OpenAI and MCP, Poking Replit, and Terminal's Launch Week
Anything that they put, like, gold foil on or something, ugh, love it. I just want that, whatever it is. Unless it's like a steak. Just give me that.
Speaker 2:Was gonna say it, but what about that? Coming to come down to Miami, they'll put gold on whatever you want. I want a sack of gold coins. How's it going?
Speaker 1:Good. How are you?
Speaker 2:I'm good. I feel like we have a lot to talk about.
Speaker 1:Yeah. I feel like it's been forever, and that was a pretty nonchalant little greeting I just did. I'm like, good. How are you? Good.
Speaker 1:How are you?
Speaker 2:Oh my god, Adam. So great to see you again. Well, has been a couple of weeks because we skipped last week and then because we were traveling and then we did and me and Sam
Speaker 1:You had Sam, the week before.
Speaker 2:Uh-huh. Uh-huh. People haven't heard you in a I
Speaker 1:know. I bet they missed me. Do you think they missed me?
Speaker 2:This is not my podcast.
Speaker 1:It basically is. You have a lot of guests on, very interesting conversations. Pop in sometimes.
Speaker 2:That's so nice. Yeah. You're still on the recurring guest.
Speaker 1:Yes. It's funny. There's so much to talk about actually. Like, an insane amount of stuff. By the way, I'm gonna this video, like, who watch this on YouTube, I'm gonna make it like cartoons or something just to fit the the current like theme of not the Ghibli style because that's overdone and I don't wanna like I don't know.
Speaker 1:People might feel bad about that. Just like some kind of like Muppets or cartoons or something. We're gonna be like South Park characters or something. I don't
Speaker 2:I'm down. Yeah. I saw some, I saw some examples of that. It was pretty funny. Unfortunately, the examples I did see that Andreessen Horowitz put out an example of like a clip from their podcast that was like Yeah.
Speaker 2:Gibbla, Gibblified. Uh-huh. Gibley, Giblefied? Whatever.
Speaker 1:I have no idea. Yeah. None of us, all of us white people out here trying to say the name this week. Gibley You know, that thing I
Speaker 2:knew all about before this week. Yeah. And then the per the clip they picked, I was like laughing so hard at it because the person was literally saying nothing. It was just like, you know, for b to c businesses, like, it's not you might think that you have to really do one thing, but, you know, in in times like this, like, you could really, like, like, look at the data and see maybe you need to do other things, you know? So we're not we're not, like, really prescriptive about what it was just something that was nonsense.
Speaker 2:Was like, I can't believe as a species we produce enough to allocate resources for that person to sit there and then just like say a bunch of crap. I think I'm being I'm also thinking I'm picking on this person a lot because I'm pretty sure they're the ones that made that MCP market map. Oh, jeez. Yeah.
Speaker 1:If I see one more, like, MCP thread, I'm gonna lose my mind. Yeah. There is interesting MCP knows news, though. Like, OpenAI did go Yeah.
Speaker 2:It's huge. It's
Speaker 1:MCP is gonna be the thing. If OpenAI accepted it, that's probably like the last straw. Right?
Speaker 2:Yeah. Yeah. The the protocol has been fixed to Oh, yeah.
Speaker 1:They made it worse.
Speaker 2:Way more. Yeah. Yeah. It has auth and stuff in it. So and everyone's like jumping on the server side.
Speaker 2:It's funny because me, Jay and Frank have had so many conversations about this thing we're building for OpenControl being like, from every angle we analyze it, we're like, this is the opposite of anything we'd ever do and we have so many reasons to not do it but we're gonna do it anyway. Like, just seems like a bad idea in so many ways just because The client side of it? Yeah. Because I'm like because like, why would we compete with ChadGBT or Anthropic or like these flagship AI products are gonna have millions of users?
Speaker 1:Yeah. I guess that's who you're competing with. Yeah. Okay. So on the client side,
Speaker 2:yeah. It doesn't make any sense. But at the same time, we're just like, people wanna use this like with their team and set it up and like add their team members Yeah. And just
Speaker 1:Oh, fits great into the the OpenControl thing fits great into an SSD app, like that's fantastic. Yeah.
Speaker 2:And even outside of it, like, because we're we're the thing that we're we've been debating about and we are building now is a hosted version, you just sign up and then you can, like, connect to your AZ account, connect whatever without having to do all that SC stuff. But, yeah, we're just like, I don't think ChatGPT is gonna build like ChatGPT for Teams where you can like have different environments and like different tools. It's just like, it's such a dumb idea. We're just like adding the basic SaaS stuff like Teams, environments, permissions to a client. But it's like so useful and and and we want it and everyone wants it,
Speaker 1:so And you guy you have this great distribution channel into a bunch of actual companies that use SST. Like, you kinda have this natural like, if it's a thing that every startup would probably wanna use or every legitimate company that uses SSCs, then, like, it's a product you should probably build because
Speaker 2:Yeah.
Speaker 1:You got all these customers ready for it.
Speaker 2:Yeah. And that's that's like the plus side, but, like, we generally are pretty extreme when deciding on what to work on. We're like, if somebody else should do this, we're not gonna do it. Or if somebody else decides to do this and it's obviously gonna be better Mhmm. We're not gonna do it.
Speaker 2:Because we're like, yeah, some YC company out there, if they just decide, oh, a % of time, we're just gonna focus on this thing. Yeah. Yeah. That's gonna be better than what we do because it's like a side thing for us. So we, like, understand that but at the same time, like, nobody's doing it for some reason.
Speaker 2:We're just focusing on the server side and nobody's on the clients and, like, all the clients are still, like like we're using it through Cloud Code which is like totally inaccessible for most people. Then the next level up is Cloud Desktop and OpenAI also said they're only supporting MCP on OpenAI Desktop. And I'm just like, why just make it a web app.
Speaker 1:Like, we'll
Speaker 2:just put
Speaker 1:in the web
Speaker 2:isn't it just a host? Yeah. Why not just a website? That doesn't make sense. Yeah.
Speaker 2:Well, I I think it's because all the MCP servers right now require, like, a thing to spawn. Yeah. Why? I still don't understand that either. It's again, it's a stupid it was like a stupid way that the protocol emerged.
Speaker 2:Like, that's obviously like a temporary thing. Eventually, Linear is gonna have a /mcp Yeah. Linear.com/mcp. Right?
Speaker 1:Yeah. I the whole thing was so ass backwards. Like, the fact that
Speaker 2:Yeah.
Speaker 1:The actual thing, the MCP server, or I guess the client what the thing that had to run like, there's this stateful piece. It had to run on, a long lived server, because I had this long lived connection. That was dumb, which I know you said they fixed. Then also, that all these things are local and, like, assumed to be local, and then, like, oh, this one's a hosted one. Why aren't they all hosted?
Speaker 1:I don't understand why we were so dumb suddenly. It's like we forgot everything about the web and HTTP requests, and we're like, let's just build something totally different.
Speaker 2:Yeah. I get why it makes sense for some things to be local. Like, obviously, like, a thing that reads like a tool that reads a file of your file system. Okay. That makes sense.
Speaker 2:Protocol needs to support that kind of thing. Yeah. But like getting someone to be some random person that's not an engineer to be like, oh, yeah, you gotta put in to your settings in Claude, NPX, you know, blah blah blah, whatever. Oh, by the way, you have to install Node. Js.
Speaker 2:Like, the this is like This is not consumer ready. We're coming in from this angle. Yeah. But yeah. Like, I think it'll all end up in the correct place.
Speaker 2:It's just like the ordering is, like, really random to me.
Speaker 1:Yeah. So I wanna I wanna pick your brain on this actually. There's there's some, like, very crazy convergence that's happening here. Like, my next task at statmuse is I'm trying to automate some of our support stuff because we get a lot of support emails, a lot of the same stuff. My cofounder has seen OpenControl stuff you've been tweeting.
Speaker 1:Like, he follows us on Twitter, and he he's very excited about doing that stuff with our internal analytics, just being able to like ask questions and get answers because it's just so much better than like we have like a QuickSight dashboard that's kind of a pain in the
Speaker 2:ass. Yeah.
Speaker 1:So the the question I have for you, like, my next task is to like do this stuff and I immediately am like, great, Dax is doing this, I'm just gonna use OpenControl. The the thing that I don't like, if we wanna build agents that aren't just prompted by us, like Mhmm. Prompted by events, like an incoming email, is that a thing that you see OpenControl dipping into or is that a totally different use case?
Speaker 2:Yeah. It's funny because it's like it's even for us, it's been so unclear about what is OpenControl. Because for a while It's like initially, it was like a way to find tools, but that was temporary because you couldn't run this stuff in a Lambda, you know Right. And we knew that was gonna be temporary. You know, MCP fixed it.
Speaker 2:So eventually, know, tool definition is gonna just be using the MCP SDK. Then it turned out, okay, OpenCode is just a client. It's just the front end part Yeah. Because for some reason, everyone is like, you have to install a desktop app to access this, so we have to build a client to make it available on web. But then it's like, what is a client?
Speaker 2:It's just a for loop. It's literally just a for loop calling the model and the model has all the magic built in, so there's not much for you to do besides to use it correctly and render
Speaker 1:stuff in an interesting deeper into that because I I know you said it's a four loop, but I need to better model that in my head. But you can keep going first.
Speaker 2:Yeah. Well, mean, I think it's worth explaining. So I think what was interesting about Cloud Code is you give it a task and it like does many, many, many steps. Mhmm. You know, try one thing, if it works, we'll try the next thing, if it doesn't work, you know, try something else.
Speaker 2:Mhmm. And maybe it's obvious in hindsight, but it felt like, oh, Claude Code is doing that's a good feature of Cloud Code, you know? That's like Yeah. That's a thing of it. It's like helping you it's letting you it's giving you access to this, like, multi step AI capability.
Speaker 2:Mhmm. But turns out that's not a feature of Cloud Code at all.
Speaker 1:That's just the model.
Speaker 2:They do a bunch of, like, nice UX things. Yeah. It's it's purely just Cloud Code is is just the UX of accessing it in your terminal. All it does is take your query, send it to the model remotely. The model replies back saying, hey, can you call these tools and give me the result?
Speaker 2:Cloud code needs to call the tools, which in this case, you know, read a file, run a bash command, whatever. Mhmm. Prompt the user to make sure that this is allowed all that all that like client facing stuff. Get the result, send it back to the model, then the model replies to the next step. Run it, send it back and forth, and eventually, the model's like, I'm done.
Speaker 1:Okay. I need to go slower for the people in the back. I'm one of the people in the back. Because I think you have some knowledge that maybe is helping you just like go pretty fast to this. So I wanna walk through it from the beginning.
Speaker 1:So Cloud Code or any of these, I guess, a cursor agent or whatever
Speaker 2:Yeah.
Speaker 1:It starts with your prompt. Like, I'm imagining one big, like, text file or markdown file or something, and at the top of it is my prompt. It sends that to the model immediately. Is that what you're saying? Yep.
Speaker 1:Starting.
Speaker 2:Okay. Along with the list of tools that are available. Okay. That's It also sends
Speaker 1:the list of tools. Okay. That's interesting because I didn't really understand how MCP works. So it sends it the list of tools, which has, like, some metadata about, like, the different, like, you can pass in.
Speaker 2:Yeah. The tool name, description of what it what it can do, the the the schema required to call it. Yeah.
Speaker 1:Uh-huh. And then the response back on that first call, is it more text or is it like always JSON? Is it like how what is the mechanism that it informs the thing how to call
Speaker 2:the tool? It'll almost always have text. So there'll be one So one part of the response will be the message which is like the text being like, oh, I'm gonna try this. Yeah. And then it also contains a list of tool calls it would like to happen.
Speaker 1:And those are formatted how?
Speaker 2:So it conforms to the schema you sent up originally. So it'll say the name of the tool, it'll say and then it'll give you the exact arguments to pass the tool in the schema that you can train it to.
Speaker 1:And is this part of the MTP protocol? Like, that is the definition in terms of those tool calls or is that totally different?
Speaker 2:No. The MCP spec is just a way to hit a server and be like, hey, what tools are available on the server? And then call those tools.
Speaker 1:Okay. Okay. Okay.
Speaker 2:So it it's actually no different than an open API spec where you hit slash doc and it gives you back everything that's available. Yeah. And then you can call, you know, those endpoints. It's it's no different than that.
Speaker 1:That's MCPA. Okay. So oh, no. I'm one of those people. I probably tweeted something about MCP during that week, and I'm one
Speaker 2:of those people that doesn't actually know
Speaker 1:what it is.
Speaker 2:Damn. I think you only made fun of it.
Speaker 1:Okay. Good. Perfect. I'll just stick to that policy on Twitter. Just make fun of stuff.
Speaker 1:So okay. So after that call, it then it's literally, like, making API calls basically into these tools, whatever the interface is into the tools. Yeah. It's getting the response back, like, standard in? Like, what are these tools sending back an HTTP response?
Speaker 1:Like, what is the response from the tool?
Speaker 2:And so now now the client's job is to call the tools before the client the model say, hey, can you call the schools? Now the client's job is to call the tools. There's a spec in in MCP to, like, be able to call a tool. Okay. So it just hits the MCP server Mhmm.
Speaker 2:Whether it's through web or whether it's through standard in for, like, the ones that spawn locally like we talked about. It's, like, kind of stupid that all of them run locally. Yeah. The the transport protocol basically doesn't matter. Right?
Speaker 2:It's just Yeah. It's It's sending a command to call the tool and the MCP server itself runs the tool. So if the server's remote, like Linear's MCP server, it might, like, query their database internally and then send the results back. If it's local, it might read a file.
Speaker 1:Do MCP servers, like, when you define an MCP server, do you define the response shape as well?
Speaker 2:No. No. No. So the MCP server
Speaker 1:Are they
Speaker 2:always just there's no nothing in the spec about response shape because
Speaker 1:It's just text assumed or could it be
Speaker 2:binary? What? Just text. Literally be JSON, it can be text, it can be like anything and you don't have to define it at all because Okay. LMs are super flexible.
Speaker 2:They get a response and then they they know they know what to look at.
Speaker 1:And that's where I was headed. So basically, this client makes the tool calls and then just dumps whatever comes back from that tool, whatever it is, into the the big file I'm imagining where our prompt at the top. Yeah. Yeah. It's just building up this huge file Yep.
Speaker 1:As you go. That's kind of the idea. It's what you get the feeling with Cloud Code. Like, it's kind of like it's truncating a lot of stuff. It's not maybe formatting it the way it's actually in the file, but you kinda get the sense that you're just working through this big, long file.
Speaker 1:And that's where eventually you hit, like, a context limit, I guess, the longer you use one session.
Speaker 2:Yeah. It's funny because it feels very inefficient. Like, every single iteration of this, it sends
Speaker 1:It's getting the whole thing.
Speaker 2:The whole thing back and forth. It's not like you create a session and you, like, append to it. It's Mhmm. I think these APIs probably will evolve to be more like that. Yeah.
Speaker 2:Because sending the whole context back and forth every time is just like, you know, it can be pretty large. Yeah. But, yeah, that's what it does. It sends it. Then eventually the model stops.
Speaker 2:At some point, it'll respond without any more tool calls. And it'll say there's basically like a stop reason. It's like why did the model stop responding? Mhmm. And, you know, for the first few iterations, it'll say stop reason tool call.
Speaker 2:I stopped because I need to call a tool and I need you to call a tool and tell me how to continue. Mhmm. And I think there's another status that's like end turn or something, which means like, at this point, I I'm I'm done. Yeah. And specifically, Anthropix models, specifically three seven, it's very per it has like a lot of perseverance.
Speaker 2:That's like the feeling I get. Right? So it'll like try back and forth a loop bunch a bunch of times before giving up Mhmm. Either because I got an answer or like it's like I'm trying too hard and I'm not getting anywhere. It'll give up.
Speaker 2:Other models, most other models rarely do more than one or two steps. Like, they just Really? Give up very quickly. I've tested all the I haven't tested the new Gemini model that came out two days ago, but I've tested all the other ones. Mhmm.
Speaker 2:They give up pretty quickly. Almost all of ChatGPT's models give up very quickly. O three mini, I think, has this capability in it and it's it's like comparable to Claude, but not as good.
Speaker 1:Yeah. Okay.
Speaker 2:So, yeah, it's just like a it's it's just a capability of the model to know that it can keep trying.
Speaker 1:Well, they announced three seven with Claude Co, didn't they? I'm realizing now. Yeah. So I okay. This makes me feel a lot better if tell me if this if you think this vibes if this is true.
Speaker 1:No cap. For real. For real. Just trying to decide if vibes was a way of, like, saying it's true, and I was like, no, that's no cap. But that's sorry.
Speaker 2:That's what I'm arguing with.
Speaker 1:So I was feeling like, man, I love Cloud Code. I love being able to use my normal text editor, and I love just a CLI interface that interacts with my file system. Fantastic. Yep. My main problem is feeling like the new models come out, and you wanna try
Speaker 2:the new model. Yep. Good. Good. Good.
Speaker 1:And if you're stuck on the Anthropic model but now I kinda feel like maybe it just doesn't work with other models because, like you're saying, they they give up too quick, and that's why 3.7 is how you use Cloud Code.
Speaker 2:Yeah. And I had the same feeling because I also wanna try different models because we're trying to build this thing, and we're like, does everyone need to sign up for Anthropic? Like, can people use separate models? So I had to build my own client to even test this to be able to swap out the models.
Speaker 1:Mhmm.
Speaker 2:Similarly, I think someone should just build a Claude code type interface.
Speaker 1:It should just be an open source thing. Right?
Speaker 2:There should be one. Unplugable. Right? Yeah. So these models will get better.
Speaker 2:There will be cheaper models that can, you know, there's some trade off there that will be made that might make sense for a lot of people. The the innovation in Cloud Code is, hey, let's build a terminal UI. Yeah. That's like that's that's innovation effectively. That's it.
Speaker 2:It's not the model.
Speaker 1:And it's good. They they they move really fast. Like, I don't know if you follow Yep.
Speaker 2:Yep. Them on Twitter.
Speaker 1:Yeah. Just constant features. They did add like a VIM mode because one of my big complaints is like, I end up writing the text somewhere else and then coming back in and pasting it because Yeah.
Speaker 2:It's just
Speaker 1:hard without VIM. You're just in there in a terminal. You're used to they added VIM mode. It is, like, missing so many things I use. I know.
Speaker 2:I turned it off useless to me. Yeah.
Speaker 1:Like, I don't and I don't even I'm not smart enough to know, like, what's classic Vim versus what's some, like, key map I have in Neo Vim that's not normal. I it's like I just I realized it's very hard for anybody to ever add Vim support that I'm gonna actually use.
Speaker 2:They should just open up Vim inside
Speaker 1:Could they do that?
Speaker 2:Oh, just inside the tube. By multiplexing, which you know, I'm one of the world's experts
Speaker 1:in multi terminal experts. True.
Speaker 2:There's room for a tool where it's like half of it's like a Vim editor and then Yeah. Or like, you know, you can turn off Vim mode if they're make it accessible.
Speaker 1:Nice. Because you write you're writing a lot and you can imagine if the models just keep getting better. Let's say Cloud four comes out and it's like, oh my god, this thing's better than me. You can imagine we're gonna spend more and more time just writing prompts like this. I know.
Speaker 1:I hate to even
Speaker 2:say it out loud. I feel like
Speaker 1:a prompt guy or like a prompt engineer truther. I just always made fun of those people. But you can imagine, like, I would be typing more in Cloud Code than I type into my actual editor, and when that happens, I'm gonna need it to be good. I'm gonna need to be able to fling some text around
Speaker 2:Yeah. Neolithm that typing speed is suddenly back back as a factor that matters, you Yeah. I'm typing out a lot of sentences lately.
Speaker 1:Yeah. It's I mean, that you get better results, at least this current version with three seven. Like, you do get a lot better results when you type out a whole lot of detail, and it saves time still. It's it's kinda like, well, if I'm gonna write out all this explanation, why not I just do it? It's like, no, this would be a really annoying task to do and it's much easier to write out a paragraph and a bullet list.
Speaker 1:Like, I'd rather do that than do the task. So yeah.
Speaker 2:I think the other thing I found is, man, does it benefit from like mature code bases with patterns? Like, the terminal code base is very simple and like we've been pretty good about it's like a pretty clean code. It's like one probably one of cleaner code bases that I work on. Yeah. And Claude is so effective in there because it just knows It is.
Speaker 2:It's like so guided.
Speaker 1:Yeah. And then in older code bases, not so much.
Speaker 2:Yeah. Not so much because there's like older patterns that you don't want it doing anymore or like Mhmm. It's like stuff that you don't use anymore that it starts to like build on top of. So, yeah, weirdly, it's all the same things as stuff that would help a junior engineer. Like a really clean code base helps a junior engineer a lot.
Speaker 2:It's the exact same situation. And I and I found that there's almost something me and Liz are working on. I realized that we should change some of where we land on UI patterns because putting a little bit more effort into making more consistent UI patterns, more composable UI patterns. If you have that in place, like Claude is so good at just building new features. Like, there's so many feature requests I see for that C console where I'm like, if the S console was like 30 or 40% better, like, in terms of stuff we need to clean up, this is just like a single message I send to Claude and the whole feature would have would have gotten built.
Speaker 2:Yeah. So Yeah. My motivation to fix tech tech debt is a lot higher now.
Speaker 1:Yes. That's exactly how I feel it. My day job Yeah. At Sammy's, like, we wanna I just wanna aggressively, using these tools even, make the code base modern, use all the same kind of patterns, like a lot of the stuff we do at Terminal. I just want it to be as effective in that code base because it's it's not when it's when it's kind of when you've let things go or it's just older patterns that trying to move away from and you're half migrated, like that kind
Speaker 2:of stuff is just a nightmare. Yeah.
Speaker 1:Yeah. But yeah, I've noticed that as well. Terminal code base, it does does a really good job.
Speaker 2:Yeah. Yeah. Like in the in the console code base, we're like halfway between two databases that are moving to Postgres and like, yeah, it just gets confused or like it just starts I'd like tell it to, oh, like make sure you update both and make sure you sync the data, it's just like a lot more work for it. Yeah. Yep.
Speaker 2:Yeah. So I'm pretty excited about that. The other thing from this week is because you know what? There's a whole so I think this whole episode is gonna be AI related. Yeah.
Speaker 2:Azure Probably
Speaker 1:probably from here on out. We should just rebrand All my thoughts?
Speaker 2:This point. All exactly. My thoughts. Every single day have just been about AI. Don't think I've Yeah.
Speaker 2:It's it's crazy.
Speaker 1:I mean, all of my thoughts about work, I do have other thoughts, Casey, in life. But, like, work related thoughts, it's AI. I mean, really, like, the it's just it's it's not gonna get any slower. Things are just gonna keep happening faster. It feels so cool to be able to watch it all unfold.
Speaker 1:Like, you just know there's another frontier model coming any every couple weeks. Like, there's all these companies competing. Yeah. And it's just like seeing the steps up, and then something big will happen even, like, after all the what? Two years of AI hype and things been going on.
Speaker 1:Has it been two years? I don't know. When was ChadGBT first out? That was kind of the big first moment. My point is, after all that time, whatever it's been, then you have a moment like this week where this image model comes out, and it's like Yeah.
Speaker 1:Still blowing us all away, and it feels like the original ChadGBT moment. Like, this is crazy. Yeah. It's nuts. Yeah.
Speaker 1:It's very exciting.
Speaker 2:It's pretty fun. But the reason I was I was saying all this is because I know you saw my post today. Me, then you, then Prime, just all was all three of us independently poked Replit?
Speaker 1:Oh, really? I didn't see Prime did. I knew you had before.
Speaker 2:So, like a few weeks ago, I posted something about how competitive the vZero Bolt lovable category of product is. There's like so much competition there and it's, like, it's really intense. And I just, like, was commenting on that. And a bunch of people replied being, like, oh, like, and also a replet. And I thought about it and I was, you know, I've only ever heard replet being mentioned by VCs, really.
Speaker 2:And I Yeah. And I posted that. And you have the same thing. I'm like, this is really just a fact about myself. Like, it's just an observation.
Speaker 1:Yeah. I didn't say any negative. Like, I didn't even say anything negative. I just said I'd because I don't. I don't know a single person that uses it, and I hear about it too much on Twitter from, like, the founder to not know anybody that's ever used it.
Speaker 1:I'd literally hardly know what it is. I guess I know it's in that category. It helps you. The cursor kind of, maybe. I don't know.
Speaker 2:It's it's it's so I posted that then I and and to be fair, like, knew I it was like, obviously, it's provocative. I'm not gonna be Yeah.
Speaker 1:I mean, it's Twitter.
Speaker 2:That's that's what we do. I Twitter. Yeah. I'm not gonna say it's not. And I know the CEO is like a hothead.
Speaker 2:So he
Speaker 1:I did not know this.
Speaker 2:He quote tweeted me being like, is this hater like, fellas, is this hater correct? Prove him wrong or something like that. And then, you know, a bunch of people were playing being like, I use Replit and I love it. I'm a big fan of Replit. Like, I use Replit every day for blah blah blah blah blah blah.
Speaker 1:You fixed my profile. VC. No joke. I I did that. I went to his page.
Speaker 1:I was like, man, is am I just missing it? Like, are these people because he kinda replied to me, and it was whatever. I actually don't really understand what he was saying to me. He said, I just said I don't know a single person that uses Repli, and his answer was like, let's
Speaker 2:keep it that way. Don't
Speaker 1:actually know what that means.
Speaker 2:Like, do do I think I think he's like, I don't even want you knowing about Repli because I
Speaker 1:hate He doesn't want me as a customer? Okay. It's like, let's keep it where I still don't know any single person that uses Replic. That sounds like a bad deal
Speaker 2:for your business. You should
Speaker 1:probably want me to know some people that use it. Yeah. I went to his page, and I, like, was, like, trying to see, you know, people that praise replicas and retweets them, and two of them were actually VCs. And I just think that's really funny.
Speaker 2:I don't know. Yeah. It's true. It's it's just like a VC thing. Well, I mean, they're also like they raised at like a 1,000,000,000 1,200,000,000.0 valuation during like peak crazy times Wailing.
Speaker 2:Like when Saudi money was just everywhere. Yeah. Then they had I think they had like 3,000,000 ARR at that point. And then seeing like the level Do people use it? Yeah.
Speaker 2:People use it. Like it's it's used. It's like lot of people use it. It's true. But at the same time, they're it's a place you go and you get free compute.
Speaker 2:So it's like, what does usage even mean? Like, they're just giving away Yeah. Free compute? Yeah. And like, of course, like, they've done some interesting things in the product too.
Speaker 2:I like I looked I looked more into it. But again, it's just funny that despite all that, you are competing with vZero lovable Bolt and I like just don't hear you guys mentioned as much in in relation. Yeah. That was my point, right? Mhmm.
Speaker 2:But of course, a bunch of people are and one specific reply was like, oh, these people don't, like, replying to the CEO, kind of supporting him being like, oh, he doesn't get it referring to me. People don't realize that like This is one of those posts that's like, oh, people haven't realized how good AI is. And I'm just like I
Speaker 1:hate that. Hate that so much.
Speaker 2:It's just it's just like, you fucking moron. I'm looking at this stuff as much as you are. Yeah. If not more. I'm probably doing more than you are in this space.
Speaker 2:Uh-huh. Anyone that It's just like people, like, are so excited to be like, hey, I know something. They assume that anyone having any kind of slightly opposite take is, like, totally clueless and lost and can't see the future while they can. Uh-huh. And it's like, that that it's just so annoying.
Speaker 2:I find that so
Speaker 1:annoying. This up before, and I've had this experience as well, where the people will also assume you're like this, like, fifteen year software engineer. I care about code quality. That's all I care about. You're just making bad code, and I make good code.
Speaker 1:Like, if you've ever looked at anything I've contributed to open source or on the public Internet, I do not care about code quality. Not even a little bit. So please don't label me as a non AI, like, AI hater that just cares about the quality of code for the quality of code's sake. Like, I don't care about products or making money, just code. That's the people they they label us as that.
Speaker 2:Yeah. They they they figure out the most stupid, like, exaggeration of the opposite side and they like, you under it just so because like, yeah, you you don't have any intelligent thoughts, so you can only, you know, win the argument if you make the opposite side look stupid. What's so funny about what you just said is think about the first half of our conversation. We're in super deep trying to like make to get the most out of these LMs. Yeah.
Speaker 2:And we've discovered that code quality actually
Speaker 1:It actually does does matter.
Speaker 2:Like, it turns out, like, yeah, like, the LMs perform better when you're more organized. And we're now trying to be even more organized than we were when we were like not using AI.
Speaker 1:So Yeah.
Speaker 2:That's a good point. This is a kind of like counterintuitive thing that shows up when you're actually doing stuff versus just being some person being like, I know the future and and nobody else does, but you're not actually doing anything.
Speaker 1:If they're gonna bucket us like that, let's bucket them. They're just these, like Unemployed. Unemployed, no life, like, big idea guys. Is that how that
Speaker 2:big idea guys. They don't even have big ideas.
Speaker 1:They think they have big ideas and that they've just been screwed.
Speaker 2:You know what they
Speaker 1:are? What?
Speaker 2:They're sports fans. They're huge fans. They're giant fans. But they have literally have never even touched a ball. Yes.
Speaker 2:They think that if they are big enough fans of a pro athlete, they're effectively like the same as a pro athlete.
Speaker 1:Yes. They're lumped in there with the successes of their team that they cheer on. That's Yes.
Speaker 2:That's a
Speaker 1:good call.
Speaker 2:That's what it feels
Speaker 1:like to don't know why there's this group of people that seem to think like they're in the select few. They're the chosen ones that will be they'll make it in the AI future. And all you software developers that we've paid so much for so long, you're all gonna burn. Why I don't know why like why that's so attractive to this group of people, but they really think somehow they're gonna be spared. Like, they're gonna have this amazing life where they live with the benefits of AI.
Speaker 2:Yeah. I don't get it.
Speaker 1:Like Developers, not those guys.
Speaker 2:If I this is like so so far from the truth. But if I'm somehow pushed out of a space that I'm in, I'm gonna enter your space. Like, that's just what's gonna happen. Right. And that's not gonna be good for you.
Speaker 1:Yeah. Okay. Okay. I'm done. Let's just get that part of it out of our system and we're gonna move on.
Speaker 1:Yeah.
Speaker 2:Yeah. Pete Bertrand messaged me today and he had a funny observation. He's like, man, bad week for CEOs on Twitter. It's just like
Speaker 1:Oh, wait. What else? I've maybe I'm forgetting.
Speaker 2:The whole Guillermo seat with Cloudflare, that whole thing.
Speaker 1:Oh, that's
Speaker 2:what I did. I made a song out of out of Guillermo's Yeah. How good was that?
Speaker 1:That was really good. There's a huge trend
Speaker 2:to your product? They
Speaker 1:really went at it. That was some, like, no holding back kinda exchange from, like, pretty big figures. That was a that was surprising.
Speaker 2:Cloudflare's held back a lot. Cloudflare's held back a while. No. Well, I mean, Guillermo has taken a lot of direct shots at Cloudflare, and Cloudflare's always, like, been like, well, we're the bigger ones, so we can't say anything. And this time around, they they went the other way.
Speaker 1:Yeah. Oh, yeah. We have we haven't done
Speaker 2:an episode since the vulnerability. Right? The big No. I don't think we have. Yeah.
Speaker 2:I mean, I I wrote like my big summary on it. I just think it was like Yeah.
Speaker 1:We can
Speaker 2:just not. Yeah. Another situation where everybody involved is just being annoying. Both sides.
Speaker 1:On both sides. And so many of the arguments, I just feel like I really just like both sides. I don't know where that puts me.
Speaker 2:All these people being like, how could Guillermo not apologize to me after this vulnerability? They handled it so badly, and I deserve an apology. Where where is he? Where is he? It's been twenty four hours and we still haven't And then he eventually did apologize.
Speaker 2:He goes like, oh, we missed the mark. Guys, we missed the mark. We're
Speaker 1:so sorry.
Speaker 2:Besides, yeah, when they probably do apologize,
Speaker 1:we're not happy still. We're just never gonna be happy. Maybe we just don't like people, Vax. Is it possible we just don't like people?
Speaker 2:Because Us, me too. For the comment. Yeah. I just think I don't like a lot of people. I I definitely I think I have tried hard in my life to be someone that and it might not seem this way, but to be someone that I know where this is headed.
Speaker 2:Isn't someone that's like, everyone sucks. I'm better than everyone. Like, I'm try I'm, like, trying to be reasonable.
Speaker 1:Tried so hard. Yeah. Mhmm.
Speaker 2:And nowadays, I'm like, you know what? I'm just better than all of you. Like, I am better than most of you people. Like, you suck. All of you guys suck.
Speaker 2:Oh, it's funny. That is me. Me and my friends. We're all better than you guys. Mhmm.
Speaker 1:Oh, man. We're doing something. Something some psychologist is listening to this, and they're like, classic whatever the phrase is. You probably know it. You always know, like, the syndromes and the Classics
Speaker 2:being right it is.
Speaker 1:Being right? Okay. Well, I'll take it. What man, there's so there's actually so many things going on. We're gonna have to limit which things we can talk about, which is is unfortunate.
Speaker 2:I I don't remember anything else, to be honest.
Speaker 1:Oh, okay. Well, I mean, the terminal stuff, we just we just did a launch week. Launch weeks suck, turns out. So we we were making fun of launch weeks. I mean, the whole thing was like, we're mocking a launch week.
Speaker 1:And then I couldn't help but get a little pissed when it didn't get that much traction.
Speaker 2:I'm like, come on. I
Speaker 1:attended our fake launch week.
Speaker 2:Oh, wait, wait, Adam, but I have the best news. I have the best news. Oh, okay. Do you know who else did a launch week this week? Wait.
Speaker 2:Who? I don't. Exactly. You don't know who else did a launch week this week. Right?
Speaker 1:Someone else did what?
Speaker 2:Someone else did a launch week this week. And and it did even worse than ours.
Speaker 1:Oh, woah. Who was it? That'll make me feel so much better. I'm gonna look at every one of their posts.
Speaker 2:It's gonna make you uncomfortable, I think.
Speaker 1:Oh, no. Is this somebody I'm friendly with?
Speaker 2:Damn it. Actually, I always do this. I always do this.
Speaker 1:I've lost friends over this podcast from
Speaker 2:this too. I do always do this. It was recent.
Speaker 1:What was recent?
Speaker 2:Resend. It was recent.
Speaker 1:Oh, recent. Oh, no. It's okay. That's the one okay. So I just keep losing the friendship of someone I wasn't really friends with.
Speaker 1:They stopped following me on Twitter, and I still use Dracula, and it just bothers me.
Speaker 2:You guys stop using Dracula. Yeah. So Resend had a launch week this week as well.
Speaker 1:That's so funny.
Speaker 2:And I was looking at I saw their post today about something, and I was like, this did even worse than ours.
Speaker 1:Hey, that's so good. I'm so glad. I mean, not for them. I feel bad for them, but, like, at least I don't feel quite as miserable about how bad launch weeks are. Well, it's not even like like, oh, it's so bad.
Speaker 1:It was just like we could tweet something random in between the launch week post, and it would blow up and like do great.
Speaker 2:Yeah. It's just like
Speaker 1:people just do not wanna hear about launch week.
Speaker 2:Yeah. By comparison, it was it was much worse. Yeah. It's just like, I think I've known for a long time that launch weeks are bad because and maybe at some time they were good, but just people's attention spans and like the way the internet works now, like Mhmm. You need like punchy, like Yeah.
Speaker 2:Hyper relevant stuff
Speaker 1:Mhmm.
Speaker 2:That's like straight to the point, you consume it very quickly. And launch features just so slow. They're so slow and like, it's very hard to grab people's attention.
Speaker 1:You're dragging thirty minutes of news out over a week. So, of course,
Speaker 2:it's slow. It has to Exactly.
Speaker 1:Yeah. Is it just like it's also just like too much navel gazing or something? Like, too much like, we did things. Look at us. Like, there's something on the Exactly.
Speaker 2:I mean, we were doing it as a parody, so it was it was fine.
Speaker 1:We have a past for everything we do because it's all fake.
Speaker 2:It's all fake. Yeah. Yeah. It's it's just funny that like I don't know. So the I mean, okay.
Speaker 2:So to and we talked about this already but we actually didn't like intend to do a launch week. We just had way too many half starter projects and we needed to force them all out. Mhmm. But we can go back to focusing on one thing at a time which I like really wanna do Yeah. Moving forward.
Speaker 1:Long as it's the one thing I want to do, then yes, I agree.
Speaker 2:We should do one thing. Yes. Should do one thing at a time. So that's why it was like an excuse to to get that out. And then this whole thing where I I I can't figure how to articulate it.
Speaker 2:So there's something so funny about us doing it because we think companies are stupid for doing it. And then we just so deeply experience the stupidity personally of it Yeah. To a point where I felt bad about the launch week. But I'm like, no, this is a joke. It, like, wasn't supposed to work.
Speaker 1:Yeah. I know. It's so stupid.
Speaker 2:And also, our sales are, like, incredible for those Oh, part's real. The end thing did work. Yeah. It's just weird by comparison where we post a stupid tweet and it gets, like, five extra impressions as anything on
Speaker 1:our on our launch week thing. It really was a lesson for all of you out there. If you have a startup, just maybe don't do launch weeks anymore. Let's just make it the new trend where we just do like a day a single day, a launch day, and just pack it all into that one day, make it really punchy, make a really cool video where there's stuff exploding or something. I don't know.
Speaker 1:Whatever whatever the kids need to get their brains engaged on Twitter anymore. Just do that.
Speaker 2:Yeah. Yeah. Exactly.
Speaker 1:We we learned the lesson.
Speaker 2:I'm looking forward to finally I mean, feels good there's a clear all that stuff hanging over our heads.
Speaker 1:Yeah. Real.
Speaker 2:Like, the broadcast extension, the Europe stuff. But now, we can just focus on just the next thing coming up, which is React Miami and making sure that's that's really good.
Speaker 1:Yeah. I feel like I'm so exhausted and just like Yeah.
Speaker 2:It's like a lot
Speaker 1:to do there. Yeah. It's in real
Speaker 2:I don't think there's that much to do. Oh, good. I mean I mean, I'm always that statement is basically always wrong. So Yeah. Maybe I
Speaker 1:shouldn't say it. It really it really is. Yeah. So I guess I'm breaking this to you on the podcast. Why not?
Speaker 1:I'm gonna come. I can't be there for seven days. I'm sorry.
Speaker 2:That's fine. Yeah. Okay. It's a lot.
Speaker 1:It's difficult to pick the days, like, how what stretch is the most important stretch.
Speaker 2:Do the first half, I think.
Speaker 1:That's what I think. Yeah. So I'll be there for the Monday and Tuesday activities, for sure. I think the most important.
Speaker 2:Yeah. Yeah.
Speaker 1:So I'll come Sunday and then I'll yeah. Okay. So I'm gonna miss like the conference, which I'm fine with. I mean, Michelle does a great job. Conferences are cool.
Speaker 2:I don't need you, basically.
Speaker 1:It doesn't need me. There's nothing there
Speaker 2:that, like not need you. Needs me,
Speaker 1:and my family needs me. So, anyway,
Speaker 2:yeah, that's gonna be just little breaking news for you there. I think what I really liked about I mean, the other thing that was going on this week is PrimeTeage and all those people were locked in a tower in terminal tower, doing their game.
Speaker 1:For twenty four seven streams.
Speaker 2:Yeah. And I think what was really good, which I didn't I didn't really understand going into it, was those recap videos they do every day.
Speaker 1:I where are you seeing those?
Speaker 2:I didn't see They're on YouTube. They're on Prime's YouTube. And they're so funny. They're really funny. And they're they're a little bit too long for me, but
Speaker 1:Yeah. I saw you complimenting them.
Speaker 2:I burst out laughing a bunch of times. Yeah.
Speaker 1:And I just had to nod and smile because I did not actually know what you're talking about.
Speaker 2:Yeah. Yeah. So they they do these recaps, which is all like behind the scenes footage cut together and they do like the reality show like confessional thing where they're like talking to the camera.
Speaker 1:Are they shorts?
Speaker 2:What? It's just one long video. It's like it's like But like back to back like short clips.
Speaker 1:Prime's other channel.
Speaker 2:And I'm like, we need to do that for React Miami where we have cameras on us the whole time, basically. Mhmm. And, you know, we're just capturing like all the stupid stuff when come out. How are
Speaker 1:we gonna how are we gonna do that? We need somebody like doing that.
Speaker 2:He's gonna do So I'm gonna talk to Vegan this week because he basically figured that out for them.
Speaker 1:Yeah. Yeah. Yeah.
Speaker 2:And we have like, we have some people coming already that are gonna help us with that, but we might just hire someone full time. The sumo based people link me to the guy that they're using and he is really cheap. So, I'm gonna see if that's like an option. Uh-huh.
Speaker 1:What is it? The PrimeTimeAgen? What what is his other YouTube channel?
Speaker 2:The PrimeTimeAgen. Oh,
Speaker 1:wow. His the PrimeTimeAgen channel is way bigger than his original channel.
Speaker 2:No. It's it's just called the prime oh, it's called the PrimeTimeAgen. You're right.
Speaker 1:PrimeTimeAgen, but it's the PrimeTime.
Speaker 2:Was his name.
Speaker 1:Oh my god. I didn't know he's getting close to a million on that channel. Like, I thought the other one was still bigger, but that was probably the information from years ago in my head.
Speaker 2:So what I wanna do is to treat React Miami as a way to kick start our YouTube. So if we just make these daily Yes. Things like this and we post them just on our YouTube and nowhere else, that'll help us at least start to build up some kind of audience. Yeah. But yeah.
Speaker 2:I love it. It it was like little blooper type things. There was a line that Ben delivered from Kerzam, Ben of Kerzam. He he goes they're making they're they're asking me if he's lefty and he's like, what? No.
Speaker 2:I'm not lefty. Like, while they're like, while they're working. Yeah. And it's like kind of funny interaction. And then it cuts to him it cuts to him like in a different room like talking to camera being like, I'm not lefty.
Speaker 2:I just like to ride the mouse goofy. I found out the way he delivered it was so funny. Oh, that's funny. Like a skateboarder? He's like, made, funny hand gesture.
Speaker 2:He's he's like, he's like, naturally, like, so funny. He's so funny. I I love that guy. Yeah. So just stuff like that He's
Speaker 1:gonna be he's gonna be in Miami, right?
Speaker 2:Yep. He's gonna be staying with us. Yeah. So, yeah, I think there's just a lot of opportunity for hilarious stuff and all of it just needs to be captured and cut up and Yes. Uploaded.
Speaker 2:And I think it'll do really well.
Speaker 1:We really need to hire someone. We're we're looking to hire. If you're a podcast listener and you're below the age of I'm sorry, this is discriminatory, but if you're below the age of, like, 28
Speaker 2:Is that a have the energy and the sense
Speaker 1:of humor and the If
Speaker 2:you're an Internet native person would be considered a Zoomer.
Speaker 1:Yeah. You know, you're you're right. It doesn't matter. If you're 50 years old, but you have, like, Internet native personality and you have media skills and social media skills, video editing skills. If you have all of that, please reach out.
Speaker 2:Have no friends in real life. Yeah. Yeah. If all you do is go the internet and make jokes. Yeah.
Speaker 1:That you're our guy or girl.
Speaker 2:Yeah. Looking to hire someone for some work.
Speaker 1:Yes. Just reach out jobs@Terminal.shop. Terminal dot shop something. Just DM us. Find a way.
Speaker 1:Life finds a way if you're really serious.
Speaker 2:You're on the Internet. You can figure it out.
Speaker 1:Yeah. Yeah. Yeah. A minute. Are we recording?
Speaker 2:Yeah. We've been recording.
Speaker 1:Okay. The button looks different. I think Riverside changed the button and that just terrified me. Like, I'm not gonna spend thirty minutes talking to Dax if it's not content. God.
Speaker 1:I'm just kidding.
Speaker 2:Well, that's kind what I'm saying. Like, I'm like, we're gonna spend seven days together, we might as well We might as well get turn to content.
Speaker 1:Yeah. It really does.
Speaker 2:Yeah.
Speaker 1:Okay. Is there not other stuff going on? I feel like there's more stuff. I feel like the AI stuff, there's just the the well, we didn't really talk about the image model. How much you've played with it?
Speaker 1:What are your thoughts on? It's another example of the zeitgeist being so intense. Every time something catches fire now, it's like literally every post on Twitter. It's nuts. But what are your thoughts on on the new the new image thing?
Speaker 2:I think I posted something today that I I think it helped me clarify a feeling really intensely, which was kind of cool. So I saw a bunch of posts being like, by the way, hate these. I was advocating for the death penalty for people that post like this. They were like Photographic. Over, we're cooked.
Speaker 2:This like graphic designers are cooked. They don't even know what's Oh, yeah. Right? A bunch of those threads showing like, hey, pick product photography and they were able to like make a model holding it and it looked like professional ish photography. Yeah.
Speaker 2:I don't really have an eye for it, so I don't know if it actually did. They're talking about how graphic designers are like or designers in general should be should be threatened by this model. So I started playing with it for some UI stuff and I did some it was pretty interesting. So I, like, gave it an image of something that I liked, like the vibe of something that I liked.
Speaker 1:Mhmm.
Speaker 2:And I was like, hey, create a paragraph. Even before I did that, I was just like, hey, based off of this, like, show me what a UI kit based off of this vibe would look like. Mhmm. A button, a dialogue, checkboxes. Yeah.
Speaker 2:And I drew something that was kind of interesting, like, incredible, but like, you know, it it was it did match it. Then I tried saying, hey, describe the vibe that this image captures in like a paragraph, describe it to me in detail. And it did a really good job with that. And I was and I iterated and I was like, okay, strip all color information, strip all specifics. It gave me like a really good description of what I liked about this image that I couldn't really articulate myself.
Speaker 2:Yeah. Then I would use that as a prompt to generate other stuff in that same style, which again was like interesting, like better than stuff that I'd seen before. Definitely helps me play with ideas I have on my own. But then at the end, I was like, you know what? I just still wanna hire a designer.
Speaker 2:I just wanna hire a designer
Speaker 1:that Yeah.
Speaker 2:Is just better than me at this stuff. Uh-huh. And I only have so much creative energy in a day. There's so much to do, so many things that I'm trying to do.
Speaker 1:Mhmm.
Speaker 2:Someone is better than me at this. I just want to continue to pay them to do And their tools might change their process of change and and whatever. Yeah. Like that cron page that David did, which looks So good. Incorporated a bunch of AI, generated imagery into it mixed with the traditional skills.
Speaker 2:Also, his ability just to prompt that image into existence is better than Yeah.
Speaker 1:Yeah. That page that that page really to me is an example of, like, AI is not doing this. It also can't do the dev side, just so you know. I mean, like, there's I tried. There's pieces of it that could get right.
Speaker 1:But, like, getting those lines in the word cron in the background on desktop
Speaker 2:Mhmm.
Speaker 1:Unbelievably hard. The number of layers of tricks because you don't have, like, native can't do those, like, inner shadows on text natively in CSS. So there's, like, a lot of tricks being done to implement it. But the design itself, like, that's just an example of, like, if you don't have David's brain, you're not making that.
Speaker 2:Like, not black quality. Yep. Like, I I this tool gave me a 10x ability I had before. It is still, like, a thousand x away from anything that he can do. And I was like, okay.
Speaker 2:That's interesting because I wonder if people feel that way about software engineers, Mhmm. Where they're like, oh, cool. This helps me get further and, like, try stuff out before I even go to the person that's good at this Mhmm. Which is useful and valuable. But ultimately, like, I just felt the desire to continue to hire for this.
Speaker 2:Like, I so Yeah. I'm so designed, like, bottlenecked with so much stuff I wanna do. Yeah. That's always a bottleneck. And despite that, the AI thing is still not making a huge difference for me.
Speaker 2:Like, I just still need a a person that's good at it to hit hit the bar that I I'd like to clear.
Speaker 1:No. I I hear you. I don't think of it yeah. I wasn't thinking in terms of design. I think it's more like there's there's a few things that are interesting to me, like memes and media, just the the quality of them increasing, like the ability to be more creative.
Speaker 1:Seeing like when people early on, people are turning, like, the president and the vice president and Zelensky in the room, turning them into, like, South Park characters or Muppets. It's just funny. Like, I enjoy that it can do that kind of thing. I know that's like one off anecdotal and it's flash in the pan. Maybe it was just the moment on Twitter that was funny.
Speaker 1:Yeah. But it does feel like it raised this like possibility of being more creative in the way we express things visually. That's exciting to me. And then I think like, I mean, I've wanted to build a game forever. We've talked about this.
Speaker 1:I think the idea of like building sprites and stuff for a game
Speaker 2:Yes. That's huge.
Speaker 1:I would love to do something like that. And it's just it's it's too hard to, like, hire somebody for something that's, like, a side project. You're just gonna Yeah. Yeah. Fiddle around with it.
Speaker 1:It's not a real job, and it's gonna take way too much time to, like, have somebody just doing it with you. It's just like it opens up those kind of possibilities that are very exciting. But I don't even know I mean, I've not tried. I don't know that it really can. It just seems like it could maybe.
Speaker 2:Yeah. I was talking to AJ about this this morning, this idea of induced demand. Like, if you there's this like classic economic phenomena where if you have a highway and it's just full of traffic all the time and you double the lanes, it continues to be full of traffic all the time. It doesn't actually make anything go faster. Yeah.
Speaker 2:Because of induced demand, there's like hidden demand basically that gets, like, once there's more capacity, like it shows up. Yeah. But with a lot of this stuff, like, it's not like you're replacing something you're currently doing, it's more that you now do a new thing that Uh-huh. That wouldn't otherwise exist before. And our our I totally see our
Speaker 1:work jar is just gonna be continued it's gonna continue to be filled up. It's a liquid. Yeah. It will expand to fill. Yeah.
Speaker 1:Or is that gas?
Speaker 2:Don't think liquids do that. Whatever.
Speaker 1:Maybe gas. I don't know. Something something expands, fills up containers.
Speaker 2:I think it's a gas. Yeah. Yeah. That makes sense. I think you're right.
Speaker 2:Oh, man.
Speaker 1:Yeah. I don't know. I I thought the image thing was exciting. The the more exciting thing I think is or not more exciting, but the more interesting observation is just how inflamed Twitter can get so quickly. Not inflamed, like, in an angry way.
Speaker 1:Just like Mhmm. How pervasive stuff can spread. And I I wonder, like, is that all of Twitter? Is it just our little circle that gets like that? Does the whole of Twitter?
Speaker 1:Because it feels like every post on Twitter right now is the same stuff, And it's just, like, so infectious. Yeah.
Speaker 2:I hear a rare moment where something crosses all the bubbles in Twitter. And I think that's what this this thing was past week. I think it was like
Speaker 1:It did just apply to everybody. Yeah.
Speaker 2:That said, I saw someone make like a crazy compilation video of all the images and there was a bunch of stuff from that was cleared from outside of our circle, but like a lot of it was from inside of our circle. Part of me wonders if there's like this tech in terms of Twitter, x whatever in general, I think the authorship from our bubble is really high. And it's like a like, there's a lot of content being produced from people in our bubble.
Speaker 1:Our little bubble?
Speaker 2:Yeah. That gets consumed by the overall overall one. But, yeah, it was crazy. It's it's fun once over that happens.
Speaker 1:You're saying maybe Us Us tech centric folk were the main characters maybe? I'll take it.
Speaker 2:I I think a little bit on extras because, like, stuff is just shifting more. Like, tech in general is just Everybody else left. Central to people's lives. Yeah. That's that's another part of it probably.
Speaker 1:The only community still standing.
Speaker 2:But, yeah, it's cool when stuff like that happens. I can't like, what was, like, the last time
Speaker 1:The last big thing was MCP.
Speaker 2:No. No. No. Something that that was, like, crossed Oh, crossed over to normal Twitter.
Speaker 1:You mean like a post from someone in our circles that like
Speaker 2:No. Not just just like a trend, like this thing where like all of Twitter were just talking about the same thing.
Speaker 1:Well, was the the single word thing. Right?
Speaker 2:Single that's that's just like a single word thing.
Speaker 1:Tool companies. Yeah.
Speaker 2:Yeah. Yeah. That was probably the last one I can remember. And Biden tweeted. What did he tweet?
Speaker 1:Freedom or something. I don't know. Something America something America would claim is, like, uniquely us. Democracy, maybe. I don't know.
Speaker 2:I don't yeah. I don't remember
Speaker 1:like that. I don't know. That
Speaker 2:was pretty crazy.
Speaker 1:I did think of one other thing, and then I lost it. But it was important. No, it wasn't. That's setting a bad expectation. Shoot.
Speaker 1:But I don't I don't even know. I can't even read. My eyes are not great.
Speaker 2:I can't even read how long we've
Speaker 1:been on here. Is it like people's eyes are rolling right now? They're so tired of listening to us? No. Can you tell how
Speaker 2:long it's been? It's only been fifty one minutes.
Speaker 1:I'm gonna have to make the text bigger. I'm getting old.
Speaker 2:Oh my god, Adam.
Speaker 1:I know. I'm sorry. Oh, there we go. Made it bigger. Oh, sure.
Speaker 1:Oh, is. Yeah. Okay. Whatever. I can't remember.
Speaker 1:I feel like it was important. Something going on this week. Oh, I was gonna ask yeah. Just I've been traveling, a little unplugged, a little focused on very specific things. How's the, like, world politics or even just American politics?
Speaker 1:How's that whole situation looking? Anything crazy? I feel like if I've been unplugged for a week, I've missed some crazy shit, probably.
Speaker 2:Yeah. The the the tariff situation continues to, like, turn back and forth. I will say relative to the previous couple weeks, it's been kind of quiet. Nothing like crazy incendiary.
Speaker 1:Okay.
Speaker 2:The tariff situation continues to be like back and forth like You can actually just If you just want to catch up, just go look at the S and P chart for like the past. And you can see how the fix Oh, sure. Everyone thought that, like, oh, turns out the tariffs are less bad than we thought. And then it was like, nope. Today, again, you know, back back back to the chaos.
Speaker 2:Yeah. Trump said something today about, like, automakers should not increase their prices because of tariffs. I'm like
Speaker 1:That's not how things work. Like Corporations should stop seeking profits now.
Speaker 2:It's like isn't that is that what the other side says? I'm like confused. It just it's both both sides are just like in favor of price controls. Now that's like causing a bunch of confusion. I am so biased now because I've taken all of my money out of the market.
Speaker 1:Oh, so you want to see it burn.
Speaker 2:A small portion of it is betting on a downturn. Yeah. So there's like crazy feedback where I'm just like, I want I want I want to get worse. Yeah.
Speaker 1:I want it to get worse. Crash and burn.
Speaker 2:I've been
Speaker 1:in that situation.
Speaker 2:I'm happy with my portfolio today. Looking looking green for me, red for everyone The
Speaker 1:VIX is quite up.
Speaker 2:A lot of volatility Exactly.
Speaker 1:It seems.
Speaker 2:Yeah. The other thing is Robinhood. Man, I have grown to really love Robinhood as a company. Oh, yeah? And I've, like, gone back and forth over the So I'm I just came out.
Speaker 1:That's all I know.
Speaker 2:Yeah. Exactly. They came out as, like, a free stock trading app in a time where everything, you know, you pay per commission. It's pretty expensive. And it was like a mass market thing.
Speaker 2:Right? They got a ton of people using it. The reality is is most people shouldn't be trading stocks like that. So Yeah. Right.
Speaker 2:Not really like massively successful from that angle. But in terms of adoption, it was hugely successful. They've just slowly been adding and expanding into other areas. And they'd execute really well. Like, their products are well designed, they feel good, they work well.
Speaker 2:Yeah. And they're taking over my financial life at this point. So they put out a credit card and I canceled my Amex Platinum, which we've had for years and I switched to theirs because I'm just tired of all the Amex Platinum. It's like you have these, like, 29 benefits, and they change every year. And make sure you're signed up for them.
Speaker 2:Oh, by the way, there's other deals here that don't get applied automatically. Have to pick them. Oh, convert your points and, like, you know, it's five x we book on travel. It's just, like, so confusing.
Speaker 1:Gotta, like, you gotta like subscribe to the NerdWallet newsletter to be able to like know anything about how to use all those benefits. Yep.
Speaker 2:It's a it's a whole thing. And if you're not doing that, you're probably like not getting that much benefit, maybe even losing money when you compare it to the the monthly monthly fee. Right? It keeps going up. It keeps going up.
Speaker 2:And I think the final straw for me was I was telling someone, like, oh, man. I just have all these Audible credits that I never use because I don't really listen to audiobooks. But, know, it's an Amex benefit, so I have not all the account. They were like, man, like, Amex canceled that, like, two years ago. Like, you've just been paying for it for no reason.
Speaker 2:And I was just like, okay, I'm done with this. And then Yeah. Robin Hood has a gold card. It's physically gold. Nice and solid, little physically.
Speaker 2:That's that's not literal gold. It it looks gold.
Speaker 1:Yeah. Yeah. Just looks gold. I love did I tell you gold's my favorite color? I finally figured this out.
Speaker 1:It's almost pretty good as well. Yeah. Anyway, keep going.
Speaker 2:Well, you should get it then. I should. Percent cash back on everything.
Speaker 1:So just a credit card. You you don't have like an account with them, like a bank account. It's just a credit card.
Speaker 2:Well, he he well, he that's that's where I was going with it. So Okay. I signed up for that. 3% cash back, very straightforward, great app, great experience, has all this stuff, amazing. Then they announced yesterday they're launching a bank and they're launching a robo So Oh.
Speaker 2:I mean, I just pull all my money out, I just said it from Wealthfront. But if I ever if when I put it back in, I'm just gonna use Robin. I'm gonna use their robo advisor.
Speaker 1:Okay.
Speaker 2:I use Mercury, which is great and I have lot of business needs, I don't know if I'll use their bank. But, know, theoretically, I would just use their bank. I think I would just use it for everything. And then that was this hilarious feature, which sounds like such a bad idea, but as someone that's been using these like new modern banks for a while now, there's just acute pain point that you just always need cash. Like, there's always situations where someone only takes cash, you're paying for something in cash, you need cash.
Speaker 2:And so we're saying we have to go an ATM and, like, do whatever. Robinhood Bank announced it's effectively Uber for cash. You can, like, order cash and someone drops it off at your house and it just seems like a horrible idea because it's like, okay, but that random person knows that you have, like Yeah. Thousands of like, it just it feels weird. Like, it doesn't feel safe.
Speaker 2:And maybe it's irrational, but
Speaker 1:Is it actual Uber drivers or it's different?
Speaker 2:It's it's their own, like, it's their own network, I think. Yeah. Of whatever. And maybe they partner with someone for this.
Speaker 1:But
Speaker 2:I can't think of a single Uber driver I've had that
Speaker 1:I would want to deliver me cash.
Speaker 2:And, like, are they Targets themselves because people know that they have cash? So are they Yeah. They driving around in a cold car?
Speaker 1:You know, like Robin Hood written on the side, like
Speaker 2:It's like those are they driving in those, like, you know, those ATM trucks?
Speaker 1:Oh, like a Brinks? Maybe.
Speaker 2:Yeah. They're gonna that would make a little more sense than they They feel
Speaker 1:a little better. Yeah.
Speaker 2:Jeez. That feels a little more
Speaker 1:I bet this is only in major cities. This is the kind of thing I won't hear about Neo's Ark five.
Speaker 2:No. No. By the way, when I first moved to Miami, we get like GTA headlines all the time. It was one of the first ones. One of the drivers of that truck locked himself in the back by accident and just started shooting.
Speaker 2:What? He started shooting at the lock from inside trying to like, just panicked. And he so people were just like
Speaker 1:Oh my word.
Speaker 2:Carrying gunshots around this ATM vehicle, but just like a guy
Speaker 1:He's trying get out.
Speaker 2:Yeah. Yeah. He's trying to shoot the lock.
Speaker 1:Couldn't that, like, ricochet or something?
Speaker 2:I don't think was really dangerous. Jeez. I think it was okay. But
Speaker 1:But, yeah, I guess it's better than suffocating. Go down with a
Speaker 2:Yeah.
Speaker 1:Bullet wound. Jeez. Okay. I love the Robin Hood website. I'm sorry.
Speaker 1:I'm just so distracted because everything is gold and I just love gold.
Speaker 2:Yes. They execute really well and I love this thing where they enter the market with this, like, kinda weird product and you just never would have guessed where they went. And I just love this. I just don't know how companies are gonna eventually make money. Like, no, I don't think anyone would have really guessed that they're gonna launch a credit card, they're gonna launch all these things and their brand is gonna be very strong so they can do it.
Speaker 2:They just kind of
Speaker 1:became, yeah, they became this brand that, like, the younger generation associates with financial everything.
Speaker 2:Yeah. Very smart. It's it's well executed, takes a long time, and they've been through ups and downs. Like, they had all that trouble with the GameStop situation.
Speaker 1:Oh, they had trouble with that?
Speaker 2:Remember they shut down trades and they were like kind of insolvent on something. They had to like get a quick like a big cash infusion from Oh, I did. There's a bunch of chaos with it. Okay. Yeah.
Speaker 2:I'm a big fan of their their pro they're giving away a thousand gold bars.
Speaker 1:Oh, my word. That's so cool.
Speaker 2:Oh, man. It looks so cool too.
Speaker 1:I love I love gold. Not like actual gold, but the color. And I guess gold bars are the actual color, so I love them too.
Speaker 2:You what's funny? I think that's actually a very not white trait, loving gold.
Speaker 1:Oh, really? I've decided, like, white and gold, like, the combination of white and gold is just like it does it for me. Man.
Speaker 2:Nice.
Speaker 1:In terms of a color, like Yeah. Combination or just it's mostly the gold. But anything that they put, like, gold foil on or something, ugh, love it. I just want that, whatever it is. Unless it's like a steak.
Speaker 1:Just give me that. Was gonna say, but
Speaker 2:what about this? Coming to come down to Miami, they'll put gold on whatever you want.
Speaker 1:I'm sure. Yeah. Yeah.
Speaker 2:Well, now you should get a necklace as you'd like to say.
Speaker 1:That's what's so sad is Casey watched that episode when I wore it, she's like, well, you're just a silver girly. Like, you need to wear the color silver. Gold is not for you. And I think it's true. I think gold doesn't look good on me.
Speaker 1:So that's why the necklace looked bad. Yeah. I think so. If it would have a silver necklace, I might have pulled it off.
Speaker 2:Yeah. Think that that's maybe why I say it's not so much like, I feel like not white people love gold. Like Not white people. Here's a funny thing. Indian people love gold, like
Speaker 1:Oh, really?
Speaker 2:To like a weird degree, like to like a it's like a like a like a like a dragon hoarding gold degree Yeah.
Speaker 1:Kind of thing.
Speaker 2:When I was little, we went to Singapore and I remember the trip because I got they bought my parents bought me a Game Boy Color on that trip. I remember playing the Boy Color.
Speaker 1:Oh, taking
Speaker 2:the But I also remember we were at some store and my parents just bought a bunch of gold coins and like, they were they were just like acquiring physical gold. And not jewelry, I think some jewelry, but they were just like they got like a sack of gold coins. And they were just like, yes, this is something we're gonna buy.
Speaker 1:That is awesome.
Speaker 2:I want a sack of gold coins. There you go. So, yeah, brown people, we love gold. We hoard it. Hoard gold.
Speaker 1:Mid the gold discussion, I'd I'd switched over to the Twitter tab and I just saw where Prime and the replet founder are having an interaction. It's quite funny.
Speaker 2:Oh, did he respond?
Speaker 1:Yeah. That there's that guy
Speaker 2:so abatable. I'm so excited. I'm so excited. No. We're we're getting in the pockets.
Speaker 2:I wanna see what it is first.
Speaker 1:Prime's saying it was his first interaction with him, but it wasn't great. And he's like, it's not the first time you did a whole video about me.
Speaker 2:Wait. What? I don't know.
Speaker 1:Should I read all the it's on Twitter. It's public.
Speaker 2:Where is it?
Speaker 1:It just feels weird. It's a reply to my tweet. Okay. So the CEO, Omjad,
Speaker 2:replied to me. Wait. What? What happened with Dave? Dave just quote tweeted my thing going, what's a replant?
Speaker 2:Never heard of it.
Speaker 1:I love Dave.
Speaker 2:Love it. No. Just TJ. TJ is the last one.
Speaker 1:No. Yeah. Prime responded to him. So this is where Prime responded to him, and then it goes on from there.
Speaker 2:I see. Good. Keep this way. This is my first interaction I had with the Internet. It's not like a token.
Speaker 2:Open source project destroyed by legal threats.
Speaker 1:Okay. I've talked about this situation Okay. More than I more than I'd like, if I'm being honest. I just stated a fact. Listen.
Speaker 1:I'm not I didn't say a single negative thing about this company. I just said I'd literally don't know what
Speaker 2:it is or know anyone that uses it. That's all. It's not even negative. It's just maybe it's
Speaker 1:not for developers. Is it not for developers? Is it for other people? Is it so VCs can build their own ideas?
Speaker 2:Is what? Replit? Yeah. VCs can build their own ideas. That's what the product is.
Speaker 1:I just don't know what it is. But I feel like if if it was for developers, I feel like I would actually know what
Speaker 2:it was because developers would be using
Speaker 1:it and I would know about it.
Speaker 2:I think it's I think broadly it makes sense. It's like, what if we built like a full unified vision for someone brand new to programming, get them going to fully deployed. Like everything from education to like development to hosting all of that, like it's a full full featured thing. Gotcha. I would say it's like for developers, it's just I think it helps grow the pool of developers that exist.
Speaker 2:And they've had to like pivot, obviously with his AI stuff to, you know, how much like, what does that look like? But Yeah. I don't know. I I just I just know the CEO is a hot head and he's always like saying all this crazy crazy stuff and he gets he gets upset pretty easily.
Speaker 1:Good to know. Wish I would have known that before I tweeted.
Speaker 2:I literally thought you did that because you saw my thing.
Speaker 1:Because for
Speaker 2:me Yeah. I knew he was like that and I knew he was gonna like jump on it. So I knew it was gonna be something No.
Speaker 1:You told me that you tweeted something and then I found
Speaker 2:it, but I I didn't know you had before that. Yeah. We're all just we're all just Terminal versus Replit. That's you know, I'm gonna DM him for a sponsorship. Hey, do you
Speaker 1:guys want
Speaker 2:a sponsor?
Speaker 1:Nobody would see that coming. That's effective marketing right there. We did it. We did another episode. This podcast isn't gone yet.
Speaker 1:You thought you could get rid of us? We're back after a week off.
Speaker 2:That's funny.
Speaker 1:I'm making it sound like we just took three months.
Speaker 2:It feels that way.
Speaker 1:Can't wait to see whatever this is gonna be animated as. People aren't gonna have to look at our faces.
Speaker 2:They're gonna look at some kind of cartoon.
Speaker 1:I bet we look so goofy right now. Do you have any rep requests?
Speaker 2:Rick and Morty?
Speaker 1:Rick and Morty. Okay. Yeah. That'd be fun.
Speaker 2:Let's do that. I'm doing Rick
Speaker 1:and Morty. Alright. This is be fun. Alright. Alright.
Speaker 1:See you.
Speaker 2:Alright. See you. You see, mister Powers, I love gold. The look of it, the taste of it, the smell of it, the texture. I love gold so much that I even lost my genitalia in an unfortunate smelting accident.
Speaker 2:Hence the name, Gold Member.
