The Rise of the Design Engineer

Welcome to Syndax. Today, we got a potluck episode. We're talking to you about the come up of the design engineers. This is a new thing you're hearing a lot about, and I quite honestly, I think this is going to be very important, role in companies going forward.

Talking about how to introduce your kids to AI or if you even should, we've got special Wes coming on, so stay tuned for that.

Update on Scott using Omarji in Linux. Is he still doing it, or is he is he switched back to macOS? Why AI has no taste? And we talk a little bit more about design in this episode, and it's not just about, like, how it looks. Design is not just like, oh, I chose red because a lot of these apps look exactly the same, but they are not the same in how they function. And, finally, we're talking about keeping coding agents efficient. What how should you tackle building your projects so that they are efficient and not just spewing out a whole bunch of extra files? I got some really good tips from Scott on that one. If you've got a question you'd like us to answer, please go to syntax.fm.

Click on potluck in the, in the navigation, and send us that. We will answer it on a future episode. What's up, Scott? You ready to get going?

Hell, yeah. I'm ready to get going. Alright.

First question from birthday girl says, I'm a designer and self taught developer with a background in design. I see myself as a hybrid, equally good at both, and don't want to define myself primarily as one or the other.

Recently, the role of design engineer has been on the rise, and they link to a Vercel blog post that essentially says what a design engineer is. A design engineer at Vercel blend aesthetic sensibility with technical skill. This allows us to deeply understand a problem, the design, and build and ship a solution autonomously.

Right? Somebody who has the eye for design and and the the UX of design JS well as actually being able to to build the entire thing. So this this term of design engineer is on the rise, which I think captures the dual expertise well. I have two questions. First, what are your thoughts on the future of design, dev separations, or convergence? And two, I feel like most engineers discount design and dev hybrid expertise if they come from a design background, but the vice versa is not true. A really talented engineer who got really good at designs seems to have credibility issues.

Are people biased at design because it is not as a technical background? I find this a bit silly given the developer community is very welcoming of self taught developers, or do I experience this more because I am a woman? Curious to hear your thoughts on this, super design savvy devs. Thanks for the pod. Okay. This is awesome, and I love this idea of a design engineer because, like, many years ago, we had this question of, like, should designers Node? And everybody would would fight about it. They said, no. No. No. I'm gonna sit in Photoshop, and I'm gonna design my things, and and then I'm gonna hand it off to somebody else to actually do the implementation of it. And then equally, you see you heard, like, devs being super frustrated where, like, I designed this thing, and then I handed it off, and they didn't do a very good job. And then as the web got more complex, design became much more about how does it work, how do you interact with it, how do these things, work together rather than simply just like, what does it look like. Right? Like, I think that's such a very small part of design.

So I love this idea of a design engineer, which is somebody who very well understands how products work, how user interfaces work, how they what happens when you click it, how what happens when you have multiple items together.

This stuff is so we have a question coming up, next that is similar to this. I won't say too much, but this stuff is so important, especially as everybody is just vibe coding their UIs right Node, and there's pooping out some dog crap. Right? So I think that this design engineer is fantastic, and and I think it's going to be even more important in the in the next coming, maybe probably six months. You know? Everybody and their mother can prompt out a half decent looking UI, but where excellent UIs really come to the surface is Wes somebody who spends time understanding how it all works together and how it interacts and what happens when you hover over top of it and the little frustrations that come with using an application. And we Scott more to say that in the next question, but I think that this skill is going to be the new full stack developer. Right? Because that was the other thing. Right? I'll let you talk in a second, Scott. But, like, that was the huge thing that everybody fought over as Wes. It's like, should you be a full stack engineer, or should you be back end and front end? That's no longer a question. In most cases, people are are doing both in the web world.

Yeah. Wow. You know what? I I and and I think, one, happy birthday, birthday girl. Two, if it is still your birthday, which I would assume it is considering that's your Node, I I would say on the regard of are people biased against design because it's not seen as technical, I think that this has always been one of bias that I've noticed, and I would like to pose that it's out of jealousy. Developers look at somebody who's good at design and think they possibly could not be as good of an engineer as me because, well, I'm not good at design, and I would, was you know, that's a a whole discipline you've had to learn. Yeah. And I've spent that time in the engineering side of things. So I I've personally seen this to be the case.

Some people just have much better design intuition too.

I've seen engineers that are are fantastic designers just because they have that really good design intuition.

I would also say it is probably also because you're a woman. I I that is also a bias that we see in this industry. So you you have, multiple factors here, and I think that's working against you. But I do think that this role in particular is such a huge role in the future now, of the world that we exist in with AI and all these things. And somebody who has design taste and design style is gonna be able to create some really amazing stuff. So I think you're in a great position if if you can read and write code as well as have a great design sense. To me, this is like this is a role that I have tried to embody my entire career. This is the the this is the Yeah. Yeah. The skill set that I wish that I had. I've always wanted to be a great designer, and I've never been able to crack that nut. I've always been a mediocre designer.

Wes, you've always been a great designer, and that's one thing I've always been jealous of your abilities on because it it could if you are a great developer and a great designer, the world Wes you already. It's un you're unstoppable when you can do everything,

and and when you can come and fix solutions where it's not just what it looks like. Like, take a look at the the Vercel blog post here I'm sharing. It is, no individual expected to have all the skills that the team collectively is able to execute on ambitious designs because we have design in Figma and code, write production code, debug browser performance. You know, that that's a huge one. Write shaders. Write copy. That that's a big one as well. It's like, can somebody just look at a website that you're building and say, that's not the right tagline. That's not the right copy. You don't have to go through this whole thing of, like, oh, let's get the copywriter involved. Let's have a meeting. No. You should just be able to, like, look at it and say, like, that's not the right thing to say. That just doesn't feel good.

And, like, I I feel like we've I've done that many times on on the Syntax team where you're just like, no. That's that that comes off as weird. Right? Yeah. Three d experiences, Blender, edit videos, using CGI for camera effects, all of that type of stuff JS just like, if you can do it all or if you have a a good taste in all of these things, you're unstoppable in creating things that will be able to capture people's attention and make an enjoyable product in this age of absolutely dog crap interfaces.

Yeah. Yeah. Totally word.

So birthday girl, I think you're in a great spot. Yes. So next question from Mike Swiatek.

Hey, guys. Not a question. Oh, Mike, what are you doing? Not a question. Just a comment. I was thinking about it as I drove to work listening to the anthropic episode where you guys kept mentioning how AI tends to lack taste, and I don't know how else to put it. I would liken it to the nineties and 2000 subdivision boom where developers basically used a template. All houses look the same. Everything looked cookie cutter, and you couldn't tell homes apart.

All of those houses lacked character, and I think that's what's missing when it comes to designs or creating things. It's missing the character that makes it unique.

Yes.

Exactly. And this goes right with a birthday girl's question where the people who have the design sense and they have the ability to see something and make it look special are the people who are are are on it. Yeah. Because we've always had this. Anybody who's been in the industry knows that back in, oh, gosh, back in the early two thousands, what we had the basically unstyled websites, links were blue, backgrounds were white or a single color. Maybe you had repeating background, GIF images or something like that. Then we had the bootstrap era where everything looked like either bootstrap or foundation, and you could see it from a mile away. And then now we have the AI slop era of the same thing where everything looks like you got unnecessary all uppercase labels. You got you you have all the design tells. Right?

It's this is nothing new. It's just a different version of the same thing. And I think this is a really good example because it's yes. They look, like, violently people always say, like, violently Ontario. Like, you you've been to my neighborhood. Right? All of the houses look exactly the same, and they are all awful looking. And and that that's one that's one part of it. They they they're not nice looking houses. But, on the same regard JS, like, there's a lot of, like, weird things about it. Like, my roof has, like, 800 peaks and valleys, and it's all unnecessary. I've been in the attic. It could just be a square. It could just be a little hut. Could be a triangle, but, no, there is probably, I don't know, eight different peaks and points that are all unnecessary, and they just added that because it it looked good.

And it was impressive to the people that wanted to buy these houses in the 2 thousands. And now, like, I we have to get your my neighbor got their roof replaced, and it Wes, like, like, $50,000 or something insane like that because of how complex it is. Right? And there's all of these, like like, usability issues where, like, you you drive into the any garage on our street, and you can hardly open your car door. Like like, did nobody think that you would need to open car doors by driving in these things? Like, there's all of these awful usability things about the houses, and I agree, where it's just like, Wes. That is technically a house, and that is technically a Node, decent looking house. But when you really get into, like, the usability and the flow and how people live in a lot of these houses, there's a lot of, like, weird oddities that the people are trying to now undo in in how these things work. So my rant here is that it's not just what it looks like. With this AI design, yes, we get it. Like, there's purple and there's, like, there's, gradients and stuff added to it. Like, that's Scott the biggest point here. The the biggest point is not what it looks like. It's it's how usable it actually is. Because if you take a look at, like, all of the apps like, if I open threads, Twitter, Blue Sky, and and all of these other, like, short form blogging apps, sometimes I cannot even tell which one I'm in. If I am on TikTok and on Instagram, I often do not know which app I am in. I have to, like, swipe up to actually see what icon it is because they look so much the same. But if I if I am using threads and I get so frustrated with it because the the actual interface of it if you go on threads desktop, infuriating experience. Even though it looks very similar to Twitter, there is so much more thought and care put into the design of Twitter than there is on the threads desktop Wes website. And that there is is a perfect example of it's not what it looks like. It's how the you interact with it, how you use it. If somebody spent the time to say, Wes you click on this button Wes it says a notification, does that notification clear? Or all of these little tiny examples as to how the interface is to to get where you want. And another example is, this is my rant hour of of bad interface. But, like, the Apple, screen time for parents, every single time I need to make a change in my screen time, I get so mad because it's just it's it's an awful like, technically, Wes. There is a list of settings, and you can click on the different categories. You can click through to them, but the usability is is absolutely awful.

And that's a perfect example of, like, somebody actually wanting to use it. Yes. Sanity over.

Brandt over. Yes. Yeah. Yeah. And it you Node, it is funny because that's it's a great analogy, like you had mentioned, but the the cardboard nature of some of the the homes that are built in that era, McMansion kind of era of homes, is very analogous to, you know, the way that AI stuff is is turned out here. So just add it on. And if you want to see all of the errors in your application, you'll want to check out Sentry at century.io/syntax.

You don't want a production application out there that, well, you have no visibility into in case something is blowing up, and you might not even know it. So head on over to century.io/syntax.

Again, we've been using this tool for a long time, and it totally rules. Alright.

Next question from TLC says, I hear a lot of people talking about how LMS have helped accelerate the speed of development, but testing is more important than ever. What I haven't heard is that people are what are people actually using to write tests? Are people writing them by hand? Are we using Claude to generate the tests on the code we just wrote? What about more specific end to end pnpm edge cases? It seems like it would be a huge token spend to explore and write the playwright tests for a complex site using an agent. So, yes, a few things. First of all, your your tests are more important than ever. As as people are just prompting and adding features to absolutely everything, you need to feel extremely confident. There's the amount of pull requests that are coming in, the amount of code changes that are came coming in have skyrocketed.

And you need to have confidence that adding these things is not going to break, other parts of your website. Right? Like, I get all of these pull requests on some of my sites that that don't have tests, and I'm just like I I don't know if you actually spent time doing this or if you just drive by prompt pull request something, and it's gonna break something else. So the tests are more important than ever. I totally agree.

So a few thoughts here.

Unit Wes first for for complex logic is something that I will do quite a bit. So one example on my personal website, I needed to be able to take in a URL for any social network and parse out the ID of that URL. Fairly simple function where I don't know. There's, like, 10 different kinds of social networks, and there's probably 20 different types of links. You know? There's youtube.com. There's u u2.be.

There's, twitter.com. There's x.com. There's there's threads. There's blue sky.

They all have six or seven different types of URLs that you can. So I basically, what I did was I took every possible example of every URL that I want, and I wrote unit Wes for it. So taking this, expect the platform to be Twitter and the ID to be A BC123.

And then having that, I was so much more confident being like, alright, AI. Please write a function that will cover every single one, and you can kinda loop through it and make sure that it hits every single one of those use cases. So unit tests, extremely important in that regard. And then the end to end integration tests are super important for these types of things because you wanna make sure that you can still use it. I don't think that the token spend is is any bigger than writing the actual code. What do you think?

Yeah.

It you know what? For me, it's like Yarn of the the benefit of this type of thing is making your your confirmations of things working more deterministic.

I personally lean more on end to end tests than I do for even unit or integration tests. End to end tests in the web context is really firing this thing up in a browser and stepping through.

Tools like Playwright are so good for this because you can, once, see it happen.

In Cypress and I'm I'm sure Playwright has this too. Cypress had a mode where you could click through the thing, and it's going to write and record your tests. Yeah. You can even put an agent browser on that to do that and have it write the test as you click through it. And then that becomes a deterministic thing where it's gonna run through these things. One thing that I've always loved in a integration Wes right? An end to end test setup is, like, just straight up an app tour Wes it's going on everything, clicking on everything, doing everything one by one.

And then that way, you can just see at a glance something break.

Another thing too with this is that with the visual ESLint. Visual testing is not something that I ever really like to do because styles change and your Wes fail and all this stuff. But sometimes, when working with CSS and agents and and all these things, if you're working at big systems, having visual testing is a a good way to make sure that nothing is breaking, really. So I I really lean hard on end to end tests Mhmm. Especially in this, and moving those possible things out of the AI's discretion. Right? Yeah. Oh, fire up a Chrome MCP or fire up an agent browser and make sure this thing works.

How about instead write the Playwright test, click through it, and then you can run those Wes. And talk about token spend. Like like, Chris or Claude, they want to open a browser to test it. Right? You you see them doing that all the time, and and that's that's way more expensive than just running a deterministic test. So that's Yes. You're in good shape. You you probably use less tokens if you have explicit ones. And, also, like, one thing people hate about tests is that when they're flaky, you start to ignore them because, like, it's actually fine, but something broke. You know? And having an agent be able to decipher, is that just it being flaky? Is that, like, a nonissue that throw them? Can I modify it? Agents are fantastic at being able to look at something that was flaky and say, this is a big Deno. Something broke or no. That that was that's totally fine. Yeah. Totally.

Cool. Next question here from Shabazz.

What are your thoughts about AI being a bubble? Do you think that the bubble will burst anytime in the near future? Is it even possible? Do you think people go back to coding manually in that case, or would it be difficult to go back to coding manually if we ever have to? I think listen. I think I think the information around AI being a bubble is very real. I think the the house of cards there in terms of, CPU or GPU ownership and things like that. I I think all of that very much makes sense.

Do I think that the bubble will pop and that AI will go away? Absolutely not. I I I think we're I think we're kind of in a world where, people will just be able to do whatever they want. And I I don't I I think there's too much money involved here. I think the companies are just gonna keep passing money around.

AI is gonna keep existing until either it becomes profitable or there's just going to be a couple of big players, and that's it. Yeah. Either way, the, what is the the cat cat going on the bag? The rabbit's not going back in the hat. The wolf Genie is in the bottle.

The genie is out not going back in the bottle.

Node. So are we ever gonna go back to writing code by hand only? No. Absolutely not. I don't see see that ever happening. I cannot imagine a world where that exists.

Folks, this is the reality. This is this is it, man. This is what's going on. So, like, I I just don't I don't see it.

So to me, you know, I I think the world that we live in is you gotta get good at working with this stuff if you want to continue to be effective. Yeah. That's just where I'm at on this. How we are building this stuff has

fundamentally changed forever, and that is not going back. The bubble itself like, CJ and I had a video called the AI bubble is bursting, and that was about how expensive these things are getting, where every single company six months ago was like, whatever.

Sign up. Spend as much as you want. Go. Go. Go. We're changing. And then here we are six, eight months later. A lot of these companies are going, oh, this is kinda getting expensive.

We're not spending this much per employee because this is this is hundreds of thousands of dollars a year.

And the bubble of Anthropic being able to charge every single engineer $5,000 a month, I think, is is going away.

And we're seeing that by people realizing, oh, some of these new open weight models are 10 times cheaper and 98% as good. So that's that's where I'm going. So that the bubble of, like, how much money these companies have raised and, like, are they going to be able to get that back? I I don't know. That's the kinda the crazy part about it is that, like, what if we all just switch to the other models that are 10 times cheaper and just as good? I

I I think that the the the way you could think about this, Shippaz, is that when the .com bubble burst, did the Internet go away? Was, like, websites gone? No. You know? Just like, pets.com or who JS it? Whoever is it? Was it pets.com? One of the they were like the poster child of the whole, the .com bubble burst. Either way, when the .com bubble burst, it's not like websites or the Internet or any of that stuff went away. It's just that, some people got shut down. A lot of people did. And And and the things of, like, people are going to be ordering dog food online

turned out to not be true. It was it was just too hard. And and I think a lot of the the things that big AI companies are saying Wes it's going to be replacing absolutely everything, and there's gonna be no jobs for for anybody anymore. I don't think that is true, and and I think that that's the bubble that is is going to be bursting.

We're already seeing people, like, having a hard time believing anything that OpenAI, Anthropic are saying anymore because it's just like are you saying that because you're trying to pump the stock value, or are you saying that because you actually believe it? Yeah. Right. Yeah. Totally. They, they kinda lost me at the AI needs a therapist one. That was that was my breaking point where I was I was like, you are so dumb. The the open claw becoming sentient and AI models needing a therapist, Wes kinda it kinda lost me there.

Yeah.

Yeah. It's it is wild. Yeah. Alright. Next one here from Anan. What are your thoughts on the negative public sentiment towards AI and data centers? Do you think it would affect vibe Node products or products made with AI even if they're somewhat decent? Yeah, man. Data centers suck. You know? I think they're a necessary, evil if AI is going to continue to exist. And I say necessary in terms of they are required for the AI to continue to expand. Data centers, quite possibly for the compute, they need to exist in that sense.

Do I like data centers popping up in different communities causing messing with water supply, you know, heat and energy usage and all that stuff? I hate that. And I I'm like, I hate how, one, I hate how it's affecting communities, and, one, I ultra hate how politicians are collecting money and then just throwing these in their communities and saying whatever. You know? I hate that. And I don't I don't I don't have anything educated to say beyond that.

You know, it, like, kind of drives me nuts when I see, like, people living in, you know, major cities being like, LOL. Let's throw some data set. Like, I don't know if you see this, Wes, but I see this in Twitter where there's, like, people like pro data center. Like, yes. Let's hack data centers here and there or wherever.

But, like, it's not affecting you. So, like, that I don't know. It should it should be up to the people of the communities they're being put in. That's just yeah. I think it's it's kinda tricky. So they tried to they tried to put one here in Hamilton,

on steel land. Like, literally, like, this nobody wants to live there. It's awful. It's polluted. It's like a old steel mill that is just, like, unusable.

And and everybody said absolutely no way. And I I think part of that is for legitimate reasons where it's like, yeah, we don't want this thing having lights and humming all day long. Like, you see it even.

You drive in Ontario and you see the, like, cannabis greenhouses, they just glow. Yeah. Like, imagine imagine living next to this glowing orb.

In the middle of the night, you can just see this That smells like weed.

Yeah. That that sucks, and and I I feel for those people that that has popped up in their neighborhood. But then, like, in in the same regard, like, you can see a lot of these people protesting against it, and they they don't really understand it. Like, part of it comes from an anti AI sentiment because they they've been told by all these wieners that their jobs are are going to to go away, and it's gonna absolutely destroy the world. Right? Like, stop being a wiener and start stop talking like that. That's why these people don't like you.

That's that's certainly one of them. But then, also, like, these people don't understand. Like, I see, like, a TikTok where someone's like, no AI data centers. AI is awful. And then they, like, put the AI captions on there, and then they put the AI green screen behind them. And, like, they don't they don't understand that. Like, for you to be able to use these these apps and these features, we need these somewhere. You know? So for some regards, yes. And, like, for, like, good things like medicine and for you to be able to use social networking and and enjoy your TikToks, you need those. So I think there's a there's a bit of a a balance here, and it's kind of a a hot hot topic. And, also, I don't even know. Like, the water is is one of them. I have this Wes, and nobody can seem to answer it. Like, where does the water go? Doesn't it just come doesn't it go back into the clouds and rain again? How does water work? How does rain work? I would like to talk to somebody that's a bit more dialed on that and not so, doesn't have, like, like, a a horse in that race.

Yeah. Well, I think that's tough because I think a lot of the people, have horses in the race. Yes.

Yeah. And and one thing that bug I think maybe I'm sensitive to this because, specifically, like, Denver, man, we we're we have a major drought this year.

You know, the I worry about the climate here. Mhmm. And Denver is worried about it too. They put a moratorium in effect, so no no more, data centers for one year. In Colorado, in in general, I think a lot of communities here are very anti the data center, specifically because we have limited water here.

Like, the water that we get comes from snowfall in the mountains and it melting. And if we have a bad snow year like we had last year, then there's, like, no water. I can only water my grass on, what, two days, two days a week, and it's dry and super a 100 degrees here. So, like, I I mean, the water stuff is a sensitive conversation. Like a like a liquid cooled

PC.

You put the the water in or, like, the cooling the coolant in, and, like, it just, like, cycles forever. Right? You're not just, like, like, putting toxic stuff in it and then just firing it into space. You know? Doesn't and I I'm I may be, like, totally up to lunch. You guys can let me know in the in the comments. But, like, are not new data centers closed looped? Meaning that, like, you would fill them with, like, water and coolant, and then they would not you're not just, like, constantly turning the tap on and refilling these things and and cooling things off. I think that's a little different than, like, like, pumping them into crops, like, pumping it out of the the ground and then, like, making, like, almonds or something like that.

Yes. I don't know the answer. Folks, we are speaking ignorantly. So if if there's any good resources on this stuff, leave them in the comments. I wanna know all about this. Next question from a a Aaron.

Can we get an update on Scott using Linux? I sold my Mac almost a year ago and went all in on Linux. I have a Linux desktop and a laptop, framework 13. I love Linux, but the thing I'm discovering, it feels like I need Node at least one stable machine in my life. I miss my MacBook and will probably buy one later on this year. I was wondering how Scott feels since the Omarchie episode.

Yes. Well, you know what? I I think my, well okay. May I answer your question, a, Aaron? I have been using my MacBook Pro full time. Reason being is that the computer that I was using for my Linux is quite old.

I think I bought it I think I bought it through ESLint. So it might be, like, from, like, 2016.

I mean, that might be a ten year old laptop. Yeah. You have the way you have the framework 13. Right?

Yeah. Yeah? No. No. I don't have the framework 13. I have the, it's made

by system seventy six anymore. Ran another random number.

I think you bought that during the March. I bet I bet it's eight years old if I remember when you bought it.

It's old Yes. For a laptop. For me, you know, I I used to upgrade my laptop, frequently. So, this this MacBook Pro is, like, the longest I've had any computer, and it is showing its age right now. But either way so that computer was not even good to start with. It was like a cheap lemur that I I got to to have around. And so because of that, it's it's not impossible it's not possible for me to use JS my full time machine. We're doing a lot of media. If I was doing less media recording, I think I'd be fairly happy working on that full time. The biggest issues I had with Omarchie Chi in general was Omar Chi Omar Chi. I have no idea how to say it still today.

Was, like, the dual monitor support I thought sucked. And and there were people in the comments being like, it's not that hard. You gotta just open up this config and write these following config ESLint. Like, brother, just let me plug it into a monitor and have me drag stuff around like that or or key easy keyboard shortcuts. Yeah.

But could I use Linux full time? I think I could. I think I could if it wasn't for my media stuff. If I was just a developer, I think I would be very happy on that. As far as what am I using Linux for today, I have a Steam machine, the video game system that I'm running Linux on and doing some non gamey stuff on it. But two, I am going to be building a CyberDeck, which is fully Linux and will be existing in a little handheld Linux computer.

And that will actually be my beefiest Linux computer even though it is going to be a handheld CyberDeck. So, folks, keep your eyes out for that project because I'm gonna be throwing Linux on that bad boy, and I'm gonna be doing some cool stuff with it. So I can't wait to show it off, actually. I

would love to do it, but there's just the media stuff. You know? Like, I need Photoshop, and I need a DaVinci. I know DaVinci Resolve runs on, like, one version of Linux, and you can, like, kinda get it running. I, like, looked into it at one point. I'm like, can I really get, can I really get DaVinci Resolve running on Linux? And, like, all the reviews are like, it works pretty good. Like, I don't want pretty good. I don't want, like, oh, you have to do x, y, or z.

No thanks. And then, also, like, a lot of the, like same with, like, Windows people. They're like, what's the equivalent of x, y, or z on Windows? And there's always some, like, half baked watered down version of something. Like, I'm not using I don't know. What's the, like, Photoshop alternative that everybody always uses? Like,

some crappy sketchy Vercel. The the one on Tolinski. Yeah. I don't I don't know. Like yeah. For I haven't used Photoshop in so long, so I understand that.

GIMP. That's the one. I'm not using GIMP.

No. I'm not using GIMP. And I you can't you can miss me on on GIMP entirely. And, like, people talk about, like, Sanity being good enough, and then it's like, I use Audacity to record, Courtney's podcast.

And, man, Audacity sucks.

And I'm not saying that's the only audio thing, but Audacity straight up sucks. It it like, there there's so many annoyances with it in general. Look like it's built in Java and Windows XP, or did they file and spin that thing? Yes? No. They they it it I would say it looks a hair better. Okay.

At least unless they're coming out with something new that looks I mean, even the logo. The logo should tell you what it looks like enough. It looks the exact same.

They've made it look a hair decent, but no. Like, that's such a thing for me that's like, the design of the application has to be nice still. I'm I'm just sorry. I I need that. I'm Yeah. I'm I'm I'm not Oh, here's the thing. Sanity four

looks a little better than I remember it. I'm just looking at it right now. But Audacity four beta two. Okay. So it's just in beta. For that type of stuff. I would absolutely love it because there's so much in macOS that I I absolutely hate, but I just don't have time to fuss with that type of type of thing. And, also, I just love a computer that you can upgrade. You Node, I'm looking at these framework laptops, like framework pro, where you can, like, put more RAM in it if you want or you can you can choose which ports you want. Like, one of my biggest qualms with my MacBook Pro is there's I'm I'm over almost over this, but there's no USB a on it, and there hasn't been forever. So you have to carry the dongle around because so much of my stuff is still USB a.

And I'm almost to a point now where everything I own is is just straight up USB c, But still, every now and then, you you run into something weird. And you're like, oh, now I gotta find the dongle and and put it in. I don't I don't use anything USB a. That port is is not something that

that's Scott something that I think about. But, like, no. Nothing. Not a single thing. Oh, maybe not a single thing. I don't know. But unless I'm at my desk. And if I'm at my desk, I do have, like, the big unit, which I need anyways to connect all my stuff because I got too much stuff to connect in. So, like, yeah, I don't know. I'm using I I I I personally I think I could I don't know. Man, I I'm happy. I think my next computer is even gonna be a MacBook Air and then get a Mac Studio and have that be chilling here. That's

we had a couple other questions of, like, like, what are you waiting out for for, like, the local the local, AI thing? And it's just, like, some b p ass Mac Studio that that you can buy and slap in I I like, it's just it's too early. People you can't really do it on hardware unless you have many, many thousands of dollars to run, but maybe new chips in the next, I don't know, couple years. It's gonna be a few years before this type of thing. And I'm sitting here on my m one MacBook Pro. It's starting to get a little bit slow as well, and I'm kinda like, I don't wanna upgrade yet. But

yes. I know. It JS it is like my fans are running all the time on this thing. And I I I'm thinking, like, taking out a,

another mortgage to get a Mac Studio with a 1.5 terabytes of RAM whenever that drops. I know. That's actually a joke. I'm not not considering that, folks. But That's insane. You know what? One thing I did is the Mac Pro, when they moved to the 16 inch from the 15 inch, it's just you can't bring it on a plane anymore. The 15 inch would just fit underneath the little notch, and, like, I cannot Node on a lap on a plane anymore with the 16 inch, and I am getting the 14

next time I get a laptop. I think this is the power of being short and and kind of slight here is that I can I can do my stuff on the laptop on the plane? I'm a little guy. I can I can crunch up in a little ball and use my laptop? I can get, you know, I can get in here and just oh my gosh.

Shortcake.

Just it's just a little guy.

Yeah.

Are you ready for this next one here? Yes.

Court, we are ready for you. Next question we have, a special guest for because Scott and I don't know anything about kids or parenting or whatever. So Syntax Face's dad says, how are you teaching and exposing your kids to AI? Are they asking to have AI chat Bos friends, or are they starting to ask Claude instead of Google? How has this latest round of tech shown up in family life, and what are you hearing from parents in your circle? So we thought, let's bring in the expert, doctor Courtney Tolinski, knows a ton about, children. So you want give us people have probably heard you before, but give us a quick rundown of your credentials and and what you do before you answer this. Yeah. Hi. So I'm court doctor Courtney.

I'm a licensed clinical and school psychologist.

And I've been working with kids and families in schools, private practice, etcetera, for over fifteen years. And now, actually, Scott and I have, well, we've started a podcast about child development.

So it's called Phases dot f m, and now I get to talk even more about all these things, and it's been great. That's awesome. So what do you think about AI in kids? Is it a good idea to expose them to it? Yeah. Well so this one's tricky, and and I think it depends on, you know, the child's it depends on so many factors. Right? But, obviously, the child's age, their maturity level, you know, and and where they are developmentally.

So this person did not leave this child's age, but let's let we'll start at, like, I guess, high school. We can start there. So high schoolers, we're imagining, are using AI more in schools. And I know at least for our kids, they have Chromebooks, and so Gemini is just kinda built into their search ESLint.

And so it's just already there. You Node? I I don't know how much the teachers are really monitoring that, but I do think it's important at home to have these conversations with your kids. So if it's a high schooler, they might be diving more into, like, the chatbot friends, and using it to supplement their actual answers or responses Mhmm. Or, you know, what they should be doing independently. So I think it's really important to teach kids to be critical thinkers, and it's something that we've been talking about here with our kids who are only, you know, seven and nine. But I think it's even more important at that high school age to really have that conversation and to let kids, you know, understand and know that just because it's a computer, you know, it doesn't mean that it's going to be accurate. Like, we do tend to think of computers as being, like, you know, accurate and error proof, but, you know, we know that AI is not. Right? It hallucinates.

It is not always correct. And the more you have conversations with it, the more it starts to make mistakes. Yeah. So, you know, having that understanding really built into your kids of it can make mistakes and it is not error proof is really important and, again, to be critical thinkers. So Mhmm. To be able to, like, think about, is this answer correct? Is this, you know, aligned with what I know? Yeah. Do I need to, like, double check this or,

you know, go elsewhere for information? I I was showing my kids the other day how AI is highly agreeable, and I showed it by just doing two separate prompts of, like, things that you thought, and and it it agreed with you in both cases. Because often, we'll we'll use it to, like, not like resolve a fight, but, like, someone someone says one thing and the other person says says another thing. And they'd be like, hey, Gorgle. Is it true that x, y, or z? And it's really funny. I Node them at how you just phrase it in a certain way, and then it will come back to you with an agreeable answer to you. And they're like, oh, wow. Like, you can it's not the definitive answer.

And I think I feel like optometrists have have a high level of skepticism towards a lot of the stuff. Maybe even too much. They think, like, the news they think the news is AI because it's so well produced or anything with, like, a, like, a blurred background, they think JS AI.

Oh, that's so funny. Yeah. So maybe not I mean, a healthy dose of skepticism.

Critical thinking because it you know, you can it's an important life skill for all adults to be critical thinkers of anything anyone is telling you.

But but I like that you're doing that

with your kids. I think that's great. Another thing is we we have it in our car now. We have Grok in our car. And it's funny because when we're in the car, we're often talking about things and questions will come up.

And I really like having it because you can find the we have a button that literally you can get an answer to anything.

And sometimes I'll I'll pause and be like, well, let let's talk about it. What do you think? You know? Like like, that there's that critical thinking of, like, don't press the button to answer absolutely everything immediately.

It's nice that we have it, but let's talk about what it have critical thinking. Think about what the answer might be.

What are your thoughts? And then we can ask Rock what it thinks, and we can kinda build off of there. We were just talking on our episode about boredom, Wes Yes. About, like, how it's good for kids to be bored

and how even with, on demand TV, it's so easy to get everything instant. And I think that, like, really aligns a lot with that. Like, sometimes you need to suffer through discussion or use your brain and figure things out. And if you're just getting everything instant, instant, instant, instant, like, what's that teaching? It's not it's like not it's not teaching critical life skills, but it's like ignoring that entirely.

Yeah. Well and I think that's so tricky too because, like, just the process I mean, if you go back to high school or college, the process of writing, I mean, it takes forever, and there's so many, like, drafts and versions that you have to do. There's so much effort, and it's just so easy to, like, put it into, you know, AI and have it do it for you. But then you're missing all of those foundational skills, you know, that you need to be a good writer or a good thinker. Mhmm. You know? So I I think it's a tricky balance because I can understand that temptation

for kids to wanna just put it in there and just get the answer. Yeah. You know? Instantness. I was even thinking this is not AI related, but I was thinking about the are you guys into, like, Neatos and all the squishies right Node? Your kids are are into those? We don't have any, but Oh. Yeah. I know it's it's a big toy right now. Yeah. Everybody loves getting these squishies, and they're they're really hard to find. And there's all different kinds, and and kids want them. The boo boo? It it's it's even crazier than that. It's it's absolutely nuts how these kids want these the latest squishy. That is you know, they they see a video about it. Their friends are getting it. And I remember that as a kid of, like, wanting the latest thing, but, like, now we have the Internet where you can just press a button, and it will come the next day, at your doorstep. And, like, I was thinking about, like, being okay with not having something.

Like, it was a was a very good lesson for me growing up where, like, I didn't I didn't have all the stuff that my friends had and simply because you could not find it. I remember I wanted a visor so bad. You know? And so stupid, but, like, I wanted a visor for, like, a year. And then finally, we found one at, like, a discount store, and I was so happy about it. But, like, that like, just hold out. We'll find it somewhere JS, like, I feel like such a good life skill.

I also wanted a visor, and I I ended up getting I got to skate at the work tour. I got to Scott at, like, do, like, Scott at the work tour, and I got, like, a a a visor and a T shirt from it. So I was like, I got my visor. I can put my, bleach tips through the top, you know, bike them up real good.

Oh my god. One other one other thing I'll add here is that I we specifically don't let our kids have the, AI image transformation, because, like, a couple times we've done it where you take a photo and then you try to make a silly thing, or, like, what would it look like if I have this haircut? But I think that that could be very damaging to children Wes Yeah. The what would I look like if I was prettier? What would I look like if I Mhmm. Had a different color hair? And then all of a sudden, you just don't start to not like yourself because the photo that the AI comes up with JS looks so much better.

Yeah.

Yeah. I I think that's a good point too. And, you know, I was talking about high school earlier, but, like, going to, like, elementary school or middle school where we are now, we monitor our kids when they're, oh, and they Computing in general. Yeah. Computing in general. But if they're using AI, we're using it with them, and we don't let them have a lot of exposure to it. But I made, like, a really cool project with my son. It was like an affirmation poster, and I had him, like, type in, you know, the things that he wanted it to say, the things that he's good at and his strengths, and we made it in his likeness. It was really neat.

But, you know, I did that with him, and we talked about the process. So I think introducing kids to AI is important because, obviously, we know it's not going away, but you have to do it in a way where you're, like, guiding their usage and monitoring it. And, obviously, with chatbot friends, you know, things like that, having them understand that it is, you know, a predictive model. It does not care about you. It does not have feelings for you like a human being would. And, you know, having those that distinction in those conversations, I think, is so important, but doing that exposure in a way that is, you know, supported and guided

by you. Mhmm. You know what we should do, Wes, is we should get CJ to do his video that he did about building an LLM from scratch, teaching LLMs and how they work. We should get him to redo that in, like, a blippy costume. Bos.

You guys? Yeah.

Do you wanna hear about tokens?

Yeah. See, we'll put them on that. Would watch that show. I know. Yeah.

Yeah. Do it in, like, a Mark Rober style. Yeah. Oh, yeah.

Oh, yeah. That would be great. I like that a lot.

Cool. Okay. Cool. Well, thank you so much, Courtney, for your time. I appreciate it. Check out faces.fm if you wanna learn a little bit more about how to raise your kids.

A lot more to that, though. It's awesome.

I'm just reading the phases website right now. Doctor Courtney Courtney Tolinski, PhD, PhD, LP, NCSP, licensed clinical in school psychologist, and her husband, Scott.

Yes.

Yes.

Just Scott.

Just me. That's me. Just Scott.

Oh, that's hilarious.

I don't have any titles. I Deno. Nothing Full stack? Yes. Yes. Full stack. Yes. Exactly.

Alright. Next question we have here from Geezy.

As someone who has spent much time learning to program the skills without AI, I was recently turned away for not having much experience using AI to build things.

What would you recommend for someone who doesn't really work with AI to you using it as a develop development tool effectively? This is crazy that we're starting to hear these things. Whereas, like, it used to be, can you code without the help of AI? And now people are being turned away because they don't know how to use these tools effectively, and and they wouldn't make for a good employee.

And I think, like, as as much as you we joke about just typing in the box what you want and it comes out the other end, there certainly is a huge skill to be able to use these effectively, be able to choose like, one weird skill right now is that you have to, like, you have to know how expensive these things are. Like, we were talking yesterday how, like, some people are spending thousands of dollars a day, and we're like, how can that even be possible? And and one of the ways is that people just continue chats in the morning. Yeah. And if if the chat is not cashed, then let's say you had a $150 chat yesterday, and you you you say, alright. Let's keep going today.

That's another $150 tokens when because it's it's no longer cash. It's no longer hot, and it has to send that whole back history, and then you you've racked it up. So, like, you have to know about these types of things in terms of how much they cost and JS well as, like, how do you you steer them? How do you how do you write plans? How do you tackle them? What skills are you using? All of that type of stuff. And how do you figure this stuff out? I I think you just gotta start using it. You gotta say, given that what I know about being a good programmer, how can I wield this tool in such a way where it will output high quality code and software for what it is that I want? So you just have to, like, really get in. Listen to us. We got lots of tips on this podcast as well.

Yes. Yes. And, honestly, I I I think, like, that is such like, the other like, we talked about in the is the the the genie going back into the shoe. It's it's gonna become a it's it's not going to become. It's become a a part of this job. Mhmm. Just JS, like, at one point, you're like, alright. Test driven development. That's, like, a part of this job you need to understand. Whether or not you're doing it, you should probably understand how to be effective with it. Or even for full stack engineers, when JavaScript frameworks came into play, it's like, alright. Now you gotta learn how to do, UI and JavaScript stuff. You Node? I mean, there's there's always going to be these shifts, and, this is just another one of those shifts. You have to understand how to be effective with it, or you will get turned away from jobs. And and I I there's going to be exceptions to that. It's not every job. But I think more and more, you're gonna see that.

Yeah. Yeah. Your your basic, like, run of the mill your your common job is is going to want to see that you are good with these tools no differently than how they wanted to see you're good with your your text editor setup, your your GitHub integrations, all of that type of stuff.

Just I I know a lot of people have been ignoring it for for so long because they thought it was was bad or whatever, but, it you're not that genie, unfortunately, is not going back in the bottle Tolinski Scott says. Alright. Last question here in this, like, all AI episode

from Keyboard Tapper.

How do you keep coding agents, clog Node, cursor, you name it, efficient, token usage, speed, hallucinations, and accurate in big code bases? Do you cap agents dot empty size? What I try to do is I try to keep code files small, small React Node components.

If there's a linting rule, rather catch it with the linter than writing it down inside of claud.md.

Store the why of technical choices in separate ADRs and link to them. Be mindful of compacting conversations.

Curious on what you guys do. Thanks for the show. It's awesome. Let me tell you. I do a lot of this exactly like you do. I try to make things deterministic. We talked earlier in the show getting, the the the correct ESLint. Take the agent out of just testing stuff to writing tests and having those tests execute and end to end setup. But, I use tools like Fallow to make sure that, code duplication is taken care of and things like that. There was a really good blog post from Ben Vinegar on the modem site, modem.dev.

I'll link to it talking about writing code that's effective and efficient for agents because you're creating functions that have unique TypeScript names rather than generic names and making your code more findable, reducing token spend in that way. I do use the ADR approach because I have found it to be good. So ADR is like a decision document that documents why you make decisions that the AI can reference when it only when it needs to instead of just throwing all that stuff into an agents dot Node. Again, take anything that can be, deterministic out of your agents. N d. Put it into hard things.

Cloud code and open these things have the idea of, like, hooks.

And if you can build in hooks to catch the AI in a deterministic way, it's trying to do something. You stop it and then step in there.

That's great too. Again, any of these tools that can be caught by linters, what what do they what is Fallow considered? Fallows static code analysis, anything that could be, caught in those ways, I do so.

Small code files, man, it loves to make a big ass file. It loves to make a big ass file.

So making a deterministic way to say, hey. I think this funnel's getting a little big. Definitely always something good to have. I I think that those are great. We've we've done several episodes in the past on, like, tips for this type of stuff.

But, Scott, nail it there, is make it as deterministic as possible.

Don't absolutely throw so much in your agents.md.

In many cases, those are going to cause slower responses and way more Node generation. Right? Like, I I when the skills first came out, I slapped so many different skills in here and had all these different, like, superuser or whatever.

I I ripped them all out. You don't need them. I certainly still have a handful of skills and and rules per project.

But Scott of this stuff, you don't necessarily need, and it's it's just making things messier than normal. I have skills for, like, like, here's how to use the Svelte,

agent tools or whatever. Okay. Here's how to use Dex, which is our thing. So that way, it doesn't have to do Dex help. Dex is a a task list thing that I use. So it's it doesn't like, JS if you don't have a skill for that, it's always going to be like, let me go to decks hyphen hyphen help and see how to use it. So I'll use skills for things like that. But, again, like like you said, Wes, yeah, I'm looking at a agents that Node file I have right now that's way too long that I need to, knock out here.

I I also, like, what people love to do is they they love to say, like, analyze my code base and then create a agents dot m d based off of that, and then it just duplicates absolutely everything. And a good coding model is able to infer a lot of that stuff without you having to, write it down explicitly.

So it's it's, in many cases, not necessary.

Yes. That's all I have here. Wes, do you have anything else for us? Do you wanna get the sick picks? And,

yeah. I do. I have the sickest pick for you.

We just spent some time up at our Scott. And one thing that I built, I don't know, probably eight years ago at our cottage Wes a floating dock.

It's like a 12 by 12 raft. It has those, like, blue barrels underneath, so it floats, and it's super fun.

And it's connected to the ground with two huge, huge chains, and then there's cinder blocks at the bottom. And over the winter, one of the chains snapped and fell to the ground 15 feet underwater. And I was like, oh, no. Like, first of all, the chain is really expensive. I don't wanna have to buy a new chain. It's a couple $100.

But, also, 13 feet is too too deep for me to dive. I'm not that good of a diver, and nobody I know can dive that deep to be able to find it. And we also didn't know where where it had fallen. Right? And it's, like, mucky or or you Node? You never know what's Yeah. It's just it's it's sandy. It's not mucky, but it Not much. Sanity enough that you can't see through it.

Yeah. So what we ended up doing is we bought something that I've wanted forever, and that is a magnet fishing set. I I don't know if you ever been on YouTube and watched, like, magnet fishing Wes they throw a magnet in the lake and they come out and there's, like, iPhones and gold on the on the bottom.

I've wanted one of those forever. So I think it would be so fun. And, we bought one because it was cheaper than buying a new chain, and we just dragged it around. Took a a couple hours. We dragged it around the lake, and then all of a sudden, click, Wes found the chain, and we're able to pull it up most of the way and then dive down and and grab it with our hands and and tie a line to it. But I that was so fun. And then we went to a couple bridges and threw it over the bridge trying to find stuff, but I'm gonna I think we're gonna have the rest of the summer doing a whole bunch of magnet fishing. So I highly recommend it's kinda scary how how Vercel, 1,400 pound magnet. I got it stuck to the bridge, and I almost could not take it off with all of my might. Gonna say. Yeah. Yeah. That sounds crazy. Yeah. You gotta be careful you don't pinch your fingers, but it's fun. Oh, yeah. Oh, that's the type of thing I would do.

That is exactly the type of thing that I would do.

Yeah. Yeah. Okay. Sick. Yeah. I love this. That sounds a lot of fun, man. Being outside, doing outside stuff. Love that.

I'm gonna pick something about inside but not computers.

I hate, kitchen brushes. We we I don't know. A lot of them don't work very well. But I got a a kitchen brush that I actually really like, a a cleaning brush for the dishes that I actually really like. I like enough to talk about as a sick pick.

I like this one because not only does it work Wes, it's $10. It's, just a kitchen sink. But it has, like, this notch on the the side of it, allowing you to hang it over the side of the sink.

So a Oh, that's cool. Drying, you're not just setting it on the counter, getting the counter on what. You hang it on the side of the sink.

And it's just a stupid little feature, but it's a delight to use. And, also, I've I don't know what is it about this kitchen brush. It's not particularly, like, brittle, but I found it's really good at getting stuff off the pans.

I have I have some

sweetest scraper head on it too.

Scott, let's scraper head. We have these stainless pans that are, like, really hard to clean sometimes. Yeah. And it's been super effective on them. So I guess I'm sick picking a $10 kitchen, kitchen scrubber, but my gosh, it's made my life a little bit better. So I'm I'm putting it in front of y'all. Make

me want to we use the IKEA dish brush because, like, they're also pretty aggressive, and they're, like, $2 a piece. But it doesn't hang over the side. Now I wanna oh, you know, ours has, like, a little suction cup.

You can put it on the side of the sink, but, like, then it's in the way of the entire sink. What this one does is it just hangs on the end. I wanna three d print something for mine now that will do this. Clip and great deal. Clip and print. I know. Our ours, like, got ours was needing

a a new one, so it was a perfect time to see what kind of advanced technology exists out there for her.

Dishbrush. Love it. I we we've talked about cleaning stuff way too often on this podcast. We got the scrub daddies because of you, and I love a scrub daddy. Scrub daddy it up. I'm going to shamelessly plug really quick Courtney's podcast, phases dot f m. That's p h a s

e s, like phases of life, phases.fm.

We're also on YouTube. And, man, she's like, she went from not doing any of this type of stuff to putting out shorts every single day.

And Just getting me out. Doing a great job at it. She's doing an awesome Bos. So check it out on YouTube as well. I've been, like, really, really impressed with her efforts there. So we got her we got her in those little DJI mics and everything like that. That's pretty dope. Yeah. I need to try one of those. My wife has the Hollyland,

Hollylark, whatever. We got it off of, producer Randy's recommendation.

And it's really cool, but I I kinda I I want one for the shorts, and I wanna hold it. Like, I you know, it drives people nuts because you're not supposed to hold them, but I I get it.

I don't get it. I don't get it. I don't get why people do that. It's because it's it's a visual thing. It's something to look at, and people clip on to funny things like a dish brush. You should try.

Clip it onto your dish brush Okay. And and record a short. I guarantee it will do well.

I've been watching you, Warhammer YouTube Warhammer painting Yarn, and I've noticed the guys clipping it onto a paintbrush and stuff. And Yeah. Now that you say that, it's like, oh, I get it. Okay. Yeah. Oh. Drives audio people nuts because it's not how you're supposed to use it, but nobody cares.

That's it for today. Thank you so much for tuning in. Send us your potluck Wes, syntax. F m.