Pod Insights Syntax - Tasty Web Development Treats 956: Should I Keep Using WordPress? About this episode In this potluck episode, Wes and Scott answer your questions about paid vs. free SSL, the state of frontend jobs, headless WordPress trade-offs, organizing TypeScript types, and more!
Show Notes
00:00 Welcome to Syntax!
00:51 Recapping the GitHub Meetup
05:14 Is there any real benefit to picking a paid SSL over Let's Encrypt?
08:03 Is the pure frontend role disappearing?
11:17 Is the gravy train over for software devs?
20:48 How Scott automates versioning with GitHub Actions
changesets
Intro to using changesets
zero-svelte
graffiti
25:16 Brought to you by Sentry.io
25:41 Thoughts on VS Code alternatives and the rise of Zed
33:01 Should I switch to headless WordPress or continue rolling my own PHP templates?
37:33 How do you organize TypeScript types in a frontend project?
40:55 How do I continue to level up as a developer?
45:36 Stay in a comfortable job or embrace new challenges?
Hit us up on Socials!
Syntax: X Instagram Tiktok LinkedIn Threads
Wes: X Instagram Tiktok LinkedIn Threads
Scott: X Instagram Tiktok LinkedIn Threads
Randy: X Instagram YouTube Threads
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Wes Bos? Welcome to Syntax. Today we got a potluck episode for you to ask where you bring the questions. We bring the answers. And if you've got a question for us, please submit it. Go to syntax.fm, click on potluck in the menu bar and submit your question. We'll answer it on a future episode. Today, we've got some really interesting questions around, should I keep using WordPress? Or should I be going headless and doing it all in JavaScript or something fancy? Is the gravy train over for software devs? A lot of people are having a hard time finding a job right now and things are just switching up. Where to store your TypeScript types? Do you put them in a separate file? Do you put them in the file where you have them? It's a big mess and we're going to figure it out for you. Finally, there's a bunch more questions,
0:43 Wes Bos? but one other interesting one is like, what is the benefit to buying SSL certificate? When you can just get a free one with Let's encrypt. Scott, we just got back from San Francisco. We did our GitHub meetup. Props to everybody who came out for that. That was a hoot, wasn't it?
0:58 Scott Tolinski? Man, a big time hoot. It was a hoot nanny.
1:02 Wes Bos? Couple hundred people there. There was a couple of funny things. Let's talk about all the funny things.
1:07 Scott Tolinski? Talk about the funny things, yes.
1:09 Wes Bos? First of all, there was another meetup, like on the other side. We like did this brewery outside and there's like a brewery in the middle, like a little, like they pour beer there and we were on one side. And then there was another meetup on the other side, which it was like a meetup for like investors or something like that. And some of those people came to our side just unknowingly. And they were just like talking to me for like a couple of minutes. And I was like, this person doesn't know what we are. Like, why are they here? And then they're just like, so are you part of Y Combinator? And what are you investing? I was like, oh, I think you're at the wrong meetup. But then they like continued to talk to me about investing, like maybe there's an opportunity.
1:50 Scott Tolinski? Maybe they're, hey, yeah, no kidding, right?
1:54 Wes Bos? What else? There was, this is not funny, but some kid who's 14 years old came, took a bus in.
2:00 Scott Tolinski? Yes.
2:01 Wes Bos? Amazing. That was amazing. So shout out to, I forget his name, but shout out to the 14 year old that came out. That was really cool.
2:09 Scott Tolinski? If you're out there on YouTube, leave a comment. That was really great of you to come. Thank you so much for showing up. We also had folks from Z editor show up. We had Robbie Russell was there. There was like big name people. A lot of Jason Langsdorf, a lot of really great people showed up, Ashley Willis. It was just like. Just really great people and a lot of great conversation, a lot of big syntax fans. We had the park security there was loving our merch. One guy, one guy on a bike, park security.
2:42 Wes Bos? It was a police officer.
2:44 Scott Tolinski? He was a police officer.
2:45 Wes Bos? Scott scared to back the blue here, but yeah, it was.
2:48 Scott Tolinski? No, he was, I don't know. I thought he was a, he was a, he was a full on police officer. Okay.
2:51 Wes Bos? Oh, yeah.
2:52 Scott Tolinski? So, yeah, he, uh, no, he wore our internet hat. There was other people who were just like coming by and being like, hey, I see these hats over here. So we gave out so much gear. It was crazy. Even the VC people were coming over and being like, oh, this gear is actually rips. So we had glow in the dark syntax stickers.
3:09 Wes Bos? Those were sick. I, you have to send me a few because I brought, I brought one home and I didn't think to grab more. And my kids were losing it. And I was like, oh shoot, I guess I should have got more.
3:19 Scott Tolinski? I, uh, I didn't bring any home. So sadly, all of the ones that I had are, are gone. So I think they, they went really fast. So I don't know if they're still there. The Sentry folks might have them. So maybe next, uh, event we can go to, or, uh, we can get them to send us some or something. Because really cool. The stickers seriously glow. Let me tell you this, when you have 300 glow in the dark stickers in your suitcase, apparently that really wigs out the security machine. The guy pulled my suitcase aside, which is like not that uncommon, right? But when he like opened it up, he was like, there's something really weird right here. It was just this giant ball of light. And I was like, I wonder if it's these stickers. He's like, I don't know about that. And I was like, well, they're glow in the dark.
4:03 Scott Tolinski? And he was like, oh, but it's just, it looked like this, like giant blob. He's like, there's this giant mass in your suitcase.
4:10 Wes Bos? I wonder what it looks like glow in the dark in a, what's that? What is an X-ray machine?
4:16 Scott Tolinski? I don't know what they use. What's the tech? Yeah.
4:19 Wes Bos? Like, I guess so. Like sometimes don't you have to like drink a glow in the dark liquid before you get? Like an x-ray on your bones or something?
4:28 Scott Tolinski? Yeah, then you can have that like skeleton, you can just go in the dark skeleton.
4:32 Wes Bos? That's cool. So don't eat our stickers, but.
4:35 Scott Tolinski? Don't eat them.
4:36 Wes Bos? Might be kind of cool. That's funny that that got.
4:38 Scott Tolinski? But shout out, we also did a little pizza at the end. Everything was great. Awesome time meeting up. So thank you, everybody who came out. Can't wait to do another one. Wes, we should do it in Toronto this year.
4:46 Wes Bos? Yeah, Toronto or like, let us know in the comments below which city we need to hit up. So we, we hear Seattle a lot. We want to do Michigan at some point. We'll do every city, but let's know in the comments below. Especially Europe. We gotta do a European tour.
5:01 Scott Tolinski? Let's do Paris. Let's do, uh, where are the beautiful places?
5:06 Wes Bos? Amsterdam.
5:07 Scott Tolinski? Let's do Mallorca. We'll do a meetup in Mallorca. Who's, who's coming?
5:13 Wes Bos? First question from debug dev says, I got an email from Bluehost extolling the benefits of paid SSL over presumably let's encrypt. Is there any real benefit to picking a paid SSL over? Let's encrypt, especially on high-end and non-commercial sites. This is an interesting question. As far as I understand, the only benefit to using a paid SSL is like organization certification, but I have not seen that in a browser in years. You remember, you used to have a browser where you would like go to eBay and then the whole bar would like light up green and be like eBay Corporation.
5:54 Wes Bos? I guess that could be a way to show that it is associated with a corporation to get that. But aside from that, and I don't even know if that's that big of a deal, right? Like the domain name is the most important verification thing. Aside from that, there's no benefit. I think that that's, that's an old thing. They used to make a ton of money selling SSL certificates and it's not necessary anymore. There's no security benefits or anything like that, you know, and as traffic that is sent encrypted over SSL is, is exactly that. It's not more secure.
6:28 Scott Tolinski? Yeah, not more secure because it just has a bad, it used to be like a big trust and verification thing. And even back in the day, I remember like being really stoked to have that badge that was like protected by Kubota SSL.
6:40 Wes Bos? Oh, yeah.
6:41 Scott Tolinski? That was like such a legitimizer that you needed that, uh, that, that badge on your website, the little lock, the web 2.0 style lock or something like that.
6:51 Wes Bos? When I see that now on a website. I think like, oh, man, these guys are like corpo, you know, like that almost makes me not want to use it.
7:00 Scott Tolinski? Yes.
7:01 Wes Bos? But like, I guess for regular people, they need to see a lock or something like that, you know, like antivirus scanned website.
7:07 Scott Tolinski? It's almost better nowadays that like, it's now just something that is to be expected and transparent in that like, people don't have to know what it is.
7:17 Wes Bos? Yeah, it doesn't.
7:18 Scott Tolinski? It warns you if it doesn't exist now, you know?
7:21 Wes Bos? Like the benefit of an SSL certificate does not protect you from like a lot of the other things, which is like, like loose database credentials or hackers and things like that. You're not being protected from any of that. You're simply, an SSL certificate stops your data from being intercepted between your browser and the server that you're sending the data to.
7:43 Scott Tolinski? Yes.
7:44 Wes Bos? And that can even happen on your work. If your work makes you install an SSL certificate on your thing, I use an app called Proxyman. And you can install your own SSL certificate in the middle and literally just intercept that traffic and you can see it.
7:58 Scott Tolinski? Yep. Beware of loose database engineers. All right. Next one. Wahid. Hi, syntax. Do you guys think the position of a pure front end developer still exists? Like two or three years ago, I would get bombarded by recruiters about front end positions and now I get zero. And when I search for front end jobs, all I see is full stack and backend jobs. And for all of the different frameworks, languages and stacks, I'm at a wit's end to know what to learn. Do I think this job still exists? Absolutely. Because I know even as far as Sentry hires front end specialists, absolutely. But front end specialists these days aren't just, you know, a CSS or an HTML person.
8:38 Scott Tolinski? There's somebody who understands, react really well. There's somebody who can really dive deep into all of that stuff. I think these jobs absolutely still exist. Full stack developers are much less common in larger companies where there are more engineering talent and the backends and the data side, the full stack side of things are much more intense. Like the scale at which the data exists in Sentry, you couldn't have.
9:06 Scott Tolinski? A front-end developer working on.
9:08 Wes Bos? No, that's not even like a back-end dev, you know, that's infrastructure engineer. It's unbelievable the scale out it is.
9:15 Scott Tolinski? Right. And so the lines are definitely blurred because even with server-side rendering or whatever, there are aspects where you're having to interact with the backend of things.
9:24 Wes Bos? Yeah.
9:24 Scott Tolinski? But to me, a lot of that is not like full stack, like full stack is doing it all, tip to tail.
9:32 Wes Bos? Yeah.
9:33 Scott Tolinski? And I do think like UI and like front-end jobs absolutely still exist.
9:39 Wes Bos? I think so as well. However, if you're, bigger corporations are going to have separate people, but I think if somebody, if you're being hired at a company that just has an app or builds apps or whatever, in many cases, you're going to be expected to do like front to back because like the lines of like what's front-end, what's back-end is relatively blurred. You know, you gotta just query the database server, render HTML, jump into the client and hydrate something, add a date picker with some JavaScript. Like we had this question a while ago, what it was, can I just do CSS anymore and not do JavaScript? I don't think that that exists. There's certainly you can be a specialist and there's people out there that are know CSS performance really well, like Harry Roberts
10:23 Wes Bos? is probably one of them, you know, just performance nut, knows everything about CSS and rendering and how all of that stuff works. Those jobs certainly exists, but they are much less common than someone is like, we just need to build an app and you have to build the app for us and whatever you need to do to make that happen, that that works. And then as it gets bigger into like bigger infrastructure and Kubernetes and clusters and sharding, all of that type of stuff, then, then it starts to drift off. But I would not scare away from just like calling yourself a full-stack dev, learn a thing or two about how servers work and how requests come in, how databases are queried. I think you're going to get a lot further, especially this next question we have here is just like.
11:08 Wes Bos? There's a lot of people looking for jobs right now. And if you can do it all, you're going to be higher up on that list. So this is an interesting question from Pookie. And this is just something that I have heard and been asked, even at the meetup alone. I was asked this question similar to this like three different times. So the question is, will this industry become any more stable and friendly for software devs? After about 10 awesome years and most of them listening to the podcast, thank you. I'm seriously considering leaving to become an electrician between AI, outsourced jobs overseas, and a return to office, which I live in a small town, which I don't have many options. I don't think the industry is stable to provide for me and my family.
11:51 Wes Bos? The job did let me go and I've been looking for six weeks and it seems terrible out there. Some engineers spending nine months looking for jobs. Even if I get a role, any boss could decide to cut costs at any moment and it seems scary. What do you think? Will the pendulum ever swing back in our favor? This is a big question. I wrote a couple of thoughts down here as well. Like, first of all, I think that we had some good gravy years in this industry. Good gravy. There was people making silly money, people doing a bootcamp and making 250 grand right out the door. Facebook, Google, just gobbling up people who knew what they're doing. They're smart people, but just didn't have the, didn't have the experience.
12:36 Wes Bos? And, and that gravy train is, is over now. So a couple of thoughts here. Everybody in our industry is, is freaking out right now, but I feel like every industry is freaking out right now. You, you think about like, teachers are scrambling to figure out what to do with assignments. How do you make assignments when people can just chat GPT something, you know? Business analysts are freaking out because you can summarize stuff really quickly. Photographers are freaking out because like styles are changing and also like the tools, it's, it's easier than ever. Like, you should see the sick ass photo I took of Scott with portrait mode. That's definitely possible.
13:20 Wes Bos? But like beautiful sunset Scott smiling thing. I'm an idiot. I just touched the portrait button and and pushed the button on my phone and it it poops out like a like a unreal image like that. That's that's pretty good, right? That's dentists are being bought out by private equity, you know, like I see so many of these dentists that are like, or like Juno vet, whatever. And it's like, I don't know, you're just like killing the entrepreneurial vet by squeezing everything out of these small vets. But I think on the flip side, there's lots of people are probably already commenting down below mean things. There's a lot of F'd up drama going on, but also lots is changing.
14:01 Wes Bos? So I think the answer is somewhere sort of in between. And and I think if you were to just try to move to somewhere like electrical, I don't think that you will feel like that is is more stable, because I think every industry is changing right now. Everybody is is feeling shifts, not just because of AI. That's that's part of it. People are freaking out and trying to figure out how to use that. But also, there's just like political things and there's manufacturing. There's all kinds of stuff going on right now. People's attention spans are shorter than ever. Short form video summarized content, you know? And I don't think the answer is going to just be moving to another industry. I feel like every industry right now is feeling a bit of a shake up.
14:43 Wes Bos? So that's that's my thoughts of what's going on here. So so what should you do? Become an invaluable person, right? People that are invaluable are people who can think big, can architect systems, they can tackle problems, they can stand out just by talking to people about like what's going on with hiring right now. I just put a thread out on Twitter being like, what's going on? And everybody's saying the same thing is that it's there's just so many people looking for jobs. There are lots of good people that cannot get hired because there's so much noise right now. And and AI is making that absolutely worse because everybody can generate all this slop and just send out 10,000 resumes. So I think that how we hire is going to change.
15:27 Wes Bos? And as always has been being able to stand out from the rest of those people who are trying to just cheat the system is going to be. It's going to be the move. And how do you stand out? Right? Think big, learn how to architect systems, build your own thing. You've you say you've been off for six weeks. Six weeks is a long time. You could build
15:45 Scott Tolinski? Yeah.
15:46 Wes Bos? like a life changing app that not necessarily life change, but you could build something really good in six months. If you can find the energy to do that, you could absolutely slam dunk it. So I don't know if changing systems is the move, but I think changing your approach to standing out and becoming hired is definitely needed.
16:05 Scott Tolinski? I think this approach that you're referring to has always existed. It's as a hack, is a hack to get more better jobs. Because there were times when like hiring was really tough before the whole Y C bump and all that stuff where there's
16:18 Wes Bos? Yeah.
16:19 Scott Tolinski? the startups coming like I remember back in the day, it was tough to get like I was fighting with a lot of people for an HTML job. The things that work then are always going to work. Like you said, becoming an invaluable person. If you have that time and you have those skills, you're making things that those things could get noticed, but they could also become a financial backing for for you. You could do so many different things. But like you said, I think it's really important to emphasize that every industry is feeling some kind of crazy craziness right now. When I talked to Dr. Sarah Bird, that's all I asked her about these changes. And she was just like, every single industry is feeling these
17:01 Scott Tolinski? changes. And it's just like, these people get it. It is. It's a situation right now. So I loved all of your advice, and I don't have anything important to add. I just want to emphasize, reemphasize it all.
17:13 Wes Bos? Totally. I just was looking through the replies on that tweet I put out. And Pierce Boggan, he is on the VS Code team. I met him at GitHub Universe. And he says, from a hiring manager's perspective, we recently redid our interviews for the code product team. So anyone who works on VS Code, they got to hire for that. They all start with an app building exercise. They can use any AI developer tool they want, which is super interesting. You get to learn a lot about how everyone uses different AI tools differently. So some spend 30 minutes planning, then pass
17:45 Wes Bos? AI and finish the exercise in five minutes. Some iterate much more with an agent. Some prefer to stay in the editor. So this is kind of really interesting is like, instead of being like no AI in the interviews. It's the opposite and being like,
17:58 Scott Tolinski? Do what you want. Yeah.
17:59 Wes Bos? How, how do you approach things? How do you tackle systems? How do you tackle? How do you, how do you solve problems? I think it says a lot about a person of like, yeah, there are people that are just gonna make me a website that does cats for whatever. And then you get some slop out the other end, but there are very smart people using these tools that are figuring out how do you approach these problems? Then he goes on to say is then we have two interviews focused around getting to the root of the developer problem. And designing a solution to solve that problem based on real feedback we've gotten from the team. So that's again, I just, did I not just say that? Problem solving, systems design, those are the bigger problems and they always have been in software design.
18:41 Wes Bos? And, and just like writing the actual HTML tags is, is probably not going to cut it anymore. Whereas previously it might have.
18:49 Scott Tolinski? I think good interviewers have always known this, that like the people who understand the fundamentals of how things work, but also how to fix things when they don't work. They understand problem solving, debugging, they understand like people who can, like you said, architect and all that. People who, who have been good hires all along have like understood that those things are super important. Which is so funny because I, I remember applying for jobs that like, they would ask me a question about, in one particular case, it was backbone. And Marionette was the stack and they were like, how would you do this in backbone? It's basically just like a, give me the syntax of how you would do this in backbone.
19:32 Scott Tolinski? And I was just like, well, I can tell you how I would do it without backbone. And I could tell you where I would go to learn to do it in backbone or whatever. It's like, they don't, they don't like that as an answer, even though that was a better answer than, oh, you write the dot set function or whatever, you know? Yeah. Shortsighted.
19:49 Wes Bos? It's never been about memorizing the method names and it's always been about tackling these problems, which gives me the question is like, how do you get better at these things? Sorry, reps. Yeah. Put in the work. Six weeks. You got six weeks to build maybe six different products. You can, you learn a hell of a lot about architecture and design in six weeks, building six different things.
20:13 Scott Tolinski? That is, that is so right. Everything I've ever learned that was valuable was through some kind of hard exploration into something, going back and forth. And, and AI makes this easier now, not because you can have AI do the work for you, but you can use it as a bit of a sounding board. You can use it to research things for you. You can use it to explain docs or blog posts. And obviously, you have to be a little bit cautious with that, but it's never been more accessible than it is right now to, to learn everything. Next question here from Bleming. Hi, longtime listener. In a recent episode about ZED, Scott talked about automating
20:54 Scott Tolinski? versioning with GitHub Actions. He talked about change sets, but looking up, looking that up brought me to solutions for monorepos. In the past, I would have used a simple npm version, major minor patch with Git tags to match, but automating this would create a merge conflict since the package.json gets modified. We currently have a Docker build flow to pushing to Docker Hub, but we have stuck to using global tags like staging latest development. We love to get versioning up and running, could properly version the project in GitHub and Docker Hub. Could Scott talk a bit more about his process and maybe share his GitHub Actions? Yeah, so I have shared the GitHub Actions. It's in the notes here. I was like, huh, change sets a monorepo tool? Because I've only ever used it outside of a monorepo. And
21:39 Scott Tolinski? I went to their, I went to the readme for change sets immediately, and woof, change sets for such a simple, stinking project? This is always drives me nuts when you go to a project like this, and it's like, here's a tome, here's 10,000 links, here's all this stuff. And it's like, the very first sentence, "A way to manage your versioning and change logs with a focus on monorepos." And like, sure, I think by putting monorepos so front and center there, yes, if you have a monorepo with a lot of packages, this can be handy. You can just always, also just straight up use this. So, Bleming, I'm going to point you to one page in specific on this, which is just the intro to using change
22:23 Scott Tolinski? That's the only link you need. They should just put that on the dang read me as the very first thing. Oh, how do you use change sets? Oh, you run npx change set CLI in it. It gets you going. Then you run npx change change set CLI. It creates a change set for you. You modify it. You take you say what the changes were, what type of change it was. Then you run change that sales to live version and then that applies it. It creates a change log. It's like a two or three step process. And this read me, granted, like, I understand that this tool is way more powerful than that. But like this read does such a good job of making this seem so much scarier than it actually is.
23:06 Scott Tolinski? I've also included in my links that my YAML for two projects, one of which is a neither of which are mono repos. By the way, in case you're wondering, one of which uses change sets just in the repo itself. And then one of which uses change set versioning within a GitHub action and I have the action linked. But in reality, the code for that is literally just npx change set version. That's it. It's like a one liner. It applies that and then it emerges that back in. So give that a check. There's a little bit more there, but change sets is a pretty simple tool. What it does is it generates change log for you. It bumps your package versions. It keeps things nice and organized.
23:47 Scott Tolinski? The thing I like the most about using change sets is now that I have a package that other people are using. I can enforce in a GitHub action, hey, you have to have your change set. So that way, if somebody wants to commit and somebody wants to help this project. All they have to do is submit a PR with a change set and that version is going to get bumped. It's going to run through NPM publish on its own and I don't have to do anything other than looks good to me.
24:12 Wes Bos? Talk about a good use case for agents though, in in a pull request. Cause like, nothing is worse than like when you send a PR to some thing like that clearly fixes a problem. And they're like, well, can you please like submit a change set and do all this other bullshit stuff? And it's just that sucks. But now you can simply just like at change set bot, please generate a change set saying X, Y, and Z. Man, like that, that is why a lot of people don't use these tools is just like the overhead is a little bit more. And then you have to go back locally and whatever. Now it's so much easier, especially like you're, oh, you're just on your phone. Yeah. Change that bot, please do that.
24:53 Scott Tolinski? Even not even change that by, you could be like, on the GitHub pull request, you can assign Copilot. So you could say, "I need a change set for this, assign Copilot," and it's gonna make one for you.
25:03 Wes Bos? Yeah, that's what I mean.
25:04 Scott Tolinski? Even if it's not perfect, chances are it's gonna be better than what I would have written by hand.
25:07 Wes Bos? Oh, yeah.
25:08 Scott Tolinski? I'm often using AI to write my change sets.
25:10 Wes Bos? Even just the fact of making the file, having to like, go locally, make the file, push the changes up, that's annoying.
25:15 Scott Tolinski? And if you want to see all of the errors in your application, you'll want to check out Sentry at sentry.io/synthax. You don't want a production application out there that, well, you have no visibility into in case something is blowing up and you might not even know it. So head on over to sentry.io/synthax. Again, we've been using this tool for a long time and it totally rules. All right.
25:39 Wes Bos? Snowy Owl says, "Using an IDE that is not VS Code, what are your views on that? I hear good things about Zed and Warp having its own IDE. Curious if you have switched from VS Code or not. Being hesitant to swap is mostly because of the rich ecosystem of VS Code extensions." Yeah, this is, we just did an entire episode on Zed. I also talked to the WebStorm people at the meetup. They gave me...
26:03 Scott Tolinski? Yeah, a lot of WebStormers.
26:04 Wes Bos? Now I have a Zed and a WebStorm sticker on my laptop, so I'm wrapping them both. The Warp one is interesting because Warp is now not just a terminal, but it's like an actual like UI for editing stuff. You know, it's kind of a mix of like Claude Code, but it has a much better UI for it, in my opinion. I don't still, don't just love working directly in the terminal, but that's pretty neat. And then Zed, yeah, I'm, man, I'm, I'm pretty excited about it, especially once we did that episode with Scott and he kind of explained that the ecosystem has caught up on a lot of the problems. So what do you think?
26:43 Scott Tolinski? Yeah, I've been using Zed ever since we recorded that episode and I'm almost embarrassed that episode came out and it did because like a day or two afterwards, they fixed so many of the things that I had that were like, like so many of the UI things that they were, they like nailed and now there is like a UI for updating your settings and stuff. Not that I don't care about, I like a, or a JSON settings page, but like sometimes the UI is nice. Now there's a UI for that. And like they've really nailed a lot of like little things here and there within the latest updates of Zed and it just keeps getting better. So I, I've been continuing to use Zed. That said, I'm the type of freak that's changing what I'm using 24/7.
27:23 Scott Tolinski? I probably changed my browser twice since we recorded this episode and when it's released. And so I'm always bouncing around on different stuff. That said, the things that I've been really liking about Zed is just how many of the things I would reach to extensions for exist inside of the editor. I'm looking for this tweet desperately right now because searching on Twitter is impossible. So I asked somebody, somebody said, I can't go to Zed.
27:49 Wes Bos? The Grok is so good for finding tweets because you could just describe what you, what you're thinking of. Like at GitHub Universe, there was this woman who, who did this like music code thing and it was like a, like a techno.
28:05 Scott Tolinski? Her name is, is it DJ L?
28:08 Wes Bos? DJ Dave. And I didn't know who it was and I didn't know what the tool was. I knew Scott knew, but before I bug Scott, I just typed into Grok being like, there's something where it's like a JavaScript to make music and they were at GitHub Universe and they're like, DJ Dave did it out with this thing called, I forget what it's called again. What's it called?
28:28 Scott Tolinski? Oh, Strudel.
28:31 Wes Bos? Strudel.
28:32 Scott Tolinski? Is it Strudel? It's a German word, right? Strudel. We love Strudel.
28:36 Wes Bos? Yeah, Grok is very good at finding tweets. Sometimes.
28:39 Scott Tolinski? Yeah, I didn't find it.
28:40 Wes Bos? What do you, what are you searching for?
28:42 Scott Tolinski? Either I had a tweet about, or we had something about the Zed episode and somebody said that they liked Zed, but. The extensions that they used did not exist for it. So I said, hey, what are these extensions that you're using? And they responded with a big old list of a bunch of extensions. I will say, most of them, if not all of them, like, were features that exist inside of Zed without needing an extension. And many of them actually exist now within VS Code without an extension. I think we've gotten into the point where VS Code has existed for so long. We're using all these different extensions that you've just had installed for so long that you don't even ever think, did the editor go back and add this feature?
29:25 Scott Tolinski? So I know a lot of people think that there's some things that they need their extensions for that they just straight up don't. Now, granted, I'm not, I'm not saying that all of their extensions do exist or everyone's extensions do exist. I will say, I'm not missing a whole lot in the Zed ecosystem lately. I'm just not.
29:43 Wes Bos? I find it's helpful to sometimes just like reevaluate before you like, just, just try it, you know? And maybe you find out you don't need it. And maybe you find out something that is like, ah, that was a killer feature from VS Code that I cannot do without. But sometimes, even with like the whole like prettier in ESLint and then switching to something like Biome or what's the new Veet? The one from Veet, what is it called?
30:08 Scott Tolinski? Oxland.
30:09 Wes Bos? Oxland, things like that. It's just like, they're trying to get to parity because that's what everybody, the whole industry is built on. But sometimes maybe it's worth just like reevaluating and saying, ah, maybe I didn't need all those 8,000 plugins for my, my Linter.
30:24 Scott Tolinski? Yes. Yeah, like some of them are like Claude code for VS Code. Oh, so Claude code is baked into Zed, Codex baked into Zed. Yeah, GitHub pull requests, that still doesn't exist, but the Git, Git stuff is good. Git Lens, do you use Git Lens?
30:37 Wes Bos? Oh, man, no, no, I don't, I use Error Lens.
30:42 Scott Tolinski? Yeah, Arilens for sure.
30:43 Wes Bos? Who needs that? Who's, who wants to see while you're coding, who authored each chunk of code?
30:50 Scott Tolinski? I think GitLens does a lot more than that, but it also is heavy. It's very heavy.
30:56 Wes Bos? Yeah.
30:57 Scott Tolinski? Yeah. So there's a lot of extensions. I personally have been, like I said, riding with, with Zed for a bit. And I gotta say, there's some really, really great things about it. My computer no longer feels like it's taking off 24/7. And I do love VS Code still. I will say that. So no shade on VS Code, but there's been something really nice about using this.
31:18 Wes Bos? That's fresh. Man, the thing I keep, keeps bringing me back to Cursor is
31:24 Scott Tolinski? Cursor.
31:24 Wes Bos? Tab completion. Yeah. And like the tab completion and the like next tab where like it knows where you want. Auto imports things when you use them. That stuff is so good in Cursor. And like I keep, I keep going back to VS Code for a day or two and their tab completion is not as good.
31:43 Scott Tolinski? Zed's isn't as good either. Yeah. Zed's auto completion or tab completion isn't good, but let me tell you what's better about Zed. Is it so much faster with code actions. There's just such an inherent latency in some of those things like auto importing or any of that stuff. I found to be exquisitely fast in Zed. Even like something as stupid as spell checking. Man, spell checking in VS Code is a pain in the ass. It is probably the extensions or whatever, but in Zed, it's just so simple just to add to global dictionary, whatever. And it's just so fast.
32:18 Wes Bos? My spell checker in VS Code or Cursor, it like the little light bulb when there's a spelling mistake. It like puts it in the wrong spot for, I have some weird zoom setting set up, but yeah, that is a pain in the butt.
32:31 Scott Tolinski? It's a pain in the butt.
32:32 Wes Bos? How is that not native in, in VS Code? It should be. Cause I have, I have a list of like 800 words or something silly like that of just coding words.
32:42 Scott Tolinski? Yeah.
32:43 Wes Bos? Don't actually exist in the English language.
32:47 Scott Tolinski? Spell check, man. Uh, agentically. Yes. Oh, there's so many. Uh, Tolinski. Yes, that's a good one. Never gets. So, sorry, snowy owl. Give it a try. Use it. You might like it. It's very good. Uh, Weigart from, and I believe this is the Netherlands. I'm doing a dumb American thing where I don't recognize the flag. Uh, Weigart from the Netherlands. "Hi, guys. I find myself, uh, listening to syntax quite a bit. Even though I'm primarily a WordPress developer, most of my customers want WordPress. I've often considered going headless and switching to a more modern stack, at least for the front end. Being a WordPress dev is not exactly something to brag about these days." Hey, Weigart, as long as you've got customers, you can brag about it.
33:31 Scott Tolinski? "But I have never been convinced that I would win anything by doing that. Sure, building a front end in a framework like Astro or 11T might provide a better DX in some areas, but it complicates things in many other ways." Yes, it does. "Queering data, handling authentication and server management. I need two servers instead of one, one for the backend and one for the frontend, all becoming more difficult. So I'm curious, would you guys recommend a switch to a headless approach? Or should I just continue rolling PHP templates like it's 2010? I lean towards the latter, but I'd love to be convinced I'm wrong." Hey, listen, I think it's more viable than ever to just ship this type of thing because we have page transitions now.
34:12 Scott Tolinski? We have way better usage for all kinds of front end stuff with CSS. So you can get really great experiences. And as long as your app flies and is fast and looks modern, it doesn't matter what you're building it in. I really don't think you need to do a headless approach here. Every single time I've done a headless frontend to a site like Drupal or WordPress or..."
34:39 Scott Tolinski? Magento even like I have regretted that shit almost immediately. Uh, because you do have to re-implement so many things. Now, there is a time when I found it to be nice. Uh, if you have like a static site and a static site works really well for this stuff. But if you have a site that's being like constantly content edited and stuff like that.
34:58 Wes Bos? You gotta rebuild it every time, every time they change something. Yeah.
35:02 Scott Tolinski? Dude, your clients hate that shit. They hate it. They also hate that they now have two URLs to manage. They hate that they have to pay or potentially host two different things in two different places. Yeah, I saved it. Why isn't it on the site? You gotta wait for the site to build. Okay, what year is this? Yeah.
35:17 Wes Bos? The huge reason that people use WordPress is that it's comes batteries included and it has a massive ecosystem of plugins. And you will be giving up all of that if you go headless. You're giving up like you would be re-implementing current page in the nav. Uh, current page parent in the nav, you know, you'd be re-implementing like you'd have to rebuild the website every single time that you have there. And that seems like it sucks, especially like, oh, I want a quick store. Boop, boop, boom, WooCommerce, done. You're selling stuff in like 10 minutes, you know? Whereas like, oh, now I have to pay $900 a month for Shopify.
35:53 Scott Tolinski? Add the card. I gotta do this or that. Yeah. To me, it like if you want to do like UIs that aren't like that are, are Astro or something like that, like I wouldn't bother with WordPress as a backend and I'd go straight to something like Superbase or something.
36:08 Wes Bos? Plus, plus you can just write like I would probably spend my time on just like getting a modern WordPress development workflow.
36:16 Scott Tolinski? Yes.
36:16 Wes Bos? In place, you know, you got Git, you have a whole deployment platform, you have a whole data syncing thing. You figure out how to do your image generation so that it's on a CDN, your proper caching, page transition. There's so much of like the web that you can just apply to WordPress and then just stick with old faithful as like a, a battle-tested CMS.
36:37 Scott Tolinski? Yeah, whatever you do, do it well. Uh, yeah, you'd love to be convinced otherwise. I'm not here to convince you otherwise. I think if you're, you're skilled in that, keep rocking it. So.
36:45 Wes Bos? No, that the headless benefit is for, for companies that.
36:51 Wes Bos? Need like a CMS and and then they need that data to be pulled into multiple locations, you know, like Yes. Sanity is a really good example. People who are using Sanity will pull that data into their mobile app. They'll pull it into their like they pull it into restaurant like signage in the past. They need like a single spot for all of their content, you know, data as well as all of their images and videos and all of that stuff. And they can pull it into multiple spaces and or they want to pull it into their own application, you know. But if you're just building a straight up website and you're not using like a like a page builder, like a Squarespace or whatever, yeah, I probably would stay there. Word.
37:32 Wes Bos? Harry says, how do you organize TypeScript types in front end projects? Do you keep them local to each file as needed or maintain a dedicated types directory shared on global definitions? As my project grows, I find my type definitions becoming more scattered and disorganized. So I'm curious about best practices. So what I'll do is if I'm like typing like an entire module that maybe doesn't have types, that goes in a type directory. If I have something that is global to the entire project, maybe it lives on global scope, maybe it lives on the window, that goes into like a types directory. So I'll almost always have a types directory in there or at the very least like a types file. I'll often lean
38:15 Wes Bos? towards just using like a types dot TS file so that it's regular JavaScript instead of just like a D dot TS file. Yeah. There are cases for both, of course. And then as I'm like building out stuff, I'll often try to keep the types that are specific to that file in that file as well. And if I find that it is like, like, for example, I'm, I'm working on something right now that has like a constants file and it has a, has a bunch of different types for what those constants look like. I'll just keep them alongside. I'll just export the type of those right, right alongside where they are defined. I feel like that's a little bit better. Where it starts to get hairy is where you're like importing like both ways from different files.
38:55 Wes Bos? That kind of gets really annoying. And if, if you're using them in many, many files, probably more than five files, I'll then do a TypeScript has very good refactoring tools. You could just use the TypeScript refactoring tools and, and put them, bring them out of that file, move them to your types folder or types file. And then you can update, it will just automatically update the rest where those are referenced in your file.
39:19 Scott Tolinski? I've been leaning more and more on like, I know this isn't practical in every use case, but like you got sources of truth. So you have like your drizzle schema and then you're inferring your select types, you're inferring your, your insert types.
39:33 Wes Bos? Yeah.
39:34 Scott Tolinski? And then you use those pretty much everywhere and then you're combining them or whatever. To me, like as few definitions as possible for the same data. Like if I have a show type, I don't want the show type to be defined in eight different places, just ever slightly differently.
39:57 Wes Bos? Oh, no.
39:57 Scott Tolinski? I'm just gonna use the show type.
39:58 Wes Bos? Once, ever. Yeah.
39:59 Scott Tolinski? Yes. Once ever for any given type, because that can happen.
40:02 Wes Bos? Or none. Infer it all. Like.
40:04 Scott Tolinski? Infer it all. Right.
40:05 Wes Bos? I'm working on a function right now that generates steps. It's an array of arrays and each item in that second array is called a step. And I didn't define a separate type called step. I simply just, I'm inferring it based on the returned value of that function. That can bite you in some cases, but in this case it's beautiful. So I don't have to explicitly go type it and then hope that I match it up in my return value.
40:29 Scott Tolinski? Yes. Yeah. So I know sometimes you can get into it though, where you're in a component and you're like, oh, I'm just going to define the types for this thing while I'm here. And then you just handwrite them and the next thing you know, there's drift and you have.
40:41 Wes Bos? Yeah.
40:41 Scott Tolinski? Things that are the same. Yeah. So.
40:44 Wes Bos? CSS variables are like that too, don't you think? Where it's like, oh, that color's a little bit different than I had previously.
40:52 Scott Tolinski? Yes, they can be a bit like that. Sadiq, I'm a developer in a midsize company in a small town in Czechia, and I've been doing this for about eight years. We've always built custom JavaScript apps, mostly React, Vite and Nest JS. We also create small libraries to make our development easier, mainly common functions, components, et cetera. And with AI, it's become even smoother, almost like autopilot. But I've always thought one day I could become not just an app developer, but also give back to the community, build some cool libraries and maybe make more low level things. I know the recommended path is to study more, dig deeper and contribute to open source, like fixing bugs and gradually leveling up. I wonder how much is actually under my control?
41:35 Scott Tolinski? Is it just about being a really good and consistent? Or is there also a big element of luck involved? For example, not every actor ends up in Hollywood, no matter how hard they try. So what percentage would you say is luck? And what percentage would you say is an effort? And how would I know if I have it in me to become that kind of expert? Siddiq, there is nothing other than just doing it. There's nothing other than just doing it. It took me a long time to learn this, but just put stuff on GitHub. I've been putting stuff on GitHub for a long time. My, one of the, like, the very first things I ever made that, like, I put on GitHub was dumb. It was not like Still is. life-changing. It's not, well, it's not still is.
42:20 Scott Tolinski? Don't say that. That's not nice. I, uh, but, you know, you're, you're probably more right than not. What it was, it was a, is a package that simply just faded in elements on a page sequentially. And I called it DTF Down to Fade. And it was a jQuery plugin. They, thank you. And you just put it on there and it just faded elements in. I could tell you offhand, like, nobody used that thing. But I, I put it out there and I did it. And now I, I just do that all the time. If I'm making something, you just put it on GitHub. Maybe you post it on like a Discord channel or something like, I, I wrote a service worker for local caching in Spellkit. What did I do? I put it on the Spellkit Discord and say, hey, I just made this. I'm not looking for any notoriety from that.
43:03 Scott Tolinski? But like, people see that and that might be like, some, that might tickle someone's fancy, right? Somebody might be like, oh, I need that. I just, the other day, I've been thinking about this even more myself because I have all these internal tools. I have my own CSS framework that I use for everything. It's like, so I, I published my CSS framework. I made a little doc site for it and I shared it. And people were, were excited to see it. So it's like, you just do it, man. There's, whether it's luck or whatever, it's not about seeking notoriety. It's not about seeking any sort of fame or whatever. It's about like, just That's what the web's for, man. You made something, just put it on GitHub, throw it up there, and then feel free to share it. If you share it and nobody likes it, no foul. Yeah.
43:45 Wes Bos? It's amazing how the people that put themselves out there are seemingly very lucky. And the harder they try, the more lucky they get. So it's not about steadying up and becoming better. It's simply just throw your stuff out there. Something is gonna catch on as useful. People are gonna start using it and you're gonna learn a ton by fire by figuring out, "Oh, man, now we need..." Like, what is the thing we talked about earlier? Contributing?
44:09 Scott Tolinski? Change sets.
44:10 Wes Bos? Change sets, yeah. Now we need change sets. Now I need a release process. Now I need like all of this stuff. And you're gonna just like, you're gonna get really good really quick once these things take on. So yeah, put yourself out there. That's probably my entire career is owed to myself just putting myself out there, probably before I should have, and it's worked out.
44:31 Scott Tolinski? Yeah, and you can feel very self-conscious about doing that. But I would recommend just trying to release some of that. Because I feel very self-conscious about anything that I publish these days, no matter what. And I publish a lot of stuff. And so there's always gonna be people who find it to be not useful or not interesting, but there might be somebody who does. And that's great.
44:52 Wes Bos? People are not going out of their way to find stuff that is not useful. Like, you know, people are not going out of their way to find stuff they don't like.
45:00 Scott Tolinski? I don't know about that. Modern internet, people's like mostly just talking about what they don't like.
45:04 Wes Bos? That's true. You ever like look at the comments of like plumbing tutorial for beginners? And someone's like, "I've been plumbing for 37 years." And like, "What are you doing watching a plumbing tutorial for beginners?" Move along.
45:16 Scott Tolinski? Right, yeah. What are you doing? Yeah, I always hated that. When I had like a tutorial that was very clearly aimed at beginners and somebody been like, "Why are you doing it this way?" "Because I'm teaching it." I'm not saying this is the best practice, but I'm teaching how to understand the concepts. That's not for you. It's not for you.
45:35 Wes Bos? Question from confused agency CTO says, I've been working at my current company for 10 years. We started as the second person that's sent us a question after 10 years. We have a small team of 10 and we're focused on custom software development for SMEs, small to medium sized enterprises. So it sounds like they're like an agency and they crank out custom software. Today, we're the digital division of a consulting firm with over 1,000 employees. I hold a CTO role. Wow, that's, that's pretty high. A thousand It's a big role. CTO, 1,000 people, mainly because I've always driven innovation and had broad knowledge, thanks to my passion for learners. However, I feel overwhelmed. Many technicians prefer sticking to familiar methods and don't share my enthusiasm for continuous improvement,
46:19 Wes Bos? which makes most of the projects rely heavily on me. Also, sometimes I wish I had someone to learn from. Right now, I learn through my own effort and research and I worry that if I ever move to a product based company, my knowledge will be, might be too general, not as specialized as someone who spent 15 years doing the same thing. This makes me consider switching from consulting to product development to gain depth and to stay competitive. I could move to a company with less workload, more pay and fewer responsibilities. Sounds good. I'd lose the flexibility and freedom that I've owned here. I'm facing the classic dilemma. Stay for comfort or make a change to embrace new challenges. Any recommendations?
47:01 Wes Bos? Now, this is always an interesting question that we get. It's just like, what do I do? Sounds like you're kind of tapped out of this. The fact that you even went and wrote this question and asked us, sounds like you're probably considering something like that. It is very tempting to always stay where you're comfortable and stay what you're always doing. But man, I know a lot of people, even at this past GitHub Universe conference, I bumped into somebody who I did consulting with like 15 years ago. And now they're doing, they're doing something totally different. It's like, man, that person does a great job at being able to adapt to these different types of products. And I also, I've never met a CTO at a product company that
47:45 Wes Bos? It feels like their their skills are not broad enough, you know, there's so many problems that you hit and so many things that you can you can get in on. So I would say go for it. I think this is probably a good, especially if you're feeling a little burnt out, you're feeling a little overworked and you want a new role. People are always looking for someone that's hungry, always learning, always wanting to try new stuff. And 10 years at an agency that doesn't like to try new stuff, that seems like you have put in more time than most would have. Agency work is tough.
48:19 Scott Tolinski? Yeah, I did agency work for a long time and I really loved it. And let me tell you that the differentiators between the agencies that I love to work in and the one that I did not like to work in, because there were two I really liked. One agency, Wes, like 2,000 people, big agency. One agency, 12 people, very small. It was the one in the middle that was like 40 people that I really didn't like to work in. And the reason why I really didn't like working for that agency was because no one gave a shit. This was like well past the era of icon fonts and vector icons. And I was still getting raster icons. And they were just like, the designers were just like, huh, this is what it is. I'm like, you didn't have a vector version of this icon for me. The icon is pixelated in your PSD and like
49:02 Scott Tolinski? some of the pixels are cut off. Like, you do not care. You're so checked out. And all the other developers were like, doing who knows what, playing on their phones at work and stuff. Like, I'm here to build good stuff. And it feels like I'm one of the only people that is here to build good stuff. So I totally get you. Go with that passion, go with it, man. Yeah.
49:23 Wes Bos? Especially right now. Like, right now, everything, we just had a question. Everything is changing right now. And if you're excited about this stuff, there's certainly a lot of people who are just like, I don't feel like having to learn all of this new stuff right now. I don't feel like changing up everything we've done. And a lot of these businesses are just like gutting core parts of their product and switching them over to a lot of this AI stuff. So if you're excited about it, man, there's probably lots of companies that are looking for someone like you. If it doesn't work out, don't blame us, because it might be a totally wrong decision as well. I don't know. This is just a short question.
49:58 Scott Tolinski? No responsibility for this.
Previous 955: SvelteKit has solved data loading Next 957: CSS: Advanced and Obscure Page updated November 1, 2026 · transcript and summary are machine-generated