Expanding WordPress Horizons Through Tooling

Carl (00:00):
Alright, welcome to the Emerging Tech podcast. I’m actually the guest. I’m Carl, but I will intro the two hosts today. We have Dave, take it away Dave.

Dave (00:11):
Hey, I’m Dave Lucky. Nice to be here with you, Carl.

Carl (00:14):
And we have Kathy.

Kathy (00:15):
Hi, Carl. It’s good to see you.

Carl (00:18):
She’s shocked. You’re like, if you could see her face, she’s just like, what the hell is happening right now?

Kathy (00:23):
I love it. No, I love this. Well, I mean, this is the Emerging Talk podcast, and this is the podcast where we talk about things that are changing the paradigm, and so of course it’s great to have changing the paradigm of the entire podcast. I think it works. It’s on brand.

Carl (00:38):
Oh yeah, that’s good. Yeah, we’re super meta right now. Excellent.

Kathy (00:43):
Yeah,

Dave (00:45):
So look, I wanted you on this podcast, Carl, because we met in person for the first time in Athens at an event. it was a great party, super nice bar, but also quite noisy and I was quite exhausted after some days of conferencing and traveling.

Carl (01:19):
I think that was my third party that night. So

Dave (01:21):
You were saying a lot of words that sounded incredibly important and interesting. I was like, okay, let’s actually have this conversation again, but let’s do it in public and so maybe other people can help me understand what all the words you say actually mean later on.

Carl (01:42):
Yeah, I mean I’ll do, I’m really good. I feel like my experience in WordPress of just teaching really complex computer science topics prepared me for where I’m at today. Okay. We’ll kind of dive into it. So for context, for the listeners, I’m working on a platform called YMIR. So it sounds with an E, it’s with a Y, it’s from Nors Mythology. It’s from Nors mythology. I’ll tell you the story later once I explain a bit what it does, but it’s a kick out. I am a nerd. I spend a ton of time thinking about names, so I thought about this name pretty hard.

Dave (02:27):
So let us have the name in our head, Phil, give us the spell it out, and also tell us the domain while you’re at it so that we can go check it out.

Carl (02:35):
Yeah, so YMIR, it’s Y M I R and the domain is YMIR app. So Y M I R A A p.com.

Dave (02:46):
APP.com.

Carl (02:47):
Yeah, a p.com. And so the goal with YMIR was, I mean, what started it was I wanted run WordPress on a specific technology called a W s Lambda. So a w s Lambda is on demand computing. So it’s this idea that if you want to, you make a request to avett website, you will just basically run some code for the website, it will return the response back and you only pay for the time that this ran. So behind the scenes, what they labeled it as is serverless. So that’s why there’s no servers. So is there no servers with serverless? There’s definitely servers with serverless. I’m not a big fan of the marketing term. The idea is basically the servers are outside your sphere of concern, so you don’t have to worry about updates, server updates, security patches, all that stuff. It’s basically completely, the paradigm’s actually completely different, but when I was at WordCampUS, I was talking to a couple of people and they were saying, okay, but what web server are you using?

(04:09):
Apache Engine X, Lightspeed. I was like, none. There’s no actual web, there’s no engine X, there’s no Apache. There’s no Lightspeed. But actually at WordCampUS I had a better, easier way to conceptualize a bit. What Emer does is that Emer makes WordPress Cloud converts WordPress into a cloud application. So what most people are used to with WordPress is that you run WordPress, you’ve got a MySQL database, you’ve got maybe an object cache, memc D or Redis, you’ve got your web server. Usually all those things are on the same machine, and that’s how you host WordPress.

Dave (04:55):
That’s your classic lamp stack, right? I mean that was,

Carl (04:58):
Yeah, it’s your classic lamp stack or a lamp stack with Engine X. I don’t know what the stack’s called with Lightspeed, but is it Loom Stack? I have no idea. But yeah, so it’s your classic stack and it’s been like that forever.

Dave (05:19):
When I first ran WordPress on a laptop, it was a Windows laptop and I used Z.

Carl (05:25):
Yeah, I used Z. Yeah, because it was the only one that worked on Windows, basically. Mamp didn’t work

Dave (05:30):
Right? And WordPress was the only CMS I could get installed on a local server window, and that was the start of my WordPress story.

Carl (05:40):
Yeah, exactly. And all the hosting companies basically do it in a similar way. I kind of joke that it’s just different flavors of vanilla, right? If you go with WP Engine, it’s like French vanilla. If you go with GoDaddy, it’s your soft serve vanilla, but essentially with a few modifications here and there,

Dave (06:03):
What type of vanilla is that with Presto comb? Is that the finest Madagascan vanilla?

Carl (06:08):
Yeah, yeah, sure, sure. Exactly. Yeah. Or maybe with some Raspberry Swirl, the ones where they put a bit of the syrup on top when it’s coming out. Yeah.

Dave (06:19):
Not Jetpack like the Raspberry Whale.

Carl (06:22):
Yeah, that’s Jetpack. Oh, that’s Jet. You got it. See it. It works itself. Exactly. But they’re all kind of these flavors of vanilla and that hasn’t really changed at all, but it still really hasn’t really changed at all because that’s why we’re talking with new, that’s why I’m on this podcast essentially, is to talk emerging tech, and what Emer does is that it changes this entire structure. The way emer the product works is that you connect, it’s only with a W Ss, but you connect your a w s account and it will manage your infrastructure on a w s. It will let you deploy WordPress on your AWSs infrastructure into what I call now a kind of cloud architected version of WordPress. So the goal with YMIR in the long term over time is to move more and more parts of WordPress off this. I mean, there’s no more servers already. It’s already serverless, so there’s no more servers, but move more and more of what WordPress does to these specific kind of cloud services so that it’s really, it becomes more and more of a cloud enabled application you would design today. If you went to the drawing board today and you were like, okay, I want to build A C M Ss, you wouldn’t arrive and be like, oh yeah, I want a C M S that just runs on one server and can’t scale.

(08:02):
Nobody in their right mind would say that. They would be like, no, we need to be able to handle, especially in the case of WooCommerce and stuff like that, we have to be able to handle thousands of sales. All these kind of scaling problems that WordPress has now would be things that people would be like, I don’t want to start with this huge limitation in place and it’s just not where software development is now either. So that’s kind of what YMIR tries to do and well, I shouldn’t say try because it does, but it’s quite a step up so that we could talk about the fact that I’m building this in public too if you want, but it’s been a struggle to, because this came from the Laravel ecosystem.

(09:03):
So if there’s any Laravel developers listening to this, or if you’re familiar with Laravel, Laravel came out with this in 2019. It’s called Laravel Vapor. I was there when Taylor announced it and I was saying to, he definitely doesn’t listen to this podcast, but he’s still doing WordPress. But I was sitting next to Josh Pollock and Josh Pollock was like basic. I turned to Josh and I’m like, I’m building this. And he’s like, okay, you crazy guy. And I started working on a proof of concept because was just like, I was like, I’m going to get WordPress to work with this, and I was like, is this even possible? Slowly over time, I got a proof of concept working and I was like, okay, this ist impossible.

(09:52):
I just started, I continued working on it, and then eventually I had the platform built. It came out, you could start paying for it in 2021, I think February, 2021, and I’ve been slowly growing. It’s only doing 1300 mr. So it’s been a really hard sell because I’m pretty sure I’m making 16,000 a year. I’m pretty sure Taylor makes a million or more a year off of the same product for Laravel, but I just thought there might be enough technical people in WordPress to understand this, but the fact is there’s not that many, and it’s been a challenge for that from my perspective, but that’s part of the problem with emerging tech is like timing, because serverless as a concept is not new. I was talking with, so I know everybody in hosting, so I was having dinner in March with Phil Morris, who’s the director of operations at Automatic, and he’s been around the block.

(11:08):
He did 16 years at Sun Microsystem. So if you notice what Sun Microsystem is, you’re an old geezer like us, but he was like, yeah, sun was playing around with serverless back then. I was like, yeah, exactly. So timing is a important thing, but I think the timing’s okay because it’s just not okay with WordPress, but WordPress is difficult because WordPress is about always about 10 years behind technological trends as a general rule, because when I explain YMIR to a JavaScript developers, I just say, it’s versal for WordPress, and they’re like, oh yeah, that’s exactly it. This makes a ton of sense, and everybody in JavaScript knows versa. They love versa. They don’t want to think about things because the thing with serverless is you don’t have to think about a lot of things

Dave (12:02):
Before we go into all of that, because we’re kind of downstream of where my brain is at the moment, which is like,

Carl (12:08):
Okay, you can dial it back. You need to dial it back. Let me go too crazy on you.

Dave (12:14):
I’m not going to, I’m hardly you explain it to me. I’m five might not be too far from the truth.

Carl (12:23):
What do you want to have explained? Like you’re five,

Dave (12:26):
How do you run WordPress site that needs a lamp stack as part of it? It’s core documentation like WordPress needs lamb to run. How do you run it without land? Then what sort of video is this, Carl?

Carl (12:39):
I mean, some of it is proprietary obviously, but it was my WordCamp Asia talk. So I’ll give you the short of the WordCamp Asia talk, but the idea is that we talked about lamp stock, right? You have your NXs, your this, your that, and architecturally it’s very similar. It’s basically we’re starting to swap these parts of WordPress of the architecture with cloud services. So your NGINX becomes like, you don’t have to necessarily, I’m going to simplify it, but basically your Nginx would become what they call an AP gateway, and then that’s what receives your, because the goal of a web server is what is, you receive a request for a website for a page. The web server basically converts this and is deciding, okay, are you actually requesting an image? Are you requesting an actual PHP page? Makes that distinction true.

(13:53):
Like the config file, that’s like your Apache config or your NGINX config. That’s what it does. It basically translates what the web server is receiving and discernments what you want to do with it, and then sends it to PHP or requests a file and then returns the file. So the API gateway does that. So at a high level, it just receives the request, and at that point, it already knows that it’s a PHP request, but it will do the negotiation with now PHP, so your engine X now receives the request. It’s like, oh, I need PHP to process this. Then you’ve got PHP, FB M,

Dave (14:36):
Let’s stop with the, so the web server, you’re basically taking this complicated web server and you’ve rebuilt it as an API server.

Carl (14:46):
It’s just a service. So it’s a service that’s configured in a w s, it’s called API gateway.

Dave (14:52):
So it’s a highly opinionated and specific way of running a web server. So do you allow for, because people tinker with engineers or Apache, there’s all sorts.

Carl (15:10):
You can’t tinker with any of that,

Dave (15:11):
Right? So that’s not down. That’s what I mean by opinionated,

Carl (15:14):
But I mean, what are you usually tinkering with when you’re tinkering with Engine X, right? You’re maybe putting things to get media from elsewhere, but that could be done at the WordPress level. You could just be like, oh, my media is not here anymore. It’s on S three.

Dave (15:32):
I’m not saying it’s a good or bad approach, I’m just trying to understand it,

Carl (15:34):
But I would say it’s not just opinionated. It’s really hyper-focusing the different parts of, one of the things that WordPress does is it kind of mumble, jumbles all these different things together. Will do. That was our podcast with Kathy the other day. It was like security, right? WordPress will do all the security things we have, but it really shouldn’t. So it’s kind of the same idea. You’re having your web server do a lot of stuff. We’re going to have it do just the web server stuff, and then when you have to just do only the web server stuff, yeah, it’s opinionated, but actually it’s also like you don’t really need that much stuff if it just does the web server thing.

Dave (16:22):
So right, you’ve taken Nginx, you’ve stripped out the bits that you want, you’ve locked it down into this service, and you’ve then taken N X around the back of the shed and misery. Okay, so that’s the web server thing’s done.

Carl (16:38):
That’s graphic. But yes,

Dave (16:40):
Now the request has moved on to PHP, so you’re running, that would be PHP F vm, as you say in Nginx, and the PHP code for WordPress is still required. You still need the PHP files, but you do something clever with

Carl (16:56):
That. Okay. That’s where a lot of the magic is. So PHP F P M still exists, but it runs in a W S Lambda now with your WordPress code that’s been sometimes modified or a bit customized a bit with either the YMIR plugin or also during the build process. Some stuff might get injected, but it runs there with PHP F P M on its own with a W S Lambda, and so there’s one PHP worker and your WordPress code, and that’s it. And what serverless does is it lets you run as many of those as you want. So effectively you have access to infinite PHP workers with this, which is what is, and that’s really the crazy thing with it, at least right now, especially because it’s a bit of a innovator’s dilemma situation, but the killer situation right now is the fact that you have PHP workers scaling that is on a scale that nobody else can do

Dave (18:29):
From zero to whatever.

Carl (18:32):
I mean, I’ve done zero to 1400 in a minute.

Dave (18:35):
Yeah, me too. Easy.

Carl (18:36):
Easy. Yeah. Yeah. I know every hosting company that I’ve talked to is like, oh, yeah, we can do that, and we don’t need $3,000 a month boxes already ready to do that. I was like, oh, I spun up this project two minutes ago and I just did it. They’re like, oh, yeah, yeah, we can do that too. No problem. Yeah, so that’s

Dave (19:00):
That’s the PHP part.

Carl (19:02):
Yeah, that’s the PHP part, but that’s also most of what WordPress is, is PHP. It’s PHP and a C A D N. Really? If you have a content site,

Dave (19:11):
Let’s not forget about MySQL. That brings so many joys to

Carl (19:14):
Well, yeah, so MySQL, so MySQL is definitely the kind of a bit trickier part of this, but again, a w s is doing a lot of work to make this kind of SQL also elastically. But yeah, obviously if you spin up 1400 PHP workers connecting to a database, you’re going to blow up that database if you’re not ready. That is a hundred percent at that point, if you’re using emea, that is your number one issue if you’re not careful, is that you will blow up the database. The

Dave (19:54):
Database is the least cloudified part of this thing.

Carl (19:59):
It’s still, and in all of tech, it’s still the most unscalable, automatically scalable thing that exists. I mean, there’s a startup called Planet Scale that’s working on this, and then a w s has this product called, or Aurora Serverless that, sorry, YMIR supports, and I did a load test. So the load test I prepared for workcamp. Asia uses it and it scaled. It went from zero to doing, I forget, I think it was doing a hundred thousand inserts a second from nothing essentially. And that was great basically, and obviously talked with people. There’s not a lot of hosting companies that use a W S and WP Engine, but I was talking with some people and obviously at more extreme scenarios you would see in a larger hosting company like them at some point, technology like that, you start seeing the gears flying out of the box and things like that and smoke coming out.

(21:08):
But in general, it’s very promising because before, this is version two and version one was completely unusable. It just literally it went offline while it upscaled, which is like crazy. It’s crazy. It was literally like, oh, one second. And then you’ve got the naked gun GIF where it’s like nothing to see here. Everything’s exploding in the back while it’s upscaling. It’s like, no, you can’t do that. So now it does it seamlessly and well. But yes, the database is definitely an issue, but there are issues for various reasons. If you’re WordPress automatic or whatnot, you have to use sharding WordPress dot org is WordPress dot com is one big. You have sharding, you have different scenarios, but yeah, the database side is still the least solved problem, but that’s at the entire tech ecosystem level. It’s not even a WordPress problem. What YMIR solves right now is a WordPress specific problem. Why can’t WordPress scale beyond one server?

Dave (22:23):
Okay, so right, we’ve got your API gateway service that is taking an inbound request, it’s feeding it to PHP PHP is doing its thing with this. It’s like just have as many workers as you like orally in the same code. They will go to the same database, but you just have to solve that one problem then. And then presumably you’re just using whatever arbitrary cloud storage or your media,

Carl (22:53):
It’s SS three. So it’s SS three. So a lot of people use SS three

Dave (22:57):
And you can stick some caching in front of all this. You would a normal WordPress site so that you actually minimize what’s coming right through anyway.

Carl (23:04):
I mean, you can still use CloudFlare. The more advanced infrastructure still uses CloudFront, which is like the AWSS CDN, so you still have to use it, but yeah, you just move stuff there. I mean, there are challenges because one of the first hurdle that I spent two, three months working on is I had to rewrite the media library in WordPress because let’s say how do you upload a file when you have no server, so there’s no place to send the file. It’s a rhetorical question, but that was the challenge. What happens in this scenario when you are trying to make WordPress not live on one machine or on a machine period on a server period, but you still can’t have a CMS without the ability to upload media. So that’s what a lot of these SS three plugins do right behind the scene.

(24:09):
What they do is they sync the files with SS three, but I had to go one step further and rewrite the media library to send it directly to SS three. Again, those are normal things that if you wrote WordPress today or wrote ACMS, today would be a day one kind of thing that you would have or have the ability to have. But because a lot of, and this is not to dig on WordPress, I mean WordPress is a legacy application. It’s 20 years old, but S three was actually S three is almost 20 years old too, which is kind of mind boggling. And they still use the same API more or less. It’s mind bogglingly crazy, but still, you just didn’t have those paradigms back then. This idea of cloud, the cloud and things, and that’s what I talked during WordCamp Asia. This is just a serverless, and what I’m doing is just the natural progression of where things are going.

Dave (25:20):
You would say that because building a business around it,

Carl (25:23):
No, but also that’s fair, but also I’m just like the person in the early two thousands talking about the cloud. Do you remember in the early two thousands, if you talk to somebody, I’m going to start a hosting company and I’m going to use the cloud. They’re like, no, you’re crazy. What the hell is the cloud? This isn’t safe. You build your own data centers and you have your own racks and you do that.

Kathy (25:44):
You do load balancing and yeah,

Dave (25:47):
You

Carl (25:47):
Do load balancing. You have your redundancies. In case I had to do that for a law firm I worked for, what happens if Montreal gets nuke? We need to be able to work because people would work if Canada, that was the nine 11 days. It’s like of course a law firm would want to work normally after there was a terror attack on Montreal, but

Dave (26:08):
They’ll be people safe.

Carl (26:10):
Yeah, exactly. They needed their exchange server to keep working. God forbid they don’t get their emails, but you 2010, but you moved to 2010 and nobody would do that, right? They’d say, you’re crazy for not using the cloud.

Dave (26:24):
I get you. And I first heard about a W Lamb, but as part of their Hosting stack, pre spinning out as which was using Lambda for image processing,

Carl (26:42):
Well, YMIR does that too. So YMIR does a better version that you just do a flag. You just say image processing. True. And it configures CloudFront to do the same thing Photon does, basically that was the same thing human made did. They made it like photon compatible. I think it was the same API as well. Yeah.

Dave (27:07):
Alright, so let’s switch gears a bit now. So we’ve talked about how this works and my five-year-old brain now kind of understands it. So thank you for that and well done. Let’s talk about, so you said that you’ve been going for a year and a half, something like that?

Carl (27:25):
I mean since I started working on it four years?

Dave (27:29):
Well, no. Well, since it was purchasable

Carl (27:32):
Two and a half years,

Dave (27:34):
Right? Maybe correct me if I’m wrong, but you said the way that YMIR works at the moment is that you connect it. So I want to use Amir. I connect it and give it privileges within my BU environment and it provisions everything it needs to, and then I’m up and running and I’m like, that is what YMIR does. Now, that seems like a really specific way to deploy this technology. I mean, if you are WP Engine, what you haven’t done is build a Kubernetes deployer thing and I go into my A W S and connect it and WP Engine does it. If you’ve built an infinitely scalable WordPress cloud application engine, what the heck? Aren’t you just offering really good, cheap, fast hosting to anyone that wants it in a managed way? Why do I need that?

Carl (28:26):
Well, because I don’t want to go crazy. Basically, the number one question I get is why I’m not building a hosting. And the first question I get often is like, wait, so I’ll go back to WP Engine, but the first question I get is like, wait, it’s just you. What happens if there’s something that happens to you? I’m like, well, I love working on this and have disaster recovery. I do enterprise, I talk with enterprise, so I have to speak the language, but I also have disaster recovery scenarios if something happens to me. But the second question they ask is, why aren’t you doing a hosting? And I’m like, well, do you want me to quit and leave? Because that’s what’s going to happen if you make me build a company that I don’t want to build.

Dave (29:14):
Do not just need a co-founder Carl to do that stuff.

Carl (29:17):
Maybe I could, but I mean I would have the co-founder discussion is we could do an entire podcast on that one.

Dave (29:26):
Well, there are lots of ways of getting this technology out there, but Delta I’m trying to solve for is this seems like a really good bit of technology. Your traction seems small so far. Growing modest,

Carl (29:42):
Well, you mentioned WP Engine. I’m in talk with WP Engine right now because what I want to do is license

Dave (29:49):
It. Yeah, okay. Smart.

Carl (29:51):
Because it’s an API product, right? There’s an S D K, so you can integrate it. I’m talking with the big agencies. Some do hosting, they can’t talk about it, but they do hosting because some of them host very big websites. I’m in talks with all the big WooCommerce agencies too. It’s just, it’s hard, right? It’s a paradigm shift. It’s technically there’s not a lot of agencies even that have the technical know-how to really leverage this product also. So that’s my fault. But I mean, the reason I called it YMIR and not WP Serverless or something like that is because I don’t have to just do WordPress. I just love WordPress the community, but I’ve had lunch and learns with a W s for the education sector, and they were like, what about Typo three? I’m like, I’m not doing typo three, but Drupal maybe, and maybe I could talk to Adobe and do Magento.

Dave (31:08):
Fine. Alright, so what you’re saying is great is coming, but you’re taking a roundabout route to getting there because you want to stay sane and happy and live a fulfilling life.

Carl (31:18):
And also because I will literally, here’s how WP Engine approached me. They’re like, wow, this looks really interesting. There doesn’t seem to be a lot of people working on this. I’m like, we’re about four or five people in the entire PHP space and I’m literally the only person in WordPress. It’s again, everything in technology can be copied, especially big companies have infinite money and essentially if they wanted to, but it’s still a very multidisciplinary project. You need to know WordPress very well. You need to know a w s really well. You need to know how all of these pieces fit together really well. So what is more interesting to them? Is it to just spend three, four years building this or just license? License car?

Dave (32:12):
Alright, let’s say I’m an nonvoting licensed Carl, obviously, but one thing I’m interested about, so let’s say I am a WP engine and I charge a thousand bucks a month for an enterprise by small enterprise hosting plan. And It’s a great service, really polished, whatever, and I’ve got a lot of time and respect for the folks at WP Engine. What happens when I shop out their existing, what sits behind the server with YMIR? What happens to my cost base as WP Engine? Does it drop dramatically?

Carl (32:55):
Yeah, it’s going to drop dramatically. Just to give you an idea. So we’ll link it in the show note, Bob, you can ask me what the link to the YouTube, but I do a little, the load testing. So I do 4,000 WooCommerce orders in 15 minutes. And to give you an idea, this goes up to about 1200 PHP workers from zero. And that cost, because again, I mentioned it briefly but I’m just going to mention it. A lot of your costs when you switch to this architecture is usage based, which means that you only pay when your code runs. So when I did these orders, I only paid for when the people visited. You pay a bit for the C D N, but you only pay for the orders that went true. How much do you think it costs me to spin up 1200 PHP workers and process 4,000 WooCommerce orders?

Dave (34:06):
Alright, let’s play this game. So let’s imagine that you could do that on what would be the equivalent average. What’s our hosting banner? Do you buy WP Engine to be able to handle that kind of volume?

Carl (34:18):
There’s no physical server on the planet that exists that can host 1200 workers. So you already need multiple ones. So you need multiple of the big ones.

Dave (34:28):
Okay, so what are we talking about as a monthly cost?

Carl (34:31):
Well, those big servers, I mean, I don’t know what they pay for volume, but let’s imagine, I think it’s like 3000 I would say.

Dave (34:46):
Okay. So let’s say we’ve got 10 machines. So we’ve got 10, so I can do that. So that’s 30,000 and then that’s times it’s 15 minutes. Wow. Yeah. 15 minutes.

Kathy (35:08):
Yeah, that’s intense.

Dave (35:09):
Quarter of an hour times 24 hours times 30 days, right? I’m not doing the math very well. I dunno. Tell me Cole.

Carl (35:20):
It costs $10.

Kathy (35:22):
Wow.

Carl (35:22):
So not only that, but now you know how much it costs you per order. You can just do $10 divided by 4,000 orders and now you know how much you’re paying for each order. So that’s one of the things when I talk to finance people, their brain starts, you see the gears start turning in their head. No, but they’re like, wait, I can charge per order now. I don’t have to actually, I can do usage-based costs for WooCommerce. And they’re like, I’m like, yeah, but there’s a lot of these weird scenarios like that where you need a lot of scaling

Kathy (36:02):
And a lot of fast scaling. This is something that you’re talking about. So you have a product that you have a Super Bowl commercial, you have this business need that is coming, you have to prepare all of these servers for, but what happens to the business that all of a sudden goes viral on TikTok or whatever and they don’t know that they have to scale?

Carl (36:26):
I mean you still have to prepare a bit for,

(36:31):
I mean, super Bowl. Yeah, you have to. I mean there’s no magic for that, but I think for a lot of the most extreme WooCommerce scenarios right now, you’re covered and pre preparing for that would cost you a lot magnitudes, orders of magnitudes less to be prepared for it. Then it would cost you to just have those machines running all the time. You have to bring up the machines, you have to do this. One of the things that I find people don’t a w s does that people don’t always appreciate is you don’t have to worry about load balancing this thing. You don’t have to worry about network issues. A W Ss handles all of that. So it’s really interesting for that. So WooCommerce is one, I mean WooCommerce is the one bad I o a constantly, but I talk a lot of people with, so WP Engine, their interest is with headless.

(37:43):
So headless is a good technology that works well with this for two reasons. One, it’s like you’re just doing API, if you have a next jss site, you’re doing API calls once in a while and then you just want to cache that and why do you need to keep a machine running 24 7 for a bunch of API calls once in a while, but also depending on the technology that you use, if you build a static site, if you’re a Gatsby site for example. So I was doing a sales call, I mean I didn’t sign anything, but it was like basically you’re from the uk so enemy, yeah. So with them and they run the site on Gatsby. And what happens with Gatsby is you have to build a site. So when you deal with that, it’s like you’re your site for the process of, so you have to have, they have to keep these big machines running just to do their builds.

(38:56):
And then when the build’s over, they don’t actually need that machine to exist at all. The site gets sent to Netlify because Gatsby clouds closing, but they send it to Netlify, it’s built, it’s not making any API calls. It’s just like a static site and that’s it. But the build process, very C P U intensive, very PHP worker intensive. So those scenarios are the ones where it really shines right now. Obviously it shines for a lot of other scenarios. We had the security talk, well, I mean me and Calvin want to do something to Kathy, but one thing we haven’t discussed too is security. So with serverless, the codes read only. So that’s been a lot of, I’ve churn a lot of customers because of that. They’re like, why can’t I install a plugin? And I was like, well, you just have to install the plugin and then redeploy your application just, but your file system is essentially immutable.

(40:02):
So we’re starting to talk more about this kind of immutability concept in the WordPress space, but basically it just means that it’s locked in space. So each time that a w s spins up a new version of your code, it’s using the same code that you saved before. So even if you somehow managed to modify the files, you wouldn’t actually be modifying the files for the next person that makes a request. You’d just be modifying for that. So there’s still ways you can hack it, but I joke that you would have to literally upload the hack to your own site, which is a much different,

Dave (40:47):
Which won’t stop people.

Carl (40:48):
Yeah, no, it won’t stop people. If you go on Code Canyon, if you go on Code Canyon and you buy a hacked plugin already, then you just, but it’s like the DJ Khaled, you just own yourself.

Kathy (41:00):
You’d be surprised at how often that happens or maybe not surprised.

Carl (41:05):
Oh, I’m sure. But because this is so technically difficult, I’m not as worried about it, but you’re still vulnerable to SQL injection. So if a plugin has an SQL injection, then you still have those and they’re less common, but they’re still there. But this whole immutability, the security around that is important.

Dave (41:28):
That’s very interesting. So let me ask you a question if I may. One opportunity that springs to mind, because we work a lot on compatibility and we know that update a pain, if you work in software, you understand updates are pain, but they’re particularly a pain in the WordPress and WooCommerce space because the effects can be complex and unpredictable and it can be very time consuming to test et cetera, et cetera. So if you are taking this paradigm where you’ve got, alright, so you’ve got an API service sitting at the front and you’ve got the database sitting at the back, but in the middle you’ve got this integral PHP application, then could you have multiple different variants of that? So here is the site with, we come with some point, not one here, is it with some 0.2?

Carl (43:02):
Yeah, I had somebody wanted to try to do that for deploy environments for a whole bunch. I mean right now that’s a very extreme use case that

Dave (43:13):
I think you’re more likely the general principle because perhaps you could use it for a testing or

Carl (43:21):
You could go even further than that. I’m not going to go beyond that. When I did the A W Ss, they’re like, what about PHP 5.6? I’m like, no, I’m not supporting PHP 5.6, but I do 7.2

Dave (43:35):
That’s right in the back of the shed with N

Carl (43:37):
X. No. Yeah, but I do 7.2 and if you, a reverse case is like what if you have this old application, it just runs on as 7.2. It’s really hard to maintain servers with old PHP versions and things like that. I mean, you shouldn’t do that, but sometime maybe you have a small brochure site or something that somebody built a while ago and it, it’s chugging along. There’s probably some vulnerability maybe depending on how many plugins you’re using. But in general it’s pretty okay and you just want to kind of keep it there

Dave (44:17):
Like it’s archival.

Carl (44:19):
So you could do that with this because there’s essentially nothing, as long as you keep paying your a w S bill, it’s going to keep running that code.

Dave (44:27):
Yeah, I think that’s a really interesting use case because you do see, I did agency for a long time and you do see sites which are fine, but hosted in a normal way that WordPress is hosted, it’s not cost effective for them to do the updates. So sometimes they churn and they go, well, let’s just build a crappy static version of this. I’m like, we’ll make do with it. And I think the longer the older we see the WordPress fleet get, the bigger that opportunity is just to go right. It’s fine. It never needs to change and I want to pay as little as possible for it and let’s just let it live.

Carl (45:06):
Yeah, I’ve been tried the oldest WordPress version that this would work on, but I feel like it would work on pretty old WordPress versions. I go pretty deep low level in WordPress to make this work and they’re old hooks.

Dave (45:25):
I dunno whether you need to worry about that too much because if it’s survived to now, it’s okay.

Carl (45:34):
Yeah, exactly. Yeah. But always the servers always have to be upgraded. Look, I’ve been assistant, one of the reasons I love this technology is I’ve been assistant men since I’m 16. I just turned 40. I’ve been assistant men for 24 years. You tell yourself, man, I’m going to do this thing. It’s going to self update and stuff like that. But you know what? At some point you’re Ubuntu, L T Ss is not supported anymore. You’re going to need to upgrade it. You’re going to need to do something. You can’t just leave. I mean some people do, but professionally, at some point that server, something needs to happen to it or God forbid it goes down like the hosting company disappears. If you’re on a smaller hosting company to get bought out by GoDaddy and then shut down just to rip Media Temple, those things happen in the long arc of time, which you have to think about as assistant men.

(46:48):
Those things happen. And you know what I love about this? The only time I have to worry about something being down is a w s is down. And let me tell you, when a w s is down, you’re the least of your little brochure site is the least of your problems A third, the Internet’s on fire. When a w s has a major outage and it happens on a scale that is smaller than probably you would be able to do as assistant admin, you’re not going to out sit in a W S or G C P or Azure

Dave (47:23):
Or WP Cloud Automatic runs a pretty good infrastructure shop too.

Carl (47:27):
Yeah, they do. I YMIR’s closest competitor is probably WB Cloud, and I think I’m a better product than WP Cloud to be honest.

Dave (47:39):
I’ll let you wrestle or Barry for that title. But what I do want to say is in the privilege of being on this podcast with assist admin is to shout out that Automatic are looking for an epically good systems engineer. At the moment we’ve got an open role. So if you, Carl, Kathy, the listeners, know somebody who’s exceptionally good CIS engineering, send ’em our way.

Carl (48:04):
Okay. Okay. Yeah. I might know somebody actually.

Dave (48:09):
Anyway, I just wanted to put that out.

Carl (48:10):
But yeah, no, CI min is hard. Cis min is hard system min is hard and you have to, the arc of time that you have to look at is way longer. So I love that I never will have to worry about an SS S L certificate update. Oh, TU is out a date. I’m like, if I stop deploying a site, it’ll run as long as I’m paying my A W SS bill, it’ll keep running. Because one of the reasons questions I get pretty often too is why don’t you support G CCP or Azure and stuff like that? And one, the services aren’t interchangeable at that level. You can’t just pop one out for the other and then it’s like willy-nilly. But second is just, if I take G C P G C P sunsets stuff all the time,

Dave (49:05):
Google kill stuff. Are you serious?

Carl (49:06):
Yeah, I know Google kill stuff. So one thing that I feel confident with a w s is that they will never, they might not let you, that will be a thing with the product itself, but they might not let you deploy the same way that you did the previous deploy. But they will never take something that was running and say, this doesn’t run anymore. They never stop maintaining what they have already deployed. So that’s good because I just don’t want to have to think about that anymore. I’m tired of it. And that was my selfish cis men side that was just like, I mean that’s what you pay hosting companies to do. But I deal with people that, again, the point of the product is that you’s for people that have their own infrastructure. So how much of that do you want to minimize?

Dave (50:04):
I’m really admiring your zen state and the lack of your very low drag coefficient that you’re designing with this startup. I can empathize with not wantingness of those different responsibilities. And you’re making some hard choices by principle, which I think is always like,

Carl (50:33):
But one of the things I was talking with also in Asia is it lets you build WordPress products without having a lot of infrastructure knowledge. Now, one of my favorite examples from the Laravel space is this company called Fat Analytics. They do basically privacy focused analytics. They basically take on Google analytics and until very recently it was two people and was one developer, and they were just running this on Laravel Vapor and handling millions of requests a second with just this without any need of, there’s still stuff, but you still have to worry about how your code scales and stuff like that. But they didn’t have to worry about, oh, how many machines, how do I deal with increasing my server fleet or stuff like that. They just could focus on what they were good at, which is writing PHP code and doing the really application level things as opposed to the infrastructure things. I was talking with Elementor at Workcamp Asia and they’re like, oh yeah, we have this amazing Kubernetes thing, but they have a team of four or five engineers working all year just to handle their Black Friday sale. That’s not realistic for a smaller plugin company that might want to develop a WordPress based product that they just want to be able to build it and grow it and not have to worry about all that stuff for a long time. So what could you do with that too? Those are the kind of interesting things.

Dave (52:22):
Yeah, I mean we have to stop, but I want to reconnect that to the mission of this forecast, which is emerging tech. And I think one of the things that we’re seeing is that AI tools, so allow individual contributors to be more productive. I’m going to stick a vote for web three there, and you can suddenly access this open permissionless system of data and interactions and relationships. So serverless is further allowing, it kind of integrates with this like narrowing focus and allowing you as an individual contributor to build stuff that scales

Carl (53:03):
WordPress applications. So speaking emergent tech, what I think and what I know talking to also the hosting company is in, you can see it in the products that are coming out, is what they think is the next decade is WordPress based applications. The LMSs, the E-commerce. But also I was saying why couldn’t Gravity form have a self-hosted version of Gravity form and take on Typeform and Google form? Why not?

Dave (53:29):
Well, that’s Matt’s vision, isn’t it? That he’s talked about is where presses the operating system on the internet. So I imagine he’d be behind that vision.

Carl (53:38):
That’s kind of it. So what could you do with that when you have the platform to let you scale? Because that’s what all these products have in common is that they need PHP worker and CPU scaling. They’re not just put a C D N in front of it and you’re good. So I think that’s all the time we have here. I’m going to finish as the host, so we’re going to wrap this up. And it was good having you here, Dave and Kathy, where can we find you on the Internets

Kathy (54:09):
Everywhere, right? Everywhere. We are the internet, aren’t we? The three of us?

Carl (54:14):
We are the

Kathy (54:14):
Internet. We’ve got it all covered.

Carl (54:15):
Yeah.

Kathy (54:16):
This was such an interesting conversation. I can’t wait to learn more about ER and maybe play with it such a great convo.

Carl (54:25):
Yeah, me and Calvin are going to talk more about it too for the security aspect. But yeah, for me, you can find me@carlalexander.ca twig press on Twitter and it was YMIR app Y M I R A P p.com.

Dave (54:45):
And you can find me Dave Lockie pretty much missed ice.

Kathy (54:49):
Same. And of course Do the Woo.

Dave (54:51):
Alright, thanks. This has fun.

Carl (54:53):
Thank you. This

Kathy (54:54):
Has been great.