Pod Insights Syntax - Tasty Web Development Treats 930: Text Editor Keybindings, WASM Replacing Docker, LLM apathy and hosting mini apps About this episode In this potluck episode of Syntax, Wes and Scott answer your questions about VS Code vs Cursor, navigating promotions and job titles, database fundamentals, avoiding decision paralysis, how AI is shaping frameworks, and more!
Show Notes
00:00 Welcome to Syntax!
00:56 Brought to you by Sentry.io
06:24 Moving from VS Code to Cursor without losing your shortcuts
12:13 Should you bring up a senior promotion at a new job?
16:32 Relying on LLMs vs. learning database fundamentals
20:42 Overcoming decision paralysis in programming
25:00 What to do when your code gets too messy
27:39 Could Wasm replace Docker and Kubernetes?
32:14 Organizing mini-apps in Express: monorepo, micro frontends, or something else?
38:49 Will AI lock us into React and make new frameworks irrelevant?
46:57 Sick Picks + Shameless Plugs
Sick Picks
Wes and Scott: Niimbot
Shameless Plugs
Subscribe to 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
Transcript 10,414 words · 2 speakers
Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Scott Tolinski? Welcome to Syntax. Today, we have another potluck for you. We had so many awesome questions that we wanted to
0:06 Wes Bos? Keep them coming.
0:06 Scott Tolinski? Take an opportunity to, yeah, to answer those. And if you have questions for us, for our Potluck episodes, they happen once every month, usually. This time, we get two. And we answer your questions on the show here. So head on over to syntax.fm and ask a question. There's a link right up there in the top nav. We're going to be covering all kinds of stuff today, from if you're moving from VS Code to something like Cursor, how do you handle weird things that might happen? Key bindings, keeping things in sync. Will Wasm replace Docker? Getting promotions? Is LLM puking out React? Are we done? Are we correct? Yeah, is that it? Is this it? Or hosting mini apps inside of a larger app?
0:51 Scott Tolinski? My name is Scott Calinsky. I'm a developer from Denver. And with me, as always, is Wes Boss. This show is presented by Sentry at sentry.io, S-E-N-T-R-Y.io. Man, Sentry is just absolutely cooking lately, and there's almost too much stuff to keep up with, whether that's their AI debugging tools, which simply rule. Finding out exactly why things went wrong is really just the core message from Sentry. And it feels like their tools are always improving this stuff, just nonstop. They even are now doing things that, like logs, like we all need logging. And now, instead of having to bring in another service, you just do your logging directly in Sentry. But it's more than just like...
1:32 Wes Bos? And then connect to your traces? So it's like part of your issues, which is, that's what you want. You don't want to have to piece it together yourself.
1:40 Scott Tolinski? Right. It's not just a thing. It's a thing that's connected. And all this data connecting really allows you to solve stuff in absolutely fast ways. There's a reason why, like, companies like Disney Plus, like Disney Plus has a lot of people clicking Play all the time. They say, "I gotta get Mickey Mouse. I gotta get Minnie Mouse." And companies like Sentry make that possible by really helping the engineers over there make that happen and fixing bugs and solving problems. So check it out, sentry.io/. Syntax, sign up and get two months for free using the coupon code TASTYTREAT, all lowercase, all one word. Let's get into it. All right, so let's get started with this part two of our potluck-a-thon.
2:23 Scott Tolinski? With me as always is Wes Boss. My name is Scott Talinski, of course. Wes, what's going on today, man? How you feeling?
2:29 Wes Bos? Good, good. We're back from the cottage right now. We're still going back up once more. It's in the middle of August right now as of recording, but we just got back from the cottage. It was awesome. It's getting really dry up there right now. Kind of scared about a couple of forest fires out of raging, but I put sprinkler systems in the cottage. For those who don't know, we have a cottage up in northern Ontario, and I installed like almost an acres worth of sprinkler systems, and then we pump water from the lake and just go nuts. So I've been just soaking the grass because it's super dry, and I don't want that catching on fire if anything were to come close to it.
3:09 Scott Tolinski? We've been having crazy wildfires here in Colorado currently, so yeah, definitely a tense situation all around with that stuff. I did see you got all dressed up today. What's the occasion for busting out the tuxedo?
3:21 Wes Bos? Oh, no, the Canadian tuxedo. No, this is actually a new selvage jacket. I've been wanting, I've been wearing selvage denim for like 20 years, and I absolutely love it. Salvage denim is like a Japanese denim that like wears, it's like really blue when you get it, and you get all kinds of cool fades and whatnot. It fits perfectly to you. So for many years, I've been wanting a selvage jacket, and I've never been able to like pull the trigger on one that I've liked. And finally, my wife is like, you need to get one, you know? And I found this one. It's got little W's on the pocket. I researched it for a year. I had it in my cart for so long, and finally, I pulled the trigger on it.
4:03 Wes Bos? And I'm just trying to wear it a little bit more right now to get the fades going.
4:08 Scott Tolinski? Yeah, I love that. I've mentioned this before, but the very first day that I wore my first pair of selvage jeans, which are definitely stiff and tight, at least mine were, I went to the big house, which is one of the largest, if not the largest, football stadiums in the country to watch a football game. And I had to walk up and down like 100 flights of stairs. And I had to like turn to the side and like walk down because I couldn't move my legs up and down the right way. It was absolutely absurd. And then Michigan proceeded to lose to Minnesota of all teams. And I was just like, what in the world is happening? Michigan doesn't lose to Minnesota. No, that doesn't happen. So doubly bad dad to sit in the heat in selvage jeans. I couldn't walk up the stairs. And we lost to Minnesota.
4:53 Wes Bos? Yeah. I've been buying them with like a little bit of elastic in them the last couple of years. Just like having kids and like really stiff pants. I used to wear the like 20 ounces. There's this company called Naked and Famous here in Canada and they make a pair called The Elephant. And it's like, it like cuts your skin. Like I literally had blisters from wearing them previously. And I stopped wearing those because it's just like, it's hard to move around in them. And especially when trying to like kneel down and tie up shoes eight times a day.
5:25 Scott Tolinski? Yeah, it's tough. Cool. Well, let's get into it. Let's talk about potluck questions. It's where you have the questions, we bring the answers. Also, folks, we want to get Dr. Courtney Tolinski, my lovely wife and doctor of psychology, back on the show to talk about mental health stuff. If you want to leave a question for the doctor, you can do so on the potluck form, leave it anonymously. I mean, you could put your name on if you want to, but I would encourage you to just leave it anonymously. And then we can get a bunch of questions for her. So that way, next time she comes on the show, which will be once we get enough questions in here, that she can answer your your mental health
6:10 Scott Tolinski? questions. Again, she's a doctor of psychology, so she can answer things about, you know, learning disabilities. She can talk about all kinds of things related To mental health, learning disabilities, those types of things. So leave your questions in the potluck form and let's get into it with the first question today, which is from DM. Hey guys, big fan of your work. Quick question, have you tried switching from VS Code to Cursor? I'm running into weird issues with shortcuts, wondering how or if you've dealt with shortcut weirdness when switching over. Did you rebind everything manually or tweak some settings? Appreciate any thoughts. I didn't have any weirdness. I don't know about you, Wes, but everything for me just synced over fine.
6:55 Scott Tolinski? With the exception of some extensions that don't exist for, what is it, Open VSIX or whatever that is. Some extensions don't exist and those extensions therefore did not sync over. But other than that, it honestly, I have VS Code Insiders, I have Kuro and I have Cursor. All just sitting there in my dock and some days I open any one of the three. Because they're always constantly being updated with new stuff and I'm trying them all out. And I gotta say, it largely just is a very fluid experience, seamless between the three of them.
7:32 Wes Bos? Yeah, I had a couple of issues. So what I do is like, cause I switch between VS Code, Cursor, Windsor for a while, don't have Kuro yet. I was on vacation while it was open. So I'm waiting for my, my invite to that. What's the one from TikTok again? Why am I blanking on that?
7:49 Scott Tolinski? Oh yeah, we just heard about, I don't know, I have not used that one.
7:53 Wes Bos? Trey, Trey, which is the, the one from ByteDance or TikTok. There's a bazillion of them, right? And I want to be able to like hit the ground running. So what I did, first of all, about a year ago, I've been running the Sublime Text key mapping. For, I don't know, when did I switch to VS Code? 2017? Like six, six, seven years I've been running the Sublime Text keymap plugin. And then I hated that because whenever I told somebody what a keyboard shortcut was, it was not the right shortcut.
8:26 Scott Tolinski? Yeah, right.
8:27 Wes Bos? Like, nobody has that installed, you know? So what I finally did is I scrapped that plugin and just relearned the default VS Code ones. And then I have in my .files repo on GitHub, I have all of my custom key mappings on there. And what I do is I simply just copy and paste or overwrite that file from one to another. And I have to make sure that I keep it in sync because like if I change something in Cursor, I want it to also reflect in VS Code. So I try to keep them all in there. The one thing that I lost in Cursor was hitting Command, like 1 through 6. Will bring you to tabs 1 through 6. I love that. And it's in the browser, you know, if you want to go to the first, second, third tab,
9:11 Wes Bos? I always hit Command 1, 2, 3. So I had to remap all of those. And then there was a couple that were kind of weird. Like, I'm pretty sure it's Command L in Cursor opens up the... There's one of them that it opens up the chat window that I was not used to.
9:32 Scott Tolinski? I still haven't gotten a, I still haven't gotten a shortcut for that I just do the Command Paladin type chat.
9:37 Wes Bos? Oh, yeah. Honestly, that's probably, probably the move as well. But yeah, I just, I just had to remap it and explicitly set them. Even if you want it to be the same as VS Code, just explicitly set them in your key bindings file and then they're, they're going to be the same across all of them. There's still like two or three that are a bit weird, but I just kind of deal with it and remember that's, that's par for the course when I'm switching between them so often.
10:03 Scott Tolinski? Yeah, yeah. It is, for me, I have set up a number of keyboard shortcuts, but I did them all so long ago that, you know, I'm not like, I guess I'm not changing that stuff that much right now.
10:14 Wes Bos? You know what? You know what I've been starting to do is installing extensions on a project by project basis.
10:23 Scott Tolinski? That's such a good idea.
10:24 Wes Bos? Instead of, because like I have like ESLint and prettier just like globally. And then when, and then there's like Dino and Bun extensions and whatnot. So what I'll do is I'll, I'll install it and then disable it. And then only enable it for that project and then it goes into like the VS Code settings, whatever. I have not found a good way to like sync my installed extensions between all of them. If anyone has has something for that, because I would love that. Like I found that like I had to reinstall a bunch in Cursor that I that didn't make it over for whatever reason, even though I don't think Cursor uses the open, like there's basically VS Code has their own marketplace. And then there's like an open VSIX, which Scott said.
11:08 Wes Bos? And that's for anyone who's using like not VS Code, the extensions have to be published on there as well. Like I have a, I have a theme and I have to publish it to both. That's kind of annoying. But Cursor is like reverse engineering the VS Code marketplace, which is apparently against terms of service. But yeah, they they reverse engineered it.
11:30 Scott Tolinski? Man, I'm stuck in a hole trying to get the syntax theme on open VS X, whatever, because it's like, I've gotten everything approved, the organization is approved, whatever. And then you click publish and it's just like, project does not exist. And then you
11:43 Wes Bos? It's a pain in the butt.
11:44 Scott Tolinski? You start the whole thing over again and it's like, it already exists. And like, make up your dang mind. It either exists or it doesn't. And I can see it there. And it's like, I clearly am caught in some sort of database limbo state, but there's no like customer support person I can reach out to. There's a GitHub issue that I filed like a month ago. So I would love that theme in the other editors if that could happen somehow. But it's, yeah, it's, it's that type of project. So.
12:12 Wes Bos? That's a pain in the butt. Daniel says, I got promoted to senior at my previous job just a week before an offer from a new company came through. I did apply for a senior, but I was offered a mid level, but the salary of a mid level was more than the senior in my previous job. So I went for it. Okay, so he got promoted to senior at his first job, got a new job as mid level, which paid more than his original senior one. Now I'm at the job. I've been doing good and enjoying it. My probation is at the end of this month. Do you think I should bring up the senior promotion or is it too early? I don't think my manager was aware that I applied for senior.
12:55 Wes Bos? Interesting. So there's, there's always two camps here. Like first of all, who cares what you're called, but a lot of people are paid according to what their actual name is. And that's why people care so much about this, right? Like, I like Netflix or whatever. They have L1, L2, L3, and, and your, your salary is based on that. Um, so some people like it just because it's, it says I'm a senior, right? Um, some people like it because it means how much they pay.
13:19 Scott Tolinski? And some people also like it because in your job, like resume, you're applying for a job and it says you're a senior, then you're mid-level. And then, you know, somebody who's just glancing at that stuff, looking at a resume might be like, what happened here? You know?
13:35 Wes Bos? Yeah. You got worse. So should, should this person bring it up? I would probably wait until your, your probation is over before, before you do that. But I, I, I don't know, I probably wouldn't do it just yet. Um, it sounds like you're making good money that you're happy with. And if it's just the, the name of it, I would probably wait six months or so and just really prove yourself. Collect a whole bunch of like stories, right? Like, I did X, Y, and Z. Here's what we did. You have to be able to, to prove this stuff when it comes time to asking for a, a pay bump or a bump in, in what you're calling yourself. You gotta sort of come to the table with like, I did X, Y, and Z. We fixed all these things. This thing is faster. The client is much happier. And if you're only there
14:20 Wes Bos? for a month or two, I doubt that you have a lot of that ammo yet.
14:25 Scott Tolinski? I think you take this up in a different way. I think you do bring it up. Cause this person says, should I bring up my, the senior promotion? I think you cut the word promotion out of this. It says, I don't think you need to be looking at this like a promotion. Here's the angle I think you take. I think you take the ignorance angle as in, and I'm talking about you talk to your direct manager in a casual way, not like, hey, I applied for this. Where's my title? Or I am listed as this. I thought I was going to be a senior. Can I get promoted? I think you take the angle of, hey, just checking in. When I applied, I applied for For the senior role and is accepted, but I actually don't see that on my job title currently.
15:11 Scott Tolinski? Do you know what might have happened there? Like, this is casual. This is like, hey, me just checking in. What's going on there? And then, because then they can come back and say, oh, actually, you know, the way that we do things around here is that this band is not listed as senior, but we know that you had that title. Maybe, you know, whatever. We'll, we'll, we'll keep that in mind. Because I think you don't want to put any pressure on this. I think you kind of want to act like, hey, not, it's not a big deal, but like, hey, I applied for senior. I'm just curious what happened there. To me, that's something that makes sense if you do it earlier, because you know that I'm just getting That's a good point. I just got here. I'm getting used to the new org. What's going on here? Yeah.
15:53 Scott Tolinski? To me, that makes sense as a way, and I think you can do it in a way that is, don't put any pressure on it. Don't make it into like a big deal. Come from a place of curiosity. I think we should all be able to talk to our managers like this, if, if I'm being honest. I know some people kind of walk on eggshells about like, whether it is job titles or pay or anything like that. But I do think, you know, having casual conversation about it isn't going to hurt anybody. You can even do it in person if that's a possibility. I don't think it needs to be something that serious. Like, you're just seeking answers. I'm just asking questions here, but you know, just wanted to, want to know what's going on there. But don't be like, hey, where's my promotion? I should be promoted. That's not going to go over well. Uh, next question from Ead, but you can call me Rusty.
16:36 Scott Tolinski? I'm struggling a lot with SQL and database logic in general. I use L L Ms all the time and feel like a fraud most of the time. And for a nearly 1.5 year experienced web developer, I'm pretty weak on database logic. Bro, at 1.5 years, you should be pretty weak on a lot of stuff. Is it normal to use L L Ms to help me in stuff like this? Or should I study more to avoid being behind in backend and database logic? You know what? Uh, this is the stuff that's kind of important. It's important for security. It's important to make sure your apps are running well. It's important to make sure that, uh, you know, data's not being leaked or you don't have any issues because, you know, the backend is that can write and access data.
17:20 Scott Tolinski? And, and those things to me are higher stakes than oops, a button is in the wrong place. So if you're asking me and you're handling a lot of this stuff. And an LLM is spitting it out, you cannot put anything into your app that you don't know what it's doing on the backend. You just straight up can't because like we see hacks all the, all the darn time because unsecured rows or whatever, unsecured tables.
17:49 Wes Bos? And like performance problems too, right? If you don't know what an index is, or you don't know like a, what the like relational keys and how to relate items to another, you can make a real mess. And that stuff sucks to have to, to like change later, you know, like I, I have a database table named customer and a database table called user. And in my head, I know what they are, but I'm so mad that they're named that because it makes no sense.
18:17 Scott Tolinski? I, I've had so many situations like that. And I'm like, yeah, I could do migrations. I could do that. But like, once you get your data in a place or whatever, and it's fairly mature, man, changing it all around sucks. And so my suggestion here is we are not in a post learning environment. I would spend some major time, get your SQL, get your database, get your backend knowledge, like locked down from that perspective into like intermediate territory where you know the best practices, you know, the entry level stuff, you know, the table stakes stuff, you know, all of the, the most important things about it before
18:57 Scott Tolinski? you just let the LLM write the code for you in that regard. It's not worth it. You're going to get yourself into trouble.
19:05 Wes Bos? So many of these questions that we're getting are like, do I, do I learn this anymore? You know, like, what's the point? And like, man.
19:13 Scott Tolinski? Bro, you are responsible for the code.
19:16 Wes Bos? Yeah, you're still responsible. Check out Aaron Francis. He's got a bunch of courses on different database stuff. And even if you just like watch and like learn about databases and how all of this stuff works, it's going to be helpful just to know what to type into the box. When you say, you know what? Use this as a primary key or add an index for the email field or for whatever reason, or you'd be able to look at your, your performance. And because before you know it, you're going to be having some. Some loop or like you're gonna have some crazy database bill because of some garbage that was written. So.
19:54 Scott Tolinski? But you're gonna expose a bunch of sensitive data and you gotta be responsible for that. Like you said, learning is fun, bro. Like the AI stuff is great that it can make you so much more productive, but at the end of the day, it should be an extension of what you're doing. It shouldn't be like the expert because it is oftentimes not doing things in the best way. I mean, at least like I know it can be very effective, but especially it would be like you have a lion and no lion tamer, right? That lion is gonna do whatever it wants. Yeah, been there. The lion's eaten people, you know? So you, you have to be a lion tamer to this thing still. And yeah, you gotta know what it's doing.
20:39 Wes Bos? You have to have some idea about what's going on. Question from Corey. I've always struggled with making decisions and unfortunately, after being in this field for a while, I realized that programming is basically a stream of constant choices and decisions. No kidding. It's like something new comes out every single day. There's so, everything is a trade off. Everything it's, it depends and it all changes.
21:02 Scott Tolinski? Do I do that?
21:03 Wes Bos? Every, every three days. So with the JavaScript ecosystem moving so quickly, every decision is followed with a dreadful self monologue. Is this the right way to go? Or am I shooting my future self in the foot? Do you ever experience decision paralysis? If so, do you have any tips over not stressing over FOMO and falling into the pit of decision paralysis? Yeah. My answer here is this stuff doesn't move as quickly in production as it does on social media and on this podcast. It's fun to, to talk about all this type of stuff, but there's a reason why 10 years later, everyone's still mostly just using
21:43 Wes Bos? React for the, these types of things. And just know that picking something that is a little bit more established and safe over something that is new is often the better choice because these things. They they stop getting maintained, they get changed, they change their name. And like if I if I went for absolutely every single brand new thing, you realize that they're especially if there's a business behind it, they're going to pivot into an AI site builder at some point. And you're gonna be you're gonna be out of luck. So just know that. No one ever got fired for choosing IBM is that's the saying, right? And and no one ever got fired for choosing MySQL or or Postgres.
22:28 Wes Bos? And no one ever got fired for like, there's so many people still using it express. And and they're totally fine with it. And it's still a very popular choice, even though I I think we're at a point where you should probably stop using express.
22:40 Scott Tolinski? Yeah.
22:40 Wes Bos? There's a reason why a lot of people are still choosing it.
22:44 Scott Tolinski? A lot of the really like like fancy stuff that people will like go on Twitter and say, this changes everything. Ultimately, can get you into trouble because the communities aren't mature. You end up hitting a edge case that hasn't been hit yet because there's not enough people using it. Or it's in beta and you hit a bug, then you gotta sit and wait. Like as somebody who, you know, really is early on a lot of stuff because that's my job. Man, it's frustrating world. And like, there's so many times when I'm like, you know what? I want to get this done. I'm picking PocketBase, I'm picking Svelte, and I'm going to crank through something because it's easy. And I know it's going to work and I'm very familiar with it. Not that PocketBase is bad. PocketBase is great. But like, I just am, you know, it's not the latest and greatest new thing, but it it it works and it's great and it's easy.
23:30 Scott Tolinski? So like, don't stress yourself out about this stuff. You know, get good at a set of tools, be effective with them. And then when you feel like they are showing their age or cracking for whatever reason, replace them here or there and keep your ears open for what is new and maybe give it a try in your spare time. But don't let that kind of stuff slow you down. As far as decision paralysis goes, again, I think that's going to come from establishing really like, what are these things I'm very effective in? And like locking them in like this is the stack. When I used to work at an agency, we were cranking out sites like every month, like we had.
24:08 Scott Tolinski? Specific stacks that we worked in for a reason because we were very good at that and we didn't have to decide, all right, what are we using for this? Uh, no, we're gonna you, we're gonna use Drupal, we're gonna use the Omega theme. I think it uses stylus, so we're gonna use stylus or SAS. Uh, we're going to use this and that and these following Drupal plugins and we're gonna knock this thing out. And there was nothing like that productivity that you could get from not having to think about all of those little pieces. And you could just get to work. I think that's the one thing, and this is not related necessarily, that AI has helped me with.
24:47 Scott Tolinski? Is that like, it's easier for me just to get to work because I'm not always thinking of the little tiny pieces that I do a hundred billion times. Every single project. I'm just getting to work on the stuff that's important to me.
24:59 Wes Bos? There was a second question here that he asked is what do you do when you're writing code and have that sudden realization that the code you have written is starting to get too convoluted? Been there, brother. That's the right time to peel it back and to refactor it. Don't just push it, push it through because it's very unlikely that you're going to come back and fix that and change it and you're gonna build stuff on top of it. So if things are getting too convoluted or if you're doing things in a weird way, um, and even talking to people about your approach, like yesterday I was working on some Cloudflare durable object stuff and I had a, I had a problem and I brought it into the Discord.
25:40 Wes Bos? And then some guy was just like, probably not the right way to do that in the first place. And I was like, oh, that, that, I'm glad that you said that, you know, like the, the solution to my problem was not like I, I had to fix some bug. The solution was you're just not, you're tackling this in the wrong way. You shouldn't be doing this. So stop then, peel it back. It's frustrating to go back a little bit, but your future self will thank you.
26:05 Scott Tolinski? The AI I want is the one that says, Hey dude, just to let you know, this is not it. What you're doing right now, you think this is it?
26:14 Wes Bos? Yeah.
26:15 Scott Tolinski? You're making a big, a big wrong choice that you're going to regret in like another couple hours. And then if you didn't commit before this, well, you're gonna, you're gonna really regret it. So, commit frequently.
26:26 Wes Bos? Yeah.
26:27 Scott Tolinski? Being able to jump back and like, yeah, sometimes a branch or sometimes a pathway in your code is a learning opportunity.
26:37 Wes Bos? Yeah.
26:37 Scott Tolinski? You could maybe cherry pick some code and sometimes you just say, uh, let me just take some wooden boards and board up that branch and say, "I'm not going back in that room. Uh, that was the wrong idea and just moving on from that." So, no, this absolutely happens, yeah.
26:52 Wes Bos? I need the little clip in the corner of my text editor that says, "It looks like you're trying to, "to write some code that will address all situations. Would you like to just create a new file and duplicate a little bit of that code?" You know, like, sometimes I try to like, make this amazing code that can, can be used in any situation where it's just like, "Ah, this would be better if I just duplicated these 20 lines and have a different implementation."
27:17 Scott Tolinski? Mine needs to be, "Hey brother, I see that you are sidestepping the purposefully built way to do something because you think you have a novel use case." Let me tell you, your use case, not novel, and you're making things harder on yourself. Uh, that's what I do. I'm always like, "Of course, no one's doing this. I need to do this on my own."
27:37 Wes Bos? Yeah.
27:37 Scott Tolinski? Yeah.
27:38 Wes Bos? Oh, that's great.
27:39 Scott Tolinski? Next question from Willie Mike. "I've been seeing a lot more hots. What do you think hots is here?"
27:45 Wes Bos? Hots around using, like, hots, like, people are like stoked about something.
27:50 Scott Tolinski? Okay, I've never heard that.
27:51 Wes Bos? The hots?
27:52 Scott Tolinski? I've been seeing a lot more of hots around using Wasm, not for websites, but for more of Docker lambda replacement. Do you see serverless standalone Wasm computing as potentially an answer to easier, more simple Kubernetes or Docker?" Willie?
28:10 Scott Tolinski? I got no clue. Wes, what do you think about this? I don't use this stuff that much. I'm gonna be honest. I hate Docker.
28:16 Wes Bos? So WebAssembly, the idea with Wasm is that you can take a some compiled code, like that's written in any language, right? Could be written in C, could be Rust, it's just compiled. And you you convert it to what's called a Wasm file, WebAssembly, right? And then you can run that code in something that run like in the browser was was the first implementation of that, right? So if you want, if you want to run FFmpeg or you want to run some Python script or some, you can even run PHP in the browser. We often do that because there are purpose-built tools that are are not built in JavaScript. Doesn't make sense to to write them in JavaScript or they're just not.
29:02 Wes Bos? They're in other languages, but you still need to run them in a JavaScript environment. That's that's where we use Wasm, right? Now you can run Wasm in in Node. And it's nice because you just take this tidy little bundle, very much like a like a Docker container. And you can, you can run it in in anywhere that you want. We use it on the Syntax website to stitch together MP3s. We have FFmpeg Wasm running. So the question was, do you see serverless standalone Wasm computing as potentially an answer to easier Kubernetes Docker? Absolutely. Not in every single use case. You can't throw an entire Linux environment inside of a inside of a Wasm container.
29:45 Wes Bos? But you can put a lot of purpose-built tools that you simply just call from from JavaScript and it works really well. I think even the creator of Docker said that like Docker might not exist if if Wasm was around. I don't know that it's going to totally replace it, but there are a lot of situations where I will reach for Wasm rather than having to try spin up a container in inside of Docker. That's cool. And also, like what I've been seeing a lot lately is now that we have WebGPU hitting all the browsers, like one of the downsides to Wasm is often significantly slower than running it native, right? Like the FFmpeg.
30:27 Wes Bos? Implementation is is much slower and also it's hard to do multi-threaded unless you get into to some some complex stuff. So now that we have WebGPU, you're seeing a lot of tools that are now built that will run on on Metal, especially like AI LLM models and whatnot running in the browser. So where sometimes you used to reach for Wasm, now there's there's often stuff that has been built and can run on the right on the GPU of your computer and that might that might even be faster.
30:59 Scott Tolinski? It's all very interesting to me, just not stuff I deal with then, right? Yeah. Or stuff I hope to not deal in, I'll say that.
31:08 Wes Bos? Yeah, you know what's one interesting use of Wasm is Shiki. So Shiki is a code highlighter. Shiki uses WebAssembly because its syntax highlighting engine is built on top of Onigumura, a RegEx library written in C and it's used by VS Code's TextMate Grammar System. So like, in order to split up, like you have some HTML and in order to split that up into like attributes and tag name and angle brackets and all that stuff, what VS Code uses is a huge RegEx library. And in order for that to be fast as heck, because it has to like every time you hit enter or type a character, it has to redo it, that's written in C. So if you want to run that in a JavaScript environment, you have to use Wasm.
31:53 Scott Tolinski? Interesting. I had no idea.
31:55 Wes Bos? Probably goofed up the the thing here. That's it sounds Japanese. How would you pronounce that? On? Onigumura? Let me send you the link.
32:07 Scott Tolinski? Uh, yes, that is Oniguruma.
32:11 Wes Bos? There we go. I don't have that.
32:14 Scott Tolinski? Cool. All right. Next one from Ben Damman Ban. Hey, Scott, Wes and CJ. CJ is not here, Ben, okay? We're tired of everybody being like CJ.
32:26 Wes Bos? CJ, all the comments. CJ, CJ is so good. Yeah, they were real quiet when I beat CJ in the last, the last coding video. So I don't. The CJ pods got turned off, which is great.
32:39 Scott Tolinski? On a major slump on those. I gotta get going, folks. If you haven't watched our CSS battles, you should tune those on there on YouTube. No, I'm just joking. We love CJ, folks. He's the best. Hey, Scott, Wes and CJ. I'm currently building a side project that's a collection of mini app ideas to make my personal life easier. Its current form is an Express app with template and nunchucks, but I'd like to practice my Angular skills more since it's the main stack for the front end at my new internship. I would like to have a structure where each Angular app can be a standalone and only accessible via route in the Express app, and each page already built can stay in their
33:19 Scott Tolinski? respective routes. This idea also keeps auth state in each page of the Express app, including the Angular apps. How would I best organize what looks like to be a micro front ends approach? Mono repos, same repos, separate apps with a reverse proxy? Thanks. This is interesting. You could do this a number of ways. And these mini apps, to me, I think turning this into, well, it depends on what you need here. In a big sense, if you need individual versions for packages in these mini apps, as in these mini apps
33:58 Scott Tolinski? have their own packages and these packages are all versioned very specifically, then a mono repo makes sense, or even like when you have like a linked repo, what do they call that?
34:16 Wes Bos? Sub repo, that's it. Yes. So where you have like one folder of your Git repo is a totally another. You can get that if you just get in it inside of a folder inside of an existing Git repo.
34:28 Scott Tolinski? Some people hate that, but I actually, I don't mind that. I've used it before successfully. That said, if these apps don't have that strict version requirements and they're just like hacking together on some stuff and you're all using the same versions of everything in all these apps. You just put it as folders in one repo. That's gonna be the mo- like, the last thing you wanna do when you have like, a for fun mini app thing where you can hack on stuff and try something new and whatever. In my mind, if you bog that down with a bunch of tooling around it, man, you could spend so much time on that tooling.
35:05 Wes Bos? Yeah.
35:06 Scott Tolinski? Sure, you could set it and forget it, but there are times when something, "Ooh, this might be nice if I have this and then this and then this." And next thing you know, you're installing cross packages and it's all just a- It inhibits you from working on your mini apps and suddenly the mini app, you know, structure and everything is now what you- your project is.
35:25 Wes Bos? Yeah.
35:25 Scott Tolinski? So if they are small apps that, you know, don't need self-contained versions and stuff like that, just throw them all in different folders and call it a day if you wanna- if you wanna have that. To me, that barrier to entry, you can't beat a small barrier to entry in this kind of stuff.
35:41 Wes Bos? This is what I do with my Hot Tips repo. So I have a huge repo full of, I don't know, like, probably 80 different apps that I've built over the, you know, just little demos to- to actual big things. What I have is I have V in the root and then I have a plugin called Vder, V List Directory Contents, and that's a plugin that I built to allow you to click through the folders and see, and then you click on one of them and it will- will actually build it. And that's why- that way you can have- you can run TypeScript and all of this stuff without having to do any build step. But if I do need a larger project inside of there, I can always put a package JSON in that subfolder and then that allows it to be its own thing.
36:24 Wes Bos? And that's- I love that because like, by default, half of them don't need their own package JSON, right? And I can share code between the two if I want to. But if there is a situation where I do need its own package JSON, then I can go ahead and run that. That might be a good way. Putting that in production is a little bit trickier because you have to build them all or I also have been tinkering with building them on demand as- as you click on one of them. I don't know if that's maybe not the best idea. I was also thinking- Maybe Coolify is a better approach here because maybe just you can keep them all in the same monorepo if you want that way you can share content between the two but simply just throwing it up on Coolify and like back in the day I used to have a PHP server where I had just a folders
37:09 Wes Bos? full of stuff that I built and you just drag a folder and put it on there that's that's the kind of the dream with this type of stuff but you can also have the the whole build system so maybe a Coolify approach is is a bit better and then your express app that you're talking about like authentication to like so that people can't access all of these different applications that might work better as just like a global middleware maybe a Cloudflare worker that sits in front of the entire domain name and checks the incoming request to see if you have you can access it or you can even use something like Cloudflare Access where you you have to log in with your your Gmail account or whatever before you it allows you to go forward and access each of
37:54 Wes Bos? those those values.
37:55 Scott Tolinski? I can't help but feel like this all seems over engineered for what it is though, you know?
38:00 Wes Bos? And then it if it breaks then you don't feel like changing those those different apps. That's a good question. I don't think I would build an express app that then then routes everything for you because then you're getting into like your Angular stuff is going to need like a build system, right?
38:19 Scott Tolinski? Yeah.
38:21 Wes Bos? Then you're getting into Terrible Repo so you don't build the things that haven't changed.
38:25 Scott Tolinski? That's what I'm talking about, the tooling.
38:27 Wes Bos? That's a bit of a pain.
38:28 Scott Tolinski? Project just becomes tooling, yeah.
38:30 Wes Bos? That's why like yeah, that's why my whole like VE thing that I built it doesn't do backend though it's just simply client side stuff but that's beautiful because I can simply just add a folder I can add a packet JSON if I want and then I just have to surf to the URL and the whole thing will build on demand. Next question from AI am irrelevant. As more developers use AI to write code and AI models improve, the framework and implementation details become less relevant. New frameworks are built to solve problems that developers Run into, but these problems become abstracted away by AI. More code on the internet will be written by AI. It's heavily tilted towards React, which will
39:13 Wes Bos? lead to there being mostly old React code that current models are trained on and implemented into the new code written by AI. New improved coding paradigms and code for frameworks better than React will decrease. And new AI models will be trained on older React code and older models. The question is like, is this just churning out a bunch of React code? And like, is that it? Like, what is the point even to creating new new frameworks when like, the reason why we switch frameworks is because like developer experience and performance and whatnot. And if you can just tell the AI model to do something, is there anything there? So I've gotten this question a lot where people just seem a little bit
39:58 Wes Bos? like apathetic about things getting better. You know, like we, we've arrived. This is, this is all that's going to be. And we're just gonna just spit out a whole bunch of React code and whatnot. So I've said this many times on this podcast. As soon as something becomes table stakes, the bar is going to be raised. Meaning that now that these LLMs can churn out a crud app with a React interface that saves to a database, we're not done. That's, that's not it. Anybody can make a crud app that saves their database and logs in and whatnot. So that's table stakes. Now the bar is raised and, and everything now needs to get better.
40:42 Wes Bos? Right? I'll, I'll give you one example, not from, this is just how technology works. One example, not from LLMs is like search back in the day and search in WordPress was simply just a like SQL query, right? You want to search your entire website? It looks through your titles and it looks through the body of your content with a SQL query where it looks for that word somewhere inside of that. Right? And that was search for, for a very, very long time. But now that that was table stakes, it's so easy to do. Now we want more. We want embeddings. We want to search inside of photos. We want algorithms that are based on our taste.
41:25 Wes Bos? You know, I want to search and I want it to show me things that I want, not just the same search result for everyone. We want similarity to things. We want to search, want to be able to upload a photo and search it. Like I get mad now.
41:36 Scott Tolinski? Yes.
41:36 Wes Bos? When I, I can't, like, I wanted to find a photo on Flickr the other day. And I was like, how do you search by photo on Flickr? They don't have that, you know? What an old website.
41:46 Scott Tolinski? Yeah.
41:47 Wes Bos? So.
41:48 Scott Tolinski? It's so accurate to how I browse search now. Yeah.
41:52 Wes Bos? A single one-line SQL query is now this massive thing that is like, it's backend, it's frontend code, it's databases, indexing, it's embedding, it's comparing, it's, it's custom algorithms. Technology gets better. Everything gets better and users are going to expect more. So that, that was my first thing. I, I wrote a whole bunch of stuff here because I thought that this was, was a really interesting one. The second thing here is that the apps that stand out are the ones where developers are rethinking how to do things. And using cutting-edge APIs, they're creating joyful experiences, right? We can all churn out a ShadCN crud app that looks like versale.com now. That's, that's not a problem for anybody to do.
42:37 Wes Bos? And, and to a point where people are, are getting tired of, of that experience. So the good websites now are, are people that are, oh, this is fast. This is local first. This is, oh, you're using a whole bunch of new CSS, you're using CSS anchor API. Oh, wow. You're using all these new browser APIs. That's the stuff that's, that's going to stand out and that stuff is hard because that's new. Uh, ask an LLM to, to write you a whole bunch of CSS that was released two months ago. It's not going to do a great job. It'll probably get better as you can give it docs and specifications and things like that. But the people who are still learning, this is the second time we've said this on this podcast, you should still learn.
43:19 Wes Bos? You know, you're not, you still have to progress past that. So that's good. And then my third thing here is that some cases, we just don't need to spend time on these things. Um, at, at some point, the LLMs are going to get very good at some things. And we simply would just not need to worry about. Spending time on those things. There are lots of other problems in this world that we can solve with technology and I don't think that you're going to be bored.
43:44 Scott Tolinski? I've been trying to come up with just how I feel about some of this stuff in regards to the same thing where like the AI has enabled people who don't know what they're doing to push out stuff that is not going to be good experiences. And so like, for instance, I just saw on Twitter somebody being like, "I really love this blur effect. How can I do this in React?" And then somebody responds, "Here's a React package to solve, to give you a blur effect." It's like, brother, it is a one line of CSS. You were installing a React package for one line Yeah. of CSS. And so the bar needs to be like, those are the people you're competing with.
44:30 Scott Tolinski? Because that is not indicative, like a React package for one line of CSS, even though that's not what it's doing, it's probably doing some JavaScript in there. Either way, like, that's not going to in itself make their app bad. But it's indicative of a bad app. Because if that is one thing in an app, then there's going to be a lot more. So like, yes, we need to focus on better experiences, faster apps, faster apps, absolutely. I mean, there's a reason why people say the web apps are not as good as the native apps. And some of that is APIs available in Safari and some of it is just straight up bad programming, low FPS,
45:11 Scott Tolinski? data that's taking too long to get into the device, React functions that are running anytime something way up here changes for no reason whatsoever, because people don't understand how it works. So we need to definitely focus on better experiences overall, whether that is joyful, whether that is UI, whether that is animations and tasteful and very specific ways to give context, but just speed, reliability, those things, that is what a good programmer is going to produce.
45:47 Wes Bos? Yes, I agree. And I also think that people are working on frameworks right now that are going to be tailored so that LLMs are good at writing them. The LLMs are very good at writing JSX, they're very good at writing tailwind code. And if people look at
46:06 Wes Bos? At how these, like what these LLMs are good at, they take a step back and say, okay, how do we now write a JavaScript framework that an LLM will be very good at generating code for, at not making a mess out of it, you know, like it's very, all of the things that the LLM are good at. What if you made a framework that was targeted towards that? Obviously, good for the developer as well to be able to go through it, but what if the LLM didn't make a mess and whatnot? I think that we'll probably start seeing that and there, people are trying to figure out, like, part of that is, like, docs are super important right now, and part of that is going to be supplying the LLM with what it
46:48 Wes Bos? needs to be able to figure out how to change things. So I'm, I'm kinda excited about what that will look like in the next little bit.
46:55 Scott Tolinski? Yeah, totally, for sure. Next is the part of the show where we talk about sick picks and shameless plugs. These are things that we've been liking, enjoying using any of this stuff. I'm gonna pick something that I have sick picked at least twice now, but we've been doing this show since 2017, so I'm allowed to do that. And I bust this thing out every once in a while, and when I do, I'm always blown away just by how freaking dope it is. The Nimbot is this little thermal printer that prints onto labels and it connects via Bluetooth, and man, it is this little thing. It's little, the labels come on a
47:36 Scott Tolinski? thing, they're super cheap, you go on your phone, you connect via Bluetooth, you type in what you want, you add images or whatever, you click go, and it prints you off a dang little label. And, yeah. Did you put a label on your Nimbot? You named it? Nice.
47:53 Wes Bos? Yeah, well, we, we bought a whole bunch of, like, you can go on AliExpress and buy, like, Different kinds of labels because it's thermal, right? And you can label, like you can buy kitty cat ones or clear ones. It's, it's pretty good.
48:07 Scott Tolinski? It's thermal and, and so therefore you're not buying ink for this thing, right? It doesn't run out. The only thing you would have to buy is labels and that is super cheap. And like, man, I, it recharges, the app is great. I just can't believe I've had this thing now for a while and it just still rules. Anytime I need to label, like I was labeling some stuff in my kid's room and it was just like, man, I love this thing. I, I like this so much. It rules. It, it's been so reliable for me. So the Nimbot is going to be my pick. I have a funny labeling story, Wes. Would you like to hear it?
48:41 Wes Bos? Yes.
48:42 Scott Tolinski? Okay. I worked at Target when I was 16 to when I went to college, pretty much. Target was a well-paying job in my city. It was like one of the better paying jobs. We'd just gotten the Target, brand new Target. And I got the job. I was on the sales floor and I was really good at my job. You know, I don't know why the ADHD and the spatial stuff, I was just really good at the job. So I was very well-liked by all of my bosses. But I had a number of, I had a number of like chaotic things I was doing behind the scenes for my own enjoyment. And one of which was naming the walkie-talkies. Everybody had a walkie-talkie and they had a label maker in the back room.
49:23 Scott Tolinski? So if you forgot your name tag, you could print out a new name tag for the day. And I just started coming up with like every day of the week, I would come up with a name like, you know, like Steve and I'd print it out and then I'd put it on the walkie-talkie. Because the walkie-talkies all just had numbers. And the next thing you know, all the walkie-talkies got names. They're all named different stuff and, and the staff members were always like, like the people were joking around about it. People loved the names. And then they had to like send out a, a mailer that was like, please, whoever is naming the walkie-talkies, we need you to stop doing that. Like that, like that is, we need you to stop. But the best part is, is they couldn't take away the label maker because we needed it. At the same time, there was no camera.
50:07 Scott Tolinski? Or anything where I was doing it. So I just didn't stop and they were getting so angry. I did this for years and I never got caught and I would wait a couple of weeks and they'd, you know, die down. They'd take them all off and then I'd put another one on and man, you could tell that was pissing off. Uh, some of the, the higher ups there. No one had ever suspected it was the employee they all liked, uh, doing something so. It was so non hurtful and at the same time I was just like, anytime we got a message about it, I was just in tears by myself. Uh, just being like, oh, this is so good. They're never gonna catch me. They're never gonna catch me.
50:42 Wes Bos? I love my, my nimbot as well. The only weird thing is that it doesn't, you can't like print it as long as you want. Like we have like a brother one with like the keypad on it and I hate that. Cause like, I hate typing. I want to type on my phone or like you can, you can also.
50:57 Scott Tolinski? Emoji.
50:58 Wes Bos? Like down. Yeah. You get emojis or you can, often what we'll do is we'll like my kid, we, we sort our kids like toys into boxes. And we have like a Minecraft box. And then what we'll do is I'll just go on Google, find the Minecraft logo, bring it into the app. And then it'll like, it'll figure out how to print it black and white. And then you can put the actual logo on the, uh, the label. It's, it's just, it's such a good thing. I.
51:21 Scott Tolinski? Also a multi-line text.
51:24 Wes Bos? Yeah. Yeah. You, it's the whole text editors is amazing. And you, there's also people have hacked them. You can use canvas, uh, to print to them now with the web Bluetooth API.
51:33 Scott Tolinski? I did know that.
51:34 Wes Bos? It's a wild world. I kinda wanna get the bigger one. Cause like this, I got the, though they have like a wide one, right? That's like maybe two and a half inches wide.
51:42 Scott Tolinski? Interesting.
51:43 Wes Bos? And they're, they're so cheap. Like this, this one is 24 Canadian, probably, probably like 15 or 20. Oh, they got tariffs now, Americans.
51:51 Scott Tolinski? It's 1999, 1999.
51:54 Wes Bos? Oh man, what a world. How much is that in Canadian? Man, stuff is cheaper in Canada than it is in the States. We've arrived.
52:04 Scott Tolinski? I'm gonna have to bust out my, bust out my loonies.
52:07 Wes Bos? Man, yeah, you gotta come. We, for my whole life, we've been driving over the border to buy beer and clothes and Target. And we would ship stuff. We have a mailbox in the States. We would ship stuff to. The tables have turned, but they have a bigger one. I kinda wanna get it. Thirty-seven bucks. Cause I, I put labels on everything. I just cannot-
52:26 Scott Tolinski? Dude, I love it.
52:26 Wes Bos? Live my life without having labels.
52:29 Scott Tolinski? My son was just having such a hard time ignore, like, cleaning his room and like putting things wherever they went. And I was just like, listen, if we name everything, you could see exactly what goes where.
52:39 Wes Bos? I don't know where things go unless there's a label. I have that problem. When we do the dishwasher and you know, like when we use, there, we have some plates where we use all of them. And then when I need to go, I don't know where they go. I need that like, um, layout. You ever go to the store and they have like, like, somebody's merchandising and they, they have like a photo of where everything goes. I need that.
53:00 Scott Tolinski? See, I'm, that's my domain. Where things go in the kitchen, that is my domain. So I, yeah, that's my, you know.
53:08 Wes Bos? Awesome. So I'll, I'll sickpick that as well. I'm a big fan of the Nimbot label printer. It's, it's been fantastic. The only thing I'll tell you is it's not good for kids, like kids' school stuff. Uh, we still-
53:20 Scott Tolinski? They, they come-
53:21 Wes Bos? Yeah. It like rubs off for whatever reason, but like the, the labels that we, we buy with their names on it, they must be like, printed on or something like that because they go through the dishwasher, no problem. But this stuff is, they fade.
53:35 Scott Tolinski? You've had a thermal receipt before. They fade, whatever. Yeah. But for the most part, just like labeling anything, they're not gonna, without a lot of-
53:42 Wes Bos? No, no. Just, just your regular stuff. I've had it like all my, hold on, lemme, lemme see here. Like, I've had-
53:49 Scott Tolinski? Yes.
53:49 Wes Bos? I've had this for like probably three years.
53:54 Scott Tolinski? Those are cute.
53:55 Wes Bos? And they're,
53:58 Wes Bos? like micro USB, USB extension and USB B. So I'm, I'm showing a, uh, like a shoebox, which is another sickpick of mine to organize all your cables. And like I've go nuts with these.
54:09 Scott Tolinski? I didn't have any, I don't have any cute labels.
54:11 Wes Bos? That's it. All right. Let's, uh, shameless plug. What do we wanna plug here?
54:17 Scott Tolinski? Sickpick?
54:18 Wes Bos? I'm, I sickpicked my Nimbot as well. I think I, I like that one. I love talking about how good that thing is.
54:23 Scott Tolinski? I know. I can't talk about it enough. Uh, shameless plug. We would love it, folks, if you were to, well, maybe just, just go to YouTube and click subscribe, please. Uh, click, you know, the bell and all that. People still click the bell. I haven't heard that being said as a thing in a while, but either way. Click subscribe, please. We do so much good stuff on YouTube beyond the podcast. So you want some more technical, long, deep dives on all kinds of topics. You want fun and games and just lots of laughs. We've got LOLs over there by the dozen. Check us out on YouTube, forward slash at Syntax FM, or just search Syntax on YouTube. We should be numero uno there. Again, we're doing all kinds of cool stuff, so check us out. We would love it. We would absolutely love it if you had.
55:07 Wes Bos? I always search for Syntax, but I'm curious. Yeah, we are. We are number one when you search it up. There's another. There's another channel, the perfect channel for survival game dinosaur enthusiasts called Syntax, with a C.
55:21 Scott Tolinski? Nope.
55:21 Wes Bos? Don't beat us out. Nope. Cool. All right, thanks for tuning in. We'll catch you later.
55:26 Scott Tolinski? Peace.
Previous 929: Cloudflare Blocks AI Crawlers × Debugging Local Data × Raising Kids with Healthy Digital Habits and More Next 931: Project Init - How to Make Good Choices When Starting a New Coding Project Page updated November 1, 2026 · transcript and summary are machine-generated