TBPN - Marc Andreessen & Ilya Sutskever Host TBPN, Mag 7 Earnings Recap | Aydin Senkut, Dean Ball, Shreya Murthy, Lulu Meservey & Gaby Goldberg, Jakob Diepenbrock, Shishir Mehrotra, Karri Saarinen

Episode Date: October 31, 2025

(13:04) - Mag 7 Earnings Recap (37:12) - 𝕏 Timeline Reactions (54:11) - WSJ Mansion Section (01:04:37) - 𝕏 Timeline Reactions (01:13:12) - Lulu Meservey, founder of the strategic communications firm Rostra, Lulu previously served as CCO & EVP of Corporate Affairs at Activision Blizzard and was VP of Communications at Substack. Gaby Goldberg is a communications professional at Rostra and former investor at The Chernin Group, where she helped build its early-stage crypto practice. (01:27:08) - Dean Ball, a senior fellow at the Foundation for American Innovation and former Senior Policy Advisor for AI and Emerging Technology at the White House, discusses the evolving landscape of AI regulation, highlighting the interplay between state and federal initiatives, particularly California's proactive stance in the absence of federal action. He emphasizes the challenges in defining and governing superintelligence, noting the lack of scientific consensus and the complexities of global policy coordination. Additionally, Ball addresses the potential for a tech backlash driven by public concerns over job displacement and rising electricity costs associated with AI advancements. (02:08:19) - Aydin Senkut, founder and managing partner of Felicis Ventures, began his career as Google's first product manager, launching its initial international sites. In the conversation, he recounts his transition from Google to venture capital, highlighting early rejections from firms that doubted his potential as an investor. Undeterred, he founded Felicis Ventures, achieving significant successes with investments in companies like Shopify and Meraki, and emphasizes the importance of resilience and belief in founders when making investment decisions. (02:34:37) - Jakob Diepenbrock, General Partner at Discipulus Ventures, discusses the recent demo day held in El Segundo, California, where nine companies presented innovations across various sectors, including energy, manufacturing, and software. He highlights a shift from defense-focused startups to ventures addressing broader industrial challenges, emphasizing the potential for disruption in legacy industries. Diepenbrock also notes the importance of building a supportive ecosystem for young founders tackling critical problems. (02:47:36) - Shreya Murthy is the co-founder and CEO of Partiful, an event planning platform launched in 2020 that has gained significant traction among Gen Z and millennials. In the conversation, she discusses the platform's rapid growth, particularly during Halloween, which she describes as their "Super Bowl," and addresses competition from major players like Apple, emphasizing Partiful's focus on creating fun and engaging event experiences. Murthy also highlights the importance of brand identity in distinguishing Partiful from competitors and shares the company's vision to become the leading platform for real-world social connections as traditional social media shifts towards entertainment. (03:03:24) - Shishir Mehrotra, CEO of Superhuman, discusses the company's recent rebranding from Grammarly to Superhuman, reflecting their commitment to empowering users with AI-enhanced tools. He introduces "Superhuman Go," a new AI assistant platform that extends Grammarly's technology to integrate proactive AI assistance across various applications. Mehrotra emphasizes their focus on embedding AI seamlessly into users' workflows, distinguishing their approach from other AI services that require user initiation. (03:10:22) - Karri Saarinen, co-founder and CEO of Linear, discusses the evolution of software development workflows in the context of AI advancements, emphasizing the need to assess and adapt existing processes rather than discarding them entirely. He highlights Linear's commitment to assisting companies in this transition by integrating AI capabilities, such as delegating tasks to tools like GitHub Copilot, enabling non-technical team members to contribute more effectively. Saarinen also reflects on Linear's design philosophy, focusing on functionality and pragmatism to inspire quality and craftsmanship in the software industry. TBPN.com is made possible by: Ramp - https://ramp.comFigma - https://figma.comVanta - https://vanta.comLinear - https://linear.appEight Sleep - https://eightsleep.com/tbpnWander - https://wander.com/tbpnPublic - https://public.comAdQuick - https://adquick.comBezel - https://getbezel.com Numeral - https://www.numeralhq.comPolymarket - https://polymarket.comAttio - https://attio.com/tbpnFin - https://fin.ai/tbpnGraphite - https://graphite.devRestream - https://restream.ioProfound - https://tryprofound.comJulius AI - https://julius.aiturbopuffer - https://turbopuffer.comfal - https://fal.aiPrivy - https://www.privy.ioCognition - https://cognition.aiGemini - https://gemini.google.comFollow TBPN: https://TBPN.comhttps://x.com/tbpnhttps://open.spotify.com/show/2L6WMqY3GUPCGBD0dX6p00?si=674252d53acf4231https://podcasts.apple.com/us/podcast/technology-brothers/id1772360235https://www.youtube.com/@TBPNLive

You're watching TBPN. Today is Friday, October 31st, 2025. It's Halloween. We are live from the TBPN Ultradome, the Temple of Technology, the Fortress of Finance, the Capital of Capital. We have a fantastic show for you today, folks. We're going to be talking about Mag 7 earnings, recapping what's happening in the public markets, the trillion dollar tech companies that have been powering and holding up the entire global economy. Will they continue to hold up the global economy? So far, the answer is yes. I'm glad the chat is enjoying the costumes. Jordy, who are you today? You got to be living under a data center to not know who I'm dressed up as today, none other than Mark and Drieson.

Spitting image. Spitting image. How am I doing? I think you look great. I think, yeah, I think you nailed it. I saw some other folks. they they they they of course i am elia siskiver uh there are a lot of people that boil down the mark in dreason uh costume to just being bald and that's of course not it you need the shirt

you need the watch you need the beard these are all the only thing i'm not going to be able to do very well is talk quite as quickly as yeah it feels like the mustache and in the outfit might be actually slowing you down a little bit. But I'm not trying to channel Ilya. We were watching the Guardian video on Ilya, his interview, which is beautifully edited. And one of the most interesting, I don't even know it's like an interview,

it's almost a podcast, but it's like documentary. It's 12 minute video by the Guardian. They cooked. And Ilya is the only one who says anything the entire time. Yeah. There's no guest or there's no interviewer. There's no host.

And it's just him talking about AGI. will be able to do anything, we'll make, it could be good, could be bad. And he, he just has these crazy long ways, these ways of answering these questions

that are really interesting. And it fits with that medium. I don't know how much editing they did to get his answers to a place where it, like, that kind of flow, but it's fantastic. Ilya's weights,

Ilya's words hold so much weight. Yeah, to that format. That format just worked incredibly well. Yeah. But anyways, if you are watching the recording of this, at some point in the future.

I recommend whenever I'm talking, put it on 3X speed to get the full mark experience. And we should talk about the Gen AI company that we work with to make this possible, of course. I think people would be wildly impressed at how little latency there is, the kind of character consistency. We are partner with Fall,

the generative media platform for developers, the world's best generative image, video and audio models all in one place, develop and fine-tuned models with serverless GPUs and on-demand-class But of course, we did this the old-fashioned way, the Hollywood way. We were at the office at 6.30 this morning. It took hours to get to this point. Yeah. And shout out to the whole team that made it possible. Yeah, it's a lot of fun. It's not a mask. It is makeup. It's a series of prosthetics that go on layered up with a wig for me and a beard for Jordan. They kind of take you through the process. so becoming fully bald and then add back whatever they need to, including the eyebrows. And the eyebrows do do a lot of the work. They do a lot of the work.

But yeah, if you're not taking Halloween this seriously, what's going on? You've got to be taking it seriously. Tyler, are you taking Halloween seriously? What did you do for? I'm taking Halloween extremely seriously. Well, who are you? Sam Alman at 2008 WWDC. Okay.

pitching looped. It's the, yeah, you're the spinning image. Were those the colors? of the polos he was wearing? Yeah, I think the camera temperature is a little bit odd. In real life, if the viewers could see, they would be... And the hair is a perfect match. And you also had to do three hours of prosthetic makeup?

Yeah, I was actually here a little earlier than you guys. I was in back, though. It is so strange looking across and seeing the spitting image of Ilya himself. Anyways. Ramp, time is money, say both. He's used corporate card, bill pay, accounting, a whole lot more. all in one place.

When we partner with ramp, they never expected to see Ilya. No. Pitching ramp. This is why you advertise on TBPN. You get ad reads from Mark and Dries. From the goats. The goats themselves.

We have a super fun show today. We have Lulu Misservi and Gabby Goldberg coming on. Later, we have Dean Ball joining. Who else we have? Tyler, do you know Ball? You're just trying to write. Every time we mentioned his name. Do you know Ball?

Well, after today, everyone will know Ball. Because Dean Ball is coming on TVPN. We're very excited for that. So let's take a second to tell you about re-stream. One live stream, 30 plus destinations. Of course, we are restreaming today. Reach your audience, wherever they are. But I think it'd be good to just kind of go through some of the headlines.

Trump cut China tariffs after he met up with Xi Jinping. in South Korea, President Trump and Chinese leader, Xi Jinping emerged from their face-to-face for their first face-to-face meeting in six years with a temporary detent in the bruising trade fight between the two superpowers. The U.S. agreed to lower a fentanyl-related tariff on Chinese goods to 10% from 20%

after China promised to crack down on chemicals used to make the often-deaddened drug. Precursors. Precursors. That would bring the average tariff rate on many Chinese goods to around 47% from 57%. So America's back in back and off a little bit. Beijing also promised to ease some of the controls it imposed on exports of processed rare earth minerals for one year. China dominates the production of the minerals, which are needed to make everything from smartphones to submarines.

Wow, that's a world tour of everything in the trade. deal. So we're slowly ratcheting things up, ratcheting things down. We had a good conversation yesterday about this with Chris McGuire. You're welcome to go check that out. He gave us more details on the history of the China trade war with the U.S. But the bigger news in tech world, of course, is earnings. Amazon posted a big jump in profit and revenue. Amazon.com reported net revenue, net profit surged 39% and revenue rose for the third quarter, propelled by strong retail sales as well as continued growth in its cloud computing and artificial intelligence business.

The e-commerce giants said Thursday that sales rose for the period. They rose 13% to 180 billion, and net profit rose to 21.2 billion in that quarter. And so you had a little bit of a take on what was. going on in Amazon world. They've been sort of, at the last earnings, they were sort of left out of the AI narrative. All the other hyperscalers in their cloud businesses were, you know, telling great stories

about acceleration and growth. And Andy Jassy was a little bit on the defensive. Well, now he's back on the offensive and the results were much stronger. So they beat on the top line, they beat on the bottom line. And they also beat on AWS revenue, which is great. The Traneum cross the board.

Two chip business grew 150% quarter on quarter. And there's still some debate over how Traynium will scale relative to TPU, whether Anthropic is going to buy a bunch of TPUs. Maybe they're going to pull back on training. But it seems like even if the top end is diversifying and you have open AI saying we want to be on every cloud or, you know, That's not necessarily a bull case because there's so many more buyers coming from behind, adding to it. So my take was a lot of fun around Traneum underperformance relative to TPU,

but they both seem to be catching up finally. I think it'd be interesting to see TPU on inference max. I mean, analysis is benchmark. Are they opting in? So TPU and Traneum, the way Dylan Patel framed it to me, at least, was they haven't said that they're out. In fact, they said that they are in, but it's going to take them time to actually provision everything because it is a significant cost and it's something that requires a decent amount of resources.

And there's a whole bunch of, it's a big company. And it will determine the fate of their product line. Potentially. Potentially. And so, yeah. I mean, people are going to be paying a lot of attention to that. It's possible that they want to like, that they want to get their ducks in a row so they don't just get like clobbered. in this race.

But interestingly, AMD and VEDI were both pretty happy with the results. There were tons of positivity about AMD in certain use cases. And Jensen on stage at GTC was spending like minutes just on inference max slides and data. So my take from the newsletter on Amazon,

I said Jassy needed this after the timeline had turned against him pretty horribly. In the last few months, people even saw the headcount reduction from Monday as bearish. I think that just shows how far ahead Sotia was. He did multiple rounds of playoffs earlier this year. And obviously, Amazon has its dance partner now.

And so we'll see how this develops. One of my favorite posts, I think from last week was somebody said, I don't care if I get caught. I was heavily involved with the NBA illegal gambling scheme. My name is Andy Jassy. I work at Amazon.com. I didn't realize that they were like, I feel like once you get to Mag 7 level, you don't have like a retail army that gets like frustrated about like, you know, middling financial performance. But I mean, so they were, you know, proud member of the lag 7. Yeah.

Look at the lag 7. What's that? They were, I mean, not the lag 7, but they were the lagging part of the mag 7, right? Oh, that's a thing? That's a meme or did you? Yeah, it's a meme. Oh, I haven't heard this one. It's a pleasure to introduce you to a new name.

Alphabet up 247% over the last five years. Amazon is still only up 62% even after the pop this week. Well, that's the story with Amazon. We also got Apple earnings after we wrapped the show. Before I tell you about Apple, let me tell you about Privy, wallet infrastructure for every bank. Privy makes it easy to build on crypto.

Real securely spin up white label wallets, signed transactions, and integrate on-chain infrastructure all through one simple AP. Yeah. Okay, Sarah Wang, GP, over at Andresen, just posted. The only thing better than Halloween is Halloween AI edition, in case you missed it, tagged Ilya and Mark. And of course, no AI here. We did this one. Did this one the old fashion way. We did this organically. We did this organically. It would be interesting to actually benchmark. I mean, obviously, the frontier video models are remarkable, but we've done this bezel bench for a while where we went and did something very weird and very specific. And on that specific use case of Tyler, a person that needs to be him, needs to look like him, and he has to be wearing six different watches and you have to zoom in from the very start all the way out. And he has to be lying down. And so you have really to understand how the anatomy comes into the picture. The models haven't really been

great. I feel like if we tried to recreate this using AI right now, it would get pretty chaotic, the fact that we're alive, three hours. It'd be doing, you know, you could do a face swap live. We've had some founders come on the show and do them, but it feels like it's almost more of just like a trippy background. Descartes. It was really cool. Yeah, Descartes was awesome. But it was still very much felt like a video game. It didn't feel photo real. Yeah, and, it didn't feel like you could reach out and touch. You can in fact touch. I'll be drinking through a straw today because it's locked going on. Is it? Is it locking yours? My take is I have the same level of prosthetic makeup on, but I'm just full sending it.

And by the end of the show, I'm expecting this to just be melting off my face base. We'll see how it holds up. We'll see how it holds up. And we'll also tell you about cognition. We're the makers of Devon, the AI software engineer. Crush your backlog with your personal AI engineering team. Yes. Did you want to go to Apple or did you want to go somewhere else?

Let's go to Microsoft. Microsoft. Okay. What's your Microsoft? Topline beat, earnings beat, stock traded down 3 to 4% after hours. It's recovered quite a bit since then. I said I was surprised to see Microsoft trade down following earnings, but it's recovered since.

Satya, in my view, has already gotten a lot more credit for the company's opportunities in AI than Amazon and even Google. People have known about Google strengths, but the big questions around search have held back performance to date. Microsoft hasn't really had any narratives holding it back, but I don't think people give them enough credit for the Open AI deal. If you believe Gen.A.I. is a sustaining innovation. and will transform knowledge work.

There is not a company in the world that is better positioned from a product and a distribution standpoint. Again, we talked about this with Satya on Tuesday. There are so many companies that are trying to build Excel plus an agent on top of it. Sure. Or agents for Excel,

which is exactly what Microsoft is also doing with Excel with co-pilot, right? And they're model agnostic. They're going to be able to bring in any kind of intelligence that they want. And so anyways, I think, you know, they continue to be super well positioned. It's also fascinating because they have access to Open AI's IP. OpenAI wants to be a big player in the workplace, right?

You can imagine them getting into a bunch of other applications there. And so it'll be that competitive dynamic is just something that we're going to be watching for a long time. Yeah, I saw some haters on the timeline talking about Microsoft launched a research product and the in the hot take was like Microsoft's late to launch this and it's somehow worse like they couldn't even just copy exactly. And I was and I was thinking like there is something real about like when a big company launches one of those features like they typically have to move a little bit slower because maybe they're regulated by Europe. They're all they're international. They have you know they obey all the different rules around accessibility and they have

they have like 25 they might be like sock two compliant plus i tar compliant plus you know it's 25 other compliant levels they just slow down like how fast you can move and how many things you can break um but that is very different like as an early adopter it's very different to look at a new product from a you know what 40 50 year old uh you know tech giant and say like yeah i i prefer like the cutting edge product, as opposed to what actually plays out in the market. Like, Slack was very much the same thing, where Slack was this product that was so much better than the alternatives. Like, if you, I don't know if you ever used, like, workplace chat pre- Slack, but it was, I was basically born on Slack. You were born on Slack. But Slack really was, like, a delightful,

like, you know how there's, there's just those companies that come out and you're like, oh, this company takes design seriously. This company is a power user of Figma. Think bigger, build faster. Figma helps design and development teams build great products together. But you know how you get that vibe? And then you get the other vibe where it's like, oh, this company is owned by private equity and they don't really care.

John, I got to give you, the chat's giving you a shout out. I never knew John was so ripped. There's been some other comments. Yeah. Put them up, John. There you go. Ilya, AGI has been achieved. Let's go.

Ben Gilbert in the chat. You guys are absolutely saying I love it. Welcome to the stream, Ben. Thank you so much for stopping by. You're an absolute legend. We couldn't do this show without you. Massive inspiring, everything that they've done over it. Acquired.

Totally. If you've been living under your data center, get over to Acquired.fm. Subscribe. Let's get into meta. Let's talk about meta. Yes. Top line beat, earnings miss. Massive miss, like a $12 billion.

People are not talking. enough about what happened. Like it is crazy that you could be operating such a big company, just be like, oops, $10 billion. And it's not like, I mean, think of how they've been spending for the last quarter. It is crazy. This is like the quarter, you know, this is like, you know, I'm sure meta and the executive team, it's like after a weekend bender checking like your credit card statement. Yeah. Right. It's like, it's like you have an idea maybe of what you spent. Yeah. But it's way beyond that. Okay. Wait. So, Tyler, can you look up what the actual tax charge was because I don't think it was about scale AI, which everyone, a lot of people would say,

oh, there was a lot of money to pay for something that maybe is an aqua hire, right? Or like the $3 billion that maybe Zuck just paid for like one AI researcher. He's been on the spree. I think this is a separate $10 billion charge. Yeah. Which is just... Came out of nowhere. I like the, I like the metaphor of like, you're just on a bender and you're just like, oh, my credit card bill is so huge this month. What? How do it? happened. Wait, so are you talking about the $15 billion dollar charge? It's five more than I thought. I thought it was 10. I think it was 10.5. 14 point something. Fourteen? Came out of nowhere. It was in the Wall Street Journal yesterday, but we'll... Okay. Yeah. So, okay, so apparently

it was a one-time tax charge related to the Big Beautiful Bill Act. What? So I'll read into this more, but it's not the... It has nothing to do with research. It has nothing to do with YAR. In the X chat says from what I read, they missed only because of the tax pay payment, which is only on paper, not actually cash. And they said it would save them on tax costs over time. Okay. Okay. Thank you. No, that's very helpful.

Yeah. Okay. So they're changing the amount of, like, how they're accounting for these taxes. And so they took the hit to earnings. But is that, is that why is the stock traded down 10%? Is it specifically because of this one-time tax charge? Like, I don't know that it is. I think it might be more.

I think it could be a little bit of vibes in the mix. I liked your vibe-based analysis. Your take was rough to be completely rebuilding an AI while simultaneously still finding footing in hardware. Core business rips, though. My take in the newsletter was the timeline loves the core meta-business, but doesn't have a lot of faith in Zuck. I think that sounds harsh, but I think that is generally real. People thought Bezos, who had a similar level of control over Amazon, was crazy to spend like he did building AWS. And he was ultimately completely vindicated.

people thought Mark was crazy going all in on the metaverse and VR, and so far that part of the business has not shown any signs of being ROI positive, right? Obviously, they're still shipping. The new meta-ray bands are like the leading publicly available pair of glasses. It just feels like the wind case there is like an AirPods-sized business, right? Yeah. It's hard to get to, okay, yes, this is going to be as big as even Waymo. I don't know. I don't know.

I don't know. I think that if you actually win that category, it could be an iPhone-sized business. Yeah. So I'm not saying that he shouldn't be investing there. But I don't think the market believes that he's going to win that category yet. Which is fascinating because what company do you think is in the lead in terms of VR? and AR right now. Rank them.

Who do you think is... I think the challenge is when products are not at the level where people are saying, I love this, I use this all the time, I use this every day, I can't live without it. You don't get credit for being at the leading edge. Yeah, people aren't saying that yet. Like, people expect Apple to come over the top and do, you know. They tried, and they didn't pencil up.

No, I know, but now they're working on a pair of glasses, and it's going to integrate with your iPhone. It's going to integrate with iMessage. that's going to be a real competitive, you know, challenge for meta. Yeah, no, no, no, no, it's a really good, it's a really good argument. So, at the same time, giving meta some credit. Reels is now a $50 billion revenue run rate business. Let's go.

Let's hit the gong for the Reels team. Threads is also, where is our, where's our mallet? Ilya gong hit incoming. The moment you've been waiting for. Clean, clean. Moving on, the market reacted positively the scale AI acquisition and hiring spree, but I think investors have the right to be concerned about meta's vision and ability to generate revenue from their Gen AI efforts, either directly or indirectly anytime soon.

I'm still shocked they release Meta vibes, knowing how much scrutiny their first release as a new team would get. The product showed a complete lack of vision, in my opinion, and just product quality when compared to SORA, which launched about a week later. So it felt like they knew SORA was coming. They wanted to be first to market with an AI native social feed, and they did not bring the heat. They just completely burn the ships on the meta-AI app. Because that app was originally, the reason that I have it on my phone is because I got meta-ray bans.

And so when I got meta-ray-bands, I had to get the meta-app to use the raybans. Then it turned into the meta-AI app, and it was like, hey, do you want to use this as an LLM? And I was like, I'm basically good on that front, but thanks. and then they were like, well, do you want some AI video? Would you like some slop in this drop? Which is just like a wild canvas to be iterating on so quickly. They used to launch different apps every single time. But not this time.

I don't know. The MetaVibes needs to turn. I still like the Mid Journey model. I think it is beautiful. I think there is something cool to be done with that. I'm not judging the outputs. It needs more UI. But the big problem is that you couldn't as a user consistently generate beautiful things.

Yeah. Whereas with SORA, it's not like they were necessarily beautiful, but at least they were entertaining. And that's what kicked off that viral. So the main thing is, like, for Microsoft, Google, and Amazon, it's a lot more clear what winning in AI actually looks like. I think MET is going to have to clarify what the wind state looks like beyond just saying we're building personal superintelligence because investors are smart enough to know that that personal superintelligence doesn't mean a lot today when the business today, is focused on monetizing the world's attention. Yeah, yeah. And so I expect them at some point to have to kind of clarify what their plans are.

Yeah. Tesla beat earnings missed, a beat on revenue missed on earnings. The stock reaction was down about 4% after hours. You reposted the Bucco Capital bloke post. I'm always so confused when Tesla goes down on earnings as business performance is completely irrelevant to the price of the stock. And it has been more of a vibes-based stock for a while. I had a more by the book analysis. I said, if Elon goes straight shot to the full self-driving equivalent for humanoid robots, basically no teleoperation, the project should be years behind what's going on with 1X, I would imagine.

Just like the Waymo dynamic, like Waymo was able to launch a lot sooner because they could always fall back on, hey, there's a human tele-operating this thing. So you can trust us on the city streets. and they've had a pretty good track record because of that. Tesla's going full moonshot, full send, but it's going to take a little bit longer. Like, no one debates the fact that, like, Tesla's Robotaxy fleet is a year or two behind Waymo's right now. Now, the Tesla bull hope is that you come from behind and start compounding

because you have just structurally better costs, way more manufacturing capacity, more of a flywheel on the actual data. and the training. There's a lot of options. I mean, I said they also need to build a full-size SUV. You can't wait for that. I just think it's so crazy not to build a full-size SUV in America. Like, Americans love full-size SUVs, and you see them, it's like, I mean, it's as crazy

as not being in the truck market for as long as they were. Yeah, it's interesting that Tesla has been so dependent on the luxury sedan for its entire history, even though the luxury sedan has, like, completely fallen off in popularity. I saw some data on Lexus's sales. Lexus used to sell one of the most popular luxury SUVs, and they sell either like single digit thousands now of them a year. So it's like that category has gotten just killed by the SUV. And Tesla, of course, has the X and the Y.

But it's not exactly what you're, if they were to ship like the Cybertruck SUV, I think that would sell significantly better than the. cyber truck. Yeah. I wonder if they'll ever do multiple brands. It's so normal at every other car company, but I see no reason why Tesla would ever do it. Yeah.

Like, Lexus is what, Toyota's upmarket brand, right? Yeah. Isn't that the linkage? Yeah. I don't know. I doubt you on whatever. I mean, Plaid was sort of like the trim level. Yeah. I wonder if they'd ever do a different trim level.

Yeah. Plaid plus, like, like, what does equivalent of like the my bach of Tesla if Plaid is their AMG right yeah I don't know but I don't think he cares about they have that in in China they have a like a my

bachified model model X or model model S that's like chauffeur where the seats the front seats are pushed forward basically I have other Elon Musk news but first let me tell you about Vanta automate compliance manage risk proof trust continuously

Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program. Elon Musk on data centers in orbit, SpaceX will be doing this.

He said he's going to do data centers in space. Didn't we say this out loud yesterday with someone? I don't know if he was on the show or not, but it just seems obviously, if there is an opportunity. If it's a thing and it's big, they're going to try and do it. It's easy to be bullish on space right now. Yep. But if you're investing in different categories, you have to imagine that if there are adjacent opportunities besides Starlink for SpaceX to do, they will enter them if they are large enough categories.

And so, again, that's... What does that mean for StarCloud? I mean, I can still think, I can still imagine them over time figuring out like niche, more niche use cases. But, you know, building the, building the satellite cloud, I think is something that Elon. is not going to just, you know, happily hand over to someone else. Yeah. So let's read through a little bit of this from Ars Technica. An artificial intelligence, as artificial intelligence drives the need for vastly more computing

storage and processing power, computing storage. What's that mean? Interest in space-based data centers has spiked. Although several startup companies, such as StarCloud, have begun to address the problem address the problem. They're like identifying the problem, maybe prototyping. This is very oddly worded.

I'm like struggling to read this. The idea has also attracted the interest of tech barons. In May, it emerged that former Google chief executive Eric Schmidt acquired relativity space due to his interest in space-based data centers. That's fascinating. I know that relativity was

was not doing well with kind of the V1 strategy, and there was this turnaround project, and Eric Schmidt came in and was, I think, paying a ton to keep the business going and really investing. I had no idea that that was his thesis. That's fascinating. Then earlier this month, Amazon founder Jeff Bezos predicted that gigawatt-scale data centers

will be built in space within the next 10 to 20 years. That's big. Now, Elon Musk, who's SpaceX, owns and operates significantly more space-based infrastructure than any other company or country in the world has also expressed interest in technology. After Ars wrote a story about the potential of autonomous assembly to construct large data centers in space,

Musk responded on X by saying that Starlink satellites could be used for this purpose. Simply scaling up Starlink V3 satellites, which have high-speed laser links would work, he said on the social media site X, SpaceX will be doing this. Very interesting. Yeah, I think it makes sense. The question is just like, yeah, what is the actual timeline for any of this?

I mean, no one really expected Starlink to ramp as fast as it did. And then it just like kind of worked. But if there is, I don't know, a bunch of, if there is a lot of like actual fundamental foundational science around like heat dissipation that needs to be done, that might be something that's a little bit bottlenecked. But at the same team, at the same time, kind of like, I would kind of bet on the SpaceX team to figure that stuff out since they already do a ton of heat dissipation and material science just on. Yeah, it's also just fascinating to look at the space industry versus the automotive industry.

Yeah. The automotive industry is the most brutally competitive, challenging business where Elon is having to compete with every car manufacturer in the world. Obviously there's export, import, import controls in automotive. But in space, it's like, you know, you can imagine that being not a perfect monopoly, but certainly having every possible advantage in order to dominate in that category when the time is right. Should we talk about Apple before we continue? Yeah, definitely. Apple beat top line, beat on bottom line. China sales decline year over year. If you've been paying any attention to Apple, you know that their China business is slowly going away.

iPhone revenue they missed just by $300 million. $50 billion or $49 billion this quarter. Yeah, iPhone revenue is $49 billion versus $49.33 estimate. Stock is relatively flat since then. Yeah, so this is the real like this is the real bombshell in their strategy. consumer intelligence research partners, a third-party analyst firm, said it observed great interest in the pro and pro-max models, while the new iPhone 17 error model showed virtually no sign of traction.

So I saw my first iPhone air in the wild yesterday. Didn't, no, you saw that one get smashed. Tell that story. Which one? Oh, oh, this was genuinely like the most. I posted about this. But during LA Tech Week, we had a dinner. Why did you put it in quotes?

It was LA Tech Week. Well, there wasn't, I mean. It literally wasn't LA Tech Week event. You're like, Los Angeles. Los Angeles Technology. TVPN. No, so we had dinner with some folks. And there was a founder there from Venice.

From Venice. And he was saying that the iPhone Air is the strongest iPhone ever made. and that he challenged anybody at the dinner. There was like 20 people. Do you remember Bendgate? Do you remember the I? Do you remember Ben Gate, Tyler? It was like the iPhone 5 or something.

I was, and I was pretty young. Okay, yeah, you probably before your time. But there was a time when the iPhone launched and you could just bend it. And it was like a lot of phones were getting bent. So this guy is going to you and he's going around and saying that no one can bend this phone, right? That's the strongest. No, no. he said nobody can break this phone.

He said this phone is absolutely unbreakable. Unbreakable. They've tested it internally. You cannot break it. It's the strongest iPhone ever. He compared it to every other iPhone, even the titanium iPhone. And so he was bending it like this, going crazy. And he challenged anybody at the dinner.

Who wants to try to break this phone? And Agundo founder, I'm leaving both of them unnamed. A gondo founder. And we got Gabriel from OpenAI in the chat. Good luck. man, you're building cameos over there, but you're going to compete with us. We're doing it the old-fashioned way. Yeah, we're doing it the old-fashioned way.

It just took us three hours sitting here in makeup. What do you mean three hours? It was more like four hours. It was like four hours. So Gabriel, you sped that up just a little bit, but, you know, I think we're sticking with this for the meantime. Okay, so anyways, this founder challenges the entire group at dinner. Can anybody break this phone?

Yeah. The Gundow founder raises his hand, takes the phone and, like, tries for a second and just snaps it basically. Just completely shatters it. Completely broken, won't even turn on. And the Venice founder sitting there saying, like, he was like in disbelief. Like, every, the whole room just went completely quiet. And he was like, I'm going to, I'm going to email Tim Cook about this.

Emailing Tim Cook, what is he going to do? You're not supposed to smash your phone. Anyways, on Apple, John, your take, you said sort of the inverse meta core business, not at risk due to AI, but not investing aggressively. consumers seem happy to just open an LLM app could take. I said Apple's trading it 40 times PE with very low growth. This feels steep even if I don't think they're threatened by AI in the near term. And then I highlighted a quote from Bernard Arnault from LVMH.

He said, I don't know if in 20 years time people will be using iPhones, but I do know they will be drinking Dom Perion. It's a good one. And hopefully they'll be using graphite.dev. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. In other news... We should close it out with Nvidia, just to finish this little earnings wrap up.

Basically, John, your take was on top of the world, but is the head that wears the crown heavy? Core AI factory build out going flawlessly, but unclear why Nvidia is pushing data centers in space, humanoid's quantum computing, mumbo-jumbo. Maybe it's real. Maybe just forward of the same old AI story? Question.

I said, I don't think Jensen would be just chugging beers in Korea if he was having a rough one. So we'll see how it nets out. Their earnings are not until November 18th. What's wrong with chugging some beers? Yeah, what's wrong with chugging some beers? Especially, I mean, you got to, like, he is superlative at this point. I believe just quantitatively, he is the most value creative founder, CEO in human history.

like no founder CEO has created five trillion I mean like no one's created five trillion in any in any capacity whether in in the CEO seat or or not because like with a with a with a Microsoft or with an apple you have to start having the conversation when you're having the goat debate on how much credit do you give to Tim Cook for market cap creation how much credit do you give to Steve jobs right and same thing with satchi Nadella and bill Gates like obviously Sotja has taken Microsoft from 300 billion to 4 trillion. Does that mean that he's 10 times better than the founders that started the company in the goat debate of value creation?

Like, he did have a big, he had a lot of great shoulders to stand up. It's not entirely about number go up. Exactly, exactly. But with Jensen, you don't even have to have that debate because he's been the founder's CEO the whole time. And so he is, he deserves a beer. Let the man have a couple beers. with his boys in South Korea.

He's been on an absolute chair. A couple more posts here. Jay Bolt Hart says, I'm sorry, but no company generating the cash meta is with the margins it has should ever be raising money. Of course,

they announced another bond sale yesterday morning to fund their buildout. Jay Bolt Hart says, absolutely pathetic. Maybe just spend a little less and don't pay every AI scientists like their show hey Otani.

Yeah. I want to understand more about the private credit dynamic here because I believe that they're like a good CFO would tell you that if you're buying land, you should get a mortgage because it can be secured. And there's just there's just some like cost of capital math that you should do. And it would actually be irresponsible. Like if they were saying, hey, we're going to build a new data center. We need to buy some land. We need to buy a building, buy a bunch of GPUs. And they're like, and guess what?

We're not going to use debt. Like, that's like paying off your mortgage, basically. And it's like, okay, yeah, you can pay off your mortgage. And you can be like, I'm one of those guys that doesn't have a mortgage. And it's like, congratulations, but you're probably not using capital efficiently. And so when I hear that they're setting up a big data center and they're raising debt for that particular deal, I don't have like red flags flying off in my mind. But I don't know.

This, this, this, this, this, this, this, bultard guy does. semi-analysis had a post here quoting Sundar said over the past 12 months nearly 150 Google Cloud customers each process approximately one trillion tokens with our models for a wide range of applications for example WPP is creating campaigns with up to 70% efficiency gains Swarovsky has increased email open rates by 17% let's give it up for Swaf. Email open rates and accelerated campaign

localization by 10X. Semi analysis says AI is not a bubble. Enterprise AI is very strong. Nearly 150 enterprises each process one trillion tokens with Gemini models

for a wide range of apps. And then they say, wait, that's less than a million bucks per year per enterprise or 150 million of annual revenue, i.e. 0.3% of GCP annual revenue. So... The fact that Google leads

with token count instead of revenue growth that tells you everything. If enterprise AI was actually strong, those 150 companies would be spending 10x more. I don't know about that. Because when you're a big company, like, you don't want to break out like more gap metrics, non-gap metrics than you absolutely have to. Like, because if you develop some monster of a business, like the AWS IPO story, the whole idea that Amazon had this amazing business and was like, oh yeah, like, just include that and everything else. Like definitely don't let everyone know. And Sotcha told this to us on the show was he was like when AWS had to disclose what was going on over there was the best day for us

because everyone looked at our business and we're like, they got the same thing going on over there. They got a monster of a business cooking. And so the question of like how Google positions this, like do they want to share dollars? But I don't know. I just wonder, I wonder what the shape of of GCP adoption is if it's nearly 150 Google Cloud customers each processed one trillion. I wonder how narrow that slice is. I mean, yeah, if Sundar would come out and says, we got Enterprise AI to $150 million run rate, everyone would be like, but it sounds really good when you say one trillion. Yeah. Yeah, that is a weird, weird way to frame it. I don't know. I mean, This is all so weird because it's like, okay, so then what's actually happening in the core business?

Because they, they crushed earnings. They did great. And it's like, so are we just in this weird world where every business is like, yeah, I totally want, like, I'll take one million dollars of tokens. But then I also need $75 million more of just traditional databases. Is that what's going on? Like, what's actually driving the growth, if not the token generation? There might just be something longer tail going on here, too. This might just be like this slice of this 150 enterprises that's actually, you know, that only makes up.

I don't know. The big question is just what percentage of GCP annual revenue is AI, is Gen AI. Like we've seen that the CAPX at meta, they have core AI, they have Gen AI. Like every, everyone wants to track that now about what workload is happening. sometimes the hyperscalers don't know. Because they just said, look,

I just rented you a bunch of GPUs. Are you inferencing a recommendation algorithm on it? Are you playing video games on it?

Are you generating tokens on it? We actually don't know because we respect your privacy and you can do whatever you want with it. So all we can tell

the market is like, hey, we sold a lot of GPU hours. Yeah. And it's all masked. Anyway, Tyler, what's got?

I mean, yeah, A trillion is just like not a lot at all. It's just not a lot of. Open AI is doing like on just on the API like something around like 250 trillion a month. So compared to that, I mean, if it's a trillion over the past like year, even if it's 150, I mean, that's like, that's pretty bad. Each 150 trillion tokens. And you said.

Open eyes doing like 260 a month. I actually can't keep these numbers straight anymore. Like there's, like, everything is like, nothing's in apples to apples. And I think that's very much intentional. I think that you don't want to be easily comped across anything. That's why everyone just focuses on valuation and then revenue and revenue scale. Dylan Patel says so many people hated on us for the AWS acceleration thesis a couple months ago. And we were right, of course.

Only we can do these calls due to our deep knowledge and insights with core research data center accelerator and tokenomics models. Amazon and they does the AMZN with and just drop sales at seminalysis.com. So I expect if you're training it, to be printing the next legendary chart 24 hours. This was crazy. I wasn't sure if this was fake news or not, but on the Coinbase earnings call yesterday, Brian Armstrong went. There was a prediction market that people were betting on which words he would say. during the earnings call, and he waited until the end, and he just read through and read off every single word that he had missed. He says, you know, I just wanted to add the words in here, Bitcoin, Ethereum, blockchain staking, and Web3 to make sure we get those all in before the end of the

call. So completely trolling. We've had that a few times during interviews. I was tracking the prediction market about what Coinbase will say on their next earnings call, and I just want to, you know, add here the words Bitcoin, Ethereum, blockchain, staking and web three. It's actually crazy that he hadn't gotten to Bitcoin until then. Like, it really tells you a lot about how mature the Coinbase business is that you can get to the end of your Q3 2025 earnings call without mentioning Bitcoin, which was like the entire reason the business exists.

But the business has grown so much. There's a lot else going on. It's possible that he said it and he just didn't remember. Maybe, maybe. But it is funny that he's just. One thing is we've had prediction markets created around different interviews we've done. When St. Albion came out, he was mentioning. It is wildly distracting when I'm looking at the chat and people are like, say this word.

Say that word again. It's crazy. Say it three more times. It's a very weird postmodern hyper techno capital accelerationist, like, aesthetic. So it would be like, people are gambling on the words that you're saying in a live stream. It's very, very weird. I don't know what his, I would love to know Brian's honest take on like how he feels about that. Like there's a little bit of just the market is going to exist.

It's going to be a thing. And so he by doing this, he's kind of trolling the market to just say, hey, you're never going to be able to reliably bet on this market because you never know what I'm going to do, which is one way to interact with it, as opposed to being like, hey, guys, I know that everyone has a prediction market up on what I'm going to say. And so please don't do that. It's like, no, just like go create chaos. and then everyone will just know that it's a meme and chaos.

Well, it's going to just drive more attention to the next earnings call because people are going to go. Oh, that's interesting. Well, I'll just bet yes on everything in case he does that again. Oh, interesting. The odds might be at like 95% for every word. And then he just says like he just doesn't mention it at all. Yeah.

And then everything, you know. It certainly, it certainly kind of cements him as, uh, as just a very, in the culture CEO and a very, a very like online founder, which I think is positive. Um, we are, of course, partner with Polymarket. as a prediction market. We're usually tracking the largest companies on there. Largest company at the end of 2025. Nvidia has been predicted, of course,

and even going into the next year. It's like 99% now. I mean, they're a trillion above everyone else. It was a horse race, though. Yeah, it was really close midsummer. Yeah. At one point. I mean, the market caps were eclipsing each other

and they were going back and forth. And at one point. Yeah, June 1st, Microsoft and, And vdia were both at 37% exactly. Yeah, it's been a pretty crazy, it's been a pretty crazy year because of the tariffs and what, and the AI narrative and the deep seek moment. The horse race has changed a few times. And Microsoft has been in the lead.

Apple's been in the lead. But Nvidia is running away with it. And that deserves a beer with your boys in South Korea. That's right. That's what he's been up to. Did you fully understand that there's an article here. Microsoft seemingly just revealed that OpenAI lost $11.5 billion last quarter. That seems like in line with what we've been kind of assuming.

It's like it's a massive loss-making business. I don't know that we'd heard that they were on like a $40 billion loss run rate. But that kind of lines up with the amount of money that they've been raising. Maybe Intel. Like, you mean the, like, the biggest cash in a setting? Who's losing more money than opening up? Do you count CapEx as losing money? Because I would almost certainly assume that these are not losses.

These are investments. So, like, if I raise $10 million for my startup, and then I'm like, okay, I'm going to buy, of office building for $5 million, like that's not negative 5 million earnings. That is negative 5 million cash flow. I've lost 5 million cash and I don't have liquid cash anymore. And so I would assume that a lot of Open AIs losses are actually investments in CAPEX one way or another. But they're also I mean, they're also probably losing. Stock based comp is factored into this as well. Yeah. And that's also in many cases, a non-cash expense. And so still important to track, still significant.

But all of this stuff, it's like I'm kind of in like a wait and see until the S-1 drops, until we can actually get data. Do you have any intel? Yeah, so apparently this is double what they've, like, what has previously been reported. So I think before there was a report it said the first half of 2025, they were going to do 3.5 in losses. So this is basically double that. But even that, it's still, I don't think this is very surprising.

It's so hard to track how much money they're actually raising because they'll raise for a particular deal, they'll raise for Stargate, there'll be a secondary transaction. It's very hard to understand, like, how much cash is actually going from investors onto Open AI's like for-profit balance sheet that is like available to burn.

And then once you have the cash, I think everyone's fine with the idea of Sam Altman burning, money for 12 to 18 months and going out and raise more. Like that's not crazy when you're growing your business, especially in a new category. In other news, Palantir sues former employees for stealing company secrets, company alleges pair violated non-competes to create copycat AI firm. Interesting. They've, the two employees, they're accusing them as stealing confidential data to build a rival company called Percepta backed by General Catalyst. I would not want to steal from Palantir. Bad idea. That seems like one of the worst companies to steal from. They will find you.

Yeah. And they will kill you. Um, the company, but, uh, it is a, is a wild thing. Yeah. You think about a company that's going to have like, you know, things pretty locked down. Their whole, their whole stick is, they knew, what's going on. They know that if I remember, I read the article yesterday, but I believe they, they, they tracked that the employee had slacked themselves a document that they then opened on their phone and like, presumably, you know, save. So, We'll see how this nuts out. They say that taking access to Palantiers crown jewels, such as source code and customer workflows

to create what it called a copycat platform. So they're just trying to block them from using the data and to force the return of stolen materials. One of the comments says, never go against a company that sued the government and won. Interesting. Sean Sanker was on the... Ross Douthit.

It was really good. I don't know how much you listened to, but he... It's funny because they opened with the classic question of, like, what does Palantir actually do? And I don't know. Like, why are people so confused by that question? It's so crazy to me.

It's like... And Sean explains it extremely well. He's just like, like, you have... data we help you understand your data we help you understand what's going on there you know it's it's he's he's like it's a data analytics firm sir it's the same other thing um when when when tray stevens went on sean ryan yeah like the questions were like almost the same exact questions like what does palantir do like and then you read the comments and then everyone's like this is the antichrist yeah people go crazy crazy crazy uh and i think it's i like shah ma's show me this great example of like when you want to make

a thing. Like if you're if you're running a manufacturing plant, you will have a system for tracking like if you're trying to make like Yerba Mote, you will have a system that stores like the specifications of the can, what type of ink you're using here, the size of the can, the shape, what materials, how this gets printed. Like that will all be managed in a in a particular software system. then how you're actually making the cans will be like, oh, okay, I'm producing 25 cans a second or whatever. That will be tracked in a different system. Then you might have a CRM to sell.

Who do I sell these cans to? I'm selling them, you know, every month, Jordy buys a bunch. So I have a CRM that tracks you. And he's just describing like the different enterprise software tooling sets, like CRM, CMS, warehouse management system, inventory management system. He's just describing those and explaining how

Palantir can tie them all together and how Palantir can build them and sometimes you need custom ones for specific for specific industries. Well, let's check him with the Palantir Bulls. Anyway, yes, what do the Palantir Bulls say? The Palantir Bulls will get exhausted at these levels. Full send. I like the guy doing it back.

Our team workouts should look like that. We don't break nearly enough of a sweat compared to this. This is so sick. let's go look at this bro backflips I like that guy who's just like

racing it like it's a chariot or like it's a horse race that's hilarious yeah this also seems slightly sped up maybe what do we think

do you think this is like 1.1 X slowed down you think it's slowed down get on julius dot AI what analysis do you want to run chat with your data

get expert level insights in seconds where to you next? We got the mansion section. There's a bunch of stuff we can go through. Did you hear that the Hampton's market is making an epic comeback? Let's go. Finally.

Home sales of $10 million or more houses are surging on the east end, even topping the COVID surge. In the Hampton, New York's affluent beach vacation destination, the luxury real estate market, has come roaring back. Have you ever been to the Hamptons? I still think I've never been. I actually, Bobby Cosmic was worried about the Hampton. Maybe as a kid, but not as an adult.

The number of home sales above $10 million on Long Island's East End has surged to unprecedented levels, even topping the area's pandemic. Boom. The uptick comes even as Florida and other locales have replaced New York City as a primary home for many billionaires. The Hamptons is for a last. lack of a better term, kicking ASS at the very high end of the market. Wow, uncensored swear word in the journal. Absolute dogs over there in the writers. Wow. I mean, out of control. Out of control.

We would never. We got to get this under control. No, I think I think that's not that bad of a word. A.S. The high-end Hampton's market was in the doldrums before COVID. Home sales of 10 million or more plummeted from 58 in 2015 to 30 in 2019 as the Ritzie area faced increased competition from areas such as Nantucket or Hudson Valley. And so there are some stories here to the Bannerborn. I've been working at my land thesis. Yes. I think I think that it's probably the most underrated asset class right now. Well, don't spell the alpha. I mean, I mean, I'm down to sure the alpha with the team. Now, I just think throughout, you know, all human history, people were like,

how do I acquire land? Yes. And then at some point or another, people just stopped, right? It became just about getting the home. Yes. And not thinking about land. So. Where?

Where do you want land? Do you want to buy Fifth Avenue to, to, to, to, to. I just want to buy up land in the area that I love the most. American jurisdiction? Of course America. California? Of course California. So Southern California?

I'll never give up on California. I still think there's, I still think there's huge alpha in just acquiring massive swaths of Alaskan land. Yeah. I still think of working this one out for a while. Yeah, I think it's massively underrated because if, if global warming is real and unstoppable and we can't fight the force of global warming, well, then temperatures will rise. And Alaska will be the last beachfront property. It'll be balmy. It'll be amazing. Everything else

will be a desert, but Alaska will be nice and temperate, and that's where you want to be. And if we cure global warming, if we stop global warming, if we wind up, figuring out how to manipulate the environment, we become, we harness the... You can manipulate time. Yeah, manipulating time, manipulating the weather. if Augustus is successful, if geoengineering is successful, well then it's not going to be cold up there because you'll just be able to transform it into a lush environment. 72 and sunny all day. So I think in either scenario, no matter where you land on global warming, is it unstoppable, is it preventable, Alaska looks great in that world. What do you think?

Did you ever resolve your beef with the Alaskans? No. I locked. So I posted on X. Actually, it was Twitter back then, like four years ago, I posted, like we got to go to Alaska. We got to build some condominiums. We got to build some strip malls. We got to build some skyscrapers. It's too nice up there. And I was basically like, we're going to, we're all the tech people, all my tech friends, we're all going to go to Alaska. And the Alaskans did not like this. And they said, stay out of Alaska, you tech boy. and I was getting I was getting attacked both the left and the right you kind of listened too I did I did I in fact I did not invade Alaska

but I was getting lots of DMs from Alaskans who are you know gun totin good old boy Americans do not step foot do not you wouldn't last a day up here you're not tough enough for the Alaskan wilderness and then I was also getting canceled by the left Alaskans who are like

this is my indigenous homeland you can't come here that type of thing And so I was getting it from both sides. I wound up locking my account because the post went majorly viral in Alaskan, like Twitter, Alaska Twitter. Of course, I was joking around, but the internet doesn't care about that. Twitter is. Twitter is such a small place already. Imagine how small Alaska Twitter is.

There's like six posters up there, regional posters. But this is the lesson for folks who post and are worried about getting canceled. Not that you should go out there and get canceled. But if it does happen, it usually starts by somebody quote posting you. And you can tell that, you know, you posted something, you got 100 likes, but the quote post is taking off and people are having a lot of fun quote tweeting you. That's when the like ratio on the original post is just abysmal. Exactly. So if that happens, you can just bail by locking your account, which then kills the quote tweet because it doesn't show up in the quote tweet.

And so no one's liking or retweeting a quote tweet with just this account. locked. And so it just kills the virality of the quote and the dunk. And so you stop getting dunked on. And then no one can really like DM you or or reply to you or dig through your account. And so usually at least in the short term, people are kind of like, uh, moving on to something else. I'm not, I'm not letting it percolate. Um, we should talk to Lulu about how to handle cancellations when she hops on. Uh, she's, of course, seen a bunch of these. Um, speaking of people who should be canceled. Someone owns a disgusting house in Los Angeles. We got to dig into this. The mansion section's really getting wild a day. They put a curse word in here. Don't like that. They also are listing an

$85 million mansion here for sale in L.A. an estate with a curated collection of S-E-X toys. What's going on, mansion section? Why are you getting so wild on Friday, October 31st? Spooky. It has a pleasure. It's not like them. A pair of Los Angeles. mansions, one with a pleasure suite, outfit with a curated collection of sex toys, is hitting the market for $85 million. The seller is British music producer, Alex De Kidd. He's 44.

He spent 14 years building one of the houses, a roughly 24,000. 14 years? Yeah, he started when he was 30, I guess. He thinks in decades. He thinks in decades. In addition to the Pleasure Suite, the Hollywood Hills home,

which DeKidd called his entertainer's fortress, has eight pools, very cool, an auto elevator, and a private nightclub. In 2020, DeKidd purchased an adjacent five-bedroom home with its own pool for around 7.6 million property record show. He bought the roughly 6,125 square foot house with the idea of adding it to the main residence, he said, but ultimately decided to use it as a guest house instead. DeKid bought the first property, a Spanish-style home in the birds. streets area for around 4.3 million in 2011. The purchase came shortly after renegotiating his first publishing. Mr. Morale in the chat, Alex DeKid, he's 44. His real name is Alexander J. Grant. I basically

got this house thinking it would be a six-month remodel, but after more than a year of renovating, DeKid decided to raise the house, destroying it, and start over. It was like making an album for me, DeKid said. I got these other ideas. McLean, who's known. for designing modern mansions is a madman just like me, DeKidd said. The four-bedroom house has a glass elevator for cars, a pleasure suite with a Murphy bed and mirrored wall, contains antique chairs customized with patent leather, upholstery, restraints and chains. Oh, yo, yo-yo.

I think we're beyond the days of it being taboo to enjoy pleasure, De Kidd said. I wanted a place where you could feel comfortable and intimate and explore. All right, let's go to the next time. We're not past it. It is taboo. Tear your house down. I don't know. No, the mansion section is all over the place today.

Shutterstock founder sells a bridge Hampton home. What a great time to get out of shutter stock with the dawn of generative imagery. This has to be one of the hardest businesses to make. He's not out yet. He's still the executive chairman. Yes, but he's clearly sold enough to. to buy $57 million. We should have John on.

I'd be interested to hear how he's, how, I mean, I'm sure they get so much traffic still. And I imagine they just help people generate assets based on other assets. I mean, the bulk case is the Reddit thesis, right?

Where you're licensing the data and that's actually very valuable and the stock sort of pops off of that. He sold his ocean front house in Bridge Hampton, New York, for 57 million, according to people familiar with the deal. He bought it for $40 million in 2014.

He's listed the house. The buyer of the identity. The buyer's identity couldn't be learned. Huh. We're going to figure that out. Get to the bottom of it. Let's see. Set on more than two acres,

the modern Bridge Hampton home has a gym, a game room, and a media room. Interesting. Outside there is an infinity pool. You'll love that. There we go. An outdoor kitchen, a fireplace, and a television.

A private lighted walkway that leads to the beach. Hmm. Sales volume in the ham. Hamptons hit a 14-year low, but it's coming back, and the founder of Shutterstock is getting out at maybe the top, but a local boom. Very fun. What else is going on on the timeline? Back to the timeline in the journal as well. BlackRock.

There we go. Search every bite of serverless vector in full-text search, built from first principles and object storage. Fast, 10x cheaper and extremely scalable. Continue. Black Rock stung by loans to businesses accused of breathtaking fraud. an entrepreneur tapped private credit funds to finance his telecom businesses. Now the lenders say he defrauded them. So stock market news over next says we have our fourth private credit collapse in two months.

And this one shows exactly why institutional lenders shouldn't be justified to verify their own collateral. Bekeem Bomb Bot ran Kariox Capital, a New York telecom financing outfit that convinced BlackRock's HPS investment partners and B&P Paribus to lead him $552 million. The entire deal was built on one claim. He had legitimate receivables from T-Mobile, Telstra, B-Vix, Telecom, Italia, Sparkle, and Taiwan Mobile backing the loans. None of it was real. Brom bought forage contracts that appeared to be signed by representatives from these

carriers. He created fake invoices supposedly issued by these companies, claiming they owed Karyok's money. email addresses mimicking these carriers real domains and sent verification emails to make the receivables look legitimate by stacking these fabricated invoices on top of each other he created what look like 500 million plus in collateral lender sought assets and funded the deal without catching any of it here's where it gets embarrassing when hps and bnp paribus finally tried to verify these receivables by actually calling t-mobile the carrier said they had no idea what carry ox was talking about no contracts no no invoices nothing existed

one phone call would have instantly revealed the entire fraud. These are supposed to be institutional-grade lenders with world-class risk management, yet somehow they missed basic due diligence. While all this was happening, Brambots, people were also stealing cash whenever payments came through the lender-controlled collection accounts. Instead of applying those funds to the debt, they diverted the money offshore, so he wasn't just fabricating collateral. He was stealing actual cash flows in real-time.

Lenders sued in August, 2025. and froze all assets. Then they filed for the company, Cariox, filed for bankruptcy with up to $1 billion in liabilities and basically zero assets remaining. HPS is sitting on half a billion in losses with nothing to recover against. The kicker is that BlackRock acquired HPS for $12 billion in July 2025, specifically to expand into private credit and get access to its $148 billion platform. Within 90 days of closing that deal, they're holding a half a billion.

fraud loss on receivables that HPS supposedly vetted and monitored. I mean, that's not that big of a deal, right? Half a billion. I mean, it's not great, but, you know, it is like 5% of the deal. Like, I feel like people get deals mispriced by a couple percent every once in a while. I mean, it's just, this is a big deal in the broader context of, like, private credit has exploded as an asset class. Yeah.

There's a lot of, you know, people of, you know, we saw this with the, what was that, what was the, A.V. in the chat. Large-scale fraud is actually a sign of good times. I mean, that's true. But, yeah, who knows? But anyways, that auto parts company that went bankrupt a while back also had a bunch of private credit in the mix. And so I think people are just waiting for more and more. We got to dig into that story of the satellite repossession.

Maybe one of the... I don't know about this. Private credit gets crazy. I mean, there's a famous example of one credit firm buying Argentinian debt and then repossessing a aircraft carrier or destroyer, like a military boat. And apparently the same thing can happen in space with satellites. Like they are, it is possible to more or less repossess them. And so I think a lot of these, a lot of these deals, a lot of these private credit deals are like, you're, you're on a knife's edge in terms of like, are you going to continue to get payment or are you going to need to get the value from the assets?

Whereas in venture capital, it's just so much, it's just so much further out on the risk curve that you're, you're never like, yeah, am I, like, a big piece of my business is, is making sure I can liquidate the Apple display monitors, you know, in the office or the Herman Miller chairs. It's like, no, that's going to be basically immaterial to, your business, it's going to be the difference between, you know, like, 99% of the value is going to come from 1% of the investments anyway, and then of the money you lose, you're not going to be recouping an extra 5% making a difference. In some more positive news. Google AI Studio. That's right.

Create an AI powered app faster than ever gets started. Gemini understands the capabilities you need and automatically wires the right models and APIs up for you. You can get started at AI Studio. AI. Studio slash build. Britain over at Microsoft says, please, Mr. President, respectfully let us buy this truck. Let's go.

He was responding to. Wow, 64,000 likes on that. And Palmer says, working on this. Working on this. It's got another 20,000 likes. So people want these little Toyota trucks. But what does this mean in the Palmer context? Does this mean he's going to like call the president and ask Trump to allow it?

Or is he going to make a mod retro version of this car? Is Anderol going to produce this? Like there's like 12 different ways that Palmer could like solve this problem of like Americans not being able to get cool trucks, which is interesting. Yeah, I don't know why. I don't know if these, these trucks are blocked. They're blocked in the U.S. Or just you would have to import one yourself.

Yeah. But they, I mean, Enios is basically, I think Enios has basically done this. They're, they have a, what do they have? It's not the main Grenadier, but they have, I'm pulling it up right now, the quartermaster. The quartermaster from Grenadier looks quite a lot like this truck, but it's certainly, I think this one from Toyota internationally sells for like 10 to 20 grand. Like it's very cheap. Yeah, people love to point out the fact that the Ford F-150 has the same. same bed length as these tiny Japanese K trucks. But interestingly, that's just why you should

always buy the Ford F-250 Superduty crew cab with the long bed. Because if you just go for the biggest truck, then you will be bigger than the Japanese K truck. People love the K-trucks. They have fun to those. Well, if you're starting a new truck company, you've got to get your brand mentioned on chat. GPT. You've got to go to profound. Reach millions of consumers who are using AI to discover new products and brands. Andrew Reed says if I ran a company, my first act would be to ban hard-boiled eggs in the office.

Art-boiled eggs are really an insane snack. I kind of put these in quotes because the smell is so unbelievably bad that you should be forced to eat them outside only. Yeah. It's, uh,

it, the, it's, is it more, it's not more protein than a steak and a bag? you got to go chess key mode. Chesky mode. That's the way, if you want to have protein on the go,

throw a steak in the bag and take it with you everywhere. The hard-boiled eggs in the office, that seems like an odd choice. Probably a time and a place. But name one time and one place. One place would be like, you know how in offices they have like, well, where you can go to the smoking area. Like I wouldn't have a problem with somebody going if there's like a cigarette smoking area, like a back alley.

And you can go into the back alley and scarf, scarves, hard boiled eggs. That seems reasonable. People are smoking less and less. Or go outside, put one in each, you know, cheek and then go back. Like chipmunk mode? Go chipmunk mode and then just pretend like you don't have them. I don't know. Yeah, I've never been big in the hard boiled egg thing.

I have, however, been big into linear. Linear is a purpose-built tool for planning and building products. Meet the system for modern software development, streamline issues, projects, and product roadmaps. Our first guests of the show are in the Restream Waiting Room. We have Lulu Misservi and Gabby Goldberg. Stop and by. Whoa!

Look at this. Oh, shit. Hey. Hey. Hey. You look fantastic. Is that a wig? too?

I don't know what you're talking about. But it's great to see you guys. We were watch shopping during our break between sauna and eating cheeseburgers for breakfast. The lore. The lore is coming through. Hey, John. John. Yes.

Yes. Oh. Yeah. Do you want to show? We're live. You want to show the watch you got? Oh, this? Oh, this is nothing special.

It's just my AP offshore, but it's been on the show of a lot. So, yeah, great to finally cross that off the list. What watch did you guys, 40? Oh, I got the Jaeger Le Cout. Oh, there we go. Okay. So it's in the same color, you know, to be clean. And it matches the new my bach that I just bought.

You know that's my word. My bach. My bach. My bach. It's the my my bach is outside. and it matches this watch. And really sorry, guys. Let's get right into the show.

John, you ready? Let's do it, Jordy. All right. You're watching TVPN. There we go. Today is Friday, October 31st. We are live from the TVPN Ultradome. We even got the horse.

Technology. The fortress of the scientists. The capital of capital. They got the whole stag. What is that? Is that the soundboard? What is that? Oh, the air horn.

Oh, the air horn. Happy Halloween. Happy Halloween. Well, well done. Good to see you. The chat is absolutely loving this. How did you? Big fancy both, Ily and Mark.

How did you guys advise? Oh, wow, with the loafers up on the desk, is crazy. The level of detail here, the schick is remarkable. Here, we'll fire. We'll fire back. Someone else is a shtick to us. It's a lifestyle. On a more serious note, what's the advice to clients on how to approach Halloween?

Is it just ignore? It's a distraction? Nope. It is with every other day a chance to only lock in more. We actually, yeah, we printed out some examples. Yes, take us through it, please. What else you got for us? John, you're disappearing.

There you. What is this? Oh, this is Halloween. The Halloween starter kit. I've seen this, but what is in the image? Okay, so this is Spirit Halloween, the Brian Armstrong set. It includes one great... For the whole head, right?

Oh, perfect, perfect. Yes, yes. The whole head, everything below the head is strictly optional. Wait, why didn't we consider that? We were in, we were in making... up for three hours this morning. We could have just shaved her head with a razor in the morning. That would have been so much easy. Would have done 90% of the work

for us, really. And I could have just, I could have just, I could have just grown a beard. Yes. What, what, what other ideas do we have here? Let's see. So we've got oh, here's, here's one that John and I like. Printed posts. The level of detail here

is amazing. This is insane. Continue. We might have to lose the virtual background, by the way. We're, like, it's really cutting in and out. Oh, yeah. Oh, I can see it. What is this? It's the stressed out guy smoking.

Gay Halloween. What do you mean? You're stressed out Haya Miyazaki. Look at this. This is a fine. Yes. That would be a fantastic costume. Oh, someone did it.

And that's what that post is. Someone's real possible. That makes sense. Yeah, you see you're kind of tripping in and out. It is a wild. And then here, we thought this was very disrespectful of you in particular, Mark. Oh, yes. You see this person, we will admonish them on the spot.

We're looking for them. We did like this comment from the random recruiter. This is my sleep paralysis. I do. I do look like the sleep paralysis demon. That's hilarious. I look a little bit scary. I don't know if I'm doing Mark justice.

Yeah. It's perfect. It's a remarkable, remarkable effort. We did have a. Yeah, who do you think won? I look more like Ilya than he looks like Mark. Who wins the costume contest today? I think it's the Ilya thing.

You think so? I think so. I was getting freaked out before the show. We had like five minutes after the costumes finish. Watching a six, eight, jacked Ilya just like come around the corner is scary. It's a wild, wild experience. Yeah. climbing the, yeah, it's easier to climb the mountain when you're six, eight.

Yeah, it's all in the eyebrows. The eyebrows are really everything. Because as they were working on getting the makeup on, I was like, there's no way this is going to work. Like, this is going to look like party city level. Like, it's fun that we did this, but this is not going to look convincing at all. And it's definitely a crazy. Your post, Lulu, this isn't reposting John's post.

This isn't AI. This is human craftsmanship. But it's actually crazy. So, like, the crew that typically works on, like, you know, normal film productions, they did, was it, like, 50 different layers of stuff that they were putting on our heads? It was one of the most complicated. They're doing, they did, like, a base layer. They did the head.

They painted over our eyebrows. They glued, you know, things over the eyebrows, and they put another set of eyebrows on. Then they painted over that. It's actually, it's going to take, like, probably an hour to. just get all of this off. They said 45 minutes. Okay. I don't know why you'd want to take it off.

Yeah. Do you think, I mean, obviously we're in a unique position because we get to kind of have some fun on the show. It's like almost a comedy show in many ways. Do you think there's any hope to bring back April Fool's Day in a cool way? Or will tech just like flop on that immediately? It's hard to do. Someone told me a couple years ago on April Fool's Day.

day that actually doing it on April Fool's is the cringy part. Yeah, interesting. They're surrounded by so many people doing it badly that it's really from that. And they were saying, if you think of a great prank or jap or joke, just do it randomly. Why wait till April 1st when everybody knows. Yeah, so there was a company that, the YC company that pissed a bunch of people off because they created an IDE that has like brain rot in the software.

So you can do like gambling on steak. you can do subway surfers and like they pissed they they because it's on the yC website i feel like they like they're putting like you can gamble while you code they piss a lot of people off and i think if they had just dropped it as like a feature that was like a real feature that you could play around with but wasn't like their entire brand it would have hit a lot harder and been and and gotten a much more like funny positive reaction versus people just being like i hate this blah blah blah like you know why does this yeah it's almost like taking Yeah, taking the joke like too seriously.

When you haven't established your brand at all, then people, there's just, you're making people do so much work. I guess that's the, I guess that's the benefit of doing something on April 1st is that you're, there's a lot less work to be like, is this a joke?

You can just kind of, you can kind of lean on, on, you know, the audience needs to like, uh, like, you don't need to,

you don't need to, you don't wind up in the situation where you have to explain to like 10,000 people like, we're kidding, we're kidding. we're kidding. Like people kind of put it together themselves and be like, okay, I'm good. Are drops dead as a meta? Do you think? Like, it feels like launch videos and drops are both kind of over. So, so I don't think that drops are dead. I think that that specific style of drop

with the launch video has been over, is completely saturated and overdone. Yeah. I think you just have to, like people are not going to stop announcing things and stop dropping things. I just have to find new ways to do it. Well, I meant drop in the context of, like, we made this product with our branding on it kind of thing. I thought base did a really good job. Did you see bases drop for their one billion dollar round? It was a newspaper. And I thought that was, like, well executed in the sense that, like, it had a lot of text.

It had a lot of content there. Yeah, Tyler has it there. Join the charge. And I just thought, like, yeah, it's like, it's not. just like a t-shirt it's like way way deeper uh and i'd say the same thing for the for the launch videos it's like uh we we found a very unique aesthetic yesterday i'm not going to share it because i think it would be a good launch video of uh but there there are there are specific formats where it's immediately identifiable as like oh that's that type of video and if you use that as

your launch video you'll break through and you'll be the first person to take you know the like right now it's like the default meta is basically like after Game of Thrones stick around for the interview with the actors like that's the aesthetic of most launch videos right now it's like cinematic couple minutes interview with a few people kind of like mini documentary they said we couldn't build enterprise enterprise category no one they said it was impossible yeah they're a little commoditized but yeah yeah fresh ideas at one point i'd said at one point that somebody should do a version of it of those sorority dance videos. That's a great one.

That's a great idea. That's a great one. If you have someone who can do flips or if you just like all wear a matching outfit, like it has to be a company. If you're like solving cancer, doing cancer, that's probably not right. But as long as the joke is not too close to home,

the problem with being a slop adjacent company that jokes about slop. Sure. Now it's so obvious that that is very close to what you're doing and now you have to defend yourself. Yep. Like what was that car company that was going to do that rebrand? a couple years ago. Jaguar? Jaguar. Jaguar was really bad, too. Yeah.

Oh, there was another? There was another one where they joked about a rebrand. I don't even know if it was a joke. They claimed it was a joke because people hated it. What? I don't remember this at all. Claiming after getting a reaction that people don't like something and then saying, it was a joke is not an effective strategy. You've seen that a little bit. The substrate launch was crazy. That was kind of like a Lulu and Gabby masterclass special.

Well, James is pretty... We'll let them know, yeah. Yeah. We'll let Lulu and Gabby know they'll feel really good about it. James Proud is in heavy founder mode, by the way. He did a lot of the writing and outreach and a lot of that themselves, which obviously we think is the right way to do this. Like the founder can't just sit it out and deploy it to do it all for them.

But yeah, substrate, very cool company. Absolute, what we call it? Absolute dog. go. An absolute dog. That's the word. That's what we love to use. He's an absolute dog. Straight from Pat McAfee. All right, guys. Well, have a great rest of your day. Yeah, get back to the gym.

You know, before we go, we should talk about Ramp. Oh, please. What's your go for us? We almost had a whole conversation. Before we go to the gym, let me tell you about Ramp.com. Time is money saved both with corporate cards, bill, paying more. The average company saves 5% worth. There we go. flawless adreed incredible we'll have you guys we'll have you guys do a proper proper guest uh guest host hosting uh session just hosting would be fantastic if uh clearly well if we ever yeah any more some morning where we're like uh we're two hours in the gym not enough today yeah i'll call you guys up and and you can fill in but uh looking sharp uh are you uh hopefully hopefully you guys have some other costumes for for later this evening trick or treating

But if not, you know, tell them to go to TBPN.com all the kids you run into and tell them to. All the children at TBPN.com. We're going to hand out the APs. Capital. Cabby's going to hand out APs, everyone. You got to get over there. Got to get over there. Great to see you guys. Talk to you.

Talk to you. See both. Cheers. Bye. Bye. Hi. We have to do a real ad read for numeralhq.com. Sales tax on autopilot.

Spend less than five minutes per month on sales tax compliance. We have Dean Ball coming on the show next. I don't know if he knows that we're dressed up his mark and Ilya. Tyler, do you know Ball? Of course I know Ball. You can't keep getting away with it. This is one of the jokes that keeps getting funnier. Dean Ball is a senior fellow at FAA, author of Hyperdimensional,

former senior policy advisor for AI at the White House and Strategic Advisor at the NSF, previously at the Mercatus Center. Do you know the Mercatus Center? Of course. Why do you know the Mercatus Center? Tyler Cowan. Of course. The absolute goat himself.

He's been at Stanford Hoover Institution, the Manhattan Institute, focuses on AI, emerging technology. He's in the Restream Waiting Room. Dean, welcome to TVPN. How are you doing, Dean? Good to meet you. Welcome to the show.

Hey, guys. Thanks for having me. Sorry to jump scare you on Halloween. We're having a lot of fun over here. I love it. I, it's crazy. You're, you, you, you look exactly like Dean Ball.

Your, your costume is flawless. Almost to a T. It's amazing. You had to. Were you in, were you in six hours of makeup this morning? Make me look like Dean Ball? Exactly.

Yeah, that's exactly what I said. Make me look like the former White House policy advice. Yes. Yes. Yes. That actually would be potentially the most, like, scary costume is getting prosthetic makeup to look like yourself. Oh my gosh.

that might feel it's a very jarring very bizarre experience looking in the mirror and actually seeing like for the first time of my life seeing something that looks materially different from the way and and i and i will say i i did i did ask john i was like is is dean ready to do it somewhat serious interview with us in extremely unsurious characters and we talked and we thought that uh we thought it'd be a really fun conversation yeah even though um even though we look absolutely ridiculous, but it's great to have you on the show finally. Yeah, great to have you on the show. Thank you so much. Featured a bunch of your writing on the show over the last year. So, yeah. I'm very happy to have you on. Well, thank you guys very much. It's great to be here.

Let's kick it off with sort of what's top of mind for you. There's a lot of different policy positions that I'm sure you could be working on. FAA is pretty big organization. It feels very, very significant to me, especially right now in this moment, but also just feels like the organization generally has been on a massive upswing. What has been the key focus this year for you? And then let's just kind of like continue to dial in the aperture and go and go deeper and deeper as we go. Yeah, absolutely. So, I mean, started the year.

I was still a kind of policy scholar and writing at think tanks, including at FAAI. Worked on state policy, you know, all the crazy bills that were coming out at that time for AI regulation. And then in April moved to the White House. And so I wasn't writing anymore. my focus was really entirely on the AI Action Plan, which the Trump administration released in July of this year. Then a couple weeks after that, once the action plan seemed like it was in a good place and well accepted within the sort of bureaucracy of the federal government, I thought it would take longer to get it rolling downhill, but it really didn't take that long. So once I felt like it was in a good place, I retired from that role and went back to the private sector where I feel like that's my strength, writing, communicating, things like that. What do you think is a bigger force in AI regulation right now?

A, constituents who just don't like AI, I'm worried about it taking my job, I don't like the way it looks, it's slop, it's bad, it's brain rot, whatever, like anti-AI for just genuine direct, I don't like it reasons, versus jockeying between labs of like, if I can get this law through, I will have an edge over you and we're kind of doing jujitsu between both people who are actually both pro AI, but just trying to get a business edge. Is there one camp that you think is like more powerful? I think for the most part, it's the first thing you mentioned. I think it's that there's generally kind of a negative AI sentiment among the public.

It's not that high salience of an issue. So people will say I want AI regulation, but then when you ask them how important of an issue is that to you, it's like 8, 9, 10, you know, immigration. economy is much higher. And once you get down to 8, 9, 10, it means it's not really an issue that that many voters are actually making their decisions at the ballot box on.

And so there's, but there's this general motivation that a lot of politicians feel of, you know, oh, well, we let the internet get away from us, we let social media get away from us, we can't let the same thing happen again. And so we have to be preemptive

and regulate this time. That's basically their attitude. Then why is California trying to kneecap AI? Like, is it just a bigger issue? issue within California? What are the dynamics in California specifically? So California views themselves much like the Europeans do. They view themselves, first of all, I mean, the one big difference is that California is also an exporter of technology, but the

people in Sacramento are not. San Francisco is the exporter of technology. Sacramento views itself really as a nationwide regulator. So their mentality is the federal government's not going to act, especially President Trump, you know, they would say President Trump was irresponsible. So we have to act for him. And, you know, one thing that a senior European diplomat told me when I was in government is when we want to talk policy in America, we fly from Brussels right over D.C. to Sacramento. Interesting. So they're coordinating, you know, the European Union has a, has a coordinating office in San Francisco where they meet with legislators and, you know, companies. And it's basically like this sort of transnational governance regulation happening between.

California and Brussels. Interesting. And so, like, is it correct to frame the idea of the motivations of Californian regulation as, well, nationally, AI regulation might be eight or nine, but in California, it's much higher because there's more visibility? Or is there something else that's driving, like, the California regulation movement? I'm not sure it's a higher salience issue for the voters in California. but I think it's a higher salience issue for the legislators

because the legislators smell power, right? I mean, you can regulate this, you know, what I think will be a multi-trillion dollar industry while it's still nascent. Before, you know, one of the things about regulating now is like if you tried to regulate the iPhone now in a major way, well, like 150 million Americans or whatever use iPhones and so they'd be like mad at you.

But like AI is still a little bit nascent and there's all kinds of use cases that might just be killed in the cradle. and you can control it, you know, much, much easier when it's a younger technology. And you can control it if you're California. You can control it for the entire country, more or less unilaterally. And it's a democratic super majority state. So they have all the cards. So if it's a low salience issue, is there a world where this is all about, like, leverage to get,

to generate more taxes, to create a more balanced budget? What are the different incentives that are working within the California political, like, jockeying that's going on? I think certainly regulatory power is one of them, but I don't really think they're after money, particularly. If you're a California legislator, the basic reality is that money kind of just prints itself. They're basically opposed to scarcity, like petro state. Okay. Because of technology. But the pet...

Yeah, the petro dollars are not oil. Instead, it is capital gains taxation. Yep. I was once talking to a California legislator who was on the budget committee, and they get these like, you know, every other week or whatever. They get these like reports from the state treasury about the state's fiscal position. This was a couple of years ago, sort of right post-COVID. And they were like, we had a $10 billion surplus in the last like month. And they were like, where did $10 billion come out of nowhere?

And it was the SNAP IPO. Yeah, wow. It was all, just one IPO generates that kind of revenue for the state. So, you know, and it's going to be like that on steroids when, you know, when the frontier lab employees start to have their liquidity events, the state is going to have an enormous windfall of capital gains tax. I mean, yeah, there was a rumor just earlier this week about opening eye going out at $1 trillion, which would be. Yeah. Yeah. And already already the employee pool of the public, the PBC is worth roughly at current marks like $130 billion.

So multiple in that, it's going to be substantial. How does this power struggle between Sacramento and D.C. around regulation play out going forward. I think people like SACs have been, you know, not only talking about what Anthropic is doing, but I imagine doesn't want AI to be regulated, you know, at a state level. But how do you see this playing out over the next, call it, six months? Yeah, I don't think anyone wants to see AI regulated at the state level, except for the California. California legislators. And so, you know, at the end of the day, I think that President Trump and AI's our SACs have made their position clear that we want a nationwide legal framework to govern AI. The question is going to be, can Congress act? You know, can Congress get together and come up with something that sort of everyone can agree to? And I think we don't know the answer to that yet. I think, you know, it's hard to see where the politics will go.

AI politics involved, they don't involve quite as fast as AI, but they have all quite quickly. And so we could be in a totally different, there could be a total vibe shift between now and say June. June's probably your cut off before the midterms to sort of get something serious done. And so yeah,

can it happen? In principle, sure. But we'll, you know, we'll see. What are the, I don't know, like what are the regulatory frameworks that either you think will actually land on or California is advocating for? I just remember,

we were watching the Elias Sitzkiver interview with a Guardian a couple years ago. And the point he was making was, like, in the long-term AGI might see humans like we see animals and just convert the entire Earth, like, you know, all the land into solar panels. And, you know, sort of like a complete AGI doom scenario. That doesn't seem like the focus of regulators today. He also was mentioning the risk of fake. news and the idea of generative content potentially muddying the waters of understanding what's real on the internet. That one feels like it's totally here today, but has also kind of always

been here because of Photoshop. It is such an interesting, you know, tech, I don't, correct me if I'm wrong, but tech has never gone through like a tech lash where there's so much video footage of the key players talking about the doom states of technology. So if you look back, go look at interview footage of Sam Altman and all these people, and they will say, like, yes, there's, you know, or Ilya that when you're referencing, like, it's very possible the world will be covered with solar panels and data center. Yeah, yeah, yeah. And we were playing that video this morning while we were getting this crazy, you know, uh, you know, uh, costumes on. And like, the, the people that were

doing it were like, that sounds horrible. Like, they listened to the whole thing and they were like, I don't want any of that. And so it's not like during social media, there wasn't footage. of Mark Zuckerberg being like, it's very possible that the future. The election could be stolen. Social media won't be social. The Facebook data could be used to target voters or something like that.

And so that's just like incredible ammo that I imagine in Sacramento that are pulling up and it's like, okay, you want to go with what this guy wants and not what are. Yeah. What's the dynamic there? Are they actually pulling up Dumer clips?

It's hard to be like, oh, I didn't mean that. It was just kind of a sci-fi thing. Yeah. It helped with fundraising, actually. It really depends. So there's people in the California legislature and elsewhere that have sort of very mundane concerns. And that's like sort of your modal policymaker, I would say. That's kind of, yeah, your misinformation, your deep-pakes. The things, not that they're not serious, but that are here today. And then there's the people that are thinking about these sort of very far-flung doom scenarios. There's not that many legislators who do that, but there are a lot of people in, like, quote-unquote AI policy who think about that. And so they do influence some of the legislation. One of the funny things, though, is that thus far, the legislation that that camp, the sort of AI safety camp, has been able to get through, has actually been like somewhat more measured than some of the crazier stuff that we've seen from the people that don't know anything about the technology. I just kind of have a generally hostile attitude, like state of Illinois banned mental health chat bots,

which is basically like if I ask chat GPT, you know, if I say like, hey, I'm having a bad day, can you like tell me something that make me feel better? in Illinois, that's like a chatbot engaging in mental health services and it's illegal. That's crazy. Yeah, it's like really, it's really crazy. Yeah, that's not like a doomer bill, you know? No, totally. That's like, that's like people with, Nashville's been going on their own path with, with, with, because of how big the music industry is there. Yeah, can you get us up to speed on that like the regulation of AI music? We were talking about Suno earlier this week. I think we're going to have the founder on, but we'd be very interested to hear the, the, the, how that landscape differs from other. regulatory pushes. Yeah, so you're talking about something called the Elvis Act, which passed last year. And, yeah, basically, that's a bill that's about, like, name, image likeness, basically saying if you're a celebrity residing in the state of Tennessee, you can't,

if someone makes an AI model that, you know, mimics your style or mimics a song of yours or your likeness or something like that, then you have a right to sue them, not for copyright, because copyright is a federal issue, but name image likeness is a sort of property right that's like done at the state level. Yeah, that's a creature of state law, but the most part. Yeah. Yeah, so that's what they control there. Yeah. And is there any early read on where that will pencil out? We were talking to, I forget who we were talking to, we were talking to somebody about how that is, oh, it was Zach Kukoff. He was talking about, the like how the elvis act plays into national geopolitics and some some players who are on the national

stage are wanting to win in the in in at the more local pretty smart pretty smart naming the elvis act sounds like something that anybody would want to pass yeah cool you don't think you would you don't think you'd go farther if it was the gunna act what's going on uh you had a post it was last week calling out uh there's a group uh max uh tag mark said a stunningly broad coalition has come out Skynet AI researchers, faith leaders, business pioneers, policymakers, NATSEC folks, and actors stand together and they say, we call for a prohibition on the development of superintelligence, not lifted before there is one broad scientific consensus that will be done safely and controllably and two, strong public buy-in. It's kind of a crazy thing to try to require,

from my view, because even the scientific community cannot agree on much and everybody has these like wildly different you know strategies and incentives and of course it's very possible that uh the public will never uh that to rely on the public of how we should develop technology that has to do with with national security and and try to make it like a a direct democracy uh but but maybe unpack uh what what the latest is on that front yeah so i mean the statement is very simple it says no superintelligence until we have agreement that it's good until there's broad public buy-in. So I think the first problem with this is, like, look, if you want to have a letter signed by a

bunch of important people or celebrities and whatnot that says superintelligence is something that has profound implications and we should really think about it and we should have groups of, you know, civil society type of stuff, that's fine with me. There have been letters like that in the past and I haven't complained about them. But when you say ban, you're making a ask about public policy. Not only that, you're actually making an ask about global public policy because, you know, the superintelligence could be trained in China, could be trained in the United Arab Emirates, Saudi Arabia. A lot of other places are amassing large amounts of compute. It's not like we're the only people that can do this. And a lot of the researchers, you know, who do this work are highly mobile.

So, like, you need a global treaty to ban superintelligence. That's what they're asking for. And the problem is that not only is there not scientific consensus about the safety of such a thing, more importantly, there's not scientific consensus about what superintelligence is. If GPT-5, or let's say GPT-7, say that it can make novel advancements in mathematics, it can come up with new theorems and whatnot, come up with new science. But say it also can't autonomously plan a wedding for me. I think that's a very plausible future, by the way. I think it's super possible that we'll get advanced, you know, genius mathematic AI

before we get sort of an agent that can like plan a wedding for you over six months? Truly, truly agentic AI. Yeah, yeah, yeah. And so like what, you know, which one of those is super intelligence, right? No, no, I mean, to take it even more extreme example, we were reflecting on the original open AI definition, which Ilya was like reflecting on, which is AI, you should be able to tell it to do a job and it should be able to do that, any job at the level of a human. And we were sitting for three hours getting makeup done. And we were like,

okay, to replace this job with AI, you need like the most flawless humanoid robots because this is not a digital task. You need dexterity, but also understanding the artistry of how the colors blend and mixing the different chemicals and stuff. Like, it's just like, it's just like, and that's for like, you know, just a, it's just, like, it's just, like, it's just, like, it's just, makeup. It's like it's not it's not as complicated as the IMO gold medal and yet because it exists in this cross-domain world, it's somehow more challenging. So yeah, it's just a fascinating thing that we do. Yeah, the other dynamic that's interesting right now is like, like, I feel like the industry like is somewhat going through a small period of disillusionment when like the big labs are,

you know, or open AI doing like SORA and announcing erotica and stuff like that and realizing like, wait, like, I thought we were building, you know, super intelligence, like, why are we, you know, creating more forms of social media? But that's at the same time that you have, like, peak sort of like fear and, like, peak pressure to make moves from a regulatory standpoint because people are afraid of, you know, runaway super intelligence and, you know, losing their job and all these things. And also, big tech companies have been doing layoffs. And the CEOs want to say, like, well, we did these, like, as a CEO, you want to say, like, well, we're getting so much efficiency out of our, you know, AI. Like, of course we're doing layoffs. But in reality, it's like

they just went through, they've been bloated for a while. There's, you know, X is a good case study of, like, you can cut, you know, like a lot of people and still have a functioning business. And so, I don't know, there's just such a complicated moment. Do you think there's a massive tech clash coming? Uh, yeah, probably. would be my guess. I don't know exactly what'll cause it. Maybe it'll be fears about labor market, which could be real, could be not, could be caused by AI, could be not. Could also be electricity prices. I think that's very possible. And so like, yeah, yeah. So again, like the super intelligence ban, like, it's doing populist politics that are based on kind of vague,

emotional sentiment. And that's like, the problem with that is that those kinds of broad statements that implicate policy tend to not lead to good places. fun the police was one of them. Net zero by 2030 was another one where it's like we have no real plan for how to do this but we're going to articulate some broad goal that sounds good that everyone can kind of you know you can just slide down the incentive curve um of politics and um it's it's it will not lead to good places I basically guarantee you yeah that makes it doesn't even if you're worried about AI safety even if you're worried about AI safety uh if you're if you're to rank the the the reasons we like if we rank like

AI safety is like the eighth or ninth most salient political topic. And then within AI, you rank the most salient worries, job losses, energy prices, those are at the top. How far down do you have to go until people start caring about dead internet theory? And like the sloppification. Because it does feel like it's aesthetically like risky because if you launch a model early, it's sloppier. It doesn't have great aesthetics.

And then you're faced with the question. of like, so my internet went, my electricity prices went up and I got slop as opposed to my internet prices went up and yeah, I actually got like a kind of superpowered Google search. Like it's actually pretty great. Like it's definitely helping me at work. Yeah. Yeah. I mean, I think it's a little bit of both. I think there's a lot of people who say like, deterioration of the internet is definitely happening. I think that's true, right? I mean, I think it's, the current experience of the internet has gotten worse, I would say. I mean, with the exception of the AI models themselves, which are great.

But yeah, I mean, that's like probably true. I think it's up there. I think it's way above, like, catastrophic risk, you know? Like, I think if you were to put like autonomous cyber attacks, you know, bioweapon development, I think that's probably lower than this sort of slop stuff. I would guess. Yeah, it's interesting to see. I don't know if anyone's ever done that, but it would be interesting to see.

So we were reflecting on this. This week we had Christina from Vanta on talking about security. And then we also had the CEO of CrowdStrike on. And we were kind of blown away reflecting on we've been in the AI boom for two years at least. And we haven't really had a moment where we're like, oh, yeah, that major cyber, that outage, that cybersecurity incident, that problem, that hack, that virus, that's generative AI uniquely. That hasn't happened. Like, ADBS went down last week and it was just misconfigured DNS on a database.

There's been a bunch of crazy cyber attack story. none of them have been, you know, related to Gen. I just yet. I still think it has to be coming at some point. But until you get that solid anecdote, it feels very hard to go and run the campaign of like, never again will we allow this to happen when it hasn't happened for the first time.

That's how it always goes with tail risks. You know, people can't really internalize them until they actually happen and then there you go. I actually wouldn't shock me if we see something like that in the next, you know, six to 18 months. Because the models are now getting good enough that that kind of thing seems possible. Cyber or biological?

Cyber. Bio is going to take a little bit longer, but cyber. Cyber. Maybe even cyber might take, or bio might take much longer because, like, actually making a bio weapon is still hard in the physical world, but cyber is purely digital. There's already many skilled cyber attackers,

so it feels, it feels like plausible. Luckily, you know, cyber defense is keeping up right now. We've seen a lot of advancements there, too. Yeah, yeah. There are definitely like script kitties that go around and just like to unplug different pieces of the internet for fun. And it's almost like it can be state actors, but it's also, it just feels like low stakes. It doesn't feel like often like these like little cyber attacks are like life and death. So they feel lower stakes.

They feel more like theft and like economic crimes almost. But being like I'm going to go release a bubonic plague. Like that feels like way more like act of war. or like a mass murder. Like it puts you, it takes you from like script kitty who wants to like, you know, deface a website with a DDoS attack to like, okay, you're doing a mass shooting now. You posted a couple. It's interesting how.

Yeah, yeah. Let's get a response to that. Yeah. Oh, I just think it's interesting how like these cyber risks though, like we think of them as like, yeah, it's not an act of war. It's just an economic crime because it's like purely digital. At the same time, like, if North Korea had a bunch of people that were like going into hospitals every other week and like stealing medical equipment or like taking people's

patient records. We'd be like, yeah, we're going to kill you. We're going to destroy your country if you do that to us. So it's weird. It's interesting how people think about these things and sort of compartmentalize them. Yeah. You posted a couple days ago, my sense is that selling Blackwell chips to China would quite possibly be the most unpopular tech policy move of the Trump admin, especially on Capitol Hill. It's plausible that the long term, really even near-term result will be much more compute regulation. Any updated thoughts on that front? Blackwells didn't, I guess,

come into the negotiation in South Korea, but you're still expecting them to be a part of some sort of deal? I think it could very well happen eventually. And by the way, the wisdom of doing it changes over time, right? Because the chips get worse over it. The chips, you know, there's new chips come in and they get much better.

And so, you know, like the idea of selling Blackwells to China in two years is very different from the idea of selling them today. I think part of the reason that it seemed like many people in the Trump administration were opposed to this move was that like, you know, we're racking Blackwells in American data centers right now, right? Like it was different with the Hopper because the Hopper was a previous generation. So the age 20, maybe the pill went down a little bit easier. But I think with this, it's like, yeah, I think I think that, you know, you have to, I think it's on the merits and politically. You got to wait. Yeah, you got to wait. Yeah. What did you think about basically giving American companies a rofer on the Blackwell? Is that something that you think makes sense? Uh, you mean like this is like the, yeah, yeah. Basically like every American, if every American company says, we're good, we got enough. Yeah. If Infinias says there's no more American companies that want to buy from us, well then yeah, you can go sell some to China. That feels so much more reason. This is like the, yeah, there's a bill in Congress right now in the NDAA. the end-of-year defense bill called the Gain AI Act that basically tries to do this.

You know, I think it's an interesting idea. I think the problem always is like, how exactly do you operationalize that? Particularly because there's things, you know, in this that are like, well, the public, the hyperscalers are going to have to disclose things kind of to the public about like, you know, that's sort of sensitive proprietary information. So like, how exactly do you do that? I personally have not been super in the weeds on gain. But like, yeah, I think that I think the implementation is a challenge there.

And that's kind of what I was saying, though, is like, you know, so many people in Congress are opposed to the idea of selling black wells to China right now that I think if the admin had done it in Korea this past week, I think there's a very high chance that you would have seen maybe some legislation with unintended consequences pass. What about the idea of you don't sell the black wells, but you do allow Chinese companies to purchase cloud computing resources from American hyperscalers? Is that more palatable? who gets upset in that world? Well, that's the status quo right now. Sure. You know, that's basically what happens right now. And so...

Wait, wait, but could DeepSeek or High Flyer actually go to AWS and say, we want to do, you know, a billion dollar training run? They could, yeah. There's nothing restricting them from that. Now, I think Deep Sea kind of rolls their own. I think they build their own data centers and do everything themselves. But, yeah, like, I mean, Chinese companies definitely do make you. of American compute. Now, that being said, I think the Chinese government doesn't like that.

I think the Chinese government does all kinds of things to incentivize them from that, so it's a complicated thing. That makes all that sense. What's your approach to truth-seeking? One of the challenges in trying to even conceptualize what the right policy approach is on a number of these issues, is that everyone is just heavily conflicted. Like, there's just so much money sloshing around. It's like the people that have the most knowledge of what's happening at the frontier quite possibly have somewhere between 10 and, you know, a billion dollars kind of like riding on a certain approach. And so I'm curious to know, like, kind of how you try to clock like the truth. I always try to start from technical fundamentals.

So when I decided to get into writing about AI policy, one of those. the first things that I learned to do was like train a toy neural network from myself, just so I could see how things worked, right? And I like to really understand the sort of roots of the technology that I'm talking about, whether that's the physical world or the digital world. And that helps a lot. That helps ground you quite a bit to understand those sort of actual constraints and where the technology might actually go. And then beyond that, you know, it's listening to a lot of different people. It's understanding financial incentives closely, right? You have to understand how everyone is incentivized. And you kind of know, like, there are a lot of areas where incentives are for you to sort of spin reality in a certain way. Those are a lot of other people who have incentives to just basically tell the truth. And you have to just be able to differentiate between those things. And so, yeah, you know, I mean, it's like, it's basically just, you know, there's no simple way. It's just read a ton of stuff, listen to a lot of people, maintain an open mind and be humble all the time. Yeah, that makes sense.

Uh, what, uh, neck, like, like, how are you forecasting next year? What do you think the most important issues? Is it more, is it more of the same? Uh, is it, is there going to be a new conversation that you kind of see on the horizon? You know, it basically like is the debate around, uh, you know, chip sales just going to continue throughout the entire admin or do you think it'll at any point settle? Uh, I think that debate will continue. And then I think, you know, more broadly, though, my guess is that that issue becomes less salient in AI policy over time. And probably more people join the sort of join the debate. Right now, AI policy is very much an elite conversation.

You know, it's not something that a lot of people, like the average American, so to speak, is thinking about that much. But I think that starts to change. Again, maybe it's the electricity prices. Maybe it's the jobs. Maybe it's a little bit of both. Maybe it's also some crisis that happens, like cyber stuff. But I think that, like, in general, the salience will go up. And it will also be quite complicated because I also expect that in 2026, just like this year, a very meaningful fraction of GDP growth will be coming from AI.

So, you know, policymakers are going to find themselves, I think, in a bind. Oh, yeah. Sorry. I know we're wrapping up, but I'd love to know your thoughts on, like, the AI as the only driver of GDP meme. real do you think that is? How bad do you think that is? Yeah, it's interesting. Some people hear that and they're like, this is bad. And then the other take is like, if we're not doing this, like, it's even worse. Yeah. Yeah. So I think it is a big chunk of it. I, you know, I think it's probably at least like 25 to 50 percent of growth right now. Another big chunk,

like the Q2 number was really high. It was like 4%. And a chunk of that is definitely companies stockpiling in anticipation of tariffs. So they're like mass importing stuff, which is going to be counted toward GDP. So it's like there's some of that going on probably. And but yeah, I mean, I think that it's almost certainly the case. The GDP, you know, if I had to put an,

if I had to guess, I would guess it's something like 40%. AI is something like 40%. And of course, that's, that's, yeah, and that's just enabling AI. That's just AI CAPEX. That's not even, you know, what is AI enabling. Like business services and accelerated revenue growth.

Productivity and stuff. How much of Google's revenue should be counted in like AI GDP now or like Instagram Reels? If there's an AI slop video that goes viral and that generates 20 million in AI ad revenue, are we counting that in GDP? We have no idea yet. It's going to be years before we understand it. Last question that I'd be curious to get your real. on is sentiment around. We're a couple months out from the Intel deal. How is Washington and like Capitol

Hill feeling about that deal? Obviously, it looks good on paper for the taxpayer, but do you expect to see more of this kind of thing? We've sought with MP materials as well. But what's your read? I think you'll absolutely see more of it. More equity positions. I'm long USG equity positions in private companies. For better and for worse, I think that will continue to be a thing. We've seen it throughout the critical mineral supply chain.

In fact, a lot of other companies, not just MP. And how new is this on, is this a Trump thing specifically? Is this a, is this a 2020s thing? Are there other examples of Americans doing this? It's not new under the, Sun, but it's new in recent history.

Okay. Particularly for companies that have like sort of consumer brands. Sure. But yeah, like it is, it is, it is, this sort of national champion model is not something that's, it's not, again, America did it, you know, a lot. You know, there are periods of American history where we did this a lot. And not in a total crisis. Like, like government motors, general motors, the bailout of the banks, like that is very different then, okay, Nvidia and AMD are doing pretty well.

Intel's not exactly going bankrupt. There's a lot of market cap there, but we're still going to come in and do something. I think Intel's position was really precarious, and you really needed that. I mean, I think you needed not just the injection of capital, which they were going to get either way, but kind of that vote of confidence

and kind of that indication to the market that, like, hey, this really is like, you know, the U.S. government is serious about this. That says things to open AI as they think about developing their ASIC. It says things to Apple. It says things to invidia. It says things to all sorts of people. And you've already seen some of these companies respond in kind with new orders to Intel or at least partnerships with them of various kinds. So I think you'll see more of that. In terms

of Intel itself, I think people in DC are pretty nervous still because Intel's got a serious hill to climb, you know? They've got a promising roadmap and they have to execute on it. And thus far, the execution has been shaky. And their track record shouldn't make you confident. Have you been surprised that we haven't seen an open AI Intel deal yet? Sam Alman seems ready to do deals with everyone. Intel feels like it needs a deal. It needs several deals.

That feels like a match made in heaven. And yet, I haven't seen it yet. What do you think? I have to think it's gestating. I have to think that probably other companies too. I mean, if they're smart, they will. Yep. Yeah, totally.

Just as a political matter. I mean, there's just so many, like, I've seen so many pundits like Asianometry laid out the thesis that Nvidia should dual source their CPU for Grace, for Grace Hopper, from Intel, make the CPUs there, don't go to TSMC for those, support Intel as a fab and a provider. So, fascinating, dynamic. And I feel like the Intel story is going to be one of the most. Interesting. Any plans to write a book? You've written a book? You know, I've been, uh, no, I haven't. Okay. No, no, no, but I'm thinking about it. Uh, we'll see. More, more to come. Maybe. I'm,

I'm pondering it. Hard to know what to say. It feels hard. It feels hard if you're kind of like writing about what's happening to like pick something and like start writing, you know, hundreds of pages when like everything's so in flux that you could write like, you know, six chapters that then maybe made like, that weren't even an important part of the story. Sure, sure. Yeah. If I picked up, the phone and called an agent today, you'd probably be looking at a publication date of like Q1 2028. Someone just the other day, I was asking like, why the hell of these supply chain so long for, for like, why is the timeline so long for books? And they told me it's because it's a manufacturing business run by humanities majors. No way. It's like, yeah, that's hilarious. Yeah, that's funny.

Maybe we, uh, maybe the answer is Kindle print on demand or something like that. There you go. Yeah. Kindle digital publishing, right? next to all the literatica. We can slot in some geopolitical analysis. The AI slot books. Yeah. Oh, there's, I'm sure there's so many. Somebody's going to hit you up and be like,

I can generate you 10 books by next week. Where can people find you? Where can people follow you in the meantime until the book drops in 2020? You heard it here for a break news. I've got a substack at Hyperdimensional and on Twitter at Dean W. Ball, ball like football. There we go. Thank you so much, Dean Ball.

Great to finally have you on. Welcome back anytime. Chat all the time. Thanks guys. Great to chat. Cheers. Have a good one. Bye.

Let me tell you about fin.a.i, the number one AI agent for customer service. Number one in performing benchmarks, number one in competitive bakeoffs, number one ranking on G2. And breaking news from Sam Altman, GPT6 will be renamed GPT67. You're welcome. He's doing the six-seven meme. But it's a joke. It's a joke. Okay.

Wow, that one really did not hit. That was remarkably, remarkably not. Oh, it's a joke. Not funny. I, I, this got 30,000, 37,000 likes. I thought you were going to say, I thought you're going to say it'll just be the chat GPT 20, 20, 6. I would like that. I think that actually would be a better naming scene.

No, he's having fun with the, with the 6-7 meme. go through some timeline with Tyler. Not my schick. Not my schick. Well, I will first go through Adio. Customer relationship magic. Adio is the AIA Native CRM that builds scales and grows your company to the next level. You can get started for free.

Let's see. What in the timeline should me and Tyler talk about? What is here? Tesla Roadster, Sam Altman has asked for a refund. he says a story in three parts, a tail in three acts. Act one, your roadster reservation is complete. This was in 2018, long time ago. Tyler was five years old.

He put down $45,000 as a reservation payment. Then Act two, Sam Altman emails. Reservations at Tesla. Hi, I'd like to cancel my reservation. Could you please refund me the 50K? and he's just trying to, he's just trying to ris an extra 5K out of Elon. Did you notice this? Oh, oh, because it was originally 45.

It was 45K. Now, okay, with interest, with inflation, Sam clearly deserves 50, but it's a wild move to just have 45K on the line and be like, can I get 50? Can we round to the nearest 50? But when you're, you know, Sam, I guess it makes sense. It's probably pretty reasonable. So he says, can you refund me the 50K?

I think he's actually just rounding. I think he's like, yeah, send me 45, but I'm going to refer to it as 50 because I'm rounding to the nearest 50k, I guess. But then he gets an address not found. Reservations at Tesla.com is unavailable to receive emails. He could use the extra five. You think so? I think he could put it to work in the business. One more blackwell.

Nobody's losing more money than. an open AI. Every 5K can go a long way. That's true. Yeah, Tesla and SpaceX seem to be doing... That's one minute. Less capital. That's one minute of SOAR costs.

That's true. Every minute counts. But, I mean, this just feels like they changed whatever the email is in the last seven years. But it is remarkable that the project has been in process so long that the email address has been forgotten. And also, I would assume that this type of stuff, happens all the time because I imagine that people are canceling and placing new Tesla Roadster orders

all the time. So you would think you'd want to be monitoring that box. Well, it's funny to read into it as like more battles between Elon and Sam, although I don't think that's what's going on here. It's more of just a funny quirk of their email configuration. Burn Hobart says, wow, Elon passes the replies to your email in under a minute test. That guy is going places. Bullish. Well, whatever you think about Tesla, you can buy it. You can sell it on public.com investing for those to take it seriously. They got multi-asset investing, industry leading yields, and they're trusted by millions. Jordi Hayes, what would you like to do now? Let's bring in our next guest. Aidan Sedka, welcome to the stream. There he is. You look fantastic. How are you doing?

Hey, Jordy and John. Happy Halloween. Wait, wait, wait, wait, what are you drinking? What is next to you? Wow. Do you recognize this drink? Of course, sir. Cheers.

Cheers to you. That is Andrew Huberman's Matayena Yerba Mata. We refer to it as a podcast in a can. It's hard to drink this on my end. The fake beard is getting into the drink. But it's great to see you. You're looking sharp. Do you wear a suit often?

You look very natural in it. You know what? I used to wear a uniform in high school. It's coming back to me. And I'm a huge Formula One fan. So I feel like I'm already wearing like this is the cap you normally get when you're on the podium with kind of the reeds. So I feel very good taste here. Championship. Championship hat. How did you actually get into venture? It's a good story. I was Google's first product manager and then first international sales manager. always wanted to start something. Both my parents were entrepreneurs, and I felt it's just kind of the way of the value you go from being an operator to being an investor. But interestingly, originally I thought,

hey, maybe the right way to do it is like go get a stint at a venture firm, and the five firms that I talked to not only told me I wouldn't make a good VC, they told me I wouldn't even make a VC. I just did not have the word at all. So I'm sure many people also told you that your idea would not run.

Yeah, what, like, did they have any good reasons? Did you agree with them on anything? Obviously, they were wrong, but... It's so funny. Like, I wasn't an engineer. I wasn't senior enough at Google. I never had done anything in Venture. Like, again, maybe on paper it makes sense. But to be honest with you, it's the best thing that happened because I always feel rejection. I call it rocket fuel. And it was such a simple situation where I had to burn the bridges. the only way to prove the skeptics wrong to actually get into the business and do well. And even after 20 years, I'm like, 10 IPOs is still not enough, 100 exits. I think it should be 200 exits, and maybe then the point will sink hit. Hit that gong. Oh, yeah. Great hit.

How big was your first fund? The first fund, actually, I started as a solo LP and solo GP, which was only $4.5 million. and then our first institutional fund in 2010 was only 41 million. Modest amounts. What a modest amount? What story did folks tell at Google about the funding of Google that helped you kind of understand how venture works? I've heard this story about, you know, Sequoia invested. That was probably more like a traditional round.

There's also this maybe apocryphal story about somebody going to their car and writing a check before the, company was even founded. Like, what was your introduction to just like, Jeff Bezos taking a flyer? Yeah, yeah, yeah. What, what stories did you draw on where you were like, this worked out for this particular person really well in the Google story that I felt and experienced, and I want to go and recreate that? I mean, look, I was there, like, right after the sequoian Klein around. First of all, the biggest thing about Google is that what people sometimes misunderstand about this business is all about outliers,

positive outliers, and the best training for that is to actually be at an outlier. And I think more than the VC story, to me, it was my experience working for Larry Page because every single thing that I've done in that first year for Larry,

he would like see it and I'm like, this is kind of 1% of what I'm expecting. Like, you need to do things 10x faster, like 10x more. And then it's like, you know what, maybe humans are not meant to do that. And my joke is like one of my projects was to launch Google in international languages because I was an engineer.

I spoke six languages. And I'm like, look, you know what I can do for Google is make sure it's launch in non-English languages. And he was frustrated with that. And so I always tell people, I'm an occasional and accidental father or grandfather of Google Translate because they're like, from now on, anything that's internationalization or translation, let's make sure machines do it and not humans.

That makes sense. What was your first true home run adventure? So the first true home run adventure, you can define it in many different ways. My first billion dollar exit was Marraqi, which started as a Wi-Fi project in Mountain View with Google. And I was one of three-x Googlers that invested in it. And the first IPO was Shopify. And that was amazing, like showing up to Nice, not for Google's IPO, but as an investor that, hey, like, you know, early believer, and the company was great. And what was crazy is like $2.7 billion valuation at IPO for Shopify. And today there are like 150, like another 50x from that.

Google was similar. Like it went 40x from the IPO. And that like, in a nutshell is like positive outliers. And I can't imagine better examples of outliers. Yeah, that's remarkable. Completely. What about more recently you've done Mercor? Super Base and others.

Maybe we start with McCor. We just had Brendan on earlier this week to talk about the new financing. You guys have been in a couple rounds now. Is that right? Two rounds. Two rounds. We are huge fans of Brandon and his co-founders. Look, I think it's another great example.

When I started right after Google, we told going zero to $100 million in 10 years was a great achievement, maybe eight to 10 years. You know, I had companies like Shopify go IPO in six years, but Mercor is a new breed of company. They went zero to 100 million in gross revenues in 11 months. They essentially have built this great platform where they can essentially find any type of talent or discipline in the world, find the people, have AI interview them, and generate expertise and data based on that.

And it's just like, again, like what's so exciting about the AI age is these companies, you never thought would be at this kind of scale are able to deliver these things because it's just that age and that opportunity. We're very lucky to be part of it. We're huge fans of Brandon and team. And then SuperBase is another great example. We are like, listen,

some of the most valuable tech companies were database companies. And somebody's going to build a developer-friendly backend. At that time, there was a front-end company with no real back-hand company. And it was crazy because when we invested in SuperBase, we did almost a growth around valuation and the traction was not even close to a million in ARR

and since then the company has grown 50, 70, 80x plus in just a few years. And that makes me think like what you're doing is really special because the private companies are growing 500 to 1,000 percent and public companies are growing at 20 percent. So it's a 50X delta. So in case anybody's wondering why it's interesting and exciting to be in venture

or to be involved with private companies, the very best are doing an insane job. And I think those are the stories that need to be out there. And that's why I love my job. I feel like it's the best job in the world. They announced their Series A in October 28th of 2021. I'm assuming you guys did the deal like, you know, some late summer or something like that.

At the time, was that a deal that you think maybe other people would have passed on because of the price? Like, it just looks crazy. Obviously it was the peak. So we did the Series B, and it was really interesting. Not only did multiple other VCs passed on price, but they went and then invested in competitors because they could get in at like half the valuation. But we never compromise what we believe is our belief in the founders and the company.

And look, I mean, at the end, this business is all about early belief in as all the founders. We get to participate in their journeys. And Super Bay is a prime example. They had all the initial signs that things could be really great. Like I said, you know, traction was really low. They were starting to, like, essentially, like, take off among developers. And it was built on Postgres. It was very technical.

And people are like, no, nothing could be worth that valuation at that stage with only that much signal. And I think that's kind of the magic of Felices is that when we back these incredible founders, you know, the success is far from obvious. But then after we invest, like these companies go through these insane hockey stick growth curves and Super Bay is a prime example of that as is Merckor. How do you think about what is different about this boom, this moment from previous cycles? The top things that have really changed in the last two years that justify more investment, that change the job of the venture capitalist are, you know, we're seeing higher growth rates. We're also seeing higher CAPEX requirements for some tech companies than before.

We're seeing just more potential new markets because you could be going after a labor pool instead of a different, like an existing market. There's entirely net new markets. There's questions about enterprise adoption, churn. Like what in your mind is different this time around from the mobile wave, the cloud wave, or anything else that, in how your underwriting deals this year? Yeah, I mean, look, I have seen the internet wave as well as the mobile wave. There are a couple things that are really different.

So one is the point that we were getting at is that venture game, it is like very basic is essentially, it's a growth game. If you find a company that is growing 50x faster than public companies, it's going to end up in a good place. And even if you're off by 10x on price, if you're in one of those companies, things are going to work out for you.

Also, interestingly, 20 years ago, like when I first got into business, a lot of people would worry about distribution. But the reality is today, like everybody's watching YouTube, everybody has a mobile device. Distribution is no longer a problem. But what is really interesting that is different

is the AI age, normally before, with software, first the budget was limited. I don't think any company is going to spend more than 10% of their budget on software, but also the value creation was different. Like software, yes, it is helping you, but still people are working with it.

You have to put data in it. And then to get value out of it, you have to look at the results and say, okay, now I'm going to do something based on what the software is telling me and the software is like streamlining it a little bit. In AI age, what's really stunning is like, you now can have a full closed loop

without humans where everything is being done by AI. So the data entry is automatic. The data output is automatic. I sometimes think of a turbo engine in a car. One of the interesting technologies is you take the exhaust gas and then if you channel it into the engine, the car can accelerate a lot faster. So in some ways, like what is happening with AI,

there is growth vectors in so many different areas. I had to make an engine example since I'm wearing a Formula One like that. But first of all, there is so much more value creation and now you're going after labor, which depending on the sectors you look at, is like 40, 50, 60, 70% of a company's budget. So the dollars at stake is so much larger. How much more important is it now to think about terminal market structure?

Because I completely agree with you. The growth rates are insane. The risk, I feel like, is not that the growth isn't real. It's just that you pick a market and there's one company that goes from one to 10 million in ARR to 100 million in ARR. and you project that out and you're like, well, the market's $50 billion, they're going to get there really fast. But if the market can support 50 companies that all grow really fast and they all take 2% of the market, then you will see some sort of S curve in the growth rate at some point.

And I feel like there's some markets that will probably be perfectly competitive, some that will be highly monopolistic and everything in between. And it feels like as a venture capitalist, it's going to pay to be the type of person can clock how markets play out. Does that map with what you're thinking, what you're trying to do? Look, I think you're getting at a very important point, which is it is not just enough to have high growth, but you need to have durable growth.

And we use the examples of Google, Shopify, even like more recently we had companies like Canva. And I think one of the things that people sometimes really underestimate is that that kind of high growth compounded will generate stunning results. And one of the things that I learned from both Google and Shopify, is that if in when you choose the right markets, you will probably get the sizing wrong,

but if you choose a mega, mega market, even like at what we think success, there is still so much of the market left. One of the things that I never realize is even when Shopify, when IPO, it was less than 1% of that market. And that's what helped it, like, get a 40x growth after that.

Same thing happened to Google, and the very best companies can also expand the market. So to your question, though, one of the things that we really pay attention, is what are the modes of these companies? It's not just that they have the chance of high growth, but how durable will it be? A lot of times it's the founders.

A lot of times it's also different elements. For instance, we very frequently talk about data modes at Felices. One of the things that helps you differentiate is if you have data that nobody else has, since data is the new oil, that's going to be a huge differentiator, because at the end of the day, there's a lot of AI tech, and that's more pervasive. But if you have data that nobody else has, that's going to be tremendous. this value, and if you start doing something where you basically become a key part of the process, people are not going to rip you off.

So we just recently won a term sheet over a healthcare AI company, and what was interesting is through AI, they've been able to service a part of the patient market that here too was not possible to be served at a positive return. And the moment they fixed that part of the market, hospitals are like, like, why am I only going to give you 35% of the market? You know what? The rest of the market, there are 19 players, but I'm just going to come straight to you. It's too much work.

You guys can take 100% of our referrals. So it's yet another good example of if you do a really great job with the hardest part of the market, you can essentially have a much larger portion of the market. Nobody wants, like, is interested in like somebody who's only going to win one to two percent. They want people that can do everything together and not just one part of it. And again, that's kind of also the big leap from software is maybe only solving one part of the problem, and they want like a full solution. I want like a full turnkey, not just like, hey, the software can get you the numbers or analyze things

or be the system of record. Now everything is about system of action, and AI can enable that. And if you have a data mode on top of it, obviously that drastically increases chances that not only are you going to have high growth, but the high growth is going to stick with you. Did the venture industry learn its lesson in 2021? Things got a little bit crazy. There was a lot of top signals. We had a crash,

but at the same time, you did companies like SuperBase at what was then a crazy valuation, and then now it looks like, you bought it. It looks like you stole it. Now this time around, it feels like, you know, that we were, I remember in 2022, it felt like,

like we were going to go into like a VC apocalypse. It certainly doesn't feel that way now. Do you think we learned any key lessons or are we going to kind of relearn them in the future? Let me tell you an interesting lesson or maybe like a key secret is that venture business is you have to be consistent. You cannot sprint and pause, sprint and pause. If you're not growing in this business, you're essentially losing. stagnating. There is no such thing as cruise control or like steady state. And I think the mistake that happens with a lot of VCs, if their conviction is not deep and they're essentially playing

the market game, there are a lot of followers and not a lot of original thinkers. What happens is the moment the market gets risky and things look dire, people pause and like, oh, you know what, like I'm going to stop making investment. And one thing I'm really proud is even in the 2021, at 2021 vintage and 22, when things were really expensive, we still kept on making investments and we never slowed down. And we were basically like steady growth. And part of that is because, look, this is not public markets where if there is 20% growth and your entry price is wrong,

you're going to lose a lot of money. But if you find the right companies and you do have 50 to 100x growth ahead of you, even if you overpaid for them, you're still going to do well. In fact, the only way you're going to do well is to have these positive outliers that are fund makers. And so I think the biggest lesson is, like, a lot of VCs, like, spent too much time of, like, this valuation high and low

and this trend and that, the reality is you've got to be a surfer that can surf any wave in any season and not just one. Like, you can't just go to the beach and call it quits. You need to go back out. And, you know, I know Georgia lives in Malibu,

so he's going to like the surfer analogy. But part of the other thing also is that you have to find the companies where there is going to be so much growth that even if you got the price slightly wrong, it's not going to slow down. So anybody who's like, oh, like, we're going to be in this business and it's all about the price, they're misunderstanding. It is all about like being early and understanding that incredible growth that's going to follow.

That's going to basically give you the long-term success. And so I feel there is definitely a delta between the VCs that are original thinkers that still had the conviction and the courage to make those bets and not just like super base for us, but also companies like in 8N, that was a sleeper company in Berlin, and all of a sudden, in fall, like multiple VCs have woke up that they're doing a great job in Agentec AI.

Yeah, that's a company. Interestingly enough, I saw it on getting, people started making videos about it on Instagram, which I look at as much more mainstream, before I saw anything about it on X, which is, I think, a good sign. It means that it actually has traction outside of our little Silicon Valley bubble.

How do you think about your philosophy on expertise versus like being a generalist? Yeah. There's a lot of expertise on like a category versus like judgment around human. But take me through like advice for a venture capitalist who's joining your firm, advice for someone who's graduating high school and thinking about how to pursue. Like, what is your philosophy on generalism versus becoming an expert in something? I mean, look, I think journalism in general, it's a really good thing. But I think, look, there are a couple really interesting secrets or insights.

Number one, just to give you real data to make this point, is that when I started in venture, the other thing that people underestimated is like, they're like, listen, all these sectors you're going to invest in, you're not an expert in any one of them, and you're not going to be able to make a good investment. I think what people misunderstand is if you're an expert in a sector, you're going to look at a company and you're only going to look at the orthodox way of doing things. And when somebody comes in as an outsider and like, I want to do something the way it's never been done before, you're going to say that's never going to work. And so, like, the paradox here is the expertise, right?

like the experts are supposed to know the field really well, but that also anchors them in the way that things are done traditionally. Versus a generalist is going to be first principles and say, you know what, there's actually a new way of doing things, and if this thing works, it's going to be amazing. And at Feliz is like the most interesting stat that I'm most proud of

is that we made early investments in companies, 53 of them that exceeded a billion dollar in valuation. Not a single one of them was led by an expert. That's amazing. Like that's the thing. Like when you guys started, you were not media experts or like proven like media people and you had a great idea. I would tell you the funny thing is in our case, like I've been working in and around advertising my entire career. John had been an expert on like content production.

But we certainly were not experts on hosting a show. And if you had asked us even literally like 12 months ago, if you had asked us like would you be. Like, would you be hosting, could you ever see yourself hosting, like a daily show? We would have been like, uh, like, I always love that Paul Graham formulation, which is like most co-founders, if you look at the co-founding teams, it's usually like two or three people where each of them are somewhat experts in like two or three disciplines. So you have someone who knows like marketing and finance and operations. And then someone who knows like front end and back end and technical design and architecture and also some of the. great offs about how the product design works. And so you wind up with like, you know, some level of expertise in 10 different categories bundled up in two to three people

on a specific team that seems to work really well. But yes, they're not like pure play experts, which is what we think of, at least what I think of when I think of expertise. A hundred percent. Look, I'm going to add two more things, which is I think one of the other elements of being successful, whether you're a founder or investor, especially as a VC, is all bought to people. And I've actually done sales at Google and you would say, how is that related to being a venture? At the end, it's a people's game. The founders are not, you know, taking money based on numbers, but if they can trust who you are. And so, like, being good with people and being able to relate to them is really important. So these days, when people are starting in the

business, whether they're a founder or future VC or emerging VC, I'm like, look, your biggest asset is your network and the people that you know. But the second thing is, especially for venture, you need to be exposed to outliers, right? Like, I feel like my training at Google, you know, every single thing that Google has done in the six years that I was there, not a single thing was standard. Every single thing we did,

we did a completely different way that has been ever done before, and the results were stunning. So if you've never been at a company where the founders and the firm was crazy enough to do things completely differently, and you kind of understand that culture, that mindset, right?

We did talk about being generalists, but it is also important to have been exposed to some of these things. And then you're like, okay, you know what? Now that I have seen it, it's in my DNA. It's in my blood. I can kind of like see other signs of that. So being good with people, being able to understand outliers, at least being exposed to them, I think are key ingredients, at least in venture. So those are things that nobody ever told me like, hey, this is going to be really important for you. but now after 20 years of doing it, I can tell you that it is very important.

And like journalists, like I worked in every discipline across four continents. I've done finance, sales, marketing, product. That was also really great. I didn't have to be deep expert in all of these things, but being able to understand so you can like see around the corners and find little things, you know, just enough to make sure that you're not going to make a stupid move, but you don't know too much that you would be too anchored in it and you would be skeptical of a brand new way of doing things.

Well said. I just can't get over how good your outfit looks and your frame looks. Yeah, you're ready to co-host with us. Yeah. Come to the altar. Thank you. Like, honestly, like, I'm down to that. I love this hat and I'm enjoying my Nogibate.

Yeah, are you an S-F? We are in Menlo Park, actually. We do have an S-F office. Okay. But we are in the Bay Area right now. Well, hopefully you can cross over soon in person. Yeah, let's hang out soon. Are you going to be at F-1, Vegas?

I'm thinking about it. So my older son is in varsity football. If there is no conflict with this game, I'm definitely going to be there. Okay, cool. Let us know if you are. We will look quite different so you can watch some older footage

to recognize us over there. Thank you for hanging out. And I'm impressed with how that you were able to do a normal interview when we look so ridiculous. But it was really fun chatting with you and thank you for sharing. That was a blast. Should I say Mark and Ilya? But very well done. A great impersonation.

And Jordy and John, it's such a pleasure. Let's do this again. The audio team is appreciative as well. You're looking absolutely crisp. Everyone's a fan. Everyone loves it. Great stuff. Awesome. Talk to you soon. Have a great, have a great rest of Friday. Have a good day. Bye. You too. How did you sleep last night, Jordy? I got a 92 on 8th sleep. You can go to 8th sleep. sleep.com. Get a pod 5.

Code TBPN. I got a, 32 minutes. Perfect. 93. Oh, you beeped by one. I smoked you. Very good job. Well, smoked you.

Our next guest is in the re-stream waiting room. Let's bring him into the TBP and Ultradome. Wait, is that who I think you're, did you get full prosthetic makeup to look like Jacob Deepenbrock from DeScipleis Ventures? Is that what's going on here? That is correct. You guys have discussions better that.

Yes. Your Jacob costume is flawless. How was Demo Day? Take us through it. What were the highlights? What are the new trends? What's the actual structure? Maybe start there.

What's the structure of Demo Day these days? Yeah, yeah. So yesterday, 2 PM got started. We did it Long Walls space here in El Cigendo. One of the OG El Cigendo companies used to be ABL. Shout out to Dan. Piedmont. Started at 2 p.m.

I think we had a hundred and probably 60, 70 people come out for it. I had a panel with Augustus, Isaiah, and Ted. The Godfathers. The Godfathers. It was a great group. It's a great group. Founding fathers. That everybody was hating on SF, which was all the investors in the audience.

Why hate on SF? Oh, well, Paul. Well, so one thing is when we were there earlier this week, I thought they had like somewhat solved the human feces on the ground thing and it was like John and I walked home from uh walked home from dinner granted it was about a mile walk and it was like at least six it was through the tenderloin which is a rough spot there are rough spots in l.A but uh elizco is beautiful at least when I walked around it was it's very clean if you're a if you're an industrialist the smoke stacks you know that's actually a plus that's actually a plus but yeah well uh so how many customers companies graduate. And it's a proper demo day. Like people are given pitches, raising money like that.

Yeah, yeah, exactly. So we had nine companies pitch yesterday from all over the country and a couple from out of out of the country as well. We start off with like said, panel, nine pitches. They all kind of go out to their booths afterwards. And then all the investors get to know and get to talk with them and meet them afterwards. And then they had a sick little after party in Manhattan Beach afterwards. So it was great. But I guess companies, I was happy to run through some. I also posted a thread as well if you want to pull it up. Yeah, yeah, yeah, yeah.

We'll have the team pull it up. But yeah, read through and let us know who's downlined. So I will say less defense companies this time than ever before a lot more. Yeah, I was going to ask, when we, across the different YC batches, like there's very clearly like some core themes and trends that you've seen. So I'm not surprised to see less. We probably like don't, we don't need the same volume of defense tech companies now, I think in general. Like there's just been so many, like every category now.

Oh, you think the job's finished? I don't think the job's finished, but I think it's like probably healthy for like every category now as a few players. Yep. They can now fight it out. One or two will rise to the top. Yep. We don't necessarily need 10 more backfilling, trying to go out for the same types of contracts. Yeah.

I think my general thought is like, I think defense, like, prove that you can kind of disrupt these legacy industries and, like, build massive companies in these kind of legacy industries that haven't really had venture interest until the annual and SpaceX. And I think it's now like, okay, like we will do these, like, hardware, we will do these kind of more, I guess, less tech, less venture heavy industries. There's a lot of stuff like that unlocked. So like with an energy, within mining, we did something like, like, so I always make this point where, you know, Sam Jameco, I mean, SF founder, but he runs Impulse Stove.

It's a stove company. And I don't want to, you know, tell his story for him. But it always feels like, obviously, that's not a defense tech company. That's not an anderial competitor. But that's the type of company that maybe doesn't exist unless we make, like, hardware cool and get found and get, and get VCs piling into the category. But take us through some of the, some of the companies from this batch. Who stands out?

Yeah, I mean, it's a small enough batch where like Take it through all. I'll go through all them. So starting off, we got Petra Power. Shout out to Aaron Goodman. Hopefully you're watching right now. He's doing basically replacing generators. So basically building a new electrical chemical process to replace the generator.

And there's like a massive bottleneck generators, obviously energy. That would be shorter is pretty real, especially with the new data centers going up. And he already has received nine. million dollars in DOD contracts, which is pretty crazy. First one. Second one is a pretty fun one. Making shoes in America, autonomous shoe manufacturing app price

in the USA. That's bold. That's very bold. I'll go to make it in America. Okay. There's all these missile drone companies. We're making shoes, guys. We're making shoes in America. I love it. I love it. Three hedges. They're one of the few software companies.

They're basically building a commodities hedging app for SMBs, typically oil and gas to start. Interesting. It costs a lot of money to actually have like a hedging team. Yeah.

And they said 95% of distributors just can't afford a full team. So they basically make it, their whole thing is Robin Hood. They call it Robin Hood for industrial companies. Yeah, yeah. Yeah, yeah.

We've seen a lot of startups do the sort of like treasury management. Just like earn a little bit of yield because maybe you don't have access to an investment bank and a whole treasury team. And so plenty of companies have offered that type of product, it makes sense that you would need something a little bit different in the industrial space.

Imagine day trading commodities on your phone from LHiganda. Those are that too. What is actinide? Actonide is pretty cool. So basically what they do is they enrich ice chips, basically. So they start off with one material. So there's three materials in the Russian, so there's Russian stations list. This material you can import or cannot import.

there's three materials that are actually removed from that sanctions list, which means you can get as much as you want because we just need them. The highest cost per gram one is a material called Turbium. It's used in cancer medication. They basically make this material. And it's a $2 per gram material. They enrich it. It's $36,000 per gram.

Wow. And then they already have, it's in the tweet as well. They already have some contracts on the very large contracts. They just signed literally in the last week. It was pretty sick. like Monday morning, yesterday, Monday morning signed two contracts,

locked in there with some companies that everybody watching knows very, very well. Yeah, the traction so far is pretty, pretty amazing. Yeah, it was great. I think we definitely seen, like, over the last couple batches,

like kind of an increase in traction just because people want to be here in El Sagano, people kind of see the brand we provide. And that's, I think, pretty evident with the company. So, techs. Is Bloomberg for Heavy Equipment, procurement. So that's, oh, dealerships, interesting. So actually figuring out, so Bloomberg, you think that means looking across all the different industries, like how you can go on

Bloomberg and find someone's auctioning off a, you know, a private jet, but they're also selling some slug of equity or something like that. So this is a platform to connect buyers and sellers. It basically brings together people who want to buy. Very cool. Miding equipment and heavy machinery more generally. There's like no central database with all this, this information. It's just like Boomer X in Texas calls Boomer Y in Arkansas. And then they hopefully make it happen.

But there's no like label database of all this equipment. Basically just like compile at all. Kind of like with like Carfax, for example. Like early on there's like no data. And like basically they just went out and like called all the dealers. Like hey, like how much does this work? How much is this worth? Is Loris sort of like the V2 of the vertical farming era?

it says they turn grocery store rooftops into autonomous farms. Yeah. So like my take here was like vertical farming obviously been tried. Yeah. It didn't really work because it was all like climate like kind of carbon credit driven a lot of the time. Or like the actual uneconomics not make sense with that gone. Basically they removed like one of the biggest cost which is real estate and they basically just put these like blow up farms on grocery stores roofs. And then it's like a pretty easy.

I mean the numbers are here like they have like they don't pay for the actual unit. They have people to go and license them and bring them out to grocery stores, and then they get, like, individual profit based off what the grocery stores make. And then the goal is bring the produce profits per grocery store up about 50%. The produce is then sold directly in the grocery store that it's sitting on. Exactly. Cool. Interesting. Co-location.

And they're already, like, in store. That's awesome. A pretty big contract recently, which is pretty exciting. But, yeah, I think we're like generally like this, again, food security is important. Yeah. The stuff that was tried last, I don't know, five, five years ago, so it didn't really work. because of the economic like this makes sense it also with these general tell ones of like need to make food in america it's like actually healthy cool um so that's them uh lark kai

legend at skunk works right out of college um there's this new mosa standard basically which is like open source like if you have a component it has to actually work with every type of system um there's no way to actually test this they basically have a software that can test the compliance for um for any component and then eventually you're going to build the hardware to test it in a more in-depth fashion starting with the software. I like Blitz panel. That's a good name. Blitz panel reminds me a mixed panel.

It's a pretty simple company. They make electrical panels for cheaper, and they have a whole automation process. He's built out. He was at Pipe Dream Labs in Texas. It's a huge problem, and now he's basically just building the... Yeah, going through all this stuff, it really shows you that there are like a lot of stuff that fit in thematically that does not feel like it

competes with Anderol, for example. Yeah. But it's all, it feels like aligned with the general theme that you've put together of cool, hard problems. There's a little bit of American dynamism, a little bit of military stuff in here. Tell us about the last company, number nine.

Yeah, yeah. So Architas Energy. Arcitas. This guy reads old German papers from the 40s in German. And he found this process that basically, What kind of German papers is he reading?

You can guess. You can guess. That doesn't instill confidence. Tell this guy to clean it up. What is going on? No, the papers are scientific papers. Okay, okay, okay. That needs to be clarified.

Good, good. That needs to be clarified. Papers about scientific processes, no people. Okay, yes, yes. They did have some good scientists. In the 40s. Exactly. And one of these papers in particular goes to this way of making hydrogen for cheaper.

Okay. And if you put hydrogen cheap, you can make crude oil for cheaper. Very cool. Well, congratulations on the demo day. Thank you so much for coming out. Yeah, really cool lineup. It's great. I mean, it's cool to see a lot of people focus on energy.

I mean, that's that like, yeah. People will criticize, like, you know, accelerators and like for having, like, being, like, too on trend. Totally, totally. but it's actually healthy. Like, we want a big group of companies, like, going after a bunch of different problems. This just doesn't seem like that narrow of a trend. It feels extremely narrow when you introduced it the first time.

And it just felt like, oh, it's just a fence tech or something. But clearly, like, it's such a big area that some of these folks, if they're both successful, like, they're not even going to do business together because they're operating completely different layers of the stack. It's just a fantastic end result and really cool to see the growth here. So congratulations. I think one last point about the general defense market that I think is interesting is like a lot of these big companies like annual like Serenac like the big like here once have already chosen like a winner and like they're not going to back another annual competitor. So it's like you need to have like the stack, like the capital stack makes sense. It's not to be every single level of it.

And like all these funds have already chosen somebody as like their winner. It's kind of hard to come in and like be back with these funds. So I think there's like so many opportunities that now to come because again legacy industries are like open for venture. There's a lot of talent flowing in. that's what we're trying to take advantage over this new movement. The best talent in capital actually interested in solving these hard to critical national problems. Thank you so much for coming on the show and breaking it down for us.

Congratulations on another successful demo day. We will see you next demo day and hopefully before that, of course. Yeah, great to see you. Have a great weekend. We'll talk to you soon. Cheers. Goodbye. Let me tell you about 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. Our next guest is a living legend. This is amazing. The founder of Part of Full. I'm sure she's dressed totally normal.

Yep. Welcome to the show. How many Halloween parties did you get invited to? Okay, is this AI? Yes. Is this AI? Hi.

I wish because it would have been a lot easier to put on. Did you do this yourself? Or did you have a team? I did do it myself, which is why I missed a few spots. Oh, it looks great. This is incredible. I mean, I feel like Halloween is probably high pressure, like, on the Partiful team. You guys are facilitating a lot of Halloween events.

Like, you kind of have to really show up and do it right. So congratulations. This is your Black Friday. This is our Black Friday. This is our Super Bowl. This is our D-Day. Yeah. Well, thank you for taking some time to talk to us on your Black Friday.

Is it actually Black Friday in the sense that, like, the servers will be melting? or is it just like it's the best opportunity to capitalize on the value prop that Partifle can deliver? Yeah, I would actually be curious what are the biggest holidays for or just days celebrations? Halloween is by far the biggest. Really? No way. People celebrate Halloween like 10 times, right? Like everyone can go to multiple Halloween parties and it's very valid.

You're only going to go to one New Year's Eve party. Maybe during the holidays you're like traveling so who knows if you'll be around. but like everyone's here, everyone's locked in, everyone knows what the job is. And so usually it's like you'll have six Halloween parties. And I don't because I like the entire Partifle team. I'm like kind of on call today this weekend.

But that's why I'm dressed up right now. Yeah, I love the idea of you being on call and then there's some serious thing happening in the business. And you're just dressed up as well. And then you're just dressed up as well. And then you dress up like this. It's very serious. Yeah.

It seems like you're not melting. Yeah. Yeah. Yeah. The servers are not melting because our engineers crept heavily for this, but it is like a triple on-call situation. How has the business been growing? I know that a couple months ago there was like, oh, Apple's coming after you. And that felt like maybe it was just always going to be.

Focus on yourself. Apple, focus on yourself. But was that even like an OSH-I-T moment or was it just kind of like. like, okay, we didn't really see any blip in the numbers whatsoever. And we've just been on a growth projector, growth trajectory the entire time. When it happened, it was definitely, are you not allowed to curse on this show? We don't curse on this show, but you're welcome to. But kids do watch.

Our kids watch. So if you, whatever you say, our kids will repeat to us later. It'd be like, what did Trey I mean by that? Yeah. It was an oh darn moment. There we go. Oh, shucks. Oh, shucks.

But then we're like, let's wait to see how this plays out because I think so many startups go through the thing where some big player comes in and they try to reheat your nachos and it either works or it doesn't. And I don't know. Have you guys gotten an Apple invite? No. No. There you go.

There you go. Yeah, I mean, it seems like an obvious. I mean, the other thing is Apple is not good at. fun. Like they did the gen moji thing and and they were laughed at for it. It was like kind of a cool product. But again, it doesn't even get real usage. And I just think Apple was always known for magical software because it was just like so polished and good. And I think they're becoming less known for that. Yeah, I have another take here. I'd like to know if this resonates with you, you're obviously on the inside of this, but I feel like the invite product has like,

there are technical features that someone can clone and check the box, but the brand is so important. And I feel like when I get a part of all, it says something about what I'm going to expect walking in there. When I get a paperless post, that's a different vibe. That's a different thing. When I get a physical piece of mail that invites me to something, That's also a different brand, a different vibe.

It sets me up. It's the welcome moment to the actual event. It's part of the event experience. And so I feel like there's a world where you could clone all the features, but it's not going to bring the same energy, which is actually what your customers are buying in some way. Part of the product. I think it's part of the product. But how do you think about brand and how you stand out from the competition?

I think it's brand. And there's something about you have to want to break the rules. because a party is not like a corporate meeting, right? A party is a party. And so we from the very beginning have always been like the party page should feel as much fun as the party itself. Like you're already asking guests to do a lot by RSVPing. Like you might as well make it fun for them and give them a reward where they get to see the guest list after they are. You got to make it something where people can actually feel bought into the experience.

And you don't do that by, you know, building something that just like, looks like an online form or gets buried in someone's email. But I also think there's an interesting like cross-platform aspect of this too. When you think about Apple, like if I want to invite people to my party, I want to invite my friends who have an iPhone. I want to invite my friends who have Android. Like I don't care what phone my friends have. They're still my friends. And so I think it's really important. And it's just it's hard for certain companies to execute on this. Does the Apple app only work if you have an iPhone? It does work for a Android.

Okay. It's just, that's just like, that's just mean, okay? Not even being able to invite Android people to parties. That's just mean. I mean, well, they're already green friends. They're already like, look like this. Yeah. What, uh, so, so we, oftentimes, you know, if we're talking to an enterprise software

founder, we don't typically ask the question, like, what's the 10-year vision? Like, where does this go? Because usually it's like more seats, more contracts, more problems, more problems, more product, margin expansion, based on products. But how do you, like, you know, and I'm sure you've just been learning about what Partifle is as you, as you've built it, but like,

has the vision, like, what was kind of the founding vision? What is it today? How is it changed? And kind of where are you going? Because I've always, I've like, I've seen like your posts or like different press hits throughout the years. And there was some really funny post

about like not having any, not making any money. that stuck with me. But, and obviously just, like, you know, joking about it or whatever. But, like, yeah, where did you start from and where are you going? Yeah, it's funny how many people took that post, like, very seriously. It's like, you can't joke on the Internet anymore. What was the exact line?

I forget. It was just, like, Partful won't make money. VCs gave us funding to help you party, enjoy it, babes, something like that. They're all these, like, very serious news. articles of like Partiful said that Vsons gave me a party like is it true that's kind of a if you guys aren't making money if you guys aren't making money it's kind of a systemic risk to the part to big party you know yeah no we we will definitely we will definitely make money uh just not right now the the long term vision is if you look at where most social apps have gone they've kind of

turned into entertainment platforms right like when when I open Instagram I'm being fed a bunch of reels I'm seeing less and less of my friends' content. TikTok is obviously huge and it's also an entertainment platform. And so when you really think about it, the content that we're consuming is more and more just like it can be Netflix, it can be YouTube, it can be reels, it can be shorts, it can be TikTok, but it's all just you're watching entertainment from people who you don't actually know. And so what's gotten lost in that is like, where do you go if you want to actually connect with your friends? Where do you go if you want to meet more of the people who could be your friends because less and less you're discovering them online. And so what we want to do is create

the home for you to hang out with your friends, see them more easily, discover new things to do and new people to do them with. And we think that as big social actually stops becoming social and starts becoming big new media, that big social deserves to exist. And right now it's small social, it's part of all. But the goal is become big social. That's me, I love it. Yeah. I mean, we were talking to Brian Chesky is somewhat similar thesis. And it's grown on me so much, the barbell idea. You're going to be, there's going to be the internet slop feed. There's going to be a lot of content and stuff.

But then you're also going to want to touch grass and both can be true and you can kind of toggle back and forth. What's the most underrated part of full that you've ever seen? Okay. So I, they're somewhat rated. It's not fully underrated. But I'm obsessed with the performative. contest so there's like performative male contest happening right now I saw in San Francisco there was like a performative reading contest okay yeah it's just like there's something really funny

that's happening where online meme culture people are using Partifle to take that offline sure so they're taking I know by the way I know I used to go to the same gym as a kid who like initially went viral for the performative reading in New York where he was reading like, uh, like, like some infinite jest or something. It wasn't that. It was like, it was like a feminist feminist feminist text or something like that. And he's like, you know, reading. And people are like, hey, he's been reading the same page for like, yeah, 30 minutes. Like what's going on? And that, I'm sure that was part of what kicked off this, this kind of thing. Yeah. Yeah. Or like group seven on TikTok. Like there's not like a group seven meet up. I think this weekend in New York.

So it's like the speed at which every single internet meme now has a real offline part of a event that people are actually going to that's actually happening. That's been pretty cool to watch. Take us through the SF party culture. I'm hearing 996. It's getting crazier. Everyone's working nonstop. Is that a bear signal for your business?

No one's going to be partying, right? Are they just using it for like happy hour? come talk about SaaS. Oh, that could be a bowlcase. Break it down. 9-96 culture is very concerning to the party community. I'm here to declare a state of emergency in the SF party scene. We have to end it.

The drinks, happy hours are coming back. It's okay to crack open a happy dad at 2 p.m. Pass me that happy dad. I'm going to crack that open right now. Let's get the party started right now. Yeah, do tech companies need to bring back boozy? brunches during weekdays? Was that ever a thing? Like what's what's kind of the solution here? Maybe that's what we should start using our VC dollars for. The power lunch?

Power lunch? Yeah, yeah. Funding everyone's boozy loonches and boozy dinners. I mean, that kind of, that was kind of the joke in 2021. It was like there was like, you know, people used to joke 30% of all dollar, venture dollars go to meta. And then in 2021, it felt like 30% of VC dollars were going to like founder dinner. Like it was just a founder. You could basically eat for free as a founder by just going to a different founder dinner every night with a with a fintech. Yeah. Okay.

I want to keep going. I know we're way over time. But I want you to synthesize something you've been thinking about, this idea of like rage bait marketing being not effective long term. Synthesize that with some. I feel like part of full has done a little bit of it, but you haven't like crossed any line. So it's always like worked out. Even that example you gave of like, oh, we're spending all our VC's money. That's kind of rage-bate marketing.

But it never, it never bubbled up to the point where people were like, oh, you don't have a sense of humor about it. That creates goodwill. Yeah. So I've always loved the part of a brand. The VCs are financing this party. And for the user, that's cool. Yeah. But that's quite different than like the memes that were going around earlier this year,

people being like, I used VC dollars to buy. this lamb. Oh yeah, and myself. Yeah. How do you think about like rage bait where the line is on your brand? Obviously, you came on a show that's, I don't know, mostly a joke, but somewhat serious, dressed up. Like you have fun, but then there is the line. How do you think about that? We, there's a little bit that I think you kind of have to acknowledge it. Like, if a company's pre-revenue, like, you know, there's clearly VC funding that's backing it. So you can't not talk about that, Right. And I'm a beneficiary of VCs funding a lot of free Ubers in my peak partying era. So I have a responsibility to pay it forward to the next generation.

Yeah. But I think we're here to make people happy. That's the entire purpose of the company. We're not here to make people angry. And so there's only so much rage baiting that like we, we a part of will can do that's authentic because it's like, no, the rest of the internet makes people angry. And we're like the one thing that's trying to not do that. So I think this part of why we tow the line. It's just not as authentic to us if, like, you know, the Twitter mobs are up in arms. There's also, John, you were writing about the come for, come for the tool, stay for the network. Oh, sure. This is like a perfect example, but a stronger, like, I think a stronger example than come for the tool, stay for the network. John Coogan, Chris Dixon.

Chris Dixon. It was me quite. Last question. We have to have you back on and go way deeper, but I'm super interested. It seems like Partifle has, complete and total domination over like tech culture in terms of like if there's a party that's happening in s f tech like you found your beachhead you've landed you've dominated uh how much time are you spending hunting around to find like these crazy adjacent markets for okay there's someone who collects some odd object in the midwest and like you just dominate in milwaukee now because like you're how dominant is part of full because i don't yeah it's grown it's grown a lot. So we've never had boots on the ground in a market like DC or Boston or Chicago. And those

are all huge part of them. Yeah. Yeah. Yeah. We're growing like crazy in London. Yeah. We do a on the ground there. Yeah. But it's all just web flywheel stuff, right? Yeah. Yeah. Just, you know, network effects. Yeah. Just good old fashion viral growth. Hit the gong for network effects. Let's hit the call. Great hit. All right. This was super. fun. Thank you for taking 15 minutes out of your out of your Super Bowl to come on. Hopefully no crises later today because the people

are depending on it. And I'm excited for you to become big social media. Yes, we're very excited for you. Thank you so much. We'll talk to you soon. Happy Halloween. Bye. Talk soon. Our next guest has been in the reaching room for far too long.

Shishir Metcora from Superhuman, formerly. Grammarly, we're going to break down the rebrand. We had him on when they acquired Superhuman. Yes. Welcome to the stream. How are you doing? Congratulations on the next iteration. We pressed you about this. We asked you, is there

going to be a rebrand? And you were like, nothing to it now? I think you did. Yeah. But congratulations is done, and it happened really fast, actually. Many corporations would take a year. I don't know. It feels fast to me. But anyway, how's it going? What are the key changes? What's the thesis? Why'd you go with Superhuman and not Grammarly as the key hero brand?

Yeah, that's great. Yeah, we did talk about it a few months back when we bought Superhuman. There's two big changes we made. We changed the name of the corporate entity to be Superhuman instead of Gramerly. It's a little bit like the Google Alphabet transition. There's still, Gramley is still an important brand in our portfolio products. We now have four products under the Superhuman brand. It's Grammally.

It's Cota. It's mail, which you used to know as Superhuman before. And the other big announcement is we added a fourth new product we call Go, Superhuman Go, which brings a new AI assistant product, takes the best parts of Gramerly and makes it a platform. Okay, explain that more to me because... Okay, one before we jump into that, I will say, superhuman is a perfect name for a suite of AI-enabled apps at a time when a lot of humans,

especially even, I would say, more so outside of the tech bubble, are like worried about being obsolete in an age of AI. And so this kind of ethos of like, let's give people superpowers very strong. It makes a lot of sense. Exactly. You nailed it. That's why we picked the name.

I had two tests. One was it need to be broad enough for the suite. First thing you said. And the second one was actually the word human is more important than the word super. And it's actually part of the Grammally DNA because Grammarly has always been about empowerment of humans.

You still, you know, we help you assist you in your writing task, but you still, at the end of the day, submit the article, you submit the essay, you write the email. We're just helping you be more human. So I love the concept of superhuman that way. Okay, so walk me through the actual instantiation of the product.

I understand Grammarly is, you know, a widget. But yeah, let's maybe talk about Superhuman Go, the new product, because that's what I have no context on. Yes, please. Right, right. Yeah, and actually we tease that a little bit

last time I was here too. And the core idea is a simple reframe of Grammarly. So Superhuman Go takes the core underlying technology layer of Gramerly and opens it up as a platform. So most people think about Grammarly as a grammar assistant, and that is a big piece of what it does. But underneath it is a layer that allows you to bring AI. to a proactive embedded experience in every tool you use.

Gramerly works in about a million different web apps, desktop apps, and mobile applications. And now you can run any agent, not just your writing assistant, on top of that platform, embedded, connected, and personal. How do you think about the wars at the various layers of the stack? There's a war. They're fighting all of them. I mean, there's an AI assistant in basically every web app now. Like if I go into Gmail, it's asking me to use Gemini. And then at the browser layer, there's Atlas versus, you know, the perplexes got going on.

And then there's also the OS layer. So I was reflecting the other day on Gmail was trying to get me to use Gemini. Chad Jeeb was asking me use Atlas. And then Apple was trying to get me to use Apple intelligence. And it's three helpful AI assistants all in just, you want to summarize this email. So you said the keyword that you said ask. Yeah. So actually, interestingly, my frame for, if you took all those AI providers and you put them into buckets, I call it the assist players, the chat players, and the do players.

Okay. So the chat players, that's what you're referring to. There's a chat bot everywhere. Chat ChpT is obviously synonymous with that category, and there's now dozens of them. Those all ask you to go chat with a virtual human assistant. There's a set of folks working on what we call do, which is I want an AI agent that can go do things on my behalf as headless tasks. And, you know, I think the most popular one right now, or the one people are talking a lot about is all the ones in the coding space. You know, the Anthropic team said 39% of their queries are for headless agentic tasks and being used for developers writing code.

But at the bottom of that stack is another category we call assist. And that's the category where we bring AI to you. So I'll give you one fun stat. We do about a hundred billion LLM calls a week. It would make us a top AI provider based on any, any, any, any, but we do that across about 40 million users, which if you do the quick math on that, that's about, that's a few thousand AI calls a day per person. So if you're a really, really good Gemini user or you're a really good chat chip Qa user, maybe use it a dozen times. Yep, we do it. Yeah, every time you press the key, every time you load a new dock, load a new app, we proactively bring AI to you and we're figuring out all the places we can invisibly insert AI so it can assist you right where you're working. Yeah. That makes sense. sense.

Will you, any plans to get into the browser game? Are you going to let people duke it out there? Yeah. That's not our plan. I mean, I think we work, our strength is that we work everywhere. By the way, I should say, I'm so excited that there is a browser game. I remember when I was at Microsoft when Chrome came out, and I worked at Google for a long time, and I watched that happened, and then it just felt like it went away for a long time.

It's pretty exciting, actually, that there's a competition in browsers. But we work wherever you work. We work in browsers. We work in the desktop. We work on your mobile phones. To be honest, some fragmentation in the browser world is probably good for us. I mean, it makes even more valuable to have an assistant that follows you everywhere, and that's the core of the technology we've built anyways.

But, you know, I don't see us in that game, but I'm very curious to watch it. Fantastic. It makes a lot of sense. Well, congratulations on all the progress. You dressing up for Halloween later? I am dressed up on the Superhuman CEO. Yeah, there we go. It's fantastic.

That's the most locked-in costume. It's incredibly locked-in costume. costume. He's super, Superman himself. Yeah. Thank you so much for coming on the show. Congratulations. Great to get that update. Love the new name. We'll talk to you soon. Cheers. If you have a new name, you need to get on adquick.com. Out of home advertising made easy and measurable. Introduce everyone to your unified brand. Say goodbye to the headaches of out of home advertising. Only add quick combined technology, out of home expertise and data to enable, efficient, seamless, ad buy across the globe. You also need to get a luxury watch.

to celebrate your new rebrand. Get your superhuman on your wrist with bezel. Your bezel concierge is available now to source you any watch on the planet. Our next guest is dressed up as the co-founder and CEO of Linear. I can't wait to talk. Come on in.

Carrey, how are you doing? Great to see you. Great to see you. Well, great, thanks for having me again. And yeah, like my plan was the same as everyone else. I could have to dress up like you guys. There you go. I saw it's already done.

so I thought like, well, I'll just do a different thing and not dress up. Well, you're locked in, locked in. So you have linear. You could plan to like if you grew out the hair on the sides. I bet by next year you could do a pretty good impression of one of us without even a hat. For sure. A pure play. Great to see you.

I think did we overlap almost it. Were you at the GitHub universe on Tuesday? No, I wasn't there. But yeah, that's why I couldn't join. but yeah, we can talk about that or something else. Yeah, yeah. I mean, we talked to, we got a good whirlwind tour of what Microsoft's doing. Obviously, Asatio was taking a victory lap on the Open AI deal.

And we talked to the Codex team over at OpenAI. We also talked to some GitHub folks. And I think the general shape of software development in 2025 is becoming clear. You have asynchronous agents that you can fire off. You're doing code review. I have a tab completion model and IDE that's getting smarter every month. But I'd really love to hear how you see linear evolving in that context. And how, and what, I mean, you can think about linear as delegating different tasks.

How do you delegate when you have super long running agents? We were talking to the Codex team. What do you say? Something ran for 60 hours. 60 hours. It's like, that's something you need linear to manage at this point. But take me through how you're seeing folks change their use of linear and what you're doing with the product.

Yeah. One of the things I've been thinking about is that a lot of linear was founded before all this like AI things started happening more. And as well as like I think like most companies out there today have like even like large companies, they started before this was happening. So I think we're all in this journey of kind of like an evolution of like how do we do these things. So like what are the workflows?

What needs to change? Like what doesn't change? I've been like thinking about like the whole. kind of writing examples. So like we used to write on paper, like we write documents on a paper, then we move to typewriters, then we move to like online,

like a text processing on a computers, and then we meant to like online writing. The writing is still, it's like you have a piece of paper on the screen, you're still doing the writing activity. And then now, like, okay, like the AI can also do some of the writing activities, but it's kind of like a lot of the tooling changed

around that activity or that concept or that need. But the activity itself didn't change. So, like, we still have, like, documents and papers as an idea because it's useful. So I see that, like, kind of, like, the chains on the software industry similar. That, like, the AI doesn't mean that we will just throw everything out of the window. Like, there's no longer, like, any of these ideas or workflows or something. But it's more like we have to kind of, like, go piece by piece. Like, does this still make sense?

Okay, it does make sense? Let's do that. And then, like, does this have to change? Okay, does has to change. Or it would be better. if it changes. So I think our thinking with that

has been that we want to help companies move to that direction as well. So we do have for example the agents platform where also GitHub copilot now integrated with so you can delegate tasks directly to

the GitHub co-pilot and it can work on the background. You can then come back to see what the results are and then review the code. We also have like other agents. Yeah. Is there any world where the ability to delegate to agents enables more non-technical people, it gives more non-technical people a reason to have a seat on linear?

Yeah. I think like it's something like we're definitely like seeing today. And I think it's one of the values we can give to, value we can give to organization. So like IDs and technical tools are used by engineers. Yeah. But like what we even see in our company, now is like designers are making fixes with the agents.

Yeah, yeah, yeah. When they see something like, oh, this is like a little bit off, I can like properly instruct the agents, like find this in the code base. Do you do like a version of this. And you can also like, in our case, you can preview the result through linear. So you can actually see like, oh, this is how it is and and tested. So I think it's like, it gives like a nice loop for a non-technical person to sometimes just like fix it. like fix things when they see something and you don't always have to like assign it to an engineer.

How do you think about like project management software broadly evolving? Like do you expect to, are you guys experimenting with any new paradigms? Like agents are great in that right now we even call them like human like names oftentimes right? Everyone has their, you know, like sometimes it's a girl name, sometimes it's a guy name. And like the way that we're using. them today is very much like a teammate. That might suggest that we're not going to see a bunch of new UI paradigms, but are you in the team kind of like what doesn't get shipped that you guys are kind of playing with, you know, behind the scenes? Yeah, I think like I like to think about there's like two areas. One is the organizational workflow. So like that goes everything from like what

linear really offers is like coordination and privatization and communication and communication of visibility to the organization. Like, this is the direction we're going, and this is how it's going, and these are the people working on those things. And so I think, like, with AI, we are looking for those, like, problems and, like, what can we actually fix there or, like, improve there with the AI? And then I think the other area is the individual engineering workflow, like, where does that change?

So I think, like, there is definitely, like, something we're thinking about that. Do the issue still make sense? to pull requests still make sense as a concept when you are more like working with the AI to develop something. I do think like there is still need for reporting things like bucks and like even and or like capturing issues having discussions around it.

So I don't think like everything is going to change that. I think like a lot of times what really help companies with this like some kind of direction and some kind of, kind of like a guidance to their teams like here's here's a like a system for you to operate in and these are the pieces and this is the goals and then you let them like do their work so i i at least like to see that it's it's not like i don't what we think about the SaaS future or like SaaS products it's it's more about the making them more proactive or self-driving like we like to say like self-driving

is that we're not yet at the stage of like, like with transportation, like we still need cars because we don't have teleports. So we need to now, like, we're at the stage of like, let's make those cars self-driving. And so I think like when it comes to software products, I think we are also there that I don't think we are at the stage yet and maybe not for a long time that we no longer need no software products.

But I think like they could be much more proactive and driving things on their own, on the background, without you interacting at all. So the idea would be that bugs just fix themselves. So like when bugs gets reported, someone like the AI will fix it. And you don't even have to see it. I think there's still an...

Eventually, I'm excited for the future when teams can just go into linear and just watch a bunch of work happening. You know, it's an agent, an agent finds a bug, another agent solves it. an agent has an idea for a feature or a user gives an idea that agent ships it it's you know that the fully autonomous future is wild i was uh i i wrote a post maybe it was a couple weeks ago at this point or or last week about the uninspired company of silicon valley and and how i mentioned at one point

that companies i'm sure now like so many companies have copied linear i'm sure there's companies now that copy linear without even knowing, like the origin of design language. They're just like, oh, this is just the way that design looks. So I was wondering, like, your philosophy on, like, you guys get a lot of credit for kind of like creating an entire way, you know, feeling for software and design language and all these different, you know, UI elements. Do you ever have a desire to like reinvent the next paradigm and, and, and, kind of ship something that would, you know, create, create an entirely new, you know,

version of the software that would completely re-skin it. Entirely skeuomorphic. It's all just, like, wooden blocks and grass simulation. Yeah, or do you just continue to own it and be the godfather of modern software as a service? Random fonts everywhere, random colors everywhere. I think, we definitely think about that, like, especially on the website side, that, like, well,

what is the next thing we can do since like everyone is kind of like doing this thing we're doing. So I think there's definitely that need. But I think in the end, like we try to be functional or pragmatic about the design, that it still should serve the purpose. I believe that old like design quote of the form follows function. So it's like you're not like with productivity tools or other tools, I don't think like the point is to be an artist or something like you're not trying to like invent something crazy things just for the sake of it.

I think it needs to be like to the purpose or the function of the tool. And then like around that like there's always like you can like try to figure out what needs to, what could be better or like maybe you can go against some convention when you when you have that reason. So I mean, we're always looking for those opportunities. And I think for us like part of the like the mission of the company is to accelerate an inspire builder. So we also do in a way like that we're accomplishing the mission of inspiring a lot of different companies on how they design or like focus on the craft or quality. So that's that's kind of like personally. What I also wanted to do that could we like shift this industry a little bit more into like we can do nice things.

And it doesn't always have to be. It's just like what metric goes up or down and like or some other reasons. Yeah. Yeah. Made the world a more beautiful place. I love it. That's a good time. Thank you so much for coming on the show.

Thank you so much for taking the time on Friday. Next time you're passing through LA, come by the ultra-dum in person. We love to hang. Ideally on a day when we're not dressed up like fools. Like a couple of characters. It can be a little bit more serious. But thank you so much for taking the time to chat with us.

And thank you for supporting the show this entire year. It's been very helpful. One of the very first few. We've had a lot of fun. And everyone in the chat is celebrating linear. Conor says he just onboarded his business onto Linear today, like 20 minutes ago. Let's hear it for that. John Axley says, Linear is TBPN royalty.

That's right. So awesome to see, Kari, in the Ultridum. Thank you for coming by. We will talk to you soon. Give our best to you today. We'll talk to you. We'll talk to you. Goodbye.

Cheers. There's one scoop I want to talk to about. Katie Roof is confirming a scoop from TechMeme that Axios is dropping. human interest, a 401k platform for small and medium businesses that aims to cut down on administrative work, raised $100 million at a $3 billion post money evaluation. This is a leak, but I know the CEO of human interest. He's a good buddy. And so I just want to say congrats to human interest because it's a very cool company.

And Elon has a new pod up on Allen. And on Rogan within hours. He's on a press store. We've got to get him. Maybe this is the time to strike. Maybe we get him on TBPN next week. the team seems to like the idea. We'll see.

We'll work on it. Thank you guys for hanging out with us on such a crazy day. Yeah. It's going to be hard to, we were, while we were putting this on, we were talking about how do we want up this next year. Yeah. What could possibly be as iconic as these two titans of industry. Didn't we do Elon and Joe Rogan?

That would be good. I think throwing in a Rogan is a good one. You as, you as Rogan, even though you're. 14 inches taller than Rogan? Yeah, yeah. I can see, I could see a Rogan. We could do a Rogan and Alex Friedman. That could work.

Maybe we could do a human man. Lex would be the experts. Yeah, the experts is good. As other podcasters, kind of hits. We could do a David Senra. We could both be David Sedra. How about that? David Senra by David Senra.

Yes, we could both be David Senra. An executive produced by David. That would be a lot of fun. I think your mustache is actually falling off your face. I feel this weird. thing where it feels like I'm sweating underneath everything. Yeah. But I, but it's, I, I've had this, like, scratch up here by this, like, fake vein.

And I'm not able to it. And so I'm looking forward to taking this off. It's actually so trippy. A lot of people are asking, are you going to, am I going to go home and trick or treat with the kids like this? I was thinking about it. No, I do not want to traumatize my young children. Yeah.

I will be dressing up as something else. But thank you for tuning. in with us today. Have a wonderful... And thank you to the team that helped us pull this off. This was a very special, funny thing. I never thought we'd be doing this, but that's the nature of this show. And so thank you for everyone that supports us. Leave us five stars on Apple Podcast. And go out and have a wonderful evening and weekend. Be safe. We will see you back here on Monday. We'll talk to you soon. Can't wait. Goodbye. Cheers.