Is React Really Dying?

All right. Okay. Yeah. Is React really dying? On fire. The React logo. So you all learned a lot of advanced React today, right?

Yeah. Okay. It turns out none of it matters because you will go on X, or you go on my comment stream on YouTube, this is the kind of stuff you get. React is dead. Crazy, right? Well, maybe, I don't know. I'm the kind of guy who wants to actually figure that out. So I figure, well, they gotta go somewhere, right? And they're not doing React. They can do Svelte, solid view, Angular maybe.

So I'm a numbers guy too. Maybe figure it out. So let's take a look at the numbers. So let's start out with Svelte. Let's see how Svelte doing? All right. Well, Svelte is getting 1.5 million downloads per week. That's real good. If you're running a open source project and you're getting 1.5 million downloads a week, you're doing pretty good, right? Absolutely. Okay. Let's try, Angular. How's Angular doing? Oh, Angular is pulling 3.5. Not bad, but a little bit of a kind of leveling there. Kind of interesting over the past couple of years. So definitely I would say probably not a competitor with React. Okay. How about Vue? And by the way, this is like the little app that I wrote, and guess what? In React, right? Of course. I didn't use HttpGrid.

That was an AD chart. Sorry about that. I should next time.

All right. So that's Vue. That's doing pretty good. I think that could be competitive with React, but do you guys think it's competitive with React?

Okay. Let's see. React dwarfs everything. 25 million downloads a week. I'm crazy. So actually I was like, well, maybe it dwarfs everything. So I added all of it up. All of React versus all of everything else, literally every other framework. And React is two and a half times more downloads than any, all of the other frameworks combined. Crazy. Right?

So let's talk about, well, actually, so React's the library. Let's talk about the framework. So let's talk about Next and Nuxt and all that and compare those, because that's, you know, when you put this into action, it's mind the gap. There is, there's React is not a standalone thing. You need to add something to it. So let's go and compare the frameworks. So let's go take like Next and React Router and compare those to how Nuxt and all of the other ones like Speltkit and all that, how are they doing? Incredible. 4.4 times the number. In fact, actually I would say React Router is a huge part of this. React Router gets a lot of downloads. Basically you've got the SPA is going to be React Router. Next is going to be your SSR. All right. But it's not just that it's us.

It's us downloading the React developer tools. 4 million users. It's 4 million developers. It's incredible. So with all this popularity. Yeah. Is it dead? Maybe, but it still sucks, right? That's another thing I get. It was dead or it sucks. Well, we say, okay, you've got all of these things. You don't use this. You spelled use, use solid, all this stuff's better than that. So do React developers think react sucks? Well, we actually had a survey for that. We had the state of JS, 2023. Everybody take the state of JS. All right. Okay, fine. All right. You really should. These are fantastic surveys. They really help folks figure out where the community is going. And the community in this case is really liking react. 71% of respondents say if they, well, that they knew about react one. And that they liked react. 28% said they didn't like react. So about three and three quarters. That's really good for our framework. That's our library. That's 4 million users. That's incredible. You can pair that to view.

You can bring that to angular. The only one that really comes close is smelt. It's got about the same number of. Enthusiastic positive reviews, but the salient big difference here is that the kind of light blue is a very odd chart. I gotta say, but people aren't actually using salt. They know about it, but they don't actually use it. Whereas we react, they use it and they like it.

So does that describe you guys? Do you like, and use react? Absolutely. Okay. Thank you.

Okay. How about pain points? So there was also the state of react 2023. It came out. Like barely at the end of 20, 23. Um, and there's an interesting thing here, right? You look at that chart and you're like, wow. People have all kinds of problems like Ford refs and memo and performance. Well, people don't notice is down there in the tiny little by the print fine print at the bottom. It says that only 1500 people out of the 13,000 that took the survey said that they hated anything about react. That's only 12%. It's actually less than we saw before with the state of JS and back to the state of JS ever since it was released 2016, people have ranked it as the number one, most positive framework over time. And that's just incredibly impressive.

All right. One last chart for you guys. So admired and desired. This is from stack overflow. They kind of have a very interesting way of thinking about this, but let's take a look at the top react, overall, incredibly positive sentiment. The next day essence third on the line. You're all these negative things about these frameworks. People really like it. I think that's something when you do think about all this stuff that you learned today, just think about that you're investing in a really good technology that people really like.

All right. So also learn that when you put react with boots on, you get a very creepy, weird image in AI or in Leonardo AI in this case. So he's not dead yet. I actually want one of these, because they're really cool. It's got like some plasticky kind of thing going on.

All right. So what's going on here, right? So this is, this is an odd difference between these things. You have an incredibly positive, everybody's using this framework and you got all these folks out there saying, ah, you know, it's dead. It's this and that. So as a YouTuber, I get to talk to folks a lot about the problems that they have in the comment stream. And I got some insights. So I've got three basic buckets.

Yeah. And a dry mouth of things that people don't like, or the issues that folks are having with react. The first is that they're just overwhelmed by the amount of stuff that you have to choose when you start working with react. So what would those be? The rendering, right? You got to choose. Are you going to use server side rendering, Klaison, SSG, static state generation, incremental site generation? There's so many options to choose from. Are you going to use a component library? Are you going to use Mui, like are down in the courtyard. Are you going to use ShadCN, Tailwind? Any fans of ShadCN out there? Yeah. Okay. All right. Cool. State managers. Anybody used Sushdand? How about Zustand? I'm kind of curious, actually. Is it Sushdand or is it Zustand? Okay. Sushdand? Yeah. Yeah. How about Zustand? Oh, okay. I've heard that. I did a video on Zustand way early days and I said Zustand because I'm an American.

I say it that way. And the German folks came out and they're like, no, no, no, no. You cannot say it that way. You got to say it Sushdand. So I learned that. And now everybody's like, Zustand? Like, what are you saying? All right. What are you going to use for an API layer? Are you going to use a API standard? Are you going to use gRPC, GraphQL, gRPC? God forbid. Sorry about that. Please don't use that. Uh, REST? If you do use gRPC, use twerp. It's really cool. It's from Twitch. Actually makes it halfway palatable to use.

A persistence layer. Are you going to use Prisma? Are you going to use, are you just going to go direct to the database? How are you going to store your data? And then how are you going to lay your project out? Is it going to be a mono-repo? Are you going to be sharing code? Are you going to use linting standards? Which linting standards are you going to use? I mean, everything is a choice and there are frameworks out there that are really nice in that they kind of make a lot of those choices for you. Like Angular is a classic example, but React having its kind of roots in working in whatever environment you have and kind of replacing your existing view management model, basically said, well, it's up to you, which is actually kind of cool, but also daunting because you might choose wrong, but I'm here to tell you that choosing wrong is actually okay.

So born out of my experience, I can tell you, first thing I want to do is use defensive programming, right? If you're going to buy into something, try to make sure that you can isolate it as much as you can. We talked about like a state manager. So you've got some cool business logic going on in your state manager. You have to like decide how to put stuff in a cart or something like that. Disconnect it from the state management mechanism as best as you can. So that if you decide later on, that you want to change your state manager to Tushdan from something else then you can do that. Another thing is that churn, that things change is going to be any part of the process of a healthy application. So you might, because of the specifications may change, you may get a cool new customer that's like, Oh, it's got mgraphql in it. And then you got to, how much speed on that. So that is part of the process. You got to make sure that you build that in. And then as things change, you got to prioritize catching up with it. So if you want to go from, you know, 4.1 to 4.2 cool, but don't over prioritize that. You don't want them to have the only thing in your tech debt be updating your package, Jason all the time.

All right. So number two is halfway through. So React has gotten too complex. Anybody hear this from folks? Oh, it's too complex. It used to be the simple statement, simple view library. It's gotten crazy. So one of the things I hear about in terms of complexities are the re-rendering mechanics. So when you think when you talk to your teams and you talk to folks who are starting out using React, this I've done a lot of videos about the basics mechanics of React. And this is what a lot of folks get stuck on. There's one thing I call like the magic template idea, which is that the return statement of that function. And the component is somehow a magic template that automatically gets rerun, but the code above it doesn't. So these are things that a lot of folks, when they're just starting out with React learn and they are and they're mistaken. So they need to figure out how to get around that. Or figure out what the right way is. Effects with use effect and also asynchronous programming. That is a big pain point for folks. And also this one, I hear all the time about client server rendering and also state management. And the confusion between the two. What does the state management certainly look like on the server? How does that sync with what's on the client? Some folks like think there's an actual, like continuous connection between those two things. There's not. Last year, I was here talking about the connection between, I think it was Redux and Next.js and how to properly use Redux in a Next.js context. It's complicated. It actually, it really is. So figuring that all out is really important. Or that's the stuff that folks struggle with.

So let me tell you a couple of strategies that I use or recommendations that I have for you. One is to know those React fundamentals cold. Like rendering, the rendering model, how components re-render, when components re-render, what triggers a re-render, what doesn't trigger re-render. You have to know those cold. Also understanding JavaScript and TypeScript, like how in particular, how objects and arrays work.

A lot of folks get messed up with memory management, objects, arrays, particularly nested objects, where they can mutate a value kind of deeply nested inside of an object and they get. Whoa, they get freaked out when like, oh, I changed it over here and it's somehow copied over there. What's happening? And so I did a whole video just on memory management, like going back to the absolute basics of like how a computer lays out memory and how that actually changes and how that actually works inside of JavaScript. That's a huge deal because folks, when they go through bootcamp, that's not something you learn. And it's something you end up learning by getting incredibly frustrated by it when you run into bugs with it. And then another thing I see is that folks overcomplicate their architecture. They bring in all kinds of libraries when they don't need to. So they'll start off with Vite and it was just, yeah, I'm not a couple of state managers or whatever, because everybody uses them and before they, you know what, they've got like just way too much code and they have no understanding what the heck a doll does. Another thing we, I see a lot, or the last one I see a lot of is innovation fatigue. The, the framework just seems like it's just moving way, way, way too fast, but if you look at the timeline, it actually isn't moving all that fast. We got hooks in 2019. We got RSC in 2022. That's three years to be able to work with hooks and integrate them into your system. We still get all this, you know, Hey, things are moving way too fast. Like I did a video recently called, is one the one? Now, one is a really cool framework. It's a single framework that you can use to build web applications and mobile applications in react native from one code base. It's really freaking cool. Did a video on it. Showed it, showed folks how it works. What I get. So many. Oh, we got yet another framework kind of things, but this is really cool. So as an older engineer, I've been doing this for like 40 years now. Learning how to cope with innovation has been something that I can now kind of pass on to you. So here's a couple of tips that I got for you first. Don't get freaked out when I, when someone shows you something really cool, like all this cool stuff that you learned today, you may not need it. You may not use it right away. So don't like, Oh man, I learned about this. So I have to use it right away. It may not be appropriate for your project, but understanding it's out there, that'll help you think about things like it could.

But think about positively, it could simplify stuff. So that might be a reason that you want to adopt it. Also, they're not trying to mess with you. Like they're not releasing stuff out in the wild. Like the react team is like adding in RSCs and the framework just to mess with you, right? You got to think like, Ooh, wow. That everybody is trying just to do the right thing. And then. So you want to see what, you know, what, what the framework and what, what this new thing is actually trying to accomplish, like look for the underlying motivation and make sure that it syncs with what you actually want to do. That's something you want to embrace.

And then finally use proof of concepts to figure out if it's actually going to work in your environment. Don't actually try and just integrate it into your app right away. That can cause a lot of problems and make it just a lot more complex. It's nice to actually work with things in isolation. So it's all about like thinking positively as opposed to negatively. And all of this stuff, this innovation and these choices and the complexity of it. Actually, when you think about it from the other way around, that it's not bad. It's actually good prove to us that react is a fearless framework, right? We've got 11 years of a library and they're still doing amazing stuff. They added react server components to an 11 year old library. That's incredible. It's also everywhere, right? Think all the places where you can use react web mobile. It's incredible. And also it's the first choice. Anytime that somebody is going to build a new state manager or framework, they're going to build it on top of react, but anything's not perfect, right? So we need governance. That's one of the things we saw, right? So Dominic was just up here earlier in the year. He had an issue with react 19 and suspense. Do you guys hear about that? Yeah. Okay, cool. Well, that actually kind of pointed out that the react team is making some decisions about where the rack framework is going that we might not appreciate. So cool. They're actually responsive to that.

And that's why react 19 is currently not out yet because they're still working through the suspense issues. So that's cool, but I think we also need a level of governance and then that showed it. Also, react has been at times personality driven. For example, they got Dan Abramov versus the world over RSCs. Wait, this is a big thing. When RSCs first came out, Dan was on Twitter. I convinced people that RSCs were actually cool and fine. And the reaction they got was really, really terrible. And that actually, it actually cost the react team some members and that's not great. So the good news is that with some change of the structure of that team now is less personality driven. We saw that with react con 2024 with the entire team was up there giving presentations about the various parts of it.

So what can you do? How can you interact with this team? How can you do these things? How can you work with react positively? Well, the first thing is to keep it positive, right? Don't don't quite like slam on react team members. Just keep it positive. The really important thing is to engage, re-engage with your community. Actually hit a Twitter poll up and ask folks, has your local meetup come back from COVID? So question for you, has your local meetup come back? OK. In Portland, mine has not. So it gets, you know, maybe 50 people where it used to get like 200. So I really, really encourage you go to those local meetups. Speak at local conferences. A friend of mine today is going to a local AI conference with like a right here in London with like 80 people in it. And those are the kind of bigger conferences that feed into the bigger conferences and the kind of place where you can get an opportunity to speak yourself.

Next is contribute to open source, right? All of the cool ecosystem that we have around react is because people like you had great ideas for cool libraries and brought them to life. All right, well, that's what I got. React is nowhere near dead. I'm Jaher on YouTube. Come on over and subscribe. What I want to know is you're here and you are a YouTuber, but that's something you can make a living off. I can switch to being a YouTuber. Can you make a living off YouTube? It's really hard. Yeah, actually.

Yeah. YouTube ad revenue is terrible. So the only way to really make a living on it is either have videos that go for millions of views, which only like basically Fireship does in our space. Or you do sponsorships. And I think that's actually like a recent kermuffle over. Fireship was the fact that you did a sponsorship with AppRite and people didn't like that all that much. But you know, I guess that's fine. People got to make money.

Yeah. Yeah. We want great content. You want great content, somebody actually has to pay for that. Yeah, at least I I'm thinking you like a roof over your head. I hope you have a roof over your head, yes. One day. Yeah. Cool.

All right. First question. Coming in from Tiago. What do you think about Next.js stable versions using React experimental features? Oh, that's interesting. I mean, I think it has to happen at the moment because of the release. You know, the 19 hasn't been finalized yet. So I mean, it's a it's basically a practical measure. I mean, if you look at like the Waku framework, they're also doing stuff on the on the bleeding edge. It just means that they're trying to keep us up to date, which I think is fine. But, you know, the AppRiter is, well, they have a six hundred now and now 15. So that's stabilized. But yeah, you know, as with everything, if it's experimental, then be forewarned. Yeah, I guess Next.js the team has a little inside information about stuff that's going on.

But this was like last year. That was a big discussion, right? And they I think last year we were here and then React server components were only available in Next.js. And now that that field has changed a lot, of course. But, yeah, they just have some inside information. There's nothing we can do.

All right. John asks, what's your favorite innovation in recent years? In React, we have to be React server components. I think that they're they're really, really fascinating. In fact, actually, let's see. So Tagus, I know you know Tagus had been doing some really cool work with AI where they're doing generative UI using AI. And that is basically enabled because of RSCs. And the AI can essentially just kind of tell the UI to the API that's responding what to render. And then it can then go and send back that flight stream like we just saw in the video and the last talk back to the client and then download it. Either use the client components or the server components and get that, you know, the nice rendering off of the server still. So I think that's really, really cool. The innovations that we're going to see around generative UI from AI are going to be fantastic in the coming years. Yeah, it's going to be an exciting next couple of years indeed.

Next question. Anything from other frameworks works that you would love inside of React? I'm I think Qwik's hydration model is really cool. I wish more folks actually tried Qwik just to see it. But it is it's really genuinely kind of, well, externally to Google, very innovative. I mean, apparently Qwik is very similar to Wix. They're in Toronto's framework, but I think it's very, very cool. So the idea being that it can. The hydration cycle is going to be a lot faster on the client because every component essentially has its own little bit of hydration and only hydrates what it's required to actually hydrate to get running. It's very, very cool. Yeah, I remember. I think it was last year at the React Summit in Amsterdam. Anyone here been to React Summit in Amsterdam? A few. Yeah, you've been.

You can remember. Yeah, we had to talk from Misko. What is his name? Oh, yeah, Misko. On Qwik. And I like I read it existed, but I remember being like it blew my socks off. Yeah, it's so amazing. If you haven't checked out Qwik yet, be sure to give it a click. I don't think it's out of the realm of possibility. You might see something like that for React, to be honest. And that's one of the things that a lot of folks struggle with is that hydration stuff. So it'd be really cool to see that actually pan out. Well, it's open source. So you can just add it. Have you looked at the react source code? Not recently. Oh, my God. It's incredible. This is like a philosophical question. How great is the react compiler? I think it's actually very, very cool because it's not just doing essentially like a very coarse AST kind of parsing and looking for particular signals. It's actually doing like a full compile of the code and actually running it through like multiple steps and you can kind of opt in and opt out of various optimizations. It's basically an entirely new generation of compiler that's sort of super, it is kind of, I don't want to throw shade on anybody else, but like the solid compiler and that sort of stuff just doesn't have that level of depth. Cool.

All right. Next question. Yeah. Well, I take it back. The inevitable one. When will AI take our jobs? When will it take our jobs? Never. Not if, but when. Okay. So no, I don't think it's ever going to take our jobs.

I mean, you guys have experienced that kind of 80, 20 thing when it comes to AI, when it comes to our jobs, right? And he's like, it gets you kind of 80% there, but you still need to have an actual developer go in and tweak it. And it's the same thing with all of these things.

No code with a bolt.dev. If you haven't had a chance to look at that, I was about maybe bolt.new. It's from the stack blitz folks. You basically say, Oh, I tried it out. I said, give me wordle. And it got really close actually, but it didn't really understand it. It like gave, you know, possible word options that were like four letters, six letters. You know, like, okay. You kind of get the idea, but you don't, don't actually know what the heck you're doing. So it's definitely the 80, 20 thing. You're going to need to go and clean up the code. You're going to need to go and add functionality. It doesn't exist. And you're going to need to add tests. Like that's all the kinds of stuff, again, that AI can help you with.

Um, I'm kind of curious though. So has anybody tried cursor? Yeah. Okay. Yeah. I'm digging on cursor. Cursors is really right. Have you, are you in co-pilot or cursor? Co-pilot. Co-pilot. Yeah. Yeah. Yeah. I was on co-pilot for a while. I definitely find that cursor.

It just, it's got this, uh, tab interf, uh, tab AI and the idea of being it looks at what you've been recently doing and kind of keeps it going. So if you, if you're going and adding like a field, it'll say, Oh, okay. You will, you probably want to add here, here, here, here, here, and then you hit tab and it just happens. It's really, it's almost like magic. It's really, really cool.

Yeah. I work in enterprise country. Have you see, but I've seen demos. Yeah. It's really cool.

All right. Next question. Yeah. When, when this is your, Dan is asking, when will it die? I would, you know, I think. Would 20 years. I mean, that's the thing with people are still coding and cobalt. After all this time and those applications are still running today. You know, what does die actually mean? I mean, something as, as large as react, I don't think we'll ever completely die.

Okay. So you just mentioned you've been developing for 40 years. So you started when you were like two, I think, um, So when I started developing, I focused my efforts on learning flash and the extra script. And I thought I was that Adobe. I'm yeah. We built a flash builder of flex builder. Three flex builder. Nice. Yeah. With the gray icon that was FX. Yeah. Uh, uh, relax.

But anyway, yeah. I thought this is perfect. Uh, I'm learning this. I'm done for life. Yup. You're good. And then Steve jobs wrote a little, he's like thoughts on flash. And I was like, yeah, see, I was like, I got this iPhone thing. You never know something better might come along.

Yeah. We don't know. Yeah. But you know, at the end of the day, like there's so many apps that are written in it and the complexity of those apps is so high, you know, you're going to, there's going to be a lot of, there's a lot of inertia there that you got to, and, and the nurse is not stopping. I mean, none of those graphs, they're all up into the right when it comes to react.

Yeah. That's true. So never is the answer. Yeah, well, I wouldn't. You got easy 20 years and you can't buy a house and put a roof over your head in 20 years. I don't know. Yeah. All right. That's all the time we have. All right. Q and a. So everyone can, I get a really, really soft applause for Jack. A bit harder, a bit harder, a bit harder, a bit harder. Nice. All right.