NEWSMAX

The Kentucky Derby, I think I don't actually know how long it takes breaking. Some breaking news. Barry Weiss to be named editor in chief of CBS News. Let's go. Congrats. People were predicting. Promotion, trade, deal. This is such a brutal deal. There's another post here. Paramount will announce deal to acquire Free Press and bring Barry Weiss to CBS News on Monday. I feel like this deal has been announced 50 times so that when it actually gets announced. I wonder why are media people obsessed with media people? Turns out, yes, yes, people love reporting on this. You could never get this much Press for a $250 million acquisition of your database, ERP Enterprise AI company. You'd be begging for Bloomberg and TechCrunch pieces. But when you're in the media, the media loves to tell a story about media and so you get a lot of press. In other news, Arvind over at Perplexity is announcing the browser, which I think.

And large chatgpt was not the catalyst. That's just like, you know, totally reducing the number of job openings. So there's another post in here from Simon Saris who says the zero interest rate era is going to become lost history because people want to make up a narrative around AI. The white collar bloodbath didn't happen because of a chat app release. It happened because of the end of ZIRP, which occurred rapidly in 2022. So literally when FTX, like ChatGPT was released, when FTX collapsed effectively during, you know, within like the same 30 day window. Yep. And so it really was a lot of CEOs woke up and said, we have to get a lot more efficient, we have to get more lean. Let's not hire, you know, a ton of new people, let's do more with what we have. The AI narrative from a, from a, from a job loss standpoint has also, it's also in some ways I feel like been a lot more narrative driven and that CEOs are, have a massive incentive to show they're getting more efficiency out of AI. And so they're just saying to their teams, like, I don't want you to hire more people, just be more efficient. Yep. And even that by itself of just forcing people to say like, no, we're not expanding your team, just do more with less. Yeah. There was also huge nomadic contagion from the Twitter buyout where X was able to run with like 1/5 the staff and the service doesn't really go down in their shipping features. And so I think a lot of CEOs just kind of adopted a year of austerity, a year of efficiency and wanted to test what they could do with less. And so they pulled back. So but yeah, you have to look at labs are scaling headcount rapidly. Anthropic is, you know, hiring teams to, you know, you know, capitalize on all the demand they have and they're some of the most, you know, AGI pilled of anyone. Right. And so you have to read into what, what people are actually doing and certainly that there's been massive job creation in the kind of startup ecosystem broadly just because of. We were talking yesterday with a friend. Seems like every, it's become so normalized, but it feels like every day there's like just on TVPN like at least 100 million of new funding announced, usually like you know, half a billion at least lately. And so a lot of that is going to go into hiring teams. Anyways. This original chart thought it was worth calling out. Yeah, it's important to reality check the put one chart over the other and see if they line up perfectly. We had some updates from.

Out so much stuff into the world that it just all gets diluted. And you don't read too much into any one video because it's not created by him. Do you believe in Cosgrove's law that. Aura farming is all you need? Break it down for us. Yeah. Always be or farming. Okay, I think I have the text somewhere, but yeah. Okay. Basically ore farming, right? Yeah. You can think of like most things are on some level like based off like interacting with other people. Like, like any kind of. If you're, you know, raising around, if you're going to make some deal. Right. It's based off like your interaction with someone else. Yeah. People raise, want to raise from tier ones because every dollar is the same. But you get to aura farm the, you know, the gp. Yeah. And then it's like in like when you're raising around, you want to be or farming the vc. Right. Because they want to be like, oh, this guy's like pretty sick. Like I want to give money or something like this. Right. So basically the idea is always be aura farming. Right. You always want to be doing something that increases your aura. Because everything you do, like any kind of interaction you have will be influenced by your personal aura. Yes, but stretch like wearing a fedora of nance. But stretch the analogy to the absolute limits. Sometimes when you farm, your lands grow fallow and you have a. You have a failed harvest. And you can also salt the ground so that nothing will ever grow. No bounty. And so there is a world where you go out to ora farm. You set up your ora farm, but someone has salted the ground and you don't grow a single crop. Yeah. So that's like a. I mean that's a failed ore farm. Right. Failed ora farm. So the question is the soar to putting yourself in the cameo. It is an attempted ora farming. Was it a successful ora farm? Yeah, I think most like I would say in most cases or farms you can only tell the success of a farm in retrospect if it's successful. Yeah. So. So it's like in six months we'll see. I think if people are still on Sora. Yeah. If the app is still like being used as a social media platform, not just like a creative tool. Yeah. Then I think it will be a successful farm because now Sam is just, you know, he's just everywhere. It does. It does feel like there's always a winner and a loser in an OR farm. And like if you are or farming. VCs will do this. They'll buy secondaries in a company to say they're an investor even though they didn't participate in a primary round. That's true. And then also for every time a founder aura farms a tier one venture capitalist and then blows the company up in a major scandal, that reduces the aura of the fund that. That put the money in. And they offend. They got farmed. They got farmed. Yeah. It makes. If you backed ftx, you got farmed, you. You look bad. Yeah. Like there's always two sides. Right. Because in Jordy's example, the VC is farming company, but when it gets. When it gets exposed that they actually bought secondary, that they weren't real investors, then it's a negative aura. The company is farming the vc. Yes. They're like, oh, we're so great that. But there must be positive sum or a farms where two people come together, they grow a beautiful garden together. When an emerging fashion brand partners with a legacy brand, let's say a streetwear company, partners with Salomon, they can both benefit because the streetwear brand is getting new legitimacy and Salomon gets to be culturally relevant and cool. And that is a positive sum or a firm. Sydney Sweeney and American Eagle. Positive sum or a farm. The stock's way up. Everyone's doing well in American Eagle world. Sydney Sweeney is back in the news and dominating. It seems like a win win. They ora farmed each other successfully. They grew a new bountiful harvest together. It was more positive for American Eagle. It was more like net neutral, I think for Sydney Sweeney in that she got a bunch of blowback from it. She got a bunch of attention, but it wasn't all positive. Whereas American Eagle just the stock ripped. Anyways, I have, I have an interesting post here that I wanted to throw in. It's from a guy named Robert.

So they're certainly getting in on the AI boom. Well, econ kids will see this and think I should give this guy money to help me get rich. And it's two of the most crazy looking individuals. I don't recognize either of these. This is. The guy on the left is a rapper. What's his name? Tyler. Oh, yes. He was the basis of Spring Breakups. This is Riffraff. Riffraff. This is Riff Raff. I don't know that he's actually done a get rich quick scheme or like. No, they just. There are some guys out there that look like somewhat. Like it's a very. Evocative by. Rob Freund. We haven't seen a course, bro. Just sell a course called how to make Money. Just Reasonably Priced and just really breaking down, you know? Here are the different ways that Econ. 101, baby. Tyler Cowan. Conversations with Tyler. That's all you need. Sell search services yourself. You can sell services that you deliver through a team. You can buy something for one price and sell it for a higher price. Yep. And that's pretty much it. Yeah. Naval kind of one shotted that entire category with the how to get rich without getting lucky thread. He should have dropped a chord. He didn't need to because he summed it all up in a series of like 20 pithy tweets. And you can just scroll and be like, yep, that's the algorithm. Yep. Anyway, mounted cash flow says always exciting when you see a new letter get added to the end of ebitda. What is he talking about? What's the.

You don't read too much into any one video because it's not created by him. Do you believe in Cosgrove's Law that. Aura farming is all you need? Break it down for us. Yeah. Always be or farming. Okay. I think I have a text somewhere, but yeah. Okay. Basically. Or farming, right? Yeah. You can think of, like, most things are on some level, like, based off, like, interacting with other people. Like, like any kind of. If you're, you know, raising around, if you're gonna make some deal. Right. It's based off, like, your interaction with someone else. Yeah. People raise. Want to raise from tier ones because every dollar is the same. But you get to Aura Farm the. You know, the gp. Yeah. And then it's like. Like in. Like when you're raising around, you want to be. Or farming the vc. Right. Because they want to be like, oh, this guy's like, pretty sick. Like, I want to give him money or something like this. Right. So basically the idea is always be Ora farming. Right. You always want to be doing something that increases your aura. Because everything you do, like any kind of interaction you have will be influenced by your personal aura. Yes, but stretch the analogy, Fedora of Finance. But stretch the analogy to the absolute limits. Sometimes when you farm, your lands grow fallow and you have a failed harvest. And you can also salt the ground so that nothing will ever grow. No bounty. And so there is a world where you go out to Aura Farm, you set up your aura Farm, but someone has salted the ground and you don't grow a single crop. Yeah, that's a failed Aura Farm. Failed Aura Farm. So the question is the sora to putting yourself in the cameo? It is an attempted aura farming. Was it a success?

I could never. What do you think about. I was, I was texting with Tyler about this. Like there's this post. Sam Altman is playing 4D chess. Sora 2 is about to take over social media. The virality is guaranteed. Once this scales, billions in ad revenue will flow straight into more compute fueling the flywheel. In a year, Sora 2 will be so efficient and cheap that margins explode. You can't out accelerate Sam Altman. And I wanted to reflect on this idea that Sam said. It's way less strange to watch a feed full of memes of yourself than I thought it would be. Not sure what to make of this. And gave. I mean he's. Again, my read on this is. I don't, I don't fully believe that he actually. I don't know. I don't know how to read into this other than he is very incentivized to normalize people being deep faked all the time. Right. I mean it's a crazy, crazy move to. To just be like somebody had to. I'm going first. My life, I'm merging. I'm merging. It's like him going first, being on. The first SpaceX rocket. Hey, it's not so bad. It's not so bad. I don't want to turn this off. I don't want to see that. I don't like the skibidi toilet thing to begin with. And then you mash it up and. It gets even more horrifying. Anyways, he went first. He's very incentivized to get a bunch of other big names personalities to agree to do the same thing. Is this. I would imagine it's very weird to watch videos of yourself committing crimes. You know, there's a video of guess a video of him in blackface floating around. There's a bunch of videos that I think would be incredibly strange to watch of yourself. So. But stranger than your expectation because I would expect it to be weird. I expected it to be weird when Tyler sent me a deep fake of myself bench pressing and then turning into a golden retriever. I was kind of just like, eh, yeah, this is below my expectation. Like it is less weird to me. I don't know. You haven't seen one of yourself yet, right? Has anyone made one of you? You haven't uploaded? I don't think I've done. I haven't. You haven't done the cameo thing? I haven't agreed to do the cameo thing. What about you, Tyler? Did you expect it to be weird to see yourself in a Sora video. And then was it weirder or less weird than you expected? I don't think it's that weird, but it's also like, it's still not like that good where, like, it's kind of me, but it's not like all the way there. Totally, totally. But I think. Okay, Jordy, do you think that him basically releasing a social media where like half the content is just him, is that an aura farming move? I think it increases his fame, which is a resource that he's able to leverage, but.

You know, was a driving force and create, creating Sora. But he's also. Videos of him are the ones. Some of the most viral videos created on the platform so far. So it really is, it really is him. But, but we are at this precipice where AI is cool and tech people are into it, but it's becoming a, it's becoming culturally like the danger here becomes like NFTs. Right? I think you were saying this yesterday. Yeah, Potentially hated. Yes. Even faster than social media. Yes. Especially with the video stuff. The video stuff has the same potential for like, oh, this just bugs me in the same way that the expensive monkey picture bugged people. People weren't that annoyed by Vitalik Buterin building Ethereum and nerding out about secure blockchains. It was when it became, okay, wait, what? The celebrity is hawking and Million Dollar monkey picture on a late night TV show and I'm losing my mortgage off of like, why is this happening? That's where people got really frustrated and I think it'll be the same thing if people are like, okay, I don't. Like this and it's increasing my life. Yeah. It's very different when it's like, like I have a job and there was a little bit of a boring thing and Now I use ChatGPT to summarize those meeting notes and it's delightful and I'm okay with that. But versus like this stuff is not good and I don't like it and also I'm paying for it. That sentiment that got 300,000 likes, that was also. There's also a growing sentiment around even the non coding use cases, what the value is. We had dinner last night with a very successful entrepreneur who's not really. He's built a widely known SAS company, but not really in the Silicon Valley bubble. Not kind of caught up in the AI hype cycle himself. And he said something, a line that stood out, which is AI is really good at helping people do unnecessary work faster. Right. Just like producing big reports, producing research. Right. What was your example? It was like, and, and you gave the example of, you know, if you're prompting an AI and you're like kind of bullet bulleting out from example an email or a document you can kind of get, you could kind of just send. Give somebody the bullet points. That's because I have ChatGPT running in my brain. Like if you send me three bullet points and you just say, hey John, imagine this is like five paragraphs. I can just do that in my head in two seconds. It's very easy for me to do. That based on facts, intelligence. Yeah. Also you know what else I can do? I can imagine that it's being read to me by a monkey on a jet ski. I can just imagine that if you tell me three bullet points. Hey, today we're going to talk about AI. We're going to talk about crypto, we're going to talk about ramp. I can just imagine a monkey saying that to me. Like you don't actually have to instantiate it, you can just share the idea and I can instantiate it in my head. But yeah, the example you had was funny, was like if you're going to a five minute meeting and you're like, I used AI to generate a 50 slide deck when in a 400 page memo and it's like, did that help? That seems entirely unnecessary. Yeah, yeah. You know, a while back when, when GPT psychosis was becoming, you know, more discussed, I was predicting that AI was going to go through a really, really rough PR on par with what social media started to go through. Yeah, I think that's generally happened. Correct. In the macro maybe. Well, no, I actually think, I think generally correct. Right, yeah. Every legacy media company has been putting out articles on people that are, you know, using AI in weird ways and going crazy or talking about how, you know, Chat, you know, ChatGPT or another product is like mentioning, you know, the word suicide a bunch to users. They, it seems like they, they released some pretty big updates recently that basically take people from 4, 0 back to Ch GPT 5 reality if they sense behavior that's unhealthy. But again, every legacy media company has been pushing out content around people going crazy through AI. Now you add in electricity costs going up and I think it's just going to continue. Well, the good news is that we've been here before and there are two positive possible outcomes here. One is just that the tech community comes together and builds a bunch of nuclear and solar and just power gets cheaper like that would be amazing. We need more energy production and clean energy production is kind of limitless. It's magical when it works. Problem is AI usage is exponential tokens. But maybe it's a sigmoid function and we wind up building even more and then prices do fall. It's totally possible. The flip side is that we've seen there are solutions to energy intense algorithms. We've seen this in past with crypto. So there was the same during the NFT bubble there was a, there was a huge push for oh, Ethereum is using so much energy for proof of work, they were able to migrate to proof of stake and, you know, energy prices fell. And it actually took a huge hit to Nvidia's sales that year because NFT companies and blockchain companies didn't need to buy as many GPUs because they were able to do the same algorithm effectively on a more efficient. So you're saying Jensen Satoshi, maybe It was extremely bullish for him for sure. But I do.

Reproduce forever. And so I don't really see my life as like a string of one hour tasks or five hour tasks. I see it as like a hundred year journey, potentially thousands of years when I think about my multi generations. And so I. It just feels like we're pretty far from something that is satisfying all the different criteria. And of course, you know, we need smell o vision and we're not making any progress there. Tyler, what was your reaction to this? Yeah, I mean it's very interesting. I think like Sutton is very rlpilled. I think like, wait, wait, RL pilled. Yeah. Sutton, are we in the RL paradigm? Yeah, but it's still, I mean we're still like very early days. Right. I think Sutton envisions this thing. It's like, it's just RL right? At its core, you start with nothing and you learn the world where the current paradigm is like first you basically pre train a bunch of stuff and you're basically like. At first when you're training a model, it's literally just like memorizing facts and then you slowly get it to generalize. Yep. So right now the way I think about it is like for a modern chatbot that I'm using, there's maybe like 300 megawatts of pre training compute and then 300 megawatts of RL thrown at it and it's like a 50, 50 balance or something. And then maybe he's saying it should be like 100% balance towards RL and like there. Well, so to ground this in Sora, part of why this new app is strategic is they're generating an asset and then they get to immediately see feedback. See feedback. Basically the reward is more engagement in the, you know, watching it longer. If you see a Sora video and you just on. On X OpenAI is not really able to leverage that. Yep. Right. You could indirectly be like, okay, this went viral. Like that maybe that's good there. But in the app, if you're in the, in the trough, you're immediately, you know, the time that you've spent watching the video, did you like it? Did you send it? All of that is providing feedback for the model R.L. Yeah, Tyler. Yeah, I mean, so I think if you look at just training like RL is nowhere close to half of compute. It's like way less. But I think it is like, it is like a very cool idea to basically have this very like pure model. In some sense it's just like experiencing the world and learning from that. Um, but I don't know. Like, if you go to the Dylan Patel kind of example of the baby, like, putting its finger in its mouth to, like, figure out where its fingers are. Yep. Like, the baby is not just, like, randomly twitching. Like, its nerves are not just randomly firing. Sure. There's still some, like, prior that comes from evolution or whatever, but there is, like, it's not just like, purely experiencing things and then updating. Yeah. There's something there I was thinking about. With the zebra example. Like, that's true, but also, like, the baby is kind of randomly twitching in the womb and. And like the baby, like, you can see on the. On the. What's it called? The oscilloscope or something. I forget what it's called. There's some sort of, like, device that you can, like, look inside and see the baby and the baby's moving, like, very basically, but probably, like, doing some reinforcement learning, essentially. Like. Like pre birth, I think. I don't know. Yeah. I mean, on some level, like, the. There is some encoding that is going on into the baby of, like, this is how, like, you can walk around or something like that. There's some, like. I think if you just had a baby and like, without any other people around. Yep. Like, maybe it could, like, learn to walk and stuff. But there's some, like. I think there is still some imitation going on generally. Like, Sutton is saying that, like, none of it is imitation. It's all just like rl. You're experiencing world and you're updating on those. On the actions you took. You have some goal. You either, like, complete the action or you don't. And then you update. Yep. I think. I think imitation is like, more important. Right. There's like the kind of like the Girardian take or whatever is that, like, humans are apes. Right. To ape is to imitate. If you drop a baby in a prison cell from birth, it'll never learn to walk because it needs to see a human walk to learn that. Yeah, something like that. I think they'd figure it out. Okay, maybe not walking but, like, doing, you know, useful things. Like, it's not going to. I think. I think if. I think if my son was born in a padded closet and just given food and an Excel sheet, he'd be writing VBA macros in five years with no external. Well, maybe it's been encoded in pure. Pure reinforcement learning experimentation from first principles. I think it's possible. I don't know. Anyway. Figma Think bigger, build faster. Figma helps design and development teams build great products together. Hey, man. Just wanted to say I think we talked about this yesterday, but I love it so much I put it back in. Hey, man, just wanted to reach out and say I loved how much you drank at the networking event last night. There. There are a few of those. This is becoming a new format. There was another one. I'll find the. Find the other one for me. In the meantime, yesterday, apparently there's a game that.

This is how we get to the Dyson sphere. Yes, this is a critical path to the Dyson sphere. Obviously, if you, if you're doubting that, you are. So what's notable here is that the guy, Chipotle Rowan. Yes. Says pretty cool that my electric bill has gone up 25% so other people can make this crap. And he gets 300,000. 300,000 likes on X. Like this is the, this is basically all the users. That's basically every user like this post. 8 million views, 17,000 reposts. It's a lot. And I, and I wanted to dig into some numbers, so I did fact check that 25%. I think it's real. So in 2021, the average electric bill for, for an average American was $1,452. It's like a significant amount. If you're, if you're an average American, you're making 60k or something like that, or 100k a year. And now that's up to 17 $28. That's exactly 25% increase. So it does seem like electricity prices have gone up by 2025%. And if you look at some of the charts, there really is a Spike right after 2020 when AI investment started ramping, electricity prices went up. Now some of this might be Covid, but it does seem like from 2010 to 2020, electricity prices were pretty stable. About 12 cents per kilowatt hour in the US city average on the St Louis Fed. But then 2020 hits, AI comes and things have been up and to the right ever since and not in a good way. JP Morgan estimated that 70% of last year's increased electricity cost was the result of data center demand. And the flow here is the logic is not quite perfect, but it's not great either. So it's that AI slop will be irresistibly dominant. We've seen this with everyone who works at our company becoming entirely one shotted by short form AI video. No one can get anything done. It's a disaster, it's a crisis.

African American over there doing work for us. Okay, question. Thank you for supporting us. Lightning round question, please. I want to hear about sourcing. You have three places where you can go to source the next great pre seed founder. You can go to a Cambridge Student Union meetup, meet young new grads. You could go to a pub outside of DeepMind after they just failed a training run. You can try and poach someone from there or you can go to Monaco during F1. Which one are you picking and why? I'm taking the DeepMind. DeepMind the pub. Hang out at the bar. Get the disgruntled DeepMind. Oh, the training run just failed. I want to start a company. I got to get out of this bureaucracy. Yeah, I think I take that in the last one, I think we invested in three ex Palantir people. So we love a good four deployed engineer. And so I think that's something we think that talent generally gravitates to, to places that have amazing cultures. And we spend a lot of time looking in the cultures of those companies, particularly in Europe. And I think you've had this one wave of, you know, the UI pass the Spotify. You're at the Kleiner ipo obviously. And that was kind of like one way which has happened and like, you know, gone public and now you're having this new ways. You've got Carl and Victor on the show. You've, you know, you've got Matty and there's others like, you know, Granola in London. And I think these are like people are just going bigger, right? They're just thinking bigger because they're learning more from these kind of repeat successes. And that's coming from all parts of the ecosystem. If you're working in a company out of a company, you're seeing these success stories and I think, I think there's a lot more to come personally. But having, you know, boots in the ground and that partnership with our American LPs, I find that, you know, typically America, the American firms eventually get in, right? Like every ipo, every exit, you'll see, you know, you'll, you win in the end. They're not always comfortable making that bet. So I think that's where we can play a role of kind of making that conviction early and going forward. Fantastic. I'm super bullish on congratulations in the fund. If I was an American, multi stage.

You're also not going to get that from Grok, apparently. No, not yet. There's also a wonderful ranking of energy drinks here. Notably missing is Andrew Huberman's Yer Mate Mateina, of course. But Goth says the Gold can is elite. If you don't believe me or don't get it, I don't have time to convince you. Sorry. And that's the caffeine free Diet Coke. I do the caffeinated Diet Coke, but this post is claiming that the Gold Can Diet Coke, the caffeine free and sugar free Diet Coke. It's just Forever chemicals, apparently, is the drink of men who are destined to find eras, change history, and who will be remembered for most pure forever chemicals. Carbonated forever. Well, if you're bullish on Coke or bearish on Coke, head over to public.com and.

Semblance of relevance. And I don't know if there's anything that only I find interesting. Maybe that is the true Hayes paradox. Maybe we're going into a world of more Hayes paradox content. Yeah, the concept. Everyone else is like, obviously we're joking around a little bit here. But I propose the Hayes paradox, which is the more funny you find something as an individual, the less likely that the masses will find it funny. It's inversely correlated, which is when you maybe have a funny idea for the post that. For a post that's cracking you up and no one else. Because it's too niche. We saw this. It requires too many niche references. Yep. We saw this with. We were joking about like Ben Thompson being on vacation a particular day and there was some joke about that. And like, unless you're, you know, unless you understand our context and then also the Ben Thompson context. There was like 100 people that would get that joke whatever we posted. So it just didn't do well. But I did notice that a few people who liked it were like my friends, my real friends, people in real life, like. And it would hit. I like finding, I like finding a video or meme that I know I'm not going to send to anybody. But you're just the six person group chat or just the family group chat or just the, you know, the small niche community on X or the bigger broader community and stuff goes viral all over the Internet. Brian Lovin says Sora 2 looks impressive. What's even more impressive is how little I care to watch anything made using it. The Notion team is just firing shots. They're going hard today. Well, is Notion. They are notions on turbo puffer search. Every byte serverless vector and full text search built from first principles and object storage. Fast 10x cheaper and extremely scalable. I knew I had to lay up there because it's such a logical partnership between those two companies. Alex Albert, who we had on the show yesterday, says over these next few years it's going to become more and more important that you resist letting slop consume you. Keep creating, keep learning, keep thinking. There is a wild, wild divide between the slop farm video apps and then the coding agents. No one's really leveling.