X’s $1M Sweepstakes, Ben Affleck’s viral take on AI, Elon sues OpenAI for $134B | Diet TBPN

0:01We're talking about the article apocalypse that's happening on X. Will it stop the timeline? Will the algorithm handle the flood of AI generated content that will be a lot of AI generated content that will be a lot of AI $1 million to the top article posted on X, you have two weeks. And very, very different structure. They thought, there's so much AI-generated essays on X, why not make AI-generated essays lottery tickets? Let's turn them into lottery tickets. I like it. I mean, they're definitely paying out more than a million dollars in creator payments across all the little$100,$1 ,000 checks that go out to folks, you have to imagine.

1:00But running a sweepstakes, I mean, that's time honored. What's not to like about a sweepstakes? It's great. Nikita's kind of taking two sides to this. He says, it's time to write, but also, he tells everyone else, please bear with article Armageddon. My question was, what's the strategy here, and what will the outcome be? The Elon taking over felt like, okay, it's an entirely new day, everything's moving way faster. Now, some of those projects had actually been been in the works for months or years before. I think Community Notes was the big one where the Twitter team had been building that infrastructure for a while, but they were just moving slowly.

1:38They had 5 ,000 people working on it for over five years. Probably, something like that. 2023 was a big year for expanding character limits. They'd also previously acquired this Dutch startup Review, which was a direct competitor to Substack, But that product was shut down in less than two years like January to January They only made it two years never really got it off day off the ground And I think it was because it was sort of counter positioned with Twitter was kind of like always Oh, well if we get people on the review platform, they'll leave or they'll take their audience elsewhere I never really felt yeah, and I don't know that a lot of people had some incredible faith that hey I want to build my business on yeah, like a Twitter Like yeah, I mean even pre Elon there had been companies startups that had built on the Twitter API, which had gone through a number of version changes.

2:26Elon made some more extreme changes to that. But there were plenty of times when people would build businesses like too closely to any platform, really, and then they get steamrolled, whether you're, you know, Zynga building on building Farmville on Facebook, and then they change the algorithm or they change the integration. And all of a sudden, your growth engine is just gone. This happened with Teespring, they were selling custom t shirts, you know, they've been working on on making x a more literate long form platform for a while, 2023. We need to teach our users the English language. We need them to read more than 180 characters.

2:58So they've obviously done a ton in video. They've never really leaned into a video library like YouTube. I used to cross post my YouTube videos on X. Sometimes there were length limits for a while, then they got rid of those. So I could just actually copy a 25 minute YouTube video, put it on X and it was cool. People wouldn't watch nearly as much as they did on YouTube. and there was no catalog value. That was like the really weird thing was that on YouTube, like my old videos that I haven't uploaded in over a year and my old videos are still probably getting like, I don't know, half a million views a month just because people are every day, they search, like what's the history of Anderil?

3:35Or what's the history of Ramp? Or what's the history of Xi Jinping? Well, they're evergreen content, so it just continues to get served. Exactly. And then once someone watches one of my videos, they watch five more. Well, and part of that, if you want to be a serious video platform, you need to be like heavily, you know, actually incentivizing people to watch on TV. Like so much of YouTube watch hours happen on TV. And X, I think has a streaming TV app, but it's certainly not, doesn't get a lot of love. Continuing with the long post journey that X has been on, 2023 was a massive year. February took it from 280 characters to 4 ,000 characters.

4:11That's more than I need for most of my posts. I don't think I've ever posted anything longer than that. April expanded it to 10 ,000 characters and finally in June they maxed it out at 25 ,000 characters. That's longer than most blog posts. That's around 4 ,000 words, maybe a little bit more. And you can still thread these posts together. So I was always wondering like why does articles exist if long posts are so long that you can put 4 ,000 words there? And now I think that essentially articles are first-class citizens in the algorithm and there's really no downside. Zach Fole had a great idea.

4:48He said, Grok, summarize this article in 280 characters. Don't make any mistakes. Thanks. So if you really yearn for the days, the pre-article era, just respond and tag Grok into every post and have it turn it back into just regular old-fashioned. 140 characters, even shorter. So when articles launched in March of 2024, it didn't really hit for me. Like I was not like, oh, I got to jump on this. I got to post. You know, we started doing the newsletter in 2025. Mid-year, we started cross-posting. Articles were available. I think we might have tested one or two as an article. Didn't really see a difference.

5:23So just kind of stuck with the long posts. So my pieces are usually around 5 ,000 characters. They fit fine in a long post. Oh, yeah, we're getting 10x more views now that we're using articles. It's kind of the same thing, but it does seem like an interesting format that maybe they will surface. I think it's a better experience as a reader. If you take advantage of the formatting, maybe, and you're doing bullet points and bolding and italics, like you can write more. Embedding posts. Yeah, embedding posts is cool and also images in the, I saw one that had like graphics and it read almost like a Stratechary piece.

5:55It had embedded images and stuff. It was cool. So strategically, this big article push, it feels like another shot at Substack, But I don't know that anyone's gonna be rushing to get off of Substack based on this because the value of Substack, the value prop is that you own your email list. So you have a direct relationship with your audience and you could take that anywhere. You could leave at any time. You can even take your Stripe account with you, which is a crazy thing. It's so creator friendly. And Substack's made that their brand and they built a great business around it. And long term, I think Substack should probably figure out how to build some sort of ads product that can run across all of the different Substacks because that would probably be a total TBPN victory.

6:42It would. It would. Very, very anti-ads. Anti-ads. But I just see, like, what does it take to get Substack into the Snapchat, Pinterest, Reddit category, sort of like Dakacorn category? And you imagine that for a lot of these, like if you're, they're clearly trying to push people into the Substack app, but even just embedding ads here and there, it could be done tastefully. I think it could be done in a way that's not too abrasive. But I think just from a business perspective, it would make sense and really open them up to some big, big dollars. You had kind of a rough theory that maybe there was some logic to doing this just for having new training data.

7:20Yeah, so there's this idea that - untitled YouTube chats is the data is worth 10 X more than one. There were a few people that were mentioning this. Talk about a company that feels very immune to vibe coding. You don't want the, if somebody comes to you and they're like, I got plat, but I'm going to charge you half as much. Yeah. It's like, well, I'll take the real one for storing and processing my customers financial information. Yeah. Noah in the chat says, I don't want to read Substack without ads. So that's advertising strongest soldier. So the question is like, you put out this million dollar bounty, how many people will show up?

7:59How many articles will actually get posted? I am seeing a lot of articles in the feed. It's also hilarious when you see like a Fortune 500 CEO post an article. He's like, oh damn, are you really down and you need a mill? Like, what's going on, buddy? And it's like, no, they're just posting news and they just need to get something out so they use an article. But it's funny to imagine that they're like, man, man, the million would change everything for me. Yeah, it's what I need, it's what I need. But - Tim Cook's third-driven article. He's just like, excellent. He's like, I need to, I can't be within the same - We're not gonna do a keynote this year for the new iPhone.

8:36We're dropping an article. And we're releasing it nine months early. The question is, is the data valuable? So a million dollars, how many people will show up? How many people will play the game? How many people will buy a lottery ticket? by writing original articles that go viral. I mean, I've been seeing so many people quoting articles, sharing them and say, you need to read this, it's incredibly well written. And then I tap in and it's just obviously slop. Doesn't seem like anyone's gonna be bailing on Substack today to move over, but there is just valuable. And then there's also the question of like, so you're probably not gonna get your own audience list on X, but will you even be able to monetize with subscriptions.

9:19X does have subscriptions, but Substack has a way more like built out. I wonder what the monthly revenue is for the person that's utilizing subscriptions the best, right? Oh yeah, that'd be interesting. Who's the highest earning subscription driven creator. But the interesting thing is that Substack offers more than just a one-time subscription monthly. You can pay annually, get a discount, participate in the live stream, get the chat. And then there's also like business plans, group plans. And so they just have a way more mature product catalog when it comes to monetizing a customer list or an email list.

9:52And then obviously you can take them elsewhere. You're like, look, my newsletter product has grown so much that I have like enterprise level needs. I have to be off Substack. You can go and build something from scratch that does exactly what you want. How did you conclude? What was your takeaway? My question was, so I don't see anyone who's a writer online who's on Substack being like, okay, now's the moment. and a million dollars is on the table, I'm shutting down my sub stack and I'm going over to X exclusively. But a lot of people cross post, that's what we do. I think a lot of people, if you have a free, if you have a sub stack that's part free, part paid, post the free part on X, why not?

10:29It's all upside, it's just more attention that can actually get you into funnel. I believe that the posting of the articles on X has been additive to our sub stack list. I think this has been net growth. I don't think people have been like, oh, I'm unsubscribing from TBPN's newsletter because I'm getting it on John's random Twitter feed that may or may not surface to me in the algorithm. So I don't think X needs writers to go all in. I just think that they want more copies of the best writing to make their way to X. And so if you have a great YouTube video, they want you to share it on X as well.

10:59So then they can surface that to people. If you have a great piece of writing, put it on X as well. And I think this does that. And I think they'll be successful there. Yeah, the question, I mean, if they're really successful and you have like YouTube for long form writing like the UGC platform for long-form writing it's probably a solid business but it's not gonna be a juggernaut like YouTube specifically because it's not gonna monetize nearly as well I want your take on whether on the actual article experience I was chatting with some folks who really really big on Twitter a little group of folks who are all well over 100k I've gotten there back in like the thread era they were like thread guys basically and And they were saying, like, I don't think Nival's How to Get Rich Without Getting Lucky thread would have worked as an article.

11:45The fact that it was all these bite-sized pieces, like, that was what was special. I think, personally, that meta has sort of died, and we haven't seen that work. What do you think? Coming back and just unironically ripping threads. Why not? Zig, everyone, and Zach. Do you miss threads? You mean just like, like, not the app. Do you miss that error at all? Can you remember any iconic threads other than Lulu's and Naval's? Naval is the main one that I remember. I probably scrolled a few threads. I actually did have a thread that actually blew up my account. So I was at 10 ,000 followers. Is this the Alaska one?

12:22No, that got me unfollows because I posted about bringing tech. I was early on the Let's Leave California train. I said, we got to go to Alaska. We're not going to Miami. We're not going to Texas. we're going to Alaska because if you have runaway global warming, it's going to be beachfront property up there, baby. It's going to be 72 and sunny every day in Alaska. Palm trees. Alternatively, if we solve global warming, we will have unlimited nuclear power. You will be able to heat, air condition, do whatever you need in your Alaska compound because you will have a one megawatt nuclear reactor dropped on your doorstep and you will have unlimited power and and you'll be lounging in a beautiful scenic mountain retreat, perfectly temperature controlled.

13:04So - And Alaska did not like that. They did not like that. I got a lot of DMs from a lot of people showing me their guns, saying, if you come up here, California boy, there will be consequences. And it was a rough day on the internet. I had to lock my account. So I closed out saying that maybe there's an article apocalypse. I'm not really seeing it. I think the algorithm is pretty well equipped to handle stuff like this. People always seem grumpy with X, generally, and honestly, Nikita specifically a lot of times. We should talk about XAI. Oh, yes. What's going on with XAI? So, Ty Morse. Yes. Did a podcast with Suleiman Khan Ghori.

13:40Yeah. Who was previously, at the time when the podcast was done. Yes. He, of course, was working at XAI. Yes. He came in. He talked about WTF is happening at XAI, predicting bottlenecks, bootstrapping off the Tesla network, how Elon deals with fires, what it's like working at XAI, Cybertruck bet with Elon, how they built Colossus, how XAI hires, experimentation, how Elon recalibrated his timeline estimates. No one tells me no. That's important. That's a wild one. Why the fuzziness between teams is an advantage, testing human emulators as employees, biggest blunders, what a meeting with Elon is like, how Elon gives feedback, figuring out what is truth for Grokipedia, what happens when Elon sees wrong grok outputs on X.

14:27Massively viral by the way, four million views just on X. I'll start by saying that this is fantastic content. Beautifully shot. Well, that very well done and great content, right? People are so curious about what's happening at XAI, right? We want to understand, you know, how they're approaching different problems. And Elon talks about it in such high level. He's talking about AI, AGI, super futuristic. And then he'll give little details on like some massive build out that he's working on. But he's not doing the nitty gritty. One small problem. Yes. Usually a$200 billion company likes to control who's going to go out and do a tell all about the company.

15:06Yes. And specifically with Elon companies, if you hadn't noticed, there's not a lot of people that work for Elon companies that go and do the podcast circuit and talk about everything going on. The funny thing is I didn't catch all the interview, but they talk about like how like HR in general is like not super formalized at the company. And so you can also imagine that like PR and the media relations team is not super active. Maybe nobody in HR said, hey, don't go do an hour long podcast. What happened? Ultimately, Suleiman. He said he's left. He left. We don't actually know what happened. But it is odd to go do an hour-long interview and then quit.

15:47Anyway, so I think good intentions potentially took the concept, which he talks about in here, no one tells me no, a little bit too far. I posted earlier, you can't even share the internal secrets of your AI lab on a cinematic podcast in this country anymore. The speculation is that this is the result of the interview he gave where he said XAI intends to run 1 million human emulations, a digital version of Optimus for online work. which will be powered by leasing compute from Tesla owners when the car is parked and charging. That's an interesting development that we actually haven't heard about before.

16:23This idea that you have a pretty advanced AI chip doing self-driving in the Teslas. They're just sitting there overnight. Why not do some inference and lease that back? Why is it a digital version of Optimus? Is it going to be sitting there in the corner like Clippy? I mean, I think realistically it's probably just like agentic rollouts. So if it's for online work, it's like, what are a bunch of tasks that you could turn Grok loose on? Everything from the classic order DoorDash in a simulation, try and buy something on Amazon in a simulation. All these things that require moving around on a screen, doing computer work, using Excel, building stuff, writing code.

17:04So a life lesson there. Let's go over to Ben Affleck. Ben Affleck. He went on Joe Rogan and gave some very good takes on AI. There's some discussion over how good of takes they are. But I ran into somebody that said that they met Ben Affleck. They were like a proper AI researcher, and they said that they met Ben Affleck, and he used compute as a noun. And they were like, yeah, that was impressive. He's talking about scaling compute. He's actually very, very deep in this. Ben Affleck is not a doomer. He's actually pretty white-pilled, and it seemed like a really good interview. Ryan says, what if he's just acting?

17:38get chat gbt or claude or gemini to write you something it's really shitty and it's shitty because by its nature it goes to the mean to the average and it's and it's not reliable and it's i mean i just can't stand to see what you writes now it's a useful tool if you're a writer and you're going ah what's the thing i'm trying to set something up or somebody sends someone a letter but it's delayed two days and gets and it can give you some examples of that i actually don't think it's very likely that it's going to be able to write anything meaningful, or in particular, that it's going to be making movies like from whole cloth, like Tilly Norwood.

18:15Like that's bullshit. I don't think that's going to happen. I think it's not, I think it actually turns out the technology is not progressing in exactly the same way they sort of presented it. And really, what it is, is going to be just like sort of visual facts. There was a line in here where, yeah, there was a line in here, I was like, that is a Carpathie line. Like he definitely heard this from Carpathie. He's listening to Dorkesh for sure. I can't. You can sue me, period. It kind of feels to me like the thing we were talking about earlier where there's a lot more fear. The whole idea of these models being mid, that's not quite true.

18:48I mean, it's hard with like to develop a reinforcement learning environment with a verifiable reward for like a great script that will break through because even the great script writers have flops that are unexpected. Literally everyone thinks it's great and then it goes out and it just is a box office bomb. Later he actually makes a stronger case I think because they're talking about Dwayne the Rock Johnson in the film The Smashing Machine and Matt Damon is recalling an interview that he did with Dwayne Johnson about you know one of the most climactic scenes in The Smashing Machine. It's very emotional and he asks Dwayne you know how did you pick this you know motion he He pulls up a sheet because he's sad.

19:31And like, what were you conjuring? What were you drawing on to create that experience as an actor and bring that emotion to the screen? Because Matt Damon's saying, like, I love that. I love that scene. I loved your performance. Where did that come from? Because you're thinking about content without ads in order to draw on the sort of like deep, sad. Yes, yes. That might be what does it for you. But I mean, Dwayne Johnson gave like a very emotional response about a family member who was diagnosed with cancer. and he was in the room when she got the news. And so her reaction, he was like channeling that.

20:03It was very emotional. And then the other one was, I think, about another family member who had substance abuse issues. And so it was just this very unique blend of human experiences that he was able to bring to bear. And I continue to think that the lore and the storytelling that happens outside of a piece of content is as important as the actual content in many cases. This was the same thing with art and the NFT boom. It's not the scarcity is important, but the storytelling behind the art piece is a lot of it. Like people like a Van Gogh because they know about him cutting off his ear and all this history and he's in all the history books and everyone just knows Van Gogh.

20:44And yes, you can go - Even with cars, right? It's a car is worth more if the previous owner was significant in some way. Yes, yes, yes. Steve McQueen is gonna sell for more. And I think that for certain storytelling elements, for certain stories, that lore of like who this person is and their likeness. And then he also makes a really good point that it's like, yes, you can, I mean you can obviously generate something that looks exactly like a Van Gogh right now. You can paint it with someone who knows exactly how to paint like Van Gogh. It's not gonna have anywhere near the value. And when he gets to the idea of just AI generating a Dwayne Johnson movie, it's like, well he'll sue you immediately and that's very easy and you've been able to Photoshop Dwayne Johnson into marketing materials for years and we've had a system for counteracting that like you have to pay licensing fees and that will happen so he has a good he has a good balance and it's and it's don't think it's very likely that it can it's gonna be able to write anything meaningful or in in particular that it's gonna be making movies like from whole cloth like Tilly Norwood like that bullshit you pause this he doesn't think that like an an iconic, amazing movie is gonna be generated by AI.

21:56And I sort of agree with that, but I think it's important to consider, what are we defining as a movie? Because we don't really watch movies anymore anyway. People watch content. And if you think about it, decades ago, people might see one movie a month or one movie a week. And that's like two to eight hours a month of screen time. Well, now people are doing two to eight hours of screen time a day, but it's these fragmented pieces. And so if AI seeps into the cracks, within the cracks, within the cracks, and the goal is not to create the experience of a two-hour movie that you show in a theater and you get everyone to come and buy tickets.

22:36Joe Rogan's really turned into Dwarkesh for A-listers. He's just come on and talk about AI, right? It's really - It's really - Yeah. He's talking about wanting this personal LLM. Oh, yeah. Where you can log the data in. Yeah, that one was a little bit rougher than this. I think it actually turns out the technology is not progressing in exactly the same way they sort of presented it. And really what it is, it's going to be a tool. Just like sort of visual facts. And yeah, it needs to have language around it. You need to protect your name and likeness. You can do that. You can watermark it. Those laws already exist.

23:09You can't, I can't sell your fucking picture for money. I can't. You can sue me, period. I might have the ability to draw you, to make you in a very realistic way, but that's already against the law. And the unions are going to, the guilds are going to manage this where it's like, okay, look, if this is a tool that actually helps us, for example, we don't have to go to the North Pole, right? We can shoot the scene here in our parkas and, you know, whatever it is, but then make it appear very realistically as if we're in the North Pole. So it'll save us a lot of money, a lot of time. We're going to focus on the performances and not be freezing our ass off out there and running back inside.

23:45Small chance that AI saves Hollywood as a place, right? You're getting clapsed. As a place. Yeah. And the reason for that is one of the issues right now is filming is so expensive in L.A. There's so much red tape that people have to go to Atlanta, Canada, Europe, whatever, international. but all the talent, the writers, directors, producers, the platforms, they're still here. It's very depressing. It's a bad vibe. All the restaurants are shutting down. Everything's cooked. But if you can start to generate the content here, just on your computer, there's a chance that you see a kind of resurgence in Hollywood, the place, right?

24:25And I think that realistically, that is the strategy that the industry should take because having this concentration here, there's already so much talent. Keeping it here and actually embracing AI is much more likely to revitalize the city. Ben Affleck on AI Software. I suspect many AI native software companies are misrepresenting their growth and quality of revenue. Use of credits, annualizing proof of concept revenue, and lack of annual contracts lowers the overall quality of revenue and ultimately the multiple. I mean, he's not that, I mean, this is a joke. I didn't actually say that, but is not that far off from what this is a great template i hope this meme runs way further because it's so good early ai the line went up very steeply and it's now sort of leveling off i think it's because and yes it'll get better but it's going to be really expensive to get better and a lot of people are like this we want you at gpd4 because it turned out like the vast majority of people who use ai are using it to like as like companion bots to chat with at night so there's no work, there's no productivity, there's no value to it.

25:30I would argue there's also not a lot of social value to getting people to focus on an AI friend who's telling you that you're great and listening to everything you say. Let's give Tyler a rebuttal. I think most of these takes are pretty outdated. Writing is actually AI writing. You can still tell that it's AI, but it's good. Okay, okay. Sell a screenplay then. Sell a screenplay. Yeah. Shop a screenplay. Do it. Do it. Just prompt it. But why would I sell a screenplay? I could just make the whole movie myself. OK. Just make the movie. He has this idea. Make the finished movie, call on Netflix, and get a million dollar advance for your movie.

26:08Instead of filming in Antarctica, you can just get in the parkas and film it here, right? But you don't need the parkas either, because you can just have them be in normal clothes. That's true. And then it's like, well, you don't actually need the people, right? Because you can just maybe fine tune a model so you get consistent characters. Sure, sure, sure. And then. Yeah, I will say just in the last two weeks, I am getting served AI content, that's pretty good. Pretty good. There's some pretty good stuff out there. I would push back a little bit on Ben saying, you're not just gonna be able to generate a film.