s03 e04 DeepSeek!

Welcome to the Big Cheese podcast. We're here downtown, Matt Indianapolis, Mass Ave. It's, I always have to look what day it is. It's Thursday, January 30th, 2025. And I feel like saying those dates is important for the pod because we keep having these, oh, this just happened. And then over the weekend, everyone loses their minds. And I'm like, well, go back in time. Listen to what we were talking about a week before you learned about this on the news, right? Yeah. And, and, and maybe you get some predictions, maybe you should, get some stock tips over here. I don't even know. But that is not financial. We knew that the deep sea, we knew that it was all going to be about the deep sea king, right? And so we've got the today's podcast is all about the just the absolute meltdown on the internet from post that the deep sea fallout. And I kind of want to start with, with a quote from Sam Altman. And it was, it was, it was a funny quote, and it was, it's older. It has an aged well. But basically, he was up there and just, you know, no one can compete with us. It's hopeless to compete with us. You will never be, you see, just as a matter of fact statement, no one's ever going to be able to create any model that can compete with open AI or keep up with their, their level of iteration on their technology. Boom, deep sea comes in. Everybody thinks they created a better reasoning model than open AI. And they did it for cheap. They did it for six million dollars. Right? But did they? And is that where? And so is, is so, so let's start. It was, was Altman wrong? What happened? And is this, is this for real? So, okay, so here's, here's a couple of interesting things. First, first and foremost, is that David Sacks, the all-in podcast person, David Sacks. So he had had been a Trump supporter for ever on the podcast. He slowly started converting every single other person on that podcast to be a Trump supporter. And, and weirdly enough, he kept promoting Trump promoting Trump promoting Trump. And now he is the Bitcoin, the Bitcoin and the AI's are of America, David Sacks. And so him and, and Altman are now convinced that deep sea stole a, a chat GPT's training data to build this model. And that they're, that this amounts to theft. You fucking assholes. You assholes. Like, how can you fucking stand there and even remotely pretend that that is a stance that has any fucking moral grounds? Like, it's, it's mind boggling that Altman can literally pretend to sit there and say, they can literally pretend to sit there and be like, they stole from us. It is no different than Steve Jobs when he got fucking angry at Bill Gates for when Bill Gates stole the UI from Apple, which Apple stole from fucking Xerox. And so the fact that these clowns are sitting here being like, they stole from us. Yeah, you stole from everybody. So, so it, it, it, it, it turns out that Altman thought he had this big moat. But let's be honest. No moat. His moat never had a leg to stand on anyways. Because they created that model off of data that they stole. Yeah. Well, and I think it's telling that he's getting so defensive about like the accusations of theft. I think implies that they are scared of well, we thought that no one could reproduce this. And to your point about their quote a year ago, the question was posed, hey, with a group of three developers and 10 million, do you think there's any chance how could we compete? And he was like, good luck. You're never going to do it. So I, it comes from the same guy that when he was asked about the, the pro version of opening eye, he was like, yeah, I set the price. I set it at $200. And they were like, and he was like, I didn't even put any thought into that. I just thought that would be a good price that people would buy. Oh, and then he was like, we're losing money. I'm like, no, this is the why guys he came from why commentator those that, that was from the golden age of sass where all it was was about was about reducing churn and increasing customer onboarding engagement and retention, right? He's a sass guy. He's, he's his whole thing is, let's just get customers. The rest don't figure it out. It's self out and I'll be able to raise money. He's just, the stakes are bigger, right? Now he's going and saying, just give me 500 bill. I'm the guy here. I'm the guy. And I think the biggest story in all of this is that saying I'm the guy is just not aging well. When you're talking about ripping off the entire human knowledge base trying to create a GI. Right? I hope that we're dating away from this like one person knows all the answer, you know, mentality. It's happening throughout and it's happening everywhere. And I thought that the best quote of the week was from I got to shout this guy out, Matt Stolar, STO, LLER on Twitter. He said, Hey, Microsoft, please stop trying to invent God and just make outlook search work. That was such a great friggin post. It was so timely. It was so timely that that so much of these companies are now spending all their time trying to do AI. And yet outlook is dog shit. But let's actually, I actually think there's a quick sidebar on this. What do you think actually is the customer sacrifice that's being made to chase AI? Because because I think that there's all this tech that everybody built and they're like, we don't have to worry about that anymore. Right. Right. You just have to put AI into it. And all of our broken stuff will magically be fixed. And I feel like in season one, we kind of a dream. That that's not going to be the case. Yeah. Right? Like garbage in garbage out. So Google just so Google just rolled out all their AI within docs within sheets within, you know, what do they call it? Slides. Right. Yeah. And but but you're not wrong. Like the idea that we can just throw AI into it and all of a sudden we have an entirely new product line doesn't change the fact that there's a lot of fundamental dog shit that you all are just completely ignoring and hoping that AI is going to kind of like cover. I think that brings up a good point. So I have seen a few people at the have this like approach that they're using Google L.O. notebook L.M. and they are basically saying, Hey, we're going to load it up with scientific documents and things that are grounded, grounding it in really, really good data. Right. Then they're asking against that. So think about is like instead of instead of trying to use an LLM as a general, we've talked about this so many times like a general search or whatever. It's like build your own data set and Google has basically given you an easy to use rag platform that has a great user interface. Yeah, I mean to think that you're just going to go on to slides and say, Hey, make me a tin, a tin, you know, slide deck that does X, Y, and Z without giving it any context or like supportive information is ridiculous. Yeah, it's going to be the same exact experience that we've had when we're trying to use AI to do the development where you can go in and be like, Hey, here's, I want you to do this, this, this, and this and this and give it five fucking steps, right. And it's going to go through and you're not going to realize all of the nuances that you didn't even think about that the AI has to kind of think through. So if you're doing any of this, so if you're using Google's AI, if you're using any of them, the last thing that you always say when you give an AI instruction is ask me any clarifying questions. And it's the most important fucking thing that you can actually do to these things because then all of a sudden what it's going to do is it's going to be like, Ooh, it's going to give you a chain of thought. And the moment that you ask it that it's going to be like, Ooh, I need to kind of think through some of these things. And it's going to be like, Oh, actually, when you said this, do you actually mean this, this, this, this and this? And it gives you the opportunity to kind of correct it. And that's been the most important thing as I've gone through this journey of building these apps just with AI. Again, I'm not, I'm not coding anything. I'm just using AI. People are like, well, you could just take over. I'm like, no, that defeats the point. The point is that AI has to build this 100 fucking percent. And I have to scream at it for weeks if I have to to get it to do it. Right? Like that is my mission with this thing. And, and, and, and I feel like I'm kind of in that same spot with this other stuff. Yeah. One other thing because you know me, I'm a Google fanboy, but I just discovered this feature that's actually been out for a month now. It's Google's, what is it? Deep research. So basically, you know, the whole search, the web stuff on chat, GPT and all that takes it to the next level. Like it'll go out, like I did one the other day and it went out and searched like 50 different websites. And then aggregated it all for me. And I'm like, oh my gosh. This is a big. It's still not magic. Like there's still a level of like at first, it seemed like magic. And then I read some of the responses. I'm like, okay, that website's probably bullshit. You know, like you still still can give you bad results. Well, there's a big distinction that I think the average consumer or even business user doesn't understand. There's a big difference. Okay, what's the benchmark on this big LLM that has a bajillion parameters and then I'm asking anything and it knows everything to you know, what's what's this smaller, you know, what's my data going to tell me? Because really at the end of the day, people would just want to have a different interface to their data. So I'm going to select star from users where you know, they want to hear a spreadsheet. Even you know, I have to do stuff. No, they just want to say, hey, what is this? You know, they want to talk to it. Well, I agree 100%. The hope in the dream is eventually all this AI translates to more conversational access accessibility. We talk about accessibility a lot on this on this podcast. But to your, your, your stuff that you, yes, that it's around, not necessarily every single thought that's ever been crit and the end of the honest, those general intelligence models, that was a big bet that only who made that bet? Who made that bet that it was going to be worth a bajillion, jillion dollars did did did Facebook make that bet? No. Or else they wouldn't have open sourced their their model. Right? Did you see Apple make that bet? Nope. Did Google make that bet? Yeah, but it kind of makes sense for them to invest in AI. Yeah, Google created it. Yeah, but they fumbled it. They literally created the whole transformer model. Are they getting people to spend more money on Google workspace? No, then they previously worked. No, probably not. They haven't. So real site, a quick sidebar on this is, is that what Apple, what people don't realize is so Apple created the transformer transformer model. So a GPT, where it's a general transformer, I don't fucking know whatever it is, but it's based on what Google, where Google came up with. Now there's a new version that's a 2.0 model that Google has created. And I think we're going to see everybody once again is going to hard on this. I'm hoping that that Google this time doesn't fumble the bag as bad as they did on the first time, which they created it. And they're like, yeah, that's cool. And then open AI basically. It's like, yep, we're going to take this GPT model and we're going to just make a whole fucking business out of it. So there's a whole new thing that's going to be happening a new and I won't even pretend to pretend like what is it? How is it different? What does it do? But we're going to be talking a lot more about it. It's going to be a new transformer version 2 model. At the end of the day, the the moat, right? The moat for these products is not what they thought it was going to be. In fact, I have another tweet that I thought was really good from this week. And it goes and it goes. And it's turned out that the GPT rapper has more of a moat than GPT because remember how all the tweets last year was like everybody was building an app that was based off of chat using chat GBG's API or whatever. And they're like, well, it's just a rapper. And then people would get on Twitter and they'd be like, oh yeah, well, the internet's just a rapper for the telephone line. The telephone line is just a rapper for the electricity. And electricity is just a rapper for the network. I think everyone was like, you know, go ahead. Yeah, I agree. We were harping on it too, but I think we meant in the sense that if all you're doing is wrapping to a chat interface, to chat GBG, you're not adding any value, right? So we're talking about like it's a rapper. Yeah, that app that Brandon wants to work on. Yeah. Brain nameless. No, I will tell you about my let me tell you about my new one that I've started. Yeah. So I'll tell me, let me tell you I'll use this opportunity to promote my own product paper keeper. It's in the app story. You actually got this built this thing and got it approved in the app store. It's like weeks. Dude, well, using a approval process was a nightmare by the way. Yeah, but you're done. I'm done. So it's in there. So now my wife is going to have to go and market the same paper. Did you actually create a monetization? Yes. So it's 1299 before a lifetime unlimited pro account that gives you unlimited profiles and unlimited documents. You just raw dog it. So I try to go through and figure out some like smart and I'm like, ah, fuck this. Let's just make it 1299. That that feels right. That's the most Brandon thing I've heard about. That's the good. It's a good. It's good. It's good. It's good. It's good. It's good. Product and I actually use it. So any time that you have. So we all done this. We've all taken photos of important documents that are in our camera roll. Right. They're in our camera rolls and we're like, yeah, that's where it's at. And then once we need it, it's a nightmare to actually find. So this is all this is literally an application that just allows you to have all of your important documents in one simple spot that you can share across all of your documents. So you open it up on your iPad. It's there. If you open it up on your iPhone. It's there. All of it's just synced naturally through all of it was written by a high agency dude. 100% prompting AI. 100% of it was so you heard it here first basically. In this world that we've basically been saying, hey, it's actually there. And you can do this now. We have witnessed someone. Not just talk about it, but be about it and get all the way through the app. Right. Yeah. Which brings me to another story. Okay. Okay. Paperkeeper.io. Paperkeeper.app. Paperkeeper.app. And I've demoed this software. It's really cool. I think I'd be 1299 for it. Can I can I put all my newty picks in there? You can get absolutely. Okay. I don't have any of those. I just want to have a ton of them. I'm sure you do. So so there's all this talk about which I think is actually funny. And I think that the younger the kids that more likely this is actually happen. This whole concept of one shoting. It's a video game thing. Yeah. Okay. I never knew that. Well, actually this first came to me from video games. And obviously you hear people talking about it on on about coding. Right. But like you know, like I play a bit a game called Elden Ring. And like you can power up your character enough. So you just walk into a boss arena and just one shot that didn't walk off and everybody thinks you're cool. Well, there was a pretty viral tweet that came came from a it was a I think it was Italian because the the tweet wasn't mostly an Italian. But it was like I one shot of this entire video game. It was like F zero, but 3D. Okay. So it was a in browser web game. But it looked legit. And the car we had there was all kinds of physics going on like the car was it looked like like a like a forza. Almost type of you. And it's all in the browser. And it was all in the browser. And the guy was like I built this in bolt. And everyone was like like all the initial replies were like this is amazing. This is great. How not even how just like well the AI is amazing. AI is everything. Okay. So this guy is going this guy is going to market saying that I built this forza. Racing game using Bolt. Yes. And for the it's bolt. I know it's well it's bolt.new really. Yeah. Okay. I don't know the main domain. Right. Right. Right. Right. Yeah. Bolt is an app where you can basically try to one shot an app. But you iterate on it. But it's it's trying to not just a cotasist or build you eyes. It's like I'm going to build the whole project. It's going to try to do the whole thing through props. So it's the comparable product would be V0 from Versailles. Has anybody have either of you guys built anything with a bolt? I've I've messed around with it. Okay. Yeah. I've scaffolded a few things. I always end up going it away because I'm just like I'd rather start from a template. Right. Then so site quick sidebar on bolt. It's really grinds my gears. I think that we're going to move more towards templating things because you think about it. Like if you have a SaaS product or a website or a CMS or a CRM or whatever name you're like starting point. Why am I going to go to bolt to recreate that from scratch even if it's going to pull from all the examples when I could just create it from a template. And then and then use that to supplement it. Right. So that was the punchline sort of on the story by that. It's a good point. It's a good it's a good segue because the guy didn't one shot shit. He didn't one shot anything. Bolt in fact, he just completely hijacked a product release from bolt bolt dot new slash get hub slash get hub repository name latches on to a repo essentially preclones it and lets you run that it tries to run that software. So he found this this this random ask guy on the internet found a contained repository that had only front end code that bolt was able to suck in and run. So basically it was the equivalent of get clone mdm run right like they bolt didn't write any of that code. There was no code written and but but here's the sensationalist AI tweet and it just keep serving these these constant constant constant social media posts like this will change everything this will change have you seen this. Well, this just got released here the top 10 things you need to know. It's just constant. I have a real problem and I know that this is kind of what we're here to talk about is like linked in YouTube Twitter whatever it is. I have a real problem with just the the shallowness of conversations that are going on in the internet right now. So someone claims to have one shot at a very complicated that no follow up questions. No one like the masses are just coming to be like first. Well, it took him a couple days after it went viral. Some people shot as a stand but it didn't matter. It did the damage was done. The general public is going and just believing which is shocking to me because I am so fucking skeptical anytime I see something on the internet that my first instinct is always well that's fucking bullshit. And it's funny that it's funny the stuff that so I quit Instagram so my addiction now is back to X. I hate I hate a lot of it. It's you know it's good for sports and it's good for coding but like the things that they choose to do the fact checking stuff on is really funny because it seems actually like like a couple things that the X will absolutely put the this is fake news if it's advertising a dropshipping product. Oh yeah if it's advertising a dropshipping product it's going to put the it's going to latch onto it if it's remotely related to left wing politics. The fact check it really will they do anything related to AI though or coding or anything like that the community is just reddit style fact checking it. You know and that's why I do think like ultimately all things need to become reddit and have that social there's just not enough social proof and social moderation. Yeah right you know have you seen reddits answers it's in beta right now so not entirely sure why this product needed to exist but it's I guess it can we'll be convenient so it's basically an LLM that only is grounded on reddit community so so opening up just. I know nothing about other than I asked it a few questions like one was like I was like hey what's a good business to start if you had a hundred K I'm like that's the thing I think people ask quite a bit. It gave it it was like what was cool was like when I go to reddit or Google and say if it's I should if you go to site colon reddit.com and then ask that same question it'll essentially do that work for you right but it doesn't pull out the comments from. Maybe it'll say it'll show the post and then you have to go find the comment that might be the most relevant this is like I asked it and it was like. Steve Salamander. It's comment. Steve Salamander. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. Hi. So. So. So. So reddit. Blue sky maybe. Yeah. I want to read a tweet from Mark Chen. Mark Chen is the chief research officer at OpenAI. rats to deep seek on producing an O1 level reasoning model, exclamation point. Their research paper demonstrates that they've independently found some of the core ideas that we did on our way to O1. However, I think the external response has been somewhat overblown, especially in the narratives around cost. One implication of having two paradigms, pre-training, and reasoning, is that we can optimize for a capability for over two axes instead of one, which leads to lower costs. But it also means we have two axes along we can scale, we and we intend to push compute aggressively into both. As research in distillation matures, we're also seeing that pushing on cost and pushing on capabilities are increasingly decoupled. The ability to serve at a lower cost doesn't imply the ability to produce better capabilities. We will continue to improve our ability to serve models at a lower cost. But we remain optimistic in our research roadmap and we'll remain focused in on executing on it. We're excited to ship better models to you in the quarter and over the year. What do you think? Fucking a horse shit is what that is. That's absolute bullshit because the fact of the matter is so my next application that I'm working on is called Tiny Labels. And the whole premise is I'm going to take a photo of a label that I can't see because my fucking vision, you can see my big ass, shitty ass vision, right? I can't see labels. So I'm going to take a photo of it. I'm going to process that label. I'm going to turn it into here's exactly what you need. Here's your warning, whatever the label is, it's going to automatically adjust to it. And you know what I'm going to use? I'm going to use this new deep seek vision model. I heard it's good because it's super cheap because I do have to think about I don't have access to it within Apple. By the way, which is horse shit, I should have this within Apple. I should have it on device. I should not have to be sending this out to some external thing. But I think what I'm thinking is, is I'm going to use base 10. And I'm going to set up base 10 with this new vision model that deep seek just came out with. And that's what I'm ultimately use. Well, too, too quick. Go ahead. I was just going to send everything to China. Oh, no, I'm going to send it to China. No, no, no, no. So, so, so let's make that big clarification. So for everyone that thinks that deep seeks evil and they're going to steal your data, they open sourced it. You can run this stuff. Yeah. Yeah. So it's a big time. If you go to their website, they're not going to tell you about TNM. Okay. So, but if you run the model locally, it will, which shows you that the fact that it's not necessarily an external China threat technology in the open source model. It's not. So, so real quick. The. I think door. Yeah, I think it's very important to say that do as an American do not use the deep seek app that's in the app store do not use the deep seek website. Just don't do it because what you what people don't understand is as you're going through and you're having these conversations with these a eyes, you're giving a treasure trove of information. And if you go and you do a comparison between open a eyes privacy policy and you go and do a deep seek privacy policy, it is a fucking nightmare. It's absolutely abhorrent. They will absolutely keep everything that you put in there. There's nothing that you will put into deep seek that they will not hold on to for as long as they possibly can do not use it. So if you want to use deep seek, you need to use it at base 10. You need to use it at at AWS. I think as you're just just came out and said, Hey, we're going to be supporting it where those models you can actually go and you can use them in that environment. And you can ask them questions about TN men square and they'll give you answers for it. When you go and you actually are using the applications, it's no different than when you use the applications of chat GBT is that there's a layer that an application layer that sits on top of the intelligence and that application layer is doing all the things that's still in all of our PII that's like really understanding what's going on. So just do not use their applications. And right now it's the number one fucking application above chat GBT. Don't use it. So I'm begging you, don't use that fucking application, even though it's free. And they're going, they're, they are 100% going to go after open AI and every single feature that open AI has that they're charging $200 for, they're going to offer it for free. There's a reason they're offering it for free. Just think about that. So remember when React came out. Oh, yeah. So React would became the ubiquitous framework for building user interfaces built by Facebook. By the way, as much as I hate on Zuckerberg in this podcast, my entire career was stuff that happened. Literally, I have the utmost respect for Facebook's engineering, their team. I've always relied on that because I mean, if you've seen the movie, the social network, the dude was, was, was, was always about the engineering. Yeah. Right. But the, the idea, where was I going with that? Oh, yeah. Yeah. So, so, so, so, so react. This unbelievable framework comes out. Well, right behind it, view, right? View. And now Evan, you, you know, he's, I mean, he's of Chinese descent, but he's not Chinese. I mean, he's, he's an American. But one of the things that I think that happened from that is that there, I think there was a scale from the Chinese open source community. Yeah. 100% with you. And view was a complete ripoff of, of, they, they, they waited for Angular and React come out. And they said, you know what? What can we do? That's kind of, we kind of like this. I love this stuff that's coming out of America. But you know, well, what can we do? What can we kind of do? And so they, what do they do? They do what Chinese do? They copy, they copy what you do and they try to improve it. And they, honestly, they, I was a huge view fanboy early on. Now, one of the issues with that, if you're an American trying to use their stuff is a lot of their documentation and, and, and, you know, informational and it wasn't Chinese. So it's, it's, I mean, one of the best state management libraries that the best state management library I ever used was VX. Yeah, VX was great. It was so good. It was so fricking good. And that's one of the hardest things to do on the web is keep track of all your data and all your state and the browser. I distinctly remember looking at Redux and being like, I don't, I don't love VX and he with Redux. He was like, I was like, I don't get what we're trying to do here. Redux was designed by fucking social. Yeah. Yeah. He designed like Adromoff. So a Russian. Yeah. He's a great guy. He's a good boy. But the, so we ended up using VX for, so the, and IMS Indianapolis motor speedway, the Indy 500 fantasy racing game. If you go use that, that's all in view. We did that when I was at fusion. And, and I think it's just an illustration of like that, that you have to know that about Chinese technology is like, they're not, they're not going to necessarily necessarily be the creative ones. But you push something out. They'll find it. They'll hack it. They'll reverse engineer it and they'll, and they'll repurpose it. And there's no stopping you, not stopping. In fact, they should have chipped, to shift them those Nvidia chips, because they would have done it on more expensive hardware. Exactly. So, so that was the other piece that we, what we did is we're like, Oh, by the way, you can't have these bad-ass hardware. So you can't build these cool models. They're like, All right, fuck off. We're going to go and we're going to figure out how to use and how to build these models without using that super hard, expensive hardware. And that's where that's, that is the limit. So we put that restraint on China. And we said, you can't have access to this. And they're like, hold my beer. They forgot the one most important thing in life when there's a will. There's a way. There's a way. Exactly. And they have a way. And they did it. And they did it for six million fucking dollars. No, that's debunked because because you can't, you can't, you can't say that, right? Because they used models that are have already existed. No, bullshit. How much how much? Okay, fuck that. How much money they would not have been able to do it out of their pocket. Did they spend six million bucks? So, but I don't give a shit that some that open AI spent 18, $18 million. No, that's, you can't think about it that way. You can, but you can't, which is they, they, they, they, they rode the backs of someone else's work. Dude, but that's what we all fucking do. Exactly. That's literally the, the public perception. I don't give a shit about the public. Well, let me finish. How much money the public perception is that they were able to create this from scratch for six million. No, bullshit. It is. It's told them that they created it from scratch as a fucking idiot. That's that's the public perception. You know that, but I agree that you're not the average. Who gets a fuck about the average Joe? Like, why do I give my mom my mom their own shit? Don't leave it out for this. Don't leave this all of you. Yeah. Why do I care what my mom thinks about fucking deep seat coming up with a new AI model that was built on nothing? It doesn't change the fact that they can, that they can offer the service for cheaper. Yeah. It doesn't, but it does, it does matter. And X cheaper. It does matter that they, they took in models and abstracted on top of that. No, no, no, no, no, no, no, no, no, no, no, no. I do, I do, I, I get you so here's the thing though is that that they're the every fucking Model that has been released exactly is trained on synthetic data from some Other model and they might not like it but you fuckers have stolen from the entire Internet so the fact that we're stealing your synthetic data from your Model to generate new models So so here's the problem with that game theory there's a thing called first mover Advantage and last mover advantage and in this game you have last mover Advantage but it doesn't you have to see through the crack a little bit right where it's like well someone else is just gonna do a Chaper cheaper cheaper the bottom line is that the the model is Going to be the commodity. Yeah, I'm also really excited because okay if these guys can do it for 10 million right? Well Facebook or meta right now is you know, they've got their war room here about the war rooms Yeah, they're doing their war rooms because Zuckerberg was probably go ahead wait Well, we talked about this a few weeks ago where we were like all right the Lord the models have gotten big enough the CPUs going as fast as we've got it Clocked as much as we can now how do we do the caching on us? How do we Create more cores. That's what these guys figured out before anyone else really was they optimized the thing That instead of just making it a bigger and what that guy's tweet was about was oh yeah Well, he's there's two attack vectors here like we have two areas we can improve upon they went down this path That was the most Like just absolute loser tweet of all time from the chief research officer of open AI Someone burned him so bad in the comments that they were like yeah, but The oh one I can only hit 10 times a day and it's like super expensive and there's I can just like hit infinitely So let's just do a day in the life of this fucking chief creatures officer versus a day in the life of whoever just grinded on deep Seek for the last six months. Okay. This guy goes to work Probably let's say 10 hours a day. Let's say he works real hard. It's a little come Valley right? He drives his Tesla model Y long range. He's got three kids He's he's literally makes 600 650 grand a quarter. Yeah Probably yeah, and he's not doing it off profit because opening I does not make any money they lose money. They are they are I start up So they're taking all this money and they're raising all this they're convincing all these people to give them all this money Saying that they have something that no one else has and they're and they're losing and they're losing Yeah, because they don't know the game the game is data and guess what the data is out there And if you if you're if your moat is something that can be scraped You're you don't have a moat. Well, but I'll by plan tear here Here's where I will disagree with you though is because they've got 500 billion reasons why they are still Have a great chance to win well. They've only got a hundred billion of them now And Larry Ellison looks like he's he's gotten us to that cash if you can convince all those people with more money than God to Invest in you you can they can beat them all day This is the one thing I I'll I'll give Trump is he's like y'all y'all you'll figure this out You know, I'm not I'm not he didn't give any all men any money literally He did his his game. I'll give you I'll give you like I'll give you a presentation on TV That was his thing. I'll be I'll you present that you're your work, but you notice that he doesn't Doesn't really they do he as part of the I guess the United States of America Didn't make a big commitment. No, they don't we talked about that last week. I think you might have been sitting in that meeting and going I'm gonna do this to piss off Elon cuz cuz I need that guy out of my office But like I don't actually know if this is something I'm going like long or he's just promised them the world and he's He'll follow through if he feels So so Vivek we already know the Vex out the Vex out we got doge now and so I watched It was the breakfast podcast that they do great great podcasts where he was talking about it It was basically just we have a different we have a different vision So he did so he did admit there was a different vision. There was a different there was a different opinion in that like Where Elon wanted to take doge to be more of a technology where he wanted it to be more law based and like passing laws and all that and he's like Well, there's a hatch act so I'm not able to actually do it because I want to run for governor of Ohio That's what I thought it was yeah, yeah, well, so so the thing that I think is important to know about Vivek and Elon and One of Elon's really good friends kind of went public against Elon. Yeah, and that is both of those guys were really bullied in high school So so if you go down to the past and you go through Vivek and and Elon have a very very similar childhood Which is they were nerds they got made fun of and they're literally especially you on Out for blood dude. I have one for I saw so real quick I'm not gonna say his name. I want to say his name so fucking bad and if you follow me on on on Facebook You know who this person is that I have one of those people and this guy was the most insufferable Tawatt you've ever met he was insufferable in junior high He was insufferable in high school He was just this and he was just this fucking he was he was a dork and now he's made a lot of money for himself He's done very well for himself and now he thinks he's just the fucking God's gift It's it's it's a dangerous. It's a dangerous gift for when you go from dog. There's nothing I don't know what you just said but sorry. I mean there's there's nothing worse again New fun of when you're a kid. Yeah, and there's nothing there I mean besides maybe your dad leaving you or your Right like like like I mean there's other works. Yeah, okay I don't want but one of the like modern American society where you generally have something to wake food to wake up to a Ride to school. Yeah, you know access to things. No, you're right. Not being accepted as a child is the ultimate like well Not the old again with all the other things considered the ultimate in the average modern Kids life they hurt your psyche long term it makes you not a very Dog yeah I want to say his last name and so fucking perfect Trump Trump got that from his dad is dad treated him like shit He treated him Donald Trump junior like shit. He was never good enough That's the point these people are all out to prove like they've done good They've done bad in all in between they're all out to prove to everyone how great they are and how they're Zuckerberg's same thing. Yeah, and Schneard didn't get girls built Facebook to spite them all. Yeah, well, did he have originally okay This is gun sideways, but did he originally build it to basically like Yeah, define the chick I built shit to find chick Literally, literally in genius app. He scraped all the all the sororities and the And he just he you could create a little Harvard book thing that they did yeah, and then you just pressed you just pressed face Mashed and it would randomly join It was you would really join up a man and a woman at the school and and people would be like oh cool Yeah, and this and and and and that was it that was why he was it yeah So I mean don't trust these people Okay You know I don't trust see I totally respect why combinator I totally respect the engineering teams at Facebook and But what's going on right now since the inauguration since the election you there's a there's a there's a grab for this power And the power is somebody figured out how to take all human knowledge and put it in a vector database And they're like who's gonna own that and really the answers Same as my favorite line from Ernest goes to camp Who can own a tree