the next bottleneck holding back revenue?

Maybe the team can’t launch landing pages quickly enough. Maybe leads sit untouched in the CRM. Maybe reporting takes three days every month. Maybe customer research is scattered across documents. Maybe the sales team lacks the right content at the right stage.

Start there. Define the problem, then find out whether AI can reduce the time, cost, or effort required to solve it.

That keeps the work grounded. Instead of reacting to every headline, you’re improving the part of the business you can actually influence over the next three to six months.

For me, website development was the bottleneck. AI-assisted coding removed much of the friction between an idea and a working page. For someone else, the answer may be a custom reporting tool, a better research workflow, or an automated follow-up system.

You don’t need to build everything. You need to identify the constraint that matters and use the best available tools to relieve it.

The next practical move

WordPress isn’t dead. It remains a sensible choice for plenty of organizations, especially when stability, familiar workflows, and a large implementation ecosystem matter more than customization.

But the default assumption that every business needs a traditional WordPress site is becoming weaker. Small teams can now build custom websites and internal tools with a fraction of the old development effort.

If you want to explore the shift, don’t begin by rebuilding your entire business. Choose one contained project: a personal site, a landing page, a calculator, or a simple internal tool. Ask an AI coding assistant to inspect your current setup, recommend an open-source foundation, and explain the deployment path. Keep the original system live until the replacement is tested and backed up.

The tools will change. Codex may not be my favorite tool six months from now. Vercel may not be the best hosting option forever. That’s fine. The lasting advantage is knowing how to move from a business requirement to a working system without treating every small change like a development project.

Find the bottleneck, start with a solid foundation, keep your work portable, and let AI handle the repetitive implementation while you keep ownership of the decisions.

Dan Sanchez: WordPress is gonna die. Like I'm I'm calling it now. Like it's gonna do slow death, Ken, but it's gonna go the way of the taxi cab, which never fully died. But, you know, like there's fraction of what they were before. And I think not only is it gonna go away, but this whole everybody having a website thing, I think is gonna go up because it's becoming easier and easier just to have Chat GPT knock it out for you. That I think even non-technical users, even only half technical users that knew how to set up a Squarespace or a Wix or whatever, are gonna be having their own personal websites doing everything. Because now vibe coding has gotten to the point where we can all build these websites and even like little private super apps that do all the things and it's easier than ever before. So that's good what we're gonna be talking about today. Welcome back to the AI Driven Marketer. I'm Dan Sanchez, friends call me Danchez and I'm joined by my buddy Ken Freire.

Ken Freire: Yo yo yo. And I can't believe you started this with saying WordPress is dead. You're one of those guys who's saying things are dead. I can't believe it. We've gotten there.

Dan Sanchez: I know. I know. Maybe it's not dead, but it's it's going to start dying a slow death. It's dead for me, Ken. It's dead for me. I killed my WordPress site last week and I'm not going back. I'm never going back to this thing. It's been getting tiresome and burdensome over time. It was great. I had a good run. I've been defending WordPress up until the end. it was a big deal for me. I I like built my career off of this thing. but I think vibe coding just got so easy And I started seeing like how great things looked and it was better. And I'm pretty good at web design when I hone into it and get into the CSS and the little particulars and all that kind of stuff. But now whether you're using Claude or Chat GPT or Codex or whatever to like code web design, it's just better. Does it help to know the CSS and code and direct it in the little tiny things? Of course. Like it is with all AI things. But for most people, even people who know what they're doing, it it's just a way better route. So I want to talk about that and then kind of talk about like the future of like what this means for marketers. You and I have done enough work together, Ken, that like we find that a big bottleneck we've had before in the past is setting up landing pages, right? And it's not just the copy, because you might even have Chat GPT do a first pass in your copy and it's pretty good. And you're like, okay, now I need to build a landing page. And it was always a pain, right? You'd be like, Hey, when are we gonna have that landing page up? And I'd be like, just give me a few more days. Just give me a few more days, Ken. Tomorrow is always tomorrow. It'd be like two weeks go by and I'd have one page done. And then you know, once you have one, you can like duplicate it to make the the funnel and all that kind of stuff. But man, it took forever because you'd have to

Ken Freire: Yeah.

Dan Sanchez: make all the images, even with the drag and drop builder like that you have in high level, it's just kinda like, Ugh, you know, I was good. I was proficient. I knew how to do all the things. I knew how to hand code it, I knew how to do it in WordPress, knew how to do it in eye level, I knew all the things, and it still just took a lot of time. But now it's easy and it's crazy how much more fun it is to run it through chat. So you are on WordPress. Are you now d doubting your decision to be on WordPress?

Ken Freire: dude, after the last two weeks of my debacle, 100%. I think you and I both started revamping our website and you're like, I'm done. And I was like pulling my hair because of all the little situations. And I got it to the finish line and I think I texted you and I said, I will never do this again. So my next website, I'm I'm going to vap code it and not try to keep WordPress. I tried so hard because I was like, everything's there, I understand it, I know it, and now I'm like, nope, I'm done.

Dan Sanchez: Yeah. So if you're listening to this, I'm gonna recommend some specific tools, but there's many ways to get this done. I'm just gonna tell you the path I went down and give some practicals. I don't think this is for everybody. I think this for small to mid sized players because there is a little bit more risk here because we're dealing with new tech and new tech can just can be unreliable. Like ChatGPT could go rogue and like just take out a section of my website and it's just gone. This is happening right now, even with people who know what they're doing, just like databases just disappearing. So like you have to do it smart, you gotta create backups, you gotta Think about the cybersecurity problem, because we're doing new things, you know. Cybersecurity can be an issue, but I could explain some of the ways I'm like hardening it and making it better to make it less of a cybersecurity problem. But like enterprise companies, no way you shouldn't do this at all. big companies, if you have more than five hundred employees, don't do this yet. Like if if you're that size of company, but I think more and more this is going to be the trend. In fact, I saw this great post. With like you remember the little Salesforce mascot and had like the big no sign, the big with that

Ken Freire: Yeah.

Dan Sanchez: cross through it and it said software. Now people are starting to run with that and it says SAS because SAS killed software, right? Killed, but software never went away. But like SaaS really took over the software market. the software of a service web subscription stuff. And now more and more people are replacing SaaS with Vibecoded. Even in my internal company that I work for, Trim Healthy Mama, they're actually highly considering getting rid of Slack. Because there's a solution out there we can just build on top of and like we won't have to pay the fifteen grand for Slack anymore and we could just have our own solution that's completely free, other than the tokens we spend to make it better and make it more specific to us, which would be awesome. Imagine like having Slack, but having EOS built into it so that you have your whole strategic system planned around it and built into it. So you're running your L10 meetings out of Slack. You know what I'm saying? Like we already do that, but it's kind of clunky. But imagine if it was like customized for that. Now you're like, Like customized for the way we run business? Come on. But we're talking about websites here. So let me tell you what I did for my website. This it's some parts of it are hard and then some parts of it are easy. So let me paint a picture for you, Ken. Then you you can ask me questions if it's like, but what about this, Dan? Essentially you open up you could do this with Claude. I'm using Codex or the ChatGPT app, ChatGPT work or whatever, the Mac app. And you literally just tell it, like, hey, I'm on WordPress, but I would like to go. with a different solution. I would like to host on Vercel. It's the most popular solution. from go from you generally push from GitHub to Vercel. You're like, I want to work with that solution. my website's hosted on WordPress and I would like to find a different solution that works with this. that's all I told it. And I'm like, I want you to go ahead and download all the content and all the look at the style, look at the pages, look how they're designed, download all the images that make up the design and all that kind of stuff. And we're gonna rebuild it locally before pushing it to Vercel. And I gave it the URL and it just ran with it. It could you could also give it API access to your WordPress site so it can check out the database more deeply. That's not hard. There's a literally you go to your user in WordPress and there's an API key. And if you have admin access, you give you give it admin access and it can go in, it can search public site pages and then it can search the database and find other things. And mine was able to do it, found the site index by itself and went and just indexed all the content, pulled it all into my little folder on my computer, and then started rebuilding it in just HTML and CSS. But of course I'm like, well, I saw this really interesting post and I'm gonna talk about this later, from Chris, Christopher Penn, great guy, worth following on LinkedIn. But he said, you know, don't, don't start with blank, don't Don't vibe code everything. Like people have already built great open source solutions you can build on top of. Why spend all the tokens on things that people have already perfected in open source solutions? So I was like, I told ChatGPT this, and it's like, you're right, we can find an open source. So we found a different content management system, different from WordPress that works better at Versell called Payload, which is made for, you know, content management and e commerce specifically, which is great. I don't have e commerce on my website, but if I needed to, it's there. And it's free. available and works. It's a whole JavaScript sub like WordPress is bail based on PHP on the back end, which is like server side rendering. It's slow, it's clunky, it's old. JavaScript's kind of the new thing. JavaScript back end, JavaScript front end. So it it took payload and then we just built on top of payload. A lot of stuff I didn't have to vibe code myself and it makes the security better. and then there was a bunch of stuff with setting up Vercell and getting GitHub connected to Vercelle that I didn't know. Honestly, it was a little more technical than I'm used to. I'm used to like one click WordPress installs and then just upload the theme kind of thing. So if you're not used to pushing to GitHub and pushing from GitHub to Vercel, that part can be new for any marketer. That's okay. Literally ChatGPT will just walk you through step by step. You're like, hey, I'm stuck on this page. You say I'd find this. There's this great new feature on ChatGPT. I forgot what it's called. Let me let me find it, because it's w it's absolutely worth like looking at and adding it immediately.

Ken Freire: Is it the snapshot or like the image?

Dan Sanchez: Yeah, snapshots. App app shots. No, is this app shots? Snapshots? It's app shots is what it's called. It's a little setting in codex. Turn it on, it's in the settings. It's it's called integrations app shot. You turn it on, and then you just have to hit both your command keys at the same time and it takes a snapshot and uploads it straight into codex immediately. So you don't have to screenshot it and then find it on your desktop and drag and drop it. Those few seconds, if you're doing this a lot, saves so much time. Absolutely turn on app shots in codex. makes the whole process faster. But it'll w literally work you through all the little steps because there's a bunch of little API keys you need to set up in Vercell that's kind of a pain. But once you have it set up, Ken, I I you can never go back. Like even you, you're like, Hey, all your dates are wrong on your blog post. I'm like, you know, I don't even like dates on my blog post. So I just go to ChatGPT. I'm like, Hey, all the dates are on the blog post. Can you take them off? Just remove it completely. I don't like dates on the blog post because Google starts judging you for content over a year old. So I just remove them all. That's been an SEO best practice for a long time. Done. No problem. Don't have to go into the settings. Don't have to log into payload. ChatGPT just took care of it. You know, the cool thing is even if ChatGPT breaks or is down, or the tokens become really expensive, I could still log into payload and still manage it like I did with WordPress. I could still go in and customize the code or do what all the things that I needed to do. So like the content management system is still there to use. But if I want to build my own features now, I can. I don't have to go hunt for plugins. I just build my own plugins. I made an importer. I was like, every time I publish a new YouTube. a a YouTube video from this podcast. Like this very episode, it's gonna go to YouTube. the website's gonna find it, pull the YouTube and the transcript that comes with YouTube and turn it into a blog post and publish it and make it all look nice and you put the thumbnail in the right place and it just happens automatically. I don't even have to think about it. It's beautiful. And it does it the way I want it to happen. Not the stupid podcast websites that are almost always c close but not there because they do some stupid things that you can't overcome because it's hard coded in. No, you got full control over this sucker. You want it to spin, you want it to be magical, you want stars to fall from the p landing page as they load it, it can be done. You want butterflies to fly on that website every time the cursor moves, it can be done. Like anything you want. It's just a request away. And if it's not quite right, you just follow up a few more times. This is the magic of having your website. And again, it's a super app. You can connect it to all your other things. I'm like, hey, build a form, connect it to high level so when people fill it out, this happens. Done. You give it the high level API so it can inject. Contact records into your CRM. You don't even have to deal with CRM's form builder anymore or Gravity Forms or whatever form software you were using. It just vibe codes the form. You want it to vibe code a survey? You want it to be a whole assessment process that spits out an awesome page at the end? Just ask it, tell it what you want. And it's done. Custom plugins. I'm like, this is the future. No one, once you've tasted this, you're not going back. Have I convinced you? Are you are you gonna stay on WordPress?

Ken Freire: Yeah. I mean I well for for the thing that I just finished, I'm just gonna keep it there for now, you know, because there's like I'm like in the postpartum of it. just like let it lie. my personal website, Kenfidata.com, I'm gonna completely vibe code it and just go go after that because I realized one, I haven't updated it in like three years. It still has like the entrepreneur podcast there, so like just know how long that took. But the thing that I was always concerned about that you highlighted was remember when we were on Replit, anytime I wanted a change on Replit, it was like I had to ask it and wait. And sometimes I'm like, dude, if I could just have access to the CMS, I could just change it real fast and it takes two seconds. Like sometimes writing all the information down takes longer than me just going in and editing it. But with payload, that changes the game. And I'm like, okay, now, now I'm ready to to to make the jump.

Dan Sanchez: And with Replit I was already doing things like having it build a custom CMS where I can have a login and change the content. And you could have done that with Replit, but why? Like just start with payload. The cool thing about it too is then once you do it with Replit, you're kind of stuck with Replit. But you can actually push from Replit to GitHub and then pull it to ChatGPT and then to Vercel. You know, so like you can you if you could start on Replit. Replit's way easier to use, by the way. Like if you're new to vibe coding, just pay for Replit and start there and do some stuff. because you can always exit off of Repli easily by just pushing it to GitHub where it c stores your code and then and then go from there. Like it's it's just really easy. Because I've already moved one of our sites for our book to Vercel from Replit and it was it was actually way easier than I thought it was gonna be. the one of my other favorite things, Ken, is it's like w what if I want to jump off a codex? crap. What if what if codecs isn't the thing anymore? It's fine. Like I could literally swap it all out because it's all on my computer locally in the little files and stuff. I could literally just assign Claude the same folder and it can take over the whole process. It's all documented right there. Like Claude can take over. If I need to go open weights model and have the model hosted on my computer, all the code is there. It's like it doesn't, it's like more future-proof. And I'm not dependent on chat. Chat's just my favorite option right now. It could change six months from now. I might, because I I was a fan of Replit six months, and now I'm on ChatGPT. So it's like, meh. So that's cool. If I wanted to ditch Versel. Cool. It's already on GitHub, which is free. And I could redeploy from GitHub to somewhere else. There would be some con settings and configuration, but it wouldn't be that hard. Like moving these things around and then AI can do a lot of the work for you and he's becoming more and more capable of handling all the work of moving the images. And right now all the images are stored on Vercell's it's called the blob or something. It's like the image a thing. The cool thing is too is Vercell's free unless you want to store a lot of images. like I do for my website because it's I've got a lot of blog posts. or you have like a database connection of some kind. So if it's just a static site that you want to update, even if it have it has a blog, you really don't need a database for a simple blog. you can use Vercel's free plan and it's very generous. You can launch whole websites on Vercel for free. I'm helping another friend launch a verse a Vercell a site on Vercel and it's going to be free because it's just a static, simple website with some forms embedded into it. No need to make it dynamic or database driven. Shoot, you can even just publish your blog posts to static pages because ChatGPT can do it for you. So I'm really liking the new setup and I think it's future proof because you could switch out the different modules anytime pretty easily. and I think once most people have tasted it, you can you can get this done. And I think it makes it if you're a small team, a marketing leader with a team of three to ten, like this will help you run so much faster. So much of it is like implementing new pages, split testing, all those kinds of things. And it took so much time to do the copy and the images and the design and the development to get it just to test one thing. Now it's like what used to take weeks can get done in a day or less.

Ken Freire: Yeah. But my my favorite part of the whole thing is the design aspect of it. Cause a as you know, Dan, I am if you look at all my skill sets, design is below bottom.

Dan Sanchez: On the bottom.

Ken Freire: Wherever the bottom is, it's below that. You know? So like that's not but I can communicate certain things and I'm like, wow, it vibe coded it. It it like created

Dan Sanchez: Mm-hmm.

Ken Freire: it for me and I'm like this is way different. I get excited now. I'm like, I could create this, I could create that. Like the sky is the limit with this, and that's what makes it fun. and you save on all of the the third-party plugins. I think that's the most frustrating thing with WordPress, was that like, you need a plugin for this, you need plug in.

Dan Sanchez: Yeah. Took all the plugins. Yeah.

Ken Freire: I will say that like what Claude did that was super helpful for me is I had probably like 30 plugins, like third-party plugins, and it dropped it down to like 14. Like it was able to look at it 'cause I did give it access to the API and it's like, You don't need this, this, this, this, this is slowing down the speed. So it just took all of it out and I was like, that's great. I you know, 'cause I didn't know which one I still needed or not. So it's been pretty cool to see at least that part. It has the the power of AI is so much more versatile than we think when it comes to developing these things. That's why I'm I'm actually excited to redo my my personal website.

Dan Sanchez: Yeah. And then you can redo the other website once you get the first swing because you won't go back. You'll want to redo everything. for those who are concerned about security, I mentioned that. That's one of the bigger drow downfalls of this. And if you're a big company and have lots of money to protect, don't don't switch over yet. It the time will come where it'll be good for you too. It just you gotta be s you gotta be more hesitant, slower to move. You got more to protect. but for the rest of us, what two things. One, start with Start with pre existing code bases like payload because payload's already used by really big companies and it's open source and a lot of people have worked on it. A lot of cybersecurity people have looked checked on it, made it good. So that's a great base to start from. that way you don't have to code it and chat's forgotten something important. So starting with a good base really helps. And then once you build on top of it and do some stuff before launching it live, you can have it just do a security audit. Because before it's building, it's not thinking of security. It's just it's focused on getting the functionality you want. So it's not it doesn't have that security hat on. So then you can have be like, okay, for vibe coded solutions, what are what are the most common hacks that people people stuff falls apart on or it like it's things they don't think about? Go and do a big research project and find all the best security for this particular setup with payload and versal, whatever you got. It goes and does the research. You're okay, based on that, audit our current setup and then does the audit, finds the weaknesses. You're like, okay, based on that. go and fix it and then tell me what are the things I need to fix versus the things you can fix. If it's you something you can fix, go ahead and fix it. And then it goes through and executes it and hardens the whole thing. So it's like it that helps a lot. You start with good software as the base, you have it vibe code and then you have it circle back on itself. You want to be extra particular, you can have Claude double check codex's work or codex ch double check claude. That's what my our director, our our CTO at our company currently does. He has that red team the other c other agents code. if you really want to go another level up.

Ken Freire: Yeah. That and I it's funny 'cause I was just literally thinking that while you were talking, I was like, I wonder if you would have Codex check Chat GPT's work. Like they're both working together. I do that now with like copying different things. I'm like, hey, this

Dan Sanchez: Mm-hmm.

Ken Freire: is what these guys said. What what do you see is missing, you know? and a lot of it too is because ChatGPT has more context of who I am and how I write. So even though like Claude may have like written this thing beautifully It's like sometimes missing my voice and I'm able to like just take it back and forth and it comes out way better.

Dan Sanchez: Yeah. You can do that with lots of things. Definitely good to do it with code, especially if you don't code. Otherwise, if you were good at code, you could just red team it yourself. that's helpful. but for the rest of us marketers, none most of us are not developers for a reason. There's a reason why we're marketers. so the last thing I know you're thinking if you're listening to this is, well, Dan, when are you gonna build a course on how to do this? I do not need to make a course on how to do this. Just Look it up on YouTube, ask ChatGPT for help on how to do this. Give it your website URL, tell it to look at it and what it would take to do this. I promise. Like there's enough videos already out there on how to do this and the different little parts of it that it's totally doable. You do not need a course from me. And it's as soon as I made a course, it would be out of date anyway, because all these little details are changing. Shoot, Codex is changing like twice a day, it seems like. So I'd be like, push this button on Codex and it would change. So I'm not writing a course. Just look it up on YouTube and have Chat GPT help you. it won't take more than a day. It'll it's

Ken Freire: Yeah.

Dan Sanchez: about a day's worth of work. And then once you do it, you learn a lot in the process. So it's absolutely worth it just for the learnings on how to do it. But then once you're done, it's done and then you don't have to deal with it again. Okay, moving on to second news. This one was kind of fun. But LinkedIn is adding a report AI slot button. I thought it was a joke when I first saw it on LinkedIn. And then some other people that I knew started posting about it and talking about it in a way that I was like, okay, LinkedIn is testing this right now. I don't have access to it, but I've seen enough screenshots of other people who have it that I'm like, this is gonna be a thing. So this is exciting, but it's also like Okay, like this has bigger implications because others might catch on to this if it works. Because LinkedIn is obviously overrun with AI content. It's just everywhere people have been complaining about it for a while now. so they needed to add this button. And they did repla the release a blog post talking about why they're doing it because they want to cut back on AI content. They they notice people are starting to check out because LinkedIn's become less helpful because there's too much AI, too much generic fluff. So they've are testing and putting out this button. I'm hoping if I get access to it, I'll let you know how it goes. But I assume it's pretty simple. You just report and it flags it and then it probably uses the reports from lots of people to verify if it is or not. So what does this mean for overall content though? 'Cause I think this is gonna hit a lot of different things. What do you think, Ken?

Ken Freire: Well, first and foremost, the one the where my mind went when I saw this was two things. One, I said, Huh, I wonder if people are gonna use this button at a spite, like vengeance to try to take down other people. And then the second thing was like, I hope they do this for comments, not just posts. Cause I think

Dan Sanchez: That would be really nice. Yeah.

Ken Freire: It's the comments that probably drive me even up the wall more than

Dan Sanchez: Yeah.

Ken Freire: the post. It's just like, dude, you just regurgitated everything I said and I could tell it's AI. Just like, stop it. That's that's even more

Dan Sanchez: Yeah. Yeah.

Ken Freire: annoying to me. But yeah, those two things, man, I found that I was like, but I I think I'm excited for it. It pushes you back to creativity. We've

Dan Sanchez: Mm.

Ken Freire: already both been slowing down on writing, you know, more with AI and like just writing us personally. Like, so I'm more excited for that to see

Dan Sanchez: Yeah, yeah. Yeah.

Ken Freire: more authentic writing. And there's there's people who are writing authentically now that like are still outperforming all the AI stuff out there. so it I'm looking forward to my feed being cleaned up quite a bit.

Dan Sanchez: Yeah. It's funny 'cause it's gonna there's some people who probably use AI but you can't tell. And those people should be fine, which is good because it really just cuts back a lot on the ones that it's like obvious AI stuff, which is what we're all tired of seeing. and I think this is actually a good balance. I think it's gonna force us to wrestle with it as content creators, especially for B to B because it's where we hang out is on LinkedIn. but I hope I hope this goes well for LinkedIn. I hope this helps balance out AI proliferation with human creativity and creates a good medium ground somewhere because obviously it won't be perfect and there's people who are gonna be flagged who were a hundred percent human and it's it it'll be it'll be strange. And I think you're right to wonder. Google had a p a thing back in the day where you could report content or flag content proactively if you thought something was spam or stolen or something in for as it for SEO. And people were nervous about that too. But Google I can't remember what they did, but they did figure out solutions to so that people couldn't just recruit all their friends to spam your website and make you go drop in your Google rankings because that would be really bad. Some companies are not ethical and will totally go after the competitor that maybe ranks number one and they rank number three and just like dogpile them in the algo. But Google I I didn't hear a lot about that happening because I think Google was proactive in giving people ways around that so that it didn't happen to them. I can't remember exactly what they did. Probably worth looking into because I'm sure LinkedIn will try to figure out how to balance that out. That's why they're probably rolling it out slow is to try to figure out the medium with that. But I think it's a good move. I think I hope the other social networks kind of figure out something too because it's becoming a problem. We all know it.

Ken Freire: I I I think it's definitely gonna become a problem with like when video AI so that

Dan Sanchez: Mm.

Ken Freire: pe because people will be confused. Like even my kids, sometimes I'll let watch some shorts while I'm like on my phone or whatever. And they're like, Was that AI or was that real? Like th they still don't have the eye 'cause they're young, right? But even I think older people are gonna be caught in this shuffle of like, wait a minute, was that real? Was that not? And I think the more you can just start flagging stuff, the better. and I'm not saying we should get rid of AI. I like AI, right?

Dan Sanchez: No.

Ken Freire: But just being aware that this is AI or if it's slop, it should be gone. the way I'm

Dan Sanchez: Sure, a lot of marketers are like, Well, but I just built my whole AI content marketing system like, Yeah, sorry, that was for six months ago and now it's not gonna work anymore.

Ken Freire: Yeah. Well, you know, like I I s I met an individual who was like, man, I I interview a a founder for thirty minutes and then I create thirty days of content for them and I'm like, Yeah, that's the AI slop that's gonna go away and like that's the stuff I'm excited about 'cause like you know thirty minutes is not enough. like

Dan Sanchez: Yeah, yeah. It there's a certain ratio there. It's where it's like it needs to be you can repurpose content with AI for sure, but there needs to be a certain amount of like original content to repurpose. Otherwise chat GPTs, you've given it ten percent of the content per post and it's gotta have to fill in ninety percent with its own thoughts.

Ken Freire: Yeah. Yeah, even now like when I'm I'm creating posts, I've been more strict with AI where I'm like, here are all my thoughts on this and I don't even want you to like make it nice and hook. I just just clean it up. Like is there anything that should be organized a little bit better? And then I go back and edit quite a bit. Or before I was probably letting it have too much range.

Dan Sanchez: Same. We all did. Because it was better writers than us sometimes.

Ken Freire: Yeah.

Dan Sanchez: And I was pretty good at writing LinkedIn posts, but I found that AI could keep it snappier, faster, and it was just convenient with my little custom LinkedIn GPT and you'd send it in and give you three and you pick the one that sounded the best. but now I'm not using that anymore because the rhythm of way AI writes, marketers in particular are very familiar with it and can see it in my a million miles away now. So I just handwrite everything now, even if it's dirty. The dirtiness actually is like People like better now. It's funny how things change. So

Ken Freire: Yeah.

Dan Sanchez: that's fun. lastly in the news, gosh, this is a big one. And it's honestly not one thing. There's probably like dozens of macro geopolitical things going on with AI now, with the US, with China, with the economy, and China trying to crash the US economy. There's so many big things happening that could be talked about. And if you listen to any other AI shows, it's what they're all talking about. All the big AI players are weighing in and Sam Altman and blah blah blah. There's just so much going on. And I listen to these podcasts and it makes you fearful and it makes you wonder. And I'm like, you know what? None of this is helpful. I'm not even going to say what the stories are because they don't matter for what we're trying to accomplish most of the time. Do you now, if you're already listening to those podcasts, and I'm telling you, like, it's like the top, the top marketing AI podcast. I it's not even really a marketing podcast. It's more of like, Here is the AI news contextualized for executives at big companies, is probably their angle now, has been for a long time. they ask a lot of questions and they don't give a lot of answers, and it just kind of keeps you updated on the news, but it's not really helping you make decisions that are gonna impact your revenue or bottom line. And that's the thing. I'm like, there's all this news out there, but for you and your marketing team, the thing I want to say is what's the bottleneck holding you back? Focus on fixing that and then see how AI can help you overcome that. I'm gonna keep talking about this bottleneck thing too. I swear this is where the problem is with most teams in AI right now, marketing teams. If you're a marketing leader, do not get distracted by all the big picture AI stuff. You have to look at the next three months and say, like, okay, what do I need to solve over the next three to six months in order to lift revenue? What bottlenecks are holding us back? How can I use the best of what's available with AI? It's because you can get lost in like the bigger picture of AI and geopolitics and what's gonna happen ten years from now. Are we gonna all lose jobs and is China gonna take over the world? And is AI gonna become god go to God mode and start taking over the world? It's like none of these things are helpful for helping you do your job, for helping you crush it in your company. That's just I'm gonna get off my soapbox now.

Ken Freire: Yeah. Well it it's like it's it's rage bait, you know, and

Dan Sanchez: Yeah.

Ken Freire: and this is what regular news do does. They're they're trying to get us to be afraid of something or they're trying to get us mad at something. It's just like, okay, cool. Like there's all this stuff happening. We have no control over any of those things. What do we do have control over is how can we use these tools to better improve our systems or our revenue or whatever project we're working on. Let's

Dan Sanchez: Yeah.

Ken Freire: let's use that for now.

Dan Sanchez: Yep. I think a lot of things are changing and it's changing fast, but most of it's just we're too small players. Maybe if you're a big like if you're at a really big enterprise company, like if you're work if you're the CMO listening to this and you work at American Express, like you probably should be paying attention 'cause it's gonna affect y the you're playing at a different level. For most of us listening to this, if you're solo or if you're a marketing leader with three people, this stuff doesn't matter. You have to focus on what you can do now. Cause th the the big things are gonna happen regardless of what you do. And I know the advocates who hate AI are like, Dan, you're just giving into the problem. We gotta speak against them. Like, dude, this is gonna happen. Stop. Like, AI is gonna be a thing. The geopolitics will play out. It is what it is. We're not gonna make a difference. Just take care of you, take care of your company, take care of your family, move forward. Optimism. Let's go. There's gonna be a lot of bad stuff with AI. There's gonna be a lot of great stuff with AI. It's it's gonna be what it's gonna be, and I'm gonna make the best of it. That's that's kind of like where I'm at.

Ken Freire: Yeah, just like our websites. We were like, hey, let's improve them and now we're doing it and it's been

Dan Sanchez: Yeah. Yeah.

Ken Freire: great.

Dan Sanchez: Yep. All right. Moving on. Off of Soapbox. Sponsor of the show, of course it's high level. If you don't use high level for a CRM, I don't know what you're doing. Of course, if you're a big company, I get it. Like if you're got five hundred plus employees, you probably shouldn't be on high level. But if you're like I don't know, I feel like more and more people really need to downgrade from HubSpot to high level. It's not even a downgrade. I think it's better. but it certainly saves a lot of money and I can do so much more with high level than But I find it more and more it's becoming one of my core tools that I just hook up and then just use ChatGPT to vibe code around because it can do a lot. I'm using less of the little features and some of the bigger features of just like its workflows, its contact management, its ability to send and emails and text messages. If you had to vibe code into Twilio and manage your own email system, that would be a nightmare. Just let high level handle those things for you. and compliance and all that kind of c crap, you know. security becomes a big problem. So you want to be able to harden it and use dedicated systems that think about that kind of stuff like high level, which is why I keep recommending it. Just go to dance.com slash high level in order to start your 30-day trial. You get an extension on their normal 14-day trial in order to test it out as an app. if you're in the market for a CRM, if you're in the market for marketing automation, shoot, it does all the other things like what Catlin Lee does, funnels, course I'm on vibe coding my funnel, so maybe I won't be using that so much anymore. But everything else, almost all of my marketing tools I'm now consolidating between high level, chat GPT versal combo and what is it, Google Suite. Like those those are like my three main tools. and Riverside, the the thing we're recording on. Those are like the four tools that I use. Everything else I'm unsubscribing from. I'm just vibe coding or just using one of those four tools. just to save money because it doesn't make sense to pay for all these different SaaS solutions. but high level is definitely a pillar. I know it is for you too, Ken.

Ken Freire: I I enjoy using at first I was like, ooh, is high level gonna be great? And then now I'm like, man. Like I'm it it just

Dan Sanchez: It's no brainer.

Ken Freire: takes care of so many workflows and so many things that it is a game changer for you guys. So kinda like Dan said, if you use those those small tech stacks, it's gonna be awesome for you.

Dan Sanchez: Awesome. Let's head into I'm gonna go move into the poll. I did a poll. It was a little late notice. I posted it like afternoon yesterday, but it's still kind of revealing. I said, Marketers, are you vibe coding yet? I've only had thirty two votes. I'm not feeling very popular these days on LinkedIn. Again, it was a little late. but this is where the response of the thirty two voting, I had every like which one are you? Ev and Top answer was not top answer. It's the first answer that they could vote on was everything every day and then we had all the time, occasionally and never. Everything every day got twelve percent. So the people are heavy in this stuff. All the time was forty six percent, occasionally was thirty one percent, and never was nine percent. Now, of course, I have a very more highly technical marketing crowd following me on LinkedIn, so there's that. But even Bill Sherman, who I've known for years now, said, I never thought I'd be spending time in code or gendering half of what I create now. And I wouldn't call Bill Sherman, based on all the things he posts around thought leadership and stuff, has never been more on the technical side. But if he's vibe coding, I'm not saying he's a laggard by any means. he's vi he's probably pretty savvy. But like If he's doing it a lot, this is becoming just a normal thing. Like more mo a lot of marketers are now using or vibe coding, I think. we crossed the chasm of most using marketers using AI as a meaningful tool on a daily basis last year, and I think we're probably have already crossed, if not this year sometime, all marketers will essentially be vibe coding. Or most marketers, let's put it that way. Where do you fall, Ken? Are you on the everything every day? Are you like all the time or occasionally?

Ken Freire: you know, I what let's define vibe coding real quick for people. Is it just that they're using cloth code?

Dan Sanchez: You're asking AI to write code for you.

Ken Freire: yeah, I would probably say lately every day, the last month for sh for sure, it's ramped up because I was doing the website, I'm doing some apps. and it's becoming more and more like my my kids even joke around the look at my computer and they're like, What's all this text like what are you looking at? It's just a black screen with white text the whole time and you're clicking yes and no, like I'm like, Yes, this is my world now. I'm just I feel like it's back into the nineteen eighties when you just had like the dots up and just like Yeah. Yeah.

Dan Sanchez: The hacker screens got yeah, DOS or terminal. which it took me a while to figure out side tangent here, but I was like I went to when every time I go into our IT director's or I CTO's room and he has like two other people working in there, they got all the their screens turned vertical and they're they're vibe coding everything through the terminal, which I didn't get until I started setting up with verse stuff with Vercel. There's a lot of things you have to do through the terminal. And AI could just do it for you. generally. But to give it access to the terminal is to pretty much give it access to your full computer and everything in your computer. So it's a deeper security risk. But it's also just way more powerful because there's a lot of things you can't do. Like it's had to tell me, hey, open up your terminal and give it this. Tell me what it says back. Okay. Okay, now do this. And I'm like copying and pasting back and forth. But if it had access root access to the terminal and could just run there, which you can with Codex or Claude, I think they call it a C L I. I don't know what that stands for, but that's like agents on your terminal. Then it could do a lot of that stuff for you. 'Cause there are things that it's just more helpful for. but again, greater security risk, a little bit more technological prowess. I don't do it enough that I run it from there, but it it's becoming more normal for me to for it to tell me to do stuff through it.

Ken Freire: I'm doing everything through the terminal right now. Yeah.

Dan Sanchez: You're doing everything through the terminal? You graduated. I'm not there yet. I'm like no. I I don't like how the terminal speaks. I don't like terminal lingo. Did you did

Ken Freire: I I don't

Dan Sanchez: you play with the terminal before you were doing all this stuff? Did

Ken Freire: Yeah.

Dan Sanchez: you actually ever use DOS DOS?

Ken Freire: no. Not I

Dan Sanchez: Okay.

Ken Freire: mean that was way back in the day, but no. It just you know, I hired a marketing coach a while back and he was like teaching all this AI stuff. He's like, Yeah, just start with the turbino. I'm like, okay. So like I've just I just jumped right into it.

Dan Sanchez: Okay. Yeah. I definitely use the Codex app.

Ken Freire: and then I've you know, I have some decent technical background that it's been helpful and I had to like relearn how to do certain things, but it's been fun.

Dan Sanchez: Yeah. I I don't know, I the more I vibe code the I'm I'm d I it's becoming progressively more I can't say I'm everything every day, but I mean last couple of weeks have been as I've been redoing the website, but it's like, it's not every day, but it's pretty pretty often. Like it's more and more frequently. I'm probably all the time, but not quite every day yet.

Ken Freire: Yeah. It I would say it it becomes more apparent when you're building a project versus doing like research. Like I'll still open up the web browser and do chat GPT when I'm asking questions. I know some people who just do it all on the in the terminal and I'm like, it's overkill. Like just, you know, do what's easiest. but if you're building stuff, it's become my go-to.

Dan Sanchez: Moving on to the viral post of the week. And honestly, this post wasn't viral. It's actually far from it, but it's probably one of the most valuable posts that I've found all year. It's like one of the posts that should have gone viral that didn't go viral. So I'm giving it more attention because more people you need to read this post. I'm gonna link to it in the show notes. I'll read the beginning to it. And I've kind of already alluded to it, but I'm gonna explain why this is so important. Especially knowing that more and more marketers are vibe coding and we don't know ac actually how to code. So this is important. Christopher Penn. Who's one of my favorite people to follow on AI says a reminder that there is a third way between vendor lock-in and YOLO vibe code as an alternative. That way is called FOSS, F-O-S-S, free open source software. Free open source software that bears either an Apache, BSD, or MIT license can be used for commercial purposes without restriction. AGPL and GPL license software can be used for commercial purposes. If you make modifications that you distribute for your for your use or for that for use other than your own, you must share them back with the source project. So let's say you wanted a new CRM. You don't love the way a commercial CRM software is going, cramming AI into every nook and cranny, playing fast and loose with your data. Could you vibecode an entirely new one? Sure. But that's a waste of tokens. A far better way would be to start with an existing FOSS CRM like Suite CRM, bottle CRM, or code Zeta. Corte Corteza and make a fork as in you take their software. I'm I'm riffing off his post now, and then you just start vibe coding on top of it, just like I did with my website. I'm gonna link to this post so you can read the whole thing, but essentially he's making the case that like why spend all the tokens vibe coding something that you could just grab for free and just start building on top of? It's a waste. So now I'm every time I vibe code a new solution I'm looking to see if there's somebody else has already solved it and I could just find it on GitHub or something else. And of course I don't have to find it. I just send ChatGPT, hey, can you go and find a free open source software that can plug into what I'm already doing or can I start with as a r base in order to build the thing that I want to build? Rather than spending a having AI do it and make the security and all this stuff and not do it half as well as all these other people have already done, just start with that. If you keep this in mind as you're entering into vibe code, like a lot of us are for marketing. Not only will you d build better software, build better apps or whatever you're building, but it takes less time because you're already starting with the base and it takes way less money in tokens in order to get it started. It's just better all around. So that's why this post was really important because I was literally starting to recode my wa s site when I saw it. And then I was like, crap, s chat GPT, can you help me find a CMS that would work for for sell on what I'm trying to accomplish here? And it's like, Yeah, let me go find. I recommend payload. I was like, Great. a lot of people use payload, it's become very popular for that reason, which is why Chat GPT found it. But I just thought this was a really important post. Does that kinda change how you now think about all your future vibe coding projects?

Ken Freire: yeah, dude. I think at first I was so overwhelmed because I was like, man, I don't wanna build a new CRM. I don't wanna build like, you know, this new thing. Cause now it's just like a race to the bottom, you know, a race to the top because everybody's gonna try to build stuff. I was like, I would just rather find something that's working well.

Dan Sanchez: Mm.

Ken Freire: even when he was talking about CRMs, I was like if I ever wanted to, because I've worked with all of the C major players out there, I'm like, if I wanted to, I could probably build my own CRM now. 'Cause I know the sales world, the marketing world so well that it's like, here are the things that actually move the needle. And there's a bunch of bloat in them that is not necessary, especially in like the circles that I run in, that I was like, this this might be an option. Now I'm not gonna build that anytime soon. It's not in the the project plan for the next few years, but the option is there, which is really exciting.

Dan Sanchez: Yeah. And of course, naturally I was like, wait, do I even need high level anymore? Even though high level is cheap at a hundred bucks a month for all the things that it does. I was like, Maybe I could do it. And then I looked into it, I like, no, high level does too much. These open source solutions are mainly just CRMs, not full marketing automation stacks with automated text messaging and emails and all the different things that it can do. I'm like, no, I can't can't recreate that. Again, 'cause you don't want to have to vibe code solutions that tap into Twilio or Mailgun or whatever email server you have. So many compliance issues and things you have to keep track of there that I'm like, no, not doing that one. But for a lot of other things, consolidation. So take that as a good nugget every time you start or start to do something major in your vibe coding out there, because as marketers, there's just a lot of things that we don't know. We don't know we don't know when it comes to vibe coding and cybersecurity and all that kind of stuff. And I think if we just lean on this using FOSS software, free open source software. It will save a lot of us, a lot of time, a lot of token costs, and a lot of headaches and cybersecurity problems we would face if we just vibe coded the entire solution for free. So even the the thing we were doing internally at Trim Healthy Mama with replacing Slack, the only reason why it even came up is because Jack Dorsey released a whole GitHub for free. That's like a Slack competitor that you can then vibe code and build on top of. He posted that maybe like a week or two ago and we were already thinking about getting rid of Slack. And our CTO was like, dude, look at this. We could just start and build on this. I'm like, genius. Like, why build all why we're now we don't have to waste all the Claude tokens rebuilding the basics that Jack Dorsey, the founder of Twitter, has already built and just released out into the world for free. We could just start with that and build on top of it as we see fit. I think that's gonna be the future. And I think a lot of software companies are gonna take a massive hit because of that. But what do you gotta do? We gotta do what's best for our companies, do what's best for our market. And if you're a software company listening to that, you better take that in mind and figure out how to build a moat beyond code. Code is becoming more and more vulnerable. And many of them h do have motes, but if your company doesn't, it's gonna be a problem. You wanna close it, Ken?

Ken Freire: You know, I I as I I think about all the things that are happening, I wanna go back to one of the things you said earlier, is that focus on the bottlenecks. Focus on what's going on in your business. And if you are like not sure, even right now, where you're like I don't know what those bottlenecks out. Man, reach out to Dan. Reach out to myself. Like we'd love to have a conversation with you and just figure out what are those major bottlenecks. Because that's the stuff that AI still can't do that I think is where we still have a lot of the competitive advantage. Figure those bottlenecks out and then let AI assist you.

Dan Sanchez: That's a good word. Till next time.