TBPN Live - xAI’s Colossus 2 Reactions, Sam Ross & Aaron Slodov LIVE in The Ultradome | Pippa Lamb, Ariana Thacker, Will Hurd

Episode Date: September 19, 2025

(01:21) - xAI Unveils Colossus 2 Datacenter (30:20) - Pippa Lamb, a partner at Sweet Capital, discusses the recent U.S. state visit to the UK, highlighting significant tech investments from companies like Microsoft, Google, and Nvidia, and emphasizing the UK's ambition to become an AI superpower. She also provides insights into Prime Minister Keir Starmer's leadership, noting his pragmatic approach and pro-business stance, and touches on the UK's focus on growth, particularly in the tech and R&D sectors, as it seeks to strengthen its position post-Brexit. (47:30) - Johnny Carson's Former Home Lists for $110M (53:56) - 𝕏 Timeline Reactions (01:28:53) - Aaron Slodov, CEO and co-founder of Atomic Industries, is leading efforts to revitalize American manufacturing through advanced technologies. In the conversation, he discusses the company's recent Series A funding, plans to expand manufacturing facilities in Detroit, and the integration of AI and automation to enhance production efficiency. He also highlights the importance of reindustrializing the U.S. to strengthen the industrial base and reduce reliance on foreign manufacturing. (01:46:27) - 𝕏 Timeline Reactions (01:51:40) - Ariana Thacker, founder and CEO of MoldCo, transitioned from venture capital to entrepreneurship after her personal battle with mold toxicity. In the conversation, she discusses MoldCo's mission to revolutionize mold-related illness treatment by offering accessible, evidence-based care through virtual platforms, including specialized lab tests and treatments guided by mold-trained clinicians. She also highlights the company's recent $8 million funding round, emphasizing the growing awareness and need for effective solutions to mold exposure and its health impacts. (02:01:57) - Will Hurd, a former CIA officer and U.S. Representative for Texas's 23rd district, is now the Chief Strategy Officer at Chaos Industries, a defense technology company. In the conversation, he discusses his journey from studying computer science at Texas A&M University to serving in the CIA, where he was stationed in India, Pakistan, and Afghanistan, and later transitioning into politics due to his experiences briefing Congress. He highlights the importance of integrating technological expertise into government to address national security challenges, emphasizing the need for innovation in defense technologies to maintain U.S. superiority. (02:27:53) - Sam Ross, co-founder and CEO of Numeral, a company specializing in automating sales tax compliance for e-commerce and SaaS businesses, discusses Numeral's recent $35 million funding round at a $350 million valuation, following a Series A led by Benchmark earlier this year. He highlights the challenges of sales tax compliance faced during his time running a vitamin gummy e-commerce business, which led to the creation of Numeral to simplify this process. Ross also touches on the impact of AI on business models, noting that while AI is central to current fundraising efforts, it has led to increased operational costs, such as higher AWS bills, though he remains optimistic about future efficiencies. (02:48:04) - 𝕏 Timeline Reactions 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.aiFollow 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, September 19th, 2025. We are live from the TBPN Ultradome, the Temple of Technology, the Fortress of Finance, the capital of capital. Thank you to everyone in the chat. It is teeming today. Love to see it. The first message I saw was thank you to Ramp for another TBPN Friday.

Thank you to Ramp.Ramp.com. Time is money. Save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more. All in one. place we have we have new uh sound effects from the team by the way hit him hit him there we go i'm hearing multiple things this is getting crazy and we have a special guest today let's go let's go to the wide yes we have our turbo puffer here look at this it doesn't show how three you is i mean it

this this this it is absolutely stunning it's functional too of course search every bite turbo puffer we're puffin. Serverless vector and full-text search built from first principles on object storage, 10x cheaper, extremely scalable. Go to the gong cam. Thank you, Turbo Puffer, from supporting us. I was excited to dig deeper into that company. There's some fun stuff to share at some point.

Anyway, the big news, Colossus 2, X-AI, it seems like it's going well. People have been saying Elon spread too thin. Then Starships, you know, crashing into the firmament every two seconds. Starships landed, came back. The haters are, I mean, pulled back from politics, it seems like there's a lot of players. The huge L for the, you know, firmament believers after the last launch. Yes. Elon made it to space and back. And he's also making progress on Colossus too, the data center.

10, 20 minutes ago, CNBC is reporting that XAI raised a fresh $10 billion at $200 billion. Wow. That just hit the timeline. There we go. Dylan Patel, some did that. Did you put a headline in the newsletter? In the newsletter today? Yes, XAI, if you build it, they will come.

Oh, okay. Because his plan is to build the biggest data center, the first gigawatt data center in the world. And semi-analysis says they have a unique RL methodology. It's focused less on these RL environments for specific business use cases. we're not hearing about them cloning DoorDash or Amazon or creating a unique browser use like go solve a very specific they're not trying to build deep research RL environments they're building

any the romantic companion which is an RL environment that feeds off of human interaction and so semi analysis has a very interesting take harvest it potentially potentially but uh Dylan says every time you think Elon has lost his touch or is failing he does some crazy stuff and you're like damn he really still got it how Elon is building the largest data center in the world faster than anyone else is amazing also what they're doing on the RL side is equally nuts there's a bunch of there's a bunch of posts here and a bunch of a bunch of interesting do not go data center for data center with a go yes so I mean he got into this game late he found he co-founded

it's gonna be the 10 year anniversary of co-founding Open AI December 2015 is when he co-founded Open AI and we're coming out up on December of 2025, and he's about to have a data center that's on par. It's close to Open AI. It's not yet bigger, but they're getting there. This is semi-analysis breaks down the megawatts of IT capacity, which is effectively the total AI training data center capacity. Open AI is still ahead of XAI, but Elon's pulling out every possible trick play to catch up

to the firm. He co-founded. They're now ahead of Anthropic and Meta-superintelligence, according. to semi-analysis's data, you can go deeper and obviously subscribe. And if you upgrade to the most high tier, you get their data center model, and they'll even tell you what Google's doing, which is much more complex because Google trains across multiple data centers. So everyone loved Colossus 1, built from the ground up in 122 days.

It had 300 megawatts of IT power, which is now small potatoes because the next one's three times bigger. Zuck has talked about Hyperion and Prometheus shortly after going on his MSL high hiring spree, Open AI is, of course, working on Stargate in partnership with Crusoe and Oracle and others. Also in Alex Heath, Alex Heath, the other day was saying that, I think it released yesterday, that he believes we might be in an AI bubble. Zuck didn't say that. That was an interpretation of the Zuck clip. Interesting.

So I saw that post. We should pull up the actual clip. Let's watch it. And review what Zuck said, because I believe that. this was somebody adding a caption and interpreting the vibes that were coming off of that particular interaction between Alex Heath, who just launched a new substack and podcast. Let's listen to this interview with Mark Zuckerberg. Yeah, you were talking about the CAPEX and the data centers. You obviously see something on the other side of that that will warrant that being worth it.

But I'm wondering, do you ascribe to these bubble fears at all that people are talking about that we're in this massive overspending, getting ahead of the skis bubble? and maybe a company like meta will be okay because you guys do have a core business that makes a lot of money, but how do you think about this bubble talk that has been going on for the last few months, especially?

I mean, I think it's quite possible. I mean, I think basic, if you look at most other major infrastructure buildups and history, whether it's railroads or fiber for the internet in the dot-com bubble, these things were all chasing something that ended up being fundamentally very valuable.

In most cases, it ended up being even more valuable than the people who were kind of pushing the bubble thought it was going to be. But in at least all these past cases, the infrastructure gets built out, people take on too much debt, and then you hit some blip, whether it's some macroeconomic thing, or maybe you just have like a couple of years where the demand for the product doesn't quite materialize, and then a lot of the companies end up going out of business.

And then the assets get distressed, and then it's a great opportunity to go buy more. So I think that that's obviously impossible to predict what will happen here. There are compelling arguments for why AI could be an outlier and basically just if the models keep on growing in capability year over year and demand keeps growing, then maybe there is no collapse or something. But I do think that there's definitely a possibility, at least empirically based on past large infrastructure buildouts and how they led to bubbles, that something like that would happen here. He's going back and forth there. Not as strong as what Sam Allman said. Yeah. It's possible we're in a bubble is not the same as, yeah, we're definitely in a bubble.

Yeah. But I think you're right. The post, whoever flipped that is right as well. And that it's much better to be in a position where you have a highly profitable advertising business that can support this. Yep. You know, all of the AI CAPEX could effectively never go to what we think of as like, new Gen A.I. products and it could just go to the core business. Yeah. It'd be great.

Yeah. I mean, when we hear about Meta spending $60 billion on CapEx, like a lot of that is still core AI. A lot of that is still just recommend better ads, recommend better reels, that type of stuff. And so a lot less, just tons more downside protection in a company like Meta versus XAI, which is using leverage, using these deals shifting pawns across the, having SpaceX invest, SpaceX. And then SpaceX employees and Tesla employees are coming over, and Tesla megapack, battery packs are going over there. Like, it is a crazy, crazy kiretsu that... Well, speaking of well-placed ads, Tyler Rongi just shared this moment earlier. We've got to pull this up.

Pull this video up in the timeline. Tyler clipped it for us. John delivered potentially the greatest host red ad in history. we're a little bit a little bit biased obviously ramp.com time is money saved both these news corporate cards bill payments accounting in a whole lot more all in one place. There we go. Let's bring him on Mark Zuckerberg live on TVPN. Welcome to the stream. How you

doing, Mark? Good to see you. This is really just the pinnacle, like truly like the pinnacle of the bit that we started maybe a year ago less than a year ago. I mean, I guess we started doing rampads. I think we started doing rampads even before we had a deal with them. We just thought it was fun.

But this whole idea of like, let's be overly commercial, let's really push the ad reads. It was started as a joke and then it just became like our thing and we normalized it to a point where we can actually go and do an ad read, a host read, right before interviewing a Mag 7 CEO. So it was a fantastic, fantastic day, fantastic week. We had a lot of fun. Anyway, continuing with what's going on with XAI, there's this thought-provoking part of the latest semi-analysis that article that kind of breaks down how Ani and the romantic companion chatbot actually flows.

We can read through a little bit of this. I mean, Elon is extremely satisfied with how the build-out's going. semi-analysis at XAI now has 460 megawatts of natural gas turbines installed and either operating or under construction. This includes 12 SMT 130 turbines at Colossus 1 and 7 Titan 350 turbines at Mississippi. Those are extremely cool names for turbines, Titan 350. And Elon says still small fry at one gigawatt for serious power in the 100 gigawatt continuous range, one-fifth of USA average power consumption. It probably needs to be solar battery powered at one terawatt solar battery is the only realistic option.

Elon companies are so good at infrastructure. They should have bought the Fire Festival IP because they actually, I believe they could pull it off for sure. They could have pulled it off on a barren island. They would have been able to do it. Might not have been super, well, you know, easy on the eyes, but it would have, the festival would have run. Yeah, for sure. Restream. One live stream 30 plus destination.

We're re-streaming right now. Multi-stream and reach your audience wherever day. Semi-analysis continues. The location of XAIs AI's AI Data Center hub in Memphis is highly strategic as it sits at the intersection of three states. Tennessee residents are already frustrated with the environmental and noise concerns caused by the gas turbines, powering the facility to address this. Elon Musk simply placed the new turbines a few miles across the border in Mississippi, running a short transmission line back to the Tennessee hub. It is so crazy that he can even build a line.

a couple miles like that like how long does it take California to build a couple miles of railroad like a decade the idea that you can actually like build a even a transmission line that quickly is remarkable i wonder i wonder if any crazy land deals got done like in that two mile stretch right like because if he had to like do you think he owned all the land across state lines or was there's some section where but it gets yeah it gets more complex so uh so uh so uh and if mississippi residents or regulators push back. Well, no problem. He can just go a few miles west, build additional turbine plants in Arkansas, and connect them with another short transmission line. Elon Musk has also confirmed that XAI is purchasing an overseas power plant to bring to the U.S. to continue powering the his physical AI industry. But he's buying a plant that's overseas and just moving it? I think so. I think just like pick it up and put it down. Seminole analysis called it another 200 IQ first print. principles moved by Elon due to push back on the Tennessee from Tennessee over environmental

concerns. He built it at this intersection of three states and can kind of move, move between them. There's some really remarkable things, this genius trick. Such a crazy move. And you have to think that he didn't run into this, like later, he thought about this. He thought that the state regulation, it would be complex. And so pick that zone maybe a year ago and has been, you know, benefited. fissuary of it ever since.

Such a fascinating story. And you can see some images from 200 megawatts to 1.1 gigawatts. And you can see the physical change of the, look at this, the physical change of the site from March 2025 to August 2025. That's just a few months and it looks like a completely different, like terraforming the earth. It's absolutely wild. GROC right now, guess GROC's revenue on Censor Tower.

so the app store last month? $100 million, something like that. Last month? I would hope that they're at a billion dollar run rate. Maybe last? Five million. So $60 million run rate in the app store. Granted, they could have other people go to grok.com and sign up

and those signups wouldn't get counted towards this. Yeah. But this is counting Japan and the United States. And there's a lot of people that use GROC within X, so they have an X premium account, and that gets them. That gets some access to GROC. But, yeah, I mean, semi-analysis shares a lot of data around the AI Lab, ARR, external revenue. Open AI is at a $14.5 billion run rate.

Anthropics at a $7 billion run rate now, and GROC is not really showing up on this particular chart. There's some discussion in here about Elon going to the Middle East. There's the KSA's Kingdom Holding Company. they own 16.87% by the public, by PIF, public investment fund. They owned and kept a $1.9 billion stake in Twitter when Musk took the company private in 2022 and also owned an $800 million stake in XAI prior to the merger. So they're building a position in XXAI. The UAE's privately owned Vi Capital brought in 700 million in 2022 to support Elon's takeover

of Twitter. It also invested in the XAI Series C alongside UAE State Fund MGX, and Qatar's QIA also owned and kept a $375 million stake in Twitter and participated in XAI Series C. And so according to the Financial Times, XAI is preparing a new round in the tens of billions valuing the company close to $200 billion with Saudi PIF. The CNBC broke that that has allegedly been closed. Yeah. So semi-analysis earlier this week was talking about a. raise as large as 30 billion. It sounds like 10 billion came in today. And remember XAI's lawyer was on the record, I think a couple days ago, saying the idea that XAI is struggling

with fundraising is entirely false. Yes. It was like they're good at fundraising, right? They were performing well in fundraising. Fundraising is what they said. But the reporter was pushing them and basically saying, yeah, I mean, you can give, you can give these labs like a little bit of leeway because Anthropic wasn't making a ton of money, and then when they hit product market fit on the actual model, it ramp very, very quickly. Same thing on Open AI, obviously. The 200 billion, semi-analysis says this is challenging because it is hard for most investors to justify XAI as having a valuation higher than Anthropic. That is tricky, considering Anthropics at like a $7 billion run rate and growing very, very quickly. Anyway, let me tell you about one of our newest sponsors,

There's Privy, wallet infrastructure for every bank. Privy makes it easy to build on crypto rails, securely spin-up white-label wallets, sign transactions, and integrate on-chain infrastructure all through one simple API. The official wallet infrastructure provider of TBPN. Proud to work with them. Beyond external capital sources, Elon could generate capital internally since merging X with XAI to form X holdings.

Semi-analysis believes there's an ever-growing piece of XAI's revenue. revenues are intercompany transfers. So when you call GROC, at GROC is this true on X to answer questions, X.com is licensing that LLM technology for functionality such as search ad and even the recommendation algorithm. Elon has said just today

that he wants it to be entirely powered by XAI, by artificial intelligence, no, they wants to rip out the older system. Tyler. And remember in July, lie the Department of Defense awarded a $200 million contract to X-AI. Oh, yeah, that's right. So that's significant.

Yeah. Significant, I mean, you add that. You add the $60 million in Censor Tower. There's probably a bunch more coming just directly. There's probably some coming over from. But, yeah, at the end of the day, if you're, but we're not seeing, like, the multi-billion dollar revenues coming. If you're a large-scale allocator and you are having the opportunity to simultaneously invest in Anthropic

Open AI and XAI. Open AI at somewhere 350 to 500 billion. Anthropic, 160 or 170, whatever it is. And then XAI at 200 billion. It's like you really got to be riding with Elon.

I'm ready to ride, baby. Semi analysis says it's moving, describes it as moving money from Elon's right pocket to his left pocket. This has happened a bunch though. It happened with Solar City. He acquired that company. And then we didn't

really hear that much about solar from Tesla. There was this vision for a while that you would buy a solar roof and get solar panels installed in your house through Tesla. That project, it was slow and then it didn't really reach breakout speed. But interesting to see how much, like, at this point, it's been, what, a decade since you want to hold on to your puffer fish? Are you very happy about that thing? It is so cool. At this point, if Elon was really like had changed his tune on solar he very much could have pivoted and said like now we're doing nuclear and we're not doing solar anymore and you know we're kind of slowly like the actual solar city acquisition was never that big and it's been sort of like written down over the time no

one's saying oh Tesla needs to deliver solar city the original vision there for that to happen but yet Elon's still beating the drum of solar will be really important long term so I wonder a lot of people were not happy about the solar city acquisition. Yeah, but over time it didn't matter because the stock didn't matter in the end. And that's why I've said, I think if XAI eventually rolls into Tesla, I think that one, Tesla will, like, probably pop on that news. And in the fullness of time, like, it probably will not be, I think it'll be somewhat of a rounding air. Yeah.

It would be interesting if it comes back. and in another decade, Tesla or Elon Inc. is a major player in solar power generation. They certainly aren't a material amount of total power generation. They're getting into a material amount of total AI IT capacity and data center capacity,

kind of bootstrapping a hyperscaler, becoming a hyperscaler just through the XAI efforts. But it would be very interesting to see if we look back and we're like, oh, they actually did wind up installing a ton of solar. They needed an excuse to do it, not just ground-up adoption from everyday people putting it on the roof, but could happen. So, semi-analysis asked the question, does X-AI have a shot at becoming a frontier lab? X-AI has been able to catch up in many aspects to the current frontier.

GROC-4 by some measures and evaluations was up there with some of the best models on release, yet semi-analysis believes they're struggling to translate eval strength into revenue. Let's discuss the talent pool. There's a bunch of interesting things here. So X-AI got folks from Google DeepMind, Greg Yang and Tony Wu. X-AI also has renowned researchers like Jimmy Baugh, co-inventor of the atom optimizer. Since then, X-AI has poached more talent from DeepMind meta. And in some cases, NVIDIA, which is interesting because they're buying a ton of ships from NVIDIA,

but they did get some folks. They got Ethan He, who was a senior engineer at NVIDIA working on their, Nemo model series. My question is, is Ethan He, him? Or is he just Ethan He? At this point, they have north of a thousand employees and are continuously expanding with a new office stated to open in Seattle. They work really hard. It's famously hardcore. I like this phrase here. Part of this is the fast, extremely hardcore pace. XAI engineers 996 look easy. I think you're saying 2xAI engineers

996 looks easy often pulling what they call 996007 which is 12 to 12 7 days a week which I think is just 24 hours a day If you were doing 12 noon to 12 Well that's not that much more than 9 to 9

12 to 12 Yeah I think what they're saying here is that basically it's like you're 100% working Even when you're sleeping Like that's part of your job The sleeping in the tent

is like you're doing that to enable your work. So you live at the office and you truly are working endlessly. But the talent Exodus is a real issue with XAI. People are churning out, but they keep finding young people to join. And two of their best data center folks just left to join OpenAI,

but Colossus 2 keeps chugging forward. So they're able to backfill and they haven't been able to be fully destroyed. There's an interesting... Yeah, I mean, there's just, There's a world where even if XAI ends up just leading on infrastructure and actually being able to deliver compute, there's a ton of value in that as well. Right? We were talking offline the other day. Is there some world where Oracle says, like, hey, we have this crazy backlog. Basically, like, if you give us access to Colossus, we'll pay you a premium for it. But who knows?

Yeah, yeah. I mean, there is a question of, like, it's super. risky, there's debt, they're building out this massive cluster, and there isn't that much demand from the actual API for GROC, and there's not this crazy compounding advantage in consumer. But at the end of the day, if you wind up with a one gigawatt cluster, there's a lot of hyperscalers that want a one gigawatt cluster, and so you could probably just go sell that to AWS or Google pretty quickly. So it doesn't seem that crazy in terms of like where the dollars are going.

It seems like an efficient buildout that creates capacity. but then we're talking about like the dark fiber world where we couldn't find a use for this today so we had to sell it kind of at cost or below cost but it's not a zero it's not it's not it's not completely spent on on you know something that might not materialize at all anyway let me tell you about cognition our latest sponsor they're the makers of devon devon is the AI software engineer crush your backlog with your personal AI engineering team speaking of coding coding coding is today the largest API use case for AI. Devin, of course. I can't get over the team. He uses AI to write code. Cowbell. The cowbell is great.

Hit that horn again, Nick. I love the horn. That's great. Grock 4 is priced at the same level of Claude Sonnet 4, yet does not show the same level of coding abilities for non-coding applications. GPT-5 is even cheaper and has better capabilities across the board. So both Anthropic Sonnet 3.7 and Sonnet 4 are both at $4 per million tokens. GROC 4 is also at $4 per million tokens, but very quickly, GAPT5 is at $2 per million tokens.

GROC code fast is much cheaper, and a lot of the explosion over GROC code fast has been on OpenRouter. Semi-analysis dug into the open router stuff, and it's said that there is, there is in a small, cheap, and performant model. But there's a catch. Semi-analysis thinks that the spike on OpenRouter comes from XAI providing free voucher credits to Kilo-Code users who call XAI models via OpenRouter. Real demand will be clearer in the long run as voucher redemptions run out.

So they've been doing a lot to just spur adoption. Hopefully those folks stick around. There was one very interesting stat in here that they linked to. linked to. So a survey from Netscope found that GROC is used in 23% of organizations. This is up from like 0.01% of adoption of GROC 2 in enterprises that were surveyed. So massive, massive adoption of GROC in companies broadly. 23% is pretty high actually. But 29% of organizations have just blocked GROC entirely, likely due to the MECA incident and some of that stuff. So So it's unclear how Elon's influence over Grok will, you know, persist in post-training quirks.

If you're just, if you just want a model to optimize a manual workflow, you might not want it to take a strong stance on some political issue. You might just want it to put the tokens in the bag. So on the consumer side, Chachip-T is truly, truly crushing. the 30-day app revenue indexed to ChatGPT at 100. GROC is at 2.6 relative to Chat-GPT is 100, so 2.6% of Chat-GPT revenue, and then Claude is at 1.3% and Gemini is at 0.5%,

according to the semi-analysis and censor tower. It is pretty wild that the Mecca incident happened in the very beginning of July, and they were still able to announce the DoD contract for $200 million in mid-July. Yeah, I mean, those probably happened in a long time. Yeah, it takes time. So everyone discusses AI's weakness from model and lack of business perspective,

but it's important to recognize that they do have a unique method for RL. Reinforcement learning, semi-analysis is continuing. So Anthropics focused on code, automated software engineering, digital productivity is a path to automate model research, infrastructure improvements. This is clearly the best path towards revenue in the short to medium term, but XAI believes this is not a valid approach to AGI. A more generalized approach is required, says semi-analysis. While X-AI is chasing code performance, they are also chasing other areas, including emotional intelligence, and empathy, Ani is key to this.

This is interesting. X-AI's A&E is set to be one of the most diverse RL environments. If the goal is to maximize engagement, then the model can attempt to maximize on that. While their idea of utilizing humans as an RL environment is a crazy thing, it will be commonplace as AI. evolves. Once XAI implements RL on human environments, they could dramatically improve engagement and eventually turn to other goals as well. Becoming really good at Slop is a potential path. I think we were talking about this earlier. It's just a fascinating concept of like where this goes. It's a potentially very dark, very black mirror, very black pilling, but

could be very cool. Creating our creating many RL environments makes spiky intelligence, but But if humans are the ones to optimize against, perhaps it will be general. It's debatable whether this leads to intelligence gains anytime soon. But if the goal is to expand beyond engagement, it could lead to impressive advancements. Or it could just lead to a dystopian future where people are addicted to some anime girl they talk to. So is Dylan Patel over at semi-analysis. It's a wild. So the next section, the future of XAI, and the next few.

to three years, XAI's compute footprint will be heavily weighted towards training. Their inference demand is rising and will benefit the neocloud market. But in the short term, it will pale in comparison to their multi-gigawatt training expansion. While that situation isn't viable from a financial perspective, we think XAI's integration into X will likely provide some cash cushion. X is already integrating XAI technologies into its core advertising engine to increase monetization per user. Again, you know, like X-A-X, Twitter has never been an incredible cash engine throughout its history, but it's something semi-analysis says, but to support its LLM training buildout, XAI will need to generate tens of billions of dollars of revenue and follow the path

of anthropic and opening I. The core question is whether XAI's success would take market share from them or create net new spending. We think it'll be a mix of both. XAI's ramp will enable Gen.A.I. to reach a broader audience, but it also increases the financial weakness of the Gen. A.I. Market and the risk of market turndown as training still outspends inference for a precise breakdown of training inference capacity. Go hit the data center industry model. Do it. Just upgrade. Pull the trigger. Pull out the ramp card. Put it down for the semi-analysis. Gigateer. Pay it the most. And then hop over to figma.com. Think bigger. faster. Figma helps design and development

teams build great products together. Get started for free. Let's go. And we have our first guest of the stream, our European correspondent. Our UK correspondent. I love this song. We actually should license that.

We definitely are going to get monetized for that. Great to see you. It's my favorite sound. I'm back in the studio. We've got the soundtrack going. It's such a bangor song. It really is.

Give us the weather report first. How is summer? How was it Wimbledon? You went to some tennis event? What did you do? I had a tennis, yeah, I had a good tennis summer. I was actually the U.S. Open as well. Is that the same as Wimbledon?

Are they the same thing? I mean, they're technically the same on like, you know, seniority, but I would say, you know, as a proud Brit, that they are slightly different. Okay, got it, got it. And then, and then take us through this week. Why are we paying attention to Europe? So I think, you know, the headline here, I know that Jordi sort of said, UK correspondent, not European, is that we've had this huge state visit this week from, you know, the White House delegation. We also had a huge tech delegation come over, which really sort of marks the first landmark second state visit that a U.S. president has been given, which President Trump liked to remind us lots of times.

So it's the first time that a U.S. president hasn't been invited for a second ceremonial visit to visit the royal family. So you had the sort of pageantry of this big White House delegation to Windsor Castle. There was a big state banquet. it. You had, you know, Benson even taking off his black leather jacket and putting on his white tie, which is, you know, one level up from black tie. You know, we get very specific here. I saw tons and tons of horses. There were a lot of horses. There was a lot of gold. We love gold. That's great. So we really had one day, which was, you know, a showcase of diplomacy and

tradition. And then you had a second day, which was at Checkers, which is the UK Prime Minister's traditional sort of country home, you know, historical diplomacy is carried out there. And the focus on that was not just, you know, UK trade and investment. If you cast your mind back to May, I think I was on the show, we had the UK and the US come up with one of the first trade agreements, which given the sort of the landscape of tariffs, I think was a demonstration by the White House that they're really keen to make trade deals where they can. So you had a continuation of that, but the real focus this time was on tech and innovation across both energy and nuclear. You had it across AI and quantum and then also, of course,

defense and national security. So that's why you not only had the delegation from the White House and a lot of key politicians, but you also had some of our favorite tech CEOs from Silicon Valley making the trip over and I'm putting on the white tie and also announcing a lot of investment. Who were, who were some of the key names? Sam was there. So Jensen. Jensen, Benioff, right? But Palantir and a role as well. Yeah. Yeah, so I guess the main companies, you had Microsoft, so you had Satya there, you had Google, you had in video. So Jensen was was extremely bullish on the UK and we'll maybe come back to that because he's sort of announced a multi-layered investment in the UK over multiple years.

You also had, you know, carp here from Palantir. You had Brian Schump from Anderil. And, yes, Sam, Altman from Open AI. So it was really, you know, and there are a few others as well across, you know, Blackstone and Black Rock. You also had a lot of the large big pharma and healthcare companies, you know, GSK, for example, Black So Smith's Klein has announced that it's going to be an. investing a lot into U.S. R&D. So it was really a big showing across major industries for investment on both sides of the Atlantic,

but there was a huge, you know, representation from Silicon Valley. And, I mean, just some of the numbers are quite impressive. So Microsoft announced that it's going to put $30 billion into the UK of the next four years. You had Google also committing around $7 billion. $30 billion. So much. That's crazy. You had Google announcing they're going to put $7 billion into a data center that they've already established.

But the real number here that was impressive was Invidia. So Jensen really wasn't just taking off the black leather jacket for nothing. He's put $670 billion into a UK cloud computing company called N-scale, which is really designed to help AI, sort of the UK become an AI superpower, as he called it. So that specifically is into manufacturing and the infrastructure. And then he's also potentially talking about putting $500 million into Wave, which is the main autonomous vehicle company here in the UK. I don't know if you've had Alex Kendall on TVPN.

He's an amazing founder, you know, brilliant UK founder. So Jensen's really, I mean, he did a whole separate event where he met with Prime Minister Kit Stama and really announced with a lot of enthusiasm, such as the Huang style, that the UK is going to be a superpower. So AI superpower. How should Americans be thinking about Kirstarmer as prime minister right now? Like, where does he fit in in the left-right continuum?

Or there's a lot of countries that have leaders who are leaning more Trumpian in their style or leaning more focused on innovation or leading less lawyerly, more engineering focus? How should people think about him who haven't dove very deep? Yeah, it's interesting. So I think the first thing to know is that Kiyosama came in on a landslide mandate similar to Trump earlier in the year.

So early in the year to Trump. So in the same way that you had a new leadership coming to the US at the end of last year and then the inauguration earlier this year, the UK had a similar thing happen about six months before. So Kirstama came in, you know, his party, the Labor Party, had been in opposition for a very long time. You'd had Boris Johnson and conservatism for a long time. So it was a real big landslide change for the UK. So what that means is similarly to the White House currently, the UK Prime Minister has a pretty large mandate for change.

So even though, you know, when I was sitting in the US at the end of last year, there was some trepidation around the fact that, you know, the party is traditional. more left-leaning. It's called the Labour Party. However, what you've actually seen is, one, a real ability to drive change. So, you know, the number one thing is can you actually get legislation passed? And number two, and this was a sentiment echoed by quite a lot of tech CEOs yesterday, was a real pragmatism to actually getting stuff done. Now, my personal view is that the UK, after leaving the European Union, is now very much standing on its own. It's been able to, you know, reassert its relationship with the US. It's been able to carve out its own opinion on things like AI sovereignty. It can be sort of pro-AI. It doesn't get held back by

some of the anti, or I should say more risk-averse EU legislation. So what you're really serious out. Yeah, exactly. So even though traditionally you would, you might think of the, the Labour Party is more left-leaning. In the UK, everything is going to. In the UK, everything is going to to be slightly more central as well. It's, you know, it's just a smaller country. Yeah. And so I think that, you know, you're not going to see as extremes, but, you know, what I've been really, you know, happy to hear from a lot of the CEOs I've spoken with is that there's this pragmatism to getting stuff done and being, you know, a pro, a pro business, open for business leadership within Europe.

What, I feel like, I feel like this is a new thing where, uh, like, like, like the entire like America is going on tour like we saw this in the Middle East where it was like all of the tech CEOs and all the government people from the United States show up and they're like we're all doing deals today and then you see the Trump uh the dinner it's a little competition yeah yeah yeah and at the dinner it's like Mark Zuckerberg throws out a number and then Tim Cook throws out a number and it feels like it was the same thing this uh this weekend or this week thank you for it was like that but with with frilly outfits yes yes you know Tim Cook was also there. I mean, I'm going to tweet an interesting list afterwards, but I highly recommend if you

have some downtime, just Googling some of the images from the last few days because they were really quite something. You know, the pageantry is quite impressive on the UK. Nobody does it. Nobody does it like the royal family. What are the top sort of like priorities just for the British government right now in the US? I mean, there's a lot of focus on tariffs, immigration, the overall economy. The economy feels like running hot. Everything's at all time highs from gold to Bitcoin to the stock market. At the same time, we're seeing weakness in the labor market. How are things feeling over in the UK broadly? I think the UK's main focus is on growth. There are definitely a lot of issues, of course, around immigration. I think some of the highlighted issues on places

like X have kind of showed that there are a lot of, I guess, questions around immigration policy across Europe at the moment. The UK is definitely part of that. I think, you know, from my perspective, a lot of the focus is really trying to drive a pro-tech, pro-growth environment. I think that traditionally, a lot of the growth in the UK has come from the city. So traditional, you know, big finance players, it was a real hub for financial services. And I think that there's a recognition that the UK is actually, if you look at it, a really big hub for R&D.

You know, you've had, you know, Google DeepMind originated in the UK. You've had a lot of advances in both synthetic bio and big pharma come out the UK. You know, hubs not just in London, but, you know, real R&D centres like Oxford and Cambridge have really been leading for a long time. You've got the big semis company, Arm, based in the UK. So in the last few years, I think you've really seen governments on both sides of the aisle try and move away from like purely financial driven growth into, you know, tech and R&D. So that's something that I think the White House and Ten Downing Street are really aligned on. And, you know, this is obviously was a big sort of PR campaign to also show that the special

relationship, as we call it, goes both ways. And I think, you know, something that was significant. And a lot of people have said to me, okay, well, what's in it for the US? You know, we're obviously investing a lot of money into the UK. But I think that if you look at how the US thinks about exporting its own AI data stack across the world, and you put that against some of the competition with, you know, the Far East and China, this was a really good opportunity for the US to kind of reassert its position as the dominant AI player. And, you know, they signed a really interesting memorandum.

It was called the Tech Prosperity Deal. lovely name. Let's give that for tech prosperity. Yeah, exactly. Which basically allows the US to align on how data models are built

and some of the kind of very cool things that the US cares about in terms of ensuring that the data stack is defined by the US and not by someone else. Take us through an update on the private markets in Europe broadly, in the UK. Who are folks focused on at the growth stage?

We were at the New York Stock Exchange. Was that last week? I keep losing track of time. But Plurna, IPOed successfully, stock traded up. How was the mood? I mean, it was obviously continued to hold up pretty well. Yeah, the mood was good. Our takeaway was kind of interesting with the CEO is that it was just another day for him.

He didn't bring his family. He just came, raised some money. got in, got out. It was a fundraise. It was a fundraise. Whereas with Figma, it was this huge moment because a lot of folks in Brooklyn use Figma and Manhattan

use Figma. It's an American company, too. It's an American company. They really focus on the story of Dylan and with Sebastian over at Klarna. It wasn't as much of like everyone is a partner in this ecosystem. But in general, it felt like a really good moment.

People had been waiting for Klarna to go out for a long time, and it was great to see them get out successfully. But who else is driving the market or the news? Yeah, do you expect any type of spinouts from DeepMind as well into more like application layer companies or do you feel like the researchers there are just going to be content to, you know, double down?

So I think just on that front, as well, as you mentioned, I think both Dylan and Sebastian, amazing founders, really exciting that we're seeing the IPO market off to a, off to a good start at the moment. And, you know, I think... Although Stubhub, I think Stubbhub is down. Stubbh had a hard time. From listing price yesterday, so...

It's okay. Well, we'll get it all back on the next one. But I think that, you know, in terms of names to watch here, the big elephant in the room for a long time has been Revolut. So for those not familiar in the U.S. market, Revolut is really the largest fintech player. here in the UK. Its last valuation, I think, was around $70 billion.

The last employee, I think that was a secondary. So that's really been the main question. Some of the news stories around that have been around, it's banking license over the years and whether they were going to, you know, go for an IPO. I think that one, you know, continue to watch on. And if you look at Jensen's comments,

even on this trip, he's even said that, you know, He's even said that, you know, NVIDIA may also invest in Revolute. No way. Jensen's really, he's really keen on the UK. We love, we love Jensen. He's pumping everything. It's great. I think in terms of Google Deep Mind, I'm unclear if they're going to be, you know, spin-outs.

I think that, you know, the UK government has always looked at deep-minded as an example of UK companies that really should have been able to stay and prosper in the UK. or within Europe. And so I think that you're going to continue to see a lot of focus on, you know, UK's own AI infrastructure, the energy stack that's required for that. So we also had quite a lot of announcements

around nuclear this week because, of course, the energy question is a big one for the UK. Interestingly, actually, one of the other things the White House got out of this week's negotiations was that there was a commitment to the UK from being completely unreliant on on any Russian nuclear energy as well. So I think that, you know, you're really seeing both the UK shore up its energy stack.

And I think a lot of that is to do with the fact that UK really wants to try and, you know, make itself a great hub for AI talent, whether that be through Google DeepMind or some of the other really exciting companies like Wave. Yeah. Makes sense. As extra context, in 2021, Demis Hesibis over at DeepMind co-founded isomorphic labs, which is a spin-out, but still, I believe, private and owned by Alphabet and DeepMind. And they're focused on artificial intelligence for drug companies, not Al-Bet-owned, that type of thing. And he's still based in the UK, which, you know, the Prime Minister always likes to point out,

but he's still working with Satya, so, as though, with Google, yeah, with Ruth. Yeah. Well, thank you so much for hopping on the show. Always great to get that date. Always great to hang out with you. Always fun to chat with you guys. We'll talk to you soon, Pippa. Cheers. Bye.

Bye. Hopefully we can have a new tennis tournament sometimes soon, so Pippa comes back to the U.S. I'd love it. Seems like that's what kids are over here. Should we talk about? Vanta, automate compliance, manage risk, and prove trust continuously.

Vanta's trust management platform takes your manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework. We're managing a complex program back in America. We should talk about this house in Malibu. Johnny Carson's longtime Malibu home lists for $110 million. I've been to this house. An early investor in Jewel and a hard rock cafe,

Aris are selling the modern California house, which the talk show icon owned until his death in 2005. It's 10,000 square foot, six bedroom house on roughly four acres. And we'll go through a little bit of this. For more than 20 years, Johnny Carson lived in a modern, triangular-shaped home perched on a bluff, overlooking little point-dune. How did you never... I've talked to you about this.

You told me that Johnny Carson lived on the bluff in point-dune. Yeah, I definitely told you this. As we were driving by it together. Interesting. Probably told you multiple times. Interesting. Who knows? You might have been locked in.

You might have been locked in. There might have been something more... There might have been a more pressing current thing to be monitoring. You might have been the monitoring situation. This house is... It's amazing. The property, which has changed. Chains multiple times since his death has been redecorated and he's going back on the market.

The price tag is more than double what the sellers paid for it in 2019. He had 40 million in 2019. I know, I know. Every time you look back at things that you could buy before COVID, it's painful. Yeah, it's crazy. The fame talk show host bought the estate in the 1980s, about a decade before ending his 30-year run helming the Tonight Show. He owned the property at the time of his death in 2005. The current owners are the venture capitalist Reyes Vallani,

an early investor in the e-cigarette maker, Jewel, and his wife, Hard Rock Cafe, Eris, Augusta Tigrette. The couple paid $40 million in 2019, as you mentioned. Look at Johnny Carson looking great. In, of course, a Johnny Carson suit. Fantastic, interesting story about this. Johnny Carson was so dominant in the culture. I mean, the Tonight Show was one of just a few.

late night shows because we didn't have the proliferation of the internet there were only a few things that you could watch every night and he was absolutely dominant how many how many viewers was he pulling nightly i believe it was like tens of millions it was it was massive i might have that wrong but uh the interesting story was that he was so focused on his craft he was not interested in business at all really and so he had a business partner who said hey you wear a suit every night, we should get you a custom suit. We'll start a D to C apparel company, effectively, Carson's suits, and we'll sell those suits, and then you'll be able to market them on the show. But he didn't really care about contracts or understand business, and so his business partner.

We've got to get one of these. I'm on eBay right now, in 1970s Johnny Carson's suit for 170. You got to get one of those. That sounds amazing. I got to get one of these. But his business partner set up the Carson's suit enterprise, and I don't believe he had equity in it. And so he would get free suits, but he wasn't making any money off of it. And for years, Carson was not monetizing well at all, but eventually he went back and renegotiated his contract, got ownership over the Tonight Show, and was one of the few late-night hosts to actually own. his own show, have full control over it. And then after he died and retired, the Tonight Show, the catalog, and there's merch,

and there's DVD sales, and there's greatest hits. All of those would be sold and continue to generate revenue, even though it was a daily show. But there were a lot of iconic comedy moments that happened on that show. So back to the house, Riaz, Valani, and Tigrette. Also have homes in London and Northern California. They viewed the estate as an investment property where they could entertain, Tegrette said. They redecorated the two-bedroom main house and had plans drawn up to

add two new houses in another pool but didn't end up building them. It is a remarkable property how large it is and then it's just a two-bedroom because, but it has all these massive open spaces you can see in those images. They are selling because they have decided to focus on real estate investments in northern California. It was designed by architect Ed Niles. The main house is about 7,100 square feet. Triangles appear throughout the house. The dining room is a triangular shape as is the wood and glass ceiling as as is the wood and glass ceiling of the central atrium where mature trees are situated the estate also has a two-bedroom guest house a tennis pavilion with two guest rooms and is said to have been a gift from from carson to carson

from nbc to grette said one of her favorite parts of the house is an outdoor dining area with a large round table where guests would congregate violating the triangle symbolism with the round table. We had so many important business meetings there, she said. The era is naturally shaded thanks to a coral tree and a pepper tree whose branches grew together over time, creating a leafy canopy. And it is beautiful. There's a massive, there's a massive tree inside the house. I don't know if you can see this in the images, but, and the tennis court is. This view of the tennis court. It's so crazy how big this tennis court is. It's wild. Yeah, you can see that. That's insane. So insane. The property is among the largest in point doom with where standard lot sizes are

roughly one acre. This one is, of course, four acres. The main house is situated. Is that the listing agent Chris Cortazzo? A friend of mine. Would you know him? I bought my house through his firm. Oh, really? That's funny. The main house is situated closer to the bluff than houses that can be built today, he noted. The $110 million asking price is warranted by recent trades. He said, including a property that recently sold over a million. Yes, there's a house right nearby this. That's like a, I think it's like a two or two bedroom house. that has similar views that sold for 80 million. It's literally a shack that's sold for 80 down the road. So this makes that makes 110 for this spot look like an absolute bargain.

It's great. You know what else is a bargain? Graphite.dev. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. You can get started for free. Road River Partners I've never seen this account before

no profile picture says in the late 1990s a pal attended a dinner for tech hedge fund CIOs all the top guys all the top guys were all the top guys at the time $25,000 of wine was consumed and portfolios

were shared my pal was sitting at a hundred and thirty-five percent gain for the year and his performance was the worst in the room by far a week later he'd gone, he'd went 100% cash. He was the only one that survived the next three years.

Insane. Pretty crazy. You're looking around too much wine. I'm up 135% and I look like the idiot. Maybe it's time to... Yeah, the account here asking coward, Mr. Coward. Coward nutlick.

But why did he go 100% cash? Because everyone else was up with. What if he just, like, gave up? He's like, I'm terrible at this selling to cash. He's going to just like distribute everything out and just quit. And then he just looking around like, wait, I'm the only one standing. Yeah. It's remarkable.

I do know. My dad always told me a story about one of his college buddies just went 100% to 100% to gold, like right before the great financial crisis. and just, like, ended up just absolutely printing. Well, whatever asset you want to invest in, go to public.com. They got multi-acid investing. It's investing for those who take it seriously. They got industry-leading yields trusted by millions.

We're not allowed to give financial advice, but I think it's fair to ask you to take it seriously. In other news, we interviewed a lot of fun people at the MetaConnect 2025 event, the one that broke through on the internet, on teapot, on X. was, of course, our interview with chief wearables officer at Luxottica. Rocco Basilico. Rocco. Malin says, I'm sorry, his name is what?

We live in a simulation dog. And this got 3,000 views. There were a couple other posts list. So we need to get some background on Rocco, 7.4. No relation to Rosco from the Roscoe chicken and waffles chain. that we really enjoyed talking to him. Also, no relation to Rothko, the artist. Oh, Mark Rothko, yes, I don't think so.

People had a ton of fun with this. Where was the other post in here? Someone else was saying something about. He's Italian king. My name is Rocco's basilisk, and I'm making artificial intelligence that you put on your body. So let's give some backstory.

Tyler, do you want to give some backstory on Rocco? Yeah, what is the joke that people are making here? What's wrong? Okay, yes. So his name basically sounds like Roco's Basilisk. Yes. Which is like this famous thought experiment from less wrong in like early like 2010s, I think. Yeah.

Basically the idea is like, okay, so there's going to be some future super intelligent AI. Yep. And like you can do this kind of like decision making where it's kind of similar to like this prisoner's dilemma thing where there's two agents. Yeah, Pascal's Wager. Yeah. is Pascal where basically you can chart it out to where like the idea is um you have this this the super intelligent AI um by the time it gets to becoming like instantiated um the idea is that it's going to like punish people that did not like accelerate it did not help

it be born and and the idea is actually it's not just like you didn't if you didn't love me at my benchmarks you don't deserve me at my ASI yeah yeah so it's basically like The whole thing is, like, it's like an information hazard, this, like, term. Eliezer actually banned, like, discussion of the topic. Wait, why? Less wrong for, for, like, five years. Oh, because it's a justification for accelerating superintelligence. I think so, but also.

Yeah, yeah, you don't want to be the safety guy. Yeah. But the whole thing is, like, if you charted out, basically, it only applies to you if you, like, learned about this idea. So, like, if you knew about the idea of Rocco's Basculance, then, and you didn't, like, start accelerating AI, then it's going to punish you. Yeah. And this is why we covered MetaConnect and we covered...

Isn't there some lore about Rocco's Basilisk? The whole idea is like it's a basilisk because like... What is a basilisk? Is that a snake? Yeah. It's like in mythology, if you look at the basilisk and the eye, then you like instantly die. Yeah. So it's like

if you learn about this thing and you don't do anything, then you like will be punished by the future AI. Apparently this is this is what this is what originally brought Elon Musk and Grimes together. Musk was going to tweet about Rococo's basilisk, which is a reference to Rococo, the 18th century Baroque art style,

and they bonded over that. He discovered that Grimes had made the same joke three years earlier and reached out to her about it. One can only speculate about it. Very, very interesting. thought experiment. Well, we should have, we should have asked him about it because I wonder, I wonder how aware he is. It can't be the first time he's, he's heard that his name sounds like Rocko's Basilisk, right? Yeah, I assume not. He must be aware. Well, we'll have to, I mean,

the story of him just like cold emailing Zuck, like, hey, I think we should do some international business together. Let's hop on a quick call. He made it happen. I wonder if Zuck knew about Rokker's Netflix when he got that, because then it's like, you're getting an email from Roeco's Vasselist asking if you want to help. help accelerate AI. Yeah, you have to say yes. You have to say yes. You have to say yes.

You have to say yes. Yeah, I think it's one of the most under-hyped, like, partnerships of the last few years, right? It's so strategic for both sides. It gives meta this incredible, durable advantage to be able to innovate on the technology side and not on the design side. Yep.

And it gives Luxottica the ability to defend itself and actually thrive in a world where, you know. We were debating. Everybody's wearing smart glasses. I think that's actually a thing, right? It's like there's people that don't wear glasses today that may or may not be using smart glasses in a few years, but Zuck brought this up.

There's like one to two billion people in the world that wear glasses all the time and looking good. Yeah, I mean, people were, we were debating with a few people about whether or not the meta-AI glasses. Oh, I just triggered it. It's talking now. Whether you can immediately clock them, whether they look like tech nerd gear,

or whether you can pass amongst normies, the unaugmented, as someone who's just wearing raybans. And I was leaning in the camp of like, if someone throws on meta raybans and they do have the cameras on there, but they're not recording, there's no light, you kind of don't notice. But other people were saying, I don't know, like you can still see the camera. lenses and it still reads as, uh, it's a little like tech nerdy, but I think it's like the specific possible. That's a specific frame though. Like this is, this is actually an oakly frame with no, uh, with no, uh, sunglasses in it. It's just eyeglasses. Um, but, uh, but the, the, the original meta ray bands, they do look exactly like ray bands, an iconic silhouette and something that, uh,

people are very confident in wearing. And I think even with the cameras on there, even though you can see them, uh, people are much more willing to, to throw. them on when they go out. And you do, you do face a little bit of like, oh, like you're like a tech nerd. You're trying like the new thing. Like it's not as it's, I mean, it's like caring. Yeah, but think of, think of the billion plus people that just wear glasses because they need them to see. They already wear glasses. And then you say like, hey, do you want a small display in those? Yes. That can give you directions. Yeah. You can use to play music. Chunky frames are kind of in right now. It's kind of a vibe and so I think

I mean I agree it's it is an underrated partnership it's a great partnership I wonder I'm I'm like dying to know what the the well first off

like what Apple's response will be like because I feel like they they could I'm imagining that they're going to be white for some reason I don't know why but I just have like Well, I also think they could do the clear one, like the old Macs. Yeah.

You could see all the internals. They've really moved away from that aesthetic, though. But if they brought it back, I think people like. It just feels like they haven't even done black AirPods, right? Like the AirPods. Maybe they could do the beats line and use some of that. By the way, apparently the new phone scratches like crazy. I saw, yeah, I saw a picture of it crack.

A bunch of people have been reporting this. But, yeah, what was it? They went back to aluminum or something? Yeah. People don't like that. Titanium. It was a good run. I don't know.

It was a good run. I think, well, you, you already purchased it, so it's in the mail? Yeah, this is, this is bullish for Apple, the fact that I'm not going to get it for more than, like, basically, I'm going to get it in like three weeks. Yeah. Well, they're selling. Julius AI, what analysis do you want to run? Chat with your data and get expert level, expert level insights in seconds. Go to Julius.

Let's pull up this video of Matthew. Matthew McConaughey. We'll be right back. Matthew McConaughey says he wants a private LLM fed only with his books, notes, journals, and aspirations, so he can ask it questions and get answers based solely on that information without any outside influence. Now, like, that's impossible, right? Just because you need to do pre-training on something or other, right? Yeah, I mean, unless his books are, like, the length of, like, the internet. Yeah. I don't think the model would be very good if you just train it

on, like, five bucks. So, so really he's, he's kind of saying, like, he wants something fine-tuned? Yeah, I mean, it seems like he doesn't really understand how they're trained. Yeah. But, yeah, you could instantiate this, I guess, through fine-tuning it. Let's play the Matthew McConaughey clip and see what he has to say. LLM. Who I can upload, hey, here's three books are written. Here's my other favorite books.

Here's my favorite articles I've been cutting and pasting over the 10 years and log all that in. And here's all my journals, whatever. or the people out and log all that in. So I can ask it questions based on that. Right. And basically learn more about myself. Right.

I stand on the political spectrum. Right, right, blah, blah, blah, blah. I'd like to, no, that's what I would like to do, which is sort of a glorified word document. But it still would hold a lot more information than just, oh, can you find this term? I would be asking it, and it would be responding to me on things that I've forgotten along the way. Load it with the information I'd like to load it with. Right.

Yeah. Maybe even, like I'm saying, in the words of belief, and the man I'm working to be, and the man I want to load it with that, load it with my aspirational. I both have a phrase for this. And then ask it, and it's giving me answers going, oh, this is, but before it's slowly learning about me

through conversations, then going, oh, I think this looks like based on a conversation. No, I want the answers based on what I've uploaded. Can you play the beginning of this again? He uses some funny phrase. Well, what does he say at the beginning? I am interested, though, in a private L. LLM. Private.

Who I can upload Hey, here's three books are written. Here's my other favorite. You can do this with

my favorite articles I've been cutting and past in over 10 years and the log all that with PDF. Log all that in. That's what he keeps saying. Okay,

yeah. I love that he says he wants to log it in. No one's talking about logging data into your private LLM. Like this is the future. This has been...

No one's tried logging content into a private LLLLLLLLLLLL. This has like been possible for like three years but no one has embedding your embedding that doesn't sound like logging something in that's what i want i want to log it in i want to log in my favorite book i want to log in my favorite article log in the holy bible sounds like that's what he's doing uh he should he should hit up sales force

for this he's he's oh true yeah with uh what do they have merlin ai or something well he's close with he's one of the faces yeah he's one of the faces and if einstein a i i he should fine tune it fine-tuned it on my, on Matthew McConaughey mode. That, I mean, Einstein AI, I feel like, you know, missed opportunity to personify the AI much more. You should have gone, you know, Microsoft's going to have Clippy. You know, we're getting ony from XAI. Every company is going to need their own personified AI.

And, you know, I'd be much more likely to use Salesforce's AI if I knew I was talking to Matthew. Matthew McConaughey. In other news, Howard Lutnik is posting screenshots of his portfolio. Wait, really? No, I'm kidding. I mean, it's close. Pull up this next post. Yeah.

It's a picture from Trump on Truth Social of an AI slop of him sitting at his desk with bought Intel at $20. Now it's $30. He's happy. Well, in other somewhat political news, TikTok sale announced by, we're tracking this on Polymarket, December 31st is now at a 79% chance that a TikTok sale will be announced. And Polymarket also got a market up on will Tesla acquire X-A-I in 2025. So the market will resolve to yes.

It is officially announced that X-A-I will be, has been, or is being acquired or merged with Tesla. I can see that happening. I'm interested to keep track. see. In other XAI news, Nikita Beer says the goal for your X timeline is to get out of the mainstream Algo and political crusades and find your niche. You should be able to post about your interests and have friendly, relevant people chime in. And I feel like I've been able to do this. Like you do have to be proactive about it still. You have to say I'm not interested in this post. You have to send a lot of likes and retweets to content that you actually like,

engage with it, reply to people that you like in real life, that you like. You like. like in your niche, but my algorithm's been doing well, and I'm hoping that they continue to improve it. Elon says the algorithm will be purely AI by November with significant progress along the way. We will open source the algorithm every two weeks or so. I wonder how that, like once it's purely AI, it's driven by just weights.

Are they going to open source the weights? Open source, open weights or fully open source the model? Like that sounds crazy. What do you think, Tyler? It's kind of unclear. I mean, by AI, I assume he means like LLMs, Because, like, if you just say, like, AI, like, it's kind of already done by AI. Yeah, it'd be very interesting to, like, open source, the open source, the algorithm.

And the algorithm is just, like, recommendation. Dot next, you know, parentheses, like, like, inference of the model is the, is the, is the Python code. And it's like, yep, we didn't change it. It still just goes. This is crazy. Breaking news. Trump is adding a $100,000 fee application fee for H1B visas in the latest. crackdown.

In other breaking news, fall.a.I. The world's best generative image and video models all in one place, develop and fine-tune models with serverless GPUs and on-demand clusters. In the Wall Street Journal, Tyler.

UA.E. Project. I want to add tbPN.com slash sounds, and it's the digital version of our soundboard that people can use in their meetings. Ooh, I like that.

Yeah, we got to do that. Given to me in a DM. Yeah. and sound bored at home so yeah just try to wrap it up in the next make it wrap it up in the next like don't make mistakes and don't make mistakes uh don't make mistakes uh the united arab emirates is buying houses in washington dc uh they have a 200 million dollar collection now the government's latest purchase is a 27.5 million dollar spec home in mclean uh virginia i don't know if i'm actually saying that correctly uh this week it added a roughly 21000 foot spec estate in

Virginia paying $27.5 million in an off-market deal. The deal is the most expensive home sale to close in the capital region this year. The UAE directly or through its employees or royals owns at least 21 properties in the area worth roughly $200 million. The Emirates have purchased so much real estate at such high prices in recent years that they've helped reset values in the Gold Coast region. Let's give it up for resetting values. According to real estate agents, the house was recently restaged to give it a modern minimalist look. Each time they purchased something, it's a big number, says a real estate agent of Compass. The estate is located near Langley where the CIA is based with a six-bedroom main house and a pool house. It was constructed as a spec home by Ambar Homes, a luxury development firm, headed by local dentist and entrepreneur, Dr. Sonu Kakar. Carr said he bought the land for $2.1 million, started construction in 2022. It took him four years to start construction, completed the house, two years later.

So he's selling this big house. It was first listed for $2,99. Came down a bit, but they got the deal done. There's a marketing video showing a motorcade of Cadillac escalades, resembling a diplomatic entourage heading up a winding driveway to the newly built modern mansion. Check this out. They made a hype video? a hype video, a startup launch video for their house. It's funny. It's like they knew

they were selling this to. They were like, we got this house and we're going to show a bunch of a home hype video at the timeline. Wait, wait, did you guys scroll that that fast? It plays it that fast? That's in the video. That's baked in. They're doing hype videos for sure. That's a vibe reel if I ever saw one. The UAE's residences are dotted in large clumps across the capital region, typically on prime waterfront lots. On Crest Lane, the country owns several homes on the Bank of the Potomac, where that's where the ambassador's official residence and guest house are located. Well, you've got to go hang out in D.C. do deals, and so you've got to have some big houses for your folks to come and hang out at. On Chainbridge Road, one of the area's most

upscale neighborhoods, the Emirates paid just under 43 million in 2020 for a riverfront estate previously owned by AOL co-founder James Kimsey in 2020. The same year, the country purchased a vacant lot nearby for 20 million. Are they actually just buying these on the balance sheet? I think so. I think they got cash to spend. And they want to come visit. When they want to come visit, they want to have a nice place to entertain, host guests,

do deals all the above. If you're a startup, you got a couple hundred million dollars burning a hole in your pocket. you're doing deals in D.C., pick up a $23 million mansion. Yeah, why not? Mark Zuckerberg did. Earlier this year, Meta founder, Mark Zuckerberg, and his wife, Priscilla Chan, purchased a $23 million mansion in Washington, D.C. Of course, that is much smaller than what the MRIs have been picking up.

But regardless, you've got to get your brand mentioned in ChatGPT, go to profound, reach millions of customers who are using AI to discover new products and brands. T.J. Parker has put it. Yeah, that's a good post. case for Apple and AI is they're going to completely avoid enormous amounts of CAPEX spend, get most of the consumer value from intelligently integrating Gemini, and get paid a boatload of money from Google for doing so. And I agree. I think, I do think it is frustrating as a user that Apple has moved slowly and somewhat ineffectively on AI, but I don't

think they're in a bad spot from a business standpoint. No, no. Nothing's gotten me to consider they're moving off of Apple. Like, even if iPhone Air doesn't have the right battery life or iPhone Pro, you know, cracks or whatever. If I need a case on it, it's like, I'll solve all of that. I'm sticking with. And that's a big question. Sticking with Cooper T.

Now. The big question with the meta, the new meta displays is, is there a world where, because right now when we demo the displays, you can get an Instagram, DM, you can get a WhatsApp message, and it will pop up in your display, nice little notification. but the question is, will they be able to figure out a deal with Apple so you can get iMessage piped into there?

So yesterday, Metacnect continued and they did talk about the developer kit, the SDK, how developers can integrate. And they had a couple of integrations. They did one with Twitch, so you can live stream from the devices, from the meta rayband displays. They did one with Stream Labs and Disney,

and there were a few others. they're still pretty niche and they're all partnered with like pretty big big companies, the type of companies that would do like pretty big press junkets for these but some cool demos I thought that they were interesting I'm very interested in the Twitch demo obviously we live stream on Twitch,

thank you to Bobby Cosmic in the Twitch chat holding it down but it seemed like the live streaming capability which we didn't expect to launch very quickly is going to launch it is going to be live you are going to be able to live stream from the glasses, but the amount of time that you can actually spend live streaming is pretty short. It didn't seem like something where you could be streaming for hours.

It was probably minutes. There were different timers that popped up when I was looking at the video, kind of showing, okay, you know, if you keep streaming at this scale, it's going to, the chips are going to be on fire. You're going to run out of battery life pretty quickly. And so, like, there might be all-day battery life, but that's more like all-day battery life

for the occasional picture and the occasional meta-a-a-I-L-LM query, not all-day battery-life stream your entire life. Like you were mentioning that the AI mode was maybe an hour max. Live AI was maybe an hour max. That's the using it to translate, you know, real time. And I think it's also taking pictures

of what you're seeing while it's doing that for the extra context there. Anyway, whatever, if you're planning to build a product that integrates with meta-AI display glasses, You've got to get on linear. Linear is a purpose-built tool for planning and building products. It's a double-kill.

You can plan and build the products. Meet the system for modern software development, streamline issues, projects, and product redmaps. Jeremy Giffon has a post. Hot-take. Wouldn't be surprised if this was printed in the legacy media soon, but for now we're seeing it on X. By hand, wrote some drafts out on a piece of paper and then transferred it.

had maybe a member of his team, one of his guys, actually. So he says the chief of staff is overrated. We need more confidants. Eminence Grizzes? I don't know what that is. Viziers, Praetorians. There is a phalanx of different guys you can have at your side. You don't need to copy the president.

Consider the king or the priest or the emperor instead. Bangor. Must have the title grand in front. I think we got to give a new type. title. He needs to be a confidant or a grand vizier. I think a vizier is a good, is a good title. Our first hire was a vice president. Our first hire was a vice president. We have a wheelman. We have a VP of logistics. We have a few other folks, a CIO, of course. Yep. Chief intern officer, Tyler Cosgrove over there holding it down.

Yep. Jim Kramer says that 30-year treasury better start behaving. That is bar. order. He also posted a picture of him hugging Tim Cook saying own it, don't trade it. I love this. He said this 10 years ago. We reacted to his interview

with Tim Cook on Squawk Box, or on mad money a decade ago. The stock had sold off. They'd hit earnings. They beat earnings. They'd generated something like $10 billion in free cash flow, but it wasn't enough for the street. And so

Kramer, did a very interesting interview with Tim Cook where he explains that he is excited about the company. He personally is bullish on the company, but he has to get the story straight for what the street is saying. And the Wall Street traders were saying, Apple, will this company go the way of HP or Compaq or IBM? Have they reached stagnation? And Tim Cook says, no, we're going to keep growing. We're going to keep compounding. you don't even know how much revenue we're going to make off of the App Store.

And he kind of teases, he says, hey, services are going to be big. The iPhone's still going to be very big. And he was 100% right. Got a great toll road going. Based 16G says. Speaking of toll roads, the tax is a toll road on your business. Don't waste any time with it. Head over to Numeril HQ.

Put your sales tax on autopilot. That's right. Numeral is the platform for sales tax and VAT compliance. And we have Sam from Numerol, the CEO, joining in just a little bit. Looking forward to that. Bay 16-Z just shared a screenshot of CNBC. They said New York will soon send its, quote, first-ever inflation refund check checks to taxpayers.

Is this the state or the city? The state of New York will be given you a refund for inflation. They're paying back. Concerning, concerning. Well, I'll see. What else we got here? There's news about OpenAI. OpenAI has approached Gore-Tech, which assembles AirPods, home pods, and Apple Watches to supply components, such as speaker modules for OpenAI's future products. One of the products OpenAI has talked to suppliers about making resembles a smart speaker without a display, the people said, would you carry this with you? Would you have this in your home?

Open AI has also considered building glasses, a digital voice recorder, and a wearable pin. and is targeting late 2026 or early 2027 for the release of its first devices. Digital voice recorder, you know who really convinced me, who was really convincing at the MetaConnect event? Was it the VP of wearables? He was saying he believes in kind of lindy form factors. You wear a watch. He had a garment on, you wear glasses.

People don't wear a lot of pins. Maybe if you're the hand of the king, you wear a person. pin, maybe a crown, a cane, a smart cane is an interesting idea. But it does seem like it's harder to get people to wear a chain of pin. Open AI is saying we're going to do a speaker, or at least that's, you know, reading between the lines, that's what they would do here. But Apple is, remember, they're building the smart lamp. Yep.

For your. Are they? Well, this was T's a while back, remember? It's not really that. It's an iPad on a robotic arm that can follow you around the room in the kitchen. I know. I mean, I just meant more of the Pixar-Lamp. It does feel Pixar-Lamp-esque, and I'm actually very excited about that product.

I think it'll be very cool. But maybe they should... There was a post a couple of days ago from someone saying, like, Apple should lean into... They should make just an actual lamp. You don't need to throw crazy AI. It doesn't need to be iPhone-level internals. Just apply your design skills to... industrial design all over the home.

We've asked for this stuff, the TV. The TV, but it's not going to happen. It's not going to happen. Unless you have the full flywheel of you can take a cut of everything that happens on that device, you have an app store, you have all the downstream stuff, you have a solid

replacement cycle. When was the last time you upgraded your TV? Happens like once every five years, 10 years. And yet people upgrade their phones, they lose their AirPods, they get new AirPods, they get the latest iPad, IMA. But Apple made it a battery powered TV, and you'd be forced to replace your TV every...

I think you're talking about the Applevision Pro, which is the vision for, you know, where that goes. Well, speaking of Hollywood Visions, NIR is shouting out Alibaba's character swap via WAN 2.2 animate, very impressive and free and open source. By the way, this lets you insert yourself as the main character or anyone. Here we go.

Let's play it. Mr. Sorkin, you are five minutes late. Is there a reason why I should let you have? I'm just trying to ditch the cops, okay? I don't really care if you let me in or not. This is going to be pretty viral. Mr. Spector will be right with you. This is going to definitely launch some startup videos. Can I get you anything?

A coffee here? Someone's going to use this to do a startup launch video very soon. Pretty cool, swapping yourself in. It really, it really does it flawlessly. Remarkable. Look at that. I'm a fam. I have to know.

Yeah, another tool in the toolkit for video creators, much like fin.AI, the number one AI agent for customer service, number one in performance benchmarks, number one in competitive bakeoffs, number one ranking on G2. We have another, the management section today, it's Friday in the Wall Street Journal. Another one. Kara Ross, the ex-wife of billionaire Stephen Ross, who created related properties, the man who created the strip mall, essentially. She is putting her New York home on the market. The gemologist and jewelry designer bought the Lenox Hill co-op around the same time she filed for divorce

from the Miami Dolphins owner. 9.75 million, she's asking. She bought it for $5.6 million in 2021. Around the same time she filed for divorce. She tapped the lauded designers Tony Ingrau and Randy Kemper, who worked on the couple's former home in Columbus Circle to reconfigure and redesign the apartment.

The apartment is, is 2,500 square feet, has two bedrooms and a large terrace. The terrace is 2,000 square feet. So she referred to it as a diamond in the rough, of course, a play on words because she is a jewelry design. Yeah, she is a jewelry designer. They've been worn, her designs have been worn by the likes of former First Lady Michelle Obama.

And so if you're looking to pick up something in Manhattan, give her a call. It was a classical apartment with hallways and drop ceilings and beams and all in, and beams in all the wrong places. They focused on opening up the layout so that the terrace is visible from the entrance to the apartment, even removing one of the three bedrooms. Now the expansive living area

has a backlit onyx bar that can change color to fit the mood. They've really got to swap out the current furniture and just redo the interior because this place is not looking as good as it could. It doesn't look homey. It doesn't look like you could just hang out here

and crash. In a short time that she's lived there, Carr says she's hosted several events including her daughter's engagement party. I love to have friends and family over to create a nice environment for them. The co-op market has been soft in recent years, but listing agent expect to see demand for this property

because it is newly renovated. Apartments like that sell for a premium because time is money. Homes that require renovation are still trading at a discount. In 2022, Stephen Ross sold the couple's former condo near Columbus Circle for about $40 million, a significant discount from its original 75, million dollar price tag. That building was completed in the early 2000s by a partnership

that included related. So he used his own firm. Stephen, who acquired the Dolphins around 2009, has since launched a new company focused on development in West Palm Beach, which is very, very popular. And if you're looking to grow your business, get on Adio. Customer Relationship Magic. Adio is the AI Native CRM that builds scales and grows your company to the next level. Fen Lang has a good purchase. Best $25 purchase at Bluetooth device to connect AirPods on any plane. A couple of people post about this.

I put this in here not for this use case, but to bring up this sort of funny state of the world where every single like hospitality-esque business just like adopted

USB. Yes, at the worst possible time. And now everything is on USBC. So I was on a plane the other week. They had a USBA it was a charter that I got through Preston Holland and I was just like the plane had been refurbished in like 2020 and it still had USBA I was like what's going on here and then we were at the hotel for MetaConnect the other day and it's like great I'm glad I have USBA charging here next to the And it seems like you could create a business just like going around to like a lot of these different types of companies that have installed USBA and just be like, look, like this is not a real, this is no longer an amenity. In fact, it's a frustration because I check into a hotel room and I'm like, great, I'm glad I could charge a USBA device here.

Thank you for this. Hopefully USBC is the final, the final format. Although you know that there are like 12 different standards for what goes in. in a USBC cable. So USBC is the shape. It's not actually the data transfer. The final shape, I just wanted to be the final shape. Which is fine, but there are slow USBC ports.

There are fast USBC ports. There's Thunderbolt 4, which is a faster protocol that goes over USBC. It's all a mess. Well, we have an iPhone review from Joanna Stern. Let's pull up this video. I thought this was important for the audience to see. So she has a new iPhone. This is the new iPhone.

This is the air? This is the air. I feel like the air had round edges. It has sharp edges. That looks like the air to me. He's cutting cheese with it. She's cutting the cheese. She has a little viral sense going.

She, of course, is the tech reviewer over at the World Street Journal. Well, without further ado, we should bring on our second guest to go and our first in-person guest. Let's go. Aaron. Welcome to the show. How you do, Aaron? show announces series A today. Congratulations. Good to see you. Welcome to the Ultradam. You got a mic here, keep it close.

Hello. Introduce yourself. How's it going? Aaron Slodov, CEO, Atomic Industries. Very, very few people have sat here at this table with us. I thought this was appropriate, right? Yeah, that's great. You got a round to announce? What's the news today? Well, the first news is I got some merch here. Let's see it. It's the first of kind merch, and you can take a look at the hats or some of the shirts. You're going with wild colors. I'm going with brown. Brown is the most underrated color on earth.

That's a clay. We've got a good clay. Oh, a clay. Nice. Here we go. Very, very cool. What's the inspiration? The inspo for these?

Yeah. So these are obviously not like our official logo. And I have, I'm going to have to air drop you some hoodies, I guess, too, but do it. I had a great guy on Instagram. who basically takes corporate logos and, you know, vintageifies them. And, yeah, so special, special edition. 2019, it's overnight success. Yeah.

Yeah, we actually met in Pasadena, we got coffee. Probably 2021, something like that. Yeah. It broke it all down. It's been. Yeah, we're absolutely brutal to start an American dynamism business like four years before the word was. Yeah, for sure. The trend was created.

Yeah. Yeah, break down the news today. All right, so series A. You can hold this. What is there? Oh, boy. So we have... A lot of gong hits.

Yeah, a lot of gong hits. We put that through its paces. So this is kind of, you know, a bunch of great market validation. We're going to dump this into R&D, a little bit of, like, expansion on the footprint. And actually, if you want, instead of doing like a fancy render, I can actually show you guys a little clip of the new factory. Let's do it. We have a video pulled up already.

Let's pull it up. Wait, this is in Detroit. This is in Detroit. Making Detroit. Bring it to the West. This is, wow,

that's huge. That is a huge. You make a lot of things in that. Yeah, it's a lot of, we're going to keep the floor dirt like that. We had you, we had you call in and you were walking around the previous facility.

It didn't look small. What goes on the floor? Are you like at the point where you were like copy-paste? the same work cell over and over again? Not quite yet, but the idea behind what we're doing is we started kind of software first, to kind of prove the concept that we could actually design something in the same clip that like a trade expert could.

So we did that first. That was enough proof to actually get the first factory off the ground. And then the design software started saturating into the molds that we would make, right, like at that factory. Now we're kind of moving into the actual production of parts. So this facility, we're going to basically have factory one and two. That's the

footprint that we're using. Factory one makes the molds. We ship it to factory two to actually make the parts. So that facility will just be nothing but part making. What's the wheelhouse part that you make? What's the canonical example? Name it all. Name every part. Yeah, name every part. You make parts. Name every part. We got a lot of parts here. Yeah, I mean, basically what we're doing is trying to focus on taking like specialist shop knowledge right like you would go to a shop that's really good at doing like automotive parts or aerospace parts and generalizing that same capability like across the board so our envelope or constraint is kind of like size instead of industry or part so we can't do something like the size of the table obviously but uh we can work up to that you know so for

now we're kind of like parts that are a couple shoe boxes are smaller like what was the thesis for your pitch to have every person who works in manufacturing wear a pair of meta glasses and start collecting data like how would that actually flow through so i started a lot of my earlier career in like self-driving cars right um i worked to google 15 years ago doing this is actually insane um i don't even like thinking about that were you like 15 years old yeah i was i was 10 was wonderkin um and you know a lot of a lot of the kind of like idea behind that is controlling a system well enough, right? Like a human control car. Same principles kind of apply to a factory, right? If you're producing the same thing over and over again, you can dial a factory like Elon does,

right? And it's just like ultra efficient. The Holy Grail is being able to take a factory and kind of adapt to different things, you know, over and over again. And mold making is kind of like that because every part has a unique geometry. So ultimately the shape drives like the construction of every mold. and it's a little bit different, right? But, yeah, ultimately that will... Yeah, I mean, actually pulling data from someone who's working on a line. I mean, they're going to see...

You're going to see someone using a screwdriver one day and then a hammer and then different tools and then pushing buttons and maybe sitting in front of their computer taking notes. Isn't that just like a ton of messy data? Like, how do you actually compress that into something that's actionable?

So the self-driving analogy is very similar to that, right? like being able to actually intuit, like, a mechanical task, right? Or, I don't know, anything like programming a machine or something like that, right? If we have the ability to actually collect the data seamlessly and relatively easily, that's kind of a floodgate, you know, to being able to do a lot of these things. Processing them is a different problem, depending on how general they get. But I do think that that's actually a pretty interesting use case for them. Yeah.

Give us the update on re-industrialization broadly. There's been some headlines on the job side this year that have been not super thrilling to look at. Actually, reduction in manufacturing workforce, but what's actually... Yeah, a lot of energy in the space. What's the vibe like on the ground? The vibe on the ground... Actually, and to be clear, I want to get a sense of what the vibe is, like, legacy manufacturing and then the re-industrialized crowd. No way.

You've got to take that. I can't. I can't do that right now. Wait, take it. Take it right now. He'll call that. He'll call back. That's actually, I mean, I don't know if you're fucking with us.

Yeah, I don't know if you're messing with this, but it's fine. It's crazy. Big if true. Big if true. The, yeah, I mean, the vibe on the ground for this stuff is, it's interesting, right? Because there's a huge need for it. And, like, economic data, you know, numbers go up and down all the time. and people will panic.

Businesses might not actually take on more because CAPEX is ultra-expensive or something, right? At the end of the day, stuff still has to get made, right? And there's a lot of work that we're doing right now with the admin, which maybe we'll talk about someday, but people are very aware of this problem, right? And we need to pump resources into the industrial base in a really meaningful and fast way.

so ultimately there's a huge demand and need even without the tariffs it was still there right and i mean yeah starting this company when i did was a really interesting moment in history and like i was trying to front run that problem and ultimately i didn't really do anything for you know 20 20 at all but yeah i think i think people you know they're actually churning out parts and components are, you know, extremely happy with where things are at, but, you know, overall, this is something that, like, we have the same conversation about over and over and over again, which is, you know, factory jobs and manufacturing jobs. Like, a lot of people don't know if that actually means a specialist, you know, C&C machinist, a cam programmer, like,

somebody just pushing a button on a machine on a line somewhere. And ultimately, you'll kind of see, you know, a collapse or you'll probably see, you know, a little bit of condensing of skilled trade labor as, like, technology flows more into the industrial base, and we get more efficient factories that are cheaper. Ultimately, it's not a zero-sum game, right? It's like Jevin's paradox for stuff. If you can figure out a way to make more, we're going to make more. I feel like there's a ton of energy, like the biggest American dynamism story or the biggest re-industrialization story has oddly been like the X-A-I Colossus, like the Neo-Clouds, what Crusoe is doing. Like, it feels like there's an immense pull from the industry to build big things.

And they're building multi-billion dollar projects in America. It's very much assembling parts from all over the place. We just got to report that Elon's potentially buying a power plant internationally and then just rebuilding it in Memphis or Mississippi. But do you think that that actually translates to going down the stack? Or do you think that all the AI energy will just wind up resulting in just build a big data center but still continue to make the transformers wherever they need to be made or make the networking equipment wherever it needs to be made?

I mean, the chips, like the actual energy. around moving chip production back to America. It's a multi-decade project. Yeah. I think there's like a secret plan here that nobody's really talking about, which you've seen Elon and, like, Jensen hint at this, right, in some of their public stuff that they've said. They're all driving for applying AI to, like, the physical.

Oh, sure. Right? Obviously, we have to play with LLMs for a little bit and, like, cool video and image tools and stuff. But ultimately, I think that, like, you know, we saw, what was that, like a week ago, Elon talking about making transformers now, like, just going straight for the kill. I think that ultimately, and like how Jensen talked about, like, you know, I mean, SpaceX vertically integrated a ton of, a ton of part making internally because they were fed up. But what does that mean for your business? Right.

So ultimately, what it means is that, like, you're going to have a tradeoff where skilled tradespeople are kind of like, you know, the replace. rate is very low. So you might actually see some kind of like conglomeration or agglomeration of these trades kind of, you know, forming like a foundational layer of building blocks where we can make components and parts. And the technology is applied to that. Right. Like that's what I was saying like Jensen, his whole thing is like using tokens for whatever, right? Sure. If we can apply that to the part making layer, right? It's just foundational kind of stuff. And then from there you can build whatever you want. Yeah.

Are there, I remember early on when we were talking one of the interesting use cases for AI in a manufacturing context was you have a part that needs to be injection molded,

you need to build the casting for that, right? And you need to figure out where you drill the holes to dissipate heat, right? And the example you gave me was like if you're molding a bumper for a car,

it's a very odd shape. It's not just like something you can roll out on a sheet or cut stamp or do anything else, right? You need to actually mold it. And does that have any indexing towards, like what goes on the data center? Like, are there parts that are benefiting in the data center context from that type

of like injection molding optimization work? Potentially. I mean, it's going to be a lot of, like, racking and componentry, right? So if there's some kind of weird new, you know, form factor, or what probably not. Yeah, most of the data center just feels like it's like vertical racks

and just bolted together. Very like, I mean, Zucks building data centers intense now. Like they're not going with like bespoke, like beautiful, like, you know, custom shapes or anything like that. I think a lot of it, like whenever you need to form a part of some kind, right, you have to pick what type of methodology you're going to use, like whatever is most cost effective and also meets the requirement. So if it's a bunch of metal stuff, you're probably going to stamp it or, you know,

it'll be made out of sheet metal or C&C or something. and then if it is a plastic thing that you can actually blast out in high volume and it's actually worth it, that's a great problem to solve, right? And what we're doing is actually trying to lower that barrier. And not everything can be a plastic part and people don't want that. But if you actually lower that barrier, then, you know, engineering changes actually. Because now you can actually re-scope what is plastic, what is metal, and consider a different production method for it.

How are the capital markets treating manufacturing businesses, and especially tech-enabled manufacturing businesses today. I'm assuming you're not getting AI multiples, but hopefully there's plenty to go around. Well, that's the thing. There's no comp for this, right? Nobody's actually done this before. So, you know, TBD on that,

but I do think that even like announcing the raise earlier, right, was a huge mistake on LinkedIn, you know, just because I now have gotten probably 100 different emails and texts from, you know, people on the other side of venture. shirt. Yeah. They're just like loans. You know, it's a wellspring of money. There's an ocean of capital on the other side of this if you can prove something

viable. Where's the business? I mean, you said $25 million series A, massive facility that you showed us. How are you tracking with like revenue, headcount, anything else that you can share? Yeah. So, with that kind of progression of going like software to mold making

to part making, we picked up like our first couple big, you know, part production contract. So interestingly enough, it's kind of like SaaS. It's just like, you know, I need a widget for five years. You're in charge of making the mold and the part now.

So we're going to start stacking like multi-year production programs. So we're kind of in the mid-eight figure range for that right now. That's good. Let's give it up for the mid-eight figures. It's an amazing place to be underrated, right? People already, to always talk about

nine figures, being in the mid-eight figure. The next time I'm back here, we'll be will be, we'll be, you know, nine figures. Fantastic. There we go. There we go. And then, yeah, we're kind of... Come back tomorrow. Yeah, we'll be back tomorrow. Yeah, answer that phone call and then... Roughly, uh, 50 people right now. Oh, wow. And it's interesting, too, because that includes people that are, like, in, in the factories themselves, like, ever helping. Yeah, yeah, yeah, yeah. I mean, it's, this is one of the main reasons

we actually wanted to build in Detroit, right? Because, you know, even though I love California, that talent is not here. So we do have to kind of like rely on that initially to build out the footprint there. And then once we actually have... Would you come to California? Yeah, for sure. But is California anywhere near the top of the list?

I feel like Texas, there's other... I mean, what's what's happening in Memphis and in Mississippi, like there's so many other states that are more like, please bring your business here. Yeah, I mean, we're going to have to take a look at statewide, you know, incentive packages and see how that's looking. but that's that's kind of the model right is like is California offer incentives for these

no things are they disincentives yeah disincentives yeah if you come here we'll make your life hell but the weather's great I was so funny I was looking at a I was looking at the weather report yesterday and I was like this is this is why we do what we do 75 today in Malibu 75 and sunny then 78 and sunny then 77 and sunny then 79 and sunny and 79 and then 77 I'm sure it's not that bad in Detroit right now.

Actually, it's probably perfect. Yeah, it's probably perfect for now. Just don't talk about winter. California people don't talk about fall in Detroit. I mean, in Chicago, it's the same thing. Chicago is fantastic summers. Yes, it's the people call Detroit the St. Tropeii of America. It is.

It is. Anyway, congratulations. Thanks so much for coming on the show. Hit that gong. Yeah, hit that as hard as you can. Oh, that was a good hit. You kept it in. That's a true test of strength.

Being close to the gong and still getting that contact. Congratulations. Yeah. Thank you so much for talking about all the progress. Go take that call. We will talk to you soon. And we have Ariana coming on the show in just a few minutes. In the meantime, let me tell you about 8Sleep.

8Sleep.com. Get a pod five, five-year warranty, 30-risk-free trial, free returns, free shipping. What did I do? You did well? I got a 94 last night, seven hours, fifty-six minutes. Oh, a 94, eh? 94. 98.

I was not expecting that out of here. Put that phone down. I beat Jordy Hayes. Great. The great sleep-off. We should run a weekend, like weekend average, see how. Yeah. See how we do.

Sleep off. Anyway, there's new paper that claims LLMs are better at selecting founders than venture capitalists. I don't believe it. V.C. Bench, the first benchmark for predicting for founder success in venture capital. This is from the University of Oxford, and Julia Hornstein's shares, thinking about this, quote, but for Andresen, there is one job that AI will never do as well as living, breathing, living, living, breathing human being his.

Think I'm kidding on an A16Z podcast last week. Andresa opined that being a venture capitalist may be a profession that is quite literally timeless. When the AIs are doing everything else, he continued, that may be one of the last remaining fields that people are still doing. Didn't Blatt say something too? Like, you could just go singularity, you just invest for a living? I agree with the sense that the high-level point, which is that, like, humans will, like, raise capital. machines will probably deploy it. But I do wonder, I would love to,

we should dig into this VC bench thing more because I do think if I just like ran, you could basically run a model against like a deck. Yes. And even, you can even analyze things like how fast a founder responds, right? I would love. People talk about some of these like smaller data points, like founders that respond fast, like generally outperform.

the like outperforming I mean I'm not exactly sure how what the outperformance was but being in the 60th percentile of venture capitalist like does not matter like you still do very poorly like you need to be at the long tail of the power law I'd be fascinated to know brand is like one of the biggest drivers brand super important but also I think I mean you read zero to one the value of secrets like secrets are not in the training data and so there's so many decisions that are made in in a business partnership or backing a founder that are secrets that never make their way onto the internet they never make it way that they never make it into the weights whatsoever maybe they make it into a biography after someone passes away like decades later but so much of

decision in in venture capital and decision making in business is really on the on the backbone of, of, of, uh, of whisper networks and understanding, being able to accurately predict the future. It's not just accurately predicting the future. I feel like accurately predicting the future like broadly like, like AI is not that bad at that. But AI doesn't necessarily know like, like all the different, all the different secrets that are going into a business before. If you're running, let's say you're running a model generally. Yeah. And, and you asked in 20, 17, do you think that bright young people in Silicon Valley are going to want to work in defense tech in seven years? A model might be like, eh, people have like pretty bad taste in their

mouth, like Google engineers are protesting, even having any contracts, right? Yeah. So it just depends on I mean, yeah, the LLMs right now are not designed to be contrarian machines. Like they're very much consensus focused and just getting to a model that thinks independently or can take a contrarian position that isn't just take your conclusion and add a minus sign in front of it. It's like, what if the average Redditor was deploying capital? That's the, that's certainly the bare case for this. At the same time, you know, potentially like the average redditor is deploying capital. And that might actually define the VC industry pretty accurately right now. And so I'm not surprised. Tyler, what's your take?

Okay, so I'm looking at the actual like benchmark. Yes. So basically, so there's 9,000, the data said is 9,000 anonymized founder profiles. And then basically they're either labeled like success or, I guess, like, failure, depending on... Do they return capital or something? Yeah, if they exited, IPOed over 500 million or raised over 500 million. And then, so it has, like, some of the stats are, like, tier 1 VCs are at predict 23%. And then deep seek chat is at 80%.

That's a huge delta. It's kind of like, it seems hard to, like, anonymize these correctly where like you're still encapsulating like what made them a good founder but also still like different enough that a model can't just like be like oh that's just Mark Zuckerberg totally his name totally so like yes he's successful yes it's I think it's the other thing is the issue is like VC is not about identifying the great founders you have to identify them and win allocation at specific moments 100% that and then yeah and then also it's it's identifying the next great founder which is usually

breaking a pattern. I mean, Paul Graham has a famous example where he over-indexed on Mark Zuckerberg met a founder who literally looked like Mark Zuckerberg wrote a check and it was one of the most underperforming adjustments he's ever made. And he was like, yeah, I died just like this guy. He went to Harvard, dropped out of Harvard, looked like Mark. We have Ariana in the Restream waiting room. Let's bring her in. Let's bring her in to TBPN, the Ultradome. How are you doing? What's happening? Hi, Jordy. Hi, John. Great to be here. Great to meet you. Okay, I will start this off by saying I hate mold. I'm one of the number one haters. I had a rental once that had a mold problem. I got terribly sick. And it ended up being this crazy saga where I was like fighting with my landlord because they were saying like there's no mold problem. I was showing them test results that showed there was a mold problem. They were kind of pushing back. I think it's a huge problem. I went down this rabbit hole. I think a bunch of people are

kind of like waking up to how prevalent the issue is, especially in like specific regions and even specific like local communities. So excited that you're building this and excited to hear about it. Thank you. Yeah, there's definitely a cultural awakening happening. You might have even seen HBO shows and Netflix and throughout the Reddit forums, Facebook communities, influencers are on Instagram, et cetera, like it feels like there's this groundswell of education and awareness happening beyond even our company. Yeah. How did you, why did you start the company, personal story? Yeah, super similar to yours, went through that journey with my landlord as well. And prior to starting mold code, didn't have health issues, then experienced debilitating symptoms and actually

developed something called chronic inflammatory response syndrome, which is the more formal name for mold toxicity or mold related illness just in a six and a half month time frame of being in my Miami unit which was quite moldy crazy what what is the I have an easy to tell what the company does at a high level but like how are you how are you kind of like tackling the problem what does the business look like yeah we're tackling this in a few different ways so the vision is to drive the standard of care for mold treatment and mold labs and what we do is offer labs both for health and the environment. So that includes four different tests available on our website today,

specifically a three panel and a 16 panel for the LabCorp LabDraw, and that's based on our 30 years of research in the field. So we have on our founding team the world experts, the folks that not only pioneered the space, but also defined chronic inflammatory response syndrome, came up with the biomarkers, came up with the treatment all in their lifetime, which is an extraordinary effort. So we have that. We also have something called HLA haplotype testing to see if you have a genetic susceptibility to mold-related illness, and that impacts one and four here in the U.S.

And we also have a test, which I frankly love over even a home inspector at times, it's just a quick litmus test, low-cost, use a swipper pad to wipe down your surfaces. So we have that on the lab side. And then on the care side, we have our own in-house providers, and we built a platform to treat patients at scale. And that includes concierge's care in our patient portal, and that includes a distillation of something called the Shoemaker Protocol, which is the only published, peer-reviewed treatment approach in the space. There's hundreds of publications making the mold health connection. There's really just a small body of work that demonstrates treatment and what's effective for patients. So that's what we've done. We've translated that into a virtual care clinic, and we've done a lot of creative things to make that scalable for patients so we can keep the cost low and affordable. What if somebody I guess like what is your recommendation you meet somebody like randomly they hear about your business like what is the first step because I think a lot of people you talked about people having like genetic susceptibility this this is I think important because it means that like a few people could be living in a house and like some people aren't having necessarily any issues or symptoms and like one person could be having a super adverse reaction and that I feel.

like can make can make people like just like not really be aware of the problem because if you have somebody's in college and they have some roommates that are doing fine and they're struggling they can it can end up being confusing but what do you recommend to people where do you recommend people start if they're kind of concerned that they might be dealing with an issue yeah absolutely so we have testing options to meet patients where they are so it really just depends on where they are in that journey and you touched upon a really important topic and that's the canary effect there could be others in your household that start experiencing symptoms before you even experience symptoms. But it really just depends on where the patient is. Have they already made the mold

health connection? Have they conducted a battery of lab test to validate that? If so, those patients are often great to start care. If the patient hasn't even made the mold health connection, but they're sure that there's mold in their home, they can probably skip the dust test and even a visual inspection or having that musty earthy odor is sufficient. So for that type of patient, and we may recommend even the starter panel, which is one of our three lab panels, and that has the highest signal biomarkers based on the research that we have to date, which I'm sure will evolve. We actually have an R&D arm with the largest data set in the world for what we're doing, too.

So we're looking to advance that and really drive the standard of care. Or if a patient doesn't know if they have mold in their home, but they're having the indicative symptoms, which includes brain fog, fatigue, gastrointestinal issues, et cetera. It's normally a multi-system, multi-symptom issue. specific but affecting multiple systems then we would recommend a lab panel such as the starter panel or a complete panel the HLA is help would you guys help cost like clients customers with with like legal because I ended up having to

hire a lawyer when I was going through my issue because my landlord was just like completely just like basically like gaslighting me not not is that something that that you guys would ever help people with or is that outside of kind of the wheelhouse? It's currently outside our scope. In terms of the next evolution of the company, the parent company is actually called the immune co, and our vision is to spin out Lymeco and Long Co next.

Our thesis is that the data set we've collected, which is looking at these microbial insults, mold and biotoxins that lead to innate immune system dysregulation leading to chronic inflammation can actually have the underlying research to help solve some of these other biotoxid-related issues. So we're more so going that direction versus expanding two horizontally within the mold economy, but I think there's definitely

opportunity there for founders. The mold economy. We're here. How are you guys acquiring customers at this point? This feels like the kind of thing that, one, there's a lot of influencers that have dealt with this and that'll talk about it, but it also feels like you could have an organic TikTok or real strategy or something along those lines. Yeah, we're doing a lot of different things.

We actually just launched something called a content farm. We had 7 million views in the last week or week and a half or so just off of that. So we are tapping into TikTok and Instagram and doing some fun things there. We had a Reddit AMA with Dr. Shoemaker, the Messiah of Mold, our founding physician, and that went viral in Reddit. The Messiah of Mold. Yeah. You got some good coinages here, the Mold economy, the Messiah of Mold.

Thanks. I don't have bots in my cap table, though. So maybe that's next. But yeah, so we've done that. We've done, we've had shout outs from patients. At least 15% of our patient base is based on referrals and we haven't even optimized for that yet. I've been on at least 30 podcasts at this point.

So we're tapping into a lot of different ways to reach patients and increase education and awareness, but there already is a large portion of the US that has made the multi-health connections. So it's been pretty streamlined targeting those patients and onboarding them into our care. Would you do anything? Last question I have, I'm curious.

Would you do anything around helping people make decisions on like specifically the moment in time where somebody's either buying a home or considering renting a property because the thing that's like tragic about like kind of the mold crisis if that's what you want to call it is there's a lot of homes that are brand new and they will literally have like a like a significant mold toxicity problem on day one on day one because of the way they were built or you think about it's like houses can take a long time to build. They could have leaks. They could have a bunch of weather, all these different things, and then you move into a house that you think is brand new, and it already is going to, like, make

you sick, which is just tragic. Thank you for touching on that misconception that new builds don't have mold issues. They definitely do. The building materials are left outside in the rain. The energy efficiency standards often have a tradeoff for having more microbial growth in the home. So I think there's a lot of dramatic irony there for home buyers and renters. It's not something we're targeting at a deep level right now I often say that the environmental side is off more intractable than the health care side just because the health care side already has the solution available and every home is unique every rental is unique so we offer free educational guidance but i think that's also another opportunity for a builder in this space if they want to help

essentially renters and owners better navigate um which homes have mold and to avoid that and there's also novel materials coming out like sire wall too to prevent mold in the first place. Awesome. Well, congratulations. Give us the fundraising news. What's the fundraising news? You announced it yesterday, right?

Yeah. Yeah, so we raised earlier this year a bit of a delayed announcement with Cantos and Club Fund as our lead. And, yeah, there we go. How much did you raise? Eight million total. There you go. Let's go. Congratulations.

Great names. Great names. I love to see it. Thanks for helping on the stream. Thanks for jumping on. We'll talk to you soon. Cheers. Bye.

If you want to get away from your moldy old home, head over to wander.com. That's right. Look, go wander with inspiring views, hotel great amenities, dreamy beds, top tier cleaning, and 24-7 concierge service. It's a vacation home, but better folks. Our next guest is Will Hurd from Chaos Industries. He's the chief strategy officer, former CIA, former congressperson. What's going on? How you doing?

And doing fantastic. It's great to be on with y'all. I feel like I'm on with the Shaq and Barkley of tech reporting. So thanks for having me. We love to hear that. Thank you. Let's go. We don't know about sports, but those sound like some big names. Fun fact. I endorsed you for president. I said, everyone else who's running is a lawyer. You're the only computer scientist. Talk to me about your journey, studying computer science through working in the government. I want to get a little bit of backstory before we go into what you're doing now. Sure. You know, it was funny when I was in high school, I had the opportunity to do an internship with a woman who was at Southwest Research Institutes. They are one of the largest private research entities in the country. And I fell in love with

computer science and decided to go to Texas A&M University, gigam Aggies, and, you know, study computer science. And when I first, when I first got to college, I thought I was going to go work at IBM. And then I let's give it up for international business machines. We love international business and machines. There we go. And and and but I had this guy who was a guest lecture in one of my classes, um, that was a former CIA officer. Yeah. And he told the most amazing stories. And I was like, I want to do that. And so when I graduated from college, I joined the CIA, and I was a case officer, so my job was recruit spies and steal secrets. It was the best job.

Yeah, and that's an important distinction, because I think people think about CIA agents, but those are the people that are recruited by the case officers who are doing the puppeteering. That's right. Or you could say the spy master. Yeah, the spy master. That's great. So, yeah, please. It was, look, it was an awesome gig, right?

Two years India, two years in Pakistan. I did some interagency work in New York City around counterproliferation, and then a year and a half in Afghanistan, where I managed all over undercover operations. And in addition to collecting intelligence on our threats to our homelands, I had a brief members of Congress. And I was pretty shocked by the caliber of our elected leaders, so I decided to run for Congress. And that's kind of how that started. That's amazing.

We had Dan Wang on the show. recently. He wrote a book Relentless, and he talks about, I hope I have that right. I might be messing some of that. But the basic thesis of the book is that China has an engineering mindset, and they employ a lot of engineers in the government, whereas America has become the lawyerly society. And we have a lot of lawmakers who are lawyers by training. And I'm wondering about your thoughts on the direction America should be going in terms of empowering folks who can think from first principles or have engineering mindsets to actually go and build things like high speed rail or we got to the moon. And we're seeing a little bit of this with Doge and more engineers

in the White House. What's your overall view on the level of engineering talent in government and the benefits and costs of that? Look, I think that thesis is actually correct. I've spent the last 25 years of my adult life in some form or fashion connect to the national security, whether it was CIA in Congress. I was on the House Intelligence Committee, seeing how things are threats to our nation evolve. I was on the Appropriations Committee making sure that we were spending trillions of dollars in the right way. I've been on the National Security Agency Advisory Board. When I was at Allen & Company, the New York Investment Bank, I was working with technology companies, and here's been one similar throughput to all this. The Chinese government has made it very clear.

The way they're going to surpass the United States of America as the global superpower is by mastering a number of different technologies. It's about 15, 16, 17, everything from AI to hypersonics. And with that as the overarching goal, there is a desire and interest in order to win. You mentioned a space race. Right now, we're rushing to go back to the moon. And the- And guess what the Chinese are going to do? The Chinese are going to do with the space race.

Chinese are going to do it the exact same way we did it in the past, right? Shocker, they're going to steal all of our plans and ideas and replicate them, you know. What? I'm hearing this for the first time. Exactly, exactly. But we're also seeing that in whether it's AI and how they're trying to dominate it. We saw them dominate 5G, the fact that, you know, the 5G infrastructure now is, I think it's like it's over a third. in the world is Chinese tech. In AI, the issue of AI, they're trying to make sure by 2030, every six-year-old is AI conversant.

And we're still having debates here in the United States of America about whether AI should be included. And then I don't even tell you about what's happening on the battlefields, right? We're seeing, they're learning from Ukraine. They're figuring out how to use autonomy in air, on land, and in the sea. They're focused on electronic warfare and how to disrupt our systems.

They're trying to not only build more than what we have. They're also looking forward to making sure that they have mass and autonomy in warfare. And so this is the world in which we're living in. And we need people that actually understand technology to make sure that we can make it happen. A perfect example. No company cannot figure out, they don't know that what their budget is going to be next year or what their plans for sales is next year. But within the federal government,

you know, we don't know necessarily from year to year what the funding is going to be. And for a trillion-dollar enterprise, you can't do that. I think at a minimum you should have a two-year budget and appropriation so that you know with some certainty of where you need to go and not trying to go month-a-month to try to get things done. And I think that's right now one of the issues

that I think a lot of venture capital and venture VC-backed companies are trying to change with DOD. It's slow. It's going. It's moving. It's changing. But we have to have that. We have to bring that mentality to government.

Yeah. Talk about your path to chaos. It's a wild name for a defense tech company, but you're doing some very cool things. How did all this come together? Look, I didn't think Sling and Radars was good. It was on my bingo card, right? But when when I got to meet Dr. Beaumar, one of our co-ceeos and founders of chaos.

He had also previously founded Epirus, John Tennant, another one of our co-ceeos, who I've known for some time. And he's like, look, Will, we're doing something interesting. Why don't you come to L.A. and see our product? And I was blown away, right? First off, unbelievable team, right? These are, you know, engineers that have been, you know,

quoted in Nobel Prizes, right? You have real operators in government, Navy SEALs, folks that worked in Delta Force, CIA, you have operators that have been in the system, and you're working on some of the most cutting edge technology, right? Like, you know, we have figured out a thing called coherent distributed networking. This is taking multiple things talking to each other, and we figured out how to do that to the picosecond. now a picosecond you know for some references one pico second to a second is like a second is to 32,000 years that's really fast and by figuring out how you have this time transfer work we're able to build a more powerful system and a cheaper system and and and so the that you know amazed me and ultimately the vision of of the company right since let's call it the end of the Korean war, right? America and our allies have had air superiority. We, and that air superiority has given our troops the ability to look forward, not above their heads at swarms. It's allowed commerce to whip around the world. And those days are over, right? Because of autonomy on the battlefield.

70% of casualties in warfare right now are from autonomous systems and drones. And the time to detect and do something about it is shrinking. And ultimately what chaos is doing is we're giving the warfighter more time, time to understand what's happening, time to decide what to do, and time to act. And so for me to get back to my roots of national security, it was exciting. And so that's how I ended up, that's how I ended up at chaos. What's the highest priority threat that radar can help mitigate? Is it, is it these autonomous drones these days? It is. Look, there's a lot of conversation right now around Golden Dome. And Golden Dome is something that we need to have.

We need to be able to defend against nuclear-tipped intercontinental ballistic missiles from our adversaries, primarily China, Russia, Iran, and North Korea. And this is a phased approach from detecting it when it launches from their homeland to try to identify it in space. the technology that has evolved and what we're able to do now sensing in space is pretty fantastic and this is an important problem but the reality in my the most recent threat like the threat we're most likely to see is a a subsonic cruise missile or a drone and this can be shot from a plane off the coast of California it could be a van that is

filled with a number of explosive drones, these are the things that we're going to see more likely than a nuclear weapon. And we call these underlayer threats. These are things that are happening under 50,000 feet. And that's something that radars is able to help detect. And one of the things that we're learning. Because of some of the technological advances we've made, we can see big, high and far, and low, small, and near at the same time. And I always remember, remind people, I think pretty much everybody in America watch Maverick, you know, Top Gun 2. Of course. That's a real thing. Flying below the radar is a real thing. And so we're able to see below that, below that level. And that's something that is going to be helpful against this drone

threat, which we've already seen. We've seen what the Ukrainians did in Russia. We've seen what Israel has done in Iran. All of that being able to take out. this this exquisite infrastructure right and and exquisite is actually a a has a unique phrasing or or definition in government that means big-ass expensive system and so so so we're you're able to take these systems out with a system that cost a billion dollars with a drone that cost 20k right so that cost asymmetry is one of the things that that the right And what we're doing in multistatic radar is something that is one of the reasons why we're getting traction and excitement within DOD and the rest of and militaries around the world, to be honest. I don't know if you can say anything here, but are you guys doing anything on the counter narcotics side?

Yeah, we've been seeing some stuff here. I've watched a bunch of YouTube videos on narco subs. Aren't true submarines, but designed to be hard to detect. Look, this is an issue that's actually, you know, close to my heart. When I was in Congress, I represented a district that was 29 counties, two time zones, 825 miles of the border between Texas and Mexico. I had more border than any other member of Congress. And so what the narco trafficantes, the drug traffickers are doing, they're sophisticated.

The narcos in Mexico alone are making roughly, and this is a conservative number, $60 billion. dollars in one year the you the entire u.s. intelligence budget is 60 billion dollars right and so you add up all the all the foundation model lab revenue and you don't get it you don't get there it's bigger than AI yeah no it's it's it's it's wild and so so yes i you know there is it is known um in the end of in the last year i think beginning of this year a border patrol testified in front of the Senate, that there were, you know, something ridiculous, like 41,000 drone sightings in just one sector of the border. So we know the narcos are using the drone warfare against themselves, right? And so, so like against rival, rival gangs, basically. Absolutely. It's like the drone

warfare you're seeing between Ukraine and Russia. They're adapting some of these tactics, techniques, and procedures to that fight. And then we also know, the only way you get 60 billion, million dollars worth of drugs into the United States of America is by you're bringing them in volume. And there is likely some coming via autonomous systems, air and and C. And we've had conversations with authorities in order to help with this project. And so once we learn more, we'll come back and I'll give you a report on some of the things that we're seeing. Can you give me some, describe the shape or the physicality of what you're building? You know, there's been, you look at those like Anderall submarines, they're square now, they look like they just fit in shipping containers.

There's a lot of logistics that you get out of the box. If it fits in a shipping container, you can load it onto a boat, an autonomous boat. You can load it on a truck. There's also stuff that can be mounted on an aircraft or something that's, you know, getting into exquisite territory, exquisite system territory. where do you see the opportunity to deploy new radar systems so one of our products we call vanquish is an expeditionary system so the three of us can can carry that it can fit on the back of an ATV and go into to really is that an invite is that an invite if we're going ATVing let's do it hey you know I think we're just right around the corner from y'all so amazing so next time next

time we'll take you out we'll hang out that'll be great and so that one is It's low form factors, small, that can be deployed in difficult environments. We have another system called Astria, which we were just celebrating. We had a major finance award from the U.S. Air Force about this product that we're deploying in Eglan Air Force Base in Florida. How much was that deal for? We like to ring the gong around here for big numbers. There we go.

Hit it. How much was it? So $2 million. Boom. Congratulations. Seven figure. We got one more gone. We got one more.

Give it to us. The House of Representatives put in 10 million in the defense bill. There we go. Congratulations. There we go. I love to see it. We like to celebrate big contracts around here. What else are you or defense tech companies or the United States government focused on in terms of actual deployment?

There's been a lot of talk about, you know, the Ukraine war. It's been going on for a long time. Very sad. It's unfortunately become a serious business for many companies. There's also Anderall announced a one billion dollar deal with Australia. There's deals going on in Taiwan, obviously in the Middle East. Where else in the geopolitical realm are, should founders be thinking about understanding the nature of the changing warfighting? and the changing battlefield, and then building solutions for?

If you're a defense tech company and you have an operator or aren't operating in Ukraine, that's going to be a problem. If you're investing in companies, if I'm going to be an investor in a defense tech company, my first question is, does your stuff work and has it worked in Ukraine in one of the most difficult environments on the planet right now? And that environment is going to make the first island chain if we get into a conflict with the Chinese look like a,

playground, by the way. And so we've been in Ukraine. We've been tested against the electronic warfare systems of the Russians. Ukrainians have asked us to come back. We're working on being able to do that. We were deployed in the Middle East during the 12-day war. We saw Shahid drones going back and forth between Israel and Iran. So we've been deployed there. Of course, the potential for geopolitical conflict in the Indo-Pacific is an important place where we're having conversations to be able to deploy against there. So we've been active. How quickly can you guys set up in a new in a new region, right? Because like CENTCOM has been quieter, but feels like that could bubble up again at any point. Look, so our system deployed,

you know, once it comes off the truck, we can put it up in about 15 minutes, right? And so doing that is something, but changing the flow within the, within the U.S. government, Look, the value of a company like chaos when we have, you know, basically half a billion dollars in our bank account from all of our, all of our raises is we're able to operate in these long sales cycles or potential changes within the federal government. And so part of us is just understanding where the threat is, making sure we're deploying. And one of the reasons we're kind of, you know, been a little bit more active in the public is that our stuff works.

we've been deployed. And it's operating. We have a year of consistent deployment in really tough places. And so that's one of the things that's important for us. And we're going to go wherever the warfighter needs us. You're building a sensor. There are other companies building effectors. There are companies that want to integrate sensors and effectors. I'm thinking of Palantir. I'm thinking of a lattice from Anderle. Are those out-of-the-box integrations or those partnerships that you're talking to? Obviously, you're super close with a lot of those folks. What does building a product that integrates with everything else because I can't imagine that you're saying, oh, on day one, we want to, you know, we, we don't, we want to be the only defense tech company. No, you're part of the ecosystem.

You're building, you know, one solution. How do you think about integrating with the rest of the ecosystem? Part of this and what the U.S. government needs and wants is completing the kill chain, right? So you can sense and do the effector. And one of the areas, holy grail may be too strong of a word, but. getting to an effector that's cheap, right, or inexpensive is the word, what I should use. And so that integration is important. And so we've already demonstrated the ability to integrate with our friends at Anderra. But Paul and I go way back. When I was in Congress, I was talking

about having a smart wall, right? And I get this phone call from some dude named Paul Marlucky. And I'm like, you know, he's like, I got your smart wall. And that began a relationship and helping those guys. Again, same with Palantir, right? You know, I've seen Palantir for a really long time, you know, even back to my days when I was in the CIA. So our ability to integrate and work together, and the unique thing about some of the technology

that we're working on, this cutting-edge stuff, is it actually can potentially be a force multiplier for other systems as well. And so the name of the game is to be able to integrate and make sure the command that we fit into the command and control infrastructure and ultimately our sensor can make every effector a better so so this is you know exploring those areas as we explore the kill chain and and all of our suite of products making sure that again at the end of the day we see farther right we see farther than everybody

else and we give that warfighter more time and when you have more time then you can decide wait a minute which tool in my toolkit do i want to use in order to in order to deal with this threat. And the more time you have, the better it is. And when you have bombs dropping on your hand or people shooting at you, even if it's five to ten more minutes, that's an eternity. And to be able to give that time back to the warfighter is what we're focused on. What's it like for people that are starting their career on the public sector side, whether it's intelligence or something like down at the border? Do they have a feeling that they have like the tools? are great or are they still or is the sense if you're if you're actually up you know the

one needing to operate and use these tools that they can and should be a lot better I look it depends on which entity you're talking about an organization like the CIA which is which is which is geared towards helping the pointing into the spear whether that's the case offer that's collecting the information or the the person doing a a covert operation there is a belief in that you have entities like in QTEL, which is focused on bringing in new technologies into help the warfighter. You know, when you look at some of the special forces units that are really good at adapting and making sure our boys and girls in those units have what it needs, but it's not it's not uniform. And we always talk about how the procurement

cycles of getting the right thing in the right hands matters. And here's where it's some startups that are VC-based or coming out of Silicon Valley, don't get wrong. Like, we talk about these long cycles, but guess what? The stuff has to work out of the box when it comes to getting in the hands of our military. It has to work every time. And so the training and evaluation matters. And that's what adds to some of that time of getting those equipment. And it has to work within our system.

So I think if you talk to someone on the border or first and early in their career, they may be frustrated with the lack of some of the tools, but some of the problems is that that tool has to be, you know, be tested, and it has to have the cybersecurity defend against the Chinese government. Talk to me about the path to program of record. You have this radar contract, more money coming from the House approval. The typical story that I hear from a lot of defense tech startups is they get a SBIR and then they sprint towards program of record, obviously over the long term

you want, that durability of revenue that you can underwrite a massive manufacturing facility so you can be making this and improving it every year. What does the future of the company look like for you? Well, look, all those things are true and we're far down. We've engaged with all of the branches on this. We're at certain levels of the buying process. You know, it starts with a user wanting your thank. And if you have users wanting your thing, then it makes it a lot easier in order to connect the dots and get the money that you need, right? And so look, I use the annual example.

When they did their first border technology, that was a $10 million appropriation in the budget, right? That led to one of their first program record. And so that's an example of how you get in it. If you have a product that works, if you have a product that is doing something unique and you have a user wanting it, and you've demonstrated that capability in the difficult environments to operate, then guess what? If finding that path to program a record is not as difficult. And again, I'm not trying to act like it's easy.

It's not. But being able to line up all of those things is. something that we that we have experience with yes uh well congratulations thanks so much for this is super fun the chat the chat loves you everyone's going crazy they're saying uh well heard's my dog yes people are big fans thank you for coming back on you have more news or if there's something in the world happening you want to chat about we would love to do that thanks fellas i really appreciate chaw we'll talk to you have a great weekend let me tell you about adquick dot 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 efficiency and those ad buying across the globe we have our second in-person guest welcome to the stream looking good in the suit our second guest thank you thanks so much for charming oh fantastic we love hats here I'm throwing it on we got our second guest introduce yourself for the stream for those who don't know 350 million dollar man Hello, how's it going? I'm Sam, co-founder, CEO of Numeril. Which you know, of course, because... Which you know and love.

We have been partnered with Numeril for a number of months now. Kick us off with the News of the Day. The News of the Day is we raised $35 million on a $350 post. Congratulations. Do the honor. Take this, hit that gong as hard as you. Subtle, but firm. Not too hard.

Coming off Series A from Benchmark as well. Is that earlier than a year? Exactly. We did a Series A from Benchmark earlier this year. And so just following on there, you know, capitalizing all the momentum. Who did the series be? We did it with Mayfield. Mayfield. Classic fun.

Got a classic fun. Got to keep it old school. They've been in the business for generations for decades. We got to update Jordy's ad read for probably the first 50 ad reads. Benchmark Series A. Don't forget, Mayfield's series. Don't forget YC as well. When did you do YC?

We did YC beginning of 2023. Yeah. Okay. I'm actually working with my old boss at Airbnb Gustav. So it was fun to get to work with him again. No way. When did you know you wanted to automate sales tax? Oh, man.

Since the womb, actually. I've been dreaming about it. You know, no, for me, you know, I was similarly, you guys actually running e-commerce businesses. Still own a vitamin gummy shop doing 10 plus million a year. Amazing. And so. What is it about gummies that just sell better than any product on the planet?

People love gummies. They don't want to pick out of it. It's the best e-commerce business because, you know, you get that. repeatability. It's easy to predict. You take two a day. And I think people don't like pills. Americans, we love sweets. We love candies. A little treat. That's great. So you discover the problem through that business. Exactly. So my co-founder there and I of the gummy business, we were just dealing with sales tax compliance for all. We have to register in 40 plus states. And so I was dealing with, you know, local parishes in Louisiana and like on the phone with some women in West

Virginia and I got like an argument about penalty waivers. I'm like this is insane that I'm trying to spend my time on growth and marketing or product and things like that. And instead I was wasting like hours just dealing with, you know, the IRS compared to these states is like working with Apple or like Google or some like amazing company. Like Department of Revenue, Alabama, like give them a call and it's insane. And so I think the thesis was you have, you've had these incumbent tools that have been around for 20 plus years. The most famous one is this one called Avalara, and the problem with them is,

it's a kind of this very narrow tool, and you have to spend hours and month dealing with them, so you get physical mail from the states, and then you have to figure out what to do with it, how to call the state, and they won't even pick up, they're owned by private equity, so they don't even pick up the phone when you have an issue.

So for us, it's, you know, we're not really trying, you know, it's software, but we really, if you're a customer, we really want it to feel like service more than software, because, you know, with AI, since we invented AGI, you know, we're using, we're using, sales tax AGI.S. Tax AGI.G. Right. And so, no, well, yeah, with AI, you can deliver what feels more like a service. But, you know, fundamentally it's still a software business.

What was the fundraising process? Like, I want to know, just take the temperature of, like, you just went out and did a fundraising round. You talked to a lot of VCs. Where people trying to pull an AI story out of you, was it vibe round, Excel round? Like, what is the focus of Silicon Valley VCs right now since AI has become a tool that pretty which every company's pulling off the shelf. Just what was the vibe on the road show? That's a good question. I think, yeah, you have to,

AI has to be front and center. Any friend that's raising and not putting their front and center, it's hard to actually get a meeting or be taken seriously. My sense is, you know, it's starting to get maybe a little too far over the edge. I've seen some insane markups. Like we intentionally actually took like a lower markup than we could have just because, you know, I just tend to run these things. a little bit more incrementally and don't want to, you know, over-extend yourself.

I've seen, I've been in startups where we went a little too crazy and, like, you know, it's easy to kind of just, you have too much money, you spend too much. So, but, but, yeah, it's if you write, if you have, if you have, if you have some good AI logos of companies you're working with, we've seen, I've been seeing some like 100x, 200x, you know, multiples out there. It feels a little bit back. Just because they have, they're working with like a lab or something like that. Yeah, exactly.

If you can close a lab and a couple labs, like, you. you're going crazy. I mean, in some cases, it's worth it, right? Look at Mercor, right? They're growing so fast, and that's their core business. That makes sense with Mercor. But, yeah, if you're more of like a software tool that could be used by any company, not directly indexed revenue-wise to the lab revenue, which is growing very fast. Yeah. Yeah, it's a little bit of a just like, it's a good, you know, they're obviously cutting-edge businesses. They probably select the best software. But it's definitely starting to get a little frothy, you know.

I'm not putting a ton of my own money into... Yeah, do you think that AI is changing any of, like, the fundamental physics of like the the business models like we we we we we cover the story with notion uh notion so they normally have 90% gross margins now they have 80% gross margins but they're growing and so it's mathing out i feel like with as we enter like the sass apocalypse where the legacy SaaS players aren't moving fast enough and we see a new generation of of companies come up i'm wondering how much people should update on the actual like foundational like is there something changing about the monopoly thesis, how companies accrue value, the margins, what the actual

long-term cash flows will be. It feels very much like it's a big opportunity to build something new, get new customers on board, grow a big business, but it doesn't necessarily change the actual structure of the business. But what's your take? Yeah, I think each business is different. So, yeah, with the gross margins, even we seem some compression because we're spending so much running agents. And yeah, we have, you know, if you have a state portal. So many people aren't willing to admit that. No, you ask them, they're like, it's like, it's actually a good thing if you're

spending a lot on inference because it means you're actually using, you're not just saying you're. Well, a lot of people were also on credits for a long time. And so they were doing, or the credits were really discounted, but now we're in more of a stable inference costs, $4 per million tokens if you're doing something serious. Yeah, our AWS bill, I think, doubled, like, relative to our revenue. So it's like our gross margins have actually compressed, but I think our product has gotten a lot better. And, you know, the hope is eventually over time, you know, the inference costs go down and the models get more efficient.

And so I think to me, that's like a little neurotic to worry so. I mean, it depends who you are, but to me that that's like a neurotic thing to worry about, like the short-term gross margins. I think for me, at least for my business, where I'm excited, is, you know, the services economy is like at least an order of magnitude larger than the software economy. and so a lot of what we're starting to replace now is not really there's these legacy software providers but again as you bigger companies usually they don't even work directly with them they'll hire an accounting firm to manage Avalara vertex of these old school solutions they're like this the software's so rough I don't even want to let's hire somebody to yeah I've literally been in that position trying to buy the competitor product years ago and and being like it's expensive and also I don't

have the manpower to do it so we'll just continue to just just like file whenever we can. And it's very rough. It's all done in a spreadsheet. Yeah, I mean, we are the official, I have the real one? Yes. I got the tin here with the official tax provider.

Of Lucy. Of Lucy. My D2C nicotine program. They're gonna get it right. We love Lucy. We are powered by a numeral. It's fantastic. Yeah, and so it's interesting, right?

So for example, we filed taxes for someone this week in Tanzania, in Kenya, and like, we're the only software provider on Earth that's, I think, ever filed taxes in Tanzania. And so, and so I think, let's go, Tanzania. Hit the gong for Tanzania. There we go.

You can send that to the office there. Yeah. Yeah, what, you guys are scaling quickly. And it's a category that's like not, it's an interesting one because it feels like investing in e-commerce infrastructure was at its hottest, like maybe like four or five years ago. Yeah. And so, do you like that it's like a bit overlooked as a category? And like, where is most of the growth is the most of the growth coming from like online retail companies or SaaS? And great question. Yeah, so we started the business fully focused on e-commerce. That's where I was coming from. And I'd worked at software businesses before. I was a PM at Airbnb. But we started just in e-com. And I think, yeah, that market's been really dry for, you know, people servicing the e-commerce space, like the, the software providing for that space because the tam is somewhat limited you know e-commerce growth

is something like 5% year over year this is a problem eventually yeah you saturated and then you're growing at 5% and no VCs are giving you call. Most of the good businesses I've seen being built are either bootstrapped or more like a venture light model which makes sense just because the tam is going to be more cap more like an app in the shopify app store plug and play do something minimal scale it but it's not really going to become like a public company exactly I think we were the first company benchmark funded that's actually been servicing, you know, e-commerce merchants, right? But again... And was that...

Do you think they had a sense of, like, this numeral, while it looks like a Shopify plugin or just like an e-commerce infrastructure tool, it's really going to be eating into the services revenue from these legacy tax, like, prep... I mean, it's so crazy. With Lucy, like, we did not rip out and replace to get on numeral. We went from nothing.

And this was like a seven, eight-year-old business at the time. Yeah. So, yeah, so like just to benchmark, like our partner, Chathan, you know, when he was early in his career, he actually did an analysis on Avalara. You know, they were started in 2006, I think, Avalara. And so they were just going public or something. He's like, wow, this business should be amazing. He checked back 10 years later and it was just being operated really poorly.

That's why they had, you know, a private equity takeover because it wasn't, they weren't operating well. So I think he had seen this category and was like waiting further to be someone to go in challenge. You know, we started fully e-commerce. Now, you know, it's a closer split. we have a lot more SaaS and e-commerce businesses, especially with our international service, with the only player in the space that can actually go and register and file in 60-plus countries.

You feel like if you're making some consumer AI app and you're selling to users that are in Tanzania or something? Exactly, right. So Lucy's probably not selling a lot in Tanzania, but we have like these hot AI companies that, like the one in Tanzania, they're like a year old, but they have $100 million of ARR. And so you have customers around the world.

And, you know, your striped dashboard is blowing up with all these warnings and people don't, you know, they get freaked out. But they don't want to do so they just kind of kick the can for a long time. But, you know, sometimes you can, you know, you do that and not get in trouble. But some places will just like, like, you know, you don't want to mess with those countries, right? You don't want to mess with Mexico. Like, the authorities there. Yeah, it is a pretty new phenomenon that, I mean, there were a lot of apps in the last boom of, you know, software that might be internationally, distributed because of AWS and because of Google on day one, but they weren't necessarily

taking money from individuals in that country on day one. It's more like there's ads that are being shown there, but the companies that are paying are in the United States or they're doing something else and they're just like, oh, we'll monetize it later. No, these companies that are doing new AI products or something like they're putting the credit card down in that local country on day one and then you need to actually process that, right? This round, this round got this round got leaked by Arfaraq, right? Oh, yes, they did. What was that, what do you think?

Who do you think Arfaraq is? What was the impact? What was the impact from that? Do that help? Was it helpful at all? You know, hard to know. I mean, the round had already closed by then, so it didn't really matter, but, you know, it was a little bit low on our revenue.

So I did retweets. Oh, yeah, you were annoyed because they clocked the revenue. They clocked the revenue. Not like crazy amount low, but, you know, definitely seven figures low. And so, give me some credit. Yeah, I was like, come on, like, we have the most ARR of all the sales tax startups in the space and the most customers, the most money raised. And so, you know, you always want to be a little bit cocky there

because it's hard to know. Yeah, what are you, like, projecting out a few years, like, what do you think happens to the category? Can the category support a number of players? Do you think some of these venture-backed competitors will just end up being, like, a decent little business? Good question. Or do you think, I mean, there's a bunch of, you know, massive, massive, you know, tax prep, you know, big, big consulting places. So maybe it can support multiple billion-dollar players. So when you look in the public markets, right, you have companies like Avalara who is taking, you know, they're going to go public again next year. People are saying like 1.1, 1.2 billion of revenue.

Thompson Reuters has a subdivision doing like indirect tax. Yeah, called one source. Secretly a billion dollars of revenue. supposedly. And like nobody, it's like this thing nobody's looked at. You know, it's like a plug-in to SAP and Oracle. And so the tool looks insane, like it's built in the 1970s. But, you know, if you're some big, if your IBM or whoever, you might be using it.

And there's another one Vertex, Inc. found in the 1970s, another public company, sales tax software. You know, it's, but companies like DoorDash still rely on Vertex and they rack physical service in San Francisco to get the right to get the right. I think they updated with like a CD-ROM. Whoa. So are you, how much time are you spending, how much time are you spending, like, basically doing these, like, much bigger, high-level partnership stuff where you're going to these sort of scaled software companies and saying, like, look, the way you're doing this is insane,

it's going to take time to switch over, but, like, there's a better way. Is that where you're spending a lot of time? Yeah, so, you know, we started more in, like, the kind of SMB, e-commerce space, and so that's where a lot of our early traction's been. Now we're starting have conversations and starting to sign customers with hundreds of millions of revenue work with a customer of you guys ate sleep right so that's another unicorn and so we're starting to work with bigger and bigger businesses still really focused though on software you know kind of startups and and direct-to-consumer businesses we haven't gone and started talking to with legacy businesses a lot that's you

know working with accounting firms and things like and like the big four that you know this just takes a long time and we've obviously had some of those conversations but it's just not our focus. What's the latter for signing bigger and bigger customers? Do you have a rule of thumb, like never jump more than one order of magnitude? Like if you have a company that's doing 100 million in sales, you feel like, yeah, we could probably do a company with 800 million in sales, but if a company with 8 billion showed up tomorrow, that's going to be after we signed the $800 million contract.

What do you think? Yeah, good question. I think, yeah, order of magnitude seems high. You try to go a little bit conservatively, right? Because we are running their tax engine. and, you know, you want to make sure that doesn't go down and lots of money on the line. So we try to have it, yeah, like, like 4X, roughly.

Or you scale up for, or you scale up with folks. We haven't, we've never had like an issue with outages and stuff. I think, you know, in 2025, it's not that hard to scale your servers and these type of things. But, you know, we do a lot of hype marketing, but we're still kind of conservative with the way we run our product just because it is, you know, a compliance product. and so we'll make a lot of noise,

but when you actually comes down to the details, you gotta be pretty uptight, you know, is it tricky to use AI in a compliance context? Like, what kind of guardrails do you have to put on the LLMs that are doing work? I can imagine a bunch of places where it's fine to have non-deterministic, you know, just data ingest or something where there's a human

in a loop, but where have you shied away from putting it in place? Yeah, where we think about it is like, there's still usually a human that does the last mile. So I would give you a couple examples. One is when we read for clients, and you get physical mail from every state and country and you'll get like a stack every month

and it's exhausting to have to read through all this physical mail. So we get it, we scan it, and we have AI, take a first pass reading it. And most of the mail, you know, it's like, this is like... If you can take physical mail that was going to somebody, like the amount that I would... I actually should find a personal service. Like, I never want to get mail like at my house, like ever.

I used a service years ago, where you send all your mail there, they scan it, and then you get like an email inbox, basically. Obviously, that's way better now with modern AI, and I can imagine that be part of the workflow. Well, it's like, you get a letter from California or somewhere. It's like, oh, sometimes you get a little, like, I get a little, you know, like Joel. Like, Joel, of stress, right? Cortisol, that's for what I was looking for. You read my mind, yeah.

You get a little cortisol from the IRS or whoever it is. But, you know, half the time it's like, we got one last week for a client who's, you know, breast pumps in California. a foreign yard tax exempt for the next three months and like Lucy probably got this letter and so you got the stress for nothing but then sometimes it is maybe a letter hey there's a lien on your business there's a big issue here and so anything that like requires human intervention we have like a mail team of six people that literally will go they open up the they'll read they'll read the letters and they'll go and like actually handle the last mile compliance similarly on the tax research side

you know how do you know that juno a lot making this up juno alaska has a four cent soda can tax law often it gets passed in some city council notes. And so the way these legacy providers do it is they have hundreds of lawyers and researchers just reading this stuff by hand. And like, we asked, we hired some, we have a couple of lawyers on staff. We asked them, like, how would you figure out the taxability of, I think it was like bagels in California? And he's like, going on the California, it was like Control F, bread, bagel, like, it didn't

work. Okay, baked goods finally worked. And so this is like such an obvious thing where AI is able to go and do that first pass of like, we put in bagels a category. Okay, it finds all the baked goods. goods. It makes the best first guess. But any single piece of tax research, it's not even that we have like one human lawyer look after. We actually have to have two legal opinions on any single fact. But it really makes them like 100x lawyers. So all the busy work they were

wasting their time doing, they don't have to spend time. And then the things that are really complex that requires going through, you know, huge adjudications and things like this, they can really use their legal minds for. Are you writing scrapers yet to scrape every local law that's passed and at least pipeline that in to say, hey, maybe we still need a lawyer to to look at this, but we noticed that something changed. Exactly. That's 100% what we're doing, right? So, yeah, we have a tax research group, and, yeah, that's a lot of the fun work is getting done.

This is the real AI research that we need. This is important. How are you thinking of scaling headcount? This feels like a business where, obviously, you probably will grow the team quite a bit post-B, but it feels like the kind of business that you should be, like, you know, not, like, running in the red for 12 years and a business. row, like, you can probably start to make, like, real profit. Yeah, hopefully. Yeah. So, I mean, even, like, our raise, so I kept 10% dilution, I, like, I could have raised, like, a bigger,

louder round, but I think for us, you know, I've seen excess before and, like, excess hiring, so, you know, we don't have, so we started the year 25 folks. We're up to 65. Um, but, you know, but we, but even that, like, everyone, it always feels like tight, you know, there's no office manager and nothing like that. I was plunging the toys yesterday. We'll, we'll, post some footage, someone got out of me. So I think, yeah, so there we go. So I think, no, the point is, though, I think we always try to run these things

as lean as possible. I think still the hardest area to scale exponentially or, like, that tends to scale more linearly is your go-to-market sales team. It's just, you know, we have some self-service stuff we're building out, but I just think in general, AI has not replaced salespeople. It's still a very human endeavor,

and so that's probably we, and that, and obviously. So AI can eat steak, play a round of golf, Not yet. You're going to say that way. We're getting there. You should cap off the show with us. Let's check the timeline and see if there's any breaking news before the weekend.

Usually these companies wait until the second they see we go off the air. And then they publish a bunch of news. The big news that we haven't covered yet is Nvidia did one of those zombie acquisitions. They bought this company in Fabrica. And they paid $900 million. They hired the CEO. And they licensed the startup technology. the CEO is Roshan Sankar, the guy behind Broadcoms, AVGO, Tomahawk switches.

So, NVIDIA is now getting into like the app-li-gamehires. Some clear chat. Yes, yes, yes. Invidios isn't just buying talent. They're locking down the AI data centers networking backbone. Well, congrats. Dylan Patel is an early investor in Africa. Somebody should make an AI text blocker for your browser so that it detects, like, chat-GBT-GBT-GBT-GBT-

It does feel like it's pretty detectable at this point. Yeah. For some reason, it's not. What else is going on in here? Oh, J.P. Morgan put out a report on NVIDIA. Invidia's investment in Intel. NVIDIA put $5 billion in, marked up the U.S. taxpayer, marked up Leopold Oshon-Brenner. Let's hear it.

Let's hear it for Leopold and the U.S. taxpayer. Fantastic. But J.P. Morgan's analysis said more value is seen accruing to NVIDIA in the Intel collaboration, there's still no commitment to Intel Foundry. So Intel's still going around trying to get enough of those customers to justify their next big fab. NVIDIA is going to benefit a bunch from this. So even though, it's just interesting because there was one framing of this story, which was like NVIDIA is making this investment as like an olive branch. They just have 60 billion sitting around. They're going to throw $5 billion into Intel, maybe just because they think the stock will go up,

or maybe they'll basically give $5 billion away as an olive branch to support this American chip manufacturer, and then they'll trade something, and they'll be able to sell into China or something like that. Well, now, J.B. Morgan's here saying that, yeah, Nvidia is actually going to get a lot of value out of this deal, which is kind of... Yeah, or they just, right day after the announcement,

they just nuke the chart and the market sell $5 billion. We've got a post here from Nat Eliasson, who I saw in the chat earlier. I don't know how they saw here. It's just the reason you can so easily pull all-nighters in college is because you're supposed to be having kids at that age, 25,000 likes. That is a true, true gigabanger.

Gigabanger. You don't see that very often. Tyler, do you think that DeepMind solved Navier Stokes? You see this post from Google DeepMine. We're announcing a major advance in the study of fluid dynamics with AI, and then they put the water emoji, in a joint paper with researchers from Brown University, NYU, and Stanford. They're announcing an announcement.

They haven't actually announced this. What do you think? Are you familiar with this problem? Can you explain the significance of Navier Stokes? Because Sam here says, don't tell me they figured out Navier Stokes. I don't know that I can do a really good explanation, but I did see in the comments, it's like, I'm pretty, like, 90% sure it's not, it's like some random math problem that's like, not that interesting.

Well, still, let's hear it for random math problems. We want to solve all the math problems. Come on, let's do it. We also have some news in the chat. Netflix is trying to buy Warner Brothers Discovery. We thought Ellison was bidding. No way. Netflix might be trying to bid.

This is from Puck News. Thank you to Ragov for sharing that news. Hopefully you guys don't figure out that we'll read basically anything you put in the chat because you could put some crazy fake news in there if you wanted to. We talked about it. Paramounts up another almost 6% today. 40% over the last month. We have one last ad read getbezzle.com.

Your bezel concierge is available now to source you. Any watch on the planet? Seriously, any watch? You got a series B. You got to pick something up at Get Bezal. Yeah, you deserve a little true. And we'll close with this question. How long is this guy messaged me to, quote,

hop on a quick call. How long are those supposed to typically last? Somebody ask you for a quick call. How long do you think that lasts? I mean, it's supposed to be 15, but it always ends up being 30, I think. 15 to 30. Okay, well, chat GPT says, a quick call can last anywhere from five minutes to two to four months.

Some quick calls have even been going on since the early 1900s with ancestors continuing the conversation to loop in key stakeholders and ensure cross-generational synergies with a warm handoff. But this would be a lot easier to explain over voice mode. Do you have five minutes to hop on a call? This is clearly Photoshop, but absolutely hilarious. I wasn't familiar. has access to, is the chat GPT5 instant a real feature? I don't have that. Mine just says

chat GPT5 usually when I go to chat. Auto. He's getting access to this seems like some new feature. I don't know if it's all Photoshop, but it's clearly a very clever Photoshop. You guys should hop on a quick call with the weekend. With the weekend. Thank you for hanging out with us. Thank you for joining Sam. It's been an honor to have you. And the progress is tremendous. and everybody in the chat, everybody listening. Have a fantastic weekend. Thank you for tuning in. Thank you.