Do You Look Up or Down When Your Job is Being Taken Away by AI?
Whoo. Whoo. What's up? I'm good. It's really hot in Miami.
Adam:No. I didn't ask you if you were good. I said, what's up? I don't care about your well-being, Dax.
Dax:I'm just kidding. I I I literally don't know how to answer that otherwise. What's up?
Adam:I don't know. Give me news. Tell me what's going on in the world. I don't I don't care about you personally. No, I'm just kidding.
Adam:What's going on? Yeah. So you're you're doing it's hot, you said?
Dax:Yeah. It's hot. Liz is two months away? Yeah. I bet it's real hot for August '20 fourth.
Dax:So I think we're we're three months into August.
Adam:I guess it is, yeah. It's just the May.
Dax:'4 months away. We just finished we had a baby shower at our house last weekend.
Adam:I saw a tweet, I believe.
Dax:All we did was we just got these four tape we rented these four tables and they were like old wooden nice looking tables but it made the whole thing feel like crazy fancy. I'm like, you know, it's not that hard to pull off a nice looking event. You see that picture that I posted with the of my backyard?
Adam:Wait, was that your backyard?
Dax:Yeah. That was my backyard. It's just like Oh. It's like a nice camera and a good angle. Makes my backyard look giant.
Adam:Was it the tweet about like, make it a wedding venue? That one? Yeah.
Dax:Exactly. Yeah.
Adam:I did
Dax:see that.
Adam:I didn't even realize that was your house. I thought you were saying like yours could pull that off. That's crazy.
Dax:No. That's literally my house.
Adam:Why is Twitter search bad? I I hear people say that and I never knew what they meant but like I'm searching go
Dax:to my profile and go to media.
Adam:It's like the first thing. Oh, media. No. The media type? Ninth thing.
Dax:Thing. You're doing a lot
Adam:of times. I didn't think I
Dax:posted that much media.
Adam:Oh, yeah. Sure enough. Is that is that like one of Liz's relatives in
Dax:the photo? Who is that? Yeah. It's her mom.
Adam:That is your house. Wow. Look at that.
Dax:Just the camera and the yeah, the tables.
Adam:Did you hire a bartender?
Dax:That was Oh, no. Guess you you you missed a party this last year. But yeah, we've been hiring a bartender every time we do a thing at our house because it's not that expensive and they do all it's not just him, it's him and a like two other people. Yeah. And they just make everything extra nice.
Dax:Like we had these mango drink because was a mango themed baby shower. We had these mango drinks to serve and instead of just like serving the drinks in a glass, they like decorated it with all this stuff that I don't even know where they pulled that out of and it just looked so nice. And yeah, they're not that expensive and yeah, they'll just like help make your party feel nicer.
Adam:Oh, man. The the tables, the chairs, the trees in your backyard, I don't realize how majestic they are when I've been there.
Dax:Yeah. When you look up.
Adam:Yeah. I just don't look up. Do you do the thing where you just like never look around you when you're in a place?
Dax:So I don't look down and Liz for some reason is always looking down when we would walk around. And I remember when we lived in New York, this was an actual problem because she would just see the most messed up stuff on the ground and I would just never see it and she'd always point it out to me. Sometimes it's useful because it's like I'm about to step on a dog poop, but a lot of times it was like, oh, look, a pigeon that was flattened over and over.
Adam:Oh, jeez. Yeah. You don't want to
Dax:see that stuff. You know, just disturbing stuff on the ground. So I I don't look down when I walk, but Liz always looks down. She always just has these Okay.
Adam:I look down.
Dax:Oh, you do?
Adam:Yeah. I only I only look down. Like, I we have a lot of geese in our neighborhood and they just poop all over the sidewalk. So I'm always, like, I'm going on walks in these sidewalks every day, and I'm looking to make sure I never step on goose poop. Never have.
Adam:I'm pretty good at this.
Dax:I literally step in dog poop, like, once a week.
Adam:Are you serious?
Dax:Just in my own backyard. Oh.
Adam:Yeah. You don't wear shoes in the house. You you do wear shoes in the house. Oh, god. No.
Adam:Oh. Hang on.
Dax:Adam, I don't wear shoes outside.
Adam:Oh, so you're stepping it barefoot. Yeah. It's even worse. You gotta put shoes on to keep your house clean.
Dax:Reaction was that I step in dog poop and then walk in my house with the dog poop on that. Uh-huh. Like, I know I stepped in it, so I, like, blast it with the hose right away. Yeah.
Adam:Fix it.
Dax:Okay. I gotcha.
Adam:But also, I'm there, so
Dax:I get to it, like, goes between my toes and I really
Adam:Oh, my my god. I just gave you like Play Doh memories, like all the
Dax:little Play Doh tools, the way they like squeeze between. Yeah. Exactly like that. And then you get instant it it like kicks up a bunch of smell too. So Yeah.
Adam:So I I only look down. I never look at like the horizon. Like when we go through town, like I could drive into my town of like 10,000 people and probably find new buildings all over the place. So I just don't like if we drive through it, I don't look at anything. Does that make sense?
Dax:I'm very absent minded and I don't pay attention. So I don't know for me it's about physically looking, but I I'm similar. I just don't notice anything. I feel
Adam:like you notice everything. You're the person on Twitter that I follow that notices things. Oh, no. But it's not things like around you. It's like things that are true in the world.
Adam:You just you're too busy thinking about big thoughts to notice like what's around you. Is that what you're saying?
Dax:Yeah. That that's that's like the absent mindedness. Like I I'm like walking around, I'm looking but I'm like my eyes aren't actually seeing anything. It's just kind of passing right through. It's kind like when people drive like they don't remember driving.
Adam:It's Right.
Dax:Similar to that, but like wildlife.
Adam:Now I'm just looking at all your tweets. Why do I get on Twitter? I should stop. I should just not do it because I have so many questions now, but they're not pertinent.
Dax:What were we talking about?
Adam:Oh, you've been doing good. You had a baby shower. It's hot there. It's rained here for like three weeks in a row. Like, the sun is out today, but it felt like a vampire movie for real for weeks.
Adam:It's awful. Like, we got back
Dax:Is that why you look so depressed?
Adam:I mean, kind of.
Dax:That's one of
Adam:the reasons. Yeah.
Dax:You kind of look like a vampire. Something about it. I I think it's because you're wearing black but you also got the the deep v going on so like
Adam:Oh, yeah. It's just a jacket.
Dax:It's like a sexy vampire. Yeah.
Adam:Okay. I am in a very dark room, like extremely dark. So
Dax:Yeah. I don't know.
Adam:I wanna I wanna talk about stuff. If we're gonna be on a podcast, should talk about stuff. What what are we talking about today?
Dax:I don't even know what's going on. I guess we haven't talked about we just talk about AI every single podcast, but
Adam:I know. I get so tired of it. I just wanna rebrand the thing, like I don't know. What are the AI podcast names? They have the dumbest names, I swear to God.
Dax:We would AI generate it, obviously.
Adam:Oh, yeah. Yeah. Of course. A NIM dash in there and everything.
Dax:But did we guess we didn't talk about Cloud four. I think Cloud four We haven't
Adam:talked Cloud four.
Dax:Or did we
Adam:There's always something to talk about with AI. Did did people doing crypto podcasts feel this way too?
Dax:Where they're
Adam:like, there's always something to talk about with the chain. Say do they say the chain?
Dax:I don't know.
Adam:I'm just guessing.
Dax:Chains. There's a new chain. Chain. And there's a there's a side chain.
Adam:Yeah. Yeah.
Dax:Yeah. So I decided to that's fine. I've done this so many times, so I'm sure you guys hear me say this a bunch of times. I canceled all of my OpenAI subscriptions and I'm trying to I'm trying I'm trying to centralize on Claude now because I saw Claude got the voice support in the mobile app, which is the main thing I use OpenAI for. And I'm just so annoyed with the infinite model selections in OpenAI and I really appreciate how Cloud four just came out and just replaced all the Cloud three stuff.
Dax:You never have to look at that again. They're not keeping it around. And it's the same price everything. So I I it's just simpler.
Adam:I guess, do you still pay for Google? You still have Gemini? Or I guess Gemini is kind of free. I can't figure out Gemini's pricing model. I don't actually know.
Dax:Think I pay for Google One which is like because I use Yeah. I think storage and a bunch of other stuff.
Adam:And Mhmm.
Dax:I think I pay for the tier that has Gemini, but I I never I never opened Gemini though.
Adam:Oh, you don't open Gemini. So I for a while, was doing like the open Grok Gemini and OpenAI all at the same time. I did like o three and then Gemini 2.5 and then Yeah. Grok three. And I just asked them all the same question.
Adam:And then I probably don't even look at all three answers. I just kind of end up on one of them. But I was doing that and it is really really stressful. Like, I don't like this feeling of I don't even know what's best for what anymore. So I would like to consolidate.
Dax:The other issue I kept having with OpenAI is like, I would just I would use it for something simple but then I would have accidentally have selected some like intense model and it would just take forever to respond. So I don't know, my Mhmm. My product experience with ChatGPT has just it like peaked a long time ago and it just feels Yeah. Bad to use now.
Adam:I hate them all so much.
Dax:Yeah. So I mean, that guy is
Adam:Yeah. I should probably just never open a p OpenAI again. Like, they there's nothing about like using OpenAI that I feel like is unique that I can't get elsewhere. So I don't know why I ever use it because I really do. I I hate that.
Dax:I just have this stupid habit of typing c h a in my URL, Bart, to go to chat.openai.com and I keep doing it. So that's it's like crazy how sticky some of that stuff is.
Adam:Yeah. I was just gonna ask you. So you consolidate on Claude, but I don't know. Is Claude four that much better? I my overwhelming feelings with Claude four are I'm a little bummed that it is very clear the gains are marginal these days.
Adam:Like, I don't I didn't feel like it's that much better than three seven. Maybe I just haven't used it enough, but it just feels like four I had really hyped up four. That could be the problem too. In my mind, I had really hyped up what four was gonna represent. I did, like, one Dario comment where he was like, saving big numbers like four for the major improvements.
Dax:Right.
Adam:And for some reason in my mind, I thought, like, we can't even imagine what this is gonna be like. And it just came out I didn't expect it to come out this soon either, so I don't know. I was overall a little disappointed, but it's definitely better, and it's better in the use case that we have particularly, but I think I built it up too much.
Dax:Yeah. I mean, this is we talked about doing building like a proper benchmark because I just have all these qualitative I'm not even gonna say qualitative. Qualitative feel is useful if it's done in a consistent environment.
Adam:Mhmm.
Dax:My issue is every time I try these new models, it's like I'm working on a different project than I was a week ago and like, I don't know. Just hard to understand what my feelings are. But my feelings are it feels faster to me. I don't know if that's like totally made up because I don't know why it would be. But I feel like both in, OpenCode and in OpenControl, like doing some of the support stuff, I'm like I feel like it's not not as slow as before.
Dax:And then I feel like it this feels like tighter. I think that's the word I would use. It feels less like it's gonna go and do some random crazy thing. It feels like narrower in what it tries to do. So it feels better to me, but yeah, it's not like fundamentally changing what we can or or can't do.
Adam:Are you using Opus at all or just Sonnet? I haven't tried Opus.
Dax:No. Yeah. Again, I I don't know when to switch to these things like it's
Adam:That's a big pain Like, knowing Sonic versus Opus is is like a thing.
Dax:I don't I
Adam:I think I just like with Claude Mac, I'm just starting with Opus and then when I run out of it, guess I'll use Sonic.
Dax:But Yeah.
Adam:Maybe that's why I didn't necessarily think it's a seat faster. Maybe Opus is not faster.
Dax:Oh, yeah. I I basically assumed Opus was gonna be really slow and the pricing just I just wasn't interested in trying it because it doesn't feel like it's something I would use every day. So switching topics a little bit, I read something interesting today. It's still in the category of AI but a little bit different. You know, people keep talking about I think Dario again was saying something about like, there's gonna be a huge hit to white collar jobs, whatever, the whole like job displacement question.
Dax:I saw an interesting comment today. So I have a Hyundai and I know that Hyundai was building a factory in Alabama, so they can take advantage of some of the tax credit stuff. Me and my infinite poor timing had to buy a car during chip shortage, during like no cars being available and right after Biden rolls back the tax credit for electric vehicles that they had to
Adam:be made
Dax:in America. So my Hyundai was the last one that was not made in America and ever since then, they've been made here. And they're so cheap now compared to what I paid for it. Anyway, still not over it. Pretty upset about that whole But they make them in Alabama now and there was an interesting stat where the facility is crazy automated, where each vehicle takes twenty four hours of human labor.
Dax:That's what the cost. And if you value that at like $30 an hour, 40 bucks an hour or whatever, it's like around a thousand dollars per vehicle of human labor.
Adam:Twenty four hours is only a thousand dollars? Oh, man. Our our tech employee is showing but
Dax:Yeah. 24 times 50 is 1,200. So let's say around a thousand, brother.
Adam:Holy cow. No wonder people have employees. Yeah.
Dax:So this is this is a reminder and I've I've heard this thing talked about for so long, like years and years this this keeps coming back where we just assume that instantly automating everything makes economic sense, but human labor is is pretty cheap already and we've automated a ton of stuff already. So if you look at that, their factory, okay, maybe you can do something to cut the human labor in half and that's worth something, you know, $5 a vehicle is is pretty But your gains are probably more on like just trying to make the machines go twice as fast and trying to like improve the automation you already have. Yeah. Even if it still takes, you know, twenty hours per car, like maybe you're maybe the automated side is producing them twice as fast or something so you have twice as many people working on them.
Adam:Interesting.
Dax:So like, I think when people talk about automation, they have such like a naive sense of it. They just imagine like, oh, like the humans are the parts that need to be worked on. By the way, AI can help with automation of the factory parts. But you can see how in this situation it's not removing any jobs or the incentive isn't to like remove the jobs. If you double the automation side and the demands there, you're actually gonna double the demand for the human labor Mhmm.
Dax:That as well. So yeah, these situations are just so much more complicated than they're portrayed.
Adam:I got so wrapped up in the, like, car factory example you gave and, like, the human wages thing that I just forgot we were talking about losing jobs. I'm like, oh, yeah, this is okay. I see how it all ties
Dax:back to
Adam:replacing humans. Man, that's all over the timeline right now. I don't know if you've noticed Twitter. It just feels like it's constant bombardment with people predicting the end of software engineering. And I know software engineering is definitely the, like, the place where AI is currently most effective.
Adam:But man, people are they're coming out of the woodworks. It's really annoying. It's just really annoying because one, it's like, can't predict any of this stuff. But two, it's like all the predictions are coming from these very loud people who, like, make platforms that are no code things. The replica guy, I swear if I see him on one more podcast, I'm gonna lose my mind.
Adam:That guy Yeah. He's me so much.
Dax:He's No. He's he's extremely unlikable. He's always been this way.
Adam:Okay.
Dax:Yeah. Then I don't feel so bad
Adam:for not liking him. I really don't like him.
Dax:No. He's a he's very annoying. So I tried v zero again yesterday and I have such a funny feeling with it where it is such a well done product. Like Mhmm. If you're gonna build a product in that shape, it's really well done.
Dax:All these nice little details, really cool UX. They've pushed as far as some of these, you know, underlying LMs can do. But my feeling around it was it's missing the most old traditional concept, which is when a user goes and uses your product, do they feel like, wow, that felt good to do right away? I because I went and tried it and the experience was despite all the polish, the great design and great UX and all that stuff, my feeling was, okay, I go try something. Wow, this is taking forever.
Dax:I'm gonna go do something else and come back when it's done. I come back, wow, that's a disappointing result. So, like, no product is successful unless people get to the moment and feel great about it Mhmm. Pretty quickly. Yeah.
Dax:Yeah. So so many of these products, AI related products, they're like missing this old foundational traditional thing. Like, you you just cannot have a successful product without pulling users to that point.
Adam:Yeah. Mhmm.
Dax:And I get that everyone is just waiting for the LLMs to like become actually magical in these tasks and then it'll feel that way. Yeah. But they really don't right now and it's kind of like it's weird that most people have a disappointing experience. Like it's tough to work a product where that's the case. Even if it is successful for some people, it kind of limits how it grows.
Dax:And if you think if you look at something like ChatGPT, that that actually fulfills that. Like, when you go to use ChatGPT and you ask a question almost always you're like, wow, what a great answer. Like, wow, that's exactly what I was looking for. But specifically for these app builder things, these like things that are trying to layer on top Yeah. It's almost never like, wow, that was a magical experience.
Dax:What's funny is what feels magical is the deploy button. Hitting deploy and having it instantly live, which is their old traditional product not having anything do with AI, that's where the magic actually is. Yeah. So it's kinda kinda weird.
Adam:Yeah. I like, what is the way that they could make it I mean, aside from the models just getting way better.
Dax:Yeah. It's on their control.
Adam:It's it's just that's that's it, isn't it?
Dax:It just Yeah.
Adam:If the first prompt is a miss, like if it doesn't deliver yeah. And it does it feels like you you just know they're all just sitting back there because I've even said it about stuff. Well, the models will get better. Like, it's like they build this whole city and it's like, well, there'll be people someday.
Dax:It's like all
Adam:the buildings are there. We're ready for it. But like, yeah, what if it doesn't get better at that particular thing? That that's the thing that's encouraging to me on the like enhancing programmers tooling, the open codes of the world.
Dax:The bar's a lot lower.
Adam:Yeah. Like, I feel like the models are good enough right now to make our job way better. And it's just now, on that side, I feel like it's a huge lack of the tooling, a huge lack of I guess cursor exists, but it doesn't feel like the cities have been built up nearly as much as the no code, like, app builder stuff. Feels like there's so many of those and less focus on augmenting engineers where I feel like the models are actually really effective.
Dax:Yeah. And I think it's frustrating to work cause I remember when we used to work on SSD, the older versions were built on top of CDK and it was very frustrating to have a user new user come in, try it, have a bad experience and likely churn forever because of something that was out of our control. Like, is the worst feeling ever. Mhmm. So when you're building these products and you don't work on the LLM, it must also be really frustrating.
Adam:Yeah. You must just wanna like get on podcasts and talk about the end of software development and get angry on Twitter or anybody who tweets anything bad about Replit. Sorry. Had someone specific in mind.
Dax:Yeah. So like it's it is not a I just know how frustrating that can be and with OpenCode, it doesn't feel that way. It feels like, oh, there's so much for us to do. Yeah. Yeah.
Dax:Before we even hit, like, okay Diminishing
Adam:returns. Yeah.
Dax:Exactly. Mhmm. But, But yeah, I don't I don't know where all those tools go because there's so many of them. They're all in heavy competition with each other.
Adam:And I I just saw a tweet today that was some guy which I don't know I made a list by the way on Twitter. Twitter meta discussion. I made a list where it's just like a private list. I didn't know you could do that. So I just have people I wanna see their tweets.
Adam:Like, only these people.
Dax:That's supposed to be your following list, but you have this whole thing that you're following social. Yeah. Unfollow people.
Adam:I can't unfollow people because that's that's mean. But it doesn't default to that tab, so I always end up scrolling through the for you first, and it pisses me out so much. But anyway, I saw this random tweet from some random person who's like, I tried all 87 AI agentic coding tools. And they always use like totally different names. Like, no one knows what this category is even called.
Dax:I know.
Adam:But they lump them all together. It's like the first one is like goes from Lovable to like Cursor to Windsurf to like Bolt. And it's like, these are not the same kind of tool. Replit and like it's just like my wife tried to use cursor and it's it's not a tool for non programmers. And Yeah.
Adam:Whereas Lovable, Bolt, these things are very much geared towards non programmers. Right? Like, it's just totally different products, but they get lumped into the same category. It's all very confusing.
Dax:Yeah. It's like where in the spectrum it's like a spectrum of like who it's for, but then if you're on certain points of spectrum, your product is fundamentally like different. I think there could be a place for
Adam:all these app builders or for that category. Like, it does make sense to me that I hate all the local shops like my barber or like Yeah. The county clerk website. All these websites are terrible, and I would love for them to just build them with these tools that are making much better websites. That makes sense to me.
Adam:Restaurants, whoever. You have a small business, you wanna have a website, or you wanna have some little app that does a thing. Sounds great. But, like, CEOs of those companies making okay. The replic guy, you do not need to go out and talk about the end of software engineering because that's not what you're making.
Adam:Like, your product has nothing to do with the future of software development. You know what I mean? Like, it's just a totally different thing. And he just thinks like programmers won't even be needed. And I don't get how they can't like, he's a CEO.
Adam:He's been doing this for how long? How long has that company been around? Eight years or something? Yeah. Like, has he not thought critically enough about the fact that his product doesn't eliminate software engineers in the future, now, ever?
Adam:It won't. Like, that's not even the category. Am I just crazy?
Dax:It's No. I mean, it's just funny. Like, this is just what these periods of time are like. Like, you remember what crypto was like? Did you see my tweet the other day where I was like, during peak crypto, there was one conversation that kept coming up over and over and over and there was debate on both sides and then CEOs, you know, were saying one thing and people were saying the other thing.
Dax:It's it felt exactly the same way it feels now. The the question was, oh, well, crypto replace dollars. Well, it will replace money, right? Mhmm. And that was the entire conversation.
Dax:You were you were like filtered to one side or the other and you argued about that. Uh-huh. And in the end, it's like, looking back on it, it was just so irrelevant. You could have been That was question. Yeah.
Dax:That crypto doesn't replace doll doesn't replace dollars. But if you use that logic to bet against crypto You got wiped out. Yeah. Yeah. Like, it's doing great regardless.
Dax:Right?
Adam:Like, it's a huge how many is it a trillion dollar market now? Like, how big is the cryptocurrency? It's big.
Dax:Whatever it is, Bitcoin is is high up. Yeah. A hundred 7 k or something.
Adam:That's such a good But then if you're
Dax:on the other side, if you're on the side of it's definitely gonna replace money and you went hard on, like projects and stuff where it was meant to be like this thing that took over all consumers and everyone was using Bitcoin, you also lost.
Adam:Yeah. So you really should sum this up in a tweet because that tweet that I saw, the one where you're like, it's the wrong question, I did not tie it all together. A lot of your tweets,
Dax:I don't
Adam:I'm not smart enough to actually understand what you mean. I'm like, that sounds very wise, Dax, but
Dax:I don't know what you mean. So so what happens is sometimes I don't I think I have like roughly a good point, but I don't know how to express it. So I literally open ended and I hope people project like whatever meaning they think is smart into it.
Adam:And then when they project a really dumb one, you just quote tweet one of your old tweets, it's like you're an idiot.
Dax:And then reply Exactly.
Adam:You're dumb.
Dax:Yeah. You missed the point.
Adam:Okay. Well, that that point you just made, feel like is the most that's tied me back to that crypto era and I can see now how AI could play out similarly where there is like this enormous value being unlocked, but we just don't even know where that is yet or it's still yet to be.
Dax:Yeah, exactly. It's it's like if if you're betting that all software engineering is gonna go away, that's the equivalent of being like the dollar's going away. And if you're betting that this is all totally a scam, it's not gonna work at all
Adam:Mhmm.
Dax:You're probably gonna lose also. So the question is like, what is the right question to ask? And it's it's just so so so difficult in the moment. I've been having so much
Adam:Is it possible in the moment? You think it's actually possible to know the right question?
Dax:Yeah. Guess someone lands on the right someone has clarity and someone gets the right did
Adam:they actually have clarity or did they get lucky? Like, did they just choose the right thing?
Dax:No. You make a bet. Like, I mean, they everyone's making a bet. I think Yeah. Some people's bets are right.
Dax:And then do they know for sure? No. Part of his luck, whatever. But I have so much more I understand something I've always was confused about for a long time. I don't if I've talked about this before but because my dad was in tech, at some point, I was looking around and I'm like, man, how did he like not build Uber?
Dax:How did he not build Airbnb? Like, why didn't he like do this? Why didn't he do that? Like, technically was in near it and he like had the skills to like all these things in hindsight seem so obvious. And you realize like in the moment, there's just so much noise and there's so much opportunity to work on things that just don't make any sense like that's what 99% of them are.
Dax:And it's such a small sliver of them that that actually makes sense and seem really obvious in hindsight. Now, going through these errors, I'm like, I'm gonna look back if I like, you know, miss a lot of these things or I don't super nail it. I'm sure my kids are gonna look back and be like, how like, were right there. Like, how come you didn't, you know Mhmm. How come you didn't found Coinbase?
Dax:How come you didn't, like, you know, do these things? But it's it's so hard.
Adam:Yeah. My kids are like, why didn't you make Minecraft? This is the best thing in the world.
Dax:And Yeah.
Adam:How could you not have done this? You program, what are you even doing with your life? Yeah. Exactly. I had this idea in my head that something you said just reminded me.
Adam:It happened in my sleep last night. I mean, like, when I can't sleep and I have lots of thoughts. This, like, analogy of, like, these big venture capital moments, like, when when venture capital focuses on a thing, like in this case, AI, where it's like everything that gets invested in is AI, And there's just so much money and so many companies creating. It reminded me of, like like, universe creation. Like, we think of, like, all the energy, like, all that crazy amount of mass, and it's just, like, explodes.
Adam:And that's like all the money. And then it like congeals into like Yeah. Stars, you know I mean? And like or like clusters of things and it's like they're all getting acquired and like focusing into the right You know what I mean? But just kinda like visually, it works for me.
Adam:That just usually makes Yeah.
Dax:That that clicks for me too. There's something satisfying about that that imagery. I think what's funny about this stuff and Jay posted a tweet about this the other day is, unlike your example where there's like an initial burst of energy and everything is just chain reaction, VCs kind of operate like that's true where they're like I forgot what Jay's tweet was. It was something like, the world is gonna change and I'm just gonna wait patiently while that happens on its own. You know, it's like it's not that like there's a force of nature changing the world, it's like people dragging like individuals one by one and dragging users one by one and like, it's such this intense, like intentional human process but investors are just like, there's a trend that I'm gonna like Yeah.
Dax:With my sales and it's gonna like carry me, you know? Take me forward, yeah. Yeah. But you just you literally do nothing. And they had this perspective that like, oh, there's these shifts that just happened, but now, like, people are, like, pulling the world through those shifts.
Adam:Yeah. I VC the whole VC thing is I hate to say I I don't know if I should say negative things about VCs on podcast. I should? Okay. Well, yeah.
Adam:Why would I send
Dax:my stuff? That picture of my cone head. Like, I don't think
Adam:Well, they just they remind me so much. They're like upper class real estate agents to me. Like, that whole profession of real estate agents is just like, what are you again? What? It's like it's like a middleman kind of, but it's like, do I even need you?
Adam:I don't know. We do need venture capitalists because I don't know a lot of old rich well, know some.
Dax:That's that's really it. The requirement for the job is you need to know LCs. Like, you need
Adam:to know Literally it. Yeah.
Dax:You can raise a fund. Mhmm.
Adam:It's interesting.
Dax:It's like, I actually know enough to raise a fund. I just You should. Think it's the most boring job in world.
Adam:Think the funniest think of the funniest outcome, would be Dax becomes like the meme. The funniest like stereotypical VC. No.
Dax:I I don't have it in me. Like, it's not
Adam:Be pretty funny.
Dax:Just don't. I don't have it in me. It would be funny.
Adam:Actually, it wouldn't be that funny if you made a lot of money.
Dax:I think it'd be confusing. If I made a lot of Yeah.
Adam:It'd be awesome if you made a lot of money.
Dax:Like, I do it as a joke, but then I have like I actually succeed. Yeah. Uh-huh. Yeah. That's funny.
Adam:You could put in your bio, like, early in x y z and all the different companies you were early in. It's so funny how like every profession has like the Twitter bio like cookie cutter template.
Dax:You know
Adam:what mean?
Dax:Like you can look
Adam:at somebody's Twitter bio and know what they do without even like knowing where they work or what they do.
Dax:There's a there's a VC that I really dislike. I found his Twitter bio really funny because I I can't remember his name or anything about him. But it's like you go to his Twitter bio and it's like, there's like 50 companies tagged. It's like the most amount of things. It's like, I'm involved in all of these.
Dax:It's like,
Adam:how does
Dax:he not realize that the more companies you put, the more of a joke it looks like that you're claiming any association with them? It's like, you can't actually be involved in all these companies. So clearly, you're just
Adam:pulling up
Dax:any possible Yeah. Yeah. It's funny.
Adam:Oh, man. I feel like I don't wanna talk about Twitter anymore. I feel like we talk so much about I feel like our podcast is becoming me and you catching up on like what we tweeted. Yeah. And like, I should just do that on Slack.
Adam:I should just ask you when I have questions about your Twitter.
Dax:Then we could talk about other things. That's true. I was imagining getting DMs from you being like, hey, nice post. Like, does this mean exactly?
Adam:Yeah. What does this mean? I just learned what Deepak means today.
Dax:I still don't know what that is.
Adam:Apparently, this part of Twitter or No. I actually don't
Dax:I mean, I I know I know the acronym. I just it's like a I don't really understand it.
Adam:Well, could you confirm? Is it this part of Twitter? Because I think I saw five different
Dax:answers like part of Twitter or something?
Adam:That part of Twitter. Okay.
Dax:Yeah. Anyway, just kind of moving away from this this zone. Yes. Did you see that Remix finally announced the thing?
Adam:Yes. Well, I got halfway through the article probably. My intention to fan is not that good anymore. But I got to the part where they explained did Ryan write it? Ryan write that article?
Dax:I think it was a team effort.
Adam:For both of and Michael? Yeah. Yeah. I got to the part where whoever explained that they're, like, starting fresh. Remix three is, like, not is it not built on top of React Router?
Adam:Is it, like, starting fresh? No.
Dax:It's it's like they're trying to build a new thing with a certain set of constraints and they're exploring that.
Adam:Is React Router, like, acquired? Like, what did Shopify acquire? The team?
Dax:Shopify acquired, yeah, the team. But React Router now has like an open governance thing that they set up.
Adam:So are they still at Shopify? Are they leaving Shopify? Yes.
Dax:Still at Shopify.
Adam:Okay. So Remix is still a Shopify thing or no? Does it matter? Maybe it doesn't matter.
Dax:They have support from the Shopify CEO to go this route. Gotcha. Which is what I think matters.
Adam:Yeah. So it's gonna be a new web framework. We catch me up. You probably read the whole article or talked
Dax:or something. I've like I've like kinda known about this for a bit. So the the main detail is this new thing they're building, it's not gonna be built on top of React. Mhmm. It's gonna be built on top of Preact, which is like in a lot of ways, it's like it's like all the best parts of React just like slimmed down.
Dax:Kinda like the React that we remember before it got really crazy.
Adam:The React I remember.
Dax:They're setting constraints themselves which is they're not gonna try to solve any problems using like compiler tricks.
Adam:Right.
Dax:Which I think people interpret in a wide range of ways. They're like, oh, they're saying that everyone using this new thing needs to like go down this no build path where it's just like JavaScript files and whatever. I interpret it a little differently. I think about it more like it's more about setting constraints themselves to push them in like an unexplored design space where they will build a framework around these constraints. They believe that will lead to something good or have these good side effects.
Dax:You can still have like a heavy build process on top of that if you need it, a lot of production apps do. Yeah. So I thought there's exploring this I think pretty much unexplored area of, hey, if we need to do SSR, an SR framework, almost everyone uses file based routing because that helps you easily split the different routes and things like that. But they can't do that. So what would an SR framework look like when you can't rely on getting the file system during Okay.
Dax:During building stuff?
Adam:For dummies, real quick. Why can't they do that? So if they don't have a compile step, they can't do file based writing? Or is that what you're saying?
Dax:Yeah. So file based writing is an example of something that needs a compile step. Right? Because something needs to scan your file system and, like, pull out the files and know that, okay, this file is for client, this file is for server, like, no stuff like that.
Adam:Oh, I see.
Dax:Yeah. Or or the RSC thing where there's like the use client thing at the top, you need some kind of compiler to go and analyze that and Mhmm. And do it. So I think about it more like for like developer facing features, there's no feature in the in their framework that's gonna require something like that. Okay.
Dax:I think a lot of conversation is now devolving into the build versus no build thing which
Adam:I saw like Evan
Dax:Yeah. Had I mean Evan's questions were good. I I think Oh, yeah. No. I'd Mhmm.
Dax:I think he had like his questions basically hinted at this thing which is, well, there's certain things that are just impractical. Mhmm. But yeah, I don't see that as a you can still do those things. It's it's more about not giving into the temptation of, hey, there's this problem. If we do some fancy compiler trick, we can actually solve it.
Dax:Yeah. And seeing what what a framework looks like when you adhere to those.
Adam:So it's a two e thing. It's like constraints just make for different design decisions and that's fun. I like that.
Dax:At least that's how I was interpreting it but then today Did they
Adam:correct you?
Dax:Well, today posted something like, who's who's telling you you need a bundler? Is it your users? Is it the browser? Or is it the guy that makes a bundler? So
Adam:So maybe there really isn't going to be a bundler. And I'd be very curious how that works out.
Dax:Like like, are they yeah. So I don't know. I I don't if they're just doing that for fun, but it's hard to tell. Because remember how old Remix was when they first came out? And they were like always starting these like
Adam:Oh, yeah. There's a lot of yeah. Since they've been acquired from Shopify, I've almost forgotten about Yeah. Exactly. Those days those were good days on Twitter.
Dax:Good Yeah. Yeah. So already it's like, you know, it's already starting up again. But yeah, they they they
Adam:yeah. They should go really hard. That that would be so much rage bait. Like, if they just went really hard on like, there will be no bundler, I swear to you. Like, no bundler.
Adam:We're going no build. Just go hard on it.
Dax:Well, it's confusing because because that's what that Michael said that. But then someone else that's working on remix had like a more detailed message yesterday, which is more in line with what I'm saying, is it's more about how
Adam:we design Well, they can do the reasonable thing in practice. They can actually implement it with a bundle. But just say on Twitter,
Dax:there will
Adam:not be a bundler over my dead body.
Dax:Yeah. But you know what's funny? It I found it really funny because he posted that and I immediately thought, oh, yeah, this is like when they would do this and they would fight with Vercel and that whole era. Mhmm. And I look at I scroll down to the replies, Guillermo already took the bait.
Dax:Jared Palmer already took the bait. Like, they both have these And so I'm like, oh, okay. I guess we're back to what it was like, you know, three years ago.
Adam:It feels good. Feels good. I'm excited.
Dax:Yeah. So I'm just I'm interested to see where it goes. I I don't know. I I don't feel I think it's gonna be interesting and I think there's gonna be some value out of it. It's just not in line with like my tastes.
Dax:Like I don't mind build steps and build processes and like tricks like that. I know it gets like, Next. Js is a good example of what that looks like taken to an extreme. Mhmm. But, yeah, I'm okay somewhere in the middle.
Dax:So Yeah. I'm not like that excited about this.
Adam:I I've seen recent like group think not group think. Just recent like murmurs and I hadn't thought about Next. Js performance in like a year. I just kind of felt like all the early when they came out with the whatever version it was, 1514, whatever, the app router. When they came out with that, it was, like, so slow in dev and everybody was just shitting on it incessantly.
Adam:I remember thinking, like, well, the they're a big engineering team. They'll figure it out. But I've seen, like, recent murmurs about how it's still really slow. Did that just like did did that never get better? Is Next.
Adam:Js just like a slow thing now? Do you know?
Dax:It's so inconsistent. There's people that still say their dev setup sucks and they like it's like a really miserable experience. Mhmm. Other people say that it's fine. They upgrade to the version like, wow, it's so fast now.
Dax:And there's a runtime like production. Some people say that it's like slow, other people say it's not. Someone put React Router inside Next. Js and I'm just like, what are you doing? Like, what the hell is the point of this?
Dax:Because because I think I might be wrong because I can't have such time keeping track of all this stuff. But I'm pretty sure an app router, any navigation requires a server round trip.
Adam:That feels like the kind of statement that you say and then you get like three tweets from Vercel employees. Like, actually, if you ISR and then
Dax:There's probably something that's not correct about that. But I know a lot of people's default experience is I'm not it's not like a full page refresh or anything. It's just a I think it's just like an asynchronous call to render to have the server render like the next fragment they need to replace whatever's on the screen. Mhmm. And that's innately slower than client side rendering.
Dax:Sure. So I think that's what people get tripped up with. That is
Adam:the routing.
Dax:It's the the client side. It's not as like a spa feel so people get tripped up. Yeah. Yeah. So I I I think that's what it is, but
Adam:I just wonder the reason I asked, I wondered if like, is it possible as a big organization like Vercel, they just got so into v zero that that became the important thing and they just said like, to hell with Next. Js and its performance issues. Don't know. I think so.
Dax:Think in terms of like resource allocation, it's like the same if not more than ever. But Yeah.
Adam:Makes sense.
Dax:I don't know. I think they're just in like a tough zone of trying to do everything and their sometimes like certain designs just don't don't work nicely. Think the feel I have from it is and it's so hard to like really ingest this, like, do you remember the feel of like legacy frameworks? Like, it's stuff that you didn't grow up with but like
Adam:What what are we even talking about legacy anymore? What like, what is legacy?
Dax:I'm trying to think of stuff from our past. Like, do you remember like working on
Adam:Create React app? Are we going further back?
Dax:No. Something that's like really dominant. I'm trying to I'm trying to think of a good example. Like, do you remember what it felt like working on like an ASP Oh, yeah. Type thing?
Dax:Where it's like, it's huge, everyone used it, like, whatever. But like, you know, it wasn't like cool or like light or like any of these things. But it was massively successful and continued to be massively successful to this day. I think that's kind of where Next. Js is now.
Adam:Interesting. Yeah. It's the ASP.net of the twenty twenties.
Dax:It's like the default choice. Lots of companies will use it. Mhmm. It's definitely way better like in terms of the the stuff that they do and like they they they do try quite hard.
Adam:But its position in the mind share and in the ecosystem is like
Dax:Yeah. Great. It's hard to see it that way because we like saw it come up, like we were around to see it come up. So it doesn't feel like it could be that but something has to be in that slot, right?
Adam:Yeah. That's a great point.
Dax:And it feels like I mean, it's part of what it is now.
Adam:I mean, it's not still just ASP.NET. Then then I wonder like, are we just in this little Twitter niche that knows what Next. Js is and like, the rest
Dax:of the
Adam:world No.
Dax:Mean, Next. Js is huge, like it's it's Okay.
Adam:People say all of that.
Dax:Yeah, like giant Like, if if there's some giant enterprise company with, you know, the team of engineers that like are not plugged into anything that's going on at all, they're gonna use Next. Js. Yeah. Okay. Yeah.
Dax:So like when when like that's When you become that choice, if you're like massively successful, it's like incredible to get there. But Yeah. Yeah. There's a good feel of that. There's a good reality that comes comes with that.
Dax:Because And and we we see this all the time where with working on SST, we're like so focused on DX things and trying to make everything better incrementally better, incrementally better. And then every once in a while, we'll have a call with like a big enterprise company and they're like, oh yeah, we love SST, it's great, we're using it more, we have these questions. And you dig into it with them, turns out like 90% of our features like aren't even working. And like, they're they're experiencing like the most degraded DX possible. But to them, they're like, this is great.
Dax:This is like better than everything else we have. So it's like their tolerance for pain is so much crazier. So again, when people on when people are saying out of the Next. Dev server is slow or whatever, it might still be like amazing for my comparison for a lot of these these teams as sad as it It's
Adam:so funny that just you recounting the enterprise. We I just got on a
Dax:call with so a few of
Adam:us on the team got on a call with a statmuse user who's having issues that were so bizarre that, like, three of us got on a phone call with him. It's like, does it this is if this is, like, even
Dax:This is, like, a hundred users?
Adam:Yeah. Like, we got on
Dax:the phone
Adam:with the end user who had reached out to support. So my cofounder and I and one of our engineers are on this call, and he the whole site, like, basically wouldn't render. I mean, like, he was missing, like, huge chunks of the site, and he, like, had to do this crazy set of steps to find a sign in button because there wasn't a sign in button, so he had to go to, like, pages that had, like, a kind of an upsell button on the table to, like, sign in.
Dax:And he was just doing
Adam:it happily because he didn't know any better. And he's like, yeah, that's great. I'm like,
Dax:it's not even half of it's not loaded.
Adam:It's crazy. So we figured out the issue, but it just reminded when you said the
Dax:enterprise thing. Like, people will put up with stuff. Yeah. When there's good product market fit, it's just funny what people put up with. Yeah.
Dax:It's like, the stuff that I feel bad about, they're like miles they're like miles behind that. It's it's just funny to get that perspective.
Adam:Are are enterprises using v three or are they still do you have a lot of customers still use v two?
Dax:A lot of v three. I think our total v three usage is like five x what v two ever was.
Adam:Really? Yeah. So like sat we still haven't moved on to v three yet. I'm working on it right now. But but sat me just on v two and I wondered if that's normal or if No.
Dax:It is it is. There's lot of people still using v two but just like the growth for v three like outpaced Yeah. Any like historical usage of v two, which is always really crazy to see because when you're doing the transition, you're like so nervous about it but then in hindsight you're like, we're this this was like so much more successful, like we could have Yeah. Pissed off every single v two user and it would have still barely
Adam:started over and still had more Yeah.
Dax:Yeah. Would have been tiny. Yeah. Yeah. Which is like, again, like the remix thing.
Dax:I think it's kind of like that where everyone wants them just to like stick to the zone and increment from wherever they are. If v three is successful because Remix v three, right? Okay. They're also doing a v two to v three, big jump. Mhmm.
Dax:Was successful in the way that it needs to be for them to continue working on it. It's gonna like make all the old remix usage seem like basically zero.
Adam:Yeah. Man, you've got me missing that era just when remix was a thing on Twitter.
Dax:I just did. Happens now.
Adam:There's a lot yes. Is it the algorithm or something? Like, I do feel like Twitter is okay. So it could be a few things. I was thinking about like that era when like everybody had an avocado in their bio.
Adam:Like, there's like a million developer advocates and developer relations and develop Yeah. That's not on Twitter anymore. And I feel like either they all meant to like went to Mastodon or whatever, and that's why things change because like just the group that stayed on Twitter is just very different than what it used to be. Or it's something like the algorithm not feeding us the same kind of stuff. I don't know.
Adam:Something's different. It does feel like there's not really much happening at any given time on Twitter, and there used to feel like there's a lot.
Dax:Yeah. It's not even just on Twitter. I just mean, like, I feel like there was there was a period For most of my career, I felt like I would keep up with hacker news and every week there'd be like something very interesting, new that somebody was trying or a new project or whatever. And it feels like that peaked in an era where a bunch of DevTool companies were founded Mhmm. And funded.
Dax:And then like, that's when like all that stuff we're talking about was going on. And then there was a consolidation period where most of those companies failed or they realized like, this giant dev dev advocate presence was like, doesn't make sense. But also, think that the company just went away. Yeah. And there's been like crazy consolidation where it feels like there's not really any room for anything new.
Dax:Like, do you remember when like, there was an era where you'd always hear about Svelte and, like, Svelte was doing something new and always a cool interesting thing. Then, oh, and Solid JS came out and that's like a cool new interesting thing. And then it feels like now all those projects kind of work in silence and they're kind of established themselves.
Adam:Yes.
Dax:I don't know. It just feels like we're very standardized and consolidated. And I think the AI thing I don't know if this is actually what's happening, but in theory, the AI thing pushes consolidation. Right? Because it look kind of like pushes people into a narrower set of tools.
Dax:Yeah. Like, people are more apt to use the defaults because the AI Exactly. Is more apt. Yeah. Like, Tailwind, like, ended up winning and, like, there was a period where Yeah.
Adam:Nobody talks about that. There's just, like, so little of the, like, framework wars or any signs of, like, just drama for the sake of drama. Like, all that just kinda went away, but it's not, like, in a good way. It's not like in a, oh, it's so great.
Dax:Feels like
Adam:stagnant. Yeah.
Dax:It feels like we, like, explored the space of ways of doing things and turns out there's just like a narrow set that most people are gonna settle on and now Mhmm. There's really nothing else to explore here, which which might be true. And I wonder if, like, scientific research works this way too where, like, you go down some path and there's, like, all this discovery and new stuff to figure out. Eventually, you've kind of roughly, like, covered that zone of exploration. Mhmm.
Dax:And the next one is gonna be somewhere completely different. I I think that's maybe what it is, but it's not just front end though, like, it's also back end. It's also like different architectures in, like, on your back end. It's also databases.
Adam:Yeah. Feels like a lot less, like, programmer just discourse.
Dax:It's just I think it's like there's no more wrong ideas, but wrong ideas are are exciting, you know. It's I remember that like, I remember I remember, like, a database was such a big question mark for so much of my career. It's like every new project, was gonna try a new database, like Yeah. Oh, graph databases. Like, those are really cool and you can do all these cool things with graph databases and that was like a thing.
Dax:But now it's just like, yeah, just use a relational database and it's probably fine.
Adam:Yeah. Oh, that perfectly sums up the database converse that just perfectly sums up this, like, era of trying new crazy things and constantly feeling like I need to do the other thing because maybe time series databases or whatever. Like seven startups for every type of database. Yeah. And then, like, this era just feels like everybody just uses Postgres and MySQL and nobody cares.
Dax:There's seven startups for Postgres. Yeah. There literally is.
Adam:Yeah. Yeah. But, like, yeah, the the conversation is no longer that that's such a good point. So it's just that we figured out all the right answers. Is that basically it?
Adam:And And now there's not that much to talk about. It's like, which
Dax:The thing is, like, it's not that because I ultimately believe that you can mo you can usually make anything work. Right? Yeah. The fact that you chose a graph database, like, things are gonna be harder, other things are gonna be easier. But, like, you would probably be fine Yeah.
Dax:For like a lot of projects would probably be fine. Nobody does that anymore, it feels like. Mhmm. I feel like there's like no I don't know
Adam:what it is No, taking the hard path?
Dax:Yeah. There's no I don't know. There's no, like, being doing something that's, like, technically not right. Yeah. You ever just,
Adam:like, wonder what does everyone work on? Like, there's so many people on Twitter that are developers and, what are they what are these what do people work on? I'm sure they think that about me too, but I just, like I sometimes step back and I'm like, I don't even there's not that many things on the Internet. Like, what are these people building?
Dax:I know. Well, sometimes I go through our SSC, like, customer list, like the console list. Because whenever I can do whenever people sign up, sometimes I'll like, once a week, like, look who's paying and like what are they building Mhmm. I go to their website. And there is a lot of stuff.
Dax:There's like so much stuff, like
Adam:I just don't get out much.
Dax:We're not that big and like there's You know, in a week there's probably like three to 4,000 production ish projects with SSC. Right? Mhmm. That are being worked on. Yeah.
Dax:Three to 4,000, like
Adam:That's
Dax:pretty what the heck? That's so many.
Adam:Yeah. Yeah. I just don't get out much. That's the point. I think I just don't explore the Internet.
Adam:I have, like, five sites.
Dax:But I kinda agree with you. Like, despite me literally seeing that, I, like, know that it's all happening, but it's almost like, what are these things? Like, this makes sense to exist? Like, there's a whole team that's making money doing this and growing and and fundraising and doing all that, but it's The world is otherwise. The world is
Adam:so much bigger than it seems. Yeah. Alright. We're not even gonna make it an hour because I gotta pee and there's just like so much to do. I feel like this is one of those podcast episodes where we're both busy and it's just kind of like, I wanna talk, I do, but also
Dax:I wanna work more.
Adam:It's It's just an unfortunate reality sometimes.
Dax:Okay. We'll see it.
Adam:We'll call it. Alright. See you. Bye.
