Baby Registries, Cold Showers, and Launching Opencode
Yeah. I I think But, like Go ahead.
Speaker 2:Go ahead. No. You go. Who knows what I was about to say?
Speaker 1:Yeah. You sleep deprived. Who knows? Did you notice that the names stay?
Speaker 2:The what?
Speaker 1:What a funny Riverside feature. So under our names, you see the descriptions?
Speaker 2:Yeah. Oh, that's from the stand up?
Speaker 1:Yeah. So they can change our descriptions and then they stick in another suit.
Speaker 2:I can't read yours because you're wearing a white shirt and the text is white. And I'm so curious. I don't remember what
Speaker 1:It's slipped. It's it says it says resident nice guy.
Speaker 2:That's a good one for you. That's funny.
Speaker 1:It's funny.
Speaker 2:Mine just says vegan vegan vegan vegan vegan.
Speaker 1:Yeah. Which is accurate? Yours is accurate and mine is also accurate.
Speaker 2:Is it?
Speaker 1:By the way, you and Casey sent us baby stuff.
Speaker 2:Yes. We did.
Speaker 1:Thank you for that. Also, how did you even think to find a registry?
Speaker 2:Casey's amazing. Yeah. Not gonna assume
Speaker 1:it all I'm assuming it was I
Speaker 2:did I did find out though before you did, so that was good. She she told me she did it. I was like, yes. So I could be like, yeah. You're welcome.
Speaker 2:And not be like, what are you talking about?
Speaker 1:Yeah. The the baby registry thing was funny because so I never done anything like that because I have a small family.
Speaker 2:You never had a baby.
Speaker 1:Yeah. Yeah. Just registry in general, like or just your family buying you stuff. Like, like, the more people you have, like, the more stuff you can get. Right?
Speaker 1:So we had a small family and we never had a we never did the wedding registry thing because we had a small wedding. So Liz was like, man, my family owes me because we haven't done any of those things And she's got a big family. So she started making this list and literally as she was making it, like, you know, one of her aunts was like, oh, like, let me see what you did so far. And she sent the list to her aunt, then it started leaking, like the list started getting around and people started buying stuff. And then my mom was like, wait, what?
Speaker 1:Like, I wanted to get that. What? I wanted to get that. And my mom felt like
Speaker 2:Oh, yeah.
Speaker 1:Annoyed that she couldn't get anything that was, like, very meaningful, you know? Like, all the good ones were taken. Yeah. Yeah. And I'm like, this is amazing.
Speaker 1:Like, boxes were just showing up to our house and, like, it's so expensive otherwise. It's really
Speaker 2:it's a really great situation. Yeah. It's a great situation. I can remember I I don't know if we did one for the wedding. I know we did for the the babies, but it's a good feeling.
Speaker 2:Yeah. You just got a lot on your mind. There's a lot going on and just having people buy you stuff and like be excited to buy you stuff is kinda awesome.
Speaker 1:Yeah. Like fighting over it. And then we have so much stuff and like, it's spread out among like, know, 50 people. So, yeah, I'll do the most expensive baby monitor. Why not?
Speaker 2:You know? Yeah. Exactly. My expensive taste really shines in these moments.
Speaker 1:We randomly came across this store like in like physically, like we saw this like baby stroller and like car seat store and we went and visited it to see like, oh, let's check out what strollers we wanna do. And they literally had like this strip of all the different terrains, which I swear they like Oh. Pluck from our neighborhood. Like, it it literally covers all the terrain in our neighborhood and you can try each stroller on it.
Speaker 2:On the different terrains. Yeah. That's amazing.
Speaker 1:And then and Liz was like, you know what? So we tried a bunch and was like, know what? I really like this one. And I'm like, okay, cool. Like, let's look at the price.
Speaker 1:Of course, it's the most expensive one. Right? So I'm like, okay, whatever. Like, goes in the registry. I think her parents got it for us.
Speaker 1:And then I'm like, okay, I didn't have time to do the car seat and get it on the registry to do research. So let me like so a couple days ago, was like looking up car stuff. Of course, the stroller we have is like the apple of strollers where like nothing is compatible with it. Like, you have to just be locked into their the stroller ecosystem and everything is like 30% more expensive because of that.
Speaker 2:So Yeah.
Speaker 1:Yeah. It's it's the it's a Bugaboo Fox Pro five, which it's a great stroller, it's a great product from what I can tell so We've we've taken it because Zuko is like a little afraid of because they've always been afraid of like shopping carts and strollers and stuff, so we were practicing it with him. We were
Speaker 2:with it. Yeah. But we
Speaker 1:like, went on walks with it and I'm like, I really hope nobody like tries to look inside and be
Speaker 2:like, oh,
Speaker 1:let me see the cute baby. And they're like, what? It's empty?
Speaker 2:Put a coconut in there or something.
Speaker 1:Great stroller, but you know, it's a it's a Strollers
Speaker 2:are a great investment. Like, I've got a five year old who still sometimes would go on family walks and he rides in the stroller. Like, he's not too big to to sit in it. So we've got
Speaker 1:Is it like the same stroller that you had from when they were a kid and then like convert to all these different stages or is it like a you
Speaker 2:know, I don't remember either I don't think either of them really did the stroller thing when they were little. Like, babies that didn't wanna be in car seats or strollers. Like, we just we didn't do it much. I think in fact, I think Casey got this stroller when Archie was already, like, two or three. I don't think
Speaker 1:it was Oh, okay.
Speaker 2:Ever in use as a baby. So maybe it doesn't even work for babies. I don't know.
Speaker 1:Yeah. I I guess our lifestyle is a little different because, like, we're gonna be like going around our neighborhood a lot. So
Speaker 2:Yeah. You can walk Yeah.
Speaker 1:And then
Speaker 2:for reasons that are not pointless.
Speaker 1:Yeah. Not forcing yourself to walk Yeah. And just like Uh-huh. Wide empty roads so you don't die at age 60. Yeah.
Speaker 2:Uh-huh. Exactly. That's the only reason we walk.
Speaker 1:Yeah. Speaking of walking, I have some things to share with you. I
Speaker 2:Nice transition. This is I'm so excited. What is this gonna be?
Speaker 1:Well, you know what's related to walking? Running. And I've gotten into running two weeks Me too. Oh, really? Nice.
Speaker 1:I mean, yeah.
Speaker 2:I've been doing it in the mornings.
Speaker 1:Yeah. Yeah, me too. Nice. Here's a funny little thing. So I got I started two weeks ago.
Speaker 1:I'm on I'm on I just finished week two. And the first week I started it, our water heater broke. So we didn't have like hot water.
Speaker 2:Did it flood?
Speaker 1:As it flooded No. Like the power went out briefly at night and came back on and something about that like, it needed like a hard reset from inside. But there were three days where we didn't have it. And as an Arch Linux user, normally I'm fine not showering for three days. But given I just picked up running and it's a summer in Miami, I came home drenched in sweat.
Speaker 1:I'm like, there's no way I just have to take a cold shower. And my whole life
Speaker 2:Cold shower.
Speaker 1:My whole life, I got to my whole life being like, cold showers are so stupid, like, why would I ever do that? It's just unpleasant. Like, I just want to be comfortable. I don't give a shit. I want to take a hot shower.
Speaker 1:And I so I I I turned on the water and I was I was annoyed, I'm like, oh, this sucks. And I get into it and it sucked for thirty seconds and something in my brain flipped where it was like, yeah, it's like being in the pool. It's hot and you get in pool and it feels amazing. And now, I'm like addicted to cold showers. Like, I look forward to the run specifically so I can get this feeling of like going into a cold pool.
Speaker 2:Yes. Like, I love sauna or exercise, get really hot and then get in a cold shower or cold plunge.
Speaker 1:Yeah. It feels crazy good. And I'm like this like, in hindsight, it makes sense because like, yeah, like when I get into a pool, I'm not like, fuck this pool. I hate it. I'm like, this is awesome.
Speaker 1:Yeah. When I get
Speaker 2:in the
Speaker 1:ocean, I I I like it too. So it's just funny, like, it's like what your brain expects and that determines what has to be your Yeah. It's like, it makes no sense that I would be drenched in sweat after running a 90 degree heat and get into a hot shower. Hot shower. But I would like you because that's what my brain expected.
Speaker 1:Right? So if these two but what's funny is if these two things hadn't happened in the same week, I would have never discovered this. Like, it's like a really
Speaker 2:happy accident. True. Like, if you hadn't been running, it hadn't been summer, and you hadn't lost your hot weather here. That's amazing.
Speaker 1:Yeah.
Speaker 2:Coincidence.
Speaker 1:I'm like a whole new person now. I feel amazing every morning. Yeah. It's great.
Speaker 2:I feel like garbage, but that's because I I can't sleep. It's been I think I'm day three of like very, very low sleep.
Speaker 1:Is it because you're so super excited about everything?
Speaker 2:Is it? It's a combination of personal struggles and work excitement. It's definitely stress. It's like cortisol in my system in the middle of the night. I just can't it's the can't stay asleep thing.
Speaker 2:I fall asleep great. I'm like a champion at falling asleep at the beginning of the night
Speaker 1:Mhmm.
Speaker 2:Like world class, five minutes tops every time. And I sleep till about two really well every night. And then after two, it's just like a coin flip. It could be a great night. It could be a night where I get up to pee like last night at 02:30 and I am up the rest of the day.
Speaker 2:I'm laying in bed till five, but I am literally awake. And that's been like two or three days. So it starts to pile on and like, I'm dragging.
Speaker 1:Do you not crash at some point?
Speaker 2:I'll sleep really well tonight. Saturday mornings because I'm not I don't work on the weekends anymore like at all. Yeah. And I used to wake up early on Saturday and Sunday and still in work those days. Now that I don't wake up to work, I somehow mentally like my brain just says, now I can sleep and I sleep super I'll sleep till like 07:30 tomorrow, which is like insane for me.
Speaker 1:Yeah. You know, it's yeah. It's funny. It's like when you get older, all this stuff change at least for me, it was I don't know how you were like as a kid, but as a when I was like a teenager, I could just sleep till like 1PM.
Speaker 2:Oh, yeah. Yeah. The afternoon sleeps.
Speaker 1:Yeah. It's like the concept of like, oh, no, I woke up. I can't go back to sleep. That's just like, was so
Speaker 2:I Like,
Speaker 1:what what the heck is that?
Speaker 2:I think about this all the time, how much like just sleeping is such a hard thing in my life and like as a kid how dumb that would have sounded. Like, what do you mean it's hard to sleep? Yeah. I think it's hard to get up ever, but like when you're a kid. Yeah.
Speaker 1:And I'm still like I I still see myself as someone that like sleeps kind of like a rock. But even I've had situations now where I've gotten to this habit of like, when I wake up and like I wake up past a certain amount, my brain just flips into you're awake, you're starting your day, you're doing everything. Mhmm. And that gets triggered by accident sometimes. Like like Liz had a Liz changed up her her alarm the other day and she like or she got up like much earlier than she normally did and my brain just acted like I was waking up and I like got like an hour and a half less sleep as I normally do.
Speaker 1:And it's like wild to me that this is like a problem. Usually, like my whole life, waking up was the biggest challenge Yes. Possible.
Speaker 2:Growing up, going to school, every morning. I hated waking up every morning. Like
Speaker 1:waking up and not going back to Yeah. It was every single morning. I'd always miss the bus. It's just a huge thing. So, yeah, I think when you get older, you're just like, I'm working now or like, I don't know.
Speaker 1:Maybe it's because this is like a positive way to spin it. I just didn't want to go to school, so that didn't really help.
Speaker 2:Yeah. Right. Yeah. The positive way to spin it is that, like, we enjoy our adult lives and we're excited to wake I wanna be awake
Speaker 1:for it.
Speaker 2:Yeah. No. It's like most of fun things in life happen when you're awake.
Speaker 1:Yeah. All of them, maybe? I don't know.
Speaker 2:I had some breakfast, so I'm feeling a little better now. I was like an hour ago, I was like, am I should I be on a podcast? I don't know.
Speaker 1:Should I be Yeah. The other thing is we launched OpenCode yesterday. So congratulations to you and to all of us. It is our biggest launch ever. I think my tweet has 2,000 likes on it.
Speaker 1:It hit a thousand likes so quickly.
Speaker 2:Our our biggest open code launch ever because I feel like we've launched three or four times.
Speaker 1:Yeah, know. No. Biggest launch biggest thing I've ever launched ever.
Speaker 2:Oh, really? In terms of likes.
Speaker 1:Yeah. In terms of the initial like just reaction.
Speaker 2:I will say, it feels like this whole project, there's been so much pent up demand. Like every time Yeah. There's murmurs, it feels like people are like very anxiously awaiting this this thing.
Speaker 1:Yeah.
Speaker 2:Which is interesting because there are established I mean, Cloud Code is, I guess, established in the AI world. It's like, it's been around for two months.
Speaker 1:I think it's a combination of to be honest, a lot of stuff we release ends up feeling this way because I do not know how other companies do it. I just leak information left and right. Like, as I work on it, I share it, I like post stuff we haven't even implemented yet. I like I'm so just watched how I do everything and I can't operate any other way. Whereas most companies like build in silence and they like do like a big Yeah.
Speaker 1:Announcement and nobody sees coming. But I think given that that's how I just I just can't operate any other way, It naturally creates like this fervor around it, where everyone's like kinda watching it and they're like Mhmm. Keeping up with it and they're like kind of anxiously waiting for me to say like, okay, go. So that's one that's part of it. So I feel like this happened a few times like when we did SCV three felt this way, OpenAuth felt this way.
Speaker 1:I
Speaker 2:remember being on the other end of it. Yeah. I remember being one of the people in the crowd, like, yes.
Speaker 1:Just just being infinitely Yeah. And it always gets to a point where I'm like, can't I gotta stop doing this and I like just like the final week I go super silent and then I release it.
Speaker 2:Yeah. Yeah.
Speaker 1:That's how it always goes. It's that multiplied by the fact that this is and I was telling the team yesterday like, it feels so good to finally build something where I'm like, 100% of the audience and reach we've built up and worked hard to build up over the last five years
Speaker 2:Mhmm.
Speaker 1:Can be tapped into for this thing. Like, this is something that Yeah. Truly anyone can use. So it's combined with that and then combined with, like, the fact that AI is like everyone's kinda really interested in figuring out AI workflows right now. So I think all of those being combined together is just making this much bigger than anything else.
Speaker 2:Yeah. And it makes it so exciting to work on. I mean, I just like being at this moment in time in these careers that we have, like being software developers, I feel like it's a very critical couple of years where things are definitely changing. Like, you can say what you want about LLMs. I know there's still a lot of like it's got that MongoDB thing where people like hate it for reasons that were like years ago.
Speaker 2:Right. Like, whatever you think about LLMs, they're at the very least, they're gonna be very good at coding. Like, they're already very good at coding, and I feel like it's going to change software development in some meaningful way over the next couple of years. It already has started. So I think being at this point in time and working on something at that intersection where you're actually getting to help shape what it could look like and think about what it could look like, Just super fun.
Speaker 2:Maybe we're way off. Maybe, you know, in two years things change so much because the models are different and Mhmm. Whatever. But it's just fun to be in this moment kind of working on this problem when you've been doing this a long time. Like you've been working as a programmer for fifteen years.
Speaker 2:It's just fun to kind of be on the front lines.
Speaker 1:Yeah. I don't know if you saw my post where I said I finally feel like I'm the right age for one of these big shifts because
Speaker 2:Mhmm.
Speaker 1:Everything you know, I think I've technically lived through a few big things, right? Like I technically lived through the internet thing. Was obviously way too young for when like Yeah. Internet took off. I lived through the mobile thing.
Speaker 1:I lived through, I guess, just like I would probably count just the SaaS thing, like, just like the era where like a million SaaS companies shot up to like a billion dollars.
Speaker 2:Mhmm.
Speaker 1:And in all of those moments, I was literally too young or even if I was like working age, I was too stupid where even if I was trying to like do something in it, would just end up working on something dumb because I couldn't see that. I think when these things happen, there's like so many things that you can do and 99% of it seems like a great idea but is kinda has like a obvious ceiling. I would have just found myself working on one of those due to inexperience. But now, there's finally a shift happening where like I have enough experience to be like, I know I can identify the nine out of 10 good ideas that are actually bad. So I can actually work on a thing that makes sense.
Speaker 1:And I have like the reach to like actually, you know, market it, distribute it. And there's something like, just just like a skill, experience, like build a good product. Mhmm. It like everything is finally aligning where in the past, it always felt way more like, oh, that's like other people doing that and I'm like kind of trying to like ride coattails, you know.
Speaker 2:Yeah. Yeah. It's a good product. Like we we made a good thing. It's really like Yeah.
Speaker 2:It's impressive in such a short amount of time. Like it was less than a month ago, started rewriting. Right? That feels like forever ago when you think of just the shift that it's taken. I mean, the direction it's gone.
Speaker 1:Yeah.
Speaker 2:Do you find it weird or do you find it like crazy that I think we work well together and like we started a podcast way before we had ever done anything together.
Speaker 1:You know
Speaker 2:what I mean? Like, do we work well together, like programming wise because we had this going on?
Speaker 1:I I mean, definitely helps, but there's some weird overlap that is uncommon. And I would even say that the overlap is actually more true with Frank and Jay than it is with me. And it's something that I had to like learn. So, my approach to product forever was to like, just get core stuff working and like never polish it or make it feel really good. And I would like kind of be annoyed doing those little details because it felt like a waste of time.
Speaker 1:Mhmm. And that totally changed when I started working with Frank and Jay because they like made me see it totally different way and I understood that we're just I I understood all the reasons why that was wrong. Like you need to take polish stuff and put out stuff that's like really well tested and handle edge cases. And you naturally go that way and I don't think most people do, especially most engineers. Like I think most engineers are like kind of like how I was originally.
Speaker 1:And that can be really frustrating when that doesn't line up on both sides because like, everyone if you don't believe the underlying reason to like go that intense on some of these details, like, it's just gonna feel very annoying to work together. Mhmm. So I think that's just like a like a random coincidence or gift or maybe the explanation is like you've, you know, built stuff and have learned the same lessons. But yeah, I think we work together in like weirdly coincidental nice ways.
Speaker 2:Yeah. Yeah. I just I was just thinking about this morning trying to figure out like like which came first. Or is it just a coincidence that we enjoy talking to each other and have to have a podcast and then also happen to work well together? Know something about it.
Speaker 2:Like the programming thing could have been totally I'm just trying to remember when we first like even worked on the same repo together. Did it was it super weird to go from it's like when you meet somebody in person that you've only DM'd or something. It's like, oh, now we're writing code together. This is weird. I don't even remember.
Speaker 2:I guess it would have probably been rebase. Shout out.
Speaker 1:Yeah. I think it's funny because I I to me, it just feels kinda like a miracle because I it's so easy for me to forget this, but my entire career, every single day's emotion was I'm getting work done but it's like so annoying and difficult and I wish everyone would just listen to me. And I was like always just like like constant friction. And then at some point when I, you know, started working with Frank and Jay, that just totally disappeared. And I forgot that that's what work feels like for most people.
Speaker 1:And that was just me finding like my right fit with other people. And the work I did was suddenly overnight like 10 x better. It's like I was always capable of that, it just I wasn't in the right environment. Mhmm. And if I had to like pinpoint specific details, I just don't it's the feeling I get when we all work together is I feel like everyone just goes off and does their own thing and doesn't really need a ton of help and doesn't care to like be overly opinionated about stuff outside of that.
Speaker 1:And even when there is like stuff that needs to be fixed, wanted to we very quickly like, even like, I think this is the feeling I've always felt working with Frank and Jay is even when there's a conflict where we have a disagreement on how something should be done, which one happens rarely because we're almost always on the same page. But when it happens, it's resolved super quickly. Like, within ten minutes Mhmm. We've we've worked out who's actually right. And everyone's like fully on board and not like That's
Speaker 2:a skill.
Speaker 1:I'll like I'll say they're right but I'm still annoyed under the hood and like, you know, I'm gonna harbor this Which is usually how conflicts are resolved. Someone wins Yeah. But not everyone's on board. Yeah. So just those are the real dynamics I've gotten so used to but it is rare.
Speaker 2:It's very very common on a team for there to be those things that drag on and there's like multiple meetings to try and figure out who's right
Speaker 1:and Yeah.
Speaker 2:What's the direction that should be gone. And then, yeah, no one even feels good at the end. There's not total buy in on thing.
Speaker 1:And I I I just take this so for granted, but it's very normal in a company for the whole like engineering team not to agree with the direction for like a year and to still go that way because they're like disagreeing, like whoever's in charge has a different view. They might still end up being right, but they don't actually have buy ins. So it's just like every day is is friction. Mhmm. These are like the details that make me like when I think about this and remember that situation I'm in, and I imagine adding someone new to the team or like starting the company and joining another team.
Speaker 1:All of that is at risk and that like scares me so much, which motivates me to like try to like make the current situation continue to work. Mhmm. Yeah. Because it is rare.
Speaker 2:That makes sense. Well, I hope our podcast listeners have enjoyed listening to us just pat each other on the back for like twenty minutes. It's been fun for me. I mean, I'm enjoying it.
Speaker 1:Yeah. But I know I I but to to continue patting ourselves on the back, it is a good product. I think people literally open it and they're just like, what the heck? Like, this is crazy. Mhmm.
Speaker 1:And I'm really just proud and excited to finally have stuff out there where people have their reaction to.
Speaker 2:Yeah. It was not a good feeling before the release. Like, it was not a good feeling when we were completely dormant like a week and a half ago. Mhmm. Not dormant, like we were in that branch and it was just very confusing from the outside.
Speaker 2:Yeah. But it felt a little better when we kind of like we merged that branch into dev and it became like there are actual releases of this thing. And then after yesterday, man, I just feel so much better. Yeah. Just feels so good to actually have it out there and just be using it.
Speaker 2:I'm using it on all my work right now and it's just it feels good. I'm actually installing it on my machine instead of just running the TypeScript file directly.
Speaker 1:Yeah, exactly. And I know, I I woke up today and there was like a million messages that were all mostly good. Just DM some like people we know, like well known people being like, hey, this launch was awesome. Congratulations. Like it's it's always a nice day.
Speaker 1:I'll I'll share some of those guys. I I haven't had time to screenshot everything, but I'll I'll share some of those with you guys later. But yeah, this is fun. But now we have, you know, another mountain of work ahead of us.
Speaker 2:There's just all yeah. There's always things that could be added. It's it's interesting that like, there's kind of this assumed roadmap where you're kind of like coming into a space where there are a few existing tools. And you just imagine like we gotta have all these features to like be at parity with everything else and then we can innovate on stuff. But then it's like and we don't need all those things.
Speaker 2:And you can kind of just like let your own open source like noise Yeah. The signal coming from open source kind of drive what's most important. Like, there's features that are pretty big features missing from something like Cloud Code, like permissions. Like, we're just right now, it's full auto. You can't there's no there's no front end for the permissions yet, so you can't, like, actually opt into that.
Speaker 2:And we've had I guess we've now had some people. But it's kinda like wait and see. Like, how long before somebody complained? Because I love it without the permissions. As soon as we have permissions, I'm we're gonna have an auto accept feature because that's what I'm turning on.
Speaker 2:I don't want it to ever change.
Speaker 1:Well, I I I actually given how and this is what's awesome about launching and release like, yes, like, have bunch of people that use it, but it's such a crazy source of data that just reorients your priorities and your perspective on things. Like, if you asked me a month ago, I would have been like, yeah, the default has to be really conservative and permissions prompt for everything. And then we have this version that doesn't do that and so many people are fine with it. And when you like run Cloud Code now, and I go through the initial setup, it feels horrible. There's like Yeah.
Speaker 1:A million danger, danger, accept this. Are you okay? And it's it's just like, just not a good feeling. So now given like, you know, the the information from people that, yes, people carry out permissions but it's not most people. I'm way more inclined to say that default should be off and people can opt into into something more shite.
Speaker 1:And if you think about it, it's like, yeah, people can make people say, that's like an unsafe default, like, know, that's gonna lead to bad things. But and that's fine. You can have tools that optimize for different things. Power user tools do not hold your hand every step of the way. Right?
Speaker 1:Like, everyone talks about how Windows is like your mom and it's always like, are sure
Speaker 2:you wanna do that? Are sure you wanna do that? Is that okay?
Speaker 1:Mhmm. Mhmm. Linux, if you do RMRF, it's just doing it, you know? Yeah. So there's just different tools for different people.
Speaker 1:And I think we're building for ourselves and I think we yeah. Like, we're power users and we don't want all that stuff in the way by default.
Speaker 2:Yeah. Yeah. Yeah. It's very nice. But then it is like, there is still just a ton of stuff that it feels like we do need to add and I just so badly I wanna play with new ideas and like, are there ways to get more performance out of it?
Speaker 2:I mean, there's just there's a constant like treadmill that all these different tools, the teams are thinking through and trying to because I just feel like the models are good enough now that there's still a lot of improvements we've made just on the tools. Like the tool side, not like tool calling tools. I just mean like our tool, open code, Clog code. I feel like there's so much room still for them to improve and to do new novel things with the models to to increase the effectiveness. So it's yeah.
Speaker 2:It's this cool unknown infinite roadmap where like, you you don't even know. Could it be two times better? I don't know. Maybe that's a stretch, but
Speaker 1:Yeah. Yeah. Frank was saying something roughly roughly the same. Like, he's like, I think we could do two times better. Probably not 10 x better in terms of performance Yeah.
Speaker 1:But but two x seems like reasonable. Yeah. There there's so many ideas given our positioning that we can we can play with. I'm mostly still excited about just the core packaging of of it like we're Yeah. We executed well on that.
Speaker 1:We're gonna continue to execute well on that, you know. It's I don't see how anyone catches up to that because it's like a pretty unique skill set. Mhmm. And I didn't have it five years ago. Like, I I gained it over the work I've done for the past five years.
Speaker 1:So, yeah, some little details that are like pretty delightful for people to find out. Like And it's cool to see people recognize it and call it out. We obviously put a lot of work thinking about configuration and how to make the whole initial flow not require any config file. And it's a great feeling, you open it and it guides you through and it's ready to go. Mhmm.
Speaker 1:Almost no CLI. CLI tools, I don't think about those things generally. And so a few people were like, wow, no config, that's awesome. So it's cool to see like people recognize the effort.
Speaker 2:It's been so long since I've run it from scratch. If you already have like an API key and you're in, does it even run through that wizard at the beginning No. Doesn't. To set up providers? It's just you're in.
Speaker 2:Yeah. That's awesome.
Speaker 1:It just works.
Speaker 2:So it's literally zero config if you have API keys like Anthropic API key in your end.
Speaker 1:Yeah. Yeah. And these are the thing going back to what I said earlier about Polish, like, yesterday I spent a bunch of time printing out really good not really good. I had to put more time into it. But decent error messages for when your config when an error in your config.
Speaker 1:And this is the type of thing where in the past I've been like, who cares? Like, people will figure it out and they will, you know, fix the config file. Like, I don't need to waste time like doing this random feature. But now I've learned a few things which is, for every person that runs into that and brings it up to you, there's probably 10 people that ran into it and just never figured it out and gave up. And two, when people bring it up, it's annoying and it's a distraction.
Speaker 1:It takes away from your time. So now I'm like, the more I can help people help themselves, like, I'm gonna do whatever it takes. It's just gonna pay back in in the long run. Mhmm. And this is not where my instincts were tuned, like, five years ago.
Speaker 2:Yeah. I'm I'm so glad you nailed that that opening sequence because it does it just makes a huge difference for people's first impressions. Yeah. Yeah. I'm feeling good about about where it's at.
Speaker 2:I mean, there's obviously there's like a million long tail things with all different providers we support because this is a tool that you can use a bunch of bad models. I mean, models that probably won't be very good at this task. But if you're doing like I mean, like this morning I had I opened open code to just make some changes to some little shell scripts that I run for, like, managing TMUX and stuff. And in those use cases, like, you probably can use cheaper, simpler, faster models, and it's it's probably fine. So there's, like, there's use cases people wanting to use this kind of stuff like in CI pipelines, just generate Git commit messages or whatever.
Speaker 2:Yeah. There's always like weird edge case usages that having a configurable model is actually a meaningful feature. Like, that that matters and it just opens it up to be used in so many other contexts.
Speaker 1:I think that's why it's always good to make stuff hackable because you will be surprised with how people use it and it usually gives you ideas. So we had someone as a PR, I'll work on this after a after the podcast. But, they want to for us to be able to detect when a model does not support tool use and not submit the tool schemas in the request because in the model there's errors. Mhmm. Which is weird because like, what the heck?
Speaker 1:Like, what do you mean? Like, what are gonna do with a model that doesn't have tool use, like, OpenCode's useless? But they have a model that they like and they just wanna use it, they just wanna talk to it. And because it's not like from a big company like OpenAI or Claude, there's not a great interface for just chatting with it. It's like a crappy interface.
Speaker 1:Mhmm. So they're just using OpenCode as a normal client for an LM, like a normal LM interaction because they want to do it in the terminal.
Speaker 2:So Interesting. Yeah. Not at least because I would have thought about it. But it makes Yeah.
Speaker 1:It makes total sense once once you see it, but we would have never thought about it ahead of time. But like making this all customizable and and hackable, you start to see people do that. The other thing that was kind of exciting for me to see is I love seeing like the boring parts of the world pop up because that's when I feel like we've really like tapped into a wider audience versus like the very internet oriented startup y people that we typically are around.
Speaker 2:Mhmm.
Speaker 1:A bunch of people are like, oh, cool. I can use this with my, like, corporate approved Azure credentials. Because like, they have like a lockdown laptop and they can't use any LM besides the one their company gives them. And now finally, they can play with this stuff. Mhmm.
Speaker 1:So it was cool seeing just all those weird weird use cases pop up. Because in my head, I'm like, oh, yeah, everyone's using Cloud Code. Like Yeah. We can all use Cloud Code. But a lot people, like, literally can't and probably, you know, it's so early as big as all this feels.
Speaker 1:The majority of people have not used Cloud Code. So, yeah, it's cool to see those people start to see this for the first time.
Speaker 2:Yeah. It's early.
Speaker 1:It is early.
Speaker 2:It's very exciting. I need to be a little less excited. I need to sleep. I need to sleep good. Wait.
Speaker 2:What yeah. What else is going on? I don't know. It feels like thing that's been pretty consuming. I've been like head down and I've just got Yeah.
Speaker 2:A lot going on. So what's what's going on on the internet outside of or the world? I don't know. Entertain me, puppet.
Speaker 1:I'm just I'm just I'm trying to I'm trying to
Speaker 2:You've been you've been head down too. I mean
Speaker 1:Well, okay. So here's the thing. You know what? I I can't bring this up every episode. My whole Twitter rework, I've continually be shocked at how effective it is.
Speaker 1:Because Mhmm. Yesterday Jay was like, oh man, these like mid journey video things are really cool. I'm like, what are you talking about mid journey And yeah, mid journey launched like a whole video generation thing. And like a month ago, I would have seen that in like four seconds after they launched it. And I went like to like a whole twenty four hours without seeing a single thing about it.
Speaker 1:They and they launched the same day we did.
Speaker 2:Oh, really? Yeah. I hadn't heard anything. But I I guess I just haven't been on Twitter probably, but because I have not done a good job with, like, curating the Twitter feed. It's mine's still a disaster.
Speaker 1:Yeah. I think I'm trying get a few other things that popped up. There was that whole I think Carpathi
Speaker 2:Oh, yeah.
Speaker 1:Gave that talk. I didn't watch it, but everyone was talking about it. They were like, software three point o.
Speaker 2:All I saw was he's jacked. Yeah. What was the talk? No. No.
Speaker 2:No.
Speaker 1:That Jacked was a that was was John Carmack.
Speaker 2:That was Carmack. Yeah. Sorry. Okay. So What what was Carpathi's talk?
Speaker 2:I didn't hear about this.
Speaker 1:I think I have to see I think he gave a talk just on his vision for software moving forward given AI. Mhmm. I think he's like trying to cat call it software three point o and like define the shape of what that looks like. I I haven't I haven't looked into it at all. He is the one that coined vibe coding.
Speaker 2:I know. Here's so here's my question. I don't I don't know a lot about Carpathi. I know he's very well respected. I know he's like an AI god.
Speaker 2:The Tesla self driving Mhmm. All these places he's been. He's amazing at it. Is he, like, actually plugged in as a software developer? Like, does he make things that are meaningful or is he just like a more data science y kind of guy?
Speaker 1:Well, here's a funny thing that happened. He posted a few months ago being like, he was trying to build something, like build like an actual web app thing. And he was like, really overwhelmed. He was like, oh man, like, what what do I even start with? Like, where do I host it?
Speaker 1:Like, what do I do? Mhmm. And I found it hilarious because every DevTool CEO in the reply begging, carpet please, please use my thing, please.
Speaker 2:Uh-huh. Yeah. This is the
Speaker 1:best way to do And I I was so proud that I had zero instinct to like get in the reply to beg. That reflects I think the state of the overall industry right now, which is, yes, he understands stuff really well and he's got a really great perspective and there's a lot that we can learn from. But how good can it be when he doesn't understand when he's not an expert in this area? Not only is he not an expert, he's a beginner. Right?
Speaker 1:Understanding what day to day building something from zero to one and launching it looks like. So most of the tools that have come out are coming from like that part of the world. Like, yes, it's coming out from software, but it feels still like external people building tools for us, the people that actually like would use them. Mhmm. So yeah, I don't I don't know how much I would put stock in his vision for.
Speaker 1:Again, I have to listen to it to see. But it definitely feels like that part of the world talking to us again.
Speaker 2:Mhmm. Yeah. The the whole like vibe coding term is forever just tainted for me. Like, there's no way to redeem that phrase. It means something to me and, like, I feel like some people try to, like, oh, well, it can just mean using AI assistance in your programming.
Speaker 2:And I'm like, nah. It's a very different thing. Yeah. But any anybody that says, I just have this visceral reaction to somebody saying, like, can I vibe code with it? Is it a vibe coding tool?
Speaker 2:Like, no. You cannot. That is not what it is. It's a a super a power user tool for experienced programmers. Okay?
Speaker 1:What was funny is that the only time I've ever vibe coded was with OpenCode because like, I was like, okay, let me like actually try doing this. Like, everyone's
Speaker 2:I've not actually tried. I should try.
Speaker 1:Yeah. That that Tetris thing that I I shared with you guys, like
Speaker 2:Yeah. That was impressive that it it just made all that, You didn't even look at the code.
Speaker 1:Just Nope. Still not that thing. I literally never opened the file.
Speaker 2:That is vibe coding.
Speaker 1:Yeah. So
Speaker 2:you did you did the thing.
Speaker 1:This is that was my first time actually trying it and it felt my feeling around it was pleasant. It was a positive feeling Mhmm. But it wasn't work. It wasn't work.
Speaker 2:It's a different it's just a different thing. Was playing.
Speaker 1:Yeah. It was like Yeah.
Speaker 2:Me It's entertainment.
Speaker 1:Playing with Legos is entertainment. Yeah. Does sound fun. I you know, I was like I've also been trying not to like ever since I got like, you know, for for a long time I only had my desktop. And a year ago I got the MacBook Air.
Speaker 1:And that a side effect of that was like, just started working in bed before sleep. Before I couldn't, so I just didn't. So now I'm trying to like, wean myself off doing that. So I was like, okay, I have my laptop but let me like, try to wind down and I'll just try to vibe cutting thing. Was really nice.
Speaker 1:Was waiting for Liz to finish her shower so we could hang out. Sitting in bed just like messing with it, had music on. It was a nice relaxing like meditative experience, like nothing I did mattered. It wasn't important Yeah. For it to be good.
Speaker 1:Uh-huh.
Speaker 2:You can turn off your brain. Yeah. Just create something.
Speaker 1:Yeah. So it was like purely play, joy, fun. And I think there's a lot of value in that. It's not whatever Yeah.
Speaker 2:You know
Speaker 1:however we're trying to use it, but there's something there.
Speaker 2:Yeah. Like create a SaaS business. Right. I think the thing that I would like to use it for now that I've thought about it I mean, play sounds fun too, like relaxing and doing something fun. Should try that sometime.
Speaker 2:But the thing I'm thinking of is like making little utility like Mac apps like in
Speaker 1:the Mhmm.
Speaker 2:The tray at the top or making like browser extensions that are just for me. Like just making things that I don't know how to make a browser extension, but if I could make one that made Twitter better or something, you know what I mean? Like Mhmm. Highly customized little like applets for those kind of context where I'd love to just like pull up in my tray and get x y z this little utility that could show me some information I want and just use vibe coding to like make that kind of stuff. I feel like it'd be good for that use case.
Speaker 2:But that's I have no experience, so I might
Speaker 1:be I think we all need to like realize we have this thing we can do now. We just need to like identify it till we actually do it. Like
Speaker 2:Mhmm.
Speaker 1:We should fix these micro cuts that we face all day because they're not difficult to fix difficult to have an LM to build something to fix it and they don't they don't need to scale. We're not trying to build a business out of it. Right. Yeah. I think Lee Lee Rob had I'm like jealous because I was thinking of like a similar term but he coined it first.
Speaker 1:He was like, this is the era of like personal software where it's like, it's software just for you. And he like wrote a really nice thing like kind of detailing that. I think the part that's confusing is I think it is true. We are, like, in this new era where we can all have personal software. I don't think 99% of the world is gonna do that.
Speaker 1:I think it's still gonna be like a small percent of people that do. And I think a lot of business are being built around the idea that everyone's gonna have personalized software. I don't think that's the case. I think most people don't really wanna do that just like how not everyone cooks. Even though we can all cook and there's infinite machines to make it easy.
Speaker 1:But, you know, for myself, like, yeah, like I think I I need to, like, use this more and, like, build these little things that could fix a bunch of my little problems.
Speaker 2:Yeah. It it kind of makes me think, like, and I've always enjoyed this pastime of like tinkering with my environment. Well, I enjoy the effects of it. I don't always enjoy how long I spend on it. So I feel like if I could use these models to just take over my dot files and like I just describe what I want, that actually sounds kinda nice because I don't I don't think it's the actual typing that I enjoy when it's configuring my environment.
Speaker 2:It's just the getting everything just how I want it. Yeah. I feel like these models could just be like, I just only wanna know what my dot files look like. I just wanna literally let them run wild and I just tell it exactly what I want. That sounds nice.
Speaker 1:Yeah.
Speaker 2:I'm gonna try that.
Speaker 1:Yeah. I actually did that the other day with trying out Weyland on Linux, which I've tried a bunch of times and I've always like failed because the whole Weyland Nvidia stuff is tricky. But OpenCode was good enough at that point where instead the the last time I tried that, I did try to use AI but it was always like, tough to clone, clot clot clot, copy paste, that whole workflow. Now, I just let OpenCode access my whole system. And it like fixed all the issues and I actually got like a very decent working setup.
Speaker 1:Still stuff didn't work where I had to like bail back into my old one, so I still failed to switch over. But, yeah, it was like, great because I could try this out and it can like debug all these things. And I don't really care about
Speaker 2:It's knowledge I don't care to Like Yeah. I don't I don't really care to to know some of these things like
Speaker 1:I still ended up learning a bunch because I was like watching what I was doing.
Speaker 2:But Yeah. Sure. You're following along. It is really effective to watch these threads. Like, if you just think of the sessions that you can share through OpenCode on the web Yeah.
Speaker 2:Reading through those is very informative as to like it you can see all the the cases it thinks through and like sometimes it's just like really nice to see how it thinks and how it works through the problem. I I love it. I think it's very educational.
Speaker 1:I'm I'm gonna write up this thing later, the thing I was telling you about. And this was such a funny workflow because I'm like, this is so crazy. We're in a time where this workflow exists. And it seems so stupid but it kinda makes sense. Liz was like implementing a feature and she got to the point where usually she has to stop because it's not she mostly does a pure UI part and then like doesn't do the data wiring or the logic.
Speaker 1:And she was like It's
Speaker 2:like me. Dax. I did the I did the front end. Made it pretty. Yeah.
Speaker 1:She was like, oh, okay, like, I wanna this needs to actually be wired up. And I was like, oh, you know, you can just tell OpenCode, like, tell it to do this and then tell it to do this. Like, I knew how to break it up into two into two tasks. Mhmm. So I, as a software engineer, gave her roughly the prompt and she went and talked to the LLM Yeah.
Speaker 1:Execute and had it execute it. Then she used the share link, sent it back to me and I was able to review and she was like, did it do it right? And the reason this is a little bit different than adjusted PR is what the LM does is every time it edits a file, makes a change, it like kinda describes like why it's doing that. It's like, okay, now I'm gonna do this and then you see the diff. So I was able to quickly look at it and be like, I very quickly understood that it went down the right path.
Speaker 1:Like, this is what I expected and Yeah. Reading the text and looking at the diffs.
Speaker 2:It's this really nice like verification of work. Yeah. Like it's better than a code review. It's it's like
Speaker 1:It's better than a code review. Yeah.
Speaker 2:Mhmm.
Speaker 1:And it's better than a pure PR because it has that those like annotations of like the thought process Yeah. And everything else. So Mhmm. Yeah. Initially the share page, I was always like, is this that useful?
Speaker 1:But then seeing that at that moment, I'm like, this is a crazy workflow because this is still having an experienced software engineer in the loop, but getting more done than we would have otherwise with like a very minimal amount of my time. And the output quality is still still great. So Okay.
Speaker 2:Can we just that little use case just made me realize we need to do like pair programming in OpenCode. Like have an invite command and you basically can like both be in the same session.
Speaker 1:Mhmm.
Speaker 2:And so you could just like read through it and send your own prompt. Yeah. I don't know. Maybe no one would use that, but I just it just occurred to me that we could totally do that the way we have things architected.
Speaker 1:Yeah. Like the share page can have like write access. We can like also also
Speaker 2:It could be on the share page or it could be that we're both in our the terminal. Yeah. And we're just connected to the same session.
Speaker 1:That that's another funny thing because I also SSH'd into her machine and just opened up T Bucks and like, it was like effectively a screen share over terminal and she was like at a coffee shop.
Speaker 2:Yeah. You got the same yeah, you got the same effect.
Speaker 1:Yeah. But yeah, we we can we can make that flow like much more native. But yeah, so I'm like, this is really I'm like, can work look like this where it's a product person still working with an engineer but like Yeah. The process looks pretty different than the engineer literally writing the code.
Speaker 2:I'm getting obsessed with this idea. Why is this the low sleep inviting someone into your session to like help you drive the session forward? Junior engineer using open code asking senior engineer to come in and help him.
Speaker 1:Yeah. I mean, that was the that was the original for me because we had some bigger companies talk to us and say, we want junior engineers to see this transcript in the PR so they can learn Mhmm. From the senior engineers and how they're using these tools. Mhmm. That was one of the big things that convinced me to even do the share page.
Speaker 1:So I was like, okay, see how bigger teams like, they think like this, they think about growing talent, like people teaching each other. So things that aren't normally on my radar, but I I totally see it. Yeah.
Speaker 2:Yeah. I wanna do this now. I don't know. Maybe no one maybe if you're a podcast listener and you use OpenCode because I think most of the OpenCode users listen to the podcast, not most of OpenCode users. I know a lot of our listeners have heard about OpenCode through the podcast.
Speaker 2:Maybe you use it and you could tell me this is a terrible idea or yes, I want it. That'd be cool. Or just go to GitHub. Just tell us. I wanna be able to invite people across the internet into my session to run stuff on my machine.
Speaker 2:It sounds kind of worse the more I say.
Speaker 1:To prompt an AI to run anything on the machine.
Speaker 2:Yeah. Right. Delete everything in the system root.
Speaker 1:Yeah.
Speaker 2:I just I just had it just it's just occurring to me how much we are placing a bet on the terminal at this point. Like I know. Developers get comfortable in the terminal, so they'll order our coffee and use OpenCode. Yeah. What's going on?
Speaker 2:It's a Boom. And it's
Speaker 1:it feels like this like, if you take a step back, feels like the dumbest bet because, like, it just feels so ancient and,
Speaker 2:like Yeah.
Speaker 1:There are people that I mean, like, Wes Boss had a post being, like, why the heck do people prefer running stuff in the terminal?
Speaker 2:I know.
Speaker 1:Yeah. And I found it funny because he posted a screenshot of Cursor or something being like, look at how much you can do, like
Speaker 2:I saw that.
Speaker 1:Yeah. Why would you do it on on the terminal? And I had two thoughts. One, I'm like, that UI looks insane. Like, there's so much stuff going on.
Speaker 1:It's really overwhelming. Like, it looks bad to me that there's so much. But also, like, you can do all that, you know.
Speaker 2:Yeah. There was nothing in that screenshot we couldn't do. Yeah. It's like,
Speaker 1:how do we want to You know?
Speaker 2:Yeah. Exactly. At the end of the day.
Speaker 1:Yeah. It's so it does feel stupid in a way because it's like, feels ancient, but somehow it makes sense and it's like resonating, which is also funny.
Speaker 2:Yeah. I mean, it it's like Yeah. It goes back to the first mentions on this podcast of like you talking about it just needs to interact with the file system. This is pre Cloud Code.
Speaker 1:That's true.
Speaker 2:Like, we just need I want cursor, but I don't wanna use Versus Code. I wanna use my current editor. And why not just have the file system as a source
Speaker 1:of truth?
Speaker 2:I gotta pull up that episode. I gotta remember exactly how that conversation went because it's so That's exactly how things have trended. I mean, I say trended in our bubbles on Twitter. Yeah. Like Claude Code has kinda had a moment and I feel like people will just continue to question like, do I need to download Versus Code to use AI tools?
Speaker 2:It just seems heavy.
Speaker 1:Yeah. The other thing I brought up yesterday was poor Frank because we all love the terminal. Frank is the kind of person whose ideal UI is a button that is like 10 words long that does a specific task. That's like, get this button pushes to production and also, like, you know, just like some big butt, he wants to click on it, like, that's just how. Yeah.
Speaker 1:Wants some personal software.
Speaker 2:He just needs to write his own with OpenCode. And, like,
Speaker 1:he just forced not with just OpenCode. SST dev multiplexer, like, he had to, like, deal with using that, the terminal, and then, like, OpenCode using the terminal. He's, like, he just wants to scroll and like click on links, you know.
Speaker 2:Oh, it's funny. Oh, man.
Speaker 1:I have to explain to him, like, like, we can't like, when a link wraps in Versus Code, like, there's nothing we can do. Like, we can't make it clickable. Have to, like, click and drag. And to him, it's just, the worst feeling ever.
Speaker 2:Oh, man. The perception I had of Frank outside from the outside versus, like, getting known a little bit better and seeing you guys interact, totally different. Yeah. And I I just feel like I would have never guessed that he's the button He
Speaker 1:he's it's just like a mix of contradicting things. Like, just is just like the dumbest user. Like, he just wants the most basic thing.
Speaker 2:Mhmm.
Speaker 1:But he does the most like comp like his work is just so like intensely at times. Like out of all three of us and like, he just does like all this crazy stuff but it's that. And then you mix in his like whole outside of work personality where did you see what I posted that he his five k time is seventeen minutes?
Speaker 2:Yeah. That's insane. Yeah. Like, I had no idea.
Speaker 1:Jay, who's extremely in shape, works out every single day at Barry's boot camp, runs like really in shape, is like 20. His his time is like 20 Crazy. And Frank is three minutes faster, which is like
Speaker 2:Yeah. I've never seen Frank say something that makes me think like he's into fitness. Or he's you know what I mean? Like, it's not a thing he talks about or Yeah. Has that, like, persona online.
Speaker 1:He he's just one of those people that I mean, so going back to the the tie with everything for him is if you give him a task, he will just like go to the to the extreme end of that task without even questioning it. Like, you know, like Mhmm. That's why this model does this models. Dev, we Like, I joke that it's a frank task because the task there was, like, go find all the providers, all the models, every single piece of information and like Yeah. Write it into a document.
Speaker 1:Like, will just go and like beast mode that and like not get tired. And just like, mindlessly do it until someone tells him stop, you're doing too much. Yeah. And now you apply that to like him working out and like running, like it's a it's the same thing. Like he just like pushes and doesn't overcomplicate it with his brain.
Speaker 2:So I love models.dev by the I have nothing to do with this. I can say I just love it. It's well done.
Speaker 1:It's yeah. That's our it's just funny. Like, we just always end up having these, like, side side side projects that Spinal.
Speaker 2:To the side to the side. Yeah. Yeah.
Speaker 1:We were talking yesterday about how, like, okay, we're, like, really gonna focus on OpenCode for the time being. And I was like, yeah, we're gonna only focus on OpenCode. And Jay was just laughing, like, we all we never do that. Like, we always like, we're already doing models. Dev.
Speaker 1:Like, it's just a
Speaker 2:Yeah.
Speaker 1:We just always end up Oh, man.
Speaker 2:The favicon on models.dev is so good, Jay. Good work. I love the styles he put he put into models.dev. Yeah. I have to pee really bad.
Speaker 2:Okay.
Speaker 1:So we'll end up.
Speaker 2:I think we should just call it. Yeah.
Speaker 1:And we'll go back to work.
Speaker 2:Back to work. Alright. Thanks, Dax. See you. See you.
