Docs in MCP, Elon and Plinko, Finances with Dax, and a Big Announcement
Maybe Chris could like cut out the middle part where I was like, maybe we'll just be done after thirty, you know. Just call it. Okay. Now, hey. We weren't recording yet.
Dax:Oh, hey. Riverside is a new process now. It, like, opens another tab
Adam:I guess. In the background? I yeah. I feel like that guy who's like, I like the old Facebook. I just every time it changes, like, when you click record now, there's like four little options you have to choose from and it's on my teleprompter so it's so tiny and I can't read any of them.
Adam:And I'm like, I don't know. I just wanna record like I used to. I just wanna hit the button.
Dax:Yeah. Now now it opens up like a background tab for you to for it to upload. That's what it
Adam:did Mine's for not doing that.
Dax:Maybe Yeah. That's just
Adam:so you don't close the tab and then like as the host, I know to upload my anyway, we're getting yeah.
Dax:How are you? I am good. Just preparing for all this stuff next week. I think it's almost there, but it's gonna be busy.
Adam:Like, are you mostly preparing for React Miami things like party?
Dax:Yeah. The party, some of the just like the tech we need next week, some of the just logistics of all the different things that we're filming next week.
Adam:So Mhmm.
Dax:I guess we can talk about it here, like, it's fine. But so we're filming this documentary on Monday. It's gonna be a mockumentary and I know those people would just like some thoughts. I think we're gonna like impro improvise most of it. But
Adam:As one does
Dax:at shop. I'm just like, don't know how much we can come with ahead of time. It's just gonna be better when we come with stuff in person. Yeah. So yeah.
Dax:But I'm excited.
Adam:Well, I'm doing everything I can to prepare for next week as well in the form of wellness shots and immunity boosts. I did a sauna session this morning. I'm just trying not to get sick. Do you feel sick at all? No.
Adam:Not yet. I which I can't figure out. Like, do I just have COVID immunity? Because my my oldest tested positive for COVID. That's what led to his pneumonia.
Dax:Did you test yourself?
Adam:No. I don't have any symptoms. I feel fine. I did sleep in bed with my five year old last night who's the worst right now of us all. My 10 year old's recovering, but the five year old is in the thick of it.
Adam:It was like last night was the worst night. Mhmm. He's just coughing all over me all night, like, just in my face. Just and I just I got terrible sleep because he's coughing all night. And I just thought, like, yeah, I'm definitely getting it now.
Adam:And I feel fine so far.
Dax:Are you gonna come in and affect all of us?
Adam:Yeah. I'm gonna get there. As soon as I land in Miami, I'm gonna start coughing like, oh, man.
Dax:We'll put you in a different different hotel.
Adam:Quarantine me. No. There's I mean, there's a very good chance I'm not coming if either I or Casey starts showing symptoms in the next forty eight hours, which is I don't know. Feels like a less likely scenario that we both come out unscathed.
Dax:But anyway People are already arriving. What? Yeah. I think so Kramer is getting here today and I'm gonna see if we want to hang out, get dinner or something. Ken was complaining to me about his flight this morning, so I guess he's gonna be here today also.
Adam:Nice.
Dax:So yeah, it's like a whole week.
Adam:We've got a whole yeah. Jeez. Everybody's converging on Daxville.
Dax:I it it is crazy because I remember when I moved here, was just like, man, it'd be cool if, like, people would ever come here. And then, like, it just it's like manifested, you know?
Adam:People keep
Dax:yeah. I am proud of that because three years ago, nobody was coming here. Two years ago, a little bit more. And then the last two years have been, like, really crazy. Yeah.
Adam:Michelle's done a great job with React Miami. It's it's a good conference. I've not been to a lot of conferences, but she does a good job.
Dax:Yeah. I also love how much people are, like, get excited to come here and everyone, like, you know, thinks about their outfits and, like, it's like a whole thing that everyone, like, plays into. Yeah. That's what I do.
Adam:I think about my outfits. I pack my Todd Shelton jeans and
Dax:Well, everyone else everyone else gets really
Adam:into Yeah. Yeah. I'm gonna show up one of these one of these times, I'm gonna show up in my What color? Wearing what you wear, like an outfit like that, like a silk onesie and you're gonna be like, woah. Adam has culture or something.
Dax:I don't think I've ever seen you wearing color.
Adam:Yeah. Mostly neutrals. Black and white. Yeah.
Dax:I Mostly the same, but I have like some stuff that's in the extreme opposite to, you know, balance it out.
Adam:I don't have any of that. Yeah. I don't have anything color wow. Can't think of I have like they're all earthy tones, like maroons.
Dax:Yeah. Yeah. Same. Mostly earth tones. Yeah.
Dax:Just the tans, the dark greens. Love the dark green. I do a lot of
Adam:dark Dark green. Okay. Okay.
Dax:Yeah. So I I get it, but, you know.
Adam:We got stuff to talk about.
Dax:That's true.
Adam:First thing I wanna talk about, this was not on my mind until right before you jumped on when I was looking at Twitter. We've unexpectedly built so many tools as a side effect of building apps. We're about to go one level deeper. Is this something you can talk about on the podcast?
Dax:It's a mix of a few things. So the stuff that you're working
Adam:on Uh-huh.
Dax:The open code stuff
Adam:Mhmm.
Dax:That's gonna be cool because it'll be like the first again, for our audience.
Adam:Talked about it last episode.
Dax:We oh, we talked about it last time. Okay. We we did talk about it last time. Yeah. So just a cloud open source cloud code, Adam.
Dax:Mhmm. And we have another guy. How do pronounce it?
Adam:Oh, Kuti Tim is making it. Yeah. Kuti Tim is making it. I am just giving advice and saying, here's what I would like. Because I'm just a big Cloud Code user, so I just I feel like I have opinions, big opinions.
Dax:Yeah. So that's gonna be under the open brand. I'm talking to a company today about building our own model. Right. And I I just wanna like get really up to speed on what like the state of the art and continuous training is because it's such a problem in open source.
Dax:Yeah. So, like, we have a vague goal of building a model that's not just for SST, but creating a nice pipeline so that all of these open source tools can easily continuously train on it. Mhmm. As time goes on, you know, React Router v seven, it's just not in like, LMs do not return React Router v seven stuff. Yep.
Dax:Tailwind v four, not included. ST v three, not included. So a single model that is good at those things Mhmm. Probably useful. I don't know if it's gonna be the model that everybody like, you know, is using primarily for coding, maybe not.
Dax:But something that people can embed in docs, something that people can use as like a bot in their Discord. So I wanna like try to go down that path and to see like what what's there. The thing that I think we're maybe all taking for granted is, I don't if I already said this last time, I'm say it again. These AI companies are massive. There's so many of them.
Dax:They have billions and billions of dollars. Mhmm. Though it feels like they're just gonna solve every problem. But the surface area is so large for what AI can do that despite all of that, despite all the competition, we just take for granted that like these little hints that they're not gonna solve. And if you like dig into it, like let's take tailwind v four, right?
Dax:How in the world is a generic LM training process gonna like properly differentiate between v three and v four and like know to like prioritize v four? Are they gonna label their datasets to that extreme where they're like, just don't think they're gonna go into that level of detail. I don't think they can. I don't think it makes any sense for them to. And I've talked to a few different open source people, they all recognize this problem.
Dax:I don't think anyone cares about it as much as we do because we're like a little more growth focused than than other people. So they're not gonna like solve it directly.
Adam:Mhmm.
Dax:I figure if we solve it for ourselves, we might as well try to solve it for everyone. Bun's another one, you know, Bun's always adding new APIs and then Yep. Yep. You know, I just don't know about them.
Adam:So Okay. I can't I I've been biting my tongue. I you should do the model thing, for sure. Like you should. But there's I just I feel like this is the most obvious thing I've ever felt convicted about that no one seems to agree with, and I'm just gonna I'm gonna build it.
Dax:What is it?
Adam:I'm going to build an MCP docs server. No. We're do that. We're gonna add We'll that?
Dax:Well well, I mean, yeah, we're gonna add we're gonna give an MCP thing for our docs. We're definitely gonna do that. We've tried some stuff.
Adam:Not just for your docs. Docs. This is just it's the it's the the what's the definitely typed thing? It's just a community project that just
Dax:Oh, I see.
Adam:MCP server is pass in a framework and a version, get back, and maybe even a search term, get back relevant markdown or something. Like, that is the MCP server and it needs to exist. And yeah, it's a lot of tedium, but like, we do that all the time. We do tedium, definitely typed. Tedium.
Dax:Okay. So yes, that's that's definitely there and that maybe is a solution to this, but it still requires a lot of setup.
Adam:Well, we're gonna build it straight into open code. So if you just use open code, you're fine. That's It'll be configured directly in
Dax:the That's true. That would make it a lot easier.
Adam:The distribution problem is And then
Dax:as long as OpenCode beats cursor, then this can
Adam:Well, mean, it could be so the cursor thing, they make it really easy to add MCP servers. Right?
Dax:Yeah. I know. I know. So if
Adam:you're serious and you're running into these issues, then I don't know. Click the
Dax:So we as I said, we are gonna, like, support those. I don't know how good it's gonna perform, though, because, like, tried
Adam:great. Sorry. This just seems like a perfect that everyone when I brought this up on Twitter, everyone's like, well, there's already, like, web fetch. It's not gonna figure out how to fetch the right docs for the right version of your framework. But, like, you're using a thing, and it's in package JSON.
Adam:We have a version. Yeah. There is no reason we can't build a Tailwind, an SSD, or whatever all inside one MCP server, like a repository of markdown that we just keep updated. You know what I'm saying? Like, why is that why is that not gonna work?
Dax:It makes sense, but then like okay. So what we found works really well is so for us, SD components, giving a doc the docs for it works well.
Adam:Mhmm.
Dax:Giving it the whole source code of the component works really well.
Adam:Mhmm.
Dax:Yeah. Like, how do you make all that easy for us to do or to provide? Yeah. And the other thing is this stuff blows up the context like crazy. That's the other issue, like
Adam:Yeah. I mean, it would be a search term. I feel like it would need to be like, I wanna get a section of the docs. I don't need all of the SSTV three docs. I need the section that deals with secrets because I need to add a secret and I don't know how.
Dax:Yeah. Like I said, I just think the performance like, it's it's low hanging fruit so we should definitely do it. Yeah. Yeah. I just don't think the performance is gonna hit the bar where I'm like, oh, feel I good about people.
Dax:Because our our primary thing, the primary piece of pain that we're feeling is building or learning SSC or building some of SSC from an LM first approach is not a good experience right now. Yeah. The thing you're talking about will improve that. I don't know if it's gonna cross the bar where I'm like, okay, this no longer feels like it's a bad experience.
Adam:Mhmm.
Dax:But yeah, mean, it's it's it's so obvious in low hanging fruit that it should just be done.
Adam:I think I think it's a few things. And I'm not trying to discourage the custom model. I think fine tuning I've seen stuff recently, like, that the major versions of these latest models have not met expectation, I guess. People are saying they're getting better results. These are like research papers.
Adam:I didn't read them. I just saw what people said about them. They're like fine tuning is this like crosses the gap, makes things really good. So I think there's plenty I think lots of companies, including SST, should be fine tuning models to see who can make the best programming model or the best like, I'm a web developer in 2025 model. It's like more specific even.
Adam:I think that's that's good. I think, to me, it would be harder to convince me to swap out Cloud three seven than to swap or to add an MCP server, I guess. Like, I get that they're both kind of tricky problems.
Dax:Yeah. But the MCP server could just call the model. Oh, I see. Yeah. That was one other
Adam:I feel like a a Docs MCP server combined with LSP diagnostics embedded in like, I'm I'm just speaking OpenCode here combined with Sourcegraph.
Dax:The Sourcegraph tool was crazy because I didn't know that was in there. Yeah. Just to recap what it was. So OpenCode just like CloudCode, but so in my head I was like, it has all the same tools as CloudCode, but it has some more tools in CloudCode. One of them is a way to search SourceGraph and I was asking it to implement something using a library, my own library which I forgot how to use.
Dax:Relatable. Yeah. And then it just did a query using source graph to find any public code using it and it found an example of what I wanted to do and then use that to to implement it. I'm like, that's genius. That's that's really smart.
Adam:Yeah. I don't know. I just I feel like I feel like these are solvable problems. Like models, I just assume models are just a brain. And they the knowledge cut off thing is not a thing that registers with me anymore.
Adam:It's like that's what fetching and and using MCP servers is for for getting, like, contextual very contextual information and putting it into the context. I also don't care about cost, so maybe that's my problem. Like, I'm like, yeah, dump all the docs in it. If it gets the right answer and it saves back and forth, then that's worth it. But I'm just assuming also model costs are gonna go to zero.
Adam:Like, that's just a long term trend is commodity. I don't
Dax:know. The cost yeah. The cost is a factor. The speed is a factor. But I I think because I've tried and the output's really bad, like even with that stuff.
Adam:What what are you talking about? What output? Like, because
Dax:because because I've tried mainly testing this, right, like, what if I gave it access to docs? What if I did x y z things? And it definitely improves the performance, but it still just makes up stuff that's like doesn't even look like SST at all, v two or v three. Like, it just like made up Mhmm. It like structures the whole app as like one JSON object.
Dax:Like, it just And
Adam:this is with this like the knowledge cut off of three seven or something.
Dax:Yeah. Like it can it can definitely write v two, s t v two code well, but then when I tell it, hey, we're doing v three here in the docs. Mhmm. It like really starts panicking. Just starts to starts to makes makes some crazy stuff up.
Dax:So Yeah. Yeah. The the doxing is, like I said, it's obvious, we should do it. I just don't know
Adam:if it'll cross that bar.
Dax:And it might, maybe with like enough tricks, but
Adam:The LSP diagnostics is also a form of docs. Right? Because that's pulling like, when you get a type error or even just the hints and stuff, don't those have like whatever docs you wrote above the function or whatever, like
Dax:That would be cool if that's in the LSP. I don't know if the LSP diagnostics contain the JS doc
Adam:Maybe not.
Dax:Information. It definitely contains the type inform or there's a type error, which is usually good enough. Yeah. Yeah. This whole context thing is is what sucks, which is why I'm looking at this continuous training thing.
Dax:Not just it's not just fine tuning, I'm trying to understand this stuff better. There's like a whole host of techniques. Yeah. I don't think people are gonna switch out their primary coding model. That's why I would want to deliver this as another AI that your primary AI can talk to and send you Got a it.
Dax:Yep. And that it's like a more specialized.
Adam:Okay. So you're I mean, we're kind of talking about the same sort of thing. The thing you're talking about would kind of be like an MCP server. I don't know if you would Yeah.
Dax:It would be an MCP server and it's Okay. And there would be like a collective effort to have data that's input into whatever this process is. And I could just be straight up a way to search.
Adam:It's just smart version of what I was saying. I was saying like dump markdown into the context. You're saying like, no, be intelligent and have a model that actually gives you the correct answer.
Dax:Yeah. I mean, it could do like I said, with enough tricks, one of these approaches might end up being good enough. But, yeah, the root problem is like getting the data together. That's really what where the tedious Oh, stuff and we have another open project that we're trying to kick off.
Adam:I I feel like you've hinted at a few and I are you gonna reveal? Can we reveal them to the podcast listeners
Dax:and Yeah. Yeah. So I posted that one yesterday because we're looking for someone to take it on and I got a bunch of responses.
Adam:That's the one I saw. Yeah.
Dax:Yeah. Yeah. So we are probably gonna move off of Astro for all of our docs and everything. Mhmm. Here's what we like about Astro.
Dax:We like that it's JSX, it's components.
Adam:Mhmm.
Dax:You can output everything statically. What we don't like is the dot astro file. We don't use like 90% of astro features. It's quite heavy for how little we use it. Mhmm.
Dax:And I was talking to Jay yesterday who's like been through like every iteration of this for the past fifteen years because he built like our SSD book, like he built that using like
Adam:Oh, yeah.
Dax:Using Jekyll and like he's just been through all these different iterations, so he has a really clear sense of what the scope of this thing is. And his point of view was that Jekyll is effectively the perfect product. And if you and his point was if you go to their site, you'll see it looks like a site from 2015 because it hasn't really needed to have been updated since since then. It's actually pretty nostalgic. Reminds you of like the beginning of really two point o.
Adam:Yeah. Transfer your plain text into static websites and blogs. Okay.
Dax:Yeah. So he's like, if we just build that, make it based off of bun so it's really fast because he'll put them in Jekyll, it was like it got pretty slow. Mhmm. Make it like very low dependency, build off of bun, JSX based, MDX based and then just take the exact scope of Jekyll, which is not very large, it's a very small scope, very well defined and just do like a modern JavaScript based refresh off of it using bun. Mhmm.
Dax:Well, I don't know, open site, I don't know. So we're gonna like probably Oh, move our docs
Adam:site. That's
Dax:to that. Mhmm. I also have been thinking about publishing some of my longer threads on my own site as well. So I probably I've kind been looking for something like this.
Adam:You can dog food it and get some good feedback.
Dax:And then a part of this would be to output the four LLMs.
Adam:Mhmm.
Dax:Whatever this pipeline is also to spit something out as structured in a way that, you know, whatever this thing is can can ingest Mhmm. Can be ingested into.
Adam:Yeah. Such an interesting time that we live in. Yeah. It's really it's fun. I I enjoy this, like, transitionary period.
Adam:It's this weird, like, you know, we need to all be building towards this, like, this world where we control a legion of little Yeah. Junior devs, and then maybe they're senior devs, and then maybe they're better than us. But it's also this feeling of, like, you don't know if at some point things do get so good that it's just, like, none of it matters. You know what I mean? Like, is is that just like not gonna happen?
Adam:Is that a sci fi reality that like software dev just vanishes because AI is so good that there's just AI? There's not software. You just talk to the AI and it does everything.
Dax:I felt this way from the beginning. I feel like every day that goes by, it seems like that's less and less the case. But Yeah. I also know that that's often also a trap. So, yeah, I don't really know how I how I really think about things.
Dax:I've been pretty underwhelmed with all the different model releases.
Adam:Yeah. Same. To be honest, the last couple of rounds, it's like it feels like three five to three seven was such a minor bump. And granted, it's just they didn't they didn't call it four o for a reason, I guess. Like Yeah.
Adam:It wasn't supposed to be a major jump maybe. I'm very keen to see the the Claude four g p t five generation. Yeah. Like, what do things look like? It's like one more data point of are we leveling off here and then we've reached peak
Dax:with Yeah.
Adam:Current And like tech.
Dax:I think it's a funny thing because we know the because it's so mathematical, they're gonna increase the amount of the the model is gonna be bigger. The next Yeah. They generate it's just gonna be bigger and therefore it's going to be like the inputs are like mathematically larger.
Adam:Mhmm.
Dax:It's question is, is output gonna be exponentially better or is it gonna be like start tapering off? Mhmm. I wonder if it's gonna turn back to the most boring problem that's existed forever which is just it's just all about your data. At this point, these models have sucked all the data that's like generally available. Mhmm.
Dax:And if they plateau of being like, yeah, we've tried every single trick, yeah, we can make them a little bit better, incrementally better just off of just like how we're training and doing all these little tricks. But there's like a ceiling here. Then I think it becomes the era of like custom trained models that are
Adam:Yeah. Vertical specific
Dax:Yeah. Like models. Yeah. Hyper trained on like certain data that's not generally available. Mhmm.
Dax:Like for us, you know, again, going back to the examples I made earlier, OpenAI is not gonna label being like, these are v two examples, never use them, these are v three things, like always prefer, you know, there's not gonna do that. And that's what it comes down to, then, you know, it's a boring old problem of like who has the most unique data and who has the most data and that that kind of goes back to that old race.
Adam:Yeah. It's interesting. I don't know the since we've talked or no, I guess when we talked last time, we did we just didn't talk about it like but we didn't talk about Llama four at all, did we?
Dax:We did talk about Llama four.
Adam:Or DeepSeax v three for that matter.
Dax:R two should
Adam:be coming out soon.
Dax:Llama four sounded like just like benchmark hacking from what I
Adam:saw That's people what I saw.
Dax:I tried it myself and I was like, this is I was excited because I'm like, oh, oh, the knowledge cut off is August 2024. Like, s t v three was, like, kinda out then.
Adam:So interesting you think about knowledge cutoff. I just haven't thought about knowledge cutoff in so long.
Dax:It just it just it's our, number one pain point. Like, someone new hears about s t, they go their LLM, try to use it, they get the worst experience ever. So it's just, like, killing us. S t
Adam:v two stuff. I I get that. I guess it doesn't affect me because, like, as a builder of things with s s c v three, I I guess I don't really ever try to have it write the infra stuff. That
Dax:just No. Because it's so minimal.
Adam:It's so minimal. Like, that stuff just Yeah. Like, I wanted to write the stupid giant component that I don't want to write from scratch, like, for a front end. That's what I want the LMs to do for me. So I but I get if you're new
Dax:If you're new, that's a problem.
Adam:Yeah. Like, it would be really cool to start with a blank directory and just create a modern repo in the vein of what your kind of pattern is. I mean, the whole SSD mono repo, the way you do things. I don't know if that's a pipe dream. I would love that.
Dax:The thing that I keep testing it with is, hey, can you add a cron job that does, you know, something? Yeah. I I listed out and I expect it to go create that cron job on the SC side and also go implement the application code for actually running it.
Adam:Yeah, right.
Dax:And that's like my benchmark. Whatever gets us there, then I'll start to feel good about, okay, this this we're like passing the bar on on what we need. Yeah. But yeah, I haven't
Adam:been able to get there yet. Well, now I wanna test. I wanna I wanna throw together a little docs MCP server.
Dax:Should do
Adam:I'm just gonna try it with SST, like, it'll just be all faked. But just to see like if you drop SST stuff markdown in context, I guess I don't even have to build MCP server. Just need to drop some markdown from your docs to your MDX files. Just drop them straight in, see what it does.
Dax:Yeah. I think Frank's gonna look at this next week. It's funny. So last year, so Jay comes out for React Miami. So it's Frank by himself in Canada.
Dax:So he's he's he's like a it's like a week where none of us are interacting with him.
Adam:Bothering him, yeah.
Dax:Yeah. Last year he shipped like all of container stuff for SD, which was like really big for us. And that that wasn't totally random, it's because we were shipping the SSH site and we were doing like a bunch of container stuff. So I was giving him stuff we needed.
Adam:Yeah. But it was like a really
Dax:productive week. Was a very good release and containers are obviously huge for us now.
Adam:What's he gonna ship?
Dax:So this week, what's up we've been like, okay, what are you gonna do this year, Frank? Like, what's your big thing? But he's gonna look into this like getting our docs in in LM format and and and all that.
Adam:Yeah. I I don't know. This feels like it's a solvable problem is my gut. I don't know what the solution will be. Maybe it's the model thing, maybe it's something else, but I do think this is the thing we can get over.
Adam:Right? Like, where we can always use the right like, where we can teach DLMs about versions of software. SCC is an extreme example because you guys changed so much from two to three. Like, it's just such a different example of code. Like, the code examples are just so starkly different.
Adam:So I I feel you guys probably feel that pain more than most.
Dax:Yeah. I mean, the but the other issue is like, it just sets you down the wrong path. Even for a minor change, like, if you're using tailwind, it's gonna set you down the v three path with like different conventions.
Adam:They have to take out of the JavaScript file, you know,
Dax:that kind of change how they think about things. So you kind of accidentally end up doing the old stuff without realizing.
Adam:Mhmm.
Dax:So, yeah. I think that the it's this weird paradox where everyone loves these coding models, they're so useful and they're so good. Every day that goes by, they get a little worse because they're a little bit more out of date. So Yeah. They need to solve the problem otherwise, the model becomes not useful.
Dax:At the same time, I don't like I can't actually see them solving it, you know.
Adam:Yeah. Yeah. I guess I'd I never thought about like I guess when they're writing React code, it's from all the samples in their in like, that they were trained on. I just I guess I think of like software development relying on fetching stuff into context. But maybe that's happening less than I realized, like, when I use Claude Claude code.
Adam:Maybe there's less fetching than I realized.
Dax:But like I said, it's just a context thing. Like, it makes everything really slow and more expensive. So as like a mass market solution, it just isn't isn't No. That makes a great one.
Adam:You're talking me into it. You're talking me into the thing you're doing.
Dax:Yeah. And here's the other thing though. Another pattern I've just seen over and over is we'll always go one step lower and then we'll realize not enough. Okay, we gotta go one step lower. Okay, we gotta go one step lower.
Dax:And we're always like, why don't we just go down all way to the bottom from the beginning? So that's kinda how I feel like with this AI thing. I'm like, yeah, there's all these like little things we could do. Mhmm. But I'm just gonna go try to do the most extreme thing of having our own model and then like build up this time instead of like and you think about SC itself, it's like, we shouldn't have built on CDK.
Adam:Right.
Dax:We got stuck with that for too long. We went one level lower now down down to Pulumi. I'm sure at some point we're gonna like have to go go lower again.
Adam:Mhmm.
Dax:Yeah. Just trying to just go to the trying to like see if there's a bottom and go straight there.
Adam:Yeah. Just avoid all the dance of, like, Plinko ing your way down. Do you remember Plinko? Wait.
Dax:Who makes that right? Is that thing that you,
Adam:like, drop a thing at the top and Right. Then was like
Dax:Yeah. Yeah. Yeah. Okay.
Adam:And, like, kind of bounces off the pegs.
Dax:I've seen that like arcades or something and like you can
Adam:like win something, but it's called Plinko and like where did you come across this in your life? Is price is price is right? The it was like one of the things was Plinko. It was like a
Dax:it was like a segment on Price is Right. Yeah.
Adam:Did that you ever watch Price is Right? Do you have the basic idea?
Dax:I think I I vaguely maybe I've seen one episode. They like guess the price and if you're too high, it's you're out
Adam:or what whoever wins the little round where they have to guess the price, somebody goes up on stage and they do like a
Dax:little game.
Adam:There's like all these mini games. Not just not just Plinko. There's lots of games.
Dax:Okay.
Adam:Plinko was one of the games that they did.
Dax:I see.
Adam:That's all.
Dax:How come you experienced this and I didn't? Is this like Price an age
Adam:is Right. Why why did I why did why is Price I think my grandma loved it. I think every time I went to my grandma's, I watched Price is Right, like, lot. I can't remember why I watched it so much, but I definitely I know a lot of Price is Right. I'm like Teej with Family Feud.
Adam:Yeah. Let's do a price is right.
Dax:Let's do a price is right thing.
Adam:For like SaaS products or something, and I'll be I'll be Bob Barker.
Dax:Gets your bill.
Adam:Yeah. Your AWS bill. There you go. Like, cloud bill, line items.
Dax:I pitched Posthog on doing something with Terminal.
Adam:I hate that name. Sorry.
Dax:I know.
Adam:I love you Post Hog. I mean, do stuff with Terminal.
Dax:Nice. Immediate that's my job. It's my job to lose sponsors.
Adam:Yeah. Yeah. That's right. Not my job. The the I don't like the name.
Adam:Sorry. Continue. Post hoc. Is this such a weird I don't know. Hog shouldn't be any company.
Adam:I feel like
Dax:we could do something funny with a name. That was, like, 80% of my reason of reaching out. I'm like, there's definitely something funny here we can do.
Adam:Mhmm.
Dax:Yeah. So let's see. Still looking for that sponsor for our rocket.
Adam:Oh, yeah. Well, let's let's go straight to the bottom of it. Let's just do the SSC thing. Let's go straight to the bottom
Dax:of Let's build our own spaceship? No.
Adam:Let's just let's get hooked up with what's that company called? Starlink. No.
Dax:SpaceX? Rocket. SpaceX. So many people screaming
Adam:screaming on the other end of the podcast right now. How could you not know? We're going to Mars. Sorry. I I don't think I wanna be near Elon right now.
Adam:So Yeah. What what's the go what's the deal? Tell me more. What do you mean? No.
Adam:I just I don't know. Is this Well, okay. Well, when you said that, it made me think of something I saw or I think Casey showed me where he's, like, had his kid with him all the time and people think it's because he thinks he's gonna assassinated or something?
Dax:To me, it lit I I did not see anyone say that and I independently had that same feeling. You had that thought.
Adam:Okay. It's like a big conspiracy or not conspiracy, just a big like theory around
Dax:It feel it like it just looks that way. Like, visually, it looks that way because it looks so forced.
Adam:I would not yeah. If I had that much money, I mean, I just I can't get in the head of somebody who has that much money. Like, I don't know. Whatever. I'm not telling him what he should do, but I just can't imagine I would be doing the things he's doing.
Adam:Does that make sense? Like, I would not put myself at risk in any way. I do all the other big billionaires just kinda, like, disappear. Right? Like, what does Larry Ellison do?
Adam:I don't know. He's just, like, living on his island in Hawaii. Right?
Dax:He was he was at the White House talking about AI, like, the day after president got elected or got got into the office.
Adam:I would just I would not be a high profile person if I had enough money to just buy an island and live there for the rest of my life. I don't know.
Dax:Yeah. I know some billionaires stay low key, but this is one of those things where I'm like, if you feel a certain way, if you're like, oh man, all the billionaires doing that, I don't get that. Like, I would never do that. I think it's somehow related to then you just if you don't if you're not
Adam:like that, you just wouldn't become Yeah. A wouldn't become a billionaire. Yeah, exactly.
Dax:People were always like, oh, I don't get why like they're still working when they have so much money. At the moment I have fuck you money, I would just not do anything. I'm like, that's why you're never gonna get it.
Adam:That's why
Dax:you're gonna
Adam:have it.
Dax:You just don't have is self
Adam:reported as a non billionaire.
Dax:Today, found
Adam:out that
Dax:Adam does not have a billion dollars.
Adam:Yeah. And won't ever. I don't have what it takes. Well, that one, I think that one is the thing you're saying that was a little
Dax:more reasonable because there are
Adam:billionaires billionaires, I think that I think they kind of like do the thing I
Dax:would expect.
Adam:Crazy person, so that's like, you know, no point looking at him for information. Yeah. What what's the deal with the tariffs? Do we have tariffs? There's no tariffs?
Adam:They pause the tariffs?
Dax:They pause. Yeah. And then the market shot up 11%. I was insane. I've never seen anything like it.
Adam:Was the whole thing just a a prank?
Dax:Is there any indication that there's a possible gambit that we can't understand. It's that they're thinking 11 steps ahead and we're never
Adam:gonna It's plan. We're never gonna follow it. Every single I have hard
Dax:time. Every single move they make is a win, no
Adam:matter which direction it goes. Playing four d chess.
Dax:An underwater five d chess now.
Adam:I guess China though, they didn't think the joke was funny and they're still raising
Dax:I saw something really funny What does it do
Adam:to us when they raise their tariffs on us? I don't understand that. What what's it do to us when China says we're gonna tariff your stuff that comes in? Do we sell stuff to China?
Dax:Yeah. We do. I think on the tens of billions, I think. Okay.
Adam:So what were you gonna say, sir? I saw something funny today because, you
Dax:know, US and China are beefing about this tariffs thing and it's like, it's so funny because it's it's indirect because they're they're never like we're
Adam:not we're not getting on the
Dax:phone with China being like, hey, fuck you. It's it's it's more like China announces to the world that they've increased its tariffs on The US and then The US announces to the world that we're you know, it's so passive aggressive and today there was like the epitome of that. I I don't know how accurate this is but someone I saw something being like, White House tells China to request a meeting with Donald Trump, which is like so passive. They're not like, hey, can we have a meeting? They're like, you need to ask us for a meeting.
Dax:We're not it's not
Adam:a meeting request. It's not a
Dax:meeting request. Don't leave us a meeting request because that
Adam:would be weak.
Dax:You need to come to us to ask us for a meeting.
Adam:Oh. Like,
Dax:goddamn. This is like
Adam:That's really annoying.
Dax:This is so lame.
Adam:Yeah. I don't know. The whole thing, it feels like it was all over the timeline and then I feel like things just kind of cooled off. People are people just exhausted?
Dax:Well, there's a pause and here's the thing. I I was saying this for a while ago, I was like, in America, the money like rules everything and we've had we've had a bunch of problems with that.
Adam:You've been consistent, you've always said that.
Dax:And I'm like, we're seeing that tested like is the the benefit of that is like, you know, things can only get too chaotic, they usually step in.
Adam:Mhmm.
Dax:And what happened, Jamie Dimon goes and gives an interview, talks about how bad this is, how talks about we need to like, you know, this needs to be like corrected, stopped. Later that day, ninety day pause, Trump even cites Jamie Dimon. That's how Oh, wow. Like, the man that will never give anyone credit for anything
Adam:Uh-huh.
Dax:Even mentioned Jamie said we needed to like, you know, pull it. Sure it wasn't just that obviously Jamie citing like real things that are happening in the market. But Yeah. Yeah. At the
Adam:end of
Dax:the day, you know, money stepped in and Cooler heads up.
Adam:Ninety days and then and then the stock market, so that we're gonna boom for ninety days and then it's gonna crash. No. We're not
Dax:gonna boom. Right now is crazy confusion. You can like if you look at the chart, you just feel the confusion that everyone's feeling because it's just like it's flat.
Adam:Oh, the VIX is way up.
Dax:Yeah. VIX is way up. Yeah. It's Mhmm.
Adam:So I know things like VIX.
Dax:Mhmm. Fear index. I don't know about the fear index. I do. I have
Adam:a fear index in the pit of I my have a lot
Dax:of anxiety. I've read the formula for the VIX like a 100 times over the years.
Adam:Oh, is is it an
Dax:I just do not understand it. I just don't I cannot get an intuitive sense of how why it works.
Adam:It's pretty accurately like predicted stuff. Right?
Dax:I don't get I don't fully get because it's it's like a volatility index. I'm not Right. Just mathematically, my head doesn't
Adam:compute it. Like, how you build
Dax:a financial instrument around volatility as opposed to just going up and down?
Adam:Can you trade it? You can actually like it's like a thing you can trade is the VIX?
Dax:Yeah. There's a VIX ETF called Oh. I forgot what it's called. Used trade I used to trade that triple leveraged one, which meant it was so fun.
Adam:Oh my god. You're just looking for chaos. You're just trading and hoping for
Dax:Yeah. It's like it's like tariff
Adam:something would happen, the market would go down 2% and then your portfolio would be down 50%. Three times leverage. That's that's intense.
Dax:Yeah. It was awesome.
Adam:You do lot of trade it seems like you're big into the trading stuff. You're always like posting your options stuff. I've been on and off over the years.
Dax:I'd like cold it for like a couple of years, so I just wasn't thinking about that at all. And I picked it back up like a year ago.
Adam:Mhmm.
Dax:And it's just fun. I never lose. Mhmm. I always make money. Never lose.
Dax:Well, that's I mean, I always make money.
Adam:If it's that easy, then maybe I
Dax:do it.
Adam:I want you to
Dax:just gamble in this and Just be really smart.
Adam:Know what's gonna happen.
Dax:Don't be smart. Just get lucky. It's I I I've just been on a really good lucky streak.
Adam:The that
Dax:that portfolio I put money into originally in 2019.
Adam:2019.
Dax:And it's now up like 1200% over what's Wow. Five, six years.
Adam:I'm like, man You're like a hedge fund.
Dax:If I was ballsy enough to put all my money in it Oh. We just wouldn't even be talking right now. Is why I'm never gonna be a billionaire. There you
Adam:go. Yeah.
Dax:That's a
Adam:really good track record. I feel like you should you should just make a course about trading or something.
Dax:I don't People keep asking me like, people are like, hey, can you like I'm gonna try and get the options trading, like, can you
Adam:you want can we like hang out and like, you
Dax:tell me about it? I'm just like, I don't know anything. I literally do not know anything. Know the basics of how it works. I have some, like, habits I've developed over the years, but I just feel like I just keep getting lucky.
Dax:Like, I I don't know. It feels like a 100% gambling to me.
Adam:Definitely happens. Somebody gets hot at the Yeah. In the casino, the roulette table. I'm saying things I don't know anything about. I've never played roulette in my life.
Adam:Is that a table or is that the thing that spins?
Dax:The thing that spins, we tried it in Vegas when we were there, like, for Reinvent. Yeah. Liz was like, I've never a game before. Let's try this. And she's like, wait, we just lost all of our money in like ten seconds.
Dax:This is not fun. I was like, yeah, this was not fun. But then we won it all back at the slot. Like, I did the slot machine once and then we won all
Adam:of it back. You won all your money back?
Dax:And then we left.
Adam:Is the is the roulette thing, is it just red and black? You're just picking No,
Dax:you can you can place on specific numbers, you can do like
Adam:Oh, there's numbers.
Dax:Odds, evens, you can do all kinds of
Adam:Oh, yeah.
Dax:And if you if in Vegas, you can whatever you want to bet on, they're gonna find a way for you to bet on it.
Adam:That's funny. Not a better bet that the table's gonna explode, like, you know, probably about anything. Bet that the tariffs are gonna be paused. Oh, there's actually that. Right?
Adam:Like there's those what's that called? A prediction markets? The prediction markets? There's like a a specific thing people
Dax:always post charts. Yeah. People always post it. Yeah. I think people always post it post screenshots of it right when right before the thing happened, like the event happens.
Dax:And it always looks like, oh, it's so accurate. But I don't think they're actually very accurate. I have to look more into this but there is like a Wisdom of a Crowd thing and that's like a real thing but
Adam:Oh, yeah. Like if you have a 100 people and they guess how many gumballs are in a jar, like you take the average of all the guesses and it's actually Yeah.
Dax:Close. And that's roughly what a prediction market is trying to recreate but I've seen it like totally swing from one extreme to another.
Adam:Yeah. Guess you've got to have the volume. You got to have like enough people.
Dax:I mean, with them though those prediction markets have decent volume. But yeah, I don't know. Whoever thought of that concept though genius, like, you basically create a betting platform for anything and you're just gonna make commission off of that forever.
Adam:Oh, yeah. That's
Dax:Jealous. Good
Adam:business. Sounds like a good to that. Why don't we do one of those? I just constantly am, like, seeing some story about somebody making a zillion dollars. It's the dumbest stuff.
Adam:Just like the stupidest ideas and they just work. Why why is it it seems like some people just have that and like, I just never make money. The things I do just struggle to make money. I guess we're making. I mean, we make money.
Adam:It's just like I feel like I work too hard for the output. You know what I mean?
Dax:I know what you mean, but there's also just like
Adam:who you are and how it feels to your soul.
Dax:Like, I could just I couldn't do that.
Adam:Like what you actually want to work on and put out in the world.
Dax:Yeah. Things have to line up. They don't line up for me when, you know, if I was working at it. Much as joke about that prediction market thing, like, yeah, I I don't think I would really Enjoy. Deal with something like that.
Adam:Yeah. I I get too much enjoyment out of just making something that I think is good. Like, it doesn't have to have any other like it didn't have money involved. No one saw it. If I just think it was good and I made something that was good, I get too much satisfaction out of that.
Adam:I think that leads too many places of, like, head down on something that's not worthwhile. That makes sense?
Dax:I think I feel that way too and I think historically, I I remember saying this years ago, was like, you know what, if I could spend all my time making stuff for myself, I would be pretty happy. That's changed now where a lot more of my satisfaction is feeling like I made something really good for myself and for a lot of people, like now that's like a much bigger factor in my motivation. But it's still the same thing, like I'm still just making stuff for myself. I'm just making sure I do all the other stuff to make it so that like a lot
Adam:of people could could potentially enjoy it as well. You've got a lot of those going. So how many how many open projects are there gonna be now? And what is Toolbeam? What's tool what's Toolbeam?
Dax:That was a temporary GitHub situation where we're moving everything under SST. By the way, did you see that NPM deleted all access keys yesterday? What? No. I need to see if they fixed it.
Dax:NPM deleted every single access key.
Adam:What?
Dax:So everyone's CI, everyone's stuffs
Adam:Is our is it still back?
Dax:It's all it's all back now. But it was down yesterday.
Adam:So they undeleted them?
Dax:Which is all gone.
Adam:Wait. What is going on with stainless? So they like, you tried to deploy yesterday, it failed, but they they restored the same keys, so no changes had
Dax:to be I mean, was like some kind of data loss because if you if you went to that page, it was just an empty list and it says something weird like showing one of zero of three. It was like
Adam:some weird thing like that. Oh, man. Now there's now I'm having issues with another unrelated project trying to see I was trying to confirm your your reporting and now I'm troubled. Anyway, what else is going on? I don't know.
Adam:I feel like something else is going I feel really busy, but it doesn't feel like there's a lot of stuff going on. Like, we're very busy, but that doesn't mean there's a lot going on worth talking about. Happens. Kinda where it is.
Dax:That's everything on my end, you know. Got the open site, got the model where you're working on open code.
Adam:Yep. Open code's gonna be great. I'm so I'm so pumped.
Dax:Yeah. I switched to using it for as my primary thing.
Adam:Yeah, me too. Until I hit like a roadblock and then I report and then I've bounced back to Cloud Code a couple times just to like remember how they do things Mhmm. And what I like about it and what I don't.
Dax:You see my thing today? Someone released something called OpenCoder, which is the exact same project
Adam:saw that.
Dax:With a similar name. Yeah. So that was annoying.
Adam:That is annoying. It's gonna cause some confusion if that gets traction. I'll crush it. I'm sure OpenCode will get traction. Crush it.
Adam:I'll throw in the I'll throw in the author.
Dax:I'll pull up her cell, and then go being like, you need to rename. You need to rename this
Adam:right now. That's what they did with Open Next, right? That's so funny.
Dax:I was gonna say, just going back to that, I will say part of me is like, I had to have an initial reaction to me like, so I messaged him telling him like, hey, we're working on this thing, you rename yours. And I remembered what a bad idea that was, how much that backfired on on Vercel
Adam:Oh, yeah. Where it
Dax:just motivated me to like never rename it. Uh-huh. Yeah. Yeah.
Adam:Yeah. You'll just embolden him. The open code thing, I think just some of the stuff if if we launch with all the stuff that is kind of, like, slated, I just can't I can't imagine Claude Co's not gonna copy half of it. They can't do the other models thing, but some of these are just really good ideas.
Dax:Yeah.
Adam:They just should have them.
Dax:Do you think there's more than one person working on Cloud Code? I feel there's like two people working
Adam:Yeah. On Maybe not very many.
Dax:My feeling is it's like a handful of people, maybe two people working
Adam:on it. Yeah. The PM lady follows me. She tweets lots of progress like they're always adding new stuff. Yeah.
Adam:Forgot her name.
Dax:I don't like head to head, I feel like whoever's working on it isn't going as fast as you guys are. So unless they like resource it differently
Adam:I'm not going fast. KudyTim is a beast. I don't know if you know this, but he
Dax:I is know. I can tell.
Adam:He's fast.
Dax:Yeah. Especially with now powered with its own tool. He's like using the tool to build itself, you know? I know.
Adam:Yeah. And I figured like, they're I think the Anthropic folks, they know programming. And and going back to the whole, like, niche models thing, it definitely feels like that's the path for Claude is to be the programming model. Right? Like, they're already kinda, like, falling into that niche.
Dax:Oh, speaking of, this thing I I found out yesterday, I don't know what the hell AWS Bedrock is doing at all. Because initially, was like, okay, they partner with Claude and if you wanna use Claude Sonnet whatever, you do it through AWS Bedrock. Impossible to get rate limit increases, like, just can don't hear about them. And then yesterday I found out, oh, it turns out they're not exclusive to AWS. You can use Claude through Google Cloud.
Adam:Oh, yeah. You told me about this. I didn't even realize Vertex was like a A thing. Yeah. So that's our Yeah.
Dax:Bedrock equivalent Google Vertex. Not only can you use it through Vertex, the rate limits that they they gave to me are higher than what Really? Not just Bedrock, they're higher than what Claude gave to me. So I'm like, what is going on in Like, this this is not making any sense at all. Like, why is Claude losing customers through Google?
Adam:Where can I yeah? Where can I see where can I see how many well, losing customers? Right? They're still like it's like Google's just reselling it?
Dax:Once you get to the mindset of like, oh, we're still winning because, you know, we're partners Mhmm. You just just eventually get squeezed on other side.
Adam:Yeah. People are using Google now and they can swap out their model really easy.
Dax:Yeah. Exactly. The moment you become a little bit commoditized, it's it's done. Mhmm.
Adam:Where where can I see what models they support in Vertex? 200 plus They're single
Dax:model garden. Oh. I heard something like
Adam:Model What? Why? Oh, Google. Oh. It's not
Dax:just Google. Like, this is like this is like a whole cloud thing.
Adam:Oh, really? Feel like they they're schizophrenic on how
Dax:they name things. Sometimes they're really technical. They're like, they just name it the thing? Like, Google SQL, you know, sometimes it's not really direct. And other times they're like, Model Garden or what was that crazy one?
Dax:AWS had a really crazy one. It was the worst name I'd ever heard. Was it the one with Was it their feature flag?
Adam:Worst named AWS service.
Dax:It was like
Adam:There's the one with the beaver for a logo. That was pretty funny.
Dax:What's what's the name of the service?
Adam:Lumberjack or something. They had like a service called Lumberjack, I think.
Dax:Is that their gaming thing? Maybe. Is what is Amazon Sumerian? What the hell is that?
Adam:Sumerian. Was like a it was like an AR or like three Yeah. D model Sumerian. Is Lumberjack a service? Lumberyard.
Adam:Sorry. Lumberyard. Yeah. There's a beaver right there with the saw.
Dax:Lumberyard feels okay to me because I'm like, okay, you're building something, you got you're like cutting wood up. Like, I get I get the feel of
Adam:It's been deprecated. So apparently, it didn't catch on. Something with three d again.
Dax:Anyway, so model garden. Oh, Elastic Beanstalk. That that's another one. Elastic Beanstalk.
Adam:That's tough.
Dax:What is that? It's like they have that and they have relational database service. That's what you can ever imagine.
Adam:We should we should do tier list sometimes or sometime of like tech names or
Dax:Yeah. Yeah. Could just be AWS. Being stuck. That's pretty funny, actually.
Dax:Yeah. Where's all
Adam:this headed? I'm just I'm just seeing like tweets from Windsurf. There's it's like cursor is not the only thing. There's like so many of that model where they're like, I'm gonna be built into an IDE. Why would that be better?
Adam:I just don't understand. Don't want it built into an IDE. There's so much shit in Versus Code that I don't know what it does or care. Like, I don't understand the idea of bolting your AI assistant onto an IDE. It just seems so much objectively better to be independent of people's But they
Dax:have the most usage, so you're kinda wrong, right? That way Wait what?
Adam:Oh, like cursors Yeah. Got the most yeah. That's a good point. Yeah.
Dax:They do. Think people use most people use Versus Code, so it kinda makes sense. The the thing with them is I don't they're trying to build a venture sale business. They have raised crazy amounts of money at a crazy valuation. You are now entering the I'm I'm gonna compete against Microsoft game.
Dax:Nobody wins that game. Yeah. Definitely not as your first time company.
Adam:Mhmm.
Dax:So Well,
Adam:they've already started adding cursor like features, right, to Versus Code?
Dax:Yeah. Heard this. And they're not doing a good job. Like, Microsoft's not doing a job. Sure.
Adam:That stuff in GitHub is not great.
Dax:Like, it doesn't feel good. But this is the thing that always happens where you're just like, wow, I can't believe Microsoft let that thing come up. They Microsoft lost. Microsoft thing sucks. And you like stop looking at it and then you look back five years later and then it's like, oh, wait, Microsoft just ended up winning all that.
Dax:So it's like they move so slow and their stuff stays bad for so long.
Adam:But they win in the end.
Dax:But in the grand scheme of things, you know, to like, our plans are measured in centuries, you know, that that type of thing. Yeah.
Adam:There's something this isn't a talking point I wanna hit on real quick because I I did have an experience with Casey where just like the vibe coding phenomenon, I I wanted to, like, sit her down and see, like, she had expressed interest in, like, trying to play with these app builder things. Yeah. And it's like, we installed Cursor on her MacBook, and it only it only then occurred to me that, like, Cursor's not made for people who don't know things about programming. It's not for people
Dax:No. It's not.
Adam:Who have no context. Because, like, immediately, we couldn't do anything. She doesn't have Node installed. Yeah. Yeah.
Adam:I'm trying to explain, oh, yeah. Yeah. Yeah. Well, if you want to build a mobile app, we should probably use XPo. She's like, what is XPo?
Adam:What are you talking about?
Dax:Yeah. Like, oh, well, we could
Adam:just use Ragnet. Let's just use Ragnet. That's fine. That's fine. It was like, oh, you don't even have Node installed.
Adam:We can't do this on your MacBook. So it's just this, like, endless so you have to have a certain amount of context, but then cursors coming in and being like, but not that much context. You don't know you don't have to know how to program. You should have to know everything about programming.
Dax:Right. Yeah.
Adam:It's like, what what is the I don't understand like it's for existing developers.
Dax:It's it's definitely for developers. It's Cursor is betting that developers continue to exist and Mhmm. There's a lot of them, maybe even more than they were before and this is how they most people build build stuff. Then you have like the whole other category of product, vZero, Bolt, Lovable. There's more now.
Dax:Firebase Studio.
Adam:So those are for non developers. I just mix them all together.
Dax:And Convex Chef. Yeah. Yeah. Those are those are for your wife would probably be able to do stuff there Okay. Without knowing much.
Adam:Okay. So it's not Kerser's fault that the vibe coding thing became this big phenomenon where people are just, like, accepting everything the LLM does. They're not really programming. They're just, like, telling it what to do, but they are programmers.
Dax:Yeah. They they know what's going on
Adam:to some degree. That confuses yeah. That muddies the waters a little bit for me, like, in terms of who is using this tool, like but I guess that makes sense. It's existing programmers who just want to turn their brain off and try and build something. Yeah.
Dax:Yeah. So I I don't know. Like like I said, it just looks like they're gonna win because Microsoft keeps sucking. But fundamentally, like, why would a fork do better than Yeah. The trillion dollar company, you know?
Dax:It's just like like, yeah, oh, like we're better we're better at product, better at building, we can move faster. But like, at this level of competition, going for a VC scale company, I don't know, like I just don't I don't see it. Yeah. And the clear thing is obviously you have to go do your own model that's specialized. Even Do there
Adam:they? I don't even know. Yeah. They I mean, the cursor
Dax:should solve this problem that I'm talking about also.
Adam:Yeah. Like, do they do they currently have their own model or is they They're
Dax:using it one. Yeah. You can see it from
Adam:their job postings.
Dax:Gotcha. But you know, Microsoft also is doing that and they have more data because they're all GitHub, you know. It's just like
Adam:That's a tough one to compete against. The Microsoft who owns Versus Code and GitHub.
Dax:And they've been building Visual Studio forever and Visual Studio is actually really good for, you know, a heavy IDE. Yeah. And the other whole JetBrains category of thing, I don't know. I was just like, what how are you, like, trying to, like, navigate through all this and be the winner?
Adam:Like, when when when we talk about them being a fork, do they, keep upstream changes merged into their fork and they're just, like, layering their stuff on top of it constantly? Or is this, a constant rebase? Because that sounds like a nightmare.
Dax:Yeah. I mean, we kinda did that with CDK. It was like I kind of relate to it a little bit. It wasn't fun and you like very quickly want to like have your own thing. Yeah.
Dax:Which I think they might do at some point. I think that that can make sense. And then Okay. So then another part of this whole space is Zed and I feel like Zed is set up really well but they're not being aggressive enough. And I like I feel this is my thing with so many companies where I'm like, you have everything except the aggression.
Dax:Like Mhmm.
Adam:Because they built this thing from the ground up.
Dax:Yep. It's much better experience Yeah. Than Versus Yeah. Like, just compete with they they could have just competed with straight up Versus Code. That was like, they had a shot at that.
Dax:But now there's like a new angle.
Adam:Yeah. Whole new How
Dax:are you just letting cursor just dominate the conversation. Like, you can't. You need to do whatever you can to, like, shit on Cursor, make it seem not cool to use, like, make it Yeah. Be like, you're an idiot if you use this. Like, there's gonna be way more aggressive versus to me, it feels a little bit passive.
Dax:It feels like, hey, we we have AI in our app now. Isn't that cool? But it's not. It's like a very competitive space.
Adam:I'm just checking the Zedd website, trying to get a vibe of like how obvious is it that this is a cursor competitor. Because it is it's like the it's like they were building a competitor to Versus Code and then, like, a way more unique competitor to Versus Code came in and, like, redefined the whole category.
Dax:Yes.
Adam:Like, AI enabled this like, the product changed fundamentally. But they have this like, as opposed to a Versus Code fork, they have this advantage of, like, they have the whole thing. They don't have to worry about this fork situation.
Dax:Yeah. And if you look at it from in terms of product quality, Zed is better. Zed is better on the editor itself. Yeah. It's as good, if not better, at the AI stuff.
Dax:The Tap Complete
Adam:Oh, really? They're they're already like Yep. Competing on that. It's you're saying it's just a marketing thing. It's not that they need to pivot their product.
Dax:Purely, they're
Adam:like it's like an aggression thing.
Dax:Like, they just let Cursor dominate too much. And it's not too late. It's not too late to, you know, still still compete. Like Windsurf is there, they're also kind of like throwing Like Windsurf is being aggressive. I feel like they're like really trying to get in there.
Dax:And I want to see Zed win because it's a better product. Yeah. A world where everyone is using Zed by default is much better than a world where everyone's using a Versus Code fork or Versus Code itself, you
Adam:know? Mhmm. Mhmm.
Dax:So, yeah, a lot of these things. I feel this way about a of companies.
Adam:Maybe you need to just you consult. You need to consult on their marketing efforts. Hit us up, Zed. Hit backs up. You can help.
Adam:He's very good at being aggressive. And he's an East Coaster, so he's an asshole. Yeah. It's perfect.
Dax:Yeah. I need more East Coasters
Adam:on your team.
Dax:It's not it's not just it's not just Zed. I feel like me, Jay and Frank talk about this a lot and I was posting about this the other day where, you know, occasionally we have companies approach us that are like that want us to join them. And it's often interesting because the companies either have a lot of resources or set up really well in a way that we're not. But the the feeling we always get is they don't really wanna win. It feels like they're coasting along, improving, but not at a magnitude more than you'd expect them to given, you know, what everything they have going for them.
Dax:Mhmm. And it just feels like, yeah, we could join but it just I just don't I'm like that's the thing that's missing for me. I feel like people aren't like really playing a win, like playing a B number one, you know? Yeah. Everything just feels like slow, incremental, like passive growth.
Dax:Some
Adam:people are just like that. You you get that sense being around people who lead things in a very different way, like a very there's someone in mind. I'm just not I'm not gonna go any further because I don't wanna that's just not fair to but I've I've like seen people who are the head of a thing and they just have a very like relaxed attitude about their things placed in the world, if that makes sense. Yeah. Not everyone has that kind of
Dax:like, it needs to be
Adam:the best. We need to go to the bottom, as Dax says, and take it to the top.
Dax:Like, I I just love the sports metaphor because I feel like there's so much to be learned, business side, looking at sports. What you're describing makes zero sense in sports.
Adam:Oh, yeah. No no chance. Every GM would be fired if they were like, you know, it's okay to be in the middle, who cares?
Dax:Yeah. And like, even like the worst player in the NBA still has what I'm talking about.
Adam:Yeah. Oh, yeah. They they have to just to be in the NBA. I mean,
Dax:they have to have that yeah.
Adam:That crazy desire and push.
Dax:So I'm just like, why why when it comes to business and especially tech, like tech just feels so like passive. It's like,
Adam:you know, we're doing okay. Yeah. What now I have so many questions about marketing theory. Like, what what is it about Zed? Surely, have people at Zed that are trying to get the word out.
Adam:Why do I never see anything about Zed? They're just not in the right places.
Dax:I I mean, I haven't thought about it too hard. But the thing is you have to give people a compelling reason to stop everything they're doing in Switch. Mhmm.
Adam:The only thing they should be
Dax:thinking about is what can we say or do or like what event can we do, what like marketing gimmick can we do that makes everyone go, when I get home today, I'm gonna try try it out. Mhmm. Because it's a good product, right? So you you imagine it'll it'll stick with some amount of people when they try it. Yep.
Dax:It's that simple. Very hard, but it's that simple. It's it's all they
Adam:need to do. Simple and hard. I hate those things. I know.
Dax:That's how it goes.
Adam:It's like a catch 22.
Dax:Yeah. Anyway, this whole editor space chaos, it's just funny because like, I'm happy with Neovim.
Adam:I know. I love it.
Dax:I'm not gonna switch. Especially with this remote, my whole remote setup, like, now I'm stuck with Neovim because it, you know, it's not as fun doing that with anything else.
Adam:Should we end it here for an hour?
Dax:I feel like I
Adam:feel like I could jump into that whole conversation because
Dax:I didn't slow, weird episode. I'm sorry for the audience. This wasn't as good as the normal ones.
Adam:Maybe maybe Chris could like cut out the middle part where I was like, maybe we'll just be done after thirty, you know. Just call it. We just both have a lot going on. I got an entire family that's sick upstairs, and what we got next week is
Dax:I'm leaving my family right now.
Adam:You're you're making oh, you you announced this on Twitter kind of? Is it reply an announcement? Sometimes I write stuff in replies I would never put, like, as a top level tweet, but you're okay. Lot of people
Dax:saw it. Like, it was a lot of people saw the message. Well, okay. So for people that didn't see it, I was talking about how Chad GPT deep research is so useful. I just, like, tell it something I want to know all about and it just, you know, gives me back everything I need to know, like writes a report, which is amazing.
Dax:Someone's like, oh, you give me an example? And I'm just like, like, I don't want to make up an example because I have the perfect example. And I was like, I'm just gonna say it. I was doing stroller research because we're having a baby.
Adam:Congratulations. I just did that. On behalf of the audience.
Dax:So I posted that thinking like, no one's gonna see this anyway. And then everyone's Everyone
Adam:starts liking it.
Dax:Yeah. Everyone's like, this is the weirdest way to announce this I've ever seen. I'm like, okay. I guess our world
Adam:is now. Yeah. Sometimes replies feel like a private little place, but I think they show up on timelines just the same. They do.
Dax:So. When you reply to it.
Adam:Just amplify. When when did she do? I feel like I should know this. This is kind of thing I
Dax:Twenty fourth, which is the same day we got married, which is the same day we started dating.
Adam:More importantly, four days after my birthday. How exciting. Wow. I just told the world my birthday. Am I gonna get, like, totally doxxed or whatever?
Dax:My my my child is gonna steal attention away from your birthday.
Adam:Yeah. What if what if your child is born on August 20? That'd be so cool if we had the same I hope not because that means that means we have
Dax:to take care of her for for four more days than we would have otherwise.
Adam:That's what I was saying
Dax:the other day, like, you know, babies when they come out at four months old, I think I think it's a good business.
Adam:You think wait. What? Babies should come out at four months old?
Dax:Don't make a business where you you have a baby, but somehow it comes out and it's like four months old already.
Adam:Oh, it's already four months old. You skip all
Dax:the yeah.
Adam:Yeah. No. Believe it or not, I don't I look back on the the first few months fondly. I mean, I know
Dax:that it's hard
Adam:and but you just like develop so many it's just so like stress, low sleep that it all gets imprinted on your brain, so you have all these memories that are most most memories are fond. I don't know. Maybe that's just me.
Dax:No. I know. I know. I know. I I get that.
Dax:But I just think even I think that'll just happen anyway, even with my genius business idea of babies that come out at four
Adam:months. So four months, I'm
Dax:pretty sure that a human baby is like, technically, like, they're not like, ideally, they're not born as early as they are.
Adam:Yeah. Like, they have to come out because their head
Dax:Their head. Yeah. Like,
Adam:all that stuff. So it's
Dax:like a so we like compensate for Yeah. That by being able to take care of
Adam:Underdeveloped. Mhmm.
Dax:They're basically like not born yet, you know, and they're and then they come out.
Adam:I mean, they're pretty worthless until I mean, I've got a five year old that can't do anything for themselves. Humans are unique. Like, they take a lot of attention and care.
Dax:Training. Yeah.
Adam:It's kinda crazy. But welcome. Welcome to the wide world of parenting. Can't wait for you to relate with parents actually instead of just pretend.
Dax:I can't I can't wait I can't wait to annoy everyone. I'm so excited Oh, yeah. To be like
Adam:You feel like I was so right.
Dax:I knew. All you guys being like, parenting is so hard and like, you have to make all these sacrifices and you can't just have everything you want. I'm gonna I can't wait till I have a child just so I can tell everyone, you guys just suck. Mhmm. Easy.
Dax:Being a parent is easy.
Adam:The whole world is rooting against you right now. Now we're gonna have no empathy when when it gets hard.
Dax:At this point, I'm just gonna lie. I'm just gonna lie no matter what happens.
Adam:I'm just gonna be like Gotta hold up the shtick at this point. Yeah. Yeah. Like,
Dax:you guys just suck.
Adam:Yeah. We might.
Dax:I parents told me parenting was hard, but it turns out you all just suck. I can't wait to post that.
Adam:You can't wait to tweet it. You're gonna go ahead and tweet it. You just can't hold back. The the thing is, and this is gonna suck for all of us that want vindication, you could have an easy baby. Like, some of
Dax:them are easy. I've got to a very easy baby.
Adam:Yeah. So if you do, you're just gonna think we're all bad at this, and I hate that. I makes me wanna throw up my mouth. I really hope you have a normal to hard baby so you can experience the true reality of baby rearing.
Dax:Based off of the sense you have of the world and me, can you see me with a difficult baby?
Adam:I don't know how to answer that. What I'm predicting can I can I see how you No? Like like like, do do you
Dax:see that that's how the universe would play things out?
Adam:It would give
Dax:me a difficult baby? Or would it give me an easy baby to piss everyone off?
Adam:Probably. What
Dax:seems more like that?
Adam:Damn it. A ladder and I don't like it. I don't like it one bit. I'm gonna be so mad. I'm gonna be, like, trying to influence your baby at a very early age to be more difficult.
Dax:My baby's gonna come out at four months old.
Adam:Yeah. Okay. My youngest was two two weeks late.
Dax:So he's a little more
Adam:two than weeks late.
Dax:That's good. Three weeks late. You you saved two weeks.
Adam:Yeah. Well, I guess. Is that how I don't think that's how works.
Dax:Yeah. It is. It's like because look at it this way. If they come out one month early, when they're two months old
Adam:My oldest was three weeks early. So
Dax:They're kind of just still like one month old, you know?
Adam:Okay. Yeah. I it kind of makes sense.
Dax:So the reverse.
Adam:I hate that that makes sense.
Dax:It's kind of the case too.
Adam:Yeah. Okay. Well, congratulations to you both and good luck. Godspeed.
Dax:It's a girl by the way.
Adam:Oh, it's a girl. I didn't know this. I didn't wanna ask in case I didn't know. No.
Dax:You didn't know. Remember you offered up your children?
Adam:Damn it. I did know it. See, I'm so glad at remembering all things like human. Oh, I hate this about me. I do know that.
Dax:Your children for marriage.
Adam:I'm gonna tell Casey I did do that. Yeah. I'm gonna tell Casey how I forgot that it was a girl and she's gonna be like, of course,
Dax:she did. I bet she remembers it's a girl.
Adam:She probably does. Oh, for sure. The number of times I've like forgotten somebody who's pregnant that I'm like related to, like a sister-in-law or something, and then like end up in the room together and it's like, I totally forgot she was pregnant. Like, I'm so bad about like people's kids and relationships and understanding or remembering any of it. There's probably something there.
Dax:Well, it's worse if you thought she was pregnant and she wasn't. So Oh,
Adam:yeah. Sure.
Dax:Of the mistakes to make.
Adam:That one
Dax:is The other one is Yeah.
Adam:Alright. We got too much to do. We gotta go.
Dax:Okay. Bye bye. See you.
