The Changelog: Software Development, Open Source - Change my mind (Friends)

Episode Date: February 21, 2025

Jerod and Adam use Chris Kiehl's post on development topics he's changed his mind on (over the last 10 years) as a proxy for discussion on dev things they HAVE and HAVE NOT changed their minds on.

Welcome to Change Log and Friends, your weekly talk show about changing your mind. A massive thank you to our friends and our partners over at fly.io. Learn more and deploy your app in five minutes at fly.io. Okay, let's talk. Well, friends, before the show, I'm here with my good good friend David Shue over at Retool. Now David I've known about Retool for a very long time. You've been working with us for many many years and speaking of many many years Brex is one of your oldest customers.

You've been in business almost seven years. I think they've been a customer for almost all those seven years to my knowledge. But share the story. What do you do for Brex? How does Brex leverage Retool? And why have they stayed with you all these years? So what's really interesting about Brex is that they are a extremely operational heavy company.

And so for them, the quality of the internal tools is so important because you can imagine they have to deal with fraud, they have to deal with underwriting, they have to deal with so many problems basically. They have a giant team internally, basically just using internal tools day in and day out. And so they have a very high bar for internal tools. And when they first started, we were in the same YC batch actually, we were both at Winter 17 and they were, yeah, I think

maybe customer number five or something like that for us. I think DoorDash was a little bit before them, but they were pretty early. And the problem they had was they had so many internal tools they needed to go and build, but not enough time or engineers to go build all of them. And even if they did have the time or engineers, they wanted their engineers focused on building external physics software, because that is what would drive the business forward. Brex mobile app, for example, is awesome. The Brex website, for example, is awesome. The Brex expense flow, all really great external

software. So they wanted their engineers to focus on that as opposed to building internal crud UIs. And so that's why they came to us. And it was honestly a wonderful partnership. It has been for seven, eight years now. Today, I think Brex has probably around a thousand Retool apps they use in production, I want to say every week, which is awesome. And their whole business effectively runs now on Retool apps they use in production, I want to say every week, which is awesome.

And their whole business effectively runs now on Retool. And we are so, so privileged to be a part of their journey. And to me, I think what's really cool about all this is that we've managed to allow them to move so fast. So whether it's launching new product lines, whether it's responded to customers faster, whatever it is, if they need an app for that, they can get an app for it in a day, which is a lot better than, you know, six months or a year, for example, having to

schlep through spreadsheets, etc. So I'm really, really proud of our partnership with Brex. Okay, Retool is the best way to build, maintain, and deploy internal software, seamlessly connected databases, build with elegant components, and customize with code. Accelerate mundane tasks and free up time for the work that really matters for you and your team. Learn more at retool.com. Start for free, book a demo, again, retool.com. All right, should we talk about change of mind,

change in our mind? Should we talk about Chris Keel? Whew, yeah. That's how you say his name. Shout out to Chris, a software developer and overall pretty cool guy. His words, not mine. I can't vouch for whether not any of that's true.

I just read it on a webpage on the internet, but Chris writes that and he also writes software development topics I've changed my mind on after 10 years in the industry. Now this did not make changelog news for a funny reason, maybe not funny, maybe an unfortunate reason. Title's too long, dude. I just couldn't figure out a way of getting that title down

to where it made any sense. And so I was like, tough nuts, I guess. You're not gonna be on Genealog News, but I thought it was a great post. Yeah, that is tough. You have to paraphrase the title to get it in there. Yeah, I couldn't even think, how would you shorten that, Adam?

Let me try. Let me test my wits here. Yeah, show us your compression algorithm. Software development topics I've changed my mind on after 10 years in the industry. That's his post title. Dev topics I've changed my mind on after 10 years in the industry. That's his post title. Dev topics I've changed my mind on after 10 years. That's pretty long still.

I mean, that's about as much as you can short developments to dev, topics has gotta be topics. I've changed my mind on it's, I mean, you can't lose that phrase. And then after 10, which is the key numeral there, of how many years. This has become a pattern, hasn't it? Something I've learned after,

and then it's like whatever many years it is. I feel like we've had a couple people who've written successful posts like that, and then other people are like, oh, I've also spent 10 years doing a thing. It's a good limiter, yeah. You can sort and limit pretty easily, you know? Cause if I've spent 10 years in the industry,

I might be like, okay, I'm inclined to read this because I may have similar or the same takes. But if I've been in the industry two years, I'm like, well, I don't have the depth. So that one's not for me. But maybe I'll watch anyways, maybe I'll check it out anyways. What's the lower limit you can put on a post like this and still get some traction?

Software development topics, I've changed my mind on after six weeks in the industry. How low can you go and still get people's attention? That's a good question. I think a year for most topics, I think AI, maybe like weeks. Right. You know, I mean, cause really,

I think I had some different ideas three months ago and some of those ideas are still the same, but they're not, they've matured or they've morphed a little bit, you know. Well, a change of mind does require some prerequisites, right, you have to have formed an opinion previously. Yeah. And then you must have been holding that opinion for some amount of time to test it against the real world.

And then you must be convinced that that opinion was wrong and change your mind, which is easier and harder to do, depending on the person, the experience level. I think people who are fresh to topics can change their mind a lot because they haven't had time to harden their heart for whatever it is that they're currently thinking of. So you and I have been in the industry

for a very long time now, did you know that? How long you been in the industry? Now he doesn't really define the industry, but I think we can just say the software world. I would say I think I entered officially in 2003 That was more on the services and hardware side of things data center side of things And it was more in the development when I say development. I mean business development side of things Yeah, but I was learning more about

technology servers firewalls virtual virtualization, etc. And then I began to build and stuff like that and so shortly thereafter I was actually into the software, not just the hardware and the biz. I'd say since 2003. So you're over 20 years. Yeah I think think so. Officially as of this year. 2023. No, it's 2025 now. So officially that's two years ago.

What year is this, man? Really? 22 years then, Jared. Jeez, I can do math. Thank you very much. Well, there was that one time vortex that happened. Some were between 2020 and now there was like a vortex. Is it really 2025?

Nobody really knows what time it is as the song goes. So I'm with you, maybe a couple of years behind. I graduated from college, I think in 06, 05 or 06. It's always very fuzzy. And went straight into industry pretty much from there. But I was already doing stuff on the web in college. Prior to college, my computer use was like basically Napster and video games.

Yeah, same. And when I amped and stuff, it wasn't really like productive creating things. I wouldn't say I was in the industry, but yeah, probably 20 years for me since it is 2025. And have you changed your mind at all? There's actually a lot that I don't think I have changed my mind on.

There's a lot that I have and a lot that I haven't, I would say. So yes and no. My list is not exhaustive, because I'm thinking like, gosh, what did I know or think I knew that I think now I don't know or know? And maybe it's changed.

You know, it's a lot of years to go through, to comb through really. Yeah. Well, we went through a little exercise, the two of us. We thought of a few things we have changed our mind on, things we haven't changed our mind on, which I think is also powerful. And Chris did the same thing.

So he starts off with a list of things he's changed his mind on, things he now believes which in the past he would have squabbled with himself. And then he also has a list of opinions he's picked up along the way. And then finally things he has not changed his opinion on. And so we've done similar, but I thought we talked through some of Chris's

because many of these things are agreeable to me. Other things I would need clarification on what exactly he means. I mean, it's a bulleted list, right? So this is how you go viral on software. Keep it simple, scannable, digestible, 10 minutes or less, 30 minutes if you wanna dig. Things I've learned after N years, make the topic too long, and then just list it up

and let the rest of us opine on it. So we're gonna start with a few of those. And where should we go first? Here's one that I thought was interesting. REPLs, this is tying now back into our discovery coding conversation with Jimmy. Chris thinks REPLs are not useful design tools, though they are useful exploratory tools.

That's a good one. To that I would say exploration is design, man. Let's go. Get your discovery coding hat on and explore your way to a design. I think REPLels are great design tools because you get to explore the space that you're trying to design for.

It's kind of like if you're trying to design the interior of a room and you're not allowed to explore that room first, you're not gonna be able to come up with a good design. But a repel is a way that you can explore the living system. It's like design, get a couch for this room. How big's the room?

Right. I can't tell you. Can I feel the walls? Yeah. No. So he's changed his mind to that. And I would like to differ. I think REPLs are useful design tools in so far as they are used exploratory.

Now, perhaps Chris thinks there's a different way to use a REPL in order to design things. And I'm not thinking of that, but that's the disadvantage of not having him here with us. And the advantage of being to assume what he thinks, because then we can just disagree with a straw man. Which perhaps- What does REPL stand for?

A REPL is a read, evaluate, print loop. And so it is an interactive environment where you can execute some code. You can read, well, you can read. Are you reading or is it reading? I never thought of it so definitionally. Maybe both. Yeah, so there's stuff you can read and it can read

and you can evaluate expressions right there and print out the results and then do it again. And so it just, the environment is modified as you run it. So if you declare a variable after the loop, you can then use the variable, et cetera. And so it's exploratory, it's interactive, it's pretty rad. It's where I do most of my design work when coding.

Oh gosh. Sometimes I will take things right out of the REPL and paste them into somewhere else, into a text file and... Call that coding. Yeah. That is coding, right? What is coding after all?

That's true. What is coding? Leslie Lamport, coding is typing. He does not like typing. He likes to think outside of the code. I think it's more than just typing. I've picked a net with him on that before, but he's way smarter than I am.

So I'll let him have that one. But I think coding is murkier today than it ever has been. Like, what does it mean to actually do a thing because everything's changing underneath us. Yeah, you really have to enjoy, I guess, typing characters into a machine. Right? Yeah, totally.

You have to think, obviously, you have to have some domain knowledge. You have to have thought through maybe a problem set, how to handle it, like middle out, something like that. Which is always the way you should. That's right. And then you gotta type it in there. Or have someone else type it for you,

like some sort of robot that you just dictate to. Yeah, that'd be nice. So here's Chris's other things he's changed his mind on. I feel like we should get Chris on with Jimmy and go and have them go at it. Because this one he says most programming should be done long before a single line of code is written. And here we have an outliner, right?

So it takes all kinds. I'm not against outliners. If this works for you, go for it. But there's also discovery coding. And I would have to understand what he means by most programming. Again, what does it mean to program? What does it mean to code?

I think Chris means you should think through things before you do things. I got no problem with thinking before you act. I just think that we think better while we're acting and while we're exploring. And so I tend to get to a line of code faster perhaps than he would, but younger him would be with me. Yeah, I don't know.

I don't fully, I don't so much disagree, but let me, it's like the spaghetti thing in your brain. I don't wanna walk around while I'm doing the dishes or cooking dinner with this spaghetti stuff in your brain, right? I don't wanna walk around while I'm doing the dishes or cooking dinner with this spaghetti stuff in my brain. Albeit, there are times when I have my biggest thoughts or biggest breakthroughs not in the moment of the action. Right, I mean, oh yeah, I didn't think about that. Let me do it like this next time

or when I get back to my terminal or whatever. In the shower? Yeah. When you're on a walk? Let or whatever. In the shower. Yeah. When you're on a walk. Let me try this new thing out. Right. So I guess if he's calling that programming,

yeah, I suppose. You know? Yeah, are we talking about sitting at the terminal and typing code into a text editor? Are we talking about actively trying to solve a problem of software? And I agree that you can absolutely, and should, and most of my good ideas come away from the keyboard.

But then I gotta come back to the keyboard and try stuff and find out. That idea kinda stinks actually. When I was walking down the road over there, it sounded like a brilliant idea. And then when I tried it, it wasn't so smart. So I think there's no... I would reword this. Most programmatic thinking or most thinking that you should do should

be done away from that keyboard long before you get to the keyboard to write it. I do think that once you're stuck you should step away. Yeah. And you can step away way faster than I usually do. I probably waste hours by not stepping away quicker. Yeah. And you can step away way faster than I usually do. I probably waste hours by not stepping away quicker. Yeah. If I stepped up. But you know the thing is, is like when you get stuck, doesn't stepping away feel like you're just giving up? Like it feels like you're admitting failure. And I don't want to be a failure. I want to succeed in life. And so I'm going to sit here and figure this out, dog nabbit. But then I do finally admit that I can't figure it out right now.

And I step away and I go take a shower, I go work out, or I go ride a bike or whatever it is. And then right there, when I wasn't thinking about it, pops in the actual other way I could do it, which is way similar and solves all my problems. That's a confounding part of the process. I would even say an energy renewal too, because there's times you get fatigued, you know. I wouldn't say necessarily physically, but maybe mentally.

You know, maybe a little both, but you step away, you get that cold water in your face if you're gonna wash your face kind of thing, you know. Girl, go wash your face. throw back to that good book title Rachel Hollis, you know go wash your face Get refreshed. Maybe you come back with some new energy. There's times I'm like trying to do something and You know, I've got the willpower. I Think I have the energy I'm making progress, but it just feels like I'm just like, I'm walking through three feet of snow, you know?

It's not even mud. It's like, imagine having to lift your leg up three feet just to get your foot out of the hole. Oh yeah, the worst. To go forward to put it into a brand new hole. It's gonna just sink right back down again. Yeah, I'm making progress. You know, I can see the finish line,

but I go away, wash my face, snow melts, oof, better world, you know? Well, some procrastination is really actually smart. The hard part is discovering what procrastination is smart and what is just lazy. I've been trying to find that out for a very long time. Cause there's times I'm like, I don't wanna do this thing until I absolutely have to do it

because that's when I'll spend the least amount of time on it. Right, under pressure. And I get it done in the moment, it's over 10 minutes versus an hour kind of thing. Some problems go away while you're not doing them. And you realize, I never had to solve that problem, it disappeared while I was procrastinating it. But other problems get bigger and hairier and worse

while you don't do them. And knowing the difference is wisdom, right? That's everything. Ooh, man. What is this, Jared? What is this you're throwing out here? All this wisdom. Just slapping us with the wisdom stick.

Well, I've been in the industry a long time, you know. I have a lot of failures under my belt. Which I can learn from. I learned something from Gerhard years ago, which wasn't really him teaching me something. It's almost like he put words to something that I already knew, but then I like, it reified to use a term.

And that was when I was visiting London for OSCON Europe or something like this. And I visited Gerhard for the first time. We saw each other IRL for the first time. It was cool. Went out to lunch. Hung out at the Pivotal office there. And played some ping pong.

And we were playing ping pong. And I don't know if they call it ping pong or table tennis over there. I can't recall. Gerhard could straighten us out on that. But he's very into it, and I like a good game of ping pong, just like anybody else, you know. For sure.

And we're playing with some of his colleagues who are on their break or whatever. And he said, do you know why we play ping pong while we're at work? And I was like, it's fun. It's fun. He's like, no, because we refresh our brains because when we're playing ping pong,

everything else in the world disappears for a few minutes and all we're doing is just playing ping pong and we can just put everything away just for a few minutes and then come back to it and your brain is refreshed. And I was like, that's right, that's true. I've known that, but he said it to me and I was like, hmm, that's a great way of looking at it because that's really what is happening.

And so useful distractions, usually physical, sometimes mental, but just in a different area. Competition, wash your face, as you said. I have a dart board over there and don't really use it anymore. But when I used to get stuck, I just throw darts. That helps. Yeah.

Take a walk, take a shower. That's so smart. I agree with that sentiment. It is like a different take on step away to get unstuck, but what you're focusing on is what the activity is when you're stepped away, you know? Because if you have to put blinders to everything else, you obviously have to focus and someone say laser focus.

And you can't think about that problem anymore. You literally have to put it down and put it away completely to focus on the collaborative process of ping pong or table tennis or washing your face or whatever the activity is that lets you kind of sort of put blinders on everything else, all the problems. Yeah, and there's something like a childlike wonder that comes out during games of competition.

Even, it doesn't have to be competitive, like even I'm sure when you go mountain biking, where it's like you're just being free and you're being a kid and you're just doing something that has exhilaration and takes concentration. And if you don't pay attention, you're gonna get hurt, so you better take it seriously. And so while you're taking it seriously,

nothing else can really fill your mind. That's the hard part about actually stepping away to get unstuck is you can step away, but you can't necessarily turn your brain off. The problem, unless you force it out you can't necessarily turn your brain off, the problem, unless you force it out by doing something that requires your brain and your body lots of times. And then you're actually stepping away.

And so I think it does force us to turn our brain off the problem, which gives us that chance to refresh. So big fan of ping pong or whatever your game of choice is. Food is all a good one. So good. Yeah.

All right, moving on. Anything else that Chris has said here that you agree with or take issue with? There is no pride, this is Chris talking, there is no pride in managing or understanding complexity. That one's probably deep for him, right? That's pretty deep, yeah. Because I can think about some ideas

where that applies to, but maybe his context's pretty deep. Yeah. Cause I can think about some ideas where that applies to, but you know, maybe his contact is pretty particular. Pride. Somebody must've been prideful around him, around managing some complexity. My guess is it was from his younger self. Like he used to take pride in it and now he doesn't. Yeah, I guess so.

Because these are things that he's changed his mind about. I would love to ask him questions about that. Cause all I have is questions. Oh, we gotta get him on the show. We can't just talk about his blog about. I would love to ask him questions about that because all I have is questions really. Oh yeah, get him on the show. He can't just talk about his blog post. We gotta talk to him. Okay. Let's switch bases then and go to second base,

which is where we always go after first base. Let's get out and talk about our own mind changes. Since here we can talk to each other. We don't have to guesstimate what we mean. You haven't changed your mind about much. I tried to code my brain for changes. Couldn't find them. Not a lot of change, honestly.

I mean, it's been a lot of the same for a decade and a half, basically. It's not a lot of change here. Do you wanna start with things that you haven't changed your mind about though? Like things that you believe in the industry? I think that would be easier for me. Okay.

I've always thought front end slash CSS is hard. And I still believe front end slash CSS is still hard. You may be more right now. Okay, let me caveat that. CSS is easier than it's ever been. It's still hard though. Just as a standalone technology. Except for that it's overwhelming now

because there's so much you can do. Yes. But it's never been. Just as a standalone technology. Except for that it's overwhelming now because there's so much you can do. But it's never been easier to accomplish stuff with CSS. However, front end, which is much bigger than just that, I think, and I've been here a long time, has never been more complicated than it is today. And so I think you might be more right about that than you were in the past. Yeah bro, I mean, oh I did conflate them,

I will agree with that Okay, so let's maybe just front end over there CSS over here. Okay, I will agree that CSS has Gotten easier and I will say that they've it's become easier potentially because of LLMs I think because you know, I'm not a daily driver on front-end building I'm not building front-end sites every day. So I't, I don't have the muscle memory every day of like how I lay something out with CSS or like I did back in the day, like back in the day I had, I really thought at one point in Jared that I could design and build an entire front end without ever looking at the browser. Like you just write the code and then call it done at some point.

You know, command R. You're like Cypher staring into the matrix. Yeah, I really felt like that. And you know, I don't think I've ever actually given myself the true try on that, but I think I may have done simple things like menus and roughly, it's not like a full on design, but like enough to be like-

So you just go from like blank page. Nothing, yes. To like writing all the HTML and the CSS. That's right. And then you just load the browser once and it looks like. Yes. I would have loved to put that to the test

back in your golden years. My golden years. Or the good old days, whatever it's called. Golden years is when you're older, isn't it? And now I don't think I could ever do that. It today is, Maybe, maybe. I mean, I'm just, my time between progress in CSS to other tasks and then back to it again

is too far in between for me to maintain that muscle memory well enough, I think. Like I used to be able to do when I was younger. Or maybe just had less things in my brain at once So I really feel like CSS is still really really hard like even with tailwind and I'm I'm pro utility classes I'm pro tailwind. I know some people are like absolute haters of it. Mm-hmm. Whatever, you know Purists. Yeah purists. Why say why not both, right? That's my favorite thing. You know?

Tailwind in the HTML, and then tailwind applies in the CSS. Why not do both? Tailwind's really good and easy to do. I mean, it's, the add apply is cool, and you can multi-line that versus like one single line. It gets a little easier. You can still define rules and stuff like that. You can include, you know,

I don't know what you call them anymore. I'm like nomenclatured out on CSS. Like you can make a class and include that class later through an apply. It's cool. You know, no more SaaS required for this. Just use Tailwind and its utilities to build your CSS. So cool. I'm not anti Tailwind at all. to build your CSS. So cool.

I'm not anti-Tailwind at all. I think that most of the little things that I build are scoped in such a way that, for me, it feels like overkill because all I need to do is ask an LLM how to get the layout I want with Flexbox or with CSS Grid, and then I can just toggle stuff. I really like the DevTools ability to change the flow and the direction of Flexbox things,

and I can just click which one I want and wait till it looks the way. It's on the opposite of you, I gotta have it in front of me looking at it. And I can just kinda twiddle bits around until I get to something that I'm happy with, and I think that it's never been easier to do that than it used to be.

But I do think on many, many projects for especially teams like you and I working together, Tailwind makes way more sense because we can just throw the utility classes in there and not have two different styles and mess, you know, like have to like synchronize the way we build things out. We just are like, use the right classes

and everything's fine. That's right, there's a way. There's a way. Yeah, I think there's a lot of good stuff in there. That's why I was so, in that conversation with Chris Coyer, I was so pro what CodePen could do because I feel like, where else can you play with HTML and CSS in particular?

And sure, you can sprinkle some JS in there if you want to. But primarily it's, you know, HTML and CSS. Where else can you do that in a way where you can show it to other people, create little things, and show off, than in CodePen, right? That is the coolest. Being able to riff like that, basically,

and show off a little bit, that to me is super cool. And I don't even know if they've, I haven't played with it in a while, but you know I'm pretty sure they have Tailwind baked in. I don't know, I would imagine they would. Chris is smart. But being able to play with that kind of stuff in the browser where you can show off to somebody and just let it be a toy so to speak or just a stab or a spike, I think you've said before. Mm-hmm. That to me is kind of cool. I also like a lot of the stuff around Tailwind

and you can define your colors and stuff like that. That to me in your Tailwind config is just the coolest. Being able to do all that like that, and you can define where you would have to do that CSS before with, in the root with variables, or not at all back before variables weren't even there. I think that's cool that you can define all these color palettes and stuff like that, and it's built into the way Tailwind works at a structural level, not just classes and utilities and stuff like that. It's part of the overarching nicety framework that it brings.

That to me, cool, but it's still hard. Yeah. I think it might always be hard, Jared. What do you think? Well, I mean, 20 years later, it's still hard for you. So you expect that to change at this point? I guess not. For some, it's easy.

Well, I think if you did it more often, I think some of it is because of your proximity to the technology not being on all the time. There are certainly people who are writing front end and specifically doing the CSS 40 hours a week, 60 hours a week. And yeah. Can you imagine? You do anything that much.

No. No. You do anything that much and you're gonna get, it's gonna become easy to you. So that's funny, I used to be like that. I would work all day and all night on nothing but that all the time. That's why I guess smaller problems, less complexity.

Well, then it's always gonna be hard until you commit. What about you? What has not changed? What has not changed? Is that what we're answering here? Is what's not changed? Right now we're doing things we haven't changed our mind about.

That's right, things we haven't changed our mind about. Thank you. So there's a few things I've been preaching for years and years and years and years, and I don't think I'll ever change my mind on them. And one of those is slow down to go faster. I have not changed my mind. I still think if you wanna go fast, you have to slow down.

And it sounds like a contradiction, but it's not. Because as you go slow, over the course of time, you end up going faster because you don't have to redo, rewrite, start over, or backtrack as often as if you were moving too fast in the first place. As Kelsey Hightower said it, let's see if I can remember his saying it was really cool as everything Kelsey says is cool. I think he said, do it right, do it light, do it wrong, do it long.

I tell that to my kids sometimes. And they're like, what's that mean? I'm like, I don't know, but it sounds cool, doesn't it? I don't know. Listen to this podcast, you'll get it. Yeah, you do it right, you do it light. You just slow down, you go faster. You know, it's the old tortoise in the air.

And so much of the tech industry is focused on raw and utter speed. And we end up foot gunning, as they say, shooting ourselves in the foot repeatedly or shooting our teammates in the foot or the future developer who's replacing us because I've spent my 12 months and I'm moving on. And we end up churning and burning a lot of our colleagues

because we're so focused on speed. And if you would just slow down and take the break or write that additional test or look at the source code or don't abstract that method yet, or insert your way of slowing down here, that those small decisions over time lead to faster. And I still believe that today. And I think I wrote that blog post like 15 years ago.

So yeah, that's one. Is that similar or the same as keep it simple? Or is like keep it simple in there of slowing down? It's definitely not the same as keep it simple. It's in there. Because you're saying don't do that abstraction now, which is kind of- Yeah, that was one example of slowing down.

Right. And in fact, many people, this leads me to another one about dry, which we can get to in a minute, but many people think that not doing the abstraction makes it less simple. But I just, I mean, I say disagree with that, which I used to not disagree with that.

I don't know, simplicity is a really tough one. Obviously it's something to strive for. Subjective for sure. But there are complicated things in the world. This is kind of where I was getting with, I can't remember who's on the show, we were talking about simplicity recently. Maybe it was Burt, I think it was.

Cause he was all about simplicity, right? Less dependencies, less abstractions, more simplicity, straightforward. And I was talking about how I think a lot of people desire that in their code. And he said, well, a lot of people like clever, complex things. And going back to Chris's point about,

don't have an ego about being able to manage complexity in your head. That's kind of a young person's concept, I think, where it's like, look how brilliant I am, because I can manage all this complexity in my head and So Burt was saying well most of us don't really want simplicity. We actually want complexity because that makes us smarter. I Understand that I think many of us do desire simple solutions though And yet we still find ourselves with complex ones over time

And so there's definitely some slowdown to go faster than that, like slowdown to keep it simple, which is incredibly difficult. And I still fail at, you know, after all these years. I think I'm being simple and then I'm actually realizing, oh, it's too simple because I wanted to keep it simple. And it doesn't handle all the things it needs to handle. Because like I said, to Bert, the real world is complicated.

And some systems cannot be simple. I mean, look at our tax code, just as for one instance. You think TurboTax can be simple software? I mean, it just can't be because it's handling a very complicated thing. Now it can be more or less simple in its approach, but you're not gonna be able to file some of these taxes. I mean, not just one person's taxes. Arbitrary people's taxes in any given year

without some serious complexity under the hood. They're just not gonna happen. And so the real world is jagged and backwards often and changing, you know, look at time zones. The real world is jagged and backwards often and changing, you know, look at time zones, political. I mean, think how much software is changing right now because of Doge, for instance. How much tech that has been wiped away or accrued. Or created.

Yeah, exactly. As there are, you know, machinations in the political. We're getting ready to ship this. Just cancel it all. Yeah. Yeah, how many of us? That code doesn't matter anymore? I mean, I've been,

I've built things that never shipped and it wasn't a software reason. It was like just business or change of mind or, you know. Yeah. It's like, that's not gonna ship. It's like, oh wow. And I'm sure our listeners are just like, everybody's lived through that project

that like you put your heart into and you're proud of it technically. And then it's just like, yeah, it's never gonna say the light of day. Yeah. And then you can think, well, at least I got paid well. Hopefully you can think that. That's the only thing I fall back on.

Or learn. Yeah, exactly. You try to get like silver linings out of it. Yeah. So yeah, simplicity is really tough. Slow down to go faster. I never really understood that. It always seemed like, you know,

obviously a conundrum of sorts or a paradox essentially. Like what? No, that doesn't make sense. Well, it goes back to one of your favorite sayings, slow and steady wins the race. Preach, bro, preach. I mean, you know, there's times, you know, I even have to re-explain that to myself because I try to go too fast, slow down and check yourself.

Right. You know, and I think I always tell people when we say or someone says, I think most people mean it this way. I know we do Slow and steady doesn't mean you literally go slow. I think even here slow down doesn't mean don't go fast. It means Slow and steady means going at a speed the fastest speed possible to remain steady Right, you know your team may be able to achieve a much greater velocity Or yourself may be able to then prior versions of you

By not also going slow. You're still sort of going fast, but you're steady Whatever that is to go forward on this thing on this mission Whatever it might be in a steady manner as fast you can go while being steady Mm-hmm. yeah, well said. And that's how you win. Yeah, exactly, and especially in a digital landscape that can change dynamically as we build it. Like we are building things that are completely malleable

and we can completely destroy our creations by just making that rash decision, you know, going through that one way door and not realizing it was a one way. And then you get out over your skis, if I might just use like all the cliches, and it's too late, like you're gonna land on your face. And if you, and now you're way backwards, right? Now you're just playing catch up.

And so many of our projects are playing catch up with a timeline because we set the timeline too drastically and so we cut corners and made bad decisions and moved too fast. And then that thing's never gonna ship because of all the tech that we've accrued. And it's like, whew, if we would have slowed down and checked ourselves, we'd actually be moving faster

in the long run, so. There you go, good summary. Okay, that's one of mine. Slow down and go faster haven't changed my mind. I did find the original blog post, it was 2010. So yes, I wrote that 15 years ago. Wow, man. What was it titled?

Slow down and go faster. There you go. So good. What makes me mad is I think, I don't think I, I'm not sure if I coined that, but I'm definitely one of the only people that have written that or like probably the first one. And if you Google it, like other jokesters show up first.

That's like, come on people. Why does McKinsey and company? This guy, Jared Santos got some good ideas. Let's just take his title. Who are these McKinsey people? And then somebody on Medium. It's like, come on, yeah, that was written in 2023. Anyways, ideas are cheap.

So I don't deserve much for that one, and I don't think I invented it either. But when you Google the exact phrase you think that my post would show up at. I'm sorry, man. It's okay. I'm sorry. I'll backlink you, bro.

I'll get you out there. Yeah, yeah, we're already doing everything we can. So well friends, I am here with a new friend of mine, Scott Deaton, CEO of Augment code. I'm excited about this. Augment taps into your team's collective knowledge, your code base, your documentation, your dependencies. It is the most context aware developer AI, so you won't just code faster, you also build smarter. It's an ask me anything for your code. It's your deep thinking buddy. It's your stand flow antidote. Okay, Scott. So for the foreseeable future, AI assisted is here to stay. It's just a matter of getting the AI to be a better assistant. And in particular, I want help on the thinking part,

not necessarily the coding part. Can you speak to the thinking problem versus the coding problem and the potential false dichotomy there? A couple of different points to make. AIs have gotten good at making incremental changes, at least when they understand customer software. So first and the biggest limitation

that these AIs have today, they really don't understand anything about your code base. If you take GitHub Copilot for example, it's like a fresh college graduate, understands some programming languages and algorithms, but doesn't understand what you're trying to do. And as a result of that, something like two-thirds of the community on average drops off of the product, especially the expert developers. Augment is different. We use retrieval augmented generation to deeply mine the knowledge that's inherent

inside your code base. So we are a co-pilot that is an expert and they can help you navigate the code base, help you find issues and fix them and resolve them over time much more quickly than you can trying to tutor up a novice on your software. So you're often compared to GitHub Copilot. I can imagine that you have a hot take.

What's your hot take on GitHub Copilot? I think it was a great 1.0 product, and I think they've done a huge service in promoting AI, but I think the game has changed. We have moved from AIs that are new college graduates to in effect AIs that are now among the best developers in your code base. And that difference is a profound one for software engineering in particular.

If you're writing a new application from scratch, you want a webpage that'll play tic-tac-toe, piece of cake to crank that out. But if you're looking at a tens of millions of line code base, like many of our customers, Lemonade is one of them. I mean, 10 million line mono repo, as they move engineers inside and around that code base and hire new engineers,

just the workload on senior developers to mentor people into areas of the code base they're not familiar with is hugely painful. An AI that knows the answer and is available seven by 24, you don't have to interrupt anybody and can help coach you through whatever you're trying to work on, is hugely empowering to an engineer

working on unfamiliar code. Very cool. Well friends, Augment Code is developer AI that uses deep understanding of your large code base and how you build software to deliver personalized code suggestions and insights. A good next step is to go to augmentcode.com. That's A-U-G-M-E-N-T-C-O-D-E.com.

Request a free trial, contact sales, or if you're an open source project, Augment is free to you to use. Learn more at augmentcode.com. That's A-U-G-M-E-N-T-C-O-D-E.com. Augmentcode.com. What else? You got another one? I think collaborating is still hard. You know, I haven't, it's always been hard.

Collaborating has always been hard, especially maybe face to face. It hasn't been that hard to collaborate with, You know, I haven't it's always been hard collaborating has always been hard especially Maybe face to face hasn't been that hard to collaborate with but I feel like as a remote team True in the groove. What is the mission? Collaboration is hard Tools change over time then you got politics over a tool like I've had them float on notion They're one of our sponsors as you may know Jerry but I I had them back as a sponsor because I'm using notion all the time, right and

For a while there We only use notion for one single thing because we tried to collaborate around certain things that just never got in a groove And I think it's mainly a tool problem not a not a consolidation or compartmentalization problem because is is the notion is notion the best place to write was the question we had and the answer is probably not like I've used obsidian and now I'm kind of like in between obsidian and notion and I don't know which one to reach for when I got to take a note anymore I was like just drives me crazy so like okay operation though or that's just for your own use?

It's for me, it's keeping my, so there's certain workflows that I have, like sponsors and campaigns. There's a lot of details involved. And I'm just like, I need to share these things. Google Docs work, but then they were too arduous, where the template wasn't that great, and then it became formatting.

And so Notion just sort of simplify some of those things And so now I'm back in this notion world because it can let me build Dashboards and workflows and operating systems basically right and it didn't have to be shared It can just be me. This is how I work and it's my environment and this is you know what works Let's just say I can't do that kind of stuff inside a docs know what works let's just say I can't do that kind of stuff inside a Docs inside of obsidian at least not that well and so notion kind of won a couple battles in terms of workflows and then I can share the things too so now I find myself in notion more all that to say is that collaborating with a team is so

hard you got Zulip you got slack you got discord you got places to go and you got opinions everywhere of like where can we talk in the real time? Where can we talk in the async? Oh, you don't like that toll gosh you have an issue with that You don't like Jira man who likes Jira. I like Jira. I met one dude is a good friend of mine He loves Jira really loves Jira because it's so powerful. Not because he like literally loves Jira. It's like, this is the best tool for an enterprise like ours. It's so flexible. It's so sturdy. It's got so many APIs and so many ways in and out of it

that you could just get things done. So many integrations. I'm like, great. Those are great reasons to love it. Those are the reasons I don't love it because I don't have those needs, you know? Right. So I feel like collaboration is just so fractured bifurcated

Fragmented that it's just like which tool where to go and then the by time by the time you get in there You're like well now we're paying 200 bucks a month if we're a small teams like hundreds of dollars a month just For like one person to do all the collaborating or a couple in a small team. I just feel like collaborating is hard. And I wish, I don't know if software can solve it because it's tried again and again and again, you know? And I feel like it's just like this drum that never gets beaten hard or good enough.

Let me hop in here. Yeah, please. Because you just said a key phrase that lines up with something I have changed my mind about. Okay. Which now I believe changed my mind about. Okay. Which now I believe and I used to not believe. Software can't solve people problems. I used to think you could just throw software darn near anything and I used to be like,

let's do this. I solve problems with software. I'm gonna solve your problems with software. Let's make some software. Yeah. And now after years of trying to do that, and sometimes it works and most of the times it doesn't, software can solve some problems,

but it can't solve people problems. And collaboration ultimately is a people problem. Now, can it help? Yes. But can it actually solve it? Can you find the one true tool that's gonna finally make us collaborate better? I just don't think you can do it.

I mean, a lot of what you're talking about here is like, people don't wanna do that. I mean, I know that I'm not a fan of this new idea to go back to Notion because we've done it before. And that's why I haven't asked you to do it. Yeah, and I'm like, well, I thought I was using Obsidian now and I'm happy with Obsidian and most of my stuff doesn't need to be shared and when I do,

it's a markdown file, so maybe I can, but I'm not dealing with third parties as much as you are, like external entities, like our customers. And so I don't have to like make it look nice. It can be, you know, H1s and H2s ultimately. And so for me, Notion is cumbersome and too much, just like Jira is for some people.

And if I can avoid it, I would love to. Because like you said, it's one more place to go. Did I put that in Notion? Did I put that in Sydney? Right now I'm living in one world, I'm happy for my notes at least. And that's a people problem. Like you have to convince me, I guess,

that it's worth it. Or I have to just be like, all right, I'll try it for you and then we'll see where it goes. And we've done this with lots of tools over the years. Trello, Notion. did we do Pivotal? Did we do Pivotal Tracker, me and you? Probably not. It was early days, just you and Gerhard.

I was in there a little bit. It was like early, early, 2016, early days. GitHub projects. Yeah. I'm just thinking off the top of my head. I'm sure there's others. And the problem is like everybody thinks differently and we think differently at different times.

Like that's the real people problems. Like my tastes have changed over the years too. To where like what I used to think was good, I don't think is good anymore. And so I have to even convince myself to stay with a tool, even if I'm not collaborating, you know? I'm like, you know what? This used to be cool, but now I'm just thinking, eh, not cool.

And then later on, I think, you know what? Trello was pretty cool. Maybe we should use Trello again. And I go back to it. I'm like, oh no, Atlassian ruined it by putting a bunch of other crap on top. Oh yeah. I'm not sure if that's true or not, but it is true. So true.

But their single sign on was such a mess. Like they consolidated things and now I have an Atlassian account that I don't care because all I use is Trello, blah, blah, blah, blah. Yeah, that's just one dynamic of people problems. You know, it's like we can't actually solve people problems with software. We have to solve people problems with people skills. And that sucks.

I can't care with that. That sucks. My resistance, I think, with that that sucks my resistance. I think With things like notion. This is not a notion problem. This is not even their fault It is my resistance is the lock-in That happens and I'm thinking okay great Let's create some workflows and we do and we start to collaborate you and I even just two people and we start to use you and I even just two people and we start

to use their commenting feature and maybe they launch a real time chat that sort of supplants some of the things we talk about in Zulip or DMs and so now we got two places we talk in real time. Well now our productivity is reliant on that particular feature set, but then the overarching tool begins to maybe not be so good or something else revolutionizes something. It's not even that it changed or got worse because they're poor developers or poor leaders.

It's just that the puck moved and that happens in tooling. That's my resistance. It's like, gosh, man, I'm just tired of, and maybe I'm just crying here, okay? I'm sorry if I'm crying here. 22 years deep, apparently, in this industry. It's time to cry. You get some scars and some calluses.

Is I am just tired of moving from tool to tool to tool over all these years. Now, I will agree, and by no means is this sponsored, they are a sponsor of ours, but this is my own true sentiment. I do think that right now, Notion is one of the best places.

It's on all the platforms. It's got some AI stuff in it. I've found that Notion AI pretty useful to summarize things and help me find things in a Massive workspace or notes like rather than search in this one weird way. I just got a notion I just search for these things so I feel like they're onto something and I really hope that they can kind of keep iterating because every year they kind of launch a

New version of it and it's gotten better and I was even surprised when I went back to it after how, how much we really, I don't want to say hate it, but we're like after that year of having all of our sponsorships in there and it just suck. And I feel like our sales suck because I didn't enjoy how it was organized anymore. So hard. It was just like every day was just like, uh, was just hard to manage our schedule. What was available to sell? Where was it at? Who had the spot, it was just hard. And so we were using probably the wrong tool for the right job, the right job we needed to get done,

but that was the wrong tool. And ultimately Google Spreadsheet was just better at it. It was just a simpler tool, so slow down, keep it simple maybe. And I- And then the other golden rule, which is like most of the time a spreadsheet's best. That's right.

Yeah, when in doubt, spreadsheet it, you know. Seriously, start with a spreadsheet. Start with a spreadsheet. And probably stay there unless there's a real reason to move out. I know you've prompted me many times to write us some custom stuff. I still want you to.

I know you do. But I resist it, because I'm like, I don't, the problem changes too frequently. And so every time you would make the software great, I think I would be like, it's gotta change. You're like, dude, this is a complete architecture change. This is a whole new feature set. And I'm like, you know how much work I had to put in

to get anywhere near what you have with Google Sheets right now? Let's just keep using that, because it's so flexible. It is. And you're gonna hate whatever I build until I dump years into it. And then what, what have you been building a backend sales system for years?

So I've been very resistant. As you know, I'm a tough sell, especially when it comes to writing code. I'd rather have no code. That's the best. No code is the best. Some of this lock-in fatigue that you have is one of the reasons why I, to this day,

and maybe won't last, but I still love Obsidian because its philosophy of file over app takes me to happy places. I just love the fact that at the end of the day, I'm writing markdown files. And I can take those anywhere and they are just store on a hard drive and I can sync them with Dropbox. And if Obsidian disappeared, I would be just fine.

I have plenty of other Markdown editors including Zed, which I write code in all the day long, weird sentence. And I love that. And Stephango, the creator of Obsidian, has this whole philosophy written out. It's called File Over App that he wrote. I'd love to get him on the show, by the way, Steph. If you listen, please come on the show.

We are fans. We'd love to talk to you. We'll get you an official invite to share your email address. We'll find you. We'll get you. That's right. And if you listen and you know Steph,

and you can help Steph know that we'd like to have him on the show, please hook us up. Because very interesting person, lots of that stuff I agree with. I'll link up in the show notes, his post called File Over App, it's a philosophy. He says, if you wanna create digital artifacts that last,

they must be files you can control in formats that are easy to retrieve and read. Use tools that give you this freedom. I like that and I like Obsidian for that reason because at the end of the day, it's just an awesome, I think, layer on top of files. And that gives us freedom and lack of blocking. That's why I love it so, I mean, don't get me wrong.

My recipes are still in Obsidian. All of my, you know, building an AI machine, creator PC, Ubuntu machine, like all of my personal docs I've created to like tell future Adam what old Adam learned, you know, is still and will be in obsidian. It's the things that I think and so this file over app scenario, there's there's certain workflows that I can see how I could possibly use it in Obsidian, but it's a personal thing. I can't share that easily with a group

and get them to just jump into a way. You know, they have to have certain plugins and community things and maybe it's a, something's gonna be in the way in Obsidian and it's gonna go beyond file because the problem set that I'm trying to communicate, the communication is beyond a simple file. as a single person or an Individual with my own things. Yeah

It works, you know it and it's so fast. I love obsidian Amazingly amazingly. It's so good, but to collaborate with a group of people around a workflow Notions better That's totally fair, that's totally fair. But lock-in sucks, man. I'm so close to asking you to come back in Notion though. I'm waiting to feel really good about it by myself though. There's a couple things I think would be kinda cool.

A couple things we do, not every collaboration, but a couple key flows that require that ebb and that flow, that yin and that yang to get done right. Let's talk about something I haven't changed my mind about. Okay. Because I know this one's gonna strike deep into the heart of many people. I've actually shared this in the past

as an unpopular opinion, and I think it was unpopular. So I mostly share this when I'm making fun of Nick Nisi because it was one of my favorite pastimes. Shout out to Nick. By the way, Friendly Feud coming soon. The cast of JS Party will be there. Thank you to everybody who took our survey. We've got a huge response.

We got way more responses than I wanted. So I'll have to comb through all those. I wanted 100, but I forgot to turn the type form thing on where you can like close it after N responses. And so we've got way more than 100, which is great. It's just more than we needed. So happy, happy that you all support us and that will be recorded early March.

And so it's gonna be coming to ChangeDog and Friends here soon. Quick update on Nick Nisi, but what I believe now, which I used to not believe, especially as a younger person with more time on my hands, is that customization and tweaking of your environment is the root of all yakshaves. I think we waste way too much time customizing,

tweaking, fiddling, changing our programming environments than we need to. I think you can get 80% of the way there with 20% of the effort and then knock it off and get some work done. Stop customizing. You're just shaving a yak, I know it feels good, but you're not doing the main thing.

You gotta keep the main thing the main thing. And I used to customize to the Hill. My.files were like pristine and I checked them in and I changed them and I loved all that stuff. And I went searching for other people's themes and I was like, you know the people who have like a comment font and then a code font and then like dark mode and light mode and they can toggle it

based on you know where they are in the world or whatever it is you know. And they've just got all that stuff figured out. It's like you don't need to do all that, you can just write code and be happy. What do you think about that, Adam? You know, there's, um, it makes me think about H-top. Vanilla H-top is not good enough for me. You know, I always have to tweak it a little bit. I always have to tweak a little bit and so I think that drastic customization is

Not good, but I think subtle Customization is good 80-20 rule right I think I'm not sitting there tweaking everything I can pot I'm just thinking of each top because like literally every machine that I command I've never I've never like launched it and thought this isn't good enough, but maybe I just don't know. Oof, well, then you should see my H-top. I would love to. I can even customize mine. I can hold on to brew install H-top.

That's how long it's been since I've used it. Let's see here. I will log into. So on my local Mac, I use iStat menus, which is rad, and provides me cool charts and stuff without having to be the terminal. Yeah, I just launched H-top. It shows me, you know,

that Brave is using all my CPU right now. And that I have nine processors and I'm using 26 gigs of RAM out of my 64. What am I missing here? I've been up for 12 days. Okay. What's yours look like? I just share it with you.

I mean, I don't know how to do this in our new world of video first, but I share with you in Zulip. We'll get it on the screen for the pod on YouTube later. But this is my H-top man. Okay, let me describe this verbally. Host name, Cineplex. Oh, can I, can I dox you?

Yeah, you can, that's cool. I just host name doxed you. Host name, Cineplex. Uptime, 16 hours? Is that days? Hours. 16 colon 11 colon 18. That's hours, because I think I downed it

and then I pulled the latest Docker images and compose up it. Now I would argue that my stock H-top is better than that because mine says uptime 12 days and yours lacks information, but I guess it looks cooler. And then it shows the standard top things of the average memory and swap usage. All of your processors are listed

along the right hand side. You have 16 processors, so that's kinda cool. Whereas the stock one lists them horizontally, I guess. And zero. And so you've rearranged the order of things. And I think maybe your colors are more vibrant. I don't know, what'd you tweak here? Mainly layout.

Okay. So host name is there, uptime is there, and there's a blank space in between those two sections. Oh, let's see, you got two columns, left and right. On the right is all the CPU cores. Right. On the left is information essentially. I like the system, host name, uptime,

and then a blank space in between to separate average CPU, the memory usage, and any swap usage. And notice that there's zero case swap. Yeah, my system's tuned, okay? I also have zero case swap. There's no swap going on here. We're gonna brag about it? Then another, yeah, we gotta brag about no swap, bro.

I mean, that's how it goes. That's right, no swap over here. The game is no swapping. That's right. Right? Then you have the blank the game is no swapping. That's right. Right? Then you have the blank space in between, then you have network and disk. Now if it's a ZFS machine and I'm running like ZFS pools,

I'm gonna have arc there and things like that to tell me about my ZFS pools. Okay, so I'm missing that section. I don't have network and disk. Okay, so you just add it, F2, and you go into your settings there and you can add it to your left. You push enter on the one on the far right,

and you arrow left, arrow left, and you put it wherever you want. And then you push enter again and you put network and disk wherever you'd like. And so- Okay, now, do you store that into like a.rc file or something? So here's the thing,

and this goes back to your principle principle of customization. Yeah to my knowledge H-top and maybe B top is better. Maybe this is where B top is better. I don't know. I like H top is as simple H-top does not have a configuration file that can translate from machine to machine for whatever reason it's a unique snowflake The layout simple is the same on all these machines, and I would go and tweak it for this one little thing. It takes me just enough time, maybe a minute or two to do this, you know, not much. But I would love it if it was a config file

that I could just paste. But no, that's not how H-top works. But H-top is simple, B-top is kind of loud and expressive and vibrant and visual. But H-top is simple, B-top is kind of loud and expressive and vibrant and visual. I think B-tops, some would say it's better. H-top to me is just simple.

So that's me. All I have to say is I like a little. I'm flabbergasted honestly. I don't want to call your expertise into question, but I just feel like there has to be a way to get a config file that's just- Please, make me wrong, I would love it, because you'll solve a big problem of mine.

The last time I checked into it, in the H-top docs, it was a unique, it wasn't a config file that was, it was like, it didn't even make sense to config file. There is one for it, but it doesn't make sense. Yeah. When you look at the format of it, it's machine to machine and it's unique. I think it's because it's like,

basing on sensors and stuff like that, because you can have like CPU temperatures in there, like per core. And like if you have LM sensors installed, you can get a bunch of stuff. We're in the weeds. Yeah. We're in the weeds. I think that, and I'm over here trying to prove you wrong. I think that you are get a bunch of stuff. We're in the weeds. Yeah. We're in the weeds.

No, I think that, and I'm over here trying to prove you wrong. I think that you are right and wrong. I think you're mostly right. I'm mostly right. Technically you're wrong, cause there is a config file. However, it's written by H-top and overwritten

anytime you hit F2. See? And it's not human friendly. Exactly. So it's not cool. Like you could figure it out. How cool would it be if it was human friendly? Yeah, I mean.

Come on. Who makes a config file? Like, you know what, human friendly, not necessary on this one, suckers. Friends the computers only, this config file. Yeah, this is computer friendly. That's just dumb, whatever that is, I don't get it. I think if all your friends are computers,

and you don't have any human friends. Every machine, Jared, is a version of this layout. Unless it's got ZFS on it that I'm adding arc and a couple other niceties. That's like particularly to that machine But it's always all the cores on the right and the left is the information every single machine So you hop in there and have to every time like once machine and it's once and done for its lifetime I'm not creating new machines all the time. I will say that Cineplex is a new machine for me I moved Plex off of

Proxmox. It was a VM in Proxmox So get this it was Proxmox hypervisor a burn to VM Docker Contain, you know a Docker application running Plex, okay. I moved it onto a new machine standalone Intel not by itself a Buntu 240 for installed Docker again

And it's running in Docker there and so now Cineplex It's the new hotness Plex was the old machine now Cineplex is the new machine I just like the name Cineplex. It's kind of cool. Cineplex is cool. It is kind of cool. So. And so this is running your Plex server. Yes. This is running my Plex server. And does Docker make that easier?

Because couldn't you just like advocate and solve Plex and be done like on top of Ubuntu? Yes. Except for I believe this has been my hypothesis that it's easier because this this application has moved from machine to machine to machine lots I feel like it would just be easier to our sync that Docker directory which is like, you know My home directory slash plex and all of it lives in there the compose file all of the data lives in there

I don't even use volumes with it That compose file and then every any data that goes with it, it's about 100 gigs, this application directory. I can just R sync it to a new machine, docker compose up dash D and it's running. So I feel like Docker makes the application composable wherever it needs to be, you know, versus like, okay, I'm app gating and installing this thing

and I've got like, it's tied to the machine more. I can leave it tied to Docker and let Docker be the runtime and no matter where Docker's at, I can run the application. Provided the hardware can support it, of course, you know, but that would be pretty easy. So that's been my, that's been my reasoning for Docker. It didn't make sense to Proxmox, VM, then Docker, you know, so I just took out a layer and got rid of Proxmox and maintain a standalone Ubuntu machine. And it's called Cineplex. There you go.

And it's got a cool H-top config. Just enough customization is that some concurrent with this. A long winded. I forgot what we were talking about. Yes, long winded. I think I'm with you I think that people I've seen this and and it's It's tantalizing you got somebody I just was watching this video this really great cinematographer

Also a developer tool kind of fella. He's got his perfect desk, right? Perfect desk. Mm-hmm. It's got this Super awesome keyboard. He built it himself. He's got these keys that don't need lube because they come pre-lubed. He can KVM between his window machines so he can game. And he can KVM between that and his Mac. And I'm like, oh dude, I just don't have that time. I just wish I could. I would love that.

I don't want to say it's a young person's game but I feel like it is because you and I are so far into our careers that we've got either more problems to deal with and we can't spend the time there or We've now let go of things like that. Then we realize this prince we have shared is truly true that overly customizing your environment is Kind of a fool's errand. What did you say it was? Is this not smart? I called it the root of all yak shaving.

Root of all yak shaving. That sounded cool. That's not cool. I mean, it is a yak shave. And, you know, yak shaving's unnecessary by definition. Well, depends on how you think about it. Sometimes a yak shave is a necessary series of events in order to do what you originally wanted to do.

But other times you're just out here shaving yaks, and you're not actually doing the main thing. That's more I'm using it that way. Yeah, I do think it's a fool's errand. And I think that if you compute as part of your joy of your life and this is your identity and stuff, have at it, have fun. But if you think that it's,

the problem is that when a lot of us see that person who happens to be also a great software developer, and then we idolize their setup, and we think like, if I could just customize mine to be like, Primagens or whoever it is, like whoever it is that you're idolizing, then I could also be a great software developer, like that person. And it's like, no dude, you're just gonna waste all your time

customizing your machine. And you're not actually getting better at the craft. You're just like tinkering around the edges of the craft, which I used to do as well. And so I've just set those things aside. And I think that we all would do well to set those things aside. In moderation, like you said,

I think you can get 80% of the way there with 20% of the effort. I customize some stuff. You know, I think you can get 80% of the way there with 20% of the effort. I customize some stuff. I got my new terminal last year and my new text editor and I've customized a few things on both of those just to make myself a little bit happier. Nothing wrong with that. You gotta get a blanket and a pillow.

But I'm not out there reading every config I can set on Ghosty in order to just like make it look like the sickest it could possibly look for instance. Right. Yeah, I think you realize there's just enough. So I'm all about good environment. And so I think what you're trying to suggest is that you should make the environment you're in comfortable,

whether it's a chair, a text editor, a new tool, whatever. H-top, you know, pick your environment of where comfortability is important and do what you gotta do to make it good for you. Your version of that and my version of that is probably two different measuring sticks. But you do you, whatever works for you, but don't do too much, don't spend all your time there,

cause you forget to do the task that was originally planned in the first place. I do wanna give this person a shout out because this person does create some amazing videos. And so I'm like mentioning this fella named Luda. If you want to check him out on YouTube, I'll link it up in the show notes, of course, youtube.com slash, and this is important, I've learned,

the at symbol. So youtube.com slash at capital L-U-D-A-L-X, Luda L- L X basically. Why is the at important? Cause you can get to ours without it. You know, I learned recently that somebody, let's see if it's important for this person. No, it's not important for this person. It is important for, no Tim.

So if you went to at techno Tim, I think, I don't know. I learned this recently cause I was trying to get there by just URL hacking so if you go to YouTube comm Www dot HGTPS Slash at Techno Tim you will get to the TechnoTim that we know. If you take the ad away,

let me see if this is still correct. You get to a whole different person. Really? Yes. So at TechnoTim versus TechnoTim is two different people, two different profiles, but that is not unanimous across YouTube so that your mileage may vary. If you do it on my-

YouTube.com slash changelog. Yeah, we're both at and on it and for my friend Luda LX It's true that he is both Luda LX and at Luda LX so pick your flavor go there and check it out his Latest video the cleanest Mac slash well It's not the latest the cleanest Mac slash PC hybrid desk setup you've ever seen now You will be enamored by the amazing cinematography you will be like I need those speakers to get by in my life if I could just customize my environment to be like that then I say quickly throw that away and just do as necessary to meet the environment yours do you remember Amy Hoy? I do, of course. Of course. Are you talking with Elmer Fudd almost?

Of course, of course. Who is that? Of course. I'm blushing over here. Yeah, of course. That was funny. So, long time creator out of the Ruby community, startup person, I don't know.

Talented individual. She puts out a lot of microsites, a lot of interesting visualizations. And I think one of the sites that she built that I still use is like everytimezone.com, et cetera. And she's always had a great design eye and the ability to bring her design eye to fruition. And I remember her complaining, this was years ago,

because every time that she would create something, usually she was creating back then, I think, a lot of cheat sheets for different tools. And they're just very well done. Like a get cheat sheet and whatever, rails cheat sheet. And they're just nice. And she says, every time I put something out there into the world that I've created for people. All these people in the comments asked me what tool I used to create this. Which is kind of the same thing as what's your text editor config?

And they're missing the entire point of how did I go about thinking through what a good design for a cheat sheet would be and how I pick my topography and my color schemes and all the things that designers think about when they're doing information design or technical writing even. And they asked me like, what tool,

like was it Adobe Illustrator? Like as if me telling them or giving my, me handing my set of tools to a neophyte would like turn them into an expert all of a sudden. It'd be similar to like seeing a master carpenter and being like, if I only I had all of his tools, I would be just as good as him at that thing. And it doesn't connect, it's not true.

And so trying to emulate or recreate an expert's environment is not gonna make you an expert. And we seem to think we can like jump in line somehow. So I don't disagree with that, but I disagree with that. And here's where I disagree. And maybe it's not a direct disagreement, it's more like a roundabout, let's just say. Okay, I'm ready for it.

There are many times, I would say in the last several years, where I've like admired somebody's work and wanted to know their tools behind the scenes. And at least from my perspective, it isn't how can I recreate you with me, or how can I be as expert level as you are. It's more like I have been down the road enough to know that there's a right tool for the right job.

And sometimes I wanna learn somebody's hacks, because their right tool is like literally that light is over there dangling off of duct tape You know, but it's in the right place and so there's things you learn by learning how they got to where they're going or what they've What they've used to get there and so at least my perspective isn't like let me become an expert like you It's and I maybe that's the line of question they've given Amy and that's just like me where I'm off Is that I would say this in particular like network Chuck

I was watching one of his videos from I think 2020 recently Because it's I stumbled on how he does what he does and I've always wondered how he did the pen thing on his screen in real time during the videos and It's so simple that you wouldn't think it's as simple and so unless you knew how he did it You might not stumble on the same simplicity Maybe it's changed since then I don't know but he uses a Wacom pen like I do and tablet and he has Photoshop open in a green screen mode so the background of Photoshop is The green screen essentially and his writing is, so it's chroma keyed.

And so he has software that watches Photoshop, and he can, in real time, during his videos, draw on the screen for you. And he sees it too, because he's seeing it in Photoshop. I would have never thought that that's how he did it. Am I going to recreate that? No, but if I think of ever wanting to do that, I'm like, I know now how he did it,

and I'll at least start there On my journey of finding out which tool helped me get to wherever I might want to go You do that with lighting you do that with You know you name it would work. I think you are talking about tools, but it's a proxy for techniques Which is entirely different thing is it yeah, I think so I mean obviously you use tools to deploy a technique sure But the how is really what you're interested in figuring out. Not exactly, and the what might be tangentially a part of it.

If I had that camera, I could be Chuck. If I had that camera, that lens, that light, I could certainly just grow my beard out and boom. No, I didn't think that. Well, for instance, you said there's a light hanging off of maybe it's duct tape over there, but they have it placed just right. And it's like, well, the knowledge or the experience

or whatever brought them to put that at the exact right place and know, you can't just transfer that by buying the same duct tape. Like you have to actually either learn it yourself or ask them like, what are some of your techniques for, you know, accomplishing such high quality? Like that would be a question that I don't think Amy would have a problem with.

But like, what tool did you create this in? Like she can just answer, it was Adobe Illustrator. And it's like, what did that do for you? It's like, okay, I should go get that I guess. Like have you ever launched Illustrator? Gosh, it's such a hard, hard tool to learn. Actually, you know the worst one is After Effects. Like good night.

Yes, well that's why I said I might be a little off of my disagreement, but that's how I approach it. No, I definitely, I don't disagree with what you're saying. I think you can certainly learn techniques from people's how I approach it. Yeah, yeah, yeah, no, I definitely, I don't disagree with what you're saying. I think you can certainly learn techniques from people who are good at it. And it maybe even just copy the way they do it until you can figure out your own process, but I think- Fake it, do you make it, emulate,

do you make it, whatever, picker. Right, but focusing on the tools is an incorrect focus in my opinion. For sure, for sure. And I like Network Chup, by the way. He's very talented in what he's done. And he will tell you like, hey, this isn't about the tools I use. He, by the way. He's very, very talented in what he's done. And he will tell you like, hey, this isn't about the tools I use.

He says the same thing. Like you can go and get all this stuff and not be the same. It's about, we know this Jared, gosh. It's about showing up. It's about being consistent. It's about caring. It's about a level of sweating particular details to iterate towards whatever your version of greatness is.

That's what it's really about. But I love it when people share some of the tools they use because I'm like, man, it's exposure, it's learning. Now I know, I've got a Wacom tablet. I knew how that worked, but I didn't think to use Photoshop with a green screen basically. That's cool. Whatever he did there was super cool very smart very simple, too Well friends, I'm here with some more obvious co-founder and CEO of temporal temporalal. Temporal is the platform developers use to build invincible applications,

but what exactly is Temporal? Samar, how do you describe what Temporal does? I would say to explain Temporal is one of the hardest challenges of my life. It's a developer platform and it's a paradigm shift. I've been doing this technology for almost like 15 years. The way I typically describe it, imagine like all of us, when we were writing documents in the nineties,

I used to use Microsoft Word. I love the entire experience and everything, but still the thing that I hated the most is how many documents or how many edits I have lost because I forgot to save or like something bad happened and I lost my document. You get in the habit when you are writing up a document back in the 90s to do control s.

Literally every sentence you write. But in the 2000s, Google Doc doesn't even have a save button. So I believe software developers are still living in the 90s era where majority of the code they are writing is there some state which needs to live beyond multiple request response. Majority of the development is load that state, apply an event, and then take some actions and store it back. 80% of the software development is this constant load and save.

So that's exactly what temporal does. What it gives you a platform where you write a function. And during the execution of a function of failure happens, we will resurrect that function on a different host and continue executing where you left off without you as a developer writing a single line of code for it. Okay, if you're ready to leave the 90s and build like it's 2025, and you're ready to learn why companies like Netflix, DoorDash, and Stripe trust Temporal as their secure scalable way to build invincible applications. Go to temporal.io, once again temporal.io.

You can try their cloud for free or get started with open source. It all starts at temporal.io. All right, should we continue changing minds or should we? Let's change some minds, man. What else can we talk about here? I got a couple more. Should we continue changing minds or should we? Let's change some minds, man. What else can we talk about here? I got a couple more.

Let me just rattle off. Yeah, let's pick a couple. You can tell me if any of them hit. Dry as we know it is a mistake. Okay, I agree. I'm just gonna change my mind on that one. I agree. SQL is good and everyone should know it,

but not really anymore. That's the other one I wrote down. You wanna talk to me about that one? Is that it? Is that your list too? Those are the changed minds. Haven't changed my mind. Convention over configuration as Rails is a great idea.

I think that's something that I thought originally and I still think today. And that's the end of my list. Like I said, I had four and three. So anything there you wanna bite off and chew or maybe we just- I'm down for any. Like whatever said, I had four and three, so. Yeah. Anything there you wanna bite off and chew, or maybe we just. I'm down for any. Like whatever you're most passionate about,

let's, you pick. I like them all. You choose since it's your list. Well, let's talk about SQL, because I think that that might lead us into another conversation. Or maybe we just end on that. So I've always been a pretty big fan of ORMs, honestly,

probably, which is object relational mappers. For the uninitiated, this is when you have a programming language library that maps on top of your database rows and columns and allows you to crud, create, read, update, and delete database records without writing native database language of SQL, structured query language. I used to really dislike SQL.

I thought it was gnarly and hard to learn and ugly. And I've always liked ORMs, mostly because ActiveRecord, which is Ruby on Rails ORM, has always been a pretty good ORM. I've also tried Datamapper and other ones back in the day. Nowadays I use Ecto from the Elixir and Phoenix people. And I think SQL is kind of like Vim. Remember Gary Bernhardt telling us, you know,

on the old Vim show where he's like, you tried this and over the years, like you've went from TextMate to Adam to this. And he's like, I've just been getting better and better at Vim. And I'm like, I've just been getting better and better of him. And I'm like, oh, you're so much better than all of us, Gary. Which is probably true, but I'm still using Zed today.

So I didn't learn my lesson. I feel like SQL is kind of like that. You can invest in the language and a library and then you can switch libraries and they all kind of map on top of what are ultimately outputting SQL expressions. And some are better and worse

and they allow you to like, you know, break out of the box and write your own fragments. And I think ECHO is a pretty good one. But after years and years and years of like ending up with SQL and then learning it because I'm looking at it now, even though I didn't write it, I wrote some Ruby code for instance.

I just think SQL is really powerful. It's been here since day one practically of most of our careers. It's not going anywhere. And it's one technology, the language I'm speaking specifically, that is worth every software developer learning because it's transferable across jobs, projects, languages, et cetera.

And I didn't really invest in it directly my career. I kind of learned it slowly through osmosis and I probably would have been better off just setting aside all the ORMs all the years and just being like, learn SQL, really, really, really good. And I'd probably be a better developer. And so I've changed my mind about that. It's probably too late for me. But for you youngsters, maybe invest in SQL.

However, my addendum to that sentence is, but not really anymore. That's cause you know what is really good at writing SQL is language models. They're just really good at it. And so do you have to learn it anymore? I don't know. I guess you have to be able to try to use it after your LLMs written it for you and make sure that it works the way it's supposed to.

But I haven't written a SQL query in a couple of years now with myself. And so there's my addendum is like maybe we don't have to learn SQL because it's so low level. So low level, that's a weird thing to say. It's low enough of a level that your everyday developer, you know, starting today and going forward won't have to just write it directly anymore. They'll have tooling that writes it for them. So that's my change of mind, but also not so sure

because of what they're doing to our workflows. Yeah, I think it's still, I think it's still valuable learning it though. I mean, cause if you write Ruby code and the ORM maps out and spits out SQL and you have a better understanding. Or that LLM is down or unavailable. And you're like, well.

It's local baby, come on. I'm kind of stuck here. Can't, yeah, I think I remember how this SQL thing works. You know, and you're sort of stuck there. I think that seems like a first principles scenario where like you eventually will have, the first principle of being a good developer would be to use a database.

If you're using a database, use a SQL-based database. And then to understand at some level how to construct SQL queries. But still be proficient in Ruby and ORMs and how you use it natively in that language and that framework is you're faster, better, whatever. And I'm with you. I think that's, that's, I think where I want to, that's where my curiosity sort of lands

at is, is give me the things that make up the foundation. I don't want to become 110% expert in that foundation. I want to become 50, 60, 70, maybe 80 if it's really important to me, but at least 50% on some of these foundational things and potentially get to 80% of mastery with it because, or awareness of it, or understanding of it, because I think that gives you a good baseline

so when you don't have your tool belt or your Your complexity killer, which is an LLM. It's like hey, I want to do this, you know query Write it for me. Like I don't want to like there's there were some things I was doing recently with a a backup and I was in bash backing up like a directory and I don't know all of seven Z's params that it can be tossed and at which case I would use them. I would never wanna do that. I just wanna know that seven Z's a tool, that's my 50%.

Seven Z's a tool, I'm using that tool to compress this thing and I can use it on Windows, I can use on Mac, I can use on Linux, it's available anywhere I want. I could use a Linux. It's available anywhere. I want But I'm not gonna go and learn all the things about this thing. So I know which params to use to keep the permissions to Do different things or certain compression levels like I know I'm gonna ask the LLM to help me with that because that's mental bandwidth I want want to give to something else. That's more important, right? You know, but I want to know enough about 7z that it can

And I don't need to know about all those swappables there same thing with SQL like you should know enough about it to be Content and I would say secure right in the code you're writing and the careers you're you're putting out there because Well, then you're not topp the careers you're putting out there because well then you're not toppling over your database or spinning up your CPU to whatever and your database machine topples over because you've written a poor query. Yeah.

No, I absolutely think that SQL is one of those foundational technologies that I think I eschewed when I was just coming out of college, because maybe I was either taught probably indirectly that it was kind of a passing fancy or it was like a thing that people are using but it's not gonna necessarily be here forever. And there are query languages that have come and gone.

I mean, I think that if you spent a lot of time investing in for instance, MongoDB's query language, that's not useless information, but it's not broadly useful because there's just not enough databases out there that speak MongoDB query language or whatever it's called, MongoQL, I don't know. I know it was JSON-esque or JavaScript-esque.

And I've been leery of any database technology that comes out with a brand new query language and not SQL in addition. I think that that's a dangerous proposition in terms of personal investment. But I'm not sure how much, like you said, like how much does each person need to become expert in the foundational technologies today

and tomorrow versus maybe just proficient enough to guide and direct the LLMs towards success. I don't know, I feel like that's a moving target. And one that I've been more skeptical of probably in the past than I am today. I think it's progressing at a more steady clip. If I'm talking about language model advancement is progressing out of a steadier clip

than I thought it would. And they're getting better. But yet still not ultimately reliable. So there you go. There's my change of mind. I think if I was coming out of college, if I would have been talking to me, coming out of college.

I would have asked myself, what should I really learn? And now today I'd say you should learn SQL for sure. Just directly, go learn it, master it. It's gonna help you. 80%, 110%, where would you land? Not 110% is technically impossible, but I go. It's a stretch.

I know, I just like to point that out and be pedantic. I'd go probably 80%, yeah, I'd get 80% of the way there, why not? Okay. Whereas I probably only got like 30% of the way there. And then took me years and years and years and years to like flush out that knowledge. Yeah, you know, the challenge though with learning like that,

I have found personally to be the case, and I'm sure you probably agree with this too, is that I can't truly learn and retain the learning unless I have a reason to learn it. And so maybe with you, it took those years and those different experiences to have the reason to learn to the depth of 80% as an example. Yeah.

Yeah, I've always been a means to an end kind of guy. So I do have to have an end goal in mind. I don't learn for learning sake very well, which is kind of what you're saying, right? Like you have to have a reason. Yeah, it's almost like I have to have a need, I have to have a problem, I have to be trying to solve something

or at least some curiosity that I'm trying to close a loop on and then I can learn it, you know? One example I can give you is that I've shunned away, if that's even the right way to say it, I've just written off windows in my life, years ago. Yeah, same. Until recently, until recently.

Now I think there's a world where I, as a computer user, slash developer, slash podcast, or slash business person, slash whatever, I find that there's a place for Windows, there's a place for Mac, and there's a place for Linux. Where's the place for Windows for you? This is still in motion. This might dovetail into a topic that is a bonus. I would say So I tried to I Prefer Docker. So let's just caveat that

for running applications, I tried to play with LLM's O llama and stuff like that open web UI on Linux and had issues with seeing the GPU and enabling the GPU to Docker on Linux. And maybe it's just me and I'm still just not there yet, but windows was super easy. It just just worked because windows is so widely used that and Nvidia

is so widely used on Windows machines that there's a perfect marriage there. I didn't have to go and load drivers and be special about it was just there. And when I launched Docker desktop on Windows, well, guess what? The GPU was available. I didn't have to do anything special. And in minutes, I was to a place that took me, you know trials and tribulations Beforehand and it could be because I was poking the dark and it could be because of X ones. I don't think it's Linux's fault but I found that getting to

Olam running accessing the GPU and Open web UI being there tail skill on the machine and then now that Essentially AI machine is available to the network and all my tail net was just like so fast compared to the same path with Linux. And so creator PC is even like I like to build I like to build machines. They're fun to build swapping out components, you know, choosing your CPU, stuff like that. I think the thing that Macs do for people like us is they really simplify it. It's like, well, I don't have to think about any of those parts because Apple solved it in a single small box

and it turns through this power. But to a tinkerer, that's not very fun. And I admit that it's a solved problem by them, but it's not that fun So a creator PC or an AI machine, I think those are the two areas where I think Windows has a Has a good chance to to kind of go there for me. Could you do the same thing with Linux? Probably but I had issues and I was like, you know what? Maybe I can solve this because I've seen other people solve it easier with Windows and it's actually not that bad It really isn't there's a cool script from

Chris Titus that will install and de-bloat certain things around Windows To make it more enjoyable as an experience less user hostile as I said on that show with more enjoyable as an experience less user hostile as I said on that show with Our good friend Tim Stewart slash techno Tim. I was like, you know, I've written windows off. It's user hostile I don't like it, you know But not given another chance and I'm further along to the AI machine running on my network and accessible via my tail net To the point where I've got, let's see here, multiple models running, you know, just too easy. It's too easy now, it's just there.

And you wanted to build that machine. You didn't want to buy something. Well, what would you buy? A Mac mini. Yeah, I suppose, yeah, like what's the, that's no fun. Cause you said Linux versus Windows for your AI machine. I think Mac OS runs a llama. Great. So I'm not even saying there's a right way or wrong way.

It's just more about exploring. I was just curious why you ruled out Mac OS for that use. I think for a creator PC, then yes. Yeah, I can't build a Mac machine. That's why. Like I can't do that. So that was a prerequisite to your decision making. Was like, I'm going to build something.

Yeah, being able to build something or repurpose hardware already had, you know, basically Into something that wasn't like if I could run if I could build my own machine and put Mac OS on it I would totally do that, but that's not possible If if I had my Rathers, I would install Mac OS over Windows pretty much any day of the week You know any day of the week, but that's just not the option that Tim Cook slash Steve Jobs and anybody else at Apple's given us. It's just not, it's just not there.

Are Hackintosh is not a thing anymore? Maybe they are, but they're always, there's a word in there, hack. It's always seemed like trouble filled and issues and. Isn't that part of the tinkering? Nah, not for me. No? No. I mean, I've never been curious to the

Hackintosh way. I've been curious to the Linux way, the stable production ready, production possible way, and I could be wrong, but I feel like the word hack in Hackintosh is what just deters me. For the same reason I wouldn't run, was it Ahi Linux or Ashi Linux or what's the name of it? Asahi? Asahi, gosh, way off. On a- It's Asahi, I don't know how to spell it.

On an Apple machine to run Linux, it just seemed very buggy. It's not what my curiosity is, I'm not saying it's wrong, it's just I'm not curious there. I would much rather, but then again I'm thinking like, gosh I would much rather, but then again, I'm thinking like, gosh. Well then you could have your cake and eat it too.

You could have your pregnancy with Mac OS. Precisely. That's where maybe I'm wrong. Maybe an M4 Mac, a base level. I'm not sure hackintoshes are still a thing or not because there was a big community back in the day, but I know with the switch over to ARM, things changed and I don't know,

I don't keep up because I just run Apple hardware but I wonder where the state of that is. Yeah. Well I would say you could probably answer that because you've got an M1 Mac. You run a llama I think. How do you run your local AI? Do you do that at all?

Do you just go to chat GBT every time now? Are you local only? It's kind of like Google and dot dot go, you know? Yeah. Where like I use dot dot go. And then I'm like, this result might not be good. I go to Google. Okay.

And I kind of do that with, I switched back and forth between DeepSeq and Llama 3.2. Okay. And now the DeepSeq is out, the DeepSeq coding latest, I think, was the one I was using. And I use that. And then when I'm like,

is this the best response that you could do? And then I go over to chatgbt and I add, you know, I paste the same prompt into chatgbt. And I think that chatgbt pretty much is the Google to my local.go currently. It's better. Not as good as change.log++, but it is better, but it is not that better. It's a different better Okay, so there's where I'm at I also wanted to be curious too. It wasn't like oh, let me build this I

Got no problem. I'm just playing devil's advocate. I like that because I mean it pushes back on your reasons why you would do X And also I was like, you know, maybe this Windows world is kind of, it's got its pros and its cons. Maybe it's cool. I'll tell you one thing, it's got some cool things in it. It just remaps certain keys. Right.

One thing I like about it is whenever you hover over the icon in the task bar, and if you got multiple instances of it, it will let you choose which one. Like if you hover over it, it gives you this new UI that shows you here's one, two, and three. For example, like FoxBor. That's kind of cool. Am I going to move there for that? No, but it's nice that it's there. And I think WSL has some really cool stuff involved and I haven't even gone as deep as I could so far but like Windows subsystem for Linux version 2 is what is being used now.

You can run Linux side by side with Windows is almost like it's a VM. Yes, but it's it's it feels more native because you can reference the file system on both. You know and I think that's super cool the way they've melded that together and I haven't explored it all but I think it's just wild the way they've melded that together. And I haven't explored it all, but I think it's just wild how they've done that. And we've known about WSL for years. And here I am just now learning it, you know? Here's me, you know what I mean?

Like that's just the jit of learning in my opinion. It's like, I've known about WSL, but I've never had a need to learn it or desire to learn it beyond just knowing it's there. Now I'm like, okay, I can probably spin up this and run Docker and Docker leverages WSL on Windows to make Docker do Docker things. But man, it was just pretty easy, honestly, pretty easy.

So let me ask you a real dumb question from somebody who hasn't booted Windows since college. You build a creator PC. How do you get Windows on there today? Do you like USB drive it or whatever? Yeah, yeah. It is a bootable ISO. And if you're doing it on Mac.

Just like it used to be. There's not like cool new ways of doing it. Same. You know, you're just booting into your BIOS. You put your USB stick in there, you're confirming which, you know, selecting that USB drive to boot from, boots, and you do the install process It's no different than the way you would do Linux or and how do you get it onto there?

Do you just download it for free off of Windows calm or yes? Yeah, you can download the 64-bit ISO from Windows and he used to cost money for that so they it does still cost money That's not cool it. I believe they'll let you use it to a certain capacity for free You can't change the desktop background. You can't do certain things, but it will be a machine I believe they can run in perpetuity without any restrictions. It's like certain things. You'd want to do you can't do without getting a license key and so I really wish like that would be Maybe they would lose tons of money, but that would be cool like make windows free like Mac OS is free Like Mac OS is essentially free to anyone who could run the hardware

I figured they would eventually get there But they also have so many enterprises paying for licenses that I'm sure there's like Some bucket of money that would just stop flowing in that's not worth doing that. So Windows by default should be free or could be free as an example in this world. Windows Pro could be the paid license. Just saying, just saying.

It's got some cool firewall stuff. Windows Firewall Defender is pretty cool. It's pretty easy to configure. It's pretty easy to hop into the command line and do like IP config to know where you're at, for example. It's pretty easy to run things in like pseudo mode or whatever you want to call that, like as an administrator. I mean, it's not dramatically different than maybe even Mac OS is in terms of what you could do.

They're just like, it's just a different flavor, honestly. No, I get it. Here's why I will never go there. Okay. Backslashes are trash, dude. They are trash, man. I haven't encountered any backslashes yet though. I mean, they're there.

You haven't gone to the C prompt, right? Don't you like, if you launch command.exe, it launches you to C colon backslash. Doesn't it? Or if things change that much. You know, I haven't navigated the file system via the terminal yet. That's uh.

So I just haven't gotten there. Don't do it man. Don't do it. Once I'm there, maybe I'll be like, this is the worst ever. Honeymoon will be over. Yeah, I'll be like upset about it. All right.

But. Well, that was a good, I appreciate that you're exploring, that you're tinkering, that you're trying out things that most of the people listening to this know very well, all the basics you've described to me, you know, about how Windows works.

You'd be surprised. Meanwhile, I don't know anymore because I haven't run since college, but cool. It's a Windows world that you're entering. I'm open to a Windows world. I'm open to a place where. I'm open to a place where- He's exploratory.

Where Windows is an option. I would love a monitor that has KVM built into it. This guy did, and it's not because I'm following him. I'm curious about this world where the monitor can be the KVM to it. You have a keyboard and a mouse plugged into the monitor or connected to the monitor. The monitor does the KVMing. You can swap between a Windows PC and a mouse plugged into the monitor or connected to the monitor the monitor is a KVM

You can swap between a Windows PC and a Mac machine now. I would love to see how well some of these LLms run via llama on like an m4 Pro or something like that Because maybe that's just the easy button Maybe that's the less fun tinkerer button where you can I mean? I'm not getting a chance to build that machine but you know that's kind of the Inquotes fun part is choosing the components and choosing the motherboard and putting it in a cool case and

Maybe some RGB if you want to now. That's not me I'm not an RGB kind of guy, but I do like things to be aesthetically nice, you know? So, you know, thus far it's just in this rack mount kind of ugly thing that was just in the rack. Like it was a different machine beforehand. It was actually my Proxmox machine that ran Plex. And I built it for Plex before, and now I've repurposed it. Gotcha.

Well, stay tuned. Will Adam's mind be changed about Windows? He's open to change. He's open. It's production ready right now. I mean, I can go to, you know, the machine's name is Genesis. I can go to Genesis right now and run OpenWebUI

and I've got Llama 3.2, 5.4, latest DeepSeq 70B, DeepSeq 32B. 70B is a little slower but 32B is pretty fast. I mean it's almost real-time results with it. 5.4 is pretty good actually. And do you run your Macs against it then? Like are you running it from your Mac over the network to your AI machine? So I'm in the browser right here in Safari.

Oh you run it in the browser. On my Mac. Gotcha. Yeah it's Open Web UI. Do you know about Open Web UI? Yeah I just think you should hook up something native to that like the enchanted one or something like some deeper Mac integrations.

Cool. I haven't used Open Web UI, maybe it's like amazing and you don't have any need for something like that. Well, Open Web UI basically is a version of what ChatGPT is to Olamma. So rather than doing it from the terminal, you're doing it from the web, like ChatGPT. And you can choose your models,

you can do code interpretation, you can do, you know, code interpretation. You can do all sorts of other cool things. You can, you can rightly and left lane two different models. So you can single prompt, add two models to the same prompt and let whichever one win and choose the one that's winning. So if you know why I will sometimes get better mileage out of five four versus, you know, llama three will sometimes get better mileage out of 5.4 versus, you know, llama 3.2, for example, you can compare those two together in the same question and

just choose which path you want to go or keep going down both paths. It's kind of cool. So are you off Chachi Petit then? No. So you're like me. That's the test. So that's my hypothesis is like, so my question really is this, this is the question thing. I'm trying to answer It's one fun to tinker but two is it worth Spending the money on hardware to local your AI Given that more and more models would become more and more open source Or is it better to pay X dollars to open AI or? whomever perplexity or You know you name the different places you can go to Claude for example

There's some cool stuff happening in clock because there's a lot of things are doing in that web UI Claude in particular with like document based Collaboration with the LLM Stephen would say collaboration given that you're you know you're so anti-humanizing these things, anthropomorphizing these things. But being able to, in the UI, iterate on the document in a way that you, as the prompter,

can know it's not being changed. It's like changing parts versus whole document kind of thing. Cloud's got some cool stuff going on, too. So to zoom out again, my question really It's like changing parts versus whole document kind of thing. Cloud's got some cool stuff going on too. So my, to zoom out again, my question really is like, is it worth it? And it may not be true today, but at some point it will be, is will it make more sense to have your own AI-able hardware

on your LAN, on your local network, to run against home assistant to run against yourself Your kids whomever could be could it be a service inside the household? Could it be that today and skip the bill and I don't think it's about cord cutting because this is like kind of cord cutting In a way, but it's like I don't always want to share all that info with someone else. It's in my own local LLM. Right. You know, and you have nothing to hide until you have something to hide.

You know what I mean? That's right. No, totally. I think that's a real good reason to want to do it. And I think that if we could, I think the models will get there in terms of their quality responses. And I think that the interfaces will get better

with how you actually like feed your life into that local network thing. But like will the devices come along? You know, like it would be rad if I could take my echo and point it at that instead of Amazon servers. Now I don't think Amazon has any reason to do that. But if there was like an echo, like, cause you know, ultimately, like you said

on a previous episode, you wanna talk to these things. And right now we're talking to an Amazon server. I'd much rather ask questions or have my kids ask questions of something that's running on my local network. So when they do ask a personal question to an LLM, it's not, you know, forever in Amazon's database. Tied to a human, you know, I think that that's really creepy.

It is. And so I think, I think that that's really creepy. It is. So I think, I think that ultimately these efforts are worth it. I think right now, the trade-offs aren't quite there yet. And there's way too much groundwork to be laid by nerds, but we'll get there. And I think the recipe though is you could do it on Linux, Mac or Windows apparently apparently is Docker

O llama open web UI Either a really powerful CPU or an IGP. I don't think O llama actually recognizes IGP is Or GPU and GPUs are scarce Expensive you can find decent ones on eBay. They're actually If you can eBay well, you can find a decent one Maybe just buy the M4 Mac Pro and call it done and maybe you'll have a limitation, right? That's essentially an IGP is an integrated GPU. I think a llama runs well there to my knowledge But if you want to like go crazy and like do multiple GPUs, you're gonna have to motherboard that thing, right?

That means you're gonna choose between Linux and Windows and in my case I was like, okay I got better mileage or easier mileage on my shoes when they were Stanford Windows and so that's the route I went but basically Docker a llama open web UI pick your model from their capable hardware and open web UI, pick your model from their capable hardware. And that's how you can get to local LLM today. You can play with DeepSeek. You can even do a smaller parameter model.

Olam obviously, or sorry, llama three two is probably the one that most people can run because it's small enough. Llama three three is pretty big. It's 70 billion parameters. It's sizable, 70 gigs I believe, which is like larger than most people can run on a single machine unless you got like multiple GPUs. I Wanted to get to that point where I can run those things and I would love it now is to is once I've sort of gotten

that mastered and Running stable to begin to and I don't even know what the terminology is there So forgive me if I'm stupid in this regard. Having done what we've done for so long, now I'm actually learning it. Is I wanna train one of these models or teach it a particular trick. The particular trick is like, okay, here's all of our agreements, here's how we price,

here's how we quantify, make that job easier. Can it make that job easier? Can it make me faster in those ways and be just as accurate and sweat the details just as much as I do in the real time doing those exercises myself? That's what I'm really trying to get to, is like I would love it if I can get to certain things I do

that are deep think things that are crucial in business to get right. Don't wanna give people the wrong price, give them the wrong contract or the wrong terms. I wanna give this thing certain parameters, certain business parameters to get to certain results. So I can kind of like word calculator my way to faster, easier, better quotes for people and stuff like that.

I don't want to write my novels. I just want it to help me make those kinds of things faster and be my word calculator and be very accurate. So that's my next journey. Cool. Let us know how it goes. Okay. I will Bye friends. Bye friends. Okay. Here in the post show. I want to talk about some bonuses, I suppose. Some extras on that last segment about running AI services locally. Now I built my AI home lab with Windows 11 Pro.

I have a GPU in there, of course, an RTX TUF 3090. And my stack is a L open web UI running in Docker Kind of easy to set up. Honestly, it was really easy with Windows drivers Nothing missing very very fast and I'm able to get some really good results I will say however that I've gone back to open the eyes For a model their oh one model and I've gone to Claude as well for some other things because there is so far not a single silver bullet for running your AI. So I'm very much in tinker mode here in my home lab, in my AI home lab, I should say.

And I want to hear from you. So if you're messing with tinkering, and learning about running AI inferences, LLMs, et cetera, locally, you're playing Linux, Windows, Mac. I want to hear from you. We want to hear from you. Hop in the Homelab channel in Zulip, go to changelog.com slash community.

And if you haven't signed up yet, that will get you in. Of course, it is free to join and we'd love to have you there. Again, changelog.com slash community, it's free. And for those who haven't caught on yet, we are publishing full length episodes with lots of clips to YouTube, fully chaptered, clips are there, lots of fun things.

Subscribe to our YouTube channel, youtube.com slash changelog, full length episodes waiting for you. Okay. Big thank you to our friends over a retool retool.com. Our friends over at augment code.com. This is your developer AI assistant that you need right now. Augmentcode.com. And of course our friends over at Temporal. Temporal.io. Solving retries for everyone. Enabling resilient non-breakable software.

Again Temporal.io. And of course to the beat freak in residence, Breakmaster Cylinder. Bring in those banging beats, the best beats in the biz as Jared says. Gotta love those beats. I love them. Hope you love them too. Okay, this week is done. The show is done.

Friends is done. And so we'll see you on Monday. Enjoy your weekend. I'm gonna be a good boy. Thanks for watching!