WP Café #136 | Moving Away From Multisite  Hello and welcome to episode number 136 of the WP Cafe Show. The show where we share the real unfiltered experience of running our WordPress business. I'm Mark Wilkinson and as always I am joined by my co-host Keith Devon. We're both from High-Rise Digital, a tiny agency building WordPress websites and products. Today we're diving into a number of topics. We're going to be looking at moving away from multi-sight, a little bit about local development and sort of other bits and bobs that have caught our attention over the last few weeks. Um, and uh, so grab a brew uh, and you can settle in and see what it's really like to see behind the scenes of running our WordPress business. You can follow us on X at WP Cafe Show. We're also on Blue Sky and the web at wpcafe.show. If you're enjoying the episode, please think about subscribing to the channel, like the video, and why not share it with someone who else you might find it useful. We'd love to hear your thoughts throughout the show, so get involved in the comments. That would also be great. Um, apologies if you can hear a weird noise outside my house at the moment. There's someone decided that now is the right time of the day, having had all day, to start chopping down some trees. So, if you can hear some sort of chainsaw going in the background, I do apologize. There's not a lot I can do about it, apart from not bothering, of course. So, we'll have to just stick with it, I think. Maybe I can mute if it gets too bad. Let us know in the comments. As always, let's just catch up with Keith. How have you been? Obviously, happy new year. It's um after Christmas. I think it's our first show. >> Uh it is. Yeah. Uh yeah, pretty good. Uh life got some life stuff going on, but uh generally had a generally had a good break. I had a couple of weeks off, spent loads of time with my daughter and uh yeah, just yeah, it was it was nice. Lots of family around and uh but and not too many plans like not too many go here for this dinner and over here and over here. It was just it felt it was well paced. It felt like a nice break and we had a bit of a New Year's shindig at our house. So, uh that was fun, too. Um, but yeah, no, just it wasn't too much of a too too difficult to get back into the swing of things at work actually. I was like I was on the Sunday I was like I'm not ready to go back to work. I feel really out of it and then I kind of just got back in and just like Yeah, same old. >> Stuck in. >> So yeah, stuck in. It's been all right. How about yourself? >> Similar. Yeah, just a chilled out Christmas. Um, I'm not a massive fan of Christmas, I've got to admit, but uh it was good. It was good. We just had some parents over and stuff on Christmas day and just a just a chilled meal and nothing exciting really. And then >> yeah, >> obviously spent some time with the family, with the kids and stuff and played a bit of golf. It was nice weather sometimes. So yeah, it was good. >> Oh, I do have to tell you, I haven't told you what happened last night to me, which was really fun. >> So I went uh I go regularly to the local comedy club. So the Folkston comedy club, which has been going for >> I don't know 10ish years. I've been going since it started. Um, and it's so it's so popular. It's so good that they release all their tickets at the start of the year and they all sell out within about half an hour. >> Wow. >> For for the whole year. So, it's it's kind of that popular. Anyway, so um my wife's on the mailing list and uh she got an email a little while ago saying they're doing an extra special show. It's on a Thursday night in January and uh it's going to be it's a big comedian who wants to test out uh material for their upcoming tour. We're not going to tell you who it is, but uh sign up kind of thing. So, my wife got tickets initially not for me, by the way. I only got to go I only got to go cuz one of her friends couldn't go, which you know, >> harsh. >> Yeah. And it would I would have been seething if I hadn't been there. And what transpired transpired. So we got there. I had my theories about who it might be. So there were names being bandied about. Uh and so we had an initial act in the first half in an interval and then he came on the compare and he's like so we've got our special guest. Please welcome Russell Howard and it was like I love Russell Howard. Like >> I hadn't seen him for ages but I used to watch his like TV show and stuff all the time. Russell Good News I think it was called. So there I am sitting like I don't know I was in the third row so like super close and he did an hour set and like literally it was like non-stop laughter for an hour. It was so much fun. >> Yeah. Really really good. So >> very good. >> He was my prediction though cuz uh I knew that the com pair of our comedy club uh and the kind of owner of it uh writes jokes for Russell Hard. >> So they they they know each other and I knew that already. So that my my theory was correct. Yeah, it was great. Really good fun. >> So So when he comes out in the next few months with his new stuff, then you've already seen it. >> I know it all. Yeah. And it's good. It's funny. Yeah, >> it's good. There you go. >> Bit of a plug for Russell Howard on this show. Who would >> if you want to share it, Russell? If you're watching, please do. >> Come on the show sometime and talk about webass. >> If you need a website, let us know. >> Yeah. >> Very good. Very good. So uh today then let me go to the notes. What have we got today? Episode 136. Migrating a site out of multi-sight to a single site. We're looking at maintaining patterns, pages, local issues with local development. Um and then a bit of AI stuff that I threw in at the end there that I've been looking at this week. Um I guess we should start with the multi-sight. >> Yeah. >> Um >> do you want to introduce it? Do you want me to introduce it? >> Yeah. I mean I think you'll be doing most of the talking so I guess I can sort of introduce it very briefly. Um >> our our big recruitment agency client um who we built site for about I don't know four or five years ago now I think um >> they've got multiple brands um and they wanted a kind of unified sort of like design that they could tweak slightly for the different brands and stuff like that. So when we built the project, I mean our head straight away just went this should be a multi-sight with a single theme, shared set of plugins, shared set of users. Multiight is an absolute no-brainer. Um we're now about to pull that multiight apart and uh put all of the sites on separate well probably on the same hosting plan but in separate WordPress installs. Um, and it's been a bit of a journey to get there. And I think we're still I wouldn't say we're 100% sure that this is the right thing to do, but I think the balance has been tipped now. And uh, the client has been convinced that this is worth trying. Um, so we're going to we're going to give it a go. There. There's your intro, Mark. What are the myths? >> So, I think I think we talked about this last week. Last last week, last show. um whenever that was before Christmas. >> And like Keith, I'm not entirely 100% sure this is the right thing to do. I'm kind of 95% there now, I think, because I think the benefits clearly outweigh the the downsides. So my my issue with the multi-sight setup is that every time you want to do something and that's whether it's use plug-in X or a feature of the hosting, they're just not set up for multicight. They're never set up for multicight. Now granted, there are some plugins that, you know, are built for multicight or have multicight in mind and do things correctly for multicight. >> Yeah. Most people, most don't in our in my experience. Multicight is more of an afterthought for them and and why would they? Because actually most WordPress users aren't using it. So I get that. So, we're having problems with things like when our biggest thing is like well my biggest thing is local development like to try and pull a site down and to work on it is just it's it's so difficult to do >> because the database is massive and the f you know what is difficult about that >> because everything's in one database. >> Yeah. >> So the database is really big. Um you know you've then got to you pull the you don't really want to pull the whole thing down because you're only working on one site. uh it becomes difficult to sort of manage in that sense and then a lot of the tools on the hosting they don't they just don't work for the multicight. So one example is the host has just introduced this really good feature where you can see site performance. It talks about page speeds and gives you lots of analysis >> but it's only reporting on the primary site in the network which we don't actually use. That's just literally like the holding site. It doesn't do anything. So it's utterly pointless. So it doesn't really help us. It's stuff like that deploying to staging, having a staging environment, creating a staging environment. >> These one-click tools they have, they just don't work for for multicight or they don't, in our experience, they don't work. >> Um, and I think if we broke them into separate sites, it would give us more options for things like if something goes wrong, you're isolated to that one instance of it going wrong. If something goes wrong with one of the sites, we want to restore yesterday's backup. We can't just do it for that site. you're restoring the whole network. >> So, you know, the brands that have done changes since yesterday are going to lose those changes when actually there was nothing wrong with those sites. >> So, I think it makes sense to split them in that instance. Um, and I kind of want someone to sort of like tell me I'm wrong, but no, no one really has. People have come up with good reasons why we should stay stay in that sense and they're all the reasons I already know and I think actually they're not as good as the reasons to move. um things like the shared users obviously shared uh code base but we can we can get by that with other ways using git and things I think so I >> think the users is the main thing for me the code base is fine because like you say >> we can set up our deployment processes to deploy code to multiple servers at once >> that's not really a problem but with users yeah I mean if they want to delete I'm I'm saying this as I'm saying this I'm realizing actually that that's not really much problem either for this particular client because there are only a handful of global users like super admins and and yes and >> they can manage we we are one or two uh we can manage having multiple passwords for six websites like that's that's not beyond the realms of like possibility and actually the the the larger user kind of uh data sets are consultants which are spec or employees team members are specific to those sites anyway. So they're not we actually don't want them to be shared. So actually it it from a user's point of view we're talking about maybe five people that access the all of the sites regularly. So is is that really a big deal? Not really. Not when you've got password managers and things like that. So >> yeah, I agree. I think that's that's again negated by those uh the lack of users that need that cross site usage essentially. Um >> so I I' I've started to research how we go about doing this and um Delicious Brains um which is now I think owned by WP Engine. I think that's how it works. Yeah, >> I think they're still Delicious Brains, but I think they're W because at the bottom of the website it says Delicious Brains and it says WP Engine and everything. So, I'm assuming it is. >> I think Delicious Brains kept Spin Up as a separate thing, didn't they? Cuz that's the same team. >> I think that's separate to Delicious Brains, though. I think >> Yes, I think you're right. But it's the same. It was the same >> same team. Yeah. Oh, well, part of Yeah, they've kept that separate. That doesn't belong to Engine because obviously it's >> But obviously a Yeah. Yeah. Yeah. Obviously some things went across. >> So they they have a plugin which many of you may have used called WP. It used to be called WP Migrate DB Pro. Um and now it's just called WP Migrate. And one of its features is that it can handle um pulling and pushing multi-sight instances in and out of networks and things like that. uh the earlier versions it would work but they were more of a manual process whereas the new ones got this multi-sight tools uh add-on if you pay for the sort of pro version and one of the things it can do is you you set up a migrate and when you add the connection details in for the the site you're pulling which is the network it then knows it's a network and it gives you a choice of do you want to pull one of the sites out of here or are we are we pulling the whole network so obviously I chose um we're going to pull one of these single sites out. And then you've got choices for things like what do you want to pull? Obviously the database and it pre-seelelects the correct tables that you need to pull. So essentially they are the tables that are prefixed with the site ID and you need the users tables. So it preselects all those for you and then you can do your search and replaces which it kind of does for you because it knows what your local address is and it knows what your address was at the live site. And then you can choose to do the media, the plugins, the themes, the mustuse plugins and so on. Uh so I've been working on pulling that down to my local machine as a single site and I came across a few hurdles. So the first hurdle was that the media library on the site is large. So there's about 13,000 items in the media library. and it kept crashing um when it was essentially working out what media there was. So, it kept timing out and I think having chatted with the support which was very good, it seems as though it runs out of memory essentially. So, they they bumped the memory on the server um and it kind of solved that problem which was good. Um, then I ran into problems where I kept getting critical errors and that was because I was pulling down host plugins that wouldn't run locally. They they they gave errors. So, as soon as they got installed and activated, it just gave gave an error. So, that was a bit of a problem. Um, and then the final thing was it kept timing out. I don't really know why. >> So, they recommended just splitting it into different requests. So do the database, do the themes and the plugins separately and then do the media separately. And that that seemed to fix the issue, >> right? >> Um so I now have a working local copy of one of the sites on my machine >> which seems to work fine. So >> um that's kind of the process now for moving it to a different site. Now obviously if we were if we were going to do this for real, we wouldn't be moving it locally probably. we would be let's say signing up with a new host or a new hosting plan with the same host and then just creating an instance and then pulling it into that instance that's already live on the internet rather than doing that way. So >> um so that's the plan. So I've put that towards the client give them we've obviously already talked about the pros and cons of this. I think we're all on board. So it's just a matter of when we want to go ahead and uh deliver it. >> So fingers crossed we can get that done and it'll uh be successful. I wonder what the best way to because one of the things we'd quite like to do is test test out a new host and see if the performance is any better. >> I wonder what the best way to do that is. Do do you use some kind of garbage domain name that doesn't mean anything >> and like I don't know protect it from search engines from bots and stuff for a little while and then try to because what you really want the two sites exact sites kind of running side by side so you can test them. >> Yeah. >> In the moment you'd have to do that. So, you'd have to most hosts will give you a like a dummy domain name that they can use, >> right? >> You know, like, you know, x197435.host.com, what whatever, you know, so you can just run it on there. >> Yeah. So you could sign up, you use that, do the pull, get everything across, uh, and then you could, you know, run some tests maybe like that, hiding it from search engines and so forth, and then see if there's any difference in in speed, which >> I think that' be amazing. >> Sure will. Um, and then if if they're happy to do that, then obviously it's a DNS change. So it's adding the the the the domain to the host if it's a new host and then pointing that domain to that new host and then search and replace to replace the database. I I would like to do that. I'd like I' actually ideally I'd like to try out three or four different hosts and like what what an amazing case study that like that would be. It's just like we've actually tested the same >> real site on four different hosting platforms. These were the results. I think that' be fascinating. >> Obviously, you'd have to pay for a hosting plan for >> Wouldn't be that expensive. >> You'd only have to do it for a month, though, wouldn't you? So, >> exactly. >> Yeah. >> Yeah. >> It' be interesting. interesting little experiment seeing what uh what works, what doesn't work, what comes out best and why. >> Yeah. >> Yeah. >> That'll be some good good content actually in that I think as well for for us, you know. It's like >> Yeah. >> Yep. So, that's where we are with that. Anyway, um if you have any questions and comments, let us know in the uh in the live chat thing. That'd be good. Um the next one on the list um this was something that you were working on I believe was about >> um how to maintain patterns or block patterns across uh pages and sites within the network. Now this might go away once we don't have a network but it won't because we will still be managing patterns across multiple sites. They just won't be in a network but they'll be across multiple sites because those sites use the same theme essentially. So >> yeah, >> um >> what what were you doing and what sort of challenges did you come up with? >> So this is actually for the same client and I we've spoken like in previous episodes about the fact that uh a new designer came in and they've redesigned a few of the homepages across the brands. Um, and they've got this new kind of homepage design that they want to roll out across the brands with some with some tweaks. Um, but essentially it's a lot of it's the same. So, taking like the hero section for example, there's a new hero section. And I built it nearly exclusively in the block editor with a tiny bit of custom CSS um to handle just a few things that the block editor is not capable of. um doing that so that it's flexible for the client moving forward. It's kind of easy for us to build. Um and so I did that for one brand. um kind of signed that off, went live, and then and then it gets rolled out for some of the other brands and then one of the final brands who are actually like the pickiest of all of them. Like um I I don't know who's in the leadership of that particular brand, but they're always the ones that are super picky and things are a bit slower with them. Um so eventually it comes to their turn and >> meticulous maybe not picky >> meticulous. Yes. Uh and the designer comes back and says actually you know can we tweak this and that and it doesn't look quite right at this break point or that break point and they're just yeah they're more meticulous when it comes to this particular brand. So of course I can make those I can make those changes. Uh, I go into the block editor and I change the padding and the margin and I I've had to add, it's been really frustrating. I've had to add some spacer blocks and things like that. Uh, and I've had to use block visibility to only show those spacer blocks on certain break points and those break points weren't necessarily aligning with like um the WordPress column like where it breaks from multiple columns into single columns and things like that. So, it it all got really messy. um making these fixes. So, possible but messy. And messy partly because I'm not just using CSS. Like it's not like I've built some custom HTML and I've got CSS and I just tweak this thing over here and add a media query here. I'm trying to do it in the block editor. So, that made it more complicated especially because of this particular layout which has like overlapping vertically overlapping images and stuff. Um, and then it's like, okay, I fixed it on this page. There's five other brands that are using this pattern that I haven't fixed all all this stuff on. And actually, some of those brands are using this page are using this hero 10 times because they've got landing pages for each of their countries that they operate in. >> Countries. Yeah. Regions. Yeah. So now I've got somewhere in the region of 20 hero sections that I now have to manually go back and tweak. And it just comes back to this idea of like what is the best way to build these kinds of things to make them maintainable. Like you could argue well they are maintainable because I can go and make those changes but it's a very painful very duplicated process rather than >> you want to make those changes once don't you >> in a in a theme in CSS or or in a theme template file like that's >> but then but then doing it that way you lose all the flexibility >> that we have with the block editor and it's just this constant ba like battle and balance and it's I I I never think I'm getting it right, you know? I never think I've found the solution or the right balance. It's always, oh, maybe I should have built it this way or that way or maybe this should have been some kind of synced block, but then if it's a sync block that doesn't work on a multi-sight level, do you know like all this kind of stuff? Like it's just it's a bit it's a bit mind-blowing really. And like should I have done way more of this with custom CSS rather than using dimension controls in the block editor because then I could just go and change the theme CSS instead of faffing about in the block editor. All these questions and I don't have the answers. Please does someone have the answers? Like please tell me. >> And I guess if it because you can put block patterns in the theme, can't you? But I guess once >> you've changed the padding, the margin >> in the editor, >> it no longer is connected to that. So if you update the pattern in the theme, I guess it's got to be split, isn't it? >> And it depends what kind of pattern it is. Because if it's if it's just a pattern in the theme that you dump onto a page and then you can edit to your heart's content, >> that's not helping >> because once those changes are made, they're made. But if it's a sync pattern and you can somehow change that in a in a PHP template or HTML template, maybe that would work. I I honestly I just don't know. I I just don't know how to do these things anymore the right way. And maybe there is no right maybe there is no right way, but um this feels like it must be solved problems like there's quite a lot of the enterprise agencies that are building WordPress websites. Feels like this should be solved. I wonder how they're solving these problems. >> Yeah. >> And it'd be really good to get some content on this like it's on like the WordPress developer blog. How how do we do this stuff >> because we don't know. Is there anybody out there at that level that is putting out that kind of content anymore >> that I'm just not that I'm not seeing because I I used to look to the likes of Bill Ericson or you know some of the yes some of the bigger agencies would >> would put out more like how we built this kind of content and I just >> maybe I'm not looking maybe I'm not you know exposed to it maybe it's out there I know humanmade put out a lot of really good kind of general content but I I think understandably it's more focused on the clients >> on the clients not yeah definitely >> yeah why WordPress is a great you know great platform for enterprise not how we built X >> in in detail for developers like me I I would just I would love for that kind of thing >> to start happening again but >> um hi Nick says use a good old fashioned ACF block instead of a pattern I mean it would work but >> I I I do wonder for that hero section for example, should that have been a custom block where you add a background image, a foreground image, a title, a subtitle, a button, and I I have fields for them all. >> But it just negates the problem of using the block editor, doesn't it? That's the whole point is like >> but you can use the block editor for like the content. Do do you know what I mean? Like in the oldfashioned sense. Um, >> and this this kind of because we use custom blocks for other complicated >> uh sections of the design. So why wouldn't why shouldn't I use it for it's cuz I knew it was possible in the block editor. That's why we didn't do it. >> I saw that I was like I can do that in the block editor. >> Maybe I should. >> When we use a custom block, it's usually because like you can't do it in the editor. >> That's what I mean. >> Whereas so so I have no other choice. I I've have to I have to go for some sort of custom solution there. >> Yeah. >> But yeah, I I I think I still want to use the block editor. That's why it's there. That's why it's been built. That's why we're >> we've gone down that road, if that makes sense. >> But looking back, was that the right decision? >> I'd say no for the hero. >> For the hero. Just had a look at it now. I mean it's just it's just classic blocks >> except it it is except when they say but the break points don't work here >> and there and that image needs to go from a portrait to a >> landscape at this specific breakpoint because it doesn't work and this overlapping section doesn't work on mobile and where it does work on desktop. I don't have the responsive controls available to me. So maybe at that point if if it's like this is complex from a responsive point of view, I can't use the block editor. Maybe maybe that's one of the kind of hard lines. I don't know. Yeah. Mark says, "Hi, Mark. A custom block with or without ACF is the right way to go for custom page sections or hero elements like this." And then he goes on to say both in a second. Um just because you can budget together with core blocks doesn't mean that it's the right approach. core block combinations and patterns aren't suitable for ongoing content maintenance. That's interesting. Aren't suitable for content maintenance. For ongoing content maintenance, >> I say aren't suitable for design maintenance. For content kind of works, >> sure there's a fundamental problem there with a block editor then. >> Or maybe there isn't. I mean, maybe it's still better than what we had >> because we were all using >> we were all using something outside of core anyway. So maybe maybe they're right. Maybe Mark and and Nick are right. Like that's just custom block territory. >> I think that I think knowing what I know today that is custom block territory. Yeah. >> Yeah. Yeah. >> Yeah. >> Okay. Well, you learn something every day. >> Yeah. >> If if you've been having issues like this as well, we'd love to love to know how you how you dealt with them and how you've uh you know, what solution you you you chose. Yeah. And maybe the reasons for that. I think that's um I think this would be such a good episode like here here's some layouts like how would you go about this in in the block editor or is that not block editor territory? Is that something completely different? >> We could do that. We could we could take take a design maybe a site that already exists. >> Invite three or four different people on the show. They don't actually have to have built it but just say >> have your have your notes ready. people send them all the the page that they're that they're building. >> Yeah. >> And uh say like just come up with a plan of how you would do that and then we'll discuss that plan on the show and each person can say okay well I would have done this with these core blocks or well I use elementary so I would have used this or >> you know this is definitely a effort. This is definitely like you know that would be super interesting to see how people handle these different challenges. >> Let us know if you want to get involved in an episode like that. I think it'd be good fun. >> Yeah. >> Anyway um right back to the notes. Moving on. Where are we up to next? Uh where we going? Uh local WP issues. So we all have a local development environment probably. Um I am a M user. Sorry, really old school. I've used it for years and years and years and it kind of just works. So I've actually just paid for the upgrade uh to M version number seven. I believe it was 7.3. am now on which I thought was worth it. Um, you get like upgraded PHP cuz I couldn't upgrade to past PHP 8.1 and I know the purists will just be saying, "Oh, you could install it yourself. You just do this." Well, it's just far easier to pay the whether it's 70 quid and then be able to get access to it and press a button. So, that's what I did. Um, you don't use M. I know you don't use M. You use local, I think it's called. And you've been having issues. So, do tell. >> First of all, I used to use M and I can't remember why I stopped. I can't remember. >> Oh, there's your problem. You see, >> it was probably I I reckon it was a similar thing, though. It just something went wrong and I couldn't figure it out >> and and there was this new tool on the block and I was like, "Okay, I'll give that a go." Cuz that's I'm in the same situation now. So, >> for the last I don't know, four, six months or something whenever I've been starting up sites on local. So, Local is a WordPress specific uh development environment uh by WP Engine. I think it was by Flywheel, but the WP Engine by Flywheel, I can't remember. >> Um, >> and is it Sorry. Is it It is WordPress specific this local. >> Yeah, it's not. >> It It's Yeah, you start up WordPress websites. You click plus and you're adding a new WordPress website. >> And you can't use it for non-WordPress websites. >> I've never tried. >> Right. It doesn't matter. But >> I I don't know. But when you hit that plus button, you're in WordPress territory. It's not like, "Do you want to install a WordPress site?" It's like, "Okay." >> Yeah. >> I I have to just jump in. My wife has just delivered me a beautiful uh >> blueberry muffin from little >> nice. It's probably free bakery item. >> Yeah. That doesn't matter. She could have eaten it. She gave it to me. >> Good. That's true. >> That's love. >> Um so yeah, so I I've been using it for ages with no problem. It does exactly what I want it to do. I want to start up a new WordPress website. I click a plus button. I give it a name like website.est. >> Yep. >> I can set the PHP version if I want to fiddle with that stuff. I create the the user and within under a minute I have a working WordPress website that I'm logging into. It's really good. Recently, I was starting to get a few errors. So, every time I'd like boot up one of my sites, there'd be an error and then I just hit okay or cancel and it >> and it would just still work. And I was like, okay, I don't understand that error, but uh but it seems to still be working. And then just recently, just before Christmas, I tried to load up one of my sites and this error came up and it just does not work. So, the site just doesn't start. Um and and then so I got into like, oh my god, what does this mean? And I the error message really didn't make any sense to me at all. It's talking about things that I don't know about. Um so I did what everyone does these days and stuck the error message into chat GBT and I I just went down a rabbit hole of Oh yeah, this looks like it might be this. Okay, let's type this command into the terminal. Okay, type that command. Okay, >> I don't know where this is going. >> That hasn't worked. Type do this thing. Type this thing over here. Okay, that doesn't work. Okay, well that makes sense. Try this thing. And we're like, okay, that and then eventually I was like, I think I've deleted the website. I think I think it's deleted the database. Cuz eventually the error stopped, but it said cannot connect the database. And I was like, have we deleted the database? Is that how it's fixed the issue? >> So you asked asked Bob whether he deleted the database. He said, yeah, yeah. Oh yeah, it's Yeah. And then I said, "Why did you get why did you get get me to delete the database without kind of warning me that that's what was happening?" He's like, "Oh, well, I didn't I didn't tell you to delete the database. You know, we simply did this, which means that the database was deleted." Like classic AI. Do you know that scenario when people say, you know, uh AI is going to help like save the world? Um, but then there's the the other scenario where you you type in AI, can you please save the world? And it does it by like killing all the humans because we're the problem. That that's that's kind of the situation I feel like. Yes, well done. You fixed the error by deleting the website. Thank you. So So that's kind of that's what happened. That was before Christmas and then I didn't touch it again and I came back, you know, after the new year refreshed like let's give this another go. I did give it another go with chatb just taking a slightly different angle and it did seem to fix it in the end but the it was such an elaborate fix that it is going to be quicker for me just to delete the sites and recreate them because if I create a new site in local it works >> it's just the ones that exist currently don't work. So, do you know the underlying problem? Like why is it that the current sites don't work and the new ones do work? >> No. >> Right. >> I don't. Something about uh MySQL, but I I I don't know what like it's not loading up properly. Something is corrupt. Something. But I I don't understand it. >> I really don't understand it. But oh my god, I was going around in circles and circles. It was so frustrating. >> It is so frustrating, isn't it? When all you want to do is just do your job. >> Yeah. Yeah, you're like, I can't do this thing cuz it's everything's just getting >> all I'm trying to do is fix a gallery issue on a client's website and I've spent two days faffing around with local and Bob. >> So frustrating. >> Yeah. Yeah, >> I know what you feel like. Um, a few comments. Uh, Exams, hi, thanks for joining us. It's quicker to set up a WordPress install on local WP. That's why I use it. And and I think your experience would be the same until it doesn't work. It's a it was it's been a really good product for I know you've used it for a long time. >> Yeah. Yeah. >> Which is um which is good. And then Mark says the same quick functionality is possible with Laravel herd. Installing WordPress is a single command in the terminal. Terminal puts me off straight away. Um usually work like WP sorry usually work like PHP version switching or database maintenance and SSL is in a guey. I I have to admit I am fully gooeyed up. I don't I don't I don't want to have to do anything in the terminal. I need a gooey um for whatever I use >> which which is what M does uh to the most. The only thing I use the terminal for is WordPress um command line stuff. >> Yeah. >> Which is quite useful sometimes. >> Um what do I do in there? Sometimes importing some posts or something if you got like a big import file or stuff like that. I have to use it these days for zipping up plugins because if you do that just with by right clicking and compressing in on Mac these days it doesn't work on WordPress anymore. >> That's interesting. It does for me. It didn't for me. >> Oh, they have they changed that. Have they fixed? >> I thought they fixed it in WordPress. Yeah. >> Okay. If anyone knows if that's been fixed, let us know in the comments. But yeah. Okay, that's good to know. >> Yeah, >> I thought I thought they'd fixed it. I know you don't always update Mac OS as soon as me though, so I wonder. >> That's true. >> It could be a Mac OS fix or it could be a WordPress fix or it could be both, I guess, but it definitely works for me. So, yeah, it's good. >> Yeah. Yeah. Yeah. Interesting. But yeah, there you go. So, I'm in that kind of world at the minute where I feel like >> feel like the next steps are just recreating everything. And in some ways, that's not the end of the world because, you know, when your local your local setup gets a little bit disconnected from reality, you know. >> Yeah. >> So, it's not clean slate and start fresh might not be a bad thing. Would if you do that would you keep the same tool though or would you go with a different tool >> I am tempted to keep it because >> you know it it works most of the time. >> Yeah. Yeah. I don't know if something must have been corrupted or some something weird's happened. When you go for something new though, you've always got that that same experience in a way that you've just had in that because I'm using this new tool, the thing that I want to do took me like 5 seconds last time, but now I don't know where it is. I know it can do it. I've got to find it. So, it now takes me five minutes and it'll take me five seconds next time I do it. But there's always that, isn't there, to try and get over. >> I think though, like you're using M, which is >> a paid for product. You get actual updates. You get actual support. Local's free. >> And there's there's no support. I mean, there's there's community support. I'm sure like I'm sure someone might jump on, but like >> it get updated. >> It gets up it does get updated. I don't know. I don't know how actively it's developed and supported though. Whereas with with M, >> you're paying for you're paying for it like and that that means something. I I trust that more. I going for >> I like it. I know it's kind of the I don't know. Oh, you're using M, you know, like don't use that. There's something cool way cooler than that. And it's like but it just it has just worked for me like all the time. >> That'll be another good video. We could do we could do you and me both setting up a WordPress site, a development site, you using M me using local and and compare and compare the process like >> is one way easier than the other. >> Maybe you can certainly go on the channel. Yeah, >> go on the YouTube channel. Yeah. Yeah. >> Cool. Anyway, so let us know what you've do to fix it. I'm sure everyone will want to know >> how you fixed it, if you fixed it. >> Yeah. Yeah. >> And then the last thing I had on the list was just like, have you built anything with AI that is interesting? Um I have this week sometime. Yes, it can't be yesterday. It must been the day before. So we had um a support question come in and the front end uh radius based search wasn't working. There was actually multiple reasons. The the the primary reason was that the the API key was like not allowed on the website. So we had to sort that out and do the correct referers. So it was allowed to use on the website. But once we' done that, the front end worked, but it wasn't actually finding any jobs. So they type in London, select London from the working front end, and then it wouldn't come up with any jobs even though they had jobs that were in London. And the reason for that is because they weren't geoloccated properly. And it was all to do with a change over in API keys and the same issue on the front end. It wasn't wasn't pulling. So a new job comes in the API key wasn't working. So it couldn't geoloccate it as that job was saved. However, they are using our job relay products and every job asterisks not every job but most jobs who come through job relay. We geoloccate the job and the data that we gather is then sent onto the website and they can use it. So the data was actually there and it was stored as post meta. But the problem was is that the tool on the front end I think it was repress was looking for different meta keys than the where the data was saved. >> So this client had a hundred jobs I think it was. So I sent them an email back and I explained what had happened and I explained the solution and I said you've got a hundred jobs. This is going to be like a a go in and copy paste job. Just go in and change the key names using the post editor. It's the easiest way to do it. I could build something, but you know, I'm going to have to bill you for it and it'll it'll it'll be a while. So, they did that. They were fine. They said, "Thanks for your help." And they went through and did it and it was great. But then I just thought, I wonder what AI can do for this. So, I just jumped into the chat GPT. I typed in the sort of thing I wanted to do. I had I had latl long saved as I think it was job relay latitude job relay longitude and I wanted as a different name. So can you just write me a script that will do this? Yeah yeah no problem off he went wrote the script. Um, and it kind of worked. And then I thought, well, what about like if I actually because I I just told it the keys, you know, this is the old key, this is the new key. I want you to copy across to the new key so it's existing. And I said, well, what about making a little settings page so I can just type in the original key and the new key I want you to save it to. Yeah. Yeah, no problem. Can do that. Off he went. Did that, saved it, tested it. Perfect. Um, then it added like batching them. So in like posts 200 posts at a time because he he told me he's like this might not work if you've got 10,000 posts because it'll time out. It's perfect. Did that. And then I was like what about if I only do it for one post type? Like let's say I've got these keys across all post types but I only want to do it for one post. Yeah, I can add that for you. So in about 20 minutes I had a working plugin that I could just >> convert one meta key to a new meta key. >> Yeah. Um, and how how long would that have taken me? It would have taken me more than 20 minutes. Maybe an hour, maybe an hour and a half. >> Maybe maybe a lot more. >> Maybe a lot more. I mean, I probably could have done the sort of rough and ready one that did the key, you know. >> But yeah, so that that was something I used it for this week. Um, didn't actually use it in earnest in terms of on that site because you'd already done it, but I thought it'd be interesting just to see how long it would actually take. And um >> stuff like that is kind of really good for because it's almost like >> it's just it's come it's like throwaway code in a way. You know what I mean? It doesn't it doesn't have to be perfect. >> It just has to do the job that you want it to do. >> Um >> yeah, >> I had a brief look through it. It looked like it was using things like nonses and and escape function. So, it was I haven't reviewed I haven't looked at it in detail, but for a quick install, do the job, take it away, it seems like a good use of it. I don't know what you've you've ever done stuff like that. >> I haven't done I haven't done throwaway stuff. I have I think we've I mentioned it on a previous show. Um I used it for some client work where they asked for like this kind of calculator. It was an admissions calculator for a school and it was uh >> they wanted they'd seen something that another school had on their website and you kind of select the year of birth of your child and then it gives you all the information about when you would apply for the school at various stages. >> Yeah. >> Um and it was remarkable. So I basically I think I gave Chat GBT a couple of screenshots of the of the tool that was on the competitor website, the other school website and said I want to I want to kind of recreate this and uh I want you to give me the HTML, CSS and JavaScript um that I that I can add to code pen and and it'll work. and it gave me the HTML, the CSS and the JavaScript. I put them into code pen in the boxes and it worked. And the UI was like similar to what was on this the school the school website in terms of like the layout and everything. It was >> it was amazing. And I've I've used it for something else for another school site. What was that? There was some popup >> some popup one of the schools asked me to do and I basically said, "Can you do this for me?" it was bang done. I was like, "Okay, that would have taken me that would have taken me a while, but um but it was done in seconds." So, yeah, there's there's definitely use cases for it. >> Yeah. >> Another completely non-code and non WordPress related. I've been using it for stuff like for for golf. So, I was looking at some new irons, some golf clubs that were they were like reviewed and stuff and they were supposed to be really good and they were quite cheap. you like a massively different entry price to like what the top brands are using. >> Yeah. >> But I was looking at I was like I don't know what like shafts I would get in them because so I literally read the specs of my golf shafts to to Chart GP and said can you tell me like what what's equivalent spec on this one. It's like oh yeah no problem. Yours are this. They've got this flex. They've got this you know this stiff length of them is this. >> These would be the most appropriate because of this and this and this and this. M >> it's like that would have taken me ages. >> Do you think it's right though? >> I do. I do. Yes, I do. >> Um I I I am suspicious like you. I thought you'd say >> Yeah. >> And I would I would never hesitate. I would never say just believe it straight away. I think if you had no knowledge of this thing, you probably don't want to trust it straight away. But I do think it's right in what it's saying. Um, >> the other thing I had was back to golfer again. So, my driver has an adjustable loft sleeve. So, you can take the take the shaft off the head and then if you turn the the shaft before you put it back on, it does things like changes the lie angle of the club and the loft on the club. >> Mhm. >> But on the end of the shaft, it just has like dot plus minus minus. So I was like, I don't really know what they're doing. >> So I just literally took a picture of it and said to to chat, can you just tell me what the difference is between these? And he's like, oh yeah, you're using the ping such and such a thing. It's got this such and such system. Give me a big table of all the options. You know, you want to use this setting if you're pulling it left. You want to use this setting if it's like that sort of stuff's really good. >> Yeah. >> Just other other uses for it that I've been been using it for. Yeah, communication for me like you've you've written a reply to a client or or to any to anybody and it's just like it's just doesn't it doesn't flow right. I've got I've got the information there but it just doesn't read properly and just dumping that in. That's I I'm using all the time for stuff like that. Um which is good. Although I don't like it when I get >> emails and you can tell they're like basically written by chatbt. We got one from a client that will remain nameless uh towards the end of last year. Um and it was just like that was 100% written by chatbt and it's someone that we know uses chatbt a lot and I really I almost just didn't give it the time of day like because I was like these aren't your thoughts like you're not telling me what you think. You just run this through chat GBT and just sent us the the output. So I don't want to be communicating with AI when I think I'm communicating with a human. So it's a really f it's a really fine balance. I think if you're using it to refine your thoughts or give you structure or stuff like that, it's okay. But if it becomes AI just talking to each other and the human sides taken out, I just I I get less and less comfortable. But >> yeah. Yeah. It's not the best. Yeah. >> Yeah. Well, that is us wrapped up. Done 49 minutes nearly. So, uh we'll we'll give it a rest there. Um thank you ever so much to everyone for listening, watching. Uh if you've got involved in the the chat, that's really good as well. Thank you. As always, if you've enjoyed the show, please hit the like button. It really does make a difference. Do subscribe if you haven't already done so. And share this around on your favorite social platform. It would be nice to get some new listeners and viewers as always. But until next time, which will probably be next week, but as always, we move around a little bit. Then we'll uh we'll see you next. >> See you soon. Bye., WP Café #136 | Moving Away From Multisite, YouTube, 01kf1ckbmf3x4r1gt8hx6r04pn.