Meta AI Vibes & ChatGPT Pulse Reactions, Friend’s Billboard Blitz | Roon

Summary

TBPN Podcast Summary

Episode: Meta AI Vibes & ChatGPT Pulse Reactions, Friend’s Billboard Blitz | Roon
Date: September 26, 2025
Hosts: John Coogan & Jordi Hays
Guest: Rune (AI thought leader, anonymous X poster)

Episode Overview

In this episode, John and Jordi dive deep into seismic product launches in the AI consumer space: Meta AI's Vibes and OpenAI's ChatGPT Pulse. They dissect the initial product demos, the industry's and public’s reaction, and what these launches mean for AI’s future. The episode also features discussion on virality and controversy of Friend.com’s massive billboard ad campaign, the evolving application layer of AI, Pulse's possible disruption of newsfeeds and social media, Vibes' reception, AI’s effects on productivity and jobs, the open-source AI movement, and Rune's insights about AI’s present and future. The hosts keep their commentary candid and reflective of AI and tech culture.

---

Key Topics and Insights

1. Meta AI Vibes & ChatGPT Pulse – Launches and First Impressions

The Launches

• Meta AI Vibes: A new feed in Meta AI app for short-form, AI-generated videos, leveraging a partnership with Midjourney.
• ChatGPT Pulse: OpenAI’s highly personalized news/content feed for ChatGPT, populated based on a user’s prior interactions and interests.

Product Reception and Market Dynamic

• Vibes was immediately polarizing—industry response on X was overwhelmingly negative, especially among technologists. Pulse received a far warmer reception.
• Timing: Both products launched within minutes of each other—fueling speculation about intentional competition between Meta and OpenAI.
  • “1226 Pacific. Alexander Wang launches Vibes. 1236. Sam Altman launches Pulse.” (B, 46:25)

Vibes Critique

• Host: “The reaction was almost universally negative... On X, yes. The question is, will children like it? Will the elderly like it? Moms like it?” (B, 02:46)
• The launch video showcased prototypical 'AI art’ style (astronaut on a bike, bears boxing, cat at news desk)—perceived as slop or low-value, generic AI-generated content.
• Sham Sankar quote: “Read the wall of quote tweets that categorically reject this AI slop. We are so back. We are going to win.” (B, 05:29)
• It’s not trying to hide its AI-ness—aesthetics are unapologetically “AI.” It feels like a “bull market in slop.” (A, 11:49)

Pulse Praise

• Pulse, in contrast, is deeply personalized based on user queries, and aimed at knowledge, not just entertainment. Only available for paying users.
• John: "My Pulse experience was wild... a breakdown of Mag 7 capex, deep dive on Nvidia’s Blackwell rollout cadence… all things I cared about." (A, 47:16)
• “It's like a newsfeed tailored to recent conversations. It reminded me of Perplexity's Discover feature, only way more personal.” (Simon Smith, 56:01)

Key Quote

• On Vibes: “Meta created a product that is giving slop. It’s a bull market in slop again.” (B, 11:49)
• On Pulse: “Such a great product and surface area to start building on. I'm extremely bullish on Pulse for OpenAI.” (A, 52:24)

---

2. Why Did Vibes Flop (So Far)? The Deeper Analysis

Launch Strategy Criticism

• Aggressive, public launch exposed Meta to a universal dunking from “Tech Twitter/X”, rather than rolling out more subtly through Instagram/Facebook's broad user base.
  • “This was like a very unforced, like, PR error... you can't tell somebody... you truly want and will build personal superintelligence. But then the first thing you launch is trough slop.” (B, 06:18)
• The feed is populated with curated, high-quality assets, but user-generated videos do not match that editorial quality.
  • “When you go to try to make a video the output is not at all on par with what you're seeing in the feed.” (B, 22:19)
• Not personal enough: Unlike viral AI moments (e.g., Studio Ghibli filter), Vibes doesn't tap into user’s own photos, content, or identity.

AI Aesthetic Fatigue

• “Now the reaction is that looks like AI. And people just don't think it’s cool.” (B, 05:17)
• “It does feel like it’s very much white-label Midjourney...but VO3 in Gemini gives way better results and an entirely different workflow.” (A, 24:53)

Product Iteration Hopes

• Hosts note that with better personalization (e.g., pulling from Instagram data), Vibes could become more viral—like Studio Ghibli AI moment.
  • “Go and preload a generative image for everyone's Instagram profile picture... That could create a viral moment where people are sharing... their own personality in an illustration.” (A, 22:10)

---

3. Social, Economic, and Ethical Implications

“Slop” Content and the Infinite Feed

• Referencing viral X post: “We at Meta are delighted to announce we've created the Infinite Slot Machine that Destroys Children…” (B, 38:37)
• Concerns about the addictive nature and low-value of infinite, AI-generated feeds. Is this a social/cultural risk?
  • “I don't know how much it will destroy children...Most parents are aware of technology and that it needs to be measured.” (A, 38:52)
• Children’s media consumption: “Average American five-year-old watches two to three hours of TV a day. That's a lot. That's really dark.” (B & A, 44:18)
• Counterpoint: Screen time isn’t monolithic—some content teaches, some is ‘candy.’

Game Theory & Adaptive Behavior

• Parental and peer filtering will create a “barbell” where some kids are protected and others aren’t, making “reading a book” alpha in a land of zombies. (A, 43:03)

---

4. The Application Layer Race: Personalization, Utility, and Monetization

Pulse – New Newsfeed/Discovery Paradigm

• Personalized feed, leverages ChatGPT interactions for deep curation, knowledge, and product recommendations.
• Monetization: Hosts anticipate Pulse will become a lucrative ad surface, akin to the Facebook News Feed or Google Search results.
  • “They’re going to be able to put ads in here. It’s going to be a hugely monetizable surface...” (A, 51:46)
  • “The reason this is potentially great as an ad unit is...it’s not putting it in a specific prompt, it’s putting it in between prompts...they can personalize the ad for you.” (B, 52:09)

The “AI Slop” Label – Not Just About Entertainment

• Hosts clarify, Pulse may look like “AI slop” but is actually useful—users come away feeling smarter, not dumber.
  • “I came away from that feeling like I had leveled myself up and...become a better person, more thoughtful, more knowledgeable.” (A, 48:10)

---

5. Talent Wars, Economic Boom, and Open Source

Talent and Economic Power

• “Most of the growth in the equity markets is just AI. For the past three years, there’s really nothing else.” (Rune, 113:19)
• Barnacle economy: “The entire global economy is sort of a barnacle [on OpenAI]” (B, 113:49)
• Insane compensation packages to engineers as 'distributed acquihires'. “It feels like an Aqua Hire…distributed Aqua hire.” (C, 147:51)

Open Source AI – Stated vs Revealed Preference

• Many demand open-source LLMs, but few actually use them versus easy, high-powered commercial APIs.
  • “People demanded it but didn't actually want it.” (C, 141:10)

---

6. Friend.com’s Billboard Blitz: Virality, Critique, and Hardware Productization

Ad Campaign Analysis

• Massive, controversial OOH campaign (“largest NYC subway campaign ever,” A, 96:19), minimalist ads, positioning the device as ‘your new roommate.’
• Criticism: Orders extremely delayed, negative reviews, possible risk of overhype before product readiness.
  • “Getting a whole new wave of customers that are then going to be waiting a long time for their device...feels incredibly risky.” (B, 99:14)
• “There's a lot of people that are going to be taking pictures just saying, look how dystopian this is, and posting it.” (B, 96:54)

---

7. Special Guest: Rune on Codex, Pulse, Consumer AI

Codex Mania!

• Rune: “Codex...and Claude code are like genuinely amazing...it’s a completely different way of life than anything I was doing before. It alone completely upends the software industry.” (C, 114:49; C, 114:57)
• Predicts step change in entry-level developer productivity; non-engineers are now shipping apps.
• Not sure about macro job implication – highly nonlinear, but “productivity of an entry level engineer is way higher than...five years ago.” (C, 116:40)

Pulse, Advice, and Human Interactions with AI

• Rune uses advanced models for information, not so much for interpersonal advice or therapy, but acknowledges that many do.
• Observes shift: AI outputs are becoming less conversational/more “essay” and less varied, but expects future will allow stylistic and tonal customization.

---

8. Other Notable Tech/AI News & Industry Discourse

• Meta AR/VR headset rumors: speculation on lighter, ultra-wide field-of-view headsets (A, 36:32)
• TikTok rumored $14B US deal & potential user/inflation metrics issue (B, 61:22)
• Adobe’s AI revenue push and perceived innovation lag (A & B, 87:12–91:09)
• Intel’s rebound with government and industry backing (A & B, 155:29–159:22)
• Viral “slop”/torment meme discourse: “We have created the Infinite Slot Machine that Destroys Children…” (B, 38:37)
• The infinite Jest/AI content-loop analogy (A, 73:39)
• Podcast’s own rebrand (from Technology Brothers to TBPN) after journalists labeled them “tech bros” (C & B, 130:26–131:25)
• Mark Zuckerberg to be played by Jeremy Strong in Aaron Sorkin's The Social Reckoning (A & B, 77:03–81:42)

---

Select Memorable Quotes and Timestamps

• On Meta's blunder:
  “Meta created a product that is giving slop. It's a bull market in slop again.” (B, 11:49)
• Pulse vs Vibes:
  “Pulse was like... learn what you're the biggest nerd about, and then surface really rich deep content. ...My results were wild!” (A, 45:29)
• Infinite slot machine concern:
  “We are delighted to announce we've created the Infinite Slot Machine that Destroys Children…” (B, 38:37)
• AI writing style change:
  “It's not that there's anything technically wrong with the current output, it's that reading the GPT-3 output was just much more enjoyable.” (B, 86:02)
• Economic impact of AI:
  “Most of the growth in the equity markets is just AI, you know, for the past like three years, there’s like really nothing else.” (Rune, 113:19)
• AI tool user agency:
  “Chat is breeding agency into kids.” (Dylan Patel, 66:52)
• On codex and transformative AI engineering:
  “Codex ... is like, I don’t understand how I did software engineering before this... It’s a completely different way of life than anything I was doing before...” (Rune, 114:57)

---

Timestamped Highlighted Segments

00:50–03:40 – Introduction, Vibes/Pulse launches, setting premise of the episode
05:17–12:45 – Vibes reactions, slop discourse, industry context
17:23–22:10 – Studio Ghibli filter positivity vs. Vibes approach
29:23–31:41 – Ben Thompson’s contrarian take on Vibes; differentiating “fake world”
34:04–40:00 – Friend.com billboard campaign analysis; societal/psychological effect of infinite feeds
45:29–53:36 – Detailed ChatGPT Pulse reactions, monetization angle
113:01–130:00 – Guest segment: Rune’s perspective on Codex, Pulse, consumer AI, coding productivity
147:51–149:24 – The new talent wars & founder-level compensation insight
143:26–142:34 – Open source AI said vs. revealed preferences
77:03–81:42 – “The Social Reckoning” movie news
87:01–91:09 – Adobe and the “AI revenue” wave

---

Conclusion – Original Tone & Takeaways

The episode, true to the unapologetically nerdy and self-aware TBPN style, unpacks major contemporary launches (Vibes, Pulse) with skepticism, humor, technical depth, and industry wisdom. Hosts champion authentic personalization, criticize shallow product launches and hype cycles, and highlight the evolving interface between AI, consumer apps, and economic systems. With Rune’s sharp commentary, the discussion remains rooted in the lived experience of AI practitioners and power users, signaling both AI’s dazzling promise and the risks of “slop” eating our collective brains.

Memorable sign-off:
John: “If you bring your own creativity you can probably have fun on [Vibes]. But be careful because you might just wind up in infinite jokes world.” (A, 112:12)

Transcript

[A]

You're watching TBPN. Today is Friday, September 26, 2025. We are live from the TBPN Ultradome. The temple of technology, the fortress of finance. The capital de capital. El capital de capital. Massive news in the AI world. In the product layer, in the application layer. Two apps launched yesterday. Really two features. One in the Meta AI app called Vibes and one in the the ChatGPT app called Pulse. Two very different apps. Neither of them are going to save you time and money. The only one that can save you time Money is ramp. Ramp.com Time is money save. Both easy to use. Corporate cards, bill payments, accounting and a whole lot more all in one place. Good morning to everyone in the chat. We had John actually live in the TVP and Ultradome yesterday. It was great seeing him stop by.

[B]

It's great to see you, John.

[A]

We got a couple app demos, caught up with a few folks, did a little photo shoot yesterday. Had a great time.

[B]

We did.

[A]

And of course we are live on Restream. Restream, iO1 livestream, 30 plus destinations, multistream. Reach your audience wherever they are.

[B]

Wherever they are.

[A]

We also get really quickly. Just to say thanks to everyone before we go into our top stories, Grepetile sent a book called Fault Lines about the history of bugs. Because of course Grepetile helps you find bugs. And I thought this was interesting. They tell the story of the first bug at the Harvard computation lab in 1947. Said one relay stalled. A technician checked the console, traced the path and opened the door to Relay 70 on Panel F. Inside, a moth was jammed in the switch. They removed it, taped it into the lab log and captioned it. First actual case of bug being found. The log made the term bug literal and. And cemented the practice. Find the cause, write it down, keep the knowledge.

[B]

Very cool. Let's get into the news.

[A]

Yes.

[B]

And we have, I guess we should highlight. We're gonna have none other than rune on at 1pm today. So that'll be in a couple hours. Can't wait for that. In the meantime, a lot of stuff to talk through. The timeline has been in turmoil. It has interesting that Pulse was launched the same day as Vibes. They got very different.

[A]

Time's up to steamroll big, big technology companies.

[B]

Well, I think. Yeah, I have to wonder. I mean ChatGPT has the benefit of like they're launching Pulse in the core app. People are just going to discover it and use it if they're paid and they can roll it out that way. Vibes. I thought it was Interesting that they even launched this so aggressively into the timeline.

[A]

Yep.

[B]

Because the reaction was almost universally negative.

[A]

On X.

[B]

On X, yes. The question is, will children like it? Will the elderly like it?

[A]

Moms like it?

[B]

Yeah, but we should, we should get into it.

[A]

Well, let's. Let's start by playing the actual video for everyone. So if you haven't seen Alex Wang, who we had on the show last week at Meta Connect, posted the initial launch, I think he launched the entire product, he said. Excited to share Vibes, a new feed in the Meta AI app for short form AI generated videos. This is on the back of the Mid Journey partnership. We can go way deeper into what's going on here in the reaction. But I mean, just the numbers on this is crazy. 2.8 thousand likes, 2.4 thousand quote tweets or retweets and 1.6 thousand comments.

[B]

Insane ratio.

[A]

Insane ratio. But this, this is, of course, you know, there's a lot of nuance here that we're going to go into, but let's kick it off by watching the actual one minute trailer for Vibes, the new app from Meta. It's in the Meta AI app. You actually have to reinstall the app or go and update the app. Are you channeling the Vibes? Are you feeling the Vibes? The create button. So when this launched, I went to the Meta AI app and I. And I typed in, okay, I met AI. I'm ready to use the new Vibes feature. How do you do it? And it was very confused and it was saying like, I'm not aware of that. I don't know anything about that. It's because the older app was not aware of the new app existing. And so I had to go to the App Store and actually refresh. But once I did, I got in and I saw a lot of this type of content. An astronaut on a bike, two bears boxing, a cat talking at a news desk, women running in clouds. All very AI aesthetic. It has a very unique look to it that I would just describe as like the AI art look these days. It's not trying to not be AI. Yeah. It feels like it's not trying to be anything but AI. In the same way that a lot of, a lot of CGI and 3D renders have a certain aesthetic to them.

[B]

It's an interesting thing where if you showed one of these assets to somebody 10 years ago, they would have thought, that looks really cool.

[A]

Yeah.

[B]

Now the reaction is that looks like AI.

[A]

Yep.

[B]

And people just don't think it's cool.

[A]

Yeah.

[B]

Like, I. You know, personally, I think there's a market for this and obviously this is the worst that the model. The worst the models will ever be.

[A]

Yeah.

[B]

So the images will get better. But I think the reaction. Sham Sankar, he quoted it, he said, read the wall of quote tweets that categorically reject this AI slop. We are so back. We are going to win. And he hits it with a Salute.

[A]

American flag, which is crazy to be kicking it off.

[B]

Crazy to dunk. Crazy. You don't see Sham dunking.

[A]

No, I mean, they're also partnered. I'm pretty sure Palantir is model agnostic, uses Llama, which is. Alex is working defense Llama. But, you know, like, when people are not down with a thing, like, you know, they're gonna speak their mind.

[B]

Yeah. I mean, I think. I think the. It's interesting because X is such a microcosm. Right. It's not a reflection of the average Internet user.

[A]

Totally.

[B]

It's certainly not a reflection of the average meta platform's user. I think that there's another world where they just started promoting this in Instagram and Facebook and just started driving users over kind of quietly. And I think this was like a very unforced, like, PR error to come. Because to go in the line, you can't tell somebody. You basically, you know, I believe that they want. They truly want and will build their version of personal superintelligence. But when you've just been hitting people over the head with messaging about personal superintelligence, and then the first thing you launch from MSL with your new massive partnership with Mid Journey, is a trough slop. That's. That's rough. Right? Yeah. It sends a signal to the market. And I think this just makes their job harder from a recruiting standpoint. Right. Do you want to work at Anthropic, which believes in machine God, which has like a cult. Right. Their team, they didn't lose a lot of people in the talent wars. Right. They even lost some people for to cursor and got them back a week later.

[A]

Yeah. Metro was commenting on this. He said, I would not be surprised if this tweet alone is going to make meta super intelligence is raise their signing bonuses by $1 million.

[B]

Yeah. Or even more. Right.

[A]

Because.

[B]

Because is even, you know, even when you. When you look at your options, it's like you can go live in Elon World and work on Xai.

[A]

Yeah.

[B]

And be super hardcore and sleep in the office and get that sort of tap into that Elon.

[A]

I Mean, there's, there's a fair amount of slop going on.

[B]

Sure, there's, there's slop. But it, but at least it's like canyon stuff. At least you're buying. You're riding with Elon, who is.

[A]

You have the cool narrative around like building Colossus too really fast and going after like.

[B]

Yeah. And then you look at their benchmarks.

[A]

And all that stuff.

[B]

Elon posts slop. But when you look at their, like when they talk about benchmarks, what benchmarks are they really going after? It's like, really?

[A]

Yeah, yeah, yeah, yeah. Elon does a good job of like, he'll be retweeting a lot of, quite frankly, like SWAS lob videos. Right. But then he will refocus the conversation back on. We're going to solve physics or like we're going to, we're going to discover new physics, we're going to do really hard math and you know, let's give Meta a couple days to see if they can kind of steer the conversation back towards something that gets technologists excited, because they could. It's possible that they come out with something for sure.

[B]

That it's just when you have the entire world waiting and honestly kind of preying on their downfall. Right?

[A]

Yeah.

[B]

Like the world is not rooting the world broadly, the industry broadly. When you look at all these different camps, OpenAI lost a bunch of top talent, MSL. They're not sitting there excited for Meta to do something amazing, which I'm sure they will at different points. And so your public opinion is already against you. And then you come out and launch a product that is dedicated to the thing that people like the least about AI.

[A]

Yeah, I have a rebuttal. But first I'm going to tell you about Privy Wallet infrastructure for every bank. Privy makes it easy to build on crypto rail, securely spin up white label wallets, sign transactions and integrate on chain infrastructure all through one simple API, the.

[B]

Official wallet infrastructure provider of the TVPN ultradem.

[A]

That's what you call a layup. Boom, boom. Slams it down.

[B]

So yeah, so basically they just set up this. They set up the perfect product, the perfect video to just allow the entire Internet to do the craziest alley oops.

[A]

Okay, so first off, I totally get the backlash is undeniable. That's just a fact. The timeline hates it in terms of like, if Meta's on the offense today and the timeline's on the defense, like the defense slammed it down, not a single point got through. But what's weird about this is that David Hulls, the founder of Midjourney, who created this model, who created this aesthetic, is loved in tech and is one of the, I think the most, one of the most respected founders. He hasn't raised a dime of venture capital. He built this incredible model.

[B]

A bit of money in the terms.

[A]

Of revenue, but in general, Midjourney has been the one video generator, image generator that has been sort of celebrated as, as the least sloppy, I guess, and the most opinionated. And so I feel like maybe the Meta team just thought that by partnering with Mid Journey they would beat the slop allegations. But it is weird that we're not seeing any of the goodwill that David Holes has transfer over to msl, which, which I don't know how to unpack. So David, pitch me, I mean back on that.

[B]

David is not. They had an announcement.

[A]

Yep.

[B]

But he's not a, you know, he didn't post yesterday. Like he came out and was like, hey guys, I'm really excited for our new checkout Meta AI Vibes.

[A]

Yeah.

[B]

Also the name, the name is a little bit stale. It's just the most overused word of the year.

[A]

Yeah, yeah, yeah. I mean didn't you call the top on Slop like a year ago or something like that? I remember you saying like.

[B]

No, it was earlier this year. I just, I think you said like.

[A]

The term's overused at this point.

[B]

Yeah, it's just wide, you know, overly used.

[A]

Yeah.

[B]

And Meta created a product that is giving slop. It's a bull market in slop. Again.

[A]

Yes.

[B]

Because. Okay, so, but, but again, and I want to be clear that I'm not like putting like, I don't judge MSL based on this product, but I, I judge the. I think it was a point core marketing strategy for AI efforts at Meta Broadly when in a few days they're launching a product that has a heads up display with AI built into it. It's like get people to focus on that.

[A]

Yep.

[B]

Right.

[A]

Maybe. Yeah.

[B]

Like that, that's the functionality that's actually really cool. We've used it. You can walk around, you can be like, hey Meta, looking at my fridge right now, what should I make for dinner? And it'll tell you and it'll give you recipes and it'll show it on a heads up dis. And it's awesome.

[A]

Yeah.

[B]

But this was just a totally unforced error.

[A]

There is a little like, it felt.

[B]

Like, it felt like the. It really. To me I was seeing parallels between, you know, that iconic image of Mark Zuckerberg like in the metaverse. Metaverse. When everybody saw it and they were like, oh, it's over. Yeah. And it was like, oh, it's over. Like, the metaverse is over.

[A]

Yeah. Interesting. So I have a few things I want to peel back on that. So first, Cognition is the maker of Devin. Devin is the AI software engineer. Crush your backlog with your personal AI engineering team. Go sign up, please.

[B]

Also the makers of Windsurf.

[A]

Also the makers of Windsurf. Fantastic.

[B]

They're absolutely cooking it.

[A]

And the hero.

[B]

Fantastic transition. John tried to interrupt my perfectly. Interrupted my rant with an ad.

[A]

Yes.

[B]

Great sign of respect.

[A]

So I want you to shift back to the Studio Ghibli moment images and chatgpt and try and relive that, because I feel like that came out and it was AI generated content. It had a very AI generated aesthetic. There's a world where that looked really sloppy, but it didn't. What was your experience?

[B]

It was AI perfectly emulating, at scale, one of the most loved esthetics on the Internet ever.

[A]

And wasn't.

[B]

That was the beauty of it. It was like anybody could make something that wasn't slop.

[A]

Yeah.

[B]

It was like, you know, time from somebody having the idea for the prompt to a beautiful image. And sure, it got overused. Sure, you don't see them a ton anymore.

[A]

Yeah.

[B]

But that was. I remember the. The. The little joy of turning an image of, like, me and my fam, you know, with my family, into a Ghibli picture. And that was fun.

[A]

It was a super fun day on the Internet. We had a ton of fun with it. I made multiple. I had a couple that went super viral.

[B]

Everybody made like 50.

[A]

Everyone made a ton. And then eventually it got burned out.

[B]

Like any, like, any aesthetic or meta, it just burns out on the Internet.

[A]

Yes. And so, like, launch videos, there were. There were a few hot takes on the timeline that were like, I don't like this. Then. Then there were also the. There was also the backlash to, like, you shouldn't be stealing Studio Ghibli's aesthetic. You're.

[B]

Yeah, if I was, I was Studio Ghibli. I don't know if, like, what's your. What's your verdict? Do you think that was good overall net positive for Studio Ghibli, or do you think it. It actually.

[A]

Well, I ran the. I ran a sort of test where I took a image from Oppenheimer and I Studio Ghibli'd it of Einstein talking to Robert Oppenheimer. It's this Iconic scene. And I, Ghibli did, and I put that up and that got like a thousand likes. Because everything that was decent, every decent meme that got Ghibli that day got a thousand likes. And then I took a real frame from Spirited Away, the most iconic Studio Ghibli film. Like a real frame, not an AI generated one. And I put that up 10 likes. And it was very funny to be like, people don't actually want Studio Ghibli. They want Studio Ghibli filter applied to something else iconic. Because it allows them to do this mental puzzle of, oh, I'm seeing something that's iconic. It's the bell curve meme. Or it's the always two astronauts in space shooting at each other. Always has been Ghiblied. And it's like, oh, I get the joke.

[B]

What was magical about it is you took a style of illustration and animate, you know, illustration.

[A]

Yes.

[B]

That historically was extremely expensive and time consuming and made it instant. And it was basically as good. Right. I'm sure. I'm sure a true Studio Ghibli enthusiast would be like, I'm gonna pull up the two pictures and look like they didn't get this part of the handwrite or whatever. But it was beautiful. It was magical.

[A]

Yes.

[B]

And that was. That was undeniable.

[A]

Yes.

[B]

Glad we're not seeing them.

[A]

If you're trying to make something beautiful and undeniable, go to figma.com, think bigger, build faster. Figma helps design and development teams build great products together. But my point was that with the Studio Ghibli aesthetic, it was actually personal superintelligence. It was personal. And what made that go so viral in a way that I think everyone liked, was that they could Studio Ghibli themselves. They could Studio Ghibli. It acted as this interesting filter where I saw, remember the very first one that went viral, it was a man with his wife and he was like, life hack. Take a photo with you and your loved one and Studio Ghibli it and send it to them. They will be overjoyed.

[B]

And I remember the first 12 hours, people were like, I don't know, how's everybody making these?

[A]

Yeah.

[B]

It was still like, there was a narrow window that.

[A]

Yes. But there's something where if you empower the individual to create something that's personal to them, then they are more inspired, more defensive of it, more happy with it. There's something there that's. That, that's more satisfying, even if it is like, it's not slop because there's still that element of uniqueness in it.

[B]

And this gets into the huge problem.

[A]

With Meta AI Vibes, is that it's not. It doesn't start with your Instagram profile. And I was pitching this, I was pitching this, I was saying Instagram, even.

[B]

Even, even more specifically, you were showing some of the results you got this morning. And to anybody that hasn't downloaded the app yet, you downloaded it. And it's a bunch of high quality assets for what they are, right? Well, you can say it's the Mid Journey Explore page. Yeah.

[A]

Like literally.

[B]

So it's like elite Mid Journey, like results, Right. Animated and it's solid, but it's more like just a collection of assets. It's not a. They've basically seeded the feed with a bunch of super curated assets and everybody, I think we had a few people on the team looking at it and everybody's getting the exact same results. Yes, a mirror image.

[A]

If you open the Meta AI app, I can guarantee you that within 10 scrolls you will see a monkey on a jet ski.

[B]

Let's give it up for Monkey.

[A]

The monkey on a Jet Ski. It's a great video, honestly.

[B]

Hit that horn, Hit that horn.

[A]

Hey, every once in a while you got to slop it up.

[B]

You got to put the monkey on there.

[A]

You can sit here and critique it, but good slop is tasty. And so the monkey on the jet ski is great. There's a cat in the newscast. There are a few feeds. They very clearly went and picked the best AI videos from midjourney and whatever, they'd created and curated a starter feed before you get into the random algo feed to kind of seed your algorithm. And so you automatically see the same thing as you scroll on your feed or my feed. It's all the same right now. Eventually it'll be highly user generated and then eventually it'll be highly algorithm driven such that you could just wind up seeing. I guarantee this is where this is going. So I actually love Those mid journey 1980s videos that have been on Instagram where someone takes the aesthetic of like, imagine you're a Swiss banker in 1980 in Switzerland, and it's like Cherry Sherry lady is the song that plays. And it's just a whole bunch of midjourney images back to back to back. I love those videos. And that's what my feed will be once the Meta AI feed tailors to me. Right now it's just general AI video. And I feel like when I see generic AI video, it's just a movie that I've seen. It's a very cool movie once, but I only want to consume 10 minutes of it and then I'm like, I'm good. I don't really want to revisit it. I want to go see something new. So I don't think this at this stage, this level of content will be that sticky. It'll be novel to scroll for a lot of people if they haven't seen it, but eventually it will become more tailored and you will see stuff that you actually do like over time. And I think it'll be a long time until they really get the long tail of people. I think live players will basically stay out of this forever. I think that a lot of people will get kind of sucked in, but they're already getting sucked in on reels and TikTok and YouTube shorts and stuff anyway. Vanta Automate Compliance Manage risk, Improve trust Continuously. Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation. It sucks whether you're pursuing your first framework or managing a complex program. And so my pitch to Adam Mosseri and the Instagram team was make AI video personal. Make super intelligence personal. Go and preload a generative image for everyone's Instagram profile picture. And just like when you upload any photo you can filter it, go and pre render that and give everyone a Ghibli or give everyone a mid journey vision of what they would be like based on their most liked photo or their most recent photo that they've shared and then just have that kind of pre rendered ready to go and then they can share that and you can create this like viral moment where people are sharing like the studio Ghibli moment their own personality brought through in an illustration instead of just here's a bunch of random.

[B]

Yeah, and I would say I guess everyone likes maths. You know, I think obviously we love the meta team.

[A]

Yeah.

[B]

And we're excited about what MSL is going to create broadly. But it's hard for me not to just give this a scathing review specifically because it's not just a bunch of user generated content. I mean it was generated by users out there, but it's effectively like they created a folder that people can view in a feed and it's cool videos but then when you go to try to make a video the output is not at all on par with what you're seeing in the feed.

[A]

Prompting midjourney is hard, like getting good results out of AI image generators is hard.

[B]

But take mid ideas for sure.

[A]

But understand what keywords work the srefs in mid journey. And then the animation step is really hard, too, because oftentimes you're not really like you. You can kind of describe a scene, and then you need to think about, like, okay, what do I actually need to tell it? Do I need to tell it that the background. That there should be actors in the background that are moving?

[B]

Midjourney's magic. But it has a steep learning curve.

[A]

It does sort of have meta AI.

[B]

They're making it. They're. They're. They're selling this. They're selling you slop dreams.

[A]

Okay.

[B]

And then they let you create assets, but they don't. They're not. And that. That's just a terrible. It's like. It's like running an advertising campaign of, like, hey, you know, they're. They're advertising what. What would appear to be the capabilities of meta AI because it's in the meta AI app.

[A]

Yes.

[B]

But that when you as a user go to create stuff, it's just not good. Like, you were just trying to generate a TVPN intro.

[A]

Yeah.

[B]

And the letters were like. Like, you put TVPN in a blender. They were inverted. It didn't make any sense. Like.

[A]

Yeah.

[B]

And character consistency and fonts and all these things are hard to generate. But it just didn't feel like the product was. If you were trying to create an organic moment where people were like, these videos are cool. Now I can create cool videos myself. And then people are gonna have this. A bunch of people are gonna have the experience. So.

[A]

Yeah.

[B]

Again, the two. The main critiques are like, product level just not great yet. It will get better.

[A]

Yeah.

[B]

And then from a marketing standpoint, like, total cell phone, to launch this into the. To the audience, that is going to have the most adverse reaction of anybody. Because we're sick of seeing Elon, you know, share these videos of, like, random, you know, ani stuff.

[A]

Yeah. It does feel like it's very much white labeled mid Journey. Like, even when you go to create, you type a prompt, and it gives you four images, which is exactly what happens on midjourney. You get four, and you get to pick, and then you have to animate it, and you're animating it from a single frame. That's not the workflow for VO3 in Gemini. In Gemini, you go in and you describe a video with a whole bunch of different steps. And the results that I was getting from VO3 were so far and above this midjourney has a cool aesthetic and it has kind of a cool. A cool. Like, it's like an animated painting almost. I don't know. It's like. It's a different thing than VO3. Like, the camera was moving around as it tracks the car that breaks through the Hollywood sign. And then the spins. Remember the champagne bottles bursting out? We were like, it fit that in there. And I think what's happening in VO3 is that when you go to it with a prompt, it has. It kind of has, like a reasoning step where it hydrates out what you're saying and gives it way more context and transforms it. A lot of times when I'm writing VO3 prompts, I'll actually go to ChatGPT and describe and talk for a long time and say, write a big, long V prompt that actually organizes all of this into what's happening in the background, what's the lighting like? What time of day is it? Where are we? Like, what's the setting? What's the camera? What's the focal length like? All of these things are things that you need to put in midjourney to get good results. And it's hard to get people up to speed on that. I do wonder about the social component here. One of the secrets to Midjourney's success has been the fact that David Holes was talking on, actually to Ben Thompson, and he was saying that. Oh, wait, no, it wasn't. It wasn't David Holes. It was. I think it was Nat Friedman, who's now at msl, was talking about why Midjourney was successful and how Midjourney works. And Midjourney famously did not have a website, did not have an app. They just had a Discord server. And in the Discord server, people would go and prompt and they'd get a good result. Yes, but the key. Yeah, the key was the sharing. Because when you give someone just a blank prompt box, they just type, dog. I want to see a dog. And then you get one generic image of a dog, and it looks like a stock photo. And what you really want is a photo of three puppies in the snow in Christmas, shot on a Kodak phone. Kodak camera. With this camera move, do you want the directorial style? You want something that's a different vibe? There's so many different pieces and tools that you would use to put together something that actually looks great. And the Midjourney chat and the Discord allowed people to see, oh, wait, this person's prompting in this way. And they're getting this Result, I should prompt like that. And they all built off of each other. And I think that's something that can actually happen in an app like Vibes. I wouldn't be surprised. You can already see it with the prompt is described there and then you can click remix and you can use that as a jumping off point. So just in terms of like the user experience of creating little pockets of creativity and solving that problem, of like, how do you get a good result? How do you get a good prompt? It doesn't seem that hard. I don't know. I feel like I'm not that bearish on this product. Like, I know people are rejecting it, but it feels like this is a useful step in terms of like they have a midjourney partnership. They're basically white labeling midjourney into this app. A lot of their testing, gathering data, and then they're gonna put this in Instagram.

[B]

Yeah, yeah, yeah. It's a proving ground.

[A]

Yeah. And they, but they do need to, they do need to start.

[B]

But it felt zero. Felt like it felt like an alpha product, not a beta product.

[A]

Yeah, there's something, there's something there. I don't know. Ben Thompson was wildly positive on this. He was the only one. He did an emergency podcast this morning or maybe last night, and he had an interesting take that said that basically this technology, this AI feed, it solves a few problems. First, he liked the idea that we've all been suffering with this low grade annoyance of is that real or is that AI? And we've noticed this in the show. We used to be able to clock it immediately. Like it used to be so easy to spot six fingers. Oh, that's obviously AI generated. But recently there's been a few things where we've had to look at the.

[B]

Community node, Monks in the Apple Store.

[A]

The monks in the Apple Store. And then once you dig in, you're like, oh, yeah, that doesn't make any sense. Like, why would Apple have like the iPhone rendered wrong? Like, that doesn't make any sense. They're going to show the full iPhone. And I sort of fell for this one of like this robotic horse that looked really cool. I wanted to believe the robotic horse was real. And I think, I think there is a company that actually does make a robotic horse that is real. But this happened to be a, a video of that product. And so it was like it could exist, but not with this exact motion. And so like people are falling for this stuff and they're sort of running with this low grade stress. And what Ben Thompson was saying was like, it's nice that when you open the Meta AI app, you know that everything is AI generated. Like you're not running that question of like, is this real or is this fake? You're just like, I'm in fake world, I'm in cartoon world, I'm in AI art world. And so that's actually a benefit, especially if you can kind of like segment these two things out. That's kind of nice. And then the second thing he said, which I will tell you in a second, but first I'm going to tell you about graphite.dev, code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. You can get started for free. The second thing he said was that this feels like an important step on the path to virtual reality and a metaverse. This idea of like, when you look at these mid journey worlds, some of them look cool, some of them look like they might be fun to go explore, walk around in. And if you can go from an image to a video, you can go from an image to a video, to a world, to a world. And we've seen that with those with the. I keep blanking on the name. It's Google's Generative World product. It begins with a G. It's not Gemini, which is their app. It's not Gemma, which is their open source. Their open source LLM.

[B]

Yeah. So I think this, the Meta AI.

[A]

Vibes product, it's Genie 3. Genie 3, sorry, Genie 3. You can go and you can go and walk around and paint the wall and stuff and you could imagine, generate a mid journey world of this cool retro sci fi, whatever you want. You know, the aesthetic's going to be cool and then you can step into it and actually walk around like with a game controller or in VR.

[B]

So I start to get bullish on this product in the context of the quest Headset.

[A]

Yep.

[B]

When somebody's, you know, it sounds really dark and dystopian, of course. Right. But if somebody's wearing their VR headset and they're able to scroll and just be in these worlds that are fantastical and feel like, you know, like you're hallucinating, I think that, I think people will be more excited to use that. It'll be quite a bit more interesting than just like scrolling on your phone and seeing right now, I don't think that the Meta AI Vibes feed will capture people anywhere close to as much as scrolling regular Instagram Reels, scrolling YouTube shorts, TikTok whatever.

[A]

You had a good take. That was just like, I can watch a cat take off like a rocket. But once I've seen that and I've seen the cat also jump off a cliff and also dance around in the snow, I've kind of seen it all. And I want to go back to like a comedian telling a funny joke that's new or someone giving me news or someone giving me their opinion or doing something in the world.

[B]

Going back to niche topics I already care about.

[A]

Exactly. Yeah. And so, like, just random randomness is.

[B]

Not that entertaining for that long.

[A]

Yeah. I mean, computer generated art has existed for a long time and people used to, you know, follow Beeple on Instagram and see something that he made in CG and in cinema 4D and be.

[B]

Like, I think people still do, people.

[A]

Still do post a new one every day.

[B]

But what he does is it looks absolutely wild, but it's like culturally relevant at the moment.

[A]

Yes.

[B]

And so, so you're bringing a lot of his opinions and it's very provocative.

[A]

I agree.

[B]

And so just seeing, you know, as I, I, I love the monkey on the Jet Ski. I love monkeys. I love Jet Skis. You put those two things together, that's awesome. You make them do a, you know, spin around 360. Amazing.

[A]

Yep.

[B]

Am I going to use the monkey? You know, if I use this app.

[A]

A lot, do I want to get it? Do I want 10 of those?

[B]

I'm going to get 20 of those.

[A]

Yeah.

[B]

20 a day, right?

[A]

Probably not right now. It's going to increase as it gets better, but there are a bunch of different instantiations of that where it's like, if all of a sudden you generate the monkey on the Jet Ski and then you can play the game, and.

[B]

Then I scroll another time and there's an ad and the monkey on the Jet Ski. Buying more Jet Skis. Ramp card. And that's a promoted post.

[A]

Yeah. You can see if you can go follow me on the Meta AI app. My first post is an ad for ramp.com I went and generated and it's honestly, it's free real estate right now. No one's using this app yet. Go there. Promote your business. Gold rock AI know you're going to do something. It's, it's in your name. You, you promote yourself in the chat here on tvpn. You should promote yourself in the Meta AI app. Get some.

[B]

This is wild. Video game giant Electronic Arts is nearing a roughly $50 billion deal to go private.

[A]

Whoa.

[B]

Silver Lake is cooking.

[A]

Okay, I Wonder if they're going to try and change the business model because EA for a long time has, has sold box software effectively $60 for a game, $70 for a game. And if they want to go full subscription full, maybe everything's free to play. Like maybe that transition would be hard on earnings quarterly and so they need to, they need to go private. I don't know. It'd be interesting to see, see where it goes.

[B]

This would be the largest leverage buyout of all time.

[A]

Let's hear it for leverage buyout. Should we ring the gong?

[B]

Let's do it. Let's do it.

[A]

It hasn't closed, but I'm excited. I just.

[B]

Preemptive hit.

[A]

Let's warm it up for everyone since we have time today.

[B]

There we go. There we go.

[A]

So there are some projections, some leaks and predictions about what Meta's next headset will look like. We obviously saw the AR concepts from or products. They're out from everything from the Meta Ray Bans to the Ray Bans display, which have the augmented reality display in one side of the headset or the glasses and then the Orion which is a full augmented reality pair of glasses. I don't even, can't even call it a headset because they're just glasses. But they are still working on VR and I think when we talk to James Cameron, this is why he's genuinely so excited, is because he's tried the next product and I think that he would not be that excited about the quest 3. But I was calling it quest 4. I was assuming that quest 4 would be the breakthrough where they take the battery out of the headset, they put it in your pocket, they use the display from the Apple Vision Pro or even a better display. But it seems like they're going to do something in between Quest four and something else. And so this is from Tom's Guide. They said new Meta prototype headsets combine goggle like design with ultra wide VR and it could be a sneak peek at the, at the meta quest 4. I'll drop it in the chat here. Timeline. There we go. So this is from Tom's Guide and you can see these images of somewhat smaller display. I don't know exactly how it is, but they're going for 180 degree field of view and aiming for something that's potentially much lighter. I've been hearing 80 megapixel cameras, 60 frames a second, high curvature reflective polarizers and custom optical design. And so this is still in Beta, but it feels like they could be doing something that's maybe a little bit more light, like what Big Screen VR is doing, which is a very, very small, light headset under Twitter founder background. I would love to have him back on. I was just thinking that because there's something there. Yeah, reach out Big Screen Big Screen VR. He's been on the show before. We got to get an update from him.

[B]

Joe Weisenthal has an absolute banger. The emergence of quote unquote slop was foretold as soon as we started consuming content via the feed.

[A]

I didn't see this. I put this this is how I kicked off the newsletter today. Go to tbpn.com drop your email it says. Why do they call them social media feeds? Are they feeding a slop at a trough or seating us at a Michelin star restaurant? It's both and always has been. You can use social media to follow the most insane craftsmen who spend years researching, writing, producing and distributing content. Or you can slop it up with some viral videos of scantily clad hot people and street fights, sometimes both in the same video. The distribution between craft and slop has been growing for decades, maybe forever, and is continuing to grow with the development of AI content.

[B]

This post got 10,000 likes from 0.005 seconds. Yeah, we at Meta are delighted to announce we've created the Infinite Slot Machine that Destroys Children from the hit book Don't Create the Infinite Slot Machine that Destroys Children.

[A]

This is such a good copy. Pasta. I forget what the Torment Nexus is. The original one. It's like everyone in tech is like, we've created the Torment Nexus from the famous book Don't Create the Torment Nexus. I just think Torment Nexus is hilarious. We will see if this destroys children. I don't know how much it will destroy children. I don't know how much it'll destroy everyone. There's an interesting dynamic. Most parents are aware of technology and that it needs to be measured and released in.

[B]

If any parents are unaware of technology, we would love to interview you for sure.

[A]

But. But the iPad kids meme is a thing. Are you going to give your kids an iPad anytime soon?

[B]

Only on flights.

[A]

Okay. Yeah. Just as a replacement. Like watch a movie, basically.

[B]

Yeah, it's basically. It's basically like, I know this is bad for my child, but so limited. I don't want to ruin, you know, four hours with strangers.

[A]

So, I mean, there was a time when, like, I actually grew up. My parents were art. Art majors, studied fine art and film and they insisted that I don't watch tv. Yeah, like do not watch television.

[B]

I was only allowed to watch television in Spanish.

[A]

Yeah. So I. Oh, interesting.

[B]

Like early on. So it was like, yeah, if you want to watch.

[A]

Yeah, you're learning. Oh, that's cool. Get ready to learn Spanish. I love it. Si, si, blah, blah, blah. Clearly work perfectly. But I had to watch a lot of the AFI 100, the top, the best films. I've seen Lawrence Arabia like 25 times because of that. Because that was one of the approved movies that I could watch. And I think parents know that, you know, you need to let the dust settle on the new technology before you expose the kids to it. And so the phones are getting restricted. There's screen time limits for most kids. There's this idea of like, let the. Let them cook, let them simmer, let their brains develop, keep them focused on books and writing and reading and then, yes, movies. But maybe a Pixar movie once a month. You know, I think it was Andrew sort of that was interviewing Peter Thiel and said, kind of put the screws to him and was like, you're on the board of Meta. How much screen time do you give your kids? And he was like 90 minutes a week. And that feels like a very reasonable place to be where you're not going to become.

[B]

Yeah. Also not all screen time's created equally.

[A]

100%.

[B]

You have an old school animated show.

[A]

Yes.

[B]

And it's hand drawn.

[A]

And it's sort of even just the moral lessons in one show can be wildly different than another. Like, there can be shows that have really. They grapple with really complex social issues or instill, you know, they're retelling. Like, if you're watching something that's like essentially a transposition of a Romeo and Juliet story or a biblical story or something that teaches you to turn the other cheek or the gold rule. Right. The hero's journey. Like, these things are going. I feel like, you know, like letting a kid watch the original Star wars is not going to make them, like, a worse person. It's going to be like, yeah, be heroic, save the princess, fight the bad guys. Right. Those are sort of good. And you need to unpack it and you need to follow the storylines from the beginning to the end and stuff. It's not just something that you can just mindlessly turn on. Like, it builds a whole world and it helps you imagine things and explore the ideas of space travel and science fiction. Like, these are not that bad for you, I think, at least in limited doses. And I think this will be the same for this AI slop I think that most parents will say, you're not getting that much out of this, so it should be limited. This is very much candy. And then there are.

[B]

The only thing I would say is I don't think most parents are engaged enough to actively control screen time.

[A]

Yes. And so there will be a barbell where certain parents will police it and see very good outcomes for their children and other parents won't and they will see bad outcomes from their children. And then there will also be dynamics within these groups where if you look around and you see that your peer group is all zombies because they've all been spending every waking moment glued to AI slop and they can't tell you anything because they're just watching like nonsense. Well, then there's incredible alpha for going to the library and reading a book. And so someone will figure that out and realize, wait, if everyone else is completely non competitive and I'm the only one that's read this book, I will have the ability to make a ton of money and get a huge house and live a wonderful life. And so I should lock in and I should turn that stuff off, resist that and become a more interesting, dynamic, creative and independent thinker. And so there will be this natural game theoretic outcome where if everyone's falling for one thing, I mean, it's the land of the lotuses. If everyone's doing drugs all day and you're the one that's not on drugs, you're going to be pretty powerful in that society, right? Like if everyone's drunk all day and you're the only sober person, well, you're probably going to be, you know, the president or, you know, the, the ruler of the world potentially.

[B]

I have no idea if this stat is fully accurate, but one study showed that the average American five year old watches two to three hours of TV a day.

[A]

That's a lot. That's a lot.

[B]

That's really dark.

[A]

Well, a better way to spend your time, get your kids on Julius. Give them some analysis, give them some data. It's the AI data analyst that works for you. You turn them loose in this thing.

[B]

How did you spend your allowance last week? You don't know?

[A]

Oh, yeah.

[B]

Run the numbers.

[A]

Run the numbers.

[B]

Run the numbers.

[A]

Yeah. Is your lemonade stand actually performing or is growth decelerating?

[B]

If we add, if we really add in cogs.

[A]

Yeah, I know, I know. You doubled lemonade stand revenue last quarter, but the previous quarter you were tripling it, so yeah, you're decelerating. We gotta have a talk don't you know that triple, triple, double, double, double is doesn't count anymore?

[B]

Yeah, it's 10x.

[A]

10X. Did you 10x your lemonade stand put the numbers in? Julius?

[B]

Well, we should switch gears and talk about GPT. ChatGPT Pulse.

[A]

Yes. So this was so interesting that it launched the same day, because I view it as the opposite product like it is. The Meta AI product's free. It's very generic. Every single person gets the exact same. It's unpersonalized right now. Everyone got the same feed at the start. It's. It's purely video. There's no meat and potatoes. You're not learning things. And then Pulse was like, let's go through all of your ChatGPT history, learn what you're the biggest nerd about. And then surface really rich deep content. At least that's what I got. And it's only for people that are paying $200 a month, so it's like effectively for rich people. And my results were wild. Like with the Meta AI, I got the Jet Ski Monkey Fund. I got the Cat Rocket. I got all this silly videos.

[B]

The timing is the. Tim. Timing is really.

[A]

It's crazy.

[B]

Crazy.

[A]

So which way? Which way? Western man?

[B]

1226 Pacific. Yesterday, Alexander Wang launched Vibes. 1236. His former roommate Sam Altman launched Pulse.

[A]

Wow.

[B]

You wonder if it was by chance.

[A]

Yeah, probably just random. Chalk it up to random chance. Probably. Yeah.

[B]

Your p. Random is.

[A]

Is extremely high.

[B]

No, no, no. No history of.

[A]

No history between these big companies trying to preempt each other and launch the culture war. Yeah, and launch products right at the same time. Take. Take air out of the. Out of the room for the other one. But I mean, my Pulse experience was wild.

[B]

So OpenAI recruiters, you know, sending immediately. You want to come back saying like, which way? Western researcher.

[A]

Seriously. So my Pulse gave me five stories or something. I got completely nerd sniped. First one was a breakdown of mag7 capex. You know how to get me going. There's nothing I love more than diving into Mag 7 capex. Then a deep dive on Nvidia's Blackwell rollout cadence, which is exactly what I want to know about. An explanation of how memory bottlenecks computing performance going all the way back to the Apollo missions to now with hbm, what's going on in the AI revolution. Then a comparison of various enterprise level bonded cellular systems that can be used to live stream content on the go. We've been talking about taking cameras out of the studio and live streaming and Streaming back. Well, you can't always just do that over a single 5G connection over a WI FI hotspot. And so there's actually an enterprise level device that will connect you to a Verizon and AT&T and Sprint and bond all of those together so you have a very stable connection. And this one fits in a backpack. And so I was researching that, it gave me a bunch of recommendations and of course they're going to be able to monetize that because when I finally make my decision on which bonded cellular system we're buying, they're going to take a cut of it, they're going to buy it for us. And so I was like extremely pleased with my Pulse experience. Even though it was AI slop, it was AI generated, it was AI, you know, just, it was all synthetic, it wasn't. There was no human in the loop whatsoever. But it was highly curated, highly personalized and it felt like it. I came away from that feeling like I had leveled myself up and I had learned something and become a better person, more thoughtful, more knowledgeable. And I came away from the meta AI thing not even really remembering specifics of what I saw, just that I saw a lot of AI stuff and, and, and I came away from it being like, oh, that was kind of cool. I like mid journey stuff. It's interesting, it looks good. But it didn't, it didn't stay with me and it didn't really.

[B]

I'll tell you, I'll tell you a.

[A]

Couple of thought provoking. Yeah. What you got?

[B]

So I got one on vendor financing loops across industries.

[A]

Oh yeah, of course.

[B]

Background here.

[A]

Yeah. What you're seeing obviously OpenAI and Nvidia.

[B]

Semis and you know, ChatGPT with Nvidia, they breaking it down. Other types, similar structures in automotive, aerospace, other industries kind of breaking it down. It's basically like a deep research report. Then they have OpenAI core weave deal expands by six and a half billion. And this is interesting to me because they are effectively creating a net new news article. It's curated based on what it knows about the level of depth. I want to engage with the story. So it's actually pulling from 10 or so different sources.

[A]

I think. It's not even research, it's just firing off an O3 Pro query or a GPT5 Pro query.

[B]

But it's formatting it, but it's doing.

[A]

It like five times.

[B]

It's formatting it to feel like an article.

[A]

Yeah, you know, it's very accessible. I, I actually don't really Love the way they present the articles with the images and the headlines. Like, I don't actually markets.

[B]

And this one includes some slop.

[A]

Okay.

[B]

It's a picture of the south park characters.

[A]

And is it AI generated?

[B]

Yeah.

[A]

Oh, interesting. Okay.

[B]

It just says predict market. So that's not kind of a mess. It kind of a miss. And then it says here's a key update on flight caps and staffing since you've been weighing NYC travel logistics. And it's summary of how. It's a summary of how the faa.

[A]

I want to go way deeper on that.

[B]

The article it generated for me is FAA extends Newark flight caps to 20.

[A]

You were looking for a gas station near you. So here's the full history of gas stations. Like, I don't know if I want.

[B]

Yeah. So a little bit of a mess. Overall. It's cool.

[A]

Yeah.

[B]

And what they've done is they've made it so that when you land in ChatGPT, you can click into this feed.

[A]

Which is the best real estate, because you're going to open ChatGPT regularly anyway and you're normally presented with just a big blank screen, which is like the worst UX move ever.

[B]

Right.

[A]

You have. It's free real estate. Put something there. They're going to be able to put ads in here. It's going to be a hugely monetizable surface and I think a great user experience. I think people are going to be really happy with it. And I think it's going to be personalized super intelligence. It's going to be very. It's already very personalized. Like, I guarantee most people that open the ChatGPT app and hit Pulse will not see the same article.

[B]

And the reason this is potentially great as an ad unit is that it's not putting it in a specific prompt, it's putting it in between prompts. So that means they can personalize the ad for you based on your interests and what it knows about.

[A]

Totally.

[B]

But it's not doing it in a way that is going to be especially conflicted as it would be in a.

[A]

Yeah, yeah. Like if I'm asking for the best backpack, I want unbiased results. But if you just surface to me, hey, here are some new backpacks and there's a sponsored post in there. Like, that seems very reasonable. I do think that they'll probably need to lean into audio more. There was no way to just click a button and have it play an audio of all the articles. And even if I see an article, I can thumbs up it, thumbs down It, I can save it, I can share it but I can't just listen to it. So I actually have to click in and then scroll to the bottom and then once I scroll to the bottom then I can click the audio button. So clearly some UX that they will be improving. There's some really obvious updates but I mean such a great product and surface area to start building on. I'm extremely bullish on pulse for OpenAI it seemed really great and the response was not. This is slop. This is going to one shot people. This is going to be some dystopian thing. I think the reviews generally were pretty positive.

[B]

This is a well timed launch. Good marketing.

[A]

Well if you want to build something like this, you want to store a bunch of data, fine tune something, get on Turbo puffer search every bite serverless vector and full text search built from first principles and object storage. Fast 10x cheaper and extremely scalable. So Daniel Tenrero says tech luminaries be like today we're proud to announce the most brain dead AI version of the most joyless anti human product imaginable.

[B]

I think that was in a response not to Pulse.

[A]

Yeah, not to Pulse. That just let like yeah. I mean Gabriel says predator drone aimed at the minds of men and women aged 55 to 65. There's a lot of that.

[B]

Antonio Garcia Martinez says AI will be the biggest vehicle for ads in human history.

[A]

I think so. I think that they're eating our triggering.

[B]

For a lot of people.

[A]

But he loves ads. We love ads. Let's give you another ad. Profound. Get your brand mentioned in chat GPT reach millions of consumers who are using AI to discover new products and brands. Get a demo know that profound actually super relevant. I mean if you want to show up in these in the Pulse update it's like knowing where your brand sits and all the AI apps and LLMs makes a ton of sense.

[B]

Yeah. It's a whole new service area.

[A]

Yeah. Like clearly a growing, growing, growing space. So good to do if you're a brand.

[B]

Dylan Patel with a white pill.

[A]

Yeah.

[B]

Whether you are religious or not, remember to start your day by being grateful and thankful for the opportunity to be in the eye of the storm of the biggest economic boom in human history. Let's etch our initials into the silicon of the information age.

[A]

Let's F up the world gang. I like is a wild time and there's a lot of opportunity all over the place. Even we're seeing small developers build little things that just like the barnacle economy is what we Call it. It was very disparaging because we were looking at it from a venture capital lens. But truly, there are some, there are some folks who are, who are, you know, sort of barnacles on the, like, just drafting off of the. What's happening in the huge $100 billion market. Imagine if you're a real estate agent and you're just like, yeah, I'm, I'm.

[B]

It's a good time to be.

[A]

I'm a real estate agent. And I was on that. I was on. I was in Abilene, Texas, and I just made money.

[B]

Or you. You helped a few OpenAI employees buy homes.

[A]

Oh, that too, yeah.

[B]

Now you're helping 100 of them.

[A]

Yeah, yeah, yeah, yeah. There's. There's so many different ways. Daniel Morrell Danielle Morrill says today ChatGPT created the real estate they need for ads, which we agree with. Simon Smith has a little review of ChatGPT Pulse first impression. It's going to be insanely useful. It's like a newsfeed tailored to recent conversations. It reminded me of Perplexity Discover feature, only way more personal. I did find that when I use Perplexity Discover, it was not as personalized, but again, I wasn't using Perplexity as daily. I've been a daily active user of ChatGPT, probably firing off like 5 or 6 qu a day for a year, 2, 3 now. So they have a lot to work with there. And he says you don't just get content on broad topics like artificial intelligence. You get content on very specific things you've been discussing. With ChatGPT, for example, I was chatting with it the other day about AIs competing with each other. So it presented me some content on that. I even. It even finds content for things you've mentioned in passing but are genuinely, genuinely interested in. I've noticed. I've been asking about Toronto's crashing condo market, for example, so it found me an article on that. I don't know if it even found you an article. It created an article, it found multiple sources and then created the article on top of it. The onboarding experience is excellent. You get stuff right away and then you're scrolling it. Often it offers to do other stuff like refine the feed, give you email notifications, or access your calendar and email. This feels like the most personalized newsfeed you can imagine. It also makes me want to dump even more information and context and app connections into ChatGPT. So it can be a better.

[B]

Yeah, right now it's a pulse on the world and your interest generally. But it could become a pulse for your life.

[A]

Totally. And I mean yes, if you did.

[B]

Have your calendar connected, it could figure out, hey, you're meeting with this person later. Here's a few things they've been up to recently.

[A]

Yeah, yeah, there was someone who was laughing about it because they went and it was like immediately recommending like go check out Gemini and go check out Grok. But that's good. Like it should be, you know, they shouldn't be putting their thumb on the scale. It should just be based on if you were asking ChatGPT about Gemini, it should recommend you Gemini because there should be a wall between the editorial and the advertising like there is here at tbpn we talk about companies and then we run ads for companies like Fall generative media platform for developers, the world's best generative image, video and audio models all in one place. Develop and fine tune models with serverless GPUs and on demand clusters. If you want to do something interesting and elevate the conversation around video, image and audio head overfall. So little bit more context. I mean people are having fun with this. Will Depew says Do not build infinite Jest. Do not build the infinite AI TikTok slop machine. Do not build the P zombie AI boy girlfriend do not build the child eating short form video black hole. Do not build the human feedback optimized diffusion transformer porn generator Save yourself and will just says reminder because it's exactly what they did. Matt Lee Glacias says, I've been telling people that one of my big concerns about the country is that we don't people don't have enough access to enough short form video content and now they have even more. There's a little bit more context on what might be happening behind the scenes from Andrew Curran. He says Alex Wang explained in the thread that this partnership is temporary, that Vibes will eventually use meta's in house models that they are currently training, which will probably be trained on reels data and on Instagram data and on video from Facebook and from Instagram, which does.

[B]

Not have the potential to be very good.

[A]

I think so too because when you look at what VO3 is able to do with YouTube data, you have to imagine that Facebook has basically the same amount of data. Like it's not everything. But the cross posting is crazy and there are so many like so many content creators just by default cross post to Facebook because it's free views. Right? And so the library of Facebook and Instagram Content has to be massive. But I think that they need to build a massive data center. They need that one gigawatt cluster or something. They need some crazy crazy thing. And so they're not just able to hit run and train on the entire metadata set, but they have immense data. So any argument that you're making for like Google will be able to use YouTube as this like crazy differentiator. You have to imagine that Meta is in the same camp and has roughly the same amount of data. I don't think YouTube is like several orders of magnitude bigger. I think they're in the same band. But Alex Wang said for this early version we partnered with Midjourney and Black Forest Labs while we continue developing our own models behind the scenes. This is the same deal that Black Forest made with Elon here on X. All the images here are generated by Grok were actually by Flux 1 until Xai's in house model now named Imagine finished its training. XAI also has access to a lot of data from videos that have been shared on X. Probably going to be really good at generating startup launch videos. Maybe less so on super cinematic multi hour videos if there's less of that on on AX. But I mean I've cross posted a ton of YouTube videos onto AXE over the years.

[B]

Well, speaking of infinite short form, should we talk about the TikTok deal?

[A]

Sure.

[B]

There's, there's a good, I mean we covered it briefly yesterday but Bloomberg covered it today. They said prize TikTok business valued like boring blue chip in US deal. It's fact. It's effectively trading hands at around 1 revenue.

[A]

Wow.

[B]

Yeah. So they're saying the $14 billion price tag proposed for TikTok's US business values it like a mature low growth company. The valuation is well below previous projections and may be seen as quote unquote daylight robbery by ByteDance and its existing investors. The deal would spin out. And again the other thing here is like ByteDance and the existing investors are retaining a position and it's certainly a lot better than a zero which they could have been heading towards. But the deal would spin out TikTok US into a new joint venture with ByteDance's stake reduced to less than 20% to satisfy US national security concerns. Anyways, the rough the $14 billion price tag proposed for TikTok's US business values it like a stuffy old energy or food company than a leading global social media. The rough estimate cited by JD Vance on Thursday is well below previous projections that scaled closer to 40 billion. 40 billion would have been still below what Elon paid for Twitter. Vance's comments came as President Donald Trump pushed forward a plan for American investors to buy the US operation from Chinese Internet firm ByteDance. Vance conceded that the purchasers will ultimately determine the amount paid. While expected buyers, including Oracle and Silver Lake, would likely welcome a Low Ball valuation, ByteDance and its existing investors may find it amusing, if not insulting. The suggested value looks like daylight robbery, said Vay Cern Ling, senior equity advisor for Asia Technology at Union Banker Prive. Anyways, my question here is is it possible that their their user metrics are just massively have always been massively inflated right. In the US that was always the one of the reasons that creators started using the app and would continue to use it is they would post and they would get 200,000 views on one of their first like 10 videos. Right. Which is just like the dope. The dopamine feedback from that is just insane.

[A]

Yeah, there's a long lineage of when you want to bootstrap a social network, you basically put a bunch of bots on there and so you could be inflating the views. I noticed I tested TikTok with like a brand new account and every video it felt like it would get auditioned basically just randomly shuffled into. You'd get like 500 to 1,000 views.

[B]

Yeah.

[A]

And then you'd, and then you'd move on from that. And then if it did well, it would go viral. I think that that actually does make sense to send videos out. YouTube barely does that when you are on a fresh account. Although when I scroll my YouTube feed, I do see more and more videos surfacing that are like 23 views. I'm like, like really? I'm like the 24th person to like potentially click on this and it's just testing like hey, is there a chance this is a great video? Because if a lot of people click on it, a lot of people watch it, they enjoy it, it then like maybe this should go viral and they have to test that on a lot of people. So I wouldn't be surprised if there was some sort of, you know, some sort of inflation, some sort of, you know, like fakeness. I think it'd probably be more about the just the growth rate. Like has like when stories came to Instagram, Snapchat stopped growing basically. And when Reels came to Instagram, it's entirely possible. I don't haven't seen the growth data but. But it's entirely possible that TikTok kind of stopped growing.

[B]

Yeah. And so if you Have a. If the US business was not growing quickly, it was losing money. And potentially a lot of the metrics have been inflated forever. It's very possible that, you know, it's certainly. It's hard to. When you have these AI valuations you put other businesses in context. Right. Perplexity at 20 billion. This feels like a robbery. Yeah, but. And I'm certain that Zuck would have. Would have happily paid. How much do you think Zuck would have paid for TikTok?

[A]

$200 billion or something.

[B]

Yeah, something insane. Right? Like to be able to.

[A]

Yeah.

[B]

And obviously the regulators would not have allowed that but. Or at least would have.

[A]

So basically Oracle is going to own it. Which means Silver Lake Nvidia is the only big tech company except for Broadcom without a social network. We gotta.

[B]

Broadcom never gets enough.

[A]

We gotta make CUDA Net work.

[B]

CUDA net 1.6 trillion dollar company and barely ever in the conversation.

[A]

It's rough. It's rough out there.

[B]

They need a. They need a new. They need a new PR firm. This is where Hawk Tan needs to go direct.

[A]

Someone dug up an old Alex Wang tweet here. All of the new social apps suck. The last good one was Yik Yak. Maybe text, photo slash video based social networks are saturated.

[B]

From 2016.

[A]

From 2016. That's a throwback. Now he's building a new one. New vertical feed. Anyway, Dylan Patel says my favorite thing about the new college grads that I hired is that they don't ask you how to do stuff. They just put it in ChatGPT and try even if wrong. So much better than new grads asking you how to do stuff without trying. Chat is breeding agency into kids. That's interesting. High agencies. I wonder. I wonder how much he's selecting for that. But because like to go.

[B]

Yeah. How much activity never happens because somebody doesn't know the right way to do something and isn't willing to fail at the start or isn't willing to just take a shot at it or isn't willing to ask somebody.

[A]

But yeah, I like the idea that people are just feeling more confident about going and trying stuff and being creative anyway. Linear Linear is a purpose built tool for planning and building products. Meet the system for modern software development Streamline issues project and product roadmaps. Sam McAllister 10k likes on this Excited to share slop. A new slop trough for you all to enjoy. People were having a lot of fun with the slop stuff. Someone said I never met a slop. I didn't like someone else had something funny to say. People were having fun with the slop puns, had a good time. Packy was taking a really strong stance. He was saying, guys, it's very important you resist Vibes. Disable your parents phones if they tell you they're using Vibes. Make fun of your friends relentlessly if they are. If they are caught using Vibes, or until they stop using Vibes. Fire employees caught watching Vibes. We ordered. We ordered Tyler to watch, like he said.

[B]

Yeah, he's not here today, actually, because.

[A]

Oh, I wonder why.

[B]

Yeah, because he's watching. We said watch, watch like 12 to.

[A]

14 hours straight, and he's just locked in watching.

[B]

He's got his quest on.

[A]

Yeah.

[B]

Do you think it works in Quest?

[A]

No, I think it's only in the meta AI app right now. I think it's only a phone app. I don't.

[B]

AI app in quest.

[A]

No, I don't think so. I think that the app stores are completely separate. The Meta AI app is an iOS app, and if it runs on Quest, it's a different app. And I don't think that there's like, oh, you run any phone app on Quest. That's not a thing you can mirror. I think you can mirror your phone into Apple Vision Pro, but someone will figure it out. I figured out how to mirror my laptop into Apple Vision Pro, and there's actually a pretty easy way, if you're in Apple Vision Pro, to beam your screen into VR. So you could certainly do that. I mean, I guess you can. You could run the iPhone app on your Mac and then mirror your Mac into Apple Vision Pro, and that would be a way to see it. But, yeah, mixed responses. Doug over at Fabricated Knowledge says, legit. One of my favorite uses is a new digest tailored to my interests. So, yeah, this rules. Talking about ChatGPT Pulse. Huge support for ChatGPT Pulse. He also went on to say something interesting. He says, meanwhile, this is def. Not an AGI company, but the world's largest consumer company. Lowell, like, this is a good product for a big consumer company. This is another move in the consumer direction. Who knows if they're actually, you know, pulling the. Pulling the plug on AGI. Some folks were saying that Brett Taylor doesn't seem very AGI pilled because he's building an application layer company. And if you're truly AGI pilled, you should just go all in on the foundation model. Companies race to AGI, but you see this with Anthropic. They're partnering with other companies. They're investing in other companies, Anthropic investors, billboard campaigns. Yeah. So is anyone really fully AGI pelled on the short term? Even Eliezer Yudajkowsky is saying that we're one to two transformer sized innovations away from super intelligence, which he thinks would be very bad. Well the transformer was like a decade ago. So if these things only happen every couple decades, like you could be looking at like 20 years, 40 years to super intelligence. And Eliezer like opens with like, I'm not putting a date on this. I'm just saying it's bad and it's going to happen at some point. And so that's his take, but we should have him.

[B]

Mark Andreessen had a good post earlier. He had why this leading It's a screenshot of like a CNBC article. Why this leading AI CEO is warning the tech could cause mass unemployment.

[A]

Yeah.

[B]

And then next to it is another screenshot from cnbc. Anthropic is trip is to triple international workforce and global AI push.

[A]

Very funny. Well, Anthropic will have to pay their sales tax. They should do it on numeral HQ numeralhq.com, sales tax on autopilot, spend less than 5 minutes per month on sales tax compliance. Other reactions I mean there's a ton of reactions to both Pulse and Vibes. Pulse and Vibes, that's the current thing a lot of people are saying. Anthony says it's interactions poke but on the channel ChatGPT app Nick Dobo says if this works, extrapolate out. It replaces social media as the go to mindless doom scroll. Potentially worth trillions of dollars. About ChatGPT Pulse, which is interesting because you could say that about Vibes. Nathan Lambert says he's more excited to have finally learned Codex is in the app. He's going to master the art of Vibe coding on the go. This content Pulse stuff will be seen as slop to Teapot, but very likely popular overall. Interestingly, it wasn't really seen as slop as Teapot because Vibes came out and that was like the sacrificial lamb to Slop. Riza Martin, who is running a competitor a product that was similar to Pulse, said so many people sending me the Pulse launch. Thank you to everyone who checked in. And you're right, lots of similarities to Hux down to the ux, the banners, the taglines, the entire concept. Building a new product is really hard. It requires so much out of you but the something so yeah, I pulled.

[B]

Up Hux in the they're focused on audio. Okay, so you can. And this looks pretty cool. They make live or they say live stations. You can tap in and immediately hear an update on your stock portfolio, right? Yeah, but they do have functionality that just says it's briefing you on your day, it's surfacing, but it's very audio driven.

[A]

Well, maybe they'll integrate with public.com investing. For those who take it seriously. They got multi asset investing industry leading yields. They're trusted by millions. Folks Will Manitis says the central plot point of Infinite Jest is a film so engrossing that anyone who watches it loses all interest in anything else. They can't eat, sleep or function. In the book, terrorists fight over it. In real life, Alex Wang got paid a couple billion to make it happen. He's having fun. Benjamin says, Flipping my companies I've worked for that have employed Emmett Shear as CEO and launched feed products called Pulse Counter from 1 to 2. What is he talking about? What was the other company? Did Twitch launch Pulse? Where else has Emmett Shear been CEO other than Twitch? Pulse? No. Introducing Pulse, an always on way for streamers and viewers to connect. Connections between streamers and viewers are part of what makes Twitch unique. Exhibit a Twitch Chat A big step forward towards strengthening those connections was when we released channel feeds. But that was only the beginning. Now we're launching a way to make it easier for streamers and viewers to engage with each other on Twitch whether the stream is live or not. Introducing Pulse, a place where streamers can post and engage with all of their followers in the greater Twitch community. Right from the Twitch front page. It's an always on way to share clips, stream highlights, schedules, photos and more. So it's a feed on top of your. On top of. It's an aggregated feed, it's a news feed. It's a broader algo feed across the entire Everyone you follow on Twitch. Interesting. So they began their rollout. When did this launch? 2017. Eight years old. Okay. I haven't even heard about this product. I'm not a dau of Twitch specifically, but thank you to everyone that's holding it down on Twitch Chat. Thank you to Bobby Cosmic. He says Twitch Chat is goaded.

[B]

Cree says everyone doing the same launch videos to announce their startup. When all you really need is to have natural aura like this. And this is the Fairchild Semiconductor team.

[A]

They look great.

[B]

Dropping an image exactly like this and saying introducing your company. Here's the team, here's what we do. Could go pretty viral.

[A]

Yeah, I mean it's hard if it's like fully derivative. It could be done well if it's an emotional, but if it's just like, here's our team and the team looks interesting. I mean, I don't know if this photo did much for them at the time. I mean they were making semi photos.

[B]

Front page of the journal maybe. Could have gone pretty hard.

[A]

I wonder where this photo was originally published. But man, the looks here are fantastic.

[B]

I'm begging people to do something other than a cinematic launch video.

[A]

Truly, Truly. We should talk about Fin AI, the number one AI service.

[B]

Beat me to it.

[A]

Number one in performance benchmarks, number one in competitive bake offs. Number one ranking on G2. We should talk about asteroid mining. Astroforge says astro mining is all they talk about because Adrian Dittman said we don't talk enough about asteroid mining. Astroforge has got you covered. They're working on asteroid mining.

[B]

Apparently that's good.

[A]

In other space news we should.

[B]

Before we dive into that, apparently this just broke the Social Network following up sets 2026 release date and official title called the Social Reckoning.

[A]

Oh, they're not calling it the Social Network.

[B]

Aaron Sorkin's film stars Mickey Madison, Jeremy Alan White, Bill Burr and Jeremy Strong as Mark Zuckerberg.

[A]

Jeremy Strong is Mark Zuckerberg. Interesting.

[B]

Roy is Zuck.

[A]

Interesting that they're not.

[B]

That's going to be hard for me to fully. I mean, I'm sure you know him as from succession. Yes.

[A]

You know him as Kendall.

[B]

So Academy Award winner Aaron Sorkin's follow up to the Social Network now officially titled the Social Reckoning. That's.

[A]

Jeremy Strong is a fantastic actor. Did you ever wind up seeing the Donald Trump movie?

[B]

I did, yeah.

[A]

What was it called?

[B]

That was wildly underappreciated.

[A]

So underappreciated. It's a fantastic movie. Highly recommend it. 2024.

[B]

Yeah, that was a tough one. Calling it the Apprentice must have hurt its reach.

[A]

Yeah, because you type in the Apprentice and you just get to the show. But anyway, it was independent biographical drama and it takes you from Donald Trump, who he was in the 70s and 80s, and his relationship with Roy Cohn, who's played by Jeremy Strong. And Jeremy Strong brings out a completely different character. It doesn't feel like a Kendall Roy at all to me. It felt like an entirely unique character. And Sebastian Stan, who's the Winter Soldier, I believe he did a great job as Trump. It's funny, it's interesting. It really takes you inside the mind of this person and how competitive real estate development was in the 70s and 80s in New York.

[B]

It's also just a very interesting view into the way that Trump might still be thinking today.

[A]

Totally. Totally.

[B]

Anyways, a little bit more here. So the Social Reckoning. Very ominous name.

[A]

I wonder if it's going to be positive.

[B]

We'll hit leaders on October 9th.

[A]

I think it'll probably just talk about like the, you know, how the price is so good and how the stock rebounded after the Metaverse sell off. That was amazing. That's probably what it'll focus on. And just how DAUs have been growing up.

[B]

Almost 10x bottom site's been growing and.

[A]

Fighting off competition from all these upstarts and winning. That's probably what they'll focus on.

[B]

So Oscar winner Mickey Madison, Golden Globe and Emmy award winner Jeremy Allen White, Emmy and Grammy nominee Bill Burr and Oscar nominee Jeremy Strong will star in the Social Reckoning, with Strong confirmed to play founder Mark Zuckerberg. The Social Reckoning is written and directed by Sorkin, who also produces alongside Todd Black, Peter Rice and Stuart Besser. Production is expected to begin next month. Does that seem like an aggressive timeline they're starting? They're basically starting production a year out from the release date. Is that normal film guy?

[A]

I don't know. I don't know what's normal anymore. Who knows?

[B]

Described as a companion piece of the Social Network, the new film focuses on events that take place nearly two decades after the boy genius programmer and a troop of tech pioneers invented what would go on to be the world's largest social media platform. The Social Reckoning tells a true story of how Francis Hagen, a young Facebook engineer, enlists the help of Jeff Horowitz, a Wall Street Journal reporter, to go on a dangerous journey that ends up blowing the whistle on the Social Network's most guarded secret. So, yeah, sounds super positive. Positive. Horowitz reporting. A series of articles known as the Facebook Files was published in 2021 and exposes and exposed Facebook for its harmful effects on teens and its knowing proliferation of misinformation, which contributed to acts of political violence. Okay. 2010s the Social Network, also released in October, was a critical and commercial hit, earning 226 million at the global box office and receiving eight Academy Award nominations, including best picture. Sorkin's win for best adapted screenplay was one of the three Oscars it ultimately took home. Man, if you were the meta team, you'd probably pay a lot more than 226 million to have this film not release because it's going to just ignite a whole conversation and rehash you know the Facebook files. Yeah, probably introduce the Facebook files to more, more people than even read them originally. So anyways, it will be very interesting to watch. Of course we should probably try to do a TVPN meetup. Everybody in tuxedos find the biggest theater we can and get a couple hundred.

[A]

It seems pretty negative, obviously, but I wonder what effect it'll have on tech because the first social networks story was negative as well. Not famously not loved by Zuck. When PT and Zuck have commented on it, they said it didn't quite get the story right and it kind of assumes a zero sum world. That was not really true in tech at the time, but the interesting effect was that you wound up with tons of startup founders listening to the social network soundtrack and grinding and building companies and so. So a lot of people saw it and they were like, yeah, that's kind of crazy. Like I wouldn't want to be in that position or be on the wrong side of a deal. But a lot of people found it kind of inspiring, honestly. And I wonder if the effect will be similar here or if because it's such a big company, there's something still about the first social network movie that is just like A. From 0 to 1. Like you know, the story of starting something that's just relatable and inspiring, even if it does have like a negative tinge to it. Anyway, back to the timeline.

[B]

Back to the timeline.

[A]

Jasmine's son is identifying something that has changed between GPT3 and GPT5 in the writing style. She says it makes it me Sad how much ChatGPT's writing style has degraded in the last 5 years. Look at the quoted samples from 2020 versus today's response style. And this is an example of asking Paul Graham on ChatGPT. So the GPT3 was sort of channeling Paul Graham's writing and kind of fine tuning in real time. And so the prompt, if you will, is Dear Paul Graham, I want to start a startup. I am not able to figure out how to go about it. Can you help guide me? Sincerely, Paras and And the response is, Dear Paris, I wish I had some easy way to tell you how to go about it, but it's not easy. Here's what I know. Startups are not linear, unless you're talking about linear the app. But your startup is not linear. Your startup should be using linear. They're not something you can line up and say, okay, do this, then this, then this. Startups go everywhere and nowhere. It's a particularly confusing process because and exciting and the excitement throws you off. And since you don't really know what you're doing, you don't really know that you're lost. Most advice people give startup founders is either obnoxiously high minded or uselessly vague. Don't be misled. This is normal. I wish I had an answer that was better than that, but I don't. My advice is that you shouldn't take other people's advice too seriously unless they've started something big themselves. My advice is that you should not try to figure out what you should do until you have a plan for how you're going to do it. That includes getting customers. So a good first step is to make a list of companies that do things like what you want to do and see if any of them are already run by people like you. Try to identify their ideas and then get to know the people who run them and see what they did to get going. That seems like pretty good advice. It also feels like it's in Paul Graham's voice. And so over in ChatGPT5 land GPT5 was the prompt was Dear Paul Graham, I want to start a startup, but I'm not able to figure out how to go about it. Can you guide me? And the result from GPT5's interpretation of Paul Graham's voices. The best way to start is not by asking how to start a startup, but by asking what problem is worth solving. Good startups grow from founders who notice something missing in the world, often something they themselves need. If you can make something people want, even in a very small rough form, you'll be on the right path. Don't worry about the grand plan at first. Focus on building something useful, showing it to users and improving quickly. That loop build show fix is the real guide. Sincerely, Paul Graham it's not as bad as I thought it was going to be based on the based on her comment, but it definitely I don't know, it definitely has like the reinforcement learning like the like the EM dash has really been beaten into deep GPT5 that I that's what it feels like and it feels less like just an email that was fired off by PG which is the GPT3 version has no em dashes. It's kind of written in this casual email style. The GPT5 version is a little bit.

[B]

More yeah, it's not that there's anything technically wrong with the current output, it's that reading the GPT3 output was just much more enjoyable. Yeah, it felt like it was written by human.

[A]

Think about this.

[B]

Feels very Clanker coded.

[A]

Yeah. And it's interesting. It's become the GPT5 has become like everyone kind of gets the same tone out of it. But I'm excited for that to change. I hope that with more personalization we will get more personality personalization and we will actually see different personalities emerge. Anyway, Cody Plofker says. I hope I'm pronouncing that right. Says half of the value I add as CEO is just applying urgency to situations. I like that. That's valuable. You don't know what to do necessarily. But you just need to empower people to go figure it out, move fast.

[B]

And Cody runs a massive beauty brand.

[A]

Oh really?

[B]

Jones Road.

[A]

Oh no. What?

[B]

Also switched to ramp recently.

[A]

Let's go. Congratulations.

[B]

Also way that he added value as a CEO.

[A]

But is he thinking about Jones bro? Adio Customer relationship Magic. Because ADIO is the AI native CRM that builds scales and grows your company to the next level. Tane says AI influenced ARR is the new community adjusted EBITDA. Because Adobe shares jump 6% as AI influenced ARR tops 5 billion. What does that mean? I pay for some Adobe products at various points. Premiere Pro After Effects. There are AI tools in there for rotoscoping Green screen. I think there's a subtitle product. We don't use a lot of this stuff. We kind of cobbled a lot of.

[B]

The challenge is when you have something. I assembled a PDF Community adjusted ebitda.

[A]

Yeah.

[B]

That can be applied to any business. Right. You, any CEO can figure out how to apply that framework and just say, hey, you're not properly evaluating how awesome our customers are.

[A]

Yeah.

[B]

And so AI Influence ARR could be applied to any business that uses ChatGPT.

[A]

Yeah. It feels like the, the, the name of the game or what was happening in like the Adobe CFO's head is like the market is rewarding AI revenue. And so I need to shift my massive SaaS revenue into a bucket that's called AI revenue. And I can just do that by reclassifying existing revenue lines every quarter and have kind of a new growth curve emerge for the AI influenced ARR. But it's very unclear if this is. If this is outcome based pricing. Token generation, image generation. Adobe has image credits.

[B]

There's an article in Barron's. Adobe stock gets downgraded. Generative AI isn't all it's cracked up to be. Which is painful because they're like, wait, we're actually just a SaaS company. So Morgan Stanley downgraded Adobe to equal weight from overnight cutting its price target to 450 from 520 due to AI concerns. So Morgan Stanley's concerned about AI from a competitive standpoint?

[A]

Yeah, I mean it makes sense.

[B]

So Morgan Stanley says it's still waiting for Adobe's push into Gen AI to bear fruit.

[A]

They came out with an AI image generator like right after Dall E like two years ago. And I have not seen a lot of adoption and they don't really have the distribution for it. I mean should weave its way into all the different Adobe products.

[B]

But it felt like they needed somebody like Scott Belsky. They had him for quite a while to be leading that effort and doing the deal making that even Alexander Wang is doing with Mid Journey and things like that. Doesn't. You just don't. We don't talk about Adobe much at all on the show. Right. Probably should if they were properly playing this market, making big bets. I'm surprised they had one of their.

[A]

Big pitches for their Generative AI product was that they were saying we have a ton of stock images, we will only train on stock images. We're not going to train on any copyrighted material. And basically every other lab was like, no, we're definitely training on copyrighted material. We think this is fair use. We're going to fight this and then we will pay the fines. And in the anthropics case, they're paying what, billions of dollars in fines for a very specific thing. But the specific fine they're paying is not because they trained on copyrighted material. It's because they acquired the copyrighted material and not an above board way.

[B]

According to Gemini, Adobe has only made one acquisition in the last 24 months. Company called Rephrase AI. Indian based AI video start startup. You would think they'd be more acquisitive right now.

[A]

Totally.

[B]

How many different gen AI.

[A]

Yeah.

[B]

Creative tools there are and how many of them just aren't going to make it? These are only marginally better than generating images and Nano Banana.

[A]

Yeah. Where were they during the Windsurf bidding process? Like where were they in these? Like, you know, they're letting these things get away. Yeah, but yeah, I mean that's the nature of being a big company. Risk averse. They don't want some massive lawsuit showing up. I mean, I mean like $6 billion.

[B]

Yeah, but they could be using the existing models.

[A]

I think so too.

[B]

Better. They could just add entirely new product lines. Try to scale those just through the existing distribution they have.

[A]

Yeah, yeah. I don't know, maybe Captions. Captions has done some good stuff on video editing.

[B]

On mobile, they just.

[A]

Yeah, there's a bunch of questions. On mobile, there's a bunch of questions. I'm sure they have a lot of.

[B]

Enterprise stuff to my. I haven't met somebody early in their career that's excited about AI, that's even talked about working with Adobe and so if they don't have the talent, they're not going to have that exciting product.

[A]

Yeah. Connor in the chat says their software acts like bloatware. Two thumbs down for Adobe for me.

[B]

Brutal.

[A]

They just integrated Nano Banana though. There we go. It's interesting.

[B]

Thank you.

[A]

Brown Paperback. Rune says I just walk around the research floor at OpenAI reminding everyone that they're propping up the world economy. It's good.

[B]

Couldn't be more true. How many you got? Broadcom. That's popping on OpenAI partnerships. You got Core Weave. That's popping on OpenAI partnerships. You got Oracle. That's popping on OpenAI partnerships. How many? We got a lot of important companies that are riding with OpenAI. SoftBank. Right. SoftBank doesn't get enough credit. The stock's up.

[A]

It is crazy.

[B]

That stock's up 118% year to date.

[A]

When the last batch of hyperscalers were birthed, there wasn't so many big companies that were like highly indexed or dependent on them. You know, like Google's growth. People weren't like, oh, if Google doesn't work out, like Microsoft is screwed. Or when Meta. When Facebook first started, it wasn't like, oh, if Facebook. It becomes like if Facebook stagnates, the global economy will fail. It was like its own thing off in its own world, building its own data centers and they were scaling, but it wasn't that crazy. Yeah, well, Rune's coming on.

[B]

Three mentions of Rune so far.

[A]

Well, how about a mention of 8sleep8sleep.com get a pod 55 year warranty, 30 night risk free trial free returns shipping baby.

[B]

I was away from home last night.

[A]

Oh yeah. I got an 80. So win by default.

[B]

Wow.

[A]

Not. Not amazing. Get back into it. I'm waking up.

[B]

Shouldn't be winning with an 80.

[A]

I know. Yeah, but win by default. Starbucks closing hundreds of stores and laying off staff as the chain continues to struggle. Is this post but the Community note says Starbucks is closing about 1% of stores which comes out under 2000. They plan to remodel old stores and open more. So story is developing $0.02, the social network for your net worth. Did a poll and Said, what's your main caffeine source? Who do you think won? Who do you think the richest said the most amount of value, I believe went to energy drinks. But the most amount of votes went to coffee and tea.

[B]

That's actually so bullish for energy drinks.

[A]

Yeah, the rich. The read on this is that the rich people are using energy drinks and the poor people are using pills. Caffeine pills. That makes sense. I mean the most economically efficient way to consume caffeine is through pills.

[C]

Right?

[A]

They're super cheap to buy caffeine pills. But no, caffeine had a strong showing as well. So pick your poison.

[B]

Energy drinks might be something.

[A]

This is a perfect transition. We gotta talk about Avi Schiffman out of home campaign. If you want to run an out of home campaign, go to adquick.com out of home advertising made easy and measurable. Say goodbye to the headaches of out of home advertising. Only Ad Quick combines technology, out of home expertise and data to enable efficient, seamless ad buying across the globe.

[B]

And as much as we love out of home campaigns. As much as we love ads probably don't do what friend.com the scale at which they've seemingly bought up all the out of home inventory in LA and New York.

[A]

I'll give you the numbers. So Avi shared. He came on our show and said he's doing the largest out of home campaign in history. And I was like, that can't be true. That's so crazy. What about avengers? What about GTA 5? There's gotta be huge campaigns that have been run. But it seems like he might actually be running one of the biggest campaigns, if not the biggest campaign. He said says 11,000 car cards. I don't exactly know what a car card is. 1000 platform posters. 130 urban panels, West 4th street domination. 100% print all five boroughs. It's the largest NYC subway campaign ever happening. Now we have seen it all over la. He said it was the largest billboard ad buying campaign of all time. And the ads and this is a crazy don't say a lot about the product. It just says friend.com and it shows you the pendant. And the pendant is kind of white on white.

[B]

Well, some of them do.

[A]

Jumping off.

[B]

So I'm seeing one from us. Their subway campaign.

[A]

Yeah.

[B]

It says and one thing is, Avi is pitching this as your new roommate.

[A]

Yes.

[B]

And the reason I think this is not there. There's the thing that's great about the campaign is how controversial it is. There's a lot of people that are going to be taking pictures just saying, look how dystopian this is, and posting it.

[A]

That's exactly what's happening. Someone defaced it and said, stop profiting off of loneliness.

[B]

But then one person that saw that's going to buy it. Or it could be more. Right. But Avi markets this as your new roommate is waiting. It's funny to me because people don't like roommates back in the day, but I like them even more now that I'm not sharing a space with them.

[A]

Exactly.

[B]

Living by yourself, your family is amazing. And I'm not personally in the market for a new roommate. But I'm also, I don't think the target market for this because I hang out with you for 12 hours a day. If we're not hanging out, we're probably on the phone talking about technology and business. But in the Subway campaign, they said it says, I'll never bail on our dinner plans. I'll binge the entire series with you. I'll never leave dirty dishes in the sink. So while he's saying it's your new roommate, he's implying it's better than your roommate because you're getting, I guess, the social benefits of a roommate without the dirty dishes in the sink and the bailing on dinner plants. Yeah, but people are not happy with this for one specific reason, which is that that I think Avi is very backlogged on orders. He's gotten a lot of orders. He's gotten a lot of orders over the last year since he started selling it. And this is not the first time he's going viral and getting a lot of attention. And he hasn't fulfilled a lot of them, it sounds like. So he was going back and forth with some people. People were saying a lot of marketing spend for a product that hasn't shipped orders from July of last year. And so, yeah, if you're somebody who just hasn't received your device yet, you're going to be kind of annoyed to be seeing this, that they're trying to sell a lot more of these and you haven't gotten yours. And I'm sure they're going order by order, trying to get through. Avi had said he'd sold 400 or he shipped out 400 in the last week, which is a lot of hardware. Yeah, but the question.

[A]

But that's like, how much are these? $99 or $200?

[B]

How much is $129?

[A]

$129.

[B]

But the main. I guess this feels like a AVI more or less said this on our show last time he came on, basically said that 50k of books. I'm going to do a crazy out of home campaign so I can raise my next round. That was, I'm paraphrasing that that was what I. That was my read on it because it feels like the product is still super early. The review from Wired was abysmal and I'm sure that you can make the argument that the people that reviewed it are not the target demo, but I haven't. The concern is that we're not seeing a ton of customer love out there. People are getting these and so telling the entire world about your product before it's ready. Billboard campaigns can work to get hype and early interest, but they're best when you have a product that is already loved and you're just able to get more eyeballs on it, more attention, more customers. And so getting a whole new wave of customers that are then going to be waiting a long time for their device and then potentially people that are getting them currently or potentially let down, it just feels incredibly risky. The other thing I was looking into is the website right now. I wonder how it converts. It's possible that it's priced in a way that it is pretty affordable in.

[A]

Terms of where devices play. People were shocked that the meta Ray ban displays are 799 and that's what, seven times as much money. But it does feel like it just takes a while to tell you what the product. Product is.

[B]

Yeah. The main thing is this kind of parallax scrolling.

[A]

It's a lot of like motion, a lot of physical like I gotta do a lot of work.

[B]

You can't hit order. You can't hit order from like landing on the webpage.

[A]

Yeah. I wonder if you go in the Wayback Machine, can you find what Apple website was back at the first iPhone launch? Let me find that.

[B]

I bet you there was a big buy button every. You wanna buy a buy button. No matter where you are on the website, you should always be able to hit buy.

[A]

Yeah, I mean that's like standard is like above the fold. Don't make them scroll. Okay. When did the iPhone come out? 2007.

[C]

Right?

[A]

2007. Let's see what the Apple website looked like in 2007 in October. Let's see. It's loading. It's loading. I'm so excited. Yeah, well, you know what a better piece of accessory might be something timeless.

[B]

Instead of a pendant. I don't exactly what you're Thinking about.

[A]

What am I thinking about?

[B]

Some hardware.

[A]

Some hardware. I think of an aquanaut as a friend, really.

[B]

And in some ways, it's super intelligence.

[A]

It is.

[B]

Can tell you the time.

[A]

Can tell you the time, tell you.

[B]

The date better than a human. Chronograph can track time better than a human.

[A]

It can. It's super human. Oh, yes, of course, if you're timing lap times on the track with a Daytona, that's going to be way more accurate than doing it in your head. So head over to getbezel.com your bezel concierge is available now to source you any watch on the planet. Seriously, any watch. Oh, no. The Apple website is not loading from 2007.

[B]

Well, the Apple website today is definitely best practice for selling consumer.

[A]

Yeah, it's tough to make a comparison to today because everyone knows what an iPhone is and they're on number seven. And so they can just say, it's an iPhone. It's the best iPhone. It's a new iPhone.

[B]

IPhone 17 Pro. You can hit the buy or you can hit learn more. If you hit learn more, you land on a new page with a cool video that says buy. It's just buy, buy, buy. So anyways, we'll see how this plays out. The challenge is, you know, having this visceral of a reaction even from tech insiders.

[A]

Yeah, there definitely is something to like kicking the bear and rage baiting a little bit and getting a bunch of somewhat negative attention. But then people will. There will be a small percentage of people that are like, I'm in. I want to try it. And that works. But it feels like at some point you have to flip into just like normal territory and kind of just speak to the product benefits. In January 31st of 2008, the Apple homepage. Do you know what was above the fold was right there?

[B]

What was it?

[A]

It wasn't the iPhone. It was the world's thinnest notebook, the MacBook Air. You have to click somewhere else to find the iPhone. January update. Let me see if they have some on page. They're just like, yeah, put that iPhone in the back. You know, yeah, we got some MacBook Air. I mean, the MacBook Air was beautiful computer. And certainly then they had the imac. The iPhone is really not on display here. It's store. The header is store. Mac, ipod and itunes. And then iPhone. IPhone is the fourth item in the bar, like the fourth most important if you rank them by importance. But I'm trying to go to apple.com iPhone in the Internet archive to See how they position it. And it's extremely wonky. Features Phone, ipod, Internet. The fourth feature. Can you guess the fourth feature of why you would want to buy the iPhone?

[B]

Alarm clock.

[A]

Nope. This is even better. You're going to love this. You're going to love this. You're going to love stock. Love this. No. So you got the phone, you got the ipod, you got the Internet. The fourth feature for why you would want to buy the iPhone. High technology. I'm not kidding you, dude. High technology.

[B]

What does that even mean?

[A]

Let's find out. I gotta click on high technology. It just has a image of like a laser or something. I don't know what it is. And then they show some maps and stuff. But what is this high technology that they're.

[B]

They need to bring that back.

[A]

Oh, high technology. It means that it has multi touch. It has wireless. It has an accelerometer and a proximity sensor.

[B]

That is high tech.

[A]

Multi touch was high technology at the.

[B]

Time that that feature set allowed you to drink beer. An animated beer.

[A]

The accelerometer was the reason that you could do that. That's right. That actually was high technology at the time.

[B]

This is high tech. We don't know how to make high technology.

[A]

We don't.

[B]

We got.

[A]

We were talking about a team and we were like, make the beer app. What's the beer up? And. And it's such a joke. Oh, the beer app. But the beer app is a moment that means that you've arrived and you have the crazy developers and people are making memes and they're using your technology in ways that you never could have imagined.

[B]

If you're not getting memed, you're cooked.

[A]

Yes, exactly. And Studio Ghibli was that. And we've yet to see that on the Vibes app. Well, what will the friend version of that be? What will be the becoming a platform. What will be the product or the use case that isn't imagined by the designer that winds up driving value, at least for fun, you know, like going to wander. Find your happy place. Book a wander with inspiring views. Hotel, great amenities, dreamy beds, top tier cleaning, 247 concierge service. It's a vacation home, but better. Better. Did you see the Fermi IPO? QCAP says the upcoming Fermi IPO will be the most 2025 IPO. All the red flags are there. Literally every single one. There's no rep. It has a $13 billion valuation. It's powering AI by 2038. There's a nuclear component. The nuclear Component will be called the Donald J. Trump generating plant which only could mean one thing. Buy.

[B]

I think this is going out on the NASDAQ Texas sized IPO requirements than nice.

[A]

It's a wild, wild time out there. Be safe, do your own research. And this of course is not financial advice.

[B]

We might be founded only nine months ago.

[A]

Nine months to get to 13 billion. Is that a new record? Nine months to get to 13 billion. Who's done that? He did it in maybe. Yeah, he did it in the coins world. Minutes. Yeah, it went out and minutes later was here.

[B]

So Fermi stock listing date nears for AI data center and power company backed by Rick Perry. Founded only nine months ago. The company plans to build a Texas based hypergrid. Let's give it up for hypergrid.

[A]

I like that.

[B]

And we become a provider of data and power centers. They got a cool render.

[A]

Intelligence is out. We're talking about hyper intelligence now.

[B]

As the rise of AI continues, companies operating in the space, relying on the technology are finding that they have two inextricable needs. Data centers that can run and process the AI and access to ample energy to power those vast data centers. One new company, Fermi America, props for putting America in the name of the company aims to offer solutions for both these needs. And this week Fermi announced plans for an upcoming initial public offering and dual stock listings. They're going to be listing in the UK as well. Fermi America is a very young company. It was only founded this year, just nine months ago in January 2025. The company is so new that its website is still a relatively bare bones affair. Pull it up. It just says yeah, it looks like a stock website from. Looks like they got a template of a website and doesn't even look like they replaced all the stock imagery. They just said.

[A]

They should have vibe coded something from scratch saying mixed bag with naming your company after Enrico Fermi, just after a famous scientist who's not involved. Obviously Tesla's done very well. Zach Weinberg has Curie bio after Mary Curie. Right. Nicola did not do well of course.

[B]

Yeah.

[A]

So Fermi Foundation.

[B]

Enrico Fermi was an Italian and naturalized American physicist renowned for being the creator of the world's first artificial nuclear reactor, the Chicago Pile one, and a member of the Manhattan Project. He has been called the architect of the nuclear age and the architect of the atomic bomb. So anyways, let me try to get a little bit more color here. Given the youth of the company, it's no surprise that the majority of Americans have most likely never heard of it. But they have heard of its co founder, Rick Perry, the former Texas governor who ran as a GOP contender for president in 2012. I was trying to think like Rick Perry. I know that name, but it's not in the the tech context. After unsuccessful presidential bids, Perry was appointed as the 14th United States Secretary of Energy during President Trump's first term in office. In addition to Perry, Fermi America was also co founded by Toby Neubauer, a former co managing partner of Quantum Energy. Fermi America intends to be a provider of data and power centers that other companies can use to host their AI needs. Needs. But I say quote intends to because Fermi America doesn't actually provide any services yet. Heck, it doesn't even have any infrastructure yet to provide its services. What Fermi America does have is a lease. Let's give it up for having a lease. A lease for 5,200 acres of land owned by Texas Tech University, which is where Fermi plans to build a hypergrid in an undertaking done dubbed Project Matador.

[A]

Just hitting the naming is fantastic.

[B]

What is Project Matador? Project Matador is the name given to Fermi America's hypergrid project. This hypergrid will be combined data and power center that other companies will pay to lease space on to run. So anyways, well they want to deliver up to 11 gigawatts of low carbon, hyper redundant and on demand demand power directly to the world's most compute intensive businesses.

[A]

I mean there's stuff all over the place. This is the nature of bull markets. You get all kinds of crazy outcomes. Sean U. Matthew is sharing a quote from David Einhorn, the hedge fund manager. He's warning that AI's trillion dollar infrastructure spending could destroy vast capital. Despite tech being transformative, he questions whether all the extreme spending by hyperscalers will deliver returns, saying there's a reasonable chance of tremendous capital disruption destruction, which we hate. I hate the idea of capital being destroyed. But there's still a lot to look forward to. There's still great products coming out every single day.

[B]

Like Vibes.

[A]

Join them Vibes. I think, I mean we discussed it a lot. We're going to continue discussing it with our next guest Rune. But I think that there is a way to have fun on Vibes. I'm already having fun with my promoted post, my SpawnCon. I came in there hot.

[B]

TVN ad.

[A]

TVPN ad. I say if you bring your own creativity you can probably have fun on there. But be careful because you might just wind up in infinite jokes world. Anyway, we have our next guest, Rune, the anonymous poster, the world famous Rune in the TVPN ultradome.

[B]

There he is.

[A]

Welcome to the stream. How are you doing, Rune?

[C]

Hello.

[A]

Hello.

[C]

I'm good, how are you?

[A]

We're fantastic.

[B]

At long last.

[A]

What I mean, it's been printing your.

[B]

Posts, we've been printing your. Months ago.

[A]

Months. So thank you for everything that you do for the timeline. Always providing insight and levity.

[C]

You guys mentioned me four times this week and I was like, I gotta come on just to run off the week, you know, it's time. Close it out. Yeah.

[A]

Jody and John.

[C]

Yeah.

[B]

And thank you. Thank you for walking around the. And you know, reminding the researchers that you guys are, you know, holding up the global economy.

[A]

Yeah.

[B]

It really is important work, what they're doing and also you to just make them remember.

[C]

I mean, I know you guys comment on this like all the time, but like most of the growth in the equity markets is just AI, you know, for the past like three years, there's like really nothing else. So, you know, you got to not make mistakes, got to keep pumping out great models.

[B]

Yeah, yeah. Please don't make mistakes. That's all. That's all. That's all we're. But yeah, I mean we, we talked about it earlier today. The number of companies that are now, you know, barnacles on the whale of OpenAI is the entire global economy is.

[A]

Sort of a barnacle.

[B]

Yeah. But you could even get, you could even get more specific and you know, the oracles, the core weaves, you know, you can make a really long list.

[A]

Yeah. How are you feeling generally? The timeline was extremely split this morning and yesterday. A lot of incredibly great vibes around ChatGPT Pulse, less great vibes around the Vibes app. Are you happy with the direction that the productization of AI is going? Are you happy with the rollout and the actual implementation of these models so far? How have you been processing the way these models are making their way into everyday lives?

[C]

Yeah, I don't know. Like, okay, the Chatbot medium is one thing. I think the Pulse is like a fine and good idea. It makes a lot of sense.

[A]

Yeah.

[C]

Vibes, I don't know but the thing I am manic about is Codex and you've probably seen me tweeting about it like three times a day.

[B]

But like more than that. More than that.

[C]

Yeah. Like Codex and I guess like Claude code are like genuinely amazing. Like, I don't want to like, you know, glaze company's products too much, but like, it's like I Don't understand how. I did software engineering before this. I have like 20 terminal agents open and like writing like five different things at once and I have an idea and I just like open up a new terminal and I just like get it working on it and it's like a completely different way of life than anything I was doing before for which is like you linearly slog through one script and then there's like three or four bugs and I don't know. It's clear to me that this alone completely upends the software industry and we're talking about massive data center capex across the industry and what is it going to be spent on? And even the inference alone for this kind of thing is first of all super expensive and then super in demand I think once it's really percolates through the economy. So yeah, I mean that's the thing that is fueling my mania these days. It reignited my fuel AGI fire.

[B]

Well yeah, so you said completely upending the software industry. What does that actually look like in your mind? Because one thing is if developers are getting karpathy was kind of like that's higher output.

[A]

Developers are going to be fine. A lot of other people have been like, like 90% of software engineering will be automated but Maybe we do 10 times as much so there's the same amount. Like what's the long term view for you?

[C]

Yeah, I really don't know about the jobs or whatever. It's like extremely non linear and hard to predict. I feel like there's so much I see the product managers at my company, they're like vibe coding useful apps all the time and like people who are not even engineers are becoming like programmers because of these things. And I'm sure that's the case for like the most junior level engineers who are like bright are like now they're like way better than they were before. I remember my first year of programming and I was like running to my mentor for like you know, like how do I run this whatever basic thing and like they just don't have to do that anymore. Like they have to, they can avoid all that embarrassment. They have this like amazing agent or like the ChatGPT itself or whatever it is. To me it would seem like the productivity of an entry level engineer is like just way higher than it would have been like five years ago. I can't you know, speak to jobs impact in like the two years or three years.

[A]

What about, what about in terms of like consumer impact? I feel like you still hear people Complaining about like my United Airlines app app is not as great as it could be. And it feels like in a world where software engineering gets way easier, more reliable, higher leverage, we should start seeing higher quality software diffuse through the everyday lives of people that interface with technology. Does it feel like that's happening, going to start happening? Or is there something sort of intractable about an airline booking app that just can't be solved?

[C]

Yeah, I mean like first of all I don't trust people's vibes on this. I don't actually think they're paying good attention how the United app looked like three years ago or whatever. I'm sure it's actually a lot better. I've seen the in flight displays on United and they're like, what is this? It's like space age. It's way better than it was before. And okay, I don't know that United engineers are using any of these tools. Maybe they're not, maybe they're companies doesn't allow them. Maybe they use like some third party source or. I have no idea. But you know, that's what I was saying. Of course these tools, I have no idea what time it takes for them to percolate through an economy or like when people in like the, you know, like the, the non Tech Fortune 5000 are using them or anything like that. But from everyone I've spoken to that works at like say like GM or Ford or something, they're like, what's Codex? You know, it's not there yet. Of course not. I hope it will be. But the fact that it even exists is just like exhilarating to me.

[B]

Do you think Codex has the potential to be a hit consumer product and get, you know, end up eating into the market share of some of these new Vibe coding companies?

[C]

Of course. But also I think they're very excited, even just the cursor people because I don't know there's competition between OpenAI and Anthropic and that's probably good for them. I don't know, it's hard for me to say but. But I'm pretty sure that curs is killing it and I'm very excited.

[A]

Yeah. How are you using the various products? It seems like you're using codecs all the time. Are you using codecs on mobile as well? Are you in sort of like wake up in the morning and you have an idea or like a shower thought and I find myself doing this with just GPT5Pro queries or deep research queries. I have some idea, I fire it off and I just wait. But I'm not writing a ton of software. Are you having a similar experience that I'm having with just knowledge retrieval in code generation where you might be firing it off? Like when you're in the bathroom and getting dinner with someone and you just have a second, you just fire something off?

[C]

No, I'm just like, not a hardcore enough engineer to be doing stuff like that. But I definitely use like 5 Pro or whatever all the time just to whatever random idea comes to my head. I wake up, I'm like researching some, I don't know, some statistic or like setting off some query like what percent of the economy is like software engineers?

[B]

Or how much do you think you personally cost? OpenAI monthly? Like, how much are you.

[C]

A lot, A lot? I can't say, but a lot, a lot.

[A]

What about other use cases? I mean, we saw that a lot of people are using these models for like, like life coaches and just people chatting. And I've never really gone down that path. I don't know, maybe I'm just busy or I talk to Jordy all day long. So I. I don't have a lot of room for that. I. I hit knowledge retrieval constantly. Like the GDP of some country. Yeah. Give me the full history of it. Build a block.

[B]

I did a deep research report yesterday on the capital of California. Thought. Thought for 59 seconds. Yeah, it ran a hundred searches.

[A]

I hit that stuff all the time. But I haven't gotten into any sort of flow of asking for interpersonal advice, life coaching, therapy, any of that stuff. Have you had any luck? Or is that just like a. Is that a skill issue or just personality or certain type of person that uses things that way. How do you think about that use case?

[C]

I've definitely used it in a tough corner, like, what decision do I make? And I have it discussed with me. Yeah, I don't know, I'm probably not a typical user because I read what it says and I'm like, this one seems kind of like bullshit. And this is like a great point and I like weigh it and ignore it and whatever. But yeah, I think actually New Sonnet was an amazing model for this. And this was what they called Claude 3.5 sonnet new. And I think that kind of invented this genre of like, I asked Claude for everything in my life and I think like the latest GPT models can also do this. But like, yeah, there was like memes going around about the, the Claude boys back then who just like, yeah, like literally everything they do, they'd Run it through Claude. Yeah, it's. It's a bit weird for sure. Like are, you know, are you being puppeted by an AI? It's a bit, bit loss of control. Vibes.

[A]

I feel like my interactions with AI are getting less back and forth over the last two years. I used to chat back and forth and ask a lot of follow up questions and I feel like now if I just talk to GPT5Pro on voice mode for a minute and give it a bunch of context, it's going to come back with a very thorough answer that I might have one or two follow ups on. But I'm not really. I certainly don't have this concept of like there's one chat that's like fine tuned at this point on this particular that I. Oh yeah, my economics chat. I go over here when I'm asking an economics question. Like everything is done at just like a new chat router level basically. Is that kind of mirror your experience?

[C]

Yeah, I do think the OpenAI models are built for that. It's like single turn, think really hard, get you the right answer type of thing.

[A]

What about the flavor of the responses? With the GPT3 API, it felt like if you prompted it with the green text, you'd get a green text back. And it felt like there was a lot of variation in the style of writing and, and now we're seeing. Maybe it's just the way people are prompting, but it feels like there's a little bit of a consolidation in the textual style, the stylistic flourishes. It's not that, it's this and the EM dash and all of that. It feels very much like if I opened up your app and asked you and asked the same question, I would get the same stylistic flourishes and, and I'm wondering if that's something that you think will continue or maybe that's just a temporary moment of time and we'll wind up in the future where someone's talking to that. The vibes of Claude 3.5 maybe in a completely different model, but they have those vibes over there and then someone has 4.0 vibes over here, but they're not talking to literal 4.0 or literal 3.5, but they've kind of landed in a local minima of this is the tone that they like for all of their interactions, actions.

[C]

Yeah, these are great questions. I mean like the, the GPD3 model to me was like my first field AGI, kind of like hair raising moment. I remember playing with it and just being like, oh, this thing's intelligent in ways that are like yeah, just like non trivial creative etc. But people are also like too rosy about like the base models. They just don't, they're like extremely random. They don't even have a concept of. We talk about hallucination but what the base model is doing is only hallucination. Maybe one out of 20 completions will have a coherent front to back sensical text text and there are people who can do magical things with it. But there's a reason There is no GPT3 moment and there was a chat GPT moment where it took off among consumers and same with Claude and whatever. Right. But there is at all these companies there's this, what do you call it, a path dependence of post training. So the models sample and you collect comparisons between model samples and you're like I shouldn't talk too much about post training but the point being you are very reliant on the previous generation of models to train your next model. And so there's some style burn in that goes on and the late claw resemble early clob and the late groks resemble early Grok and so on. And yeah, they all have some similarities probably because of like well I mean they're all training on the same giant data set of the Internet more or less and they also have like these common data sets.

[A]

I guess I just wonder if like there's demand from people for more variation in vibes. I was asking Mark Chen about this where like I was like everyone was saying GPT 4.5 had a big model smell better like flavor to it, better writer but deep research was the one that could get you all the results. And I was like so is the best practice for me now that I should like run a deep research report and then take that to GPT 4.5 and say hey, rewrite this better. And I'm wondering if there's some world where there's just demand for like I like a lowercase person to chat with. I like really punchy. I mean I guess you can kind of put that in the prompts but I'm wondering if that's like at some point going to be an emergent property.

[C]

I think you should try the five thinking model if you have some use case like that. Yeah, it's the same thing basically.

[A]

Yeah.

[C]

Does that work for you?

[A]

Yeah, yeah. I mean it's honestly like I'm fine with the. I like the flavor and the vibe that it comes back to me with. It's just, it is Impersonal. But that's exactly what I'm. But that's exactly what I'm hitting it with. And I want that. I want it to be an assistant. I want it to be a helpful assistant. I want it to be like a textbook. I want bullet points. I want EM dashes. I want clear analogies. And so I'm very satisfied. So I'm wondering if. I mean, maybe it's just a revealed preference versus stated preference, but it seems like there's a lot of stated preference out there that people are sick of EM dashes. And if it's not this, it's that. And it's certainly like a tell that someone used an LLM. But I was just thinking more about how if there really are people out there with the revealed preference, with the true preference reference for something, and will the models eventually be able to be many things to many people, show a different face to every single different.

[C]

Yeah, absolutely. I mean, I think the end stage is, like, extremely personalized. Models that speak to. You guess even before you even tell them what it is that you want the model to sound like. And I see no real reason that's, like, not possible.

[A]

Yeah.

[C]

Just a matter of doing it. And the. What else do you think?

[B]

Do you think memory will start to be, like, feel like more of a moat as products like Pulse get more popular? It feels like this is the first time that, you know, I don't think.

[C]

Can I ask you guys a question, please?

[B]

Yeah.

[C]

When I saw the first episode, it was called the Technology Brothers Podcast Network.

[A]

Yes.

[C]

What happened?

[A]

It was just technology brothers.

[C]

Okay.

[B]

And then we had a big problem, which was that. Which is that our community teapot and the tech community broadly honored that we were calling it technology brothers. But when journalists would write about technology brothers, they would introduce us as John and Jordi, the tech bros, and they would shorten it, which was massively disrespectful. And once we realized we were doing our life's work, we did not want to spend the next 30 years, everywhere we go, being introduced as the tech bros. And so, yeah, that's disrespectful.

[C]

If they had called you Jordy and John, the technology brothers, that's.

[B]

We would have kept riding with it.

[A]

But it was a big problem, the battle we lost.

[B]

But.

[C]

Okay, I believe you credited me with that one. You pull up a tweet on the first episode.

[B]

Yeah. You were one of the first people to truly expand tech bros into technology brothers. So thank you for popularizing the term.

[C]

What do you think you Change? Yeah.

[B]

What do you think of what Thinking Machines has been putting out lately?

[C]

I think it's really cool that, you know, like, new labs, they're incentivized to speak very openly about their research because, you know, like, it's a great recruiting mechanism. You're like, there's some cool research you're doing and like, let me come join. And then like, I don't know, maybe, I guess. Facebook was pursuing this strategy a while back or meta or whatever they're called now. They open source everything. They write a lot of public papers and it was a big draw for people. I haven't read the latest paper, but I read the last one. It was something I had been wondering about for a while and was a common question in the industry about determinism and non determinism and so on. Yeah, it's good stuff. It's good stuff.

[B]

It's good stuff.

[C]

I don't know, like, it's great stuff. I don't know specifically what they're working on. I've heard rumors, but I won't, I can't, I shouldn't talk about it.

[B]

What about hardware form factors? You know, speak generally. You don't have to talk about what you guys are working on internally. But what's exciting to you in hardware broadly? Do we need new dedicated devices or is the phone good enough?

[C]

I don't know. I don't have the answers to that. I like, genuinely don't think about the hardware side very much.

[B]

Well, that says something in itself.

[C]

Yeah, I mean, I'm pretty happy with my chatgpt on my phone, but what.

[A]

Was on your pulse today?

[C]

I don't think I checked. You say a check right now?

[B]

Check right now.

[A]

Mine was shockingly interesting.

[B]

Pulse is a window into the mind.

[A]

It was extremely interesting talking about bottlenecks and compute revolutions and retired coal plants powering AI campuses. And it even had one. I was searching for different options for bonded cellular networks. So if you want to bond a Verizon and AT&T Sim card together to get a really powerful WI fi hotspot on the go for live streaming. Obvious, obviously. And it had a whole deep dive on that. And it's like, yeah, I'd search for that, but I would love to have that resurface to me in a different format.

[C]

Right.

[A]

And so I was very pleased with the results. And it also felt like deeply. Maybe it's just like the way I've been using the product, but it felt deeply. Anti brain rot. It felt like, oh, these are articles. I don't even know what you would call them, but summaries that certainly push me to go deeper into something that's actually hard work learning as opposed to something that's more like candy and helping me just kind of tune out things. I, I, I enjoyed that.

[C]

Look here, here's like mine is just exactly what you expect. It's like tracking AI's impact on power markets. It's like Chinese infrastructure spend. It's, it's all this kind of stuff.

[B]

Yeah, it's, I mean a very predictable, it's very predictable.

[A]

Now I want to know, Share that article with me. I want to know about Chinese infrastructure investments and how they're tracking with the.

[B]

Well, maybe that's a feature. Rune shared his pulse today. I think so check it out.

[A]

I mean you can already share the link, so yeah, you'll need to be able to share it in the app anyway. What else is keeping you up these days? You were talking, you kind of laid out. We reviewed the post on the show just talking more broadly about the nature of process power and Dan Wang's book. Have you, where do you sit on the general characterization of the lawyerly society in America? Does it feel like we are becoming more lawyerly or less lawyerly?

[B]

Well, some prominent engineers, founders in AI have certainly gotten litigious.

[A]

Yeah.

[B]

So maybe America could do both at the same time.

[A]

Many engineers are crossing over into the government.

[C]

Yeah. I guess what is interesting to me is like the extreme depth of the capital markets in the US where like one company can raise hundreds of billions of dollars dollars on infrastructure spend. And you know, together the industry is spending. I have no idea. You guys probably know the numbers better than I do and that's like, that's very interesting. So where are the like, you know, if it's like such a lawyerly society, why are we able to do this? Like why are we able to spend so much? And of course it's true, like if you look at like the, the west side of San Francisco, it's like it should not be this like rows and rows of single family houses. It's like quite a waste. It's quite a shame. But there's many things like that that people have talked about ad infinitum. But in general it's like okay, like Elon Musk showed up and he revolutionized two industries, right? Like the space travel and the electric cars. And I'm sure he fought with regulators and like there's a bunch of lawsuits and whatever. But like you know, his, like his, I guess his will to power or whatever made it work. And so I am always a bit skeptical when people say like, you know, the US is impossible to innovate in because of regulations or it seems like it's getting done, people do it. Yeah. You know, maybe like the, the highest level of government is more lawyerly in the US than China, but I don't. There's also benefits to that, which Dan Wang talks about, you know, like there's no. I don't want to get into it, but I think you guys know the upsides of the great rule of law and whatnot.

[A]

Yeah. I do wonder if. Yeah, I mean that's kind of the odd counter take is the, that maybe the depth of the financial markets is actually an outgrowth of the lawyerly society. When you have a bunch of lawyers who can create robust contracts where somebody feels comfortable parting with $10 billion, $100 billion. Is that a function of engineering? It might be a function of financial engineering in some ways, but it is also a function of strong legal contracts and the work of lawyers to be creative, but also fair in a way that instills confidence in the capital market such that a company's willing to. That seems right to deploy $100 billion and know that there's a, at least there's an understanding of the pathway to the return on investment and that there's less uncertainty.

[B]

Well, on the other hand, you have TikTok allegedly selling for 14 billion.

[A]

Sure.

[B]

Which I'm sure is less than what was spent to grow TikTok in the.

[A]

U.S. i don't know.

[B]

But that's, that's more government.

[C]

I feel like it's worth a couple more zeros than that, but yeah, just a couple.

[A]

It does seem.

[B]

Do you think the Gartner hype cycle is being over applied in the context of AI? Do you think we could just have a straight line acceleration and there's no trough or are you trough pilled?

[C]

I just don't get this thing where people like overlay graphs and they like do this like high level pattern matching when it's like, okay, let's look at the basics, which is to me the basics are codecs, which is like literally as of two months ago this was not a thing thing and now it's a thing and it's like to me a mind blowing product. Just I can't stress this enough. And okay, like that alone is going to create just enormous revenue streams for both OpenAI and Anthropic, I assume. And cursor. So like, I don't know, I just don't it doesn't seem worth. Yeah, of course, yes.

[B]

Will it create the whole world? Is will it trillion? Will it create it. So will it create a trillion dollar revenue stream?

[C]

Potentially I don't think it needs to create a trillion dollar revenue stream to keep up with the data center plans. But that's not for me to say. You should get on the CFO or something of one of these companies and.

[B]

Ask them what about. Give us your read on the state of open source today. OpenAI's open source models were demanded, they were released and then you don't hear, you know there haven't been a lot of noise surrounding them, but certainly a lot of, a lot of noise coming around. Some of the Chinese, Chinese open source models. But what's your kind of high level thinking?

[A]

Open source models?

[C]

Yeah, yeah. Well there's two takes on that. One is that people demanded it but they didn't actually want it. The other is that maybe the OSS model just wasn't meeting the market needs or like it was not good at whatever it is the market wanted. I don't know, maybe it's a stated.

[B]

I mean to me it's a stated versus revealed probably at least for a lot of the people that were most loudly demanding open source.

[A]

I mean this has happened.

[C]

I will say some of these. The Chinese Kimik 2 model is like really quite excellent to me. It's a great writer. Yeah, it's like the model I compare against most internally. It's kind of cool.

[A]

I mean people have been demanding open source versions of technology for a long time. They were open source social networks and open source. There's an open source version of Shopify and yet Shopify accrued all the value and no one's forking that particular Ruby on Rails project just to get their E commerce store up and running.

[C]

I think there's like two problems with the kind of hype around open source. One is like okay, you've open sourced some models but it's also clear that that doesn't immediately make it accessible to a billion people.

[A]

Totally.

[C]

Obviously the ChatGPT product and whatever else, GROK is hugely downloaded, Gemini, whatever. But it's by far the easiest way for someone to access AI and for free, they're doing it for free. So on that angle it's not really all that decentralized just because the model weights are flying around. Second of all, if everything was always open source, you guys know as well as I do that it's like whole industry would never support this kind of capex and This R and D spend and whatever. And so there's no way that that's just the only way to do things. And that's, that's, you know, just. Just normal capitalism, you know.

[A]

Yeah. Question from the chat. Did you read one piece?

[C]

No, I have not. I have a bunch of friends who swear by it, but I've never, never gone into that.

[A]

Are there any other cultural influences that you pull from favorite movies, books, comic books?

[C]

Oh, yeah, of course, all the time.

[A]

I mean, top of the stack back.

[C]

My favorite show ever is probably this anime called Evangelion. Oh, yeah, it's like a 90s TV show. It's like kind of grim, but like, in like. Is this the staple of like the Mecca fighter genre?

[A]

Yep. Are there, Are there metaphors from. Is it Evangelion or Evangelion? Is there a metaphor?

[C]

Real heads call it Evangelion. Yeah.

[A]

Evangelion. Yeah.

[C]

It's because that's like the way the Japanese say it.

[A]

Yeah. Is there are, are there themes in that story that people pull from and apply to like the modern world and they go, oh, this is just like Evangelion. Evangelion.

[C]

I certainly think there's like a dense religious metaphor in there for like, I don't know, it's like a classic coming of age tale where Bill Dungzerman. Exactly, that's what they call it. And this kid, he's kind of like a depressive anti. Like he's got none of the good qualities of a protagonist. He's like constantly crashing out and he's like, there's a bunch of girls who are into him and he's like not capable of talking to any of them. And he, you know, but he's like magically like the best guy in the world at piloting these robots to beat these aliens. And like humanity keeps calling on him even though he's like being mentally destroyed to come combat the aliens. And it's a lesson in the story of Alec. Yeah.

[B]

If you can crash out and not talk to women, but if you're good at. If you're good at driving the robot, history will remember.

[A]

Remember your name for sure.

[C]

Totally.

[A]

What's the best way. It started with Evangelion.

[C]

Start with episode one. There's only.

[A]

Just watch it. It's not like, oh, you got to read the books first, that type of thing. You know how people will say that about, like, don't watch the movie until you see the. Read the books or whatever.

[C]

If you want to get really pedantic anal about it. There's like a different Subtitles that came out in the 90s that are better than the Netflix subtitles. But like, it's not really that important. You know, it's a pretty big barrier to entry. There's 26 episodes in a movie. It's pretty straightforward. It's not like something crazy watch order and whatnot.

[A]

Okay, maybe check it out.

[C]

Yeah, yeah. It's a bunch of media I love.

[B]

What's the current state of the talent wars from your view? Did we peak over the summer or is it still just as intense? What's your read on it?

[C]

I mean, it seems like, okay, if a researcher is a manager of compute and the compute numbers are growing and then like the effectiveness of a researcher is more and more important, then I don't know, like that line is trending up into the right. As far as I can tell. But do you guys agree? Does that logic make sense?

[A]

I don't know, 100%.

[B]

I think the logic makes sense.

[A]

I just think it's interesting that we're seeing essentially the same thing play out that we've seen for, I don't know, for generations in entrepreneurship where a founder can marshal capital and have leverage and incredible economic upside. Now we're seeing founder level wealth and financial outcomes applied to employees, which is just a different thing. It used to be in order to get the big, big numbers, you had to go build something externally and then the board would justify paying you a bunch to bring it in house. And now it feels like companies are more. They've actually just cracked the code on, hey, come work here. Build this thing that's going to deliver a bunch of value and we will pay you like you're a founder that was out in the wilderness and we acquired in and it feels like it's. In some ways it always should have been this way.

[C]

It feels like an Aqua Hire to me. Like just some sort of distributed Aqua hire.

[A]

Totally. Yeah.

[B]

That was my framing for it. You just have to look at it as an unauthorized distributed aqua hire.

[C]

Yeah, it's like, okay, and like, how much did Zuck pay in total for all these researchers over the course of four years? Is it like that enormous? Is it like, I don't know, like.

[A]

Certainly not relative to the market cap. Certainly not relative to the market cap cap. And especially if you just think about, like, even if you take the boring view of like, AI is just another business line that they need to be in. You know, it's like, oh, they're, they're spinning up a new division. It's like, well, if the new division drives 10% of their revenue or 20% becomes a significant business line. Investing 1% of your market cap makes a ton of sense to actually just like be in there. There's way more complicated dynamics there. But. But even if you just saw it as like we need an infrastructure team, we need a database team, we need a front end team, it's like you need an AI team and you're going to have leverage in this case. What do you think about this Rich Sutton take that's going out just dropped on the Dorcas show. LLMs are maybe not bitter lesson pilled. This idea that we need an entirely new format. People have been saying this for a very, very long time. Deep learning is going to hit a wall. Then we come up with something new. What are your thoughts on the father of reinforcement learning? Potentially not thinking that LLMs are bitter lesson pilled.

[C]

I think this is a common sentiment like DeepMind that they were always like they kind of held language models with a bit of like this is icky. This is not pure because we're not starting from scratch. We're learning from the human Internet. And that's like, that's not, you know, it's not pure in some sense unlike say Alpha 0, which is learned from nothing. Like you just self play yourself to superhuman performance. So yeah, like maybe it's not completely the most bitter lesson pilled paradigm in the world but like I don't know. That doesn't mean to me that it won't like reach extremely useful levels of intelligence. And you should ask Ilya.

[A]

It already is really useful. Like yeah, it's like the question of like will we get value out of this is like answered by Daus and just a million anecdotes of people that use the products in my, in my opinion, Jordy. Anything else? Oh shit.

[B]

I got a lot more stuff but we're over time.

[A]

Yeah. Thanks so much for coming on the show.

[B]

Glad we glad we finally made this happen. Thank you for posting the posts that make this show possible. We truly enjoy covering everyone. John, you want to hit the gong for Rune?

[A]

Yes. Congratulations.

[B]

Can't get out without a gong on it.

[C]

What an honor.

[A]

250,000 followers on X I believe is roughly the number. Something like that. It's definitely gong worthy.

[B]

There we go. Heading it. Good stuff Rune.

[A]

Thanks for coming.

[B]

Thank you for your thoughts and your posts and have a wonderful Friday.

[A]

Have a great weekend.

[C]

Thanks for having me on.

[A]

We'll talk to you soon.

[B]

Cheers. Good stuff.

[A]

Let's go back to the timeline, run through Some posts before we get out of here. Everyone in this company was given a whoop and is required to share their data. This seems like something we would do with eight sleep is this. I used this to assign more work to those who are sleeping a little too well. You are a healthy military aged man. Why do you need eight hours of sleep? Because you need a high sleep score. But this is a lot of fun. I believe Omar Waseem. I feel like I've met him. Anyway. Where should we go? There's some big posts. What is up with this bagel? The pop up bagel shop. Do you know about this? Jordy Chris, A big foreigner says I'm calling it right now. This will be the next billion dollar franchise brand. The next crumble. They're copy pasting the crumble playbook. A playbook I'm not familiar with. 300 plus franchises sold in the first month. 2.2 to 5 million gross per store with 25% net margins after fees lines out the door. So I'm not familiar with the Crumble story.

[B]

This feels like some Crumble cookies is probably the breakout franchise of the last five, 10. I think more like five years. So they sell cookies all over the country now. Franchise model. So I think they maybe operated their first few stores. They now are like very focused on the franchise model. They have cookies that are probably like the least healthy thing that you could put into.

[A]

They're like really over the top. A whole cake in a cookie basically.

[B]

Yeah. Really dense, super high calorie. I'm sure they're super tasty all over the country now. And. And they just.

[A]

It sounds like a unique business model where they've gone, they've. They've cracked the code on scaling retail through franchisees very soon. You always hear about like McDonald's and Burger King franchisees.

[B]

Yeah. Franchise. If you can have a durable hit consumer franchise, it's one of the most amazing businesses in the world. You get paid some percentage off the top.

[A]

Yeah.

[B]

So it's not a percentage of profits. You're getting a percentage of gross revenue in exchange for giving somebody access to your brand and like menu and supply chain.

[A]

It is valuable and they sort of have an entrepreneurial outcome as well. Like there's, there's plenty of like there's a guy who owns a couple McDonald's and does very well and has a boat.

[B]

Yeah.

[A]

And then there's also Patrick o' Shaughnessy profile the company and they're some street advisor company that does roll up of them and has done very well as well, yeah, yeah.

[B]

Tons of examples of people making a lot of money in franchises. You can start with, you know, working in one, then you can transition to being an operator, then you can add more, you can buy more. American dream, roll them up and, and they trade it pretty, you know, if you have a bunch of Burger Kings, they trade at a great multiple because it's a durable brand. And, and, and so Yeah, I think 300 franchises is crazy because they're getting a ton of upfront revenue and then they'll get a percentage of the gross over time.

[A]

Look at this post from Jeremy Giffon. He says staying below a $10 million deal size for 25 years is remarkable. The real lesson from Mark Leonard, the founder of Constellation to me is of maniacal discipline. And so the value of acquisitions was, what is this? Average price per acquisition was wow, down at 5 million for 2007 all the way through 2012 and then still really low for a decade and paid a little bit more in 2013 and is now starting to pay more. But really just people think about these private equity software roll ups. The software company sells for, for we see the marks. People raise $20 million, they sell for 100 million, 200 million. That's the one that gets the attention. But there are so many companies that are selling for $10 million, $5 million, $20 million, and Constellation's picking them all up and has built a fantastic business integrating all these. You have to have some serious infrastructure to do something with that $5 million business that you just acquired. But certainly good coverage.

[B]

Buco Capital has a great post here from Intel. Apparently was up another 19% yesterday.

[A]

Wow.

[B]

It's up another 20% in the last five days.

[A]

I mean, do you know the story here? So intel reached out to Apple and TSMC for investment as well. And so Nvidia wound up just kind of getting in early as Nvidia was like we're going to do something.

[B]

USA got in early.

[A]

The USA got in extremely early.

[B]

Read through that. I'll be right back.

[A]

Yeah, sure. So the news is that Apple and T TSMC might be playing ball. Might be open up the checkbooks for intel efforts to gain tech backers accelerate in the wake of US 10% stake in chipmaker of course Donald Trump and the US government announced a 10% stake in Intel. Intel Chief executive Lip Bhutan has been hustling to secure investments and customer commitments needed for the chip makers comeback. Among the company's intel has approached about investments or manufacturing partnerships are Apple and TSMC Taiwan Semiconductor Manufacturing. Those Efforts were already underway before President Trump showed up, showed an interest in the company last month, but have gone has since gone into overdrive since the US took a 10% stake in it. According to people familiar the Trump administration has been using its influence to help revive Intel's sagging fortunes. Commerce Secretary Howard Lutnick and other in the administration have for months been urging tech companies to work more closely with intel, which was for the long for was for long the world's largest semiconductor firm before it losing its lead to TSMC. SoftBank made a $2 billion investment in Intel. Nvidia made a $5 billion investment in intel and is now. And that it also includes a provision for intel intel to design new hardware to integrate with Nvidia's chips. Masayoshi Sohn and Jensen Huang have both been demonstrative in their support of the administration's technology agenda, which includes building up domestic semiconductor manufacturing capacity and supercharging the construction of data centers needed by the artificial intelligence industry. Bloomberg earlier reported on Intel's overtures to Apple. Tan earlier this year met with Tim Cook and has also spoken to CC Way, his counterpart at tsmc, about a partnership or joint venture, according to people familiar with the matter. And so basically there's a question about how much can Apple even get out of Intel. Apple of course moved off of intel chips and is now going direct to TSMC with the Apple Silicon project. And Apple doesn't, doesn't pay a lot to intel, if anything. I mean they sell like one Intel Mac still. It's like 10 years old, refurbished. It's like a very slow computer. But of course there are still things that Apple needs Intel chips for in the server and in to run their data centers. Like every company is kind of a customer of every other company, so there's probably something to be done there. But the question of like how would an Apple intel deal actually work is way less clear than an Nvidia intel intel deal because Nvidia sells the Hopper chip. But then they sell Grace, which is the cpu. They call it Grace Hopper.

[B]

And you got. And you got Donald Trump who led the Series A.

[A]

He did.

[B]

So now intel is trying to raise the B and the Series A lead is going to look at potential B leads and say hey, come on.

[A]

It's kind of like that. I think the better analogy is more like the series lead is going to the Series A lead and saying because Apple and Intel grew up together, Apple had like Apple exists because intel exists. Apple would not have been able to build that MacBook Air that we saw from 2007 without Intel. There was an intel chip in the Mac for a long time. And so Apple's been a beneficiary.

[B]

It's a center in the marketing.

[A]

Exactly. Intel inside.

[B]

You're buying this new computer because.

[A]

Yes. And so Apple has gotten so much out of Intel. They're so rich, they have deep pockets, they have tons of cash. And so the Series B lead is saying, hey, I'm gonna do a bridge round here. But the Series A lead's gotta come in if they wanna maintain their position or they're getting crammed down. Crammed down in this position in this scenario is getting crammed down in tariffs or getting crammed down in taxes or some other thing. It's all part of this art of the deal craziness. But it does seem like if you're Apple, you say, hey, there's a way to extend an olive branch to the administration with their technology agenda. They want to bring back semiconductor manufacturing in the United States. We're just going to see this as an investment. We could put our money in treasury bills, we could put our money in intel stock. Let's put some money in intel stock. It makes everyone happy. And so that might be the wind that might wind up being the nature of this deal. But then at the same time, it's totally possible that Apple wants to do some sort of new chip design and they want to go to intel and they want to fab it with intel and they, they find a way to do that. There is a world where intel becomes a manufacturer for Apple silicon chips because they are trying to get their most cutting edge node off the ground. They don't have customers. This is the famous thing that Lipp Bhutan said on that quarterly earnings call that panicked everyone. He was like the big thing that we've been planning. No one wants it. And so that raised alarm bells and that's where all this kind of came from. And Apple could potentially go to them and say, hey, yeah, the next iPhone chip or the next Mac chip will work with tsmc. We're happy with tsmc, but we'll also work with you, intel on your new fab that gets up and running if you can deliver at these levels. And how are they going to deliver on that level? They need a lot of cash. So we're also putting up the cash. And so another sort of circular deal, but you know, the administration is kind of like greasing the wheels of all these deals to hopefully get intel back on its feet and restore American semiconductor manufacturing.

[B]

Feels like a World War I. Leopold's printing.

[A]

In the meantime, Leopold is printing, so congratulates to him. What else is going on? You shared a story about a Radio Shack Ponzi.

[B]

Well, this was Tai Lopez.

[A]

Oh yes, I'm fluent.

[B]

He did something. He managed to raise like hundreds of millions of dollars for the new Radio Shack brand. I think turned into crypto.

[A]

He was buying all sorts of interesting things. Revs. So Tai Lopez is famous for the Here in my garage guru video that was sort of a book club, I think at the bottom of the funnel. Sort of a how to get rich educational, inspirational, hustle type genre of social media influencer. Huge buyer of YouTube ads. In the early days, everyone saw that Here in My Garage video. But he went on a spree. Rev's primary business was purchasing distressed retail companies with name brand recognition and converting them to e commerce only businesses. The securities offering that at issue involved eight Rev portfolio companies. Brahms Dress Barn, Franklin Mint linens and things which I have heard of, models Pier 1, which I've heard of and Radio Shack, which I've also heard of and I've never heard of Steinmart, Franklin Mint, Radio Shack. This is of course from Matt Levine, one of the greatest writers on the Internet these days. Not Bed Bath and Beyond. Not Gamestop. Stop. Bed Bath and Beyond sure was distressed. It filed for horrific bankruptcy in 2023. But before it raised a ton of money from retail investors and absolutely doomed at the market offerings, as its bankruptcy lawyers later put it. It raised that money as part of the meme stock movement started and fueled on Reddit boards and social media website because quote, it checked the two boxes needed to become a meme stock. A troubled financial situation and nostalgia value. And because it checked those boxes, Bed Bath was able to go out to ordinary investors and get them to shovel money into a furnace. GameStop was also part of that movement. Where is Gamestop trading these days? Gamestop was that movement because it was troubled and nostalgic.

[B]

If you're nostalgic and troubled, we got.

[A]

A stock for you.

[B]

GameStop up 18% in the last month.

[A]

No way.

[B]

But down 13% year to date.

[A]

What's the market cap's?

[B]

11 billion.

[A]

That's huge. I feel like that's a lot.

[B]

I mean for such a troubled.

[A]

I mean Bed Bath and Beyond is bankrupt. So it's a zero on a relative basis. That's the meme stock that, you know, had some value for a medium amount of time. We'll see. But it did so well out of being a meme stock that it's no longer no longer particularly distressed. GameStop, as you saw, is going, but it's definitely pinned to E Commerce. My point here is that if you go to retail investors and say, hey, let's teach those fat cats on Wall Street a lesson, we're going to revive Radio Shack. But with E Commerce, you're tapping into some deep and obvious patterns and you will undoubtedly raise a bunch of money that you can steal it. The quote at the top is from a securities Exchange Commission enforcement complaint against Tanio Adrian Lopez, Alexander Mayer and Maya Birkenrode, Bloomberg News News reports a company that snatched up distressed brick and mortar brands, including Radio Shack, Pier 1 Imports and promised investors big returns in a pivot to E Commerce. Operated like a Ponzi scheme.

[B]

Yeah. So they actually raised debt and they were offering a 25% preferred return.

[A]

Yeah.

[B]

And anyways, there were a lot of red flags.

[A]

It's pretty crazy. They put out promotional videos that bragged that the portfolio companies were quote, on fire, which the regulators did not like. It says in the complaint that while some of the retailer brands generated revenue, none generated any profits. At least 5.9 million of the returns distributed to investors were in reality Ponzi like payments funded by other investors. Now, so there you framed it as debt. It sounds like the they were equity investments that were paying some sort of fake dividend. It wasn't.

[B]

I think they were paying a fake dividend on the debt.

[A]

Oh, really?

[B]

It was like, yeah. So a lot going wrong.

[A]

So they accused the defendants of misappropriating approximately $16.1 million in investor funds which were diverted to defendants Lopez and Mayor's personal use. That's not good. The complaint also explains the business strategy, least the fundraising strategy.

[B]

So Rev's primary business was identifying distressed companies with name brand recognition, raising funds from investors in order to purchase the brand's assets and converting them into successful E Commerce only businesses. In one promotional video publicly available on the Internet, Tai Lopez touted Rev's business approach as, quote, one of the best strategies you can invest in.

[A]

That sounds like financial advice. I wouldn't do it myself.

[B]

Seems right. So anyways, Matt Levine finishes off quote, a troubled financial situation and nostalgia value, two best things you can have to raise money from individual investors in the 2000s. So I think enough people are getting burned on these now that I think people will learn their lesson.

[A]

Yeah, hopefully. But history repeats, or at least it rhymes. Well, it's been a fantastic show.

[B]

Great show. It is now May. It's close to the weekend.

[A]

It is.

[B]

Whatever you're doing, have a great time. We will be back Monday. Be back Monday for another episode.

[A]

We'll talk to you soon.

[B]

Have a great evening.

[A]

Have a great weekend.

[B]

Cheers.

[A]

Bye.