866: 2025 Web Development Predictions

Scott and Wes look into their crystal ball to predict what’s coming in web development next year. From the rise of on-device AI to the vanilla CSS comeback, Bun’s big moves, and React’s evolution, this episode is packed with bold predictions and hot takes! Show Notes 00:00 Welcome to Syntax! 02:00 Brought to you by Sentry.io. 03:09 The agenda. 03:40 Temporal Javascript api will ship in Safari and Chrome. Temporal Proposal. 06:23 On device AI. WebGPU API Dawn Native WebGPU 10:26 Models will plateau. Bolt.new, v0, Lovable.dev. 13:40 Web Awesome will become the most used web components library. Web Awesome. 15:57 We will be using more web components. 16:59 A push towards the ‘standard stack’. 19:38 We can really use relative color. 21:39 Vanilla CSS comeback. 23:35 A complete Mixins / Functions API for CSS. 24:27 Conditionals will ship in all browsers. 25:50 People will still make vertical centering jokes. 27:08 VSCode will be feature parity with Cursor. 28:22 Framework choice will matter less with AI tools. 29:12 OpenAI will launch a browser. Dupe.com. Buy Now! The Shopping Conspiracy. Krazy Binz. 37:18 React will drop Babel. BabelJS. 38:05 React Server Components will pop. 39:46 Remix will relaunch as something entirely different. 41:11 React Native will have it’s time. 42:06 Svelte will get component-based islands or data loading. 44:19 Server Runtimes, Bun will continue to do non-standard, lovable things. 44:44 Bun will release a PAS to compete with NPM, Vercel, and Vite. 46:06 Laravel will release a CMS. 47:57 Vite will stay king. 48:03 Rolldown ships in the next version of Vite. Rolldown. Statamic. 49:35 Sick Picks & Shameless Plugs. Sick Picks Scott: PHILIPS A19 Ultra Definition Dimmable Light Bulb. Wes: Stats App. Shameless Plugs Scott: Syntax on YouTube. 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

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Wes Bos?

  Welcome to Syntax. Today, we have a predictions episode for you. Every single year, we do a predictions episodes early in January to talk about what we think is going to happen in web development, what we think is going to change, things that will go away, things that will get popular. And then at the end of the year, we always review them and see how right or how wrong we were with these types of things, because we're usually pretty good at guessing these things, but often it's we're hilariously wrong, which just goes to show that you shouldn't listen to anyone. Everybody's just guessing, you know, especially if someone has something to sell you as well. My name is Wes Boss.

• 0:42Wes Bos?

  I'm a developer from Canada. With me as always is Scott Tolinski. How you doing, Scott?

• 0:47Scott Tolinski?

  Oh, happy new year. Happy new year, new year, new me. I am, yeah, I'm dealing with some struck 30 kids at home. It feels like kids are sick just 24/7. That's like the name of the game.

• 0:58Wes Bos?

  Like that. Yeah.

• 0:59Scott Tolinski?

  Yeah.

• 0:59Wes Bos?

  That sucks.

• 1:00Scott Tolinski?

  But yeah, no, feeling, feeling like I'm stoked for, well, even just like whenever you go over the contents of this episode and you start to like make your predictions. It's like, yeah, it's like to me, I get really excited because I'm like, oh, I could potentially get all these things this year. That's really exciting to me. And just even talking about it makes it feel like they're already here. So yeah, no, I'm stoked. Yeah, for sure. And I want to, I want to push back was I think we have a actually a very good hit rate on these episodes. I think.

• 1:31Wes Bos?

  Oh, yeah.

• 1:31Scott Tolinski?

  I think we almost always get them right.

• 1:33Wes Bos?

  Yeah. What percentage would you say our hit rate is?

• 1:36Scott Tolinski?

  That's actually great. Great question. Over the years, I don't know. I would say we get close to 80.

• 1:42Wes Bos?

  Yeah.

• 1:42Scott Tolinski?

  Yeah.

• 1:42Wes Bos?

  Yeah, 82 is I think where we're at. So 20% of the stuff we tell you will not happen. Maybe it's just because we're optimistic, but let's, the way we're going to do it is we're going to go back and forth, giving each other our predictions, and then we're going to sort of react to each other's and say what we think is, does that sound good?

• 2:00Scott Tolinski?

  Yes, sounds great. But you know what else? We have a greater than 82% of success rate in closing our bugs with Sentry. That's right. We have like a, if we actually look at our Sentry, which we do very frequently, we have 100% hit rate because it makes it easy to find and solve our bugs. Uh, just kind of endlessly easy. Not only do you get a code replay, which is, I think, just a supremely underrated feature of being able to see exactly what happened, when it happened, so that way you can fix it. Man, that, that feature alone, because you know, it's as a visual person, just reading through a bunch of stack traces and logs and stuff like that, isn't my favorite way to solve bugs. But if I can see a little, somebody clicked here and this broke.

• 2:41Wes Bos?

  Or even just the, the weird states. Remember we had that one with the player? We're like, how the hell did somebody make that error happen? Like, we could not do it. And we just watched them click around the website. We go, oh, I see what they were doing. You know, I didn't, I didn't expect somebody to do that.

• 2:58Scott Tolinski?

  Yeah, totally. For sure. So if you want all that and more, head on over to sentry.io/syntax. Sign up and get two months for free with the coupon code TASTYTREAT, all lowercase, all one word. So Wes, would you like to make the very, you want to just put your flag in the ground and make your, your first prediction here? And I'll tell you, actually, sorry, before we get going here, I'm going to explain to the audience how we're doing it this year. Wes and I both not seen each other's prediction lists. We both have our independent prediction lists. We're going to go back and forth, giving one of our predictions. If I also have it on my list, I will declare that I also have it on my list. Uh, and if it's on Wes's list and only Wes's list, then we'll just go keep going back and forth until all of our predictions are, are done.

• 3:40Wes Bos?

  All right. So my first one is one of the things we got wrong last year. And I am predicting that the temporal JavaScript API will ship in both Safari and Chrome by the end of the year. And we had predicted this last year and it did not happen a whole lot. Part of which was because the API was waiting on a standards body to implement time zone standards. But now I feel like it's, it's going to happen. And for those who don't know, the temporal API in JavaScript is an API for working with dates, times, durations, without having to necessarily tack them on to a specific date in time. So you could just say,

• 4:24Wes Bos?

  August 30th, right? That's, that's a date, but it's not attached to a time zone. It's not attached to a year, doesn't have a specific 12 a.m. attached to it. You can attach all of those things and there's amazing APIs for that type of thing. But the temporal API is going to, is absolutely massive and going to be such a game changer. To a point where I think like other languages will be jealous of how amazing temporal is.

• 4:49Scott Tolinski?

  Yes, it is quite truly the, the date system of my dreams. I've been using temporal via the polyfill for a bit. It is fantastic. There's great learning resources on how to use it. But yeah, you can use something like an instant, a plain date, a plain date time ISO. I mean, there's, it's really fantastic API. And I want to say, Wes, on my list, temporal will actually ship. So that is my, and I don't, I don't necessarily mean ship in all browsers, but I think it will ship somewhere. Because this thing is just too good and I think people really want it. And, and I'm going to will it into existence once again by putting it on my list here as well. So.

• 5:28Wes Bos?

  Yeah, I can't imagine how hard it is to implement though. Cause you know how much of a pain in the butt it is to work with time zones? Imagine having to write the logic for actually writing.

• 5:39Scott Tolinski?

  I don't want to do that. That said, you know, I will say that usage with the polyfill today is super clean. It's super good. It works very well and the polyfill isn't even that large. So if the polyfill works now, granted that's a far cry from getting it in the browser. I do think that, you know, technically these problems have been solved and it just needs to be agreed and implemented. And yeah, I, you know.

• 6:02Wes Bos?

  That's, it's already agreed on. It's, it's as far as I know, it's done. It's just browsers need to implement it.

• 6:08Scott Tolinski?

  Okay. Well, let's go, let's go already. People who aren't me, who are actually doing the work. That's how it is, right? Easy for me to say, let's go, let's push button. But yeah, so prediction number one, we both had it. Temporal will actually ship. Put it down. It's going to happen, folks. Here's a, here's one that I think is a pretty interesting. I think we'll get more use case and more commonly seen on device.

• 6:34Scott Tolinski?

  AI being used. I think right now, right? Anytime. And I guess that's, you know, you'd call it small models or whatever. I think right now, we are using on-device AI to do some things, but you know, I think it's going to become way more usable and ubiquitous. I mean, for instance, just recently I've been doing OCR and I don't know if you would technically consider that AI on-device AI.

• 6:58Wes Bos?

  Oh, yeah. It's, it's a, it's technically object detection, which is a form of AI.

• 7:05Scott Tolinski?

  Yeah. So I believe that that stuff is going to be way more ubiquitous and easy to use for developers. And therefore we will see people using on-device AI features way more frequently.

• 7:16Wes Bos?

  I agree. And the kicker for this is not only the size, like we're seeing a lot of like models where it makes sense. Oh, I could load that. That's not that big, you know, that makes sense to load that for that feature and run it either in the browser or on your, your user's device. But the other part of it is, is it fast enough, you know? Like, do you need a beefy server to run this? And we have a new feature being added both to the browsers as well as to server JavaScript, which is called WebGPU. And that's going to allow us to run these models on the user's GPU.

• 7:54Wes Bos?

  And, and then you get major performance benefits out of that because currently most of them are running in a Wasm build. And, and they can be fast, but they also can be a little bit slow, depending on obviously what the model does.

• 8:08Scott Tolinski?

  Totally. And for those of you wondering what the status of WebGPU is, is it's in technology preview in Safari and it's behind a flag in Firefox and it's been in Chrome for a bit. So yeah, WebGPU, it's coming for sure. And I think that will actually get us a lot of cool things beyond AI stuff.

• 8:26Wes Bos?

  So Google has a project called Dawn, which is a native GPU implementation. And then you're allowed, you can then use that. It's written in C++ and C, but then you can just use it via like native add-ons in Node and Bun. And then also Dino already has WebGPU shipped. So.

• 8:48Wes Bos?

  I'm just hoping it means I have to write less Python because every time I go down that route of trying to run some, I was trying to run Whisper X the other day to do some transcription, which is like a, like a better whisper. And like, I got like, like an hour in and I was just ready to throw my laptop out the window. All these pip, pipe, pip, install, everything. And it just, it made me so mad that it, like, people talk about no JS dependencies. Python is a freaking Wild West.

• 9:18Scott Tolinski?

  It is. And it's like you said, I went to pip install something and it was like, this package doesn't exist. I'm like, what do you mean it doesn't exist? I'm looking at the pip page right here. It's clearly there.

• 9:27Wes Bos?

  I've used like Conda and all these other third-party managers, but then I have like, competing versions and then I have to pip three and then things aren't alias properly and...

• 9:39Scott Tolinski?

  It's like, oh, this is available for only pip version 3.1.10 and lower, but not 3.12, even though, you know, 3.12 is what came on my computer by default or whatever. And it's like, come on. Yeah, I know. It's just weird. I don't know. Maybe it's probably just a skill issue, obviously, but yes, I do find it to be extremely obnoxious working in that stuff.

• 10:00Wes Bos?

  It is a skill issue, but I, I almost want to just run Docker for all of us. But then when you're running Docker, you can't access your computer's GPU all the time. That's a whole nother setup. So, pain.

• 10:13Scott Tolinski?

  Yeah. Did you have this one on your list explicitly?

• 10:16Wes Bos?

  I had just, I had the web GPU thing in there, which is similar. I think that, yeah, we're going to see, see that. But I also, here's another one of my AI predictions is that we're going to see a plateau of AI, meaning that we will, these models aren't going to get a thousand times better every single year. And a lot of the products, and I guess we're starting to see this already with things like Cursor, the products are really focusing on, how do you get the best possible result out of the models that we already have? And, and like, what does that mean for things like v0 and bolt.news?

• 10:58Wes Bos?

  Another one, lovable.dev that just popped up the other day where you can... Write text, text to UI as prototyping. I think that those will still become a lot more popular for quickly prototyping applications out. There's, there still will be a plateau at some point where you have to, you can no longer make the AI fix what it is that you want. But that's my predictions. AI is going to plateau, but the text to UI is going to get even, even better and even faster to a point where maybe even people will stop using like a Figma to prototype things. And maybe that's even already happening.

• 11:38Scott Tolinski?

  Yeah, I know. A lot of the stuff, I think you're right in the regard that it's not going to be the AI that itself, but like the, the way that the AI is being used and linked together to do different things. I mean, I mean, I saw a crazy demo today of one that was able to take video and separate video into separate layers. So if it's like, if it was, you know, recording a video of us in our office right now, it would be able to cut me out and have me on a layer and then my background on another layer.

• 12:10Wes Bos?

  And it would like fill in the background?

• 12:12Scott Tolinski?

  Yes. It was, it was wild. It showed some like kids playing on a beach and it had all, both the kids separate on their own layers and the beach. I just think about like the amount of work to mask and do stuff like that in the past. I think just things like that will continue to be wild. That might seem like wild that are using the current tools, but like really taking a different look or approach to it. I had access to GitHub. I guess not GitHub, it's Copilot Workspace. They're, they're very clear that you, you're not supposed to call it GitHub workspaces. You're not supposed to call it Copilot workspaces. You're supposed to call it Copilot Workspace. And Copilot Workspace is a, you know, basically text to PR kind of

• 12:52Scott Tolinski?

  system. And even though it wasn't perfect when I used it, I gotta say the general idea that-

• 12:58Wes Bos?

  You forget, it deleted the readme.

• 13:00Scott Tolinski?

  Oh, yeah. Oh, it did. It did much more than that, Wes. It did. It did a lot more destructive things than that. It was kind of like a, you know, a little demon in our codebase for a bit. But I gotta tell you, I do think that, yeah, I do think they're going to, I think they're gonna get it right. And I think that kind of thing is gonna be like, oh, you have a bug? Just toss it to Copilot Workspace and let it fix it for you, which, hey, I'll take that, right? Here's, here's one that is not AI related. That way, we can maybe go back and forth on some AI stuff, because I'm sure you have some more.

• 13:32Wes Bos?

  I think that's it for my AI stuff.

• 13:34Scott Tolinski?

  Okay, I have more, but I'll, I'll save them because I don't want to inundate people with too much AI stuff. Web Awesome will become the most-used web components library, and maybe not the most-used right out of the gate. But what I do think, maybe the language on this needs to be tweaked. I think Web Awesome is going to take off. I think people are going to use it. I think they're going to like it. I think they're going to talk about it. And I think it's going to cause usage of web components for little tiny interface things to really be utilized in the way that I personally feel like web components are best utilized. I had some chance to work with the alpha of Web Awesome, which is a-

• 14:13Wes Bos?

  Yeah, you want to give a quick rundown of what it is for anyone who didn't hear that episode?

• 14:16Scott Tolinski?

  You bet. Web Awesome is basically, you could think of it as like a web components component library. And these things have styling. They, they, they can have some default styling, but it's mostly UI styling. They can basically, you say, oh, I want a star rating, you know, like five stars. You click on one, it has an event. They feel like actual HTML elements. So I drop in, I want a rating component. I tell it that it's, you know, on, on select event or whatever, it passes an event. And bingo, bango, it feels like a real native element. Now, there are so many of these awesome, awesome components within Web Awesome

• 14:57Scott Tolinski?

  already. And I'm in the alpha, like I mentioned. And I used that one. It worked flawlessly. And it's, it's alpha. There's color picker, copy button, details, dialog divider, dropdown, format bytes, format date, icon. I could just keep on going, pop-up progress bar. I mean, a lot of these things are things that we have, but they're like- Nicer versions of them with more options that does more.

• 15:19Wes Bos?

  Yeah.

• 15:20Scott Tolinski?

  A split pane, a skeleton screen. There's just some really great, really great little web components here that are things that are going to be perfect for dropping in, especially because these work so well by just dropping in an individual script. Anytime you need one of them means you're not installing some big old library. You're not having it to be tied to React. These things can be taken other places. I used it and it was an extremely smooth HTML like experience. And I think other people are going to be like, oh dang, this is actually a really good use case for web components here. So yeah, I'm a high on web awesome for sure.

• 15:56Wes Bos?

  I, I had one in there. Let me look for it is we will be using more web components. So I don't think that the listener will necessarily even care that things like web awesome are web components. They just know that they just pop the tag and the tag goes bloop, bloop, bloop and does its thing. And I don't know that everybody listening will also author their own web components. But I think that these libraries of simply just being able to use them between like the Mux Player, the web awesome stuff and a couple of other implementations that I've been seeing it even like the Stream Deck. The Stream Deck, you build your own plugin for it.

• 16:38Wes Bos?

  All the UI components for working with text inputs and number draggers and things like that are all web component based. And I think that we are going to love using them. I don't know that we'll all write our own, but they will love to use them.

• 16:53Scott Tolinski?

  I will say I probably will not author a web component in 2025, but I will use them absolutely.

• 16:59Wes Bos?

  Uh, so along that sort of idea of web components, mine was, uh, there will be a push towards, and I don't even know if this exists, but it's called the standard stack, meaning that when all the features that you use is you try to opt for standards based first, meaning that like you try to use CSS scoping and CSS starting style and all of those like new CSS things that we got, right? And then you flip over to HTML and we have dialogue and pop over and CSS anchor, right? You flip over to. JavaScript, both on the client and the server, we have streams and fetch and web

• 17:40Wes Bos?

  requests and web response and you know, like just defaulting. It doesn't matter that you are, or even ESM, you know, just maybe just shipping ESM straight to the browser. And it doesn't matter that if you're using React or whatever, I'm not saying people are not going to use that. I'm just, I'm saying that the choices that you make are now often moving out of library or framework land and moving a little bit further out to the standard stack.

• 18:08Scott Tolinski?

  Totally. I could see that. Doing less on the framework, less in the framework, less very specific things. I think we all have a bit of fatigue from having to support the React way of doing all of the things, whether that is virtual DOM or synthetic events or any of that stuff, right? You're having to bring that along from React version to React version, updated. I know that people in the past probably felt more pain with this when we had to go from classes to functions, et cetera. But I do agree, and this is something, this is how I know you're right, is because this is something I was on for the past like two years now. And that's typically, you know, I'm typically a couple of years ahead of everyone else is how I like to think about it. I'm joking, folks, but I do think that

• 18:51Scott Tolinski?

  part of me working in Svelte so much is that you do utilize browser APIs so much more. You do utilize standards-based things more. And I do think that that is, people will open their eyes like, oh, wait, the browser has this included. Maybe we should be using this.

• 19:05Wes Bos?

  Yeah.

• 19:05Scott Tolinski?

  Because it's, it's nice, it's fast, it's easy. Yeah.

• 19:08Wes Bos?

  The amount of people that were excited about the form data API when React server components came out, it just shows you that, yeah, there's, not a lot of people know that there are primitives built into the browser for these things and you should certainly use them when possible.

• 19:27Scott Tolinski?

  Yeah, for sure. Cool. Yeah, I like that one. Not something that I had on my list, but I do, I do really enjoy that idea. And it's personally something that resonates with me very, very much. Here's one that is, let's do some CSS features. Here's, here's one. We'll see, we can do a couple of CSS features. You tell me if you have this one. We can actually use relative color.

• 19:51Wes Bos?

  Oh, that's, I didn't have that on there, but I'm, I'm now wondering, cause I am, what's relative color at right now?

• 19:59Scott Tolinski?

  I think it's here, but not, uh, it's here, but not, it's 88% support. So that feels like a lock. This feels like, this almost feels like cheating.

• 20:09Wes Bos?

  Yeah, because I was like, oh, should I just shift a website using relative color? Yeah, it's in, it's in every single browser and it has been for...

• 20:18Scott Tolinski?

  Yeah, not that long, but...

• 20:19Wes Bos?

  It's been in, in Safari for almost two years by the time you're listening to this.

• 20:25Scott Tolinski?

  No, it was released September 15th, 2024.

• 20:30Wes Bos?

  No, you see, the thing about you, you guys who just glance at can I use and then say no way.

• 20:36Scott Tolinski?

  You're talking about, I like, I like, I like my features fully implemented, Wes.

• 20:40Wes Bos?

  Okay, so the only part that you couldn't use from current color, or on relative color since March 26, 2023, was you can't use the word current color.

• 20:51Scott Tolinski?

  I like current color.

• 20:52Wes Bos?

  Absolutely. Everything else is, was fully supported for almost two years now, so...

• 20:57Scott Tolinski?

  But Opera mini does not have support for it.

• 20:59Wes Bos?

  Opera mini, get out of here. No one's using Opera mini.

• 21:02Scott Tolinski?

  I know, no kidding. But it is, I would like to see this at like 94% or something. I would like to see that percentage be a little bit higher. That's it, I am the type of, I, I would like to see that other people, other than us, uh, you know, boundary pushers will be able to, to use here.

• 21:17Wes Bos?

  Yeah, because a lot of people still are, like in the last episode we recorded, I talked about a bug I had to fix on iOS 15. And you certainly can provide fallbacks for older browsers, but that's a bit of a pain, you know?

• 21:31Scott Tolinski?

  Yeah, relative color is tough for that.

• 21:33Wes Bos?

  Yeah, all right. I think, I think that one will certainly happen. We're gonna start using that in all of them. My one along this thing is that vanilla CSS is gonna have a comeback.

• 21:44Wes Bos?

  Meaning that there are so many things that have been added to CSS, you know? Scope being relative color, uh, what are, there's a CSS anchors API, uh, starting style, what are the other ones? I, we, we go over this list every single episode.

• 22:02Scott Tolinski?

  We do, there's two.

• 22:03Wes Bos?

  There's at least... 10 features in CSS that are absolutely amazing. And as browser support becomes like, not just they're in all the browsers, but they've been in the browsers for a while now. I think we're going to see a lot of people being like, oh, wow, this actually does everything that I want. And, and just like you see people shipping ESM straight to the browser. I think we'll start to see people just shipping straight up style tags to the browser, which is, I know a lot of people still do that already, but I'm talking like no utility framework, no transpile or compiler, or nothing like that, just straight up.

• 22:43Scott Tolinski?

  That's me. That's how I do it. Again, I'm a little bit early on this one. I agree.

• 22:47Wes Bos?

  On what? On where do you, where do you ship straight up CSS?

• 22:50Scott Tolinski?

  Uh, everywhere on my own, on my own work.

• 22:53Wes Bos?

  Name one website.

• 22:54Scott Tolinski?

  Talindot ski. Yeah. Uh, Bboy Tools.

• 22:58Wes Bos?

  Okay, I'm going to make our new rule. You cannot use your, your personal blog as a, uh, an example for something like that.

• 23:06Scott Tolinski?

  Habit path.io. No framework, nothing.

• 23:09Wes Bos?

  Aren't you using, you're not using Svelte there?

• 23:12Scott Tolinski?

  Svelte, but what's Svelte doing for the CSS?

• 23:14Wes Bos?

  But are you, are you writing CSS in Svelte? And Svelte is processing and figuring out which bits go where?

• 23:21Scott Tolinski?

  It's just adding, it's adding classes for scoping, but.

• 23:25Wes Bos?

  Yeah. There, there. That's what I'm talking about. Are you pre-processing?

• 23:31Scott Tolinski?

  I'm not pre-processing it. It's just adding classes. That's, I don't know if that counts. Here, here's one I have. We'll actually have a complete mixes and/or functions API for CSS.

• 23:43Wes Bos?

  Yes, I had that as well. CSS mixes and functions will be firmed up, but not in any browser. So the, it's currently in intent author. I forget the names of it, but basically it's, the W3C has said, okay, we will now spec this out. And then, once it's been specked out, then it's, the browsers start to implement it to see, is this really how it should work? So we're still very early in mixes and functions, but I think it will be firmed up to a point where browsers will be able to start implementing it in 2026.

• 24:18Scott Tolinski?

  Yes, I'm very stoked for that and it does seem like there is some real work being done on that as we speak. So I think that's a pretty good prediction. I also have another one that conditionals, AKA, style queries will ship in every browser. So, style queries, container style queries. And the reason why I'm referring to these as conditionals, folks, is because you can basically use a media query or a container query, as you might say, a style query, using a variable as a value to check to see if this variable has a value of blank, then apply the CSS. That's conditionals to me, baby. So, yeah, conditionals will ship in all browsers.

• 25:00Wes Bos?

  It really replaces, any time you like add a class for like a version, like the simplest one is simply light or dark. Is this container light or dark? But there are many other use cases where you could set like extended or open, you know? You don't have to add an open state. You set an open CSS variable and then you can query the container. So when this container is open, then apply the rest of the styles. I think that's pretty nifty to be able to extend it past just classes and into custom properties. So, yeah, it's adding if statements to CSS.

• 25:41Scott Tolinski?

  Right. Function is in its states. Programming language. CSS becomes a programming language 2025. You heard it here, yeah. Do you have any other CSS ones?

• 25:50Wes Bos?

  My last one is people will still make jokes about vertically centering. This always kills me because we now have probably, I don't know, if I were to guess, probably 20 different ways to center something in CSS. And I even made a video that says, "I will die if someone says this joke anymore", because it just goes to show how out of touch a lot of people are with the capabilities of CSS. And here we are talking about starting style and all these amazing things that are being added to CSS. And then like the majority of the world is still talking about, "I can't center something in CSS." And that kills me because it's clear that people don't even know what they don't know.

• 26:35Scott Tolinski?

  Yes, I totally agree with this, specifically because even like I did a video on what? Using translate and rotate instead of transform properties and all the comments were, I didn't know you could do this. And that's been in browsers forever. Yeah, yeah, at this point, so, oh, yeah.

• 26:53Wes Bos?

  Yeah, that was, that was a really good example because you showed that you can apply two totally separate animations to have nothing to do with each other to the same element.

• 27:01Scott Tolinski?

  Yeah, you can compose them, it's great, totally. Alright, let's move into some non CSS ones. I have a VS Code one here. VS Code will be at feature parity with Cursor, AKA it will do everything that Cursor can do. And yeah, I don't know if that means I will use VS Code again because maybe Cursor will get better, but who knows?

• 27:22Wes Bos?

  Yeah, I agree with that one. I didn't have that on here, but I'm, I'm, somebody, I asked on Twitter for people's predictions and someone says, there will be no more VS Code forks. Let's guess now, will there be another VS Code fork?

• 27:37Scott Tolinski?

  Yes.

• 27:37Wes Bos?

  I'm gonna say yes. There's people getting some funding now, I bet.

• 27:41Scott Tolinski?

  Yeah, I heard something about like a React IDE, so at React Summit, there was a mention of something React IDE. And I don't know if that's reality or not, but I could see that.

• 27:56Wes Bos?

  It could be, you know why? Because like someone said, like React has such a benefit over many of the other frameworks. Because the AI has been trained on what 15 years of React?

• 28:09Scott Tolinski?

  Yes.

• 28:10Wes Bos?

  I don't know that that like may, maybe like a lot of these code generation IDEs spit out React as well as a default. So maybe we'll see something like that. I can't see that being a good idea, though.

• 28:22Scott Tolinski?

  I have actually related to exactly what you're saying here.

• 28:25Wes Bos?

  Yeah.

• 28:26Scott Tolinski?

  I have a prediction that framework choice will matter less with AI tools. AKA the tools are going to get better at scooping up documentation. They're going to get better at being able to know what version of what library you're using and maybe it won't have as much. Stuff to pull from like the amount of React code that's out there probably makes the AI tools much better, but being able to work with, let's say, Svelte 5 will be less painful than it is today because the AI tools will be better at consuming documentation.

• 29:03Wes Bos?

  And it could also be a benefit that there is not an extra 10 years worth of code that we don't use anymore to train the AI. I have on here, OpenAI will launch a browser.

• 29:15Scott Tolinski?

  And I think that is kind of interesting because people are trying to get as close to the user

• 29:22Wes Bos?

  as possible. And Apple probably has a pretty good spot here because they are the OS. You gotta think that Raycast is sweating a little right now as everybody's launching launchers right now. What happens when you install the app and it overwrites your user's keyboard shortcut? And then they no longer open your app anymore. So I think OpenAI will launch a browser so that they can really compete with Google as well as augment the entire browsing experience. It's kind of, I saw Microsoft launched a AI model that will take like a screenshot and

• 30:06Wes Bos?

  give you a list of like buttons and actions out of that screenshot and things to do. So I was like, oh, that's the AI is now driving the website. And then like even on Microsoft Edge now, when you search for something on the page, it will now suggest to you alternative searches. Because like when you search for something, it's just like a like a text match, right? But it now it understands what's on the page and what you're searching for, which I think is pretty nifty.

• 30:35Scott Tolinski?

  Yeah. And hey, I will want to say, I think that is, we can even put this, which isn't on my list, but I do totally agree with that sentiment is that the race for an AI browser, I think all of these companies understand that the AI browser and the AI search are

• 30:54Scott Tolinski?

  Is going to be who controls the next generation of that that's starting point of the internet. You open your browser, you start to search, or for me, I just swipe down and start searching. Whoever controls that experience is going to be the next Google Chrome or the next Google or whatever. And I do think all of these companies understand that, which is, it's actually very frustrating to me because that's one of the reasons that propelled the Arc browser to move on from the Arc browser to their new AI browser. But, you know, it's so funny, I'm on the the the Arc subreddit and those people are delusional. They're like, this thing is gonna flop so hard. I'm like, no, folks, you don't understand.

• 31:38Scott Tolinski?

  They they see a bigger picture than you see it. Their their niche browser is doing nothing for them. And I love that niche browser.

• 31:44Wes Bos?

  Yeah.

• 31:45Scott Tolinski?

  They they want to control much more than a niche browser. And then the only way they're gonna do that is by the next generation AI browser, whatever that may be. And and I'm sorry, folks, it's it's coming whether you want to resist it or not.

• 31:57Wes Bos?

  Yeah, it's it's the same reason why not a lot of people use Firefox. Because like Firefox is the superior browser for all technical people. But like regular people aren't using it as much, right? And because it's either not built into the device that they have, or it doesn't have the the features that that they're they're trying to use. And I think once all the AI stuff gets even further, you know, like people want to be able to search a screenshot, a TikTok and find the dress that somebody was wearing in their browser, you know, and have that suggest them to them.

• 32:34Scott Tolinski?

  Those tools are so crazy good now.

• 32:36Wes Bos?

  They're so good. Yeah.

• 32:37Scott Tolinski?

  Courtney is so much better at using them than me. We were at dinner the other night and our kids were drinking out of these like, they look like, you know, the honey bear, like the honey bear things. And they were glasses that they got lemonade and the kids were loving them. And Courtney is like, are we gonna be ordering these these bears? And we, sure enough, she did one of those apps and it found them on a restaurant supply store. And that's the only place they sold it. And I was just like, this is amazing. And it wasn't even a good picture. When she took that picture, I was like, there's no way the angle is not good enough for the AI to tell what that is. And it still found it, so.

• 33:09Wes Bos?

  I do I do all the time where we're me and Caitlyn are watching TV and I'll just, oh, that's a cool shirt. Where's that from?

• 33:16Scott Tolinski?

  Yeah.

• 33:17Wes Bos?

  Bam, like blurry sideways photo and it finds it. Amazing.

• 33:21Scott Tolinski?

  Amazing.

• 33:21Wes Bos?

  There's also this website, dupe.com, a lot of people are starting to use as well, where you can like, like, the whole furniture resale game is wild. And if you want to find like, oh, Wayfair is selling this thing for 50 bucks more, or six times, you know, a lot of these like fancy stores are just reselling stuff. But also, dupe culture is also bad as well, because it's ripping off designs.

• 33:48Scott Tolinski?

  Yeah, right, yeah, it's a, did you, there's a, there's a weird documentary on Netflix right now about like shopping culture.

• 33:54Wes Bos?

  Oh, I watched that, yes.

• 33:56Scott Tolinski?

  You watched it, it's, it like, we put it on thinking it was just going to be a normal documentary, and then it's like, very, I don't want to call it artsy, but it is very like, produced, yeah, it's very produced in an interesting way. And I can't help but feel like they would have had a bigger impact had that been a more normal documentary than if it was like, just kind of off-puttingly weird. That said, it was interesting to see like, all the extra clothes and all the washups. I don't buy any of that junk anyways, I'm definitely more of a save up, spend more on something that I'll have for a long time.

• 34:31Wes Bos?

  Buy once, cry once, yeah.

• 34:31Scott Tolinski?

  Yeah, like, how long have you seen me wear this hoodie for? Probably like, four or five years at this point, but it was a nice expensive hoodie, so.

• 34:40Wes Bos?

  I took my kids to one of these, have you ever been to one of these like, Amazon return bins places?

• 34:46Scott Tolinski?

  No, I don't even know anything about that.

• 34:48Wes Bos?

  So we have these stores called Crazy Bins here, and it's essentially, on Saturday, they basically, it's a warehouse, and it's just rows and rows and rows of bins, and it's all Amazon returns, or like, Gap returns and things like that. And we have a couple of stores that will sell stuff, but what, the way that this one works is they just dump everything into these aisles, and that, like, there's, there's like, four foot wide bins, and then on Saturday, it's 25 bucks, and on Sunday, it's 20 bucks, and then on Monday, it goes down by $5, until Thursday, it's a dollar. So you, you just walk in, and it's just, it, it made me like, like physically, not

• 35:31Wes Bos?

  physically, but it made me a little like sad or ill, being like, this is where all the, I don't care, I'll return it, and you know, or, oh, this thing broke, I'll return it, or, a lot of people, oh, this thing broke after a year, I'll just buy it again and switch it, put the broken one back in the box and send it back, and. And just garbage. And like you can see all the like TikTok popular items from six months ago. You know, all these like knockoff Stanleys that are no good or things that say like, I'm a Swifty on it. And, and, uh, yeah, I mean, made me a little sad coming through all that garbage. It's like, oh, some good stuff, but.

• 36:10Scott Tolinski?

  It is a, it is like, it is a problem. Um, I, I did find out, I never heard of this before. And it turns out there, there are only a few in Colorado. Closest one to me is in Aurora. But, I don't know. Yeah, I don't know. That's crazy. It's called Black Fridays, which, um, I suppose every day is Black Friday there or something.

• 36:28Wes Bos?

  There you go. Yeah. Yeah. It's wild. But like, uh, on, on the flip side though, there, it's, it's almost a cottage industry. Of every morning, there's people lined up and, and people run in and try to grab the stuff. And then it, it shows up on Marketplace, you know, it's like, uh, they're using every, every part of the animal, you know? There's this, like this sub industry of people that buy stuff and then flip it.

• 36:51Scott Tolinski?

  And whatever can prevent it from ending up just sitting in the ocean.

• 36:55Wes Bos?

  The market is being efficient, I guess. Cause I walk through and I'm like, there's nothing here worth the day, the day's dollar amount. You know, but I guess at, at the, we went on a $1 day as well. And it was just like garbage.

• 37:10Scott Tolinski?

  Yeah. All right. Well, let's get back to the web garbage.

• 37:13Wes Bos?

  Okay. Well, let's talk about React.

• 37:15Scott Tolinski?

  Hey, I got a React one.

• 37:17Wes Bos?

  No, just joking.

• 37:18Scott Tolinski?

  I got a React one. React will drop Babel in their compiler. The React compiler will eventually drop Babel here as a requirement.

• 37:26Wes Bos?

  Yeah. So.

• 37:27Scott Tolinski?

  What are we doing here?

• 37:28Wes Bos?

  The React compiler, which is used for primarily like memoizing components and, and making your things render less. It's currently just a, a Babel plugin. And that's not because it's, it's written in Babel. It's just like the easiest way for them to initially get it out. But I agree, it, we will probably see this integrated into the like Rust compilers and whatnot this year. And that was always the case as well. They said, this is just a Babel plugin right now. Cause it's able to get, we're able to get it up and running as fast as possible. But it's not the plan long-term.

• 38:04Scott Tolinski?

  Word.

• 38:05Wes Bos?

  I have React server components are going to pop. There's a lot of. Like, eh, not sure about React server components, you know, like a lot of like, not, not goodwill, but a lot of people are a little exhausted by the transition over to them. And like, like, I'll tell you, I'm on the other side as, once you understand how they work and the oddities and the weird rules, they're pretty awesome. And I think we're going to start seeing like, universal React components where it never before were you able to ship a component that has server logic, client logic, styles, API calls, you could just, you could pack that into a single component

• 38:50Wes Bos?

  and, and someone can just take it and put it into their own application. There's no setup or, or anything like that. So I think that we'll start to see some of the benefits of that type of thing. Also, I think part of the, the hardest part is people don't understand the benefit that it gives you. Like I, I put a video up on 10 stack start, which I think is another one of my predictions is going to pop. And then I was like, this is a lot simpler, you know, you can put it all in the same file. And it's just too many video. I didn't dive into the big differences of that. But a lot of people are like, yes, this is the way they should have done it. But like, the, the reasoning behind why React server components works that way is because of

• 39:32Wes Bos?

  waterfall and, and round trip to the client and back. And it's intentionally that way. And a lot of people either don't have those problems or don't fully understand the problem that they may be having.

• 39:43Scott Tolinski?

  Yeah. Okay. Sick. I like that. Uh, do you have any other React ones? Cause I have a, so, okay. You go for your other React ones. Yeah.

• 39:51Wes Bos?

  So React remix is, is going away and React router seven now has all the remix features. So React router seven is now a

• 40:03Wes Bos?

  Nextjs equivalent, right? And it's been kind of like unclear as to what will happen to Remix. And my guess here is that they are going to relaunch Remix as something entirely different. Just using the name for it. Whether that's going to be, part of my mind goes, maybe it will be like a Shopify thing. You know, like a headless Shopify React server component based. Maybe that's it. Maybe it's a Shopify React server component library. Or I can't, I can't fathom what else it would be.

• 40:36Scott Tolinski?

  Yeah, because if you think about it, what Remix became React Router, which you would think of as being like a smaller level component. So if they do end up thinking, like rethinking what Remix is, you'd think it would be something more encompassing rather than another small piece. Because that

• 40:51Wes Bos?

  Yeah.

• 40:51Scott Tolinski?

  React routing is

• 40:52Wes Bos?

  Like a CMS? Probably not a CMS.

• 40:55Scott Tolinski?

  Probably not a CMS. I do wonder if that's going to be the right call if they choose to do that to reuse the same name. So I, I have, I don't know. You know, we're in a world now where you AI something and it is a great name, but you know, you gotta be cognizant of those things. And maybe it goes along, you know?

• 41:11Wes Bos?

  The last thing I have here is React Native will have its time. Meaning that like Blue Sky's React Native, some of the top apps in the App Store are React Native.

• 41:20Scott Tolinski?

  Some of my best friends are React Native. I, um, I don't, I don't love React Native. Um, I'm going to be honest, I spent a lot of time with it. But I have to agree, there's some nice things about it and it does, um, definitely make the barrier to entry smaller. I have been using Tori to do web apps lately or, um, native apps, even though it's using a web view. It's not using native stuff and it's tough, man. It's a tough world overall. So yeah, I, you know, you'd hope that React Native would take off to some degree. I just don't know. I don't know if I also agree with that one, but I respect it for sure.

• 41:56Wes Bos?

  Yes. How are we going to measure it? You know, is it going to be NPM installs? We'll see. We'll decide next year.

• 42:02Scott Tolinski?

  We'll decide whatever makes us look better and we'll pick that one. I have one for Svelte. I think Svelte will get some kind of component data loading or Islands or something like that. I know those are kind of two different concepts, but I think Svelte, Svelte Kit specifically, now that Svelte is, is out and there's all these cool new things coming out for Svelte with the, the advent of Svelte. I, I do think that Kit is next to get not modernized because it's already fairly modern, but you see so many other frameworks now able to load data at a component base level or are able to do island-based architecture and choose what is server-side rendered and not. I just feel like they have to do something like that at some point and it feels like next year would be a good time for that to

• 42:47Scott Tolinski?

  happen. Me personally, I would love, here, here's what I would love. I would love more control over hydration. I'm going to say this component, just straight up text. No need to hydrate it. No need to have JavaScript be concerned with this at all after server-side or any of that stuff.

• 43:02Wes Bos?

  Yeah.

• 43:02Scott Tolinski?

  So, yeah.

• 43:03Wes Bos?

  I agree as well. And that was after trying out Tanzact. Very bullish on Tanzact. I think it's going to be very big. But that was the one thing that is not implemented yet, which is I, I want to mark something as a server component, right? Maybe that's, I just use server components for that. But

• 43:20Scott Tolinski?

  Yeah.

• 43:20Wes Bos?

  Yeah. I don't want this thing to be hydrated. I just want to send HTML from my server and have you put it on the page.

• 43:27Scott Tolinski?

  Correct. Yeah. And along that same regard, and in an opposite direction, I do, I do think local first stuff will like really take off. Now granted, we saw a big increase of people writing local first software in the past year. It went from like nobody talking about it to a lot of people talking about it, to even influencer types talking about it. I think that the tools here are on the precipice of being

• 43:53Scott Tolinski?

  Accessible enough for your average developer to pick up and just use that once it's like that, people are going to be really impressed by the experience. And because of that, for app-based type of things where you don't need server-side rendering, I think you're going to see a big resurgence in CSR for local-first apps and apps that are running entirely on the client.

• 44:18Wes Bos?

  Love it. Let's talk about server runtimes, Bun, Dino and Node. My, I don't think a whole lot will change in this space, but I think Bun will continue doing non-standard things that people love. So Bun, obviously, is working towards Node compat and standard space compat, but they also just roll out these random features where it's like, hmm, that's actually kind of nice, you know? So we'll see that. And I think Bun will release a platform as a service that will compete with like NPM, Vercel and Veet, all in one.

• 44:56Scott Tolinski?

  Interesting.

• 44:57Wes Bos?

  I, I think, well, that's gotta be the, like, they have VC funding, you know? You often think like, why are they, why are they funding a Java, open source JavaScript runtime? And it's, it's gotta be that they, they set their sights on like a, like a Vercel type company where you can, you can charge people to host stuff.

• 45:17Scott Tolinski?

  Yeah, that's an interesting idea. And I do think you're right to go down that rabbit hole. I do think that's probably something that's likely to happen. It'll be interesting to see. Bun, I feel like Bun is a crazy world. Every time I see something being added to it, I'm like, it's not, that's a wild choice. Whether that is like a positive thing or not, I do think it's a wild, wild thing.

• 45:39Wes Bos?

  I just went to the Bun website and right on the homepage, they're hiring a site reliability engineer.

• 45:45Scott Tolinski?

  Okay.

• 45:46Wes Bos?

  Somebody with expertise in designing and maintaining CI pipelines, especially for cross-platform builds.

• 45:53Scott Tolinski?

  Look at you.

• 45:54Wes Bos?

  Yeah. It's happening. I bet it will happen this year.

• 45:58Scott Tolinski?

  Look at you. Yeah. Huh. Cool. You know what, Wes? I don't have any else. Do you have any other predictions?

• 46:05Wes Bos?

  One or two more. Laravel is gonna release a CMS. You, and I'm saying this because I'm not gonna name any names, but somebody who was big in the CMS world just left their job. No, no. Someone who's big in the JavaScript CMS world knows a lot about this type of stuff. Just left their job. And if you look at their online postings, the type of people they're replying to, I think it's, they're talking to a lot of Laravel people, which makes me think Laravel is either going to launch

• 46:45Wes Bos?

  CMS or acquire Statomic or something in that world, you know? So I think it's gonna happen. You know, Laravel is making some strides right now and you gotta think that's part of the, a piece of their pie after they raised all that money.

• 47:03Scott Tolinski?

  Yeah. I think that, uh, that tracks. You know, there, again, VC money, follow the money. Something has to happen there. And I do feel like with all of this WordPress self-implosion stuff going on, now granted, most normal people don't care, but you know?

• 47:19Wes Bos?

  Yeah.

• 47:19Scott Tolinski?

  It even affected my wife and her, uh, her small little webpage for their testing center. So the fact that it's like reached outside of, you know, web development and is affecting normal, normal people, webmasters around the world. You gotta imagine people are gonna be annoyed with that.

• 47:35Wes Bos?

  Imagine Laravel forked WordPress? That would not, this is not gonna happen ever, but...

• 47:41Scott Tolinski?

  I think they should do something different, but yes, I agree.

• 47:43Wes Bos?

  Yeah, they probably, you don't need the, what, 24 years of baggage to come along with it?

• 47:49Scott Tolinski?

  There are some odd APIs with WordPress. I will say that it was never my favorite, even though it was so, you know, universally used and loved.

• 47:56Wes Bos?

  Yeah. I'm gonna ask you real quick about Bundlers. Um, I think VEET will stay king. We're gonna see VEET's Rust implementation come out.

• 48:03Scott Tolinski?

  Yeah. I think, you know, you said VEET's Rust thing, RollDown? I bet RollDown ships. I bet the next... Next version, here here's a here's a here's a risky prediction. I bet the next version of Veet, which they just they just released a major version. So I would imagine the next version of Veet will be using roll down for their build and their dev server. That is a risky risky bet here, but I'm willing to I'm willing to just make a little risky one here.

• 48:32Wes Bos?

  There you go. Oh, speaking of of I asked on Twitter like what people's predictions were. And I just got a reply from Statomic, which is a Laravel CMS.

• 48:47Wes Bos?

  Somebody predicted that Statomic will become a go to CMS over WordPress as we go behind the Matt Mullenway chaos and Statomic says we approve this prediction.

• 49:00Wes Bos?

  Something's cooking, something's cooking. Let me tell you, and I got I got no insights into any of it, but something's cooking, I bet.

• 49:05Scott Tolinski?

  It feels like a little insider cooking, if you're asking me.

• 49:08Wes Bos?

  Yeah.

• 49:09Scott Tolinski?

  I mean, insider is a you know, you know, something that you're not.

• 49:12Wes Bos?

  Oh, me? No, I don't. I don't.

• 49:14Scott Tolinski?

  Who me? No.

• 49:15Wes Bos?

  Stuff. Yeah.

• 49:16Scott Tolinski?

  Couldn't couldn't. Yeah.

• 49:20Wes Bos?

  All right, that's all of our predictions. Hopefully we are right on at least 82% of those. Tweet us at Syntax FM or or skeet us at syntax.fm on Boosky. Tell us what your predictions are. We'd love to hear them.

• 49:35Scott Tolinski?

  Yeah, that awesome. And yeah, that's all I got, too. Check us out on YouTube. We're all over that place. I got a sick pick here. I got a sick pick while Wes is looking up his. Folks, do you have dimmer bulb, dimmer switches and new new LED bulbs can get some kind of a buzz going on? And, you know, bulbs overall, there's like a whole world of good and bad bulbs. I have found the bulb that I prefer the most, so I will sick pick the bulb that I have no problems with, no troubles with. It's the Philips Ultra Definition Dimmable Bulb. I have these bulbs all over my house now. They you have them in several different brightness configurations or daylight or soft light. Either way, I've never had an issue with these bulbs.

• 50:20Scott Tolinski?

  They last forever. They're wonderful. They work with all dimmers, smart dimmers, any of that stuff. No buzzing whatsoever. No weird flickering or anything like that. They're great, fantastic bulbs.

• 50:33Wes Bos?

  All right, I'm gonna say pick something that you picked out in the desktop apps episode that we did, or two of the episodes. I keep going back to that episode, be like, Scott said one of these apps in that episode, and I need it. And it was, I was using iStats menu for the longest time, and I switched over to using one that, it's an open source one, just called Stats. Is that what it's called?

• 50:56Scott Tolinski?

  It's called Stats, yeah.

• 50:58Wes Bos?

  Stats. Just Google Stats GitHub, and it's gonna come up. And this, this is, it's really good. It's, it's just as good as iStats menu, has all the features that I want. I'm able to, I specifically use it for watching my network traffic and watching my CPUs. And then I also have been using it for when I'm testing out any AI stuff, and I wanna know, is it running on the GPU or not? I'll just pop open the little GPU meters, and if you see them go up or down, then you know it's actually using it. And also with the CPU, if, if you're trying to run, I was running conversions, and if you're trying to run too many at a time, it actually wasn't maxing out my CPU,

• 51:40Wes Bos?

  because it was trying to do concurrency, I guess, or how our computers work. But awesome little app. It's totally free. It gives you, put it, throw it in your menu bar, and you can see everything that's going on with your computer. And like, it shows battery level and RAM, and it shows you what app is using uploads. So, big fan.

• 52:00Scott Tolinski?

  Yeah, sick. Cool. Well, that's all we got for you today. What are your predictions? Leave them in the comments below. Do you think we totally missed on any of our predictions? Because let's face it, I don't think we did. But let's say, let's say you think we missed on any of our, leave, leave a comment. Let us know why you think we missed, or let us know what you think your predictions are, something that we totally neglected to mention. We'd be happy to hear that, because there's, this, this industry has a wide, wide, wide view of a lot of things. So their chances are, we did miss some stuff. But as always, thanks for listening. Thanks for watching. Catch you in the next one. Peace.

• 52:36Wes Bos?

  All right, peace.