ChatGPT
That's obviously an embarrassing group of people to be a part of.
Adam:Yeah. I mean, like, it's like the first time in human history, right, that we've been like, let's pause on technology. Slow down. You're going too fast.
Dax:What exactly has happened since the last time we've talked about it? Because it feels like forever ago, and it was only, like, maybe a month ago.
Adam:So we were just talking about what we talked about in the last podcast. It was mostly, like, the generative stuff, the AI, image generation stuff. Now I feel like ChadGBT is the giant elephant in the room, OpenAI. I mean, ChadGBT has been around for a while, but, like, they just started shipping stuff that seems to, like, fill a lot of the gaps in. So they announced the plug in API and this Internet connectivity where you can, like, write plug ins that can then go out and fetch real time data, not just, you know, stuff prior to the knowledge cutoff.
Adam:There's that date where the model was trained, 2021 or something. And before this, you you sort of, like, all answers are in the past before that date. Now the plugins allow you to fetch up to date information. So, hopefully, filling in the time gap. And then also, like, it's allegedly eliminating a lot of the hallucinations because now they can go out and fetch real data.
Adam:I guess you have to install the plugins. I'm not really sure how it's all gonna work. But they showed, like, examples of it running, like, FFmpeg commands in open or in ChatGPT. Like, it's running Python code natively in ChatGPT to solve things that you ask it, which is just wild.
Dax:Yeah. That that stuff is really cool. I think people and I think that's why it feels like it's fast because when OpenEye AI came out, people were like, oh, you obviously need this capability in other places. So people started to build all these little mini tools. Then ChatJPG was like, no.
Dax:Hold up. You don't need to do that. It's just gonna be in our platform. You're never gonna have to leave. The whole model is actually reversed, which, again, it's a great example of why I feel like there's a lot of chasing that happens whenever there's something new and revolutionary.
Dax:And it's obviously a lot of opportunity, with everything. But if you're just gonna hop on something because, like, you know, there's some, like, excitement around it and you just do the first idea you think of, things are moving quickly and, like, you probably your idea might not even be relevant. And it's kind of what's what's happened, I think, with a lot of those simple startups that were sort of, like, two weeks ago or now. It's not super relevant anymore.
Adam:Yeah. I had this thought even before they announced all that stuff that, like, I would not wanna build something on top of ChatGPT. I see a lot of cool examples. People post them on Twitter. I would not wanna build, like, a thing that just, like, relying on ChatGPT so much as I would like to say, okay.
Adam:This is a new era. We can build products way faster than we ever could. Can't what what I wanna build that's not related to AI, but AI is just gonna help me build it way faster. I think that's the more exciting thing to me as a technologist. I mean, it's not helping me make rebase any faster, I guess.
Adam:Sorry, Dax. But allegedly, like, theoretically, we could build it way faster.
Dax:I think we are building it faster. I have because I and both of us, I've seen I've seen you on stream, and you've probably seen me. Like, I open up chat GBT for everything. Like, whenever I need whenever I know what I need done and I just need to do it and I don't remember the exact details on how to do it, it is flawless and it's such a habit now. It's so integrated into my workflow.
Dax:And I think that's that's actually where the where the shift was for me. That's how I've been feeling about this for a while, which is, wow, it's making me so much more productive. And that was, like, only positive. Right? It only made me feel like, I was already pretty good at what I do.
Dax:Now I feel like I'm, like, a 100 times better at what I do. I think it also is weirdly creating like, if the gaps between people, like, if one person was, like, you know, twice as experienced as someone else, I feel like ChatGPPUs amplifies that, and now they're, like, a 100 times better because now there's better
Adam:So the 10 x engineers, the mythical 10 x becomes, like, a 100 or a thousand x.
Dax:Yeah. Exactly. Because they're just amplifying their judgment. And, like, people that are good at what they do is just they have good judgment and good instincts. And when they see the code that spits out, they immediately know what to refine, like, what's right, what's wrong.
Dax:Yeah. So they're just faster at, like, the whole iteration loop.
Adam:So with that in mind, tell me, Dax, how does this not eliminate jobs? I feel like everyone comes out and wants to defend and be like, you're an idiot if you think this is gonna take jobs. How does it not take jobs?
Dax:It's so there's it's so layered. Like, I feel a lot of people that are saying it's gonna take jobs, are kinda saying in a dumb way, and I don't agree with them. But then I also find people that are saying it's not gonna take jobs also kinda misguided. Like, the people that are saying it's gonna take jobs or I think they're just kinda in panic mode and not really thinking about it too deeply and they're they're scared, which makes sense because this is all, like, very disruptive. And I myself have had some of those feelings.
Dax:And I think that people that are saying it's not gonna take jobs, this is just another hype thing. Also feel like they're panicking, but they're, like, really good at being in denial. So I feel like the general takes I've been seeing have just been just kind of annoying, to be honest. It's I I don't know where I said this, but I was like, there's people literally saying I saw this is, a real tweet I saw. They were like, intelligence is no longer valuable.
Dax:What? Like, we live in a post intelligence world. This is this is no longer a thing that is gonna help. The only thing that matters is having an audience. I'm just like, okay.
Dax:That guy he does have an audience, and I'm like, I I see why he wants to live in a world where intelligence doesn't matter. I
Adam:don't think he had intelligence to start with. I mean, he's just excited I so he doesn't need
Dax:know. So there's that, and there's other people that are completely in denial that are like, this is just a dumb machine that is, like, emulating intelligence and is useless. So I feel like both sides of those takes are kind of I don't know. I'm not, like, super interested in that. Yeah.
Dax:I would say these are negative feelings. I think I've processed them and now I feel excited again. But there was a moment when I realized, because me and Liz, you know, work on a product together, like a very traditional SaaS application.
Adam:Boomi.
Dax:Yeah. Boomi. And we think it's very good and we spend a lot of time really understanding our users and building something like way better than everything else in the space to solve a similar problem. And we were
Adam:able to do
Dax:that because because we have just years of building SaaS application. Like we just picked a picked a space and just put our skill set in there and it we knew it would guarantee good results, right? Because we've done it before. But then when I was looking at this stuff and I'm like, wow, the way I build products and all those skill sets I rely on to be better than other people to like deliver products that are really good, it's a complete reset. Because now you can solve the same problems in a completely different way.
Dax:And it's a whole new world of like learning UX things. Right? There's all these little UX things that I know that work really well on a traditional product. But now there's like, okay, how do you integrate this stuff into a product? How do you deal with nondeterministic output?
Dax:Right? Like, now you have stuff where you can ask ChatGP to do something, integrate that into your product, but you don't know exactly if it's gonna be correct when it comes back. Like, how do you build these interactions, these interfaces? I realized, like, what I'm this this thing I was relying on to be better than other people, I am now I can still rely on it somewhat, but we're kinda all beginners in this world of building problems with ChatGPT, and we're all kind of at the same level. And it's like, the race resets, now we're starting again to to be to to compete with each other.
Adam:Yeah. I think, like, that it's hitting me most at home with stat muse. Like, I feel like so stat muse. Sorry. It's like a sports statistics search engine.
Adam:It's statmuse.com. Check it out. So statmuse, we're, like, basically dependent on Google. We get tons of traffic from Google, and I just don't know, like, where it's all people joke about, like, ChatGPT being the end of Google. I don't I just don't know if there's any sliver of truth there if, like, will Google's numbers start dipping because people go to ChatGPT instead of searching on traditional search indexes?
Adam:That's, like, that's a big deal for us. Statmuse, like, we depend heavily on getting lots of traditional SEO stuff. And if people start searching for sports questions in ChatGPT or elsewhere, how does that affect the landscape? I I don't know if you have thoughts on any of that. Like, is that stuff gonna change?
Adam:Are people actually using Google less? I guess I do.
Dax:So Yeah. I mean, I think I think I definitely use Google less. I think the statmuse case I think that's actually a great demonstration of what will continue to exist and what is different. So statmuse, you guys have two parts. Right?
Dax:You have the work of collecting and organizing information that is accurate, true, and correct. And that's a lot of work and that's where like a lot of the value is and you guys are putting a ton of work into organizing that. Then there's the interface to accessing that data. I think a lot of products fit, like, for how this dichotomy, like, even with Boomi, we help our our businesses, like, just collect accurate information about their customers and make sure everything's organized and structured correctly. That stuff I feel like is never gonna go away because ChatGPT gets trained on correct information.
Dax:So part of your business is like creating that and creating the source of that information. I feel like that's still gonna be the same and be valuable and ChatGPT might help you in like building the tools for that of course, but Yeah. It's not gonna replace that. People the interface for stat muse might end up being chat j p t. Right?
Dax:I don't think you guys have a problem there because they would just find stat muse through chat j p t if that's where they their entry point is.
Adam:And I'm so I'm very interested in the plugins thing when that came out for that reason because I I don't know, like, could we have a statues plugin? And I've applied, like, to do it, but I don't know, like, when we if we do have a statues plugin, will we get, like, attribution? Are people still gonna end up on statues.com, or is it just gonna feed chat chat jbt all this stuff and then we're just we're no longer relevant? I don't know.
Dax:Yeah. And that's why we're kind of a it's kind of a reset. Right? It's like business models need to change. How do you monetize this stuff?
Dax:Is it end up is it gonna have to end up being like a licensing thing? You also have to pay to use ChatGPT currently, so it's not like like Google where literally anyone can just use it. Like, does that model even make sense long term? Like, you do want people just accessing it without having to pay for it upfront. Right?
Adam:Right.
Dax:Yeah. So I have no idea. Like, business models had to completely be rediscovered again. Jay SST, he made a funny parallel. He's like, when the iPhone first came out, it was all fart apps.
Dax:Like, despite the power of the iPhone iPhone apps, it was really hard for people to think about anything more grand than fart apps. And he's like, right now, we basically have a bunch of fart apps on ChatGPT effectively, like not literally, but, like, in terms of, like, level of sophistication. Yeah. But now there's, like, you know, if you think about iPhone apps, there are billion dollar companies with, like, million line code bases that build iPhone apps. And we could have no one none of us could even conceive of it.
Dax:We had kind of the scaffold our way up there with, like, over time, we all start to understand stuff that we can do, patterns we can do. I talk about the pull to refresh thing a lot. Like, pull to refresh wasn't a thing for a very long time on apps, someone had to invent that. Now it's everywhere. So, yeah, I think we have to go through the same thing with with all this.
Dax:Like, we we don't know what to do with this stuff yet.
Adam:That it just reminds me that you we've talked about Simon Wardley before. Big, like, serverless prognosticator. Is that the word for somebody who makes, like, predictions?
Dax:I I've never heard the word prognosticator before. It's Alex. That up.
Adam:I might have made that word up. Prognostican?
Dax:Sounds medical?
Adam:I don't know. Maybe. He's a person who, like, thinks a lot.
Dax:Oh, like, prognosis? Prognos yeah. I see. Like, you give a
Adam:yeah. Okay. A maker of prognoses? Pro prognosis? Pro prognosi?
Adam:Okay.
Dax:We'll fact check this later. We're gonna ask Chad JPT.
Adam:Yeah. So he makes, like, a lot of predictions about the world. He maps. He does mapping, which is like a thing. I don't I don't know much about it.
Adam:But I just realized he talks about, like, the thing after serverless is conversational programming or something. And I never knew what that was. And it just occurred to me, this is that, isn't it?
Dax:He's right. Isn't He he, like, he laid that out ahead
Adam:of Five years ago, he had diagrams that were, like, mapping how the next thing after serverless is conversational programming. And when he says that, I just thought he meant, like, speaking English to program stuff. What the what is that? But this is it. This is like we're trying to figure out how to program on ChadGBT now.
Adam:And, like, what experiences can you build? Maybe? I don't know. I just did I miss it?
Dax:No. I mean, that sounds exactly like what we're doing. Yeah. Wow. That's impressive.
Dax:Didn't he post something the other day where it was like a chart of, like people are like, when's the singularity gonna happen? And there's a good chart of, human intelligence going up. Oh, yeah. You I found that because you retweeted it. And then it was like human intelligence going down to meet AI at the intersection.
Dax:Yeah. And someone was like, what evidence do you have of this? And they were like and Simon Worley was like, the fact that Jersey Shore is in its fifth season.
Adam:Do you really watch that show? What is it? You gotta tell me what that is. You said you re You
Dax:don't know what it is.
Adam:No. You re did you reply to me and say you watch it? Is that true?
Dax:Yeah. I I I watch every episode of Jersey Shore. I've seen every episode of Jersey Shore, and I love it. It's fantastic. Is it just
Adam:like Ken Wheeler? Is that basically what the show is? Just like a bunch of Ken Wheelers? That's what I imagine.
Dax:Here's a story here's a story about it. So I'm from New Jersey, and there was a really big reality show called Jersey Shore that was documented a very specific stereotype that does exist in parts of New Jersey. It's not, like, broadly in New Jersey. It's, like, the Italian, like, Guido, like, specific lifestyle thing. And these these people were just, like, insane, drunk, like, just not people you would look up to at all, and they did, like, two or three seasons of the show.
Dax:Then ten or fifteen years later, they rebooted it, and now they're in, like, a different part of their life.
Adam:Okay.
Dax:And they're all so mature, like, smart, and they've all, like, grown so much. And, like, they followed their lives now and, like, you know, they've had all these bad experiences because if they got famous and, like, rich and they did all this stuff, and but now they're in, a better place. Every single one of them, all of them in therapy, like, like, doing great.
Adam:Is it still entertaining?
Dax:I find it I find it very entertaining. Again, it's close to me because, like, I saw it when I was younger and I was in New Jersey, and they all live in New Jersey.
Adam:So Okay. This it all makes a little more
Dax:sense. That But anyway, apparently, that's proof that human intelligence is declining, which I'm not even gonna I don't even disagree.
Adam:Yeah. I well, is there any truth that so kind of going back to the AI thing, you said this person that tweeted that, like, intelligence isn't worth anything anymore. Is there actual is there actual data on, like I think there is, like, test scores and stuff going down for, like, the last fifty years or whatever in The US. Math scores have never been worse. I just saw this in the last couple days.
Dax:Yeah. I I mean, I think I think my whole life I've been seeing that information about that that going down. I don't know I don't know. I think there's, like, different components to this. Like, is biological intelligence, like, capability for intelligence going down?
Dax:You know, technically, I think there is, proof that our, like, ancient ancient ancestors were a lot more individually smart than us because they had to be. They had to, like, know all this stuff just to survive. Like Yeah. What plants are safe? Like, what can like, where where do I where do I need to migrate?
Dax:Like, where the weather pattern like, just as an individual, you have to be so much smarter. Now, we're collectively so much smarter that we individually don't have to be as smart. So I can see an argument for individual intelligence going down mostly because we have so much structure in place that compensates for that. And I guess that's kinda what his chart shows, like, as AI gets better, we individually Yeah. Don't have to be as smart.
Adam:That makes sense. It reminds me of the have you ever read Rational Optimist? I think it's that book.
Dax:Yeah. I love that book.
Adam:Yeah. Okay. So in that book, is that the book where he talks about, like, he's sitting at his desk eating breakfast and it's like, the plate was manufactured in Sweden. The bagel is made from ingredients from, like, seven different countries. Like, just all the different how our whole world is our our whole framework that we live in is made up of, like, collectively thousands or hundreds of thousands of people.
Adam:It's just wild when you think about it, actually. Yeah. So collectively, we accomplish a whole lot.
Dax:Yeah. Like, way, way, way, way more than what the smartest person who's ever lived can Yeah. Can do on their own. Like like, a 100 dumb people can do better than.
Adam:But, like, individually, we just watch Jersey Shore. That's how smart we are.
Dax:Yeah. Exactly. And that's kinda what ChatGPT is. It it is kinda dumb in some ways because it doesn't, like, know a lot of obvious things. But it also has, like, infinite memory.
Dax:It's basically, like, a billion dumb people.
Adam:Interesting.
Dax:It's kinda what you're tapping into.
Adam:So does it have infinite memory when you say infinite memory, like, you just mean it's trained on so much more data than any hurt one person could actually store in their head?
Dax:Yeah. Yeah. Just relative to a human, it's effectively it's like if we could just remember literally everything.
Adam:Yeah. Yeah. No. That's a good point. I've heard a lot of people say, like, it's like having an intern or, like, a thousand interns.
Adam:But it's like one person who has, like, bits of a thousand people in their head or more than a thousand, I guess. Yeah. Interesting. It's all hard to think about. It just kinda hurts my head.
Adam:I'm not very smart. We're not collective we're not individually very smart anymore. It's very hard for me to wrap my head around it.
Dax:So how do you feel now? Do you feel excited, worried, nervous?
Adam:I mean, I I don't know that I feel anything. I mean, this is just a me problem. I don't really have emotions. But I think there was a day where it was when they announced the plugins. I got a little anxious.
Adam:Not like I thought I was gonna lose my job, but more just like, I don't know what the future looks like, and that that just freaks me out
Dax:a little bit. Mhmm.
Adam:I'd also had coffee for the first time in months, and I was there was, like, coffee anxiety. But I think now I'm just kinda settled back down to a normal plane of existence. I think, like, I'm somewhat excited. I don't know that I have an emotional like, I don't feel excited day to day, but I think I am excited to see where it's all headed. It does feel like this is different than any of the big fads or like, people wanna compare it to, like
Dax:Mhmm.
Adam:NFTs and Web three, and that stuff just drives me crazy. This is not like that. This is, like, actually useful right now for lots of people. But I do wonder, like I I guess I go to the pessimistic side on some things, like, societally. I I just can't imagine if, like, some people are a thousand times more productive than they were before that employ like, employers will still continue to hire the same number of people.
Adam:Like, it just seems like fundamentally, we have bigger levers, so we need fewer people pulling those levers. I don't know. Maybe that's you always say, like, there's infinite opportunity or infinite desire. What do you say? Make me feel better.
Dax:There are just infinite things we have to do. I I think, yeah, in the short term, there's just gonna be some reshuffling, which is kind of already happening. Just ignore the AI thing. Just talk just looking at just the just like those shrinking in some of these bigger companies. So it takes time for that stuff to reshuffle, but I don't know.
Dax:Like, it's possible that this time it's different. Just but historically, like, when we've gone through phases like this, yes, there was a little bit of chaos, but then, like, stuff emerged from there that ended up being, you way crazier than we can imagine. Again, I wanna believe that it's true, so I I do believe it. But I also think it makes sense and there's a history in it. I also feel a little like, going back to, like, how we feel it, I feel a little humbled in some ways because, so my dad is a software engineer.
Dax:So he's been in tech through like other big moments. And looking back, I was kind of like, oh, man, like you were so close to all this. Like, how didn't how did why didn't you have the idea for like Uber or, like, all these, like, very seemingly obvious things? Like, why like, how come you would like, didn't work at Google, like, when they were early? Like, they all seem so obvious to me.
Dax:And I was like, oh, he was right there and he had the skills and he could have done it. But, like, now being at a similar point, it's like, oh, you just have no idea at all what's gonna happen. Yeah. And you just, like, try to, like, blindly figure it out. Yeah.
Dax:And it just seems obvious in hindsight, but it really isn't.
Adam:No. This is the most like, in terms of my entire career, fifteen years, whatever it's been, like, this is the most uncertain I've been about what my job will look like in five years, like, what all of this looks like. Yeah. It's it's it's just so fast. I think that's the thing is, like, you can look at OpenAI's roadmap and, like, yeah, they're shipping impressive stuff and doing it quickly.
Adam:Just looking around on Twitter, just like every day, there's some new, like, media thing that somebody put out that's, like, totally transforming video editing or totally transforming audio editing. It's just, like, completely changing industries, it and feels like it happens every single week. There's these major products coming out. And they look polished and, like, they're all because these AI models have gotten so advanced. So you do wonder, like
Dax:Mhmm.
Adam:What is the curve like with these models? Do they continue to progress? What is AGI? Does it matter if humans just keep getting dumber? Like, I don't know.
Adam:Is the target moving? Yeah. I don't know. It's all very interesting. It's hard not to be excited about it.
Adam:But at the same time, I understand when people are anxious. Like, I understand those feelings. And I do think it's interesting with the macro stuff, like just the the economy kinda slowing down at this time. Feels like lots of tech layoffs. Will they ever return to that peak?
Adam:I don't know. I'm a pessimist.
Dax:As nice as it was to know that, okay, worst case, there's always some, like, overpaid job at Google or whatever that I can, like, go go get. That was nice knowing that that existed. I think that's effectively gone now. I don't think that is gonna come back. But more broadly, if I really think about what I believe in, I don't actually believe that should have existed.
Dax:Like, I am much more into there being a 20 person companies than, like
Adam:Mhmm. Mhmm.
Dax:Way less of those and just like a few massive ones. I think we both like have been trying that for a while. So this could be the thing that helps that happen. Yeah. And that makes that kinda gives more people opportunity to work in a way that I think both of us have had the fortune to work in.
Dax:It just is way better and way more fun. So I think after the chaos and people there there are people that can, like, you know, figure out opportunities and and build stuff, I think they will build, like, very solid 10 to 20 person companies that that do very well. So I'm just hoping this just makes it all way more possible for them.
Adam:Yeah. That's probably the thing I feel most certain about is that small teams will build better products because of all this. Like, it would just be more common to see the really powerhouse small teams, which is exciting because I do love that environment. Yeah. This is good.
Adam:I always love talking about this stuff. I'm not an AI bro or whatever they're called. This is very exciting.
Dax:Is that a thing, AI bro? I feel like
Adam:I don't know. I I think a lot of the people that were into web three kinda shifted over to AI all of a sudden.
Dax:And I was telling you this, like, that literally happened in physically happened in Miami. There was a bunch of people that physically left that came here for the NFT hype that are now in SF for the AI hype. And it's like, man, what's it like to live life like that?
Adam:Yeah. What do you what do you think about, like, OpenAI as,
Dax:I don't
Adam:know, our warlords in five years? Like, are they just gonna be the major, major tech company? Because you could kind of imagine you could plot that out and imagine that they're just gonna be this powerhouse.
Dax:Yeah. We'll see what happens. I think if you play back some of the stuff in previous eras, it was kind of news to people that a company like Google could start out so small and end up being as big as they are. We had a bunch of examples of those like Google, Facebook, etcetera. So now we know very quickly that if something starts to get traction, like, oh, they have a shot at being a unicorn.
Dax:That I just feel that just wasn't as well understood before. Yeah. Because that's more understood now, the competition they're gonna face is really intense. I still think, like, they're very good and, like, the the team behind it is very good and they have a head start, of course. We'll see what ends up happening, if they are able to like I said, the business models around this stuff have not been invented.
Dax:It's a very expensive business to run at least today. And they get read that their next model is rumored to be trained on, like, a $225,000,000, woah, like, computing thing. So if GPT five costs that much, that's, like, insanely capital intensive. Yeah. And if it turns out, like, there isn't a great business model around this stuff, you know, that that's, like the part that needs needs to be figured out.
Dax:I think right now, people are willing I'm willing to pay, like, crazy amounts of money because this is super, useful for me, but you have to figure out how to make it effectively free and derive the money from somewhere else. That always falls back into advertising. But with something like this, advertising so clearly dilutes the value in a lot of ways.
Adam:Oh, wow. Yeah.
Dax:Because the whole pitch is, like, insane accuracy that you don't get. So it's like, what is the business model? And we'll see. Like, it might yeah. We have no idea.
Dax:I'm okay with them growing and being really successful. I know there's general anti corporate sentiment, which I get. But, you know, they're gonna create they probably already how much value do think they've already created? Like, they've probably saved billions in dollars already Mhmm. Just in in, like, people's times.
Dax:They probably will end up saving trillions and trillions of dollars. If they capture one or 2% of it, that's that's fine by me. So, yeah, I don't know.
Adam:We'll see. One of my favorite things on this podcast, X, is when we've basically landed the plane and then I ask you a question like that and just, like, totally spark things back up. It's probably, like, a terrible thing to do as a podcaster. But I love doing that.
Dax:We should I mean, we let's just double the length of our podcast. Let's go to let's go to
Adam:one hour. Weird how many times we hit, like, right on twenty four minutes. Not like 24, but, like, they round up to twenty four minutes every single episode. This one's gonna be a little longer, but like, that happens a lot. It's interesting to me.
Adam:We're just getting so efficient. We're just like little AI monkeys or something. Okay. We should probably get off here. This has been Fundex.
Dax:Okay. Say it again. Yeah. See you. See you.
