🔴 CODE RED 🔴, AWS CEO Joins, Tae Kim Tells All | Matt Garman, Tae Kim, Tarek Mansour, Matt Mullenweg, Jason Fried

28:18I believe so. And even though they've been doing it for this long, it's not like Sundar's going out there and saying, guys, we're actually going to do it. We're going to do an extra 100 billion next year because we're automating science. Okay. and we're doing this new consumer electronic device. Yeah, no, no, it is crazy. Let's continue. First, let me tell you about Privy. Privy makes it easy to build on crypto. Rail securely spin up white-label wallets, sign transactions, integrate on-chain infrastructure all through one simple API. I have a plan. And this comes from the chat, of course. If Sam Altman really wants to set the record straight, everyone's saying, Code Red, oh, Code Red, it's so bad.

28:55He needs to come out with a statement. we're going to Baja blast Gemini out of the app store. If he says our plan is to Baja blast Gemini and Anthropic into the minor leagues of AI research, I think he just wins completely. What do you think, Tom? I think people are underestimating the possibility that code red, it was actually red was past tense of read. They're talking about the code that was red by the model. Oh, yes, yes, yes. Oh, what was the code read in this scenario? Yes, it might have been RE80. They were talking about the next agent, Mark, that was doing code. Yes, yes, yes. I have read the code, and we're ready for the next pre-training run.

29:37I listened to Mark Chen on Ashley Vance's Core Memory podcast. It's very good. You should go listen. Also, Ashley has a new YouTube channel for Core Memory podcasts. So if you want to find it, head over there. And it was interesting. Mark Chen, I really like the way he runs that organization. I liked a lot of things he had to say. He had some funny, funny takes, some funny anecdotes, basically just saying, you know, he's extremely competitive. He doesn't want to lose. He's, he's, he's, you know, going all out right now. And that one of the ways he's dealt with the talent wars is to just go to everyone on his team and say, Hey, uh, I'm not going to match dollar for dollar with meta.

30:19Like if you want to make 10 times as much money, yeah, you're free to leave. Like you can just go. but we are on a mission here. We're a team. And we think what we're building is so big that in the long term, we will be better. We will be bigger. And he also clarified, interestingly, that although there was a big raid and a lot of people from OpenAI did go to meta, he was saying like, there's been, he was basically like, there's a lot of poaching that's happened from OpenAI generally. Like whenever someone starts a new lab, they always go to OpenAI. They're like, we need at least one OpenAI guy to know how they do it.

30:53Right? It makes a lot of sense. He also said he didn't lose a single direct report. I don't know exactly how many direct reports he has, but he was saying that he didn't lose a single direct report. So maybe that's like, maybe he's trying to say, okay, there were people that were - His loyal lieutenants. Yeah, his lieutenant stuck around. It was sort of interesting. But he did say that he also, he sort of echoed Shalto and said that he believes that pre-training, there's still low-hanging fruit there, that OpenAI will be doing new pre-training runs, that they have seen that scaling is holding, that there's no plateau.

31:29He also said they have models internally that outperform Gemini on benchmarks. Yes. And obviously he caveated that by saying benchmarks aren't the only thing that matter. So I do think it's worth sharing some more. Let's actually play this clip from Ashley Vance here. It says, OpenAI has seen Gemini 3 and is both moved and not. We sat down with OpenAI's research chief, Mark Chen, 90. Yeah, so to speak to Gemini 3 specifically, it's a pretty good model. And I think one thing we do is try to build consensus. The benchmarks only tell you so much. And just looking purely at the benchmarks, we actually felt quite confident.

32:17We have models internally that perform at the level of Gemini 3, and we're pretty confident that we will release them soon and we can release successor models that are even better. But yeah, again, kind of the benchmarks only tell you so much. and I think everyone probes the models in their own way. There is this math problem I like to give the models. This is funny. I think so far none of them has quite cracked it, even the thinking models. Can you just tell us? So yeah, I'll wait for that. Is this like a secret math problem? Oh, no, no, no. Well, if I announce it here, maybe it gets trained on.

32:53It's going to get so saturated. To speak to Gemini theory specifically, it's a pretty good model. And I think... I think this is looping. One thing we do. Yeah, I loop. Having a secret math problem that you give every model to assess it is pretty elite. I keep reflecting on like, so let's read what Prince is saying here. So new interview with Mark Chen from OpenAI. Ashley Vance, the interviewer, has apparently been spending a lot of time at OpenAI, including sitting in on meetings. He seems to be writing a book. and he seems to think that OpenAI has made some huge advance in pre-training. Pre-training seems like this area where it seems like you've figured something out.

33:35You're excited about it. You think this is going to be a major advance. Mark doesn't spill the beans, though. He says, we think there's a lot of room in pre-training. A lot of people say scaling dead is dead. We don't think so at all. Big question about what that means. Is that scaling RL? Is that scaling dollars in? Is it, oh, yeah, if you invest$100 trillion, you can give it one more IQ point. It's like, yeah, that would be an example of like scaling, holding, but like no one's going to make that trade off. No one is going to be like, yeah, I'm down. Totally. Spend the$100 trillion. Okay. So what Sam said in the internal Slack memo.

34:10Oh, it was a Slack memo. Yeah. Okay. Because he was directing more employees to focus on improving features of ChatGPT, such as personalizing the chatbot for more than 800 million people. And again, we've seen them like launch more functionality around this. I think the theory is that this could be a very, like, make the product really, really sticky. Whether or not that's true generally is still unclear. It's certainly people have been very loyal to 4.0. Altman also said, this is in the information piece. Did he mention Baja Blast? He hasn't specifically said Baja Blast, but I think he's kind of alluding to it.

34:52He's warming up to talking about Baja Blast. Other key priorities covered by the code red include image gen, the image generating AI that allows users to create a variety of photos. You had included in your newsletter last week that you've been going over to Gemini specifically for Nano Banana. Yes. So I wonder if this is a broader trend. Does this actually matter? I think it does. I think that the image generation functionality, like fundamentally what LLMs are doing, what these chatbots are doing is they're basically instantiating full web pages. They should be able to instantiate anything that you could possibly land on, whether it's a video, an image, a blog post with images embedded, an audio format.

35:39Like it should be able to like not just understand everything and give you the answer, but it should be able to contextualize that answer in any format. And so I do think being able to generate images at the top shelf, top tier way, the big question we were talking to Tyler was, should they say, Hey, we're just going to use Nana banana, which is like a crazy thing. But you know, there is a world where they say like, Hey, yeah, like we're not going to focus on that. We're going to, we're, we're actually going to just bend in Nana banana, but we are going to be the front door, the aggregator, and we're just going to be the actual - Use runway in the background, right?

36:17Yeah, yeah, yeah. Hand it off to a different team, potentially. I don't know. It seems like that's probably a little bit too close to home. But Ben Thompson has had this claim for a while that potentially OpenAI has a strong hold on the consumer market to the point where if they swapped out the underlying model, they would still accrue tons of the value because people don't really know what model is which. I think the average user doesn't do it. But first, Tyler has a... Yeah, I mean, I think that especially makes sense in the context of images and video because they're just so expensive. I think a Nano Banana Pro image is like, I think it's like 10 cents.

36:55No way. Okay, that might be per like 1 ,000 or something. But it's still, they're really expensive. Videos are even more expensive. Videos are like really, really expensive. So I think it makes more sense in that scenario because you would imagine that it's just so expensive to vend it yourself. It's like you're spending so much resources on that. We have to look at this. I believe this is Nano Banana. Let me see if I can find this. This Nano Banana Pro image. Let me see if we can pull this up. It's from John Gregorchuk. It says, architects are cooked. AI is coming for you. prepare accordingly.

37:36Have you seen this, Jordy? I did see that. You did see this one? Yeah. Did you look at the image closely? Uh, no. Okay. So is it, is it, it's one of the funniest images I've ever seen. So basically this image, it's like a, it has a walkway with like a 40 foot drop to the ground. I mean, it's not quite that bad, but it's, it's close. Yeah. I just, I just, I didn't I don't buy the theory that architects are cooked just because you can generate like a floor plan or designs for a home just because the actual process is you're dealing with a city basically right and you're trying to get things permitted it's not like the problem is just making pretty designs right it's the classic let me see.

38:26I'm trying to put this in the, in the chat. It's the classic, like, you know, is the radiologist's job just to look at images and detect cancer? No, it's way more than that. Okay. So this is the image. And I was actually crying laughing because the, the, the, the, the, the tagline is architects are cooked. AI is coming for you. Prepare accordingly. And you see this, and it's like this AI generated image and it looks like remarkable. It looks like a floor plan. It looks like a floor plan. It looks amazing. Like it looks like, okay, yeah, that's like all the lines are straight. We used to be in the era of like any text would be typoed and there would just be crazy lines everywhere.

39:06But you zoom in and it's like one of the funniest layouts ever because you realize that it's just one massive room with like three or four. Okay, so first off, okay, So you come in through the two-car garage. Then there's a powder room. The mudroom. So first off, there's this mudroom and laundry with two bathtubs in it. Scroll up to the right. Okay. Yeah, right there. So why do you have two bathtubs next to your coat closet? In the mudroom. In the mudroom. And also, like, you can't go. Normally, you come out of the garage, you go straight into the mudroom. But here, you have to go into the main area, which is the gallery hall.

39:45And then you go from there into there. And so scroll to the left a little bit so we can see the floor. What is the coat? What is the coat bathroom? There's a powder room, and then there's the mask. The coat bathroom, two toilets. Why is there two toilets next to each other? Remember, we were touring that facility, and it had two bathrooms right next to each other with no line next to it? Yeah, yeah, yeah. We were in the office, in the crazy office that had the machine. One of the bathrooms just had, it was meant to be a private bathroom, and it just had two toilets there. We were like, why the two toilets?

40:20Yeah, so it's like, so you come in through your main foyer. Then there's a master bathroom. Then there's a coat bathroom with two more toilets. And then there's a huge walk-in closet, which isn't even directly attached to anything else. So you have to like go through this corridor to get to the rest. And so this master suite has three toilets. But then it gets better, dude. It gets better. So go over to the top right-hand side. So look at bedroom number two. It's just like off the center. Then bedroom number three is there. Then there's a Jack and Jill bath. Then scroll down. Wait, three sinks?

40:59Three sinks, no toilets. And then there's another bathroom. And then there's a third bathroom with five sinks. You might not like it, John, but this is this is this is architecture at its best you have you have five you have five five sinks next to your two bedrooms which and then also bedroom two doesn't have a doesn't have anything it just connects it it opens into the gourmet kitchen but then if you scroll down if you scroll down you can see that there's like this huge walk guest suite what is the huge walk guest suite and then you have this like massive dining room that just makes no sense and then down at the bottom to kick it off uh there's uh of course like the great room that's directly tied into the kitchen with there's just the most open floor plan you can possibly imagine and then if you scroll down you'll see that there's like just these windows that like like all of a sudden Trevor in the chat says bathroom scaling laws like why why are why are all of a sudden the doors like vertical instead of this is supposed to be a top-down image and Now I'm looking at these doors and they're like presented.

42:12What did the comments say? Did the comments say like, hey, buddy, why, why'd you put, you know, three sinks in that one bathroom? Well, everyone, everyone gets that it's a joke. Um, oh, it was meant to be, it was meant to be a joke. Yeah. Yeah. Yeah. Yeah. Uh, this, this John guy like totally thinks it's so funny and, uh, and he's just like joking around. And so everyone's just like nightmare fuel. Like this is crazy. And, uh, John's making the same joke. It's super convenient off the open floor plan. no kitchen toilet like you know people like and then people just joking about all the different stuff uh and uh i don't know i mean you know is is is ai going to help with you know architectural design of course uh is it is nano banana going to randomly one shot like the perfect floor plan no also no uh but you know of course there's there's stuff that's that's the funniest image it's so funny so funny uh anyway i was i was actually dying laughing at this thing um um okay back back to the code red dance uh automate compliance and security ai that powers everything from evidence collection to in continuous and continuous monitoring to security reviews and vendor risk uh dd das is adding fuel to the fire he says this is why open ai is in code red in the two weeks since the Gemini launch, ChatGPT unique daily active users, a seven-day average are down 6%.

43:36He is sharing, to be clear, web traffic data. These traffic sources are so rough. I just feel like people use apps. Like the web traffic is probably a good proxy. It's probably a decent proxy. But even then, I don't know how high intent those users are. because it's like, do you think you're being tracked by SimilarWeb that effectively? Like, I would hope that I don't have that much spyware on my Chrome browser that knows exactly where I am. Maybe it does, but I would think that, you know, OpenAI and Gemini and Google would be like, yeah, we're not letting you put a pixel on our site. Do you know, do...

44:20We should have someone from SimilarWeb on the show explain it to us. Tell us your sources. Yeah, how do you actually... Tell us everything. How do you actually calculate all this stuff? because, I mean, you could just poll people. You could just ask a million people, hey, what are you using? I don't think that's how this works. But the app store. I wonder if any of the Chrome extensions sell your data. I'm sure a number of them. Yeah, I have this Chrome extension installed right now, TBPN Timeline Viewer. It was vibe-coded by someone sitting over there. He doesn't even know what programming language it was written in.

44:56This is not true. did i ever tell the story on the show on the show i don't know so tyler doesn't want to tell it so tyler gives us this uh we use a chrome plug-in to to like track the show when we're sharing posts between us and uh this chrome plug-in he like vibe coded it and he sends it over and i unpack it to install it and i'm like why are there like node modules here like uh like that's usually for like node.js javascript on the back end uh and he's just like what are you talking about i was It's like you don't know that you're using Node.js? It's a good extension, sir. Because I think it was just so...

45:33I trust Claude. I trust Claude to make the right decision. I don't even specify what programming language it uses, which is like pretty sick. It's actually extremely bullish for Claude and Claude code. It's really good. Anyway, part of the Code Red, of course, is that OpenAI's Sora app has fallen out of the top 20 most downloaded apps in the United States on both the App Store and Google Play. and so things are falling. I actually opened up Sora today. I looked at it and there was some cool stuff happening. This is a little bit of a hot take. Like it was not, there was still a lot of slop, which I would define as like, you know, it's a POV video of a bus driver with a bunch of cats on the bus and it's like cute and funny or like, you know, it's a chipmunk water skiing, like that type of stuff.

46:22Is that why you were late for the gym today? No. but I was sneaking a peek at Sora while I was driving if I die and crash because I'm looking at slop this would be extremely depressing at a stop sign? yeah I was stopped but there was one cool one which was more like pixel art actually and it was interesting because you remember the OpenAI Super Bowl ad? yes if you prompt Sora to make that type of content it actually is really cool and you can remix it in a very interesting way. And so, uh, like Sam had taken, somebody else had done like a bunch of geometric shapes, uh, pulsing to like electronic music.

47:06And then Sam was able to take him to say, make it orchestral music and make them pastel colors. And he was able to like remix off of that. And that felt like, okay, maybe we're getting into Suno territory. Very odd that Suno and Sora are so close in names. I don't know how that happened. Maybe they should team up or something. Wouldn't be the first time OpenAI has named something directly similar. I don't know. But I don't know. I was seeing like, I don't think it's fully over. I think it's like it's in a, it might be in just a trough of disillusionment. You don't know. This could be a trough.

47:45The trough is in a trough. The trough is in the trough. It's entirely possible. But clearly the vibes are rough and people are taking shots. Terminally online engineer says, just put the ads in the chat, little bro, in the chat bot. Because Sam Altman says OpenAI is making a very aggressive infrastructure bet with new partnerships. To be fair, this clip was from a long time ago that was like at least a couple months ago. Yes, yes, yes. And also you can do both. What is interesting is that it's maybe - And it sounds like they're delaying ads. Which feels odd because I personally was, I'm maybe the only person that's really excited about ads in ChatGPT.

48:28I think it's a good thing for the business. I think it makes a ton of sense. And I was excited to see where that rolls out. I hope that they don't delay it. I think that that's where they should be running. But if they really are losing ground to Google and Gemini and like the Gemini app so quickly. I'm shocked because it feels like the Gemini 3 News, like the launch went well. People were excited about the model. The model card looked good. The benchmarks looked good. But you still have to be pretty tuned in to understand the nuances of the model one way or another. Like it's just not like the big model smell and the vibes.

49:04Like your average AI user doesn't care if the model responds with, it's not just this, it's that. Like most people clearly, that's why it wound up getting RL'd into the model. Most people are like, wow, contrastive parallelism. This is epic. I love it. Thank you. Like this is really - Contrastive parallelism. Yeah, antithetical parallelism. Like I've never, this is like a big, big, big phrase, big word, like this is amazing. And so I'm shocked that there would be such a – I'm not shocked by like a vibe shift on X and in Teapot with regard to how people have been skeptical of the open AI financing.

49:49And so they've been looking for a crack to show. And Gemini coming out and leapfrogging a little bit, even if it's just on some obscure benchmark that the end user might not even care about. I was really interested in, I understand that like X would jump on that narrative, but I'm surprised to see if it's true, this idea that like there's actually some sort of consumer shift. And I mean, it seems like with the red alert comments, like maybe, maybe it is, maybe it is. So you think they have to explain the funding gap at this point, or can we all just agree that maybe everyone got a little too excited?

50:32Yeah, I don't know. I don't know. I feel like everyone's sort of repriced everything already with the Oracle round tripping and just this idea that, you know, some of the equity investments, like they are circular, but it's basically just like a discount on their purchases. And, you know, these things probably aren't as binding as we rethink. And so I feel like the open AI is going to blow up the economy narrative. I feel like that was really oversold and is much, it should be fading in my opinion, but I don't know. Buko Capital Bloke has been digging into the funding hole. Apparently, ChatGPT is also down right now.

51:16I just tested it. It's not down for me, but the chat says it's down. And X is saying it's down. Oh, really? Oh, wow. Time to Baja Blast those surfers back online, brother. It's time to rock. We need a pump-up speech that doesn't include any negative phrases that can be taken out of context. We need to be Baja Blasting. We have to Baja Blast. We have to Baja Blast our way to the top of the App Store. Sora team, I need you to Baja Blast. Bill? It's time to Baja Blast to the top of the App Store. You have to Baja Blast to the top of the App Store. You have to. And we're going to have to Baja Blast some funding into this company because apparently there's a$270 billion funding hole here.

52:08This is from a podcast between Ranjan, who writes at Read Margins, and Alex Kantrowitz at Big Technology. They did a podcast together. and here's the quote from Buco Capital Bloke. It says, squaring the total, it leaves OpenAI in a$207 billion funding hole. The math doesn't work. Maybe OpenAI should release to the world. Here's how the math can work because I haven't seen anyone state how this can actually work. And so even if you get there, OpenAI does fall$207 billion short of the money. It needs to continue funding its commitments, right? So it has in 2030, OpenAI free cashflow will be about 287 billion.

52:48That's like insane. That's, this feels like silly to me because if you're in a situation where you have$287 billion of free cashflow, like you can't raise more debt on that. Like I feel like math tends to work out when you go from a nonprofit to a$300 billion cashflow a year in 10 years. Like it just, everything just forms in front of you. Like, yes, you are building the bridge as you're driving, but like that tends to happen when you're on that much of a tear. The bigger question is like, can they actually free cash with$287 billion in 2030? So Amazon's free cashflow for 2024 was 38 billion. And let's see what Google's.

53:41Yeah, this is like a - 72. So saying that they're going to do three times more than Google and Amazon. So this is the HSBC report is modeling$386 billion in annual enterprise AI revenue by 2030. Enterprise AI revenue. Huh. These are just huge numbers. It's almost not worth analyzing. I still think the biggest thing is just understanding how tied up are these contracts. Well, let me tell you about FALL, the general media platform for developers. Develop and fine-tune models with serverless GPUs and on-demand clusters. So what else is going on? We should read through Ben Thompson's latest piece because he's provided a lot more context on Google, NVIDIA, and OpenAI with a post called Google, NVIDIA, and OpenAI.

54:41Would you look at that? And we thank Ben Thompson for always having an even keel. Highly recommend subscribing to Stratechery. It's a fantastic publication if you're not subscribed already. And he's a former guest of the show. So let's read through his latest Monday piece. He says, a common explanation as to why Star Wars was such a hit and continues to resonate nearly half a century on from its release with everyone except Jordy Hayes, who hasn't seen it. He hasn't seen any movies. I've seen Star Wars, John. How many Star Warses have you seen? I have to have seen all of them except some of the more.

55:19All of them except some of them. Well, no, the more recent ones. Hasn't there been a new Star Wars in the last few years? How many Star Wars are there, Jordy? Is there six? Six? There's six Star Warses. That's how many movies they've made. Six like real Star Wars. There's six. There's six. Hasn't there been like six? Everyone calls it the septilogy. Yeah, there's six. Wait, there are six. There's nine. There's three trilogies. There's the original trilogy, the prequel trilogy, and then the sequel trilogy. And then there's also two spinoffs. Okay, so I didn't watch any of the new ones. But you watched the prequel trilogy.

56:02The Rise of Skywalker. You're talking about George Lucas directed. Yeah. Okay. Okay. So he's a Lucas head. Yeah, exactly. Okay. So you've seen A New Hope. You've seen Empire Strikes Back. I look at those as like real Star Wars movies. You've seen Return of the Jedi, and then you've seen Phantom Menace and Revenge of the Sith and Return of the something. I can't actually. I actually don't know that much about Star Wars. But anyway, you should know enough to follow along with this analogy from Ben Thompson. And he says, you have Luke, bored on Tatooine, called to adventure by a mysterious message born by R2-D2 that he initially refuses, refusing refusal of the call.

56:39This is the classic hero's journey. So he refuses the call. A mentor in Obi-Wan Kenobi leads him to the threshold of leaving Tatooine and faces tests while finding new enemies and allies. He enters the cave, the Death Star, escapes after the ordeal of Obi-Wan's death. Spoiler alert, Ben, what are you doing, brother? What if somebody hasn't seen it and they don't know that Obi-Wan dies? It's crazy. And carries the battle station plans to the rebels while preparing for the road back to the Death Star. He trusts the force in his final test and returns transformed. And when you zoom out to the original trilogy, it's simply an expanded version of the story.

57:24This time, however, the ordeal is in the entire second movie, The Empire Strikes Back. The heroes of the AI story over the last three years have been two companies, OpenAI and NVIDIA. The first is a startup called, with the release of ChatGPT, to be the next great consumer tech company. The other was best known as a gaming chip company characterized by boom and bust cycles, driven by their visionary and endlessly optimistic founder, transformed into the most essential infrastructure provider for the AI revolution over the last few weeks. However, both have entered the cave. They're in the cave.

58:02This is the cave of disillusionment and are facing their greatest ordeal. The Google empire is very much striking back. And I believe, didn't Anjane over at A16Z coin that, the empire strikes back? Formerly A16Z, when independent. Oh, he's independent? Yeah. Oh, I had no idea. So is this what Sam meant when he tweeted the picture of the Death Star? I feel like we never really figured out what he meant by that. That was before GPT-5, I think. Yes. I think he wanted the 2025 vague post of the year. No. It's so vague that even after the release, he's still a no-no. No, no, no. I think this is it.

58:41I think this is it. It's Google's the empire, and he's launching the thing that will take a shot at Google's director. It still didn't make sense at the time because OpenAI was clearly in the lead of the models. Like 2.5, I think 2.5 was the best gem I've ever had. Not on cash flow. Who has more soldiers? Who has more researchers? Who has more TPUs?

59:03It'd be fair to characterize Google as the empire the whole time. I guess the founding of OpenAI, then Google was definitely the Death Star. Interesting. Isn't that their kind of origin story of OpenAI? Yeah, it was. They were worried about that. Anyway, I enjoy the Star Wars-based analogies almost as much as I enjoy Numeral.com. Compliance handled. Numeral worries about sales tax and VAT compliance so you can focus on growth. So Google strikes back. The first Google blow was Gemini 3, which scored better than OpenAI's state-of-the-art model on a host of benchmarks, even if actual real-world usage was a bit more uneven.

59:46Gemini 3's biggest advantage is its sheer size and the vast amount of compute that went into creating it. This is notable because OpenAI has had difficulty creating the next generation of models beyond the GPT-4 level of size and complexity. What has carried the company is a genuine breakthrough in reasoning that produces better results in many cases, but at the cost of time and money. Time and money? Throw in a ramp.com ad right in the middle of the story article. I love it. Gemini 3's success seemed like good news for NVIDIA, who I listed, Ben, listed as a winner from the release. Quote, this is maybe the most interesting one.

1:00:28NVIDIA, who reports earnings later today, is on one hand a loser because the best model in the world was not trained on their chips, proving once and for all that it is possible to be competitive without paying NVIDIA's premiums. On the other hand, there are two reasons for NVIDIA's optimism. The first is that everyone needs to respond to Gemini, and they need to respond now, not at some future date when their chips are good enough. Did you know that Sundar was, like people were claiming that he had used the phrase code red back in 2022 at the Chad GPT launch when Bard was popping? Yes, yes, yes.

1:01:02And so I'm remembering that now, but I missed it. Sundar came out and said he didn't use that exact term, but there was reporting that he did. And I heard a rumor that he also said that he wanted to Baja Blast Sam Altman out of San Francisco, out of the atmosphere with a Death Star laser. I want to try to find more historical examples. Of Baja blasting, folks. We got a Baja Blast. You got a Baja Blast sometimes. So Google started its work on TPUs a decade ago. Everyone else is better off sticking with Nvidia, at least if they want to catch up. Secondly, and repeatedly, Gemini reaffirms that the most important factor in catching up or moving ahead is more compute.

1:01:55This analysis, however, missed one important point. What if Google sold its TPUs as an alternative to Nvidia? We're going to talk to Tay Kim, author of The Nvidia Way, about that. He's going to tell all. He's going to tell all. He's breaking his silence. So that's exactly what the search giant is doing. first with a deal with Anthropic, then a rumored deal with Meta, and third with a second wave of NeoClouds, many of which started as crypto miners and are leveraging their access to power to move into AI. So a lot of those NeoClouds, they found a bunch of power and they don't really have the right chips yet, or maybe they're upgrading their chips.

1:02:31They might be in a new cycle. And so TPU could be at the top of the menu for them. Suddenly it is NVIDIA that is in the crosshairs with fresh questions about their long-term growth, particularly at their sky-high margins. If there were, in fact, a legitimate competitor to their chips, this does, needless to say, raise the pressure on OpenAI's next pre-training run on NVIDIA's Blackwell chips. The base model still matters, and OpenAI needs a better one. And NVIDIA needs evidence that it can be created on their chips. What is interesting to consider is which company is more at risk from Google and why.

1:03:04On one hand, NVIDIA is making tons of money, and if Blackwell is good, Vera Rubin promises to be even better. Moreover, while Meta might be a natural Google partner, the other hyperscalers are not. They're not going to be selling. We're going to have the CEO of Amazon Web Services, Matt Garman, on the show in just 30 minutes. And AWS announced a new chip today. And I don't think AWS is going to be buying TPU anytime soon, but we will be asking him that question exactly. And I want to get to the bottom of it. So OpenAI, meanwhile, is losing more money than ever and is spread thinner than ever, even as the startup agrees to buy ever more compute with revenue that doesn't exist yet.

1:03:45That's a wild sentence. Not mincing words. And yet, despite all that and while still being quite bullish on NVIDIA, I still like OpenAI chances more. Whoa. Oh, Ben Thompson likes OpenAI's chances. Indeed, if anything, my biggest concern is that I seem to like OpenAI's chances better than OpenAI itself. Whoa. NVIDIA's MOTES. Wait, he wrote this before? He wrote this before the red alert memo. Interesting. He's really got a crystal ball over there. So NVIDIA's MOTES. If you go back a year or two, you might make the case that NVIDIA had three MOTES relative to TPUs. Senior superior performance, significant more flexibility due to GPUs being more general purpose than TPUs, and CUDA and the associated developer ecosystem surrounding it.

1:04:32OpenAI, meanwhile, had the best model extensive usage of their API and the massive number of consumers using ChatGPT. The questions then is what happens if the first differentiator for each company goes away? That, in a nutshell, is the question that's been raised over the last two weeks. Does NVIDIA preserve its advantages if TPUs are as good as GPUs? And is OpenAI viable in the long run if they don't have the unquestioned best model? So NVIDIA's flexibility advantage is a real thing. It's not an accident that the fungibility of GPUs across workloads was focused on as a justification for increased capital expenditures by both Microsoft and Meta.

1:05:10TPUs are more specialized at the hardware level and more difficult to program for at the software level. to that end, to the extent that customers care about flexibility, then NVIDIA remains an obvious choice. The interesting thing about the flexibility is that isn't SSI a big TPU buyer? I feel like SSI was maybe going big on TPU. And I think of SSI as very much like, we're going to experiment. We need maximum flexibility. By default, I would assume that they're a heavy consumer of GPU because they want as much flexibility as possible. But maybe the nature of Ilya's research is flexibility within that is still afforded within the TPU ecosystem.

1:05:52They're using TPUs through Google Cloud. Yeah. But there's been no. Oh, yeah, they're not buying them. But still, it just implies more flexibility because you can just turn it on or off. Yeah, I'm talking about the actual like, like the TPUs is an ASIC. It has, is like literally like less features than the GPU. Like a gaming GPU, the TPU doesn't, I think it doesn't support like FP4, right? Or something like that. There's some type of math that is harder to do on a TPU because it's making trade-offs, right? And so even though I don't understand it fully, I understand that TPUs are more specialized at the hardware level.

1:06:31And so if you were to be in like the era of research, maybe you would want something that's less specialized because you'd be like, I'm going back to exploring all sorts of different types of math that aren't necessarily optimized. Yeah, but again, if you're buying TPUs through the cloud, you're effectively just buying cloud services, you do have more flexibility because you can say, hey, we want to use more of this or we want to use less. You're not like buying a bunch of servers and chips. On the scale thing, that makes perfect sense. Yeah, I think also like historically TPUs have definitely been more restrictive just because of the software was just not as good or it was closed source or whatever.

1:07:09And then yesterday, Dillon Patel was talking about how Google is slowly trying to open source more and more stuff for the TPU. So you would imagine that in the future, it should be much easier to use TPUs generally. Sean in the chat says, flexibility in terms of, hey, I have this new architecture. Do I need to write kernel code from scratch or is there a nice CUDA module I can use just time, right? Flexibility is engineering work from NVIDIA. Yeah, yeah. No, that's a really good point. So CUDA, meanwhile, has been a critical source of NVIDIA lock-in, both because of the low-level access it gives developers, but also because there is a developer network effect.

1:07:41Dylan Patel was talking about this. You're just more likely to be able to hire low-level engineers if your stack is on NVIDIA. The challenge for NVIDIA, however, is that the big company effect could play out with CUDA in the opposite way to the flexibility argument. While big companies like the hyperscalers have the diversity of workloads to benefit from the flexibility of GPUs, They also have the wherewithal to build an alternative software stack. That they did not do so for such a long time is a function of it simply not being worth the time and trouble. When capital expenditure plans reach the hundreds of billions of dollars, however, what is worth the time and trouble changes.

1:08:21A useful analogy here is the rise of AMD in the data center. That rise has not occurred in on-premises installations or the government, which is still dominated by Intel. Rather, large hyperscalers found it worth their time and effort to rewrite extremely low-level software to be truly agnostic between AMD and Intel, allowing the former's lead in performance to win the battle. Uh, and so AMD better performance, better efficiency per dollar, but didn't have the best software. And now, but now because there's so much, uh, on the line, so many, so the spending amount is so high companies will go and work around all the bugs, develop new software that allows them to take advantage of AMD's better performance.

1:09:06Uh, in this case, the challenge NVIDIA faces is that its market is a relatively small number of highly concentrated customers with the resources, mostly as yet unutilized to break down the CUDA wall as they already did in terms of Intel's differentiation. It's clear that NVIDIA has been concerned about this for a long time. This is from NVIDIA Waves and Motes, which he written at the absolute top of the NVIDIA hype cycle after the 2024 introduction of Blackwell. This article takes full circle. This takes this article full circle. This is from the previous Yes, Ben Thompson, article in Stratechery.

1:09:42It says, in the before times, i.e. before the release of ChatGPT, NVIDIA was building quite the free software moat around its GPUs. The challenge is that it wasn't entirely clear who was going to use all of that software. Today, meanwhile, the use cases for those GPUs is very clear, and those use cases are happening at a much higher level than CUDA frameworks, i.e. on top of models. That, combined with the massive incentives towards finding cheaper alternatives to NVIDIA, means both the pressure to and the possibility of escaping CUDA is higher than it ever has been, even if it is still distant for low-level work, particularly when it comes to training.

1:10:22NVIDIA has already started responding. I think that one way to understand DGX Cloud is that is NVIDIA's attempt to capture the same market that is still buying Intel server chips in a world where AMD chips are better because they have already standardized on them. NIMs are another attempt to build lock-in. In the meantime, though, it remains noteworthy that NVIDIA appears not to be taking as much margin with Blackwell as many have expected. The question as to whether they will have to give back more in future generations will depend on not just their chip's performance, but also on re-digging a software remote increasingly threatened by the very wave that made GTC such a spectacle.

1:11:02So Blackwell margins are doing just fine, I should note is he's back to the original article, the modern article, as they should in a world where everyone is starved for compute. Indeed, that may make this entire debate somewhat pointless. Implicit in the assumption that GPUs might take share from GPUs is that for one to win, the other must lose. The real decision maker may be TSMC, which makes both chips in his position to be the real break on the AI bubble. Interesting. So, ChatGPT and Moats resiliency. I can read through this one. ChatGPT, in contrast to NVIDIA, sells into two much larger markets.

1:11:43The first is developers using their API and according to OpenAI. Anyways, this market is much stickier and reticent to change, which makes sense. Developers using a particular model's API are seeking to make a good product. And while everyone talks about the importance of avoiding lock-in, most companies are going to see more gains from building on and expanding from what they already... Always Baja Blast. And for a lot of companies, that is OpenAI. One, I would caveat here, we were talking to a founder yesterday who, off the show, who was saying he immediately, as soon as Gemini 3 launched, spent like 12 hours just moving over to Gemini from OpenAI.

1:12:27So depending on the product, I don't know that API is always going to be super sticky. I say winning business one app by one app by one will be a lot harder for Google than simply making a spreadsheet presentation to the top of a company about upfront costs and total cost of ownership. Still, API costs will matter, and here Google almost certainly has a structural advantage. The biggest market of all, however, is consumer, Google's bread and butter. What makes Google so dominant in search, impervious to both competition and regulation, is that billions of consumers choose to use Google every day, multiple times a day, in fact.

1:12:59Yes, Google helps this process along with its payments to its friends, but that's downstream from its control of demand, not the driver. What is paradoxical to many about this reality is that the seeming fragility of Google's position, competition really is a click away. It is, in fact, its source of strength. And then there's an excerpt from the United States. We can skip this one and continue at the bottom. The CEO of a hyperscaler can issue a decree to work around CUDA. An app developer can decide that Google's cost structure is worth the pain of changing the model, undergirding their app. Changing the habits of 800 million people who use ChatGPT every week, however, is a battle that can only be fought individual by individual.

1:13:41This is ChatGPT's true difference from NVIDIA in their fight against Google. yeah and so this i think is the most important takeaway is just ben thompson creates aggregation theory this idea of like it's so important to aggregate demand in the modern internet world it's potentially the only thing you can do you can't really monopolize supply it's very hard to monopolize supply but monopolizing demand is something that happens um and and and the the strength of habits is significant. Like we're watching this stuff every single day. So we can take the time to, okay, yeah, we should test out this other, you know, model.

1:14:24We should daily drive this app. But for a lot of people, if they've, if they have a map that's installed and they've been using it for a year, they're never changing. Yeah. Even if the model is slightly better over there, they're just not even going to hear about it. Cause they're just like, this is the thing that I use to plan my vacations. And the thing, the thing that I've heard come up multiple times is people that when Gemini 3 launched, they switched to Gemini 3 on desktop, but they stayed using ChatGPT on mobile. And to be completely transparent, the Gemini mobile app is really, really struggling to stay connected.

1:14:57There's something when you fire off a prompt, it doesn't save it locally and then cache that and then send it off, inference it, and then come back. Like, unless you keep the app open, like, it will just give you, like, a server disconnected error. Like, I've gotten, like, dozens of these. And that's going to be a real, like, I think it should be something they should be able to fix in, like, a weekend. But, you know, hopefully it's soon. But for a lot of people. Logan. Yeah, well, they'll get it. They'll get it. But back to the moat and the map, the moat map and advertising. This is, I think, a broader point.

1:15:37The naive approach to moats focuses on the cost of switching. In fact, however, the more important correlation to the strength of a moat is the number of unique purchasers to users. The strength of the moat is increased by the number of buyers. Okay. So you can see where this is going with like NVIDIA has five buyers and ChatGPT has a billion buyers. And once it has that. Or 20 million. Yeah, advertisers might be even more. So this is certainly one of the simpler charts I've ever made because it's literally just one line. But it's not the first in the moat genre. So he talks about the moat map.

1:16:19I argue that you could map large tech companies across two spectrums. The degree of supplier differentiation from Facebook where the supplier is completely commoditized, just your friend on Facebook, to Microsoft and Apple where the suppliers are somewhat more controlled. Yeah, there's the... What a chart. The more unique buyers of your product you have, the stronger your moat, because you have to convince each one of them. And then second, the extent to which a company's network effects were externalized. Internalized network effects are Facebook again, and then externalized as Microsoft. And so putting this together gave the moat map.

1:16:57So who has the network effect versus the supplier's map. If we scroll down there, You can see them. What you see in the upper right are platforms. The lower left are aggregators. Platforms like the App Store enable differentiated suppliers, which allows them to profitably take a cut of purchases driven by those differentiated suppliers. Aggregators, meanwhile, have totally commoditized their suppliers, but have done so in the service of maximizing attention, which they can monetize through advertising. It's the bottom left that I'm describing with the simplistic graph above. The way to commoditize suppliers and internalize network effects is by having a huge number of unique users.

1:17:32And by extension, the best way to monetize that user base and to achieve a massive user base in the first place is through advertising. It's so obvious the bottom left is where Chachabee T sits. I wonder what he thinks then about them potentially kind of delaying ads and delaying ads. Probably punching the air. Him and Eric Sufer could probably, oh, no. And I'm right there with them. I completely agree. Boo! Launch the ads product. Launch the ads product. Get it out. Come on. Don't delay that. That's the most important thing. So at one point, it didn't seem possible to commoditize content more than Google or Facebook did, but that's exactly what LLMs do.

1:18:17The answers are a statistical synthesis of all the knowledge the model makers can get their hands on and are completely unique to every individual. At the same time, every individual's user usage should, at least in theory, make the model better over time. It follows then that ChatGPT should obviously have an advertising model. This isn't just a function of needing to make money. Advertising would make ChatGPT a better product. It would have more users using it more, providing more feedback, capturing purchase signals, not from affiliate links, but from personalized ads would create a richer understanding of individual users, enabling better responses.

1:18:53And as an added bonus and one that is very pertinent to this article, it would dramatically deepen OpenAI's moat. Yeah, I keep going back to this idea that OpenAI needs personalized ads like Instagram. That's what Sam said when he was interviewed multiple times on his ad strategy. He was like, you know, like ads can be bad, but these Instagram ads are pretty good. You know, And people are like, oh, he's backtracking. It's like, no, that's fine. Do the proper business model. Please implement the correct business model. I'm happy about that. But the interesting thing is that Instagram does not serve as ads necessarily when you search for something.

1:19:36Like if you're on a video for like a Ferrari, like you don't just immediately get an ad as your next thing for like Ferrari of Hollywood. like a Ferrari of Beverly Hills. Like, no, you get an ad for the toaster that you were about to check out on. I actually do like half, half, half the ads I get on meta are local dealerships in LA. Yes. But, but importantly, not when you're like, it's not tied to search. And so, uh, ChatGPT can do the same thing where they can clearly show you, um, uh, where they can clearly show you something that you are about to check out. You're shopping for Christmas for this thing.

1:20:18You're searching for the Roman empire. Let's show you the ad for the thing that you're shopping, that you're shopping for like right next to it. It's fine. Uh, and so I think that can work very well. Anyway, uh, let's go to Google's advantages. It's not a question that Google can win the fight for consumer attention. The company has a clear lead in image and video generation, which is one of the, uh, one of the reasons why I wrote about the YouTube tip of the Google spear. I mean, And Google's advantage in data is insane. Like YouTube, so massive. And that's got to be just a compounding compound.

1:20:45I mean, we're uploading another three hours of video to YouTube today. You're welcome. This one's for you, son. This one's for you, Dennis. But the flip side is like they also see the entire internet because the way the Google bot scrapes, like the Google searching in Gemini is such a killer feature. Like it's such a, it's such a killer feature if they can keep that on and they can actually surface that. And the AI search results are obviously going to get good. They're going to figure out how to surface it. I think. I'm still pretty optimistic. But let's see what Ben Thompson has to say. And let's also tell you about Adio, the AI native CRM.

1:21:19Adio builds, scales, and grows your company to the next level. So Google is obviously capable of monetizing users. Even if they hadn't turned on ads in Gemini yet, it's also pointing out, as Eric Sufer did in a recent Stratechery interview, two of the best, collabing. We love to see it. That Google started monetizing search less than two years after its public launch. It is search revenue, far more than venture capital money, that has undergirded all of Google's innovation over the years. And it is what makes them such a behemoth today. In that light, OpenAI's refusal to launch and iterate on an ads product for ChatGPT, now three years old, is a dereliction of business duty.

1:21:56He's calling them out. Whoa. Particularly as the company signs deals for over a trillion dollars of compute. What are you doing? Sam, get the ads out. Put the ads in the chatbot. We love it. Just put the ads in the chat. And go, you got to Baja blast some ads into that app. You got to. I want ads in ChatGPT, please. On the flip side, it means Google has the resources to take on ChatGPT's consumer lead with a World War I style war of attrition. Ryan Mittal, call back. OpenAI's lead should be unassailable, but the company's insistence on monetizing solely via subscriptions with a downgraded, degraded user experience for most users and price elasticity challenges in terms of revenue maximization is very much opening the door to a company that actually cares about making money.

1:22:43To put it another way, the long-term threat to NVIDIA from TPUs is margin dilution. The challenge of physical products is that you do actually have to charge people who buy them, which invites potentially unfavorable. I always, so yeah, who, who both Gemini and, and chat GPT will have ads eventually. Yes. Right. Yes. You can bet on that. Yes. Uh, who will ramp ad revenue faster. It's hard not to bet on Gemini, even with a smaller user base because they have the ad network. They have all the, they have all the customer relationships already. You can, they can just say like, Hey, here's a pop-up.

1:23:21Do you want it? Here's, here's$10 ,000 of free ad credits. Try it out. Right. It's like native. It will already be in AdWords. The example that I use is like how, you know, Zuck was able to take Instagram, which had a lot of users, plug it into the meta ads platform and just scale revenue like crazy and then do it again with Reels. And so, yeah, very clear that they both will have ads. And, again, if Gemini can really ramp that quickly, they could, again, I do feel like we're moving towards a world where every consumer will be able to get the best LLM for free. I don't necessarily believe that every American will be paying for an LLM in five years.

1:24:08And so if Google can get there first and then keep Gemini on the frontier and deliver the best, free, fastest text and image model, that's going to be very, very difficult to compete with. And so again, like getting to ads faster, it feels like it makes more sense. I like that point. I have a rebuttal. First, I'm going to tell you about Figma. Think bigger, build faster. Figma helps design and development teams build great products together. So getting to ads first is an advantage. That's your take. I like it, but there is a little bit of a risk with launching ads first because you could forever be branded with you're the ads one.

1:24:54And we saw this when Arvin from Perplexity came on the show and he mentioned this idea of ads in LLM queries. And we all agreed on the discussion. We all agreed that ads were going to come to AI tools because that is the way to get the most people using them and make intelligence free. And you got in a debate with Mark Cuban over this, for example. And it seemed very logical, but there is the fact that the first major chat app to put ads in their app is going to be a massive news cycle. It's just going to be like national news. It'll be Sam Adman. Exactly. It'll be OpenAI has ads now, or it'll be Gemini has ads now.

1:25:40and so you don't necessarily if you like it's much easier to be the second mover there because it's going to be less of a new cycle and so you kind of do want to there is there is a little bit of advantage to being the second mover there right because you're you're going to get sort of branded as like oh that's the ad supported one and the other one can add ads and people be like oh yeah like i guess but that's like standard yeah there's going to be like a backlash and people will be like oh no i don't like this company blah blah blah blah like yeah i mean the The harder thing is just how do you do it, right?

1:26:12Like, I do feel like it's different. You're going to an LLM. People are going to an LLM for advice and recommendations. That's different than going to Google and searching and seeing ads at the top. I mean, there's plenty of surface area. No, no, I'm not saying there's no surface area, but I'm just saying, like, the right way to do ads in LLM is not clear yet. Yeah, yeah. I mean, we'll need to do some experimentation. But I mean, just starting with, you know, like Google already has retargeting information. I actually went to Gemini with a question and it clearly knew everything about me. And it said, you're the host of TVPN and you've started, you founded these companies.

1:26:49And it had, it already knew that probably just because I authenticated with something else. I don't know, but it should know, okay, hey, we could retarget you with this. Let's put this in. And, and the same thing with, with OpenAI, like with ChatGPT, there's plenty of spaces where it's like you're waiting for it to give you the answer. Okay, hey, we're generating you the image. Why don't you show me other images of ads right there? There's tons of surface area. I agree that there will be a whole bunch of iterations on like what the ideal ad looks like. But yeah, you can clearly get out. So let's go back to Ben Thompson, close this out.

1:27:26Says the reason to be more optimistic about OpenAI is that an advertising model flips this on its head because users don't pay. there is no ceiling on how much you can make from them, which by extension means that the bigger you get, the better your margins have the potential to be, and thus the total size of your investments. Again, however, the problem is that the advertising model doesn't exist yet. So he started this article recounting the hero's journey in part to make the easy leap to The Empire Strikes Back. However, there is a personal angle as well. The hero of this site has been aggregation theory and the belief that controlling demand trumps everything else.

1:28:03Google was my ultimate protagonist. Moreover, I do believe in the innovation and velocity that comes from a founder-led company like NVIDIA. And I do still worry about Google's bureaucracy and disruption potential making the company less nimble and aggressive than OpenAI. More than anything, though, I believe in the market power and defensibility of 800 million users, which is why I think ChatGPT still has a meaningful moat. At the same time, I understand why the market is freaking out about Google. Their structural advantages in everything from monetization to data to infrastructure to R &D is so substantial that you understand why OpenAI's founding was motivated by the fear of Google winning AI.

1:28:46It's very easy to imagine an outcome where Google's inputs simply matter more than anything else. Which is to say, one of my most important theories is being put to the ultimate test, which perhaps is why I'm so frustrated at OpenAI's avoidance of advertising. Google is now my antagonist. Google has already done this once. Search was the ultimate example of a company winning an open market with nothing more than a better product. Aggregators win new markets by being better. The open question now is whether one that has already reached scale can be dethroned by the overwhelming application of resources, especially when its inherent advantages are diminished by refusing to adopt an aggregator's optimal business model.

1:29:27I'm nervous and excited to see how far aggregation theory really goes. Fascinating. That's his baby. Yeah. I agree. It is the correct framing. It'll just be very interesting to see. I really wonder who's going to take the leap first, who is going to jump and put ads in the app first. It feels like Google should do it. It feels like Google will be able to do it. Yeah, nobody's going to be like, what? Google is putting ads in a product? Yeah. It won't be that surprising. So they should probably move faster. We have some breaking news. What's the breaking news? Jason Freed is joining the show at 2 p.m.

1:30:10I can't wait. Surprise guest. Surprise guest. He's launching Fizzy today. Yes. Kanban as it should be, not as it has been. I will wait we'll wait to talk about this until he joins in an hour and 20 minutes so little surprise guest appearance from a legend. He's calling out his competitors directly I love when the founders do that founder mode founder mode should we talk about John Giandrea leaving the company? He's out I quit you like that one I think that's probably been used before some headline but But Amar Subramania. So, of course, Mark Gurman has the scoop, I believe. The Gurmanator. Gurmanator's at it again.

1:30:55He says, Apple AI chief John Giandrea is leaving the company. Amar Subramania from Microsoft is joined to lead AI under Craig Federighi. And so we should dig in a little bit to this history. So Swix has a little bit of a deep dive here. He says, Amon brings a wealth of experience to Apple. He's quoting here. Having most recently served as CVP of AI at Microsoft and previously spent 16 years at Google, where he was head of engineering for Google Gemini. Wait. Oh, I guess at the end of that, because Gemini is not 16 years old. This is bearing the lead. He joined Microsoft AI four months ago. Wow.

1:31:39What a crazy turn of that. LinkedIn says six months ago. but who's counting? That's pretty fast. But this makes sense considering Apple is partnering with Gemini and not a lot of people are going to be in a better position to help integrate that into Syria than Amar. Yeah. I mean, I don't know. Maybe there's something to, you know, just having a taste of all the different big tech companies. Oh yeah, I've been at Microsoft. I know how they work. I've been at Google. I know how that works. I'm ready to rock over here. Gurman does need to come back on ASAP. I agree, Raghav. He was fantastic. We'll get him in person, hopefully, before the end of the year.

1:32:18So Gurman continues. He says, strange hire for a number of reasons, but it's hard to argue that the Apple job is a bad one. Anything is an improvement at this point, so the bar is as low as it comes, easy to lay up on the resume. So that'll be fun to see. I'm personally just excited to actually test drive what Gemini, how it works in Siri, how seamless that is. Because if it really is just raise, press the button, get Gemini, and it's linked up properly and it doesn't have timeouts and it gets back to you pretty quickly, like that's going to be a pretty powerful experience. That's definitely going to cut down on ChatGPT app usage for iPhone users, I would imagine.

1:33:00Underrated threat. I would think so. Like there are so many moments where People are counting out Apple. It's not. Yeah. I don't even know that Apple will benefit massively from this. It's not like they're going to sell twice as many iPhones. They're already so big. It's not like they're going to charge for it. I don't think it's necessarily like especially bullish for Apple. It's an underrated threat for open AI. I would think so. There's a lot of queries that all hit chat GBT on mobile that are not even like super economic, but just a lot of my usage around like, hey, just trying to learn about something or research a product, et cetera.

1:33:41And if that's just like, again, one tap and you're in there. Yeah. I mean, yeah, the original promise of Siri was, you know, not just, hey, what's the weather today? But really asking anything. And Gemini clearly solves that for 99 % of knowledge retrieval queries. I would be, I think I'm going to be using that a lot unless they really botch it. I don't know how they're going to botch it. Yeah, I mean, I think the... If anything's possible, yeah. Apple's like, hold my beer. They're going to be like, for privacy reasons, every time you press the button, you have to e-sign. And it's like, why are we doing that?

1:34:19Yeah. The original Siri kind of vision was this very conversational AI, right? Yeah. But I don't think Gemini has a real-time voice model yet. I'm pretty sure OpenAI is the only one that has that. Really? Huh. I don't think that matters at all. Really? Yeah, I really think that... Because that form factor feels like that would be the best thing to be on. You press the button on your iPhone, you're immediately talking. And then it's just on. Yeah. I mean, I wouldn't be surprised if they can get that model, that version of the model out. Because it's really just distilled a little bit faster. It's not some uncanny breakthrough that the Gemini team will never be able to crack, right?

1:35:00So they just have to build that. But honestly, I don't know that that's certainly not how I would use it for most things. Most things I would say, okay, I have one question, get me an answer within a reasonable amount of time, and maybe read it off to me or produce an article that's a pretty readable article summarizing the answer to my question. And then, yeah, maybe there is a back and forth, but I don't know. Oh, you're getting truth zoned in the chat. Gemini does have a real-time voice feature. Gemini Live. Yeah, I think it's on the app. I've used it. I'll try that. I don't have an app. Large journalistic force headed door to you.

1:35:39Stand by. Tyler, Anthropic is acquiring Bun. What do you have to say? Yeah, I mean, this is definitely in line with their, like, you know, focus on dev stuff. Okay. What is Bun? Bun is a – it's like a – I don't know. It's like a bundler for JavaScript. It's like a very dev. It has dramatically improved the JavaScript and TypeScript developer experience. They're going to make Claude Code even better. I like that Gabriel from OpenAI here is OMG in the chat, which is like a pretty crazy thing to say. But I appreciate it. So Claude is one of the world's smartest and most capable AI models for developers, startups, and enterprises.

1:36:26Cloud Code represents a new era of Agenta coding, fundamentally changing how teams build software. In November, Cloud Code achieved a significant milestone just six months after becoming available to the public. That's crazy. It's only been six months. Wow. It reached a$1 billion revenue run rate. We were always struggling to understand what that meant, right? Yeah. So there are two ways to pay for Cloud Code. There's either with your Cloud subscription where you get like Cloud Pro or Cloud Max, and there's a certain amount of tokens you can use. And then there's also, you can just directly wire up APIs calls, essentially, to CloudCode.

1:37:00And then you're being charged directly based on usage. So that's probably what that revenue is from. Yeah. And then, yeah, I also have thought about this thing where, like, oh, you can break down the number of tokens from the subscription. So it's like your$20 subscription, three-fourths of your tokens are on CloudCode. I mean, three-fourths of your$20 counts as CloudCode revenue. Okay. Yeah, yeah. That makes sense. I'm not sure exactly. Yeah, I mean, I'm sure that they can account for it. So it was founded by Jared Sumner in 2021. Bunn is dramatically faster than leading competition. They say it's a breakthrough JavaScript runtime.

1:37:32Does it compete with V8? I'm very interested in like, it's Node.js. It competes with Node. The thing that Tyler needs to learn about. What was that company that OpenAI acquired earlier this year for like a billion? I know the one you're talking about. Isn't that analytics or something? Yeah, but again, I remember at the time people were like, oh, OpenAI's competitors are not going to be happy about this acquisition. So that comment from Gabriel. Well, congratulations to the Bunn team. Congratulations to Anthropic and everyone on the Cloud Code team. Very excited that you're getting to work together for your massive deal.

1:38:12Speaking of other massive deals. Speaking of size. Alfred Lin. Get that gong ready, John. You were built like that. What did he do? Alfred Lin comes in on the board of DoorDash, buys$100 million of DoorDash. Lin Sanity. He's not done stewarding DoorDash. He's continuing to steward the company with a$100 million buy. And, of course, sends the stock up almost 6 % on that. Pretty remarkable. People are excited about it. He ripped. He ripped. what is this another 4.6 million to donated to shrimp welfare okay so yeah basically the story is in Theropic they were doing some like research and about smart contracts and so they had Claude Code try to figure out like you know issues in smart contracts and then I'm not sure exactly where the money came from.

1:39:16Maybe it was for bounties. But there was some way in which Claude Code basically generated $4.6 million in cash from finding these exploits. So then they just... Did it actually generate real money? Or is this the hypothetical? It's simulated testing. Okay. Huh. Simulated. So it probably means that they could have basically stolen$4 million from people, but they don't want to do that. Maybe they should have. If they really want to get up the. In other anthropic news, David Sachs, as he's still waiting on Dario's support after the New York Times piece was published. Sam Altman, of course, came in and said, David Sachs really understands AI and cares about the U.S.

1:40:00leading in innovation. I'm grateful we have him. Of course, Dario and Sachs, not the biggest fans of each other. So I don't expect that one coming through anytime soon. And while we wait for our first guest, Matt Garman, let's pull up this clip from Huberman Lab. If we can play this, Rob Moore is highlighting, Dr. Jeffrey tells Huberman that LED lighting in buildings is a public health crisis that could be on par with the use of asbestos. Many building contractors slash designers are coming to him worried they're going to be sued and asking how to start fixing the issue. So let's pull this up when we have a second.

1:40:46Because I am very concerned about the amount of short wavelength light that people are exposed to nowadays, especially kids. The group of us that are shuffling around, some of them are saying this is an issue on the same level as asbestos. This is a public health issue and it's big. LEDs came in and people won the Nobel Prize for this very rightly at the time because they save a lot of energy. The LED has got a big blue spike in it, although we tend not to see that. And that is even true of warm LEDs, and there is no red. The light found in LEDs, when we use them, certainly we use them on the retina looking at mice, we can watch the mitochondria gently go downhill.

1:41:32They're far less responsive. Their membrane potentials are coming down. The mitochondria are not breathing very well. You can watch that in real time. Under LED lighting. And LED lighting at the same energy levels that we would find in a domestic or commercial environment. This is why I want to rig the studio with incandescent light. Incandescent. We're going back to candles. Candle, Max? Let's do candlelight. How about a hearth? This is the way... If we put a hearth so we have lights above our heads, I'm sure, or LEDs, killing us slowly and softly. If we put a bonfire right above us... That would be the way.

1:42:11And then we just, when the wood kind of burns out, the show's over. We just go until the... That would be good. I like that. Let me tell you about Turbo Puffer, serverless vector and full-text search, built from first principles and object storage, fast, 10x cheaper, and extremely scalable. And Turbo Puffer is at reInvent. Amazing. Fantastic. Well, we are joined by the CEO of Amazon Web Services, Matt Garman. Thank you so much for taking the time to come and chat with us. How are you doing? Hi. Thanks, guys, for having me. Please take us through some of the high-level announcements. Obviously, it's reInvent.

1:42:47Very exciting. Congratulations on all the progress. We'd love to know what's at the top of your mind. What's on the top of your presentations over the course of the event? Yeah. We had a couple of really exciting announcements today. A couple I'd highlight. First, we introduced this idea of Frontier Agents. These are agents both in Kiro for software development as well as in operations and security. And these frontier agents are meant to accomplish much, much more than customers were ever able to do in the past, where we have these autonomous agents that can help customers really turbocharge their software environment.

1:43:23So super excited about that. We had some announcements around Nova, which is our frontier AI models that we announced. We announced Nova 2 in our new sets of models. And one of the things I'm in particular really excited about is Nova Forge, which allows customers to actually bring their own data to pre-training checkpoints, mix in their data with Amazon data, finish training the model, and at the end of it have a custom model that deeply understands their own enterprise data and is just for them. So that's another thing that I'm excited about. And then the third thing is we announced a new chip around Tranium 3 to really turbocharge the next generation of training and inference for our customers.

1:44:04And so quite excited to get that. And that went GA today as well. That's very exciting. Let's go back and start with the first one. Let's talk about coding agents and your own proprietary models. How are you thinking about positioning those to potential buyers? Do you like the benchmarks these days? Do you think that we're sort of like post all the benchmarks? Or do you think those are still useful tools for a buyer who's making a decision? Is it about integration? Is it about cost? How are you positioning them? Yeah. When you think about software development, it's not about pure benchmarks. It's really about what is going to allow you to get the most amount of work done.

1:44:45And when we think about our offering, which is called Kiro, it's really focused on in an enterprise or environment where somebody is doing high-velocity development, they actually need more structure. People love VibeCoding, and it's exciting, but you can actually get down a path where you get stuck. and you'll often find actually that you spend just as much time trying to get back to where you were before as if you had just coded it from the beginning. We have this idea of specs that gives you structure to what you're trying to build. And so you can have agents go and start to build around those specs together with you and your team.

1:45:17And it gives you the structure that allows you to go really fast, can undo if you need to, can make sure that you're hitting your design requirements. And it really allows you and the agents to operate in conjunction with each other and move really, really fast. And we're starting to build these much more capable agents that can go and actually do long-running tasks for you on your behalf. But all of it kind of ties into this structure. And we view that as a way to deliver kind of real development that's going to be meaningful on a large code base with large teams and enterprises where they have existing things, not just kind of single individual people sitting there kind of doing vibe coding.

1:45:52Which, you know, you can do vibe coding on Kiro as well, by the way. We think that that's just not sufficient for what real development is going to need. Sure, sure. That makes sense. about what it actually looks like to set an agent off and say, hey, I got a task for you, come back to me in a few days, which it sounds like that's where we're going. We've been tracking the meter benchmark and it seems like we've been seeing doublings there. But again, a lot of those have been, the benchmark has been, how long would it take a human to do this task? The actual agent might've done it faster. And so it's not necessarily that you're actually letting something cook over the weekend.

1:46:30What's the experience been like and what have people been reporting about these long running agents? Yeah, I think the first and actually most important thing is thinking about how you actually kind of have a mind change on how you think about software development, where you think about not about do this task, get it back, look at it, do this task. But how are you thinking about directing a lot of agents to go out there and do lots of different things and let those run for long periods of time where they can kind of have amorphous tasks, like instead of go write me this function, like try to go solve this problem for me.

1:47:00And then it'll come back. And then, but you can, but if you send out two or three or 10 or 20 or 50 of those things, then your job as a software developer and as a product leader is actually much more around coordinating those when they come back, troubleshooting, make sure that, you know, directing them, course correcting, et cetera. And so I'm excited about that. We've already seen these processes go off and work for multiple hours at a time on particularly really hard, tricky, amorphous tasks. And we think those things are going to continue and be more the norm of how software developer teams change what they accomplish.

1:47:33We think Kiro is going to be the engine that's going to drive a lot of that. Yeah, yesterday we were talking to Vincent from Prime Intellect, and they do some of this fine-tuning on smaller models. And he has this thesis, I think, that you share that a lot of businesses will need to take a pre-train and then bring their own data, fine tune it, not just because it's important from performance and output, but also from cost. But I'm interested in understanding how you think the market will shape out. Do you see implementation partners and like consulting firms coming in and doing that? I was asking him like, you know, there's a lot of tech startups that are going to be able to do that.

1:48:15They're going to understand I need to build an RL environment around my app. But for larger legacy companies, they might not understand. So how are they going to wind up using that tool in particular? I think they will. And I actually just want to highlight one piece there where some of what we announced today is a little bit different. We announced this idea. It's an open training model with Nova. And so the difference of what you just said is people take a pre-trained model and they'll do RL after the fact and they'll try to do some fine tuning, which is great, but there is actually limits to where that does.

1:48:46In fact, if you do too much post-training, oftentimes those models will forget what they've done at the beginning. they'll start to lose some of their reasoning and their core intelligence. I mean, this is an unsolved problem, except when you go and insert your data in the pre-training phase. And so what we do with Nova is we expose checkpoints. You can take a 60 % trained or an 80 % trained model, pre-trained model, insert your data into that pre-training phase, mix it in. We then expose actually Amazon training data to you via an API that you can then mix it together. And so it's like you said, here's all my corpus of corporate data.

1:49:20Here's everything that I need to know about my industry. We then mix that in and then finish pre-training the model. So you get a pre-trained model that totally understands your company and your data. And then you can go do fine tuning. You can go do reinforcement learning gyms. After that, you can shrink them down and distill them. You can do all of those things, but on a pre-trained model that deeply understands what your company does. And is that called mid-training now? Is that the right buzzword for that? It's not like we're, and mid-training is a different thing. This is the first time that anyone's ever exposed this idea.

1:49:49Okay, thank you. to deliver pre-training checkpoints where we can mix in your data. No one's ever done this before today. Very cool. It's the first time. Great. Well, then, yeah, on market structure, do you think it's self-serve enough that large corporations will do it? Or do you need an AI lab? Do you need an AI scientist? Do you need someone who can write TensorFlow or PyTorch or something to implement this? Or is it something where just a normal software engineer at a large company could go and pull this off the shelf and implement it? Yeah, we'll see. I think we're going to keep working on the tools today.

1:50:23I do think that for some enterprises, they'll want to have some consulting folks that help them with this. I think we'll have some people where you have some experts that should come and teach how to do this. And I think we'll quickly get the tools to a point where, you know, it's not somewhere where, you know, a non-technical person is going to go do this for sure. But it may be a software developer that tends to be a little bit more on the AI or ML side that we hope is going to be able to go do this without having to have a whole bunch of expertise about how to go free train a frontier model.

1:50:52Yeah. On the cost side, obviously you're working, you announced a new chip. I imagine that there's, you know, the emergence of some synergies across the models that you're developing, the software you're deploying, the cloud, and then also the chips. how are you positioning the Tranium ecosystem? Is this something that you're planning on really doubling down on across the entire stack? Or do you want to be more chip agnostic? Are we going to see you buying TPUs in the future? No, we definitely, well, a couple of things that I'll unpack. The first is we're very excited about Tranium and think it has enormous potential.

1:51:32And we absolutely think there's a benefit to optimizing every single layer of that stack where we have the best cost performance that we can deliver at Tranium. We have optimized models for you to use and applications and agents at the top of that that we talked about. So we think that whole optimization of that stack is going to be critically important. And of course, we're going to support choice for our customers as well. And so we'll continue to offer GPUs from NVIDIA as an example. And we have a very tight partnership there. But we do think and we're quite excited about what Tranium 3 is going to offer for customers.

1:52:07And I do think that we're going to see an explosion of that ecosystem as more and more people get access to those chips and are able to take advantage of the pretty significant cost performance benefits that you can get from running on Tranium. How are you thinking about the open source ecosystem that you need to build around Tranium? That's the big discussion with the TPU right now, the question of Google has some amazing folks. They have some amazing software folks, it seems like. They don't necessarily need to open source everything. And so a lot of people are waiting to see how much the industry builds open source alternatives independently versus how much does Google just give away.

1:52:48What's your thought process on building an open source ecosystem or even just giving developers access to closed source software to run efficiently on Trinium? Yeah. No, we're all in favor of having an open set of software to run on Tranium. In fact, we have our Neuron SDK, which is open source today and allows everyone to contribute to that. We think that the more that we can collaborate on that software ecosystem to make it easier for people to use chips, and we, of course, support the broad set of whether it's PyTorch or other kind of open frameworks as well. So we collaborate across the industry on that and are big advocates of contributing to and supporting that open ecosystem.

1:53:33Jordy? Love to get your insight on just like general constraints for AWS as a business, what you guys are doing on the power side. Is that a real constraint? Anything that you can share there? Yeah. You know, look, as we're scaling incredibly rapidly, we recently announced that we've added 3.8 gigawatts of data center capacity in the last year alone, which is just an insane amount of data center capacity. Thank you. You're welcome. And so it's ramping incredibly fast. And it is a constraint. You know, we have more demand than we have supply today for AI. Sure. And as we ramp up the supply chain, we think about all of the constraints.

1:54:18We think about chip constraints. We think about networking constraints. We think about power constraints. We think about networking constraints, data centers, et cetera. And so we're working really hard to try to remove every single one of those. And with an industry that's growing as rapidly as the AI one is, there's always going to be some constraints. And we work really hard to keep removing blockers every time so we can keep growing fast. Makes sense. Well, we have a hard stop. So thank you so much for taking the time on such a busy day to come chat with us. Would love to have you back on the show and go way deeper.

1:54:50But thanks so much, and congratulations on all the massive releases. We're excited to dig in deeper and keep chatting about them. But have a great rest of your day. We'll talk to you soon. Thanks for having me on. Goodbye. Cheers. Let me tell you about public.com. Investing for those that take it seriously. They've got multi-ass investing, and they're trusted by millions. We have Tay Kim, author of The NVIDIA Way, and a Barron's senior writer joining the show. Take Kim, the author of the NVIDIA way. Take Kim on the beat. Thank you for joining the show. And I'm sorry it took us so long. We've exchanged posts on X many times, and we wanted to have you on the show earlier, but I'm so glad you found it up when it did, because it's the perfect time to talk to you.

1:55:33Do you have roommates? I do not. No roommates. No roommates. Long-time listener, first-time caller. I'm so excited to be on this. I'm so excited to have you here. Reminder, everyone, go buy the book Seriously, Christmas is coming I can't imagine a better gift For a 10 year old? For a 4 year old I know what my son's getting He's getting the NVIDIA way Teach him young Multiple copies No, seriously, get 10 copies A lot of teenagers have read it It's amazing It's an inspirational, entrepreneurial book That a lot of parents are giving to their kids I love your headset, by the way. We're actually developing our own TBPN over your headset because this is just like, this is the ideal setup.

1:56:26I was telling my friend, I don't care if I look like a dork, my hearing is going. So, like, I get locked in when I have the headset so I can hear everything. No, and it's wired. The worst is AirPods. AirPods have a tiny lag. If you're doing Zoom calls, it just ruins it. We feel like you're right here at the table with it. It's not code red over there. He's Baja blasting. Yeah. You know he's Baja blasting. Anyway, let's, I mean, we have some time. Let's run through, I want to know a little bit more about your perception of Jensen, your perception of NVIDIA, and just set the table for us. We know how, what is his management style?

1:57:06How does he, he has all the direct reports. He reads everyone's like to-do lists every day. They have tons of employees. He doesn't fire people. They never fire people. Like what makes NVIDIA's culture unique set the table for us so then we can go into the opportunities and challenges with that framework in mind. So the first thing I found about their culture is it's very blunt. Like I think in most companies and you guys have done startups, but I don't know if you work for large corporations. Bureaucracy builds a process that gets ossified. NVIDIA is the complete opposite. Like if things are not going well, he'll chew you out in front of the whole company.

1:57:43And that kind of blunt mentality, I think, sparks better performance because you don't want to be embarrassed in front of Jensen's in front of the whole company. But also it just sparks agility. Like when I talk to people at Intel or Google, like the biggest problem they have is meeting paralysis and you need to get sign offs from like five different executives. At NVIDIA, you have a meeting, Jensen makes the call, he seeks out the right information, and you move. So there's agility at NVIDIA. The other thing is just a meritocracy. Even from the beginning, like 30 years ago, Jensen's always asking, who's the smartest person that you work with?

1:58:23Who should I try to recruit? And from the beginning, like Dwight Dirks, who's one of the top people right now, he recruited him because he talked to this other guy, and he said, oh, Dwight's really smart. I really enjoyed working with him. And he's just almost ruthless in a sense. He just goes after them and recruits and brings people aboard. So this meritocracy, agility, speed, and just getting rid of the internal politics, I think, really separates NVIDIA. How has most of the team becoming millionaires affected the culture? if at all. I think they have that thing in the first 10, 15 years, people started getting sports cars and putting them in the parking lot.

1:59:08But a lot of people... Let's hear it. Sports cars. Let's go. That's the best news I've ever heard. That's fantastic. Sounds like it's had an incredible impact on the culture. Job's finished. Okay. You don't need to respond. You don't need to say anything more. I think winning culture breeds winning, and people want to stay with winners. Yeah. Right? You want to win on the track then. A lot of the people I meet, in terms of colleagues, they work for a chip company for five years, and the chip doesn't work out. Yeah. And you just wasted five, seven years of your life. So you want to stay within winning company.

1:59:38And NVIDIA has been winning for 30 years. So it's kind of like winning if it gets winning. You get the talent. And then the talent stays. There's so many top executives at NVIDIA that stayed there for 25, 30 years. And there's some benefit if somebody doesn't have a scarcity mindset, right? And they're just playing to win. They're no longer thinking like, oh, if we can just get to that next mile, and get this secondary sale. And if I can participate and if I can vest my two years and sell into the next tender offer, then they're just like, we're good. All that matters is just being as elite as we possibly can be and doing it for the love of the game, basically.

2:00:21And a lot of these people that, they made it, they could retire. They could have retired 10, 20 years ago, but they stay and still work 80 hours, 100 hours a week because that's what's expected. Even the marketing people at NVIDIA are working 80, 85 hours a week. And that kind of mentality, I think, is different at the companies. Okay, let's shift into the competitive dynamic. I mean, NVIDIA has been – I was revisiting the performance of the Mag 7 since the dawn of ChatGPT. It's been three years exactly. NVIDIA, by far the winner, top 10x on market cap. the next closest company, I think, maybe 4X'd by comparison.

2:01:02And so the clear AI winner in the public markets, the most obvious AI trade that just completely ripped. Now, you know, there's this whole narrative of like, how strong is their moat? What does the TPU mean? Is the TPU going to be significantly competitive? Is there going to be margin compression? How have you been processing this new narrative that NVIDIA might face serious competitive threats because they're so on top of the world that everyone owes them so much money that people are saying, I got to get a discount from somewhere and I'll go to Google maybe. I want to talk about this 10X move.

2:01:41It hasn't been like straight up to the right. There's always been, every three, six months, there's always a reason to sell NVIDIA. The H100 problem, the transition to Black Whale, China, ASIC, Procom competition. So this stuff has been happening this entire 10x move up and the media loves to latch on to the latest thing to worry about. We had DeepSeek earlier this year that the entire media establishment was it's over for the AI tree. AI models have become so efficient when it was actually the opposite because the reasoning models and there was an exponential demand for compute. So I find it amusing like the whole world kind of discovered that Google had a really good chip in the TPU.

2:02:25Which they've been working on for a decade, too. Yeah. They've had it for 10 years, right? They've offered it to clients in 2018. This is nothing new. The Ironwood specs, which I always take with a grain of salt with specs. Even in NVIDIA, they talk about 25X improvement with Blackwell when it's more like... I'm really just focused on the name Ironwood goes pretty hard. It's pretty good. They got some good names over there in Google. Ironwood came out, all the specs came out in April. Like, this is not, all this stuff isn't new. And another thing I want to say is TPUs, their chips, Morgan Stanley estimates there was a huge decline in TPU shipments in 2025.

2:03:07And Google, at Google Cloud, NVIDIA GPUs took more share than TPUs this year. Like, no one talks about this, right? And now everyone's going, Ironwood is going to take over the world and NVIDIA's in trouble. Is that just because we're still early in the Gemini AI? No, it was a Gemini 3 thing. People are like, Gemini 3 is the best model in the world. Yeah. It was the best model in the world for a whopping six days. And ChatGPT is still number one on the App Store. Yeah, but I mean, six days. It was unseated by Claude, which was also Anthropic, which is also TPU potentially in the future. Yeah. And ChatGPT, Microsoft, the head of Microsoft AI Cloud, Tony Scott Grethory that OpenAI is training their next models on the GB300 and BL72 that just went live in October.

2:03:55Actually, earlier today the NVIDIA CFO said it's going to take six months, so I was a little disappointed in that. Wait, six months for the training run? Yeah, well, she said the first models on Blackwell on the superclusters are going to take six months, so it's going to take time. Delay the singularity. It's going to be another six months. OpenAI is going to get there. The Claude and Gemini benchmark gains with pre-training. That's the most bullish thing for the whole AI industry. AI adoption, the scaling laws are intact. Everything's going to work out. And OpenAI is going to get there when they build their next training on the next model.

2:04:34So going back to TPUs, thanks goodness. A shout out to SemiAnalysis. They do the best channel work in the industry. Everyone freaked out on Friday, right? They read that semi-analysis. Oh, no. Total loss of ownership. TPU V7. They're going to destroy it. But people that actually know the industry, it was flaming bullish for NVIDIA. Yes. It was so obvious in my face because Dylan and the semi-analysis, they said, wait a minute. The next TPU V8 is not going to be that great. They lost a ton of people and the set function up in performance is not going to be that great. So you know what's going to be great?

2:05:16NVIDIA's Verirubin, which comes out at the end of next year. So no company is going to switch over. People are saying it's the Rick Rubin of chips. Are they related? Anyway, so Verirubin is going to be dramatically better at the end of next year. And even the Ironwood, which just became generally available and they're ramping right now, no one's going to switch over for one. It's a huge endeavor to put workloads from CUDA, NVIDIA GPUs, and put them on. There's always problems when you put them on a new chip. And let's talk about TPU customers, right? Everyone freaked out that Meta might spend a few billion dollars in 2027.

2:05:58That sounds like a lot, right? That's less than 1 % of NVIDIA's expected revenue. It's nothing. And Ben Thompson was very smart and astute. He's like, who's going to buy the TPU? Who are the biggest buyers of AI chips? They're the hyperscalers, right? So maybe Meta will put a portion of their workloads 1 % in the user revenue. Is Amazon going to buy TPUs? Well, John just asked the CEO of AWS if he was going to buy TPUs. He dodged that question. He didn't say yes. I mean, there's no way in hell they're going to buy TPUs. They have their own cranium. They're not going to support their number one, like, one of the number one.

2:06:36Their arch rival. Yeah, they're not doing that. So is Microsoft going to buy TPUs? You should have been like, yes, I'm going to buy one so I can study it. Microsoft's not going to buy TPUs. They're the number two player in cloud computing. And they're not. Are the NeoClouds going to buy TPUs? Now you're going to say, yes, Google has some NeoClouds. You know what happened with those NeoClouds? They're financially backstopping those NeoClouds. So Google is financially giving money and backstopping the debt for those NeoClouds. So there's a handful of small NeoClouds. but is CoreWeave going to buy TPUs?

2:07:09Probably not, right? Who are the other customers of AI chips? Enterprises, companies, sovereign AI. Yeah, anybody that wants to run a fine-tuned model, some small model, something like that. 90 to 95 % of those. Without a bite. So, like, if you just go down on a first principles basis and look at the customers of AI chips, like, they're going to stick with NVIDIA. The millions of developers There's no CUDA. So you don't have to, you really need, like Dylan talked about this, is you really need like top-notch, like software sophisticated engineers that can like work with TPUs and learn jacks and all that stuff.

2:07:47So most people aren't, don't have those crackerjack engineers, right? So they're going to stick with NVIDIA because everyone's used to NVIDIA. NVIDIA is backwards compatible and forwards compatible. So like 20, 30 years of this stuff. And if you buy it, the CFO talked about this morning. You can use it for training and you can use it for inference. It's all on the same architecture and it's going to work. I talked to an AI startup CEO a few months ago. He tried AWS training. It looks a lot cheaper. Total cost of ownership. But then it crashed. There were bugs. There were reliability. They couldn't figure out what happened.

2:08:28They just threw up their hands. I give up. No one is going to... If you have reliability problems, bugs, crashes, the best thing about NVIDIA is all that stuff has been ironed out over the last 10, 15 years. If you have a problem, you can figure it out. Yeah, it's like giving an F1 driver a car that is unreliable and saying, hey, go race. Go race. Have fun out there. And then it's like DNF immediately. Specs look awesome. It seems great. But then when you actually build your business on it, You put the future of your business onto something. The number one thing, it's not price. It's like, it better work.

2:09:09It better work. And with NVIDIA, it works. Yeah, but react to this idea that Dylan Patel was joking about as TPU is a stalking horse. So this idea that Sam Altman is already saving 30 % on his NVIDIA purchases effectively because just the threat of going to TPU is enough to get NVIDIA to make an investment or slightly discount in one way or another. I don't think that's reality, and I don't think that math actually works because I think he's confusing the AMD deal where AMD gave free warrants to OpenAI. First of all, the deal's not done. It's a letter of intent. None of these deals are done. AMD's done.

2:09:54They signed an agreement where they're giving away a percentage of their company through these free warrants. The NVIDIA deal hasn't been signed yet. And they actually language in the 10Q that it might not happen. Doesn't OpenAI have to buy AMD chips in order to get the warrants? Yes, yes, yes. And so it still could be that they don't actually end up going through with the purchase and then they wouldn't get the X. That's a good point. Yeah, yeah. So the whole, like, let's do a side step here with the circularity and all that stuff. All this stuff, it's like one gig at a time. There's milestones on OpenAI.

2:10:28There's milestones on AMD, technical milestones. They have to achieve certain targets. So all this talk about, you know, everyone loves the big number that adds up five years of CapEx. A lot. Like that, it could get, the leverage could be up or down depending on how things happen each year of the way. So, you know, it might not be that big number if OpenAI doesn't come out with an amazing model or AMD isn't able to hit the milestones they set up for their next 450 MI chip, right? So, like, you know, don't worry about five years. Like, take it one year at a time. Right now, demand is off the charts.

2:11:07Now, going back to the NVIDIA 30 % discount, like, that's not how equity investments work. If NVIDIA does invest$10 billion,$10 billion, up to$100 billion, say that NVIDIA gets ownership of the company. It's not like a freebie, right? You're giving away ownership of your company. So it's not really a discount. You're getting ownership of the company. So I don't really believe in this 30 % discount thing because NVIDIA, Jensen will say they're investing to accelerate OpenAI. And they're looking forward to OpenAI going back to OpenAI. I mean, it's definitely creative. It's definitely a new structure.

2:11:45I'm just trying to, I would steel man in that, like, if I'm an entrepreneur and somebody comes to me and they're like, I'm going to invest$100 million in your company over a series of milestones. You're also going to buy something from me. I'm like, yeah, I'm taking some dilution. But realistically, like, this is way less of a headache. Like, where else was I going to get$100 billion from in OpenAI's case? It's a great source of funding that, yes, it will be diluted, but the whole structure is all diluted all the time because of all the different ownerships. Except for this kind of sentiment thing that we had the last few weeks.

2:12:21OpenAI hasn't had any problem in raising money for venture capital. Yeah, that's true. It's not like NVIDIA is the only source of funding for OpenAI. Everyone wants in. That revenue run rate, it's like$5 billion to$20 billion at the end of this year. So what was your take on the Code Red? What do you think about the Code Red? So I saw that you showed the Ashley Vance interview. Yeah, Mark Chen. I thought it was really interesting. Mark Chen was talking about how they kind of focused a little too much on reasoning and their retraining muscle wasn't there. Reasoning, we could talk about this later.

2:12:59Reasoning is like the biggest kind of accelerant of AI demand in the past year. So I think it's actually really good. And supposedly Ilya was doing the research for reasoning. Reasoning is awesome, right? But they kind of focused on reasoning the past year with 01 and 03. And now they're like, okay, we have to go back to pre-training. So OpenAI knows that pre-training still works because Gemini 3 had great pre-training results and Cloud Opus 4.5 did. So now they're going to do the pre-training. So they have their focus on one thing, and now they're going to do the other thing and make their model much better.

2:13:35I do agree that OpenAI has been a little too maybe diluted. Like they're doing apps. They're doing hardware. They want to compete in AI infrastructure against Microsoft and Oracle. They want to compete in AI chips against NVIDIA. Like I thought it was really interesting. Like Satya repeatedly said, he wouldn't name who he's talking to, but it's like, I think it's important that we realize this is not a zero-sum game and this could be a partnership. That was during the Anthropik-NVIDIA deal with Microsoft, right? Sure. He's got that a couple times. And I think the person that he's talking to is Sam Olman.

2:14:17Right? Yeah, excellent. NVIDIA and Microsoft made OpenAI as successful. They were the partners. Why are you competing with the partners that brought you to the dance, right? Let's go back, focus on making the best AI model in the world, and don't compete with NVIDIA and Microsoft. Maybe five years from now, but it seems a little aggressive to compete with them right now. Let's talk about China. There's been a ton of debate over NVIDIA selling chips to China, legacy chips, older chips. We've gone back and forth on it so many times. What's your current thinking about the best policy for NVIDIA exporting chips to China generally?

2:15:05I think the best thing is to keep it one or two generations behind the current state of the art. Yeah. Like, this is a really nuanced policy that people, you know, everyone's either hawkish or dumbish. Totally. Whatever. The best policy is to keep, I don't want to use the word that Howard Lutnick used that got China very upset and forced. Force China to tell its companies not to buy H20 chips. But the best policy is to get China still on the NVIDIA sack. So NVIDIA gets$50 billion of revenue per year that can help R &D and fund R &D and make the chips even better. NVIDIA and the U.S. already won, right?

2:15:49They had 95 % market share. Like, why are we going to give$50 billion of oxygen to Huawei and all these other Chinese AI chips companies that now Chinese companies that need to buy AI chips are going to buy Chinese AI chips? Like, why not keep China on the NVIDIA tech stack one or two generations behind? Don't give them the best stuff, but maybe one generation behind. I think that would be the best compromise for both sides. But I don't know what's going to happen. What do you think? Where do you place the likelihood that the Chinese market has opened up again at some point in the next 12 months?

2:16:29Maybe 50-50. I mean, that sounds like a cop-out. Like, I was much more positive six months ago. But, you know, this has been just so crazy. First, the Trump administration banned H20. Then they didn't ban it. They said it was fine. But then China was like, no, you hurt our feelings. we're not going to let companies buy the H20. And then maybe Trump is going to let NVIDIA sell the H200 or a Chinese-specific version of Blackwell that's kind of hobbled a little bit. Who knows? NVIDIA needs to convince the Trump administration and then China to buy the chips. The worst part of it is China was willing to buy the H20 and it's just all the kind of geopolitics and hurt feelings.

2:17:16you know, that ship has sailed. So I don't know what's going to happen. But I do think the ideal situation is NVIDIA could sell one generation behind, make$50 billion a year, and keep the competition from Chinese AI startups out of the way. And even understanding that at some point in the future, China's buying effectively zero chips from NVIDIA, but it would be five, ten years in the future? It's like you have to assume that like... Right now it's zero. Go for it. The amazing thing is, sorry, it's 0%, right? And NVIDIA's revenue accelerated. For the first time in two years, NVIDIA's revenue accelerated in this latest quarter.

2:18:07And this is like not talked about enough. This is the first quarter that the NVL72AI server has been available in volume. And then revenue just skyrocketed without China, which is incredible, right? And that's why I'm so bullish over the next few quarters, let's say. Because this product cycle is going to last at least three, four quarters. The key tell is the revenue acceleration, first time in two years. And the other key tell is that the networking segment for NVIDIA was up 162 % year over year. And typically, a lot of these data centers and these neoclouts buy the networking stuff six months ahead of time.

2:18:46So the next six months from now, the GPU numbers for Blackwell and the NVL72 server is just going to be bonkers. It's going to be off the charts. And people don't talk about these NVL72 AI servers. They're$3 to$4 million. They're 72 GPUs, 144 dies, 1.5 tons, 5 ,000 cables. and the prior version was 8 GPUs. So these AI servers, I call it the iPhone 3G moment. Do you guys remember the iPhone 3G? This is big for the Christmas shoppers out there. If you want a gift that's a step up from the NVIDIA way, I recommend picking up an NVL 72. It's only$4 million. Yeah, but for the right 12-year-old in your life or the potential of the internet.

2:19:35I think Tyler wants one. You can put it under the tree, but it could be fun. Keep it in the garage. It's like a Lexus. You put it in the driveway with the bow on top. Yeah, yeah, yeah. That's the way it does. With a fork, you drop the forklift, come drop it off. Exactly. I got you an NBL 72. Enjoy. And then we have the reasoning model thing where exponential compute and companies are actually seeing huge cursor, 40 % productivity gains, Yeah. TH Robins and 40 % shipments. Rock and Mortgage, 80 % reduction in paperwork cost processing. This is like the next year because of AI reasoning, because of the MDL 702.

2:20:17That's why Amazon and Microsoft said every quarter this year, they raise their CapEx. Everyone's like, they're going to cut their CapEx. No, every single quarter, they raise their CapEx. That's because they're seeing the demand. And that's why Amazon and Microsoft are going to double their data center capacity over the next few years. That's crazy, right? In September quarter, there are more data centers leased than entirety of 2024. This is like exponential step function up. And people aren't talking about it. They want to talk about TPUs like destroying Invidia. What about some of the kind of demand guarantees that have been happening?

2:20:57Is that a concern at all? Do you think about it much? Is it? Not really. I mean, demand guarantees, like, you're talking about NVIDIA and CoreWeave. Like, when that happens, analysts every quarter are on the conference call, like, did you use that, you know, demand guarantee? Like, it's not happening yet. Like, CoreWeave's five-year-old GPUs that everyone says are useless are 100 % utilized, right? H100, a massive cluster, before it expires, probably a three-year contract, they got like 95 % of the pricing. This is like unheard of. And the reason is there's overwhelming AI demand. And there's not enough capacity.

2:21:41Overwhelming AI demand, not enough capacity. And people just don't care about what's happening in the real market. This is real-life facts, evidence, numbers. NVIDIA going from 56 % revenue growth to 62 % revenue growth on$57 billion with zero China revenue. I mean, these are bonkers numbers. We talk about the stock price being up 10x. Their revenue is up 10x in like two years. This is like beyond history. The last 30 years of following technology. I love how you're the only person without NVIDIA fatigue. You're just like, you're not bullish. David Goggins of the David Goggins. I mean, they can't keep running.

2:22:26They can't keep running. This is not me just like making stuff up. This is like numbers are there right in front of everyone. You make good points. You make good points. I like it a lot. We'll have to have you back on the show soon. This is a lot of fun. Yeah, let's do it. Let's make this a regular thing. Yeah, this is fantastic. Super fun to have you on finally. Thanks so much. Thanks for having me. Thanks for all your reporting time. The book is the NVIDIA way. Get it wherever books are sold. Get 10 copies. Give it to everyone in your life. Also, give the gift to Finn.ai, the number one AI agent for customer service.

2:23:00Automate your most complex customer service queries on every channel. Up next. Oh, yeah. We can just go straight into our next guest. Let's bring in. We have Tarek from Kalshi with some massive news. Tarek, great to see you. How are you doing? From the office floor. Good to see you. Hey, guys. Thanks for having me. Very excited to be here. You are locked in. Look at that backdrop. Fantastic. Please introduce yourself. You've been introduced. Give us the update. What's the news? Let's ring the gong. Well, we just raised our Series E. We just raised a billion dollars, $11 billion. $11 billion valuation.

2:23:40Great windup. Great windup. Honestly, I was waiting for a gong. Congratulations. For a fucking sick moment. How's it going, guys? Yeah, great to have you on. I don't, I was thinking over the, I don't know if anybody had a crazier Thanksgiving holiday than you, it was, there's a lot going on last week. So nice to come out of that with a big announcement, but yeah, maybe kind of just update us on, I think everybody has been following the prediction market wars. The more important story I think is just how - Some people are calling it a blood bath actually. Yeah, I mean, just like it's been a battlefield.

2:24:17on the timeline. But yeah, I think what's happening in the background is this explosion of this new asset class that, again, I think in your announcement earlier, you were saying a few years ago, nobody really cared at all. And now you and the industry broadly have millions of users. So it's pretty unprecedented. But yeah, what's been the latest on your mind? I mean, I think the thing that's happening right now is personal markets. I think I've gone mainstream. I think every inch of evidence is pointing towards that. And I think that the thing that we're seeing is there's sort of one of these rare shifts in consumer behavior that you don't see often.

2:25:02Like they don't happen. Like changing the behavior of a customer, the habits of a customer is a rare thing. And it's unique. And when you see it, you have to really go after it with all your might. and it's you know there's like a number of things that have to align for that to happen and i think they're aligning for prediction markets i think it's happening and i think there's you know one factor is the fact that people are not really trusting um the sort of legacy media and legacy sources of information and they go to prediction markets to get smarter the other one is that they're legal now you know cashes took on this sort of battle over years to legalize this entire market and set it up as a legitimate financial asset so that anyone can participate.

2:25:43And three, I mean, I think we're all sort of, we sort of caught wildfire this year. I mean, I think we're seeing people, there's a little bit of this phenomenon where you cannot watch a sports game without looking at the Calci odds live and the Calci charts. You cannot talk, debate about a topic, about the future without, you know, uh, uh, talenting somebody to put a position on Calci on the app. So, um, it's, it's a huge announcement. We're very excited about it. Um, and, and honestly, it really feels like we're just, we're just scratching the surface of what prediction markets can be. One thing I've noticed when, uh, I'm watching a sports game is there's sometimes an integration with Calci, sometimes with a competitor, uh, what's actually going on, you know, legacy sports book.

2:26:26Yeah. What, what, what is actually going on? I feel like a lot of people who are just passively observing the time, Timeline are seeing a lot of like announcements and partnerships with the partnership economy, and people are joking about it. Like what's actually going on? What's at stake with some of these partnerships? What have you done? And what does it actually mean? Because it feels like if you do a partnership with a specific league, that doesn't necessarily mean that I can't get odds on that event somewhere else. So what is actually going on with the partnership economy? I mean I'll tell you kind of our approach to this so we are building you know our focus is building on a business it's very metrics driven you know and sort of for context so we're doing a billion and a half of volume a week now and you know we're market leader by meaningful margin I think depending on sort of how you measure it so we're something around 80 to 90 percent market share now and I think any partnership we do we bucket them in a bunch of categories but they're all focus on actually driving legitimate volume and legitimate use case into the product.

2:27:26So our partnership with platforms like Robinhood, Coinbase leaked, it's coming in December. And PricePix, Webull are kind of in that bucket. Then we have partnerships with a series of partnerships coming around news. One of them leaked this morning in the New York Times article. But they're also very... One sentence, ten leaks. Everything leaks these days. Nothing is news anymore. It's all leaks. But the point is, we're focused on things that drive legitimate use to the products and then drive legitimate utility to the partner. And so whether it's a broker, obviously, this could be a big revenue line for them.

2:28:09And if it's a news network, it's a complement to the reporting that actually makes the reporting more accurate. And reporters love truth and prediction markets bring truth. So you can see the synergies and how they fit. Okay. Yeah. Yeah. Yeah. That makes a lot of sense. Yeah. What, yeah, I think the, some of the, some of the big news out of last week is that Robinhood is, is entering and kind of potentially trying to verticalize the product experience on their side. What can you say about the, I guess, like how you see the structure of the market evolving? You guys are an exchange. Robinhood is a brokerage.

2:28:47Sounds like they're trying to actually build an underlying exchange themselves. How much should observers of the industry look to how stock trading and stock markets, stock exchanges have evolved versus prediction markets? What does this market look like in five years, 10 years, as much as you can pull out a crystal ball for us? I mean, maybe the basics is like, and you've seen this a little bit in AI, right? after you see the success we've had,

2:29:22it's basically indicative of like, okay, there's a massive market opportunity ahead of us. And when that happens, I think you're going to inevitably see a ton of competition. And generally in those markets, like this sort of massive surge of competition, whether it's brokers, there's some of the sportsbooks like Draxons and FanDuel coming in. It's just usually a sign that there's a lot of good things to come for that market, right? It's a sign that you have big companies reprioritizing their entire roadmaps to go all in after this. And that's a positive for us. We're a market leader in a market that everybody is starting to believe is going to be ginormous.

2:30:00In terms of the specific question of market structure, we have obviously the exchange. We also have our direct product in some ways is competitive with some of our partners. And I think the same way that we're working with a lot of different brokers, over time some brokers are going to sort of diversify and work with a number of different exchanges. And that's how these sort of market structure evolve over time. And the only thing that matters, the kind of the thing that stands out is similar to any other market is product and product velocity is, are you putting out products faster than anybody else?

2:30:28And are you putting out products better than everybody else? And I think Cache has had a pretty incredible track record of setting the pace in the industry. At least if you look at the last year, we've set the pace in the industry and everyone's following. And I feel pretty good about us continuing to do so in the next 24 months. How do you think about the market structure? I think everyone's wondering, like, obviously, this is a new market. it's unlocking entirely new sort of asset classes. And it's obviously big. Everyone's excited about the numbers, but is this a natural monopoly? Is this duopoly?

2:30:58Like how many winners will there be? How do you even think about the market structure? Is there some return to scale?

2:31:07It's interesting. Like I kind of like don't think much about that. Like I think investors love to sort of, investors do this thing where they're sort of gonna to rationalize all of it in five years you know everybody's going to be super smart about how they like all figured out but like look i i think that like it's a very nascent thing right it's it's it's you know like it's it has some similarities to ride share it has some similarities to the drafting's fan duel era when that happened it has some similarities to the online brokerage industry and it has some similarities to financial exchanges like cme so where does it fall it probably somewhere in between all of these sort of buckets uh and probably not exactly the same as any of these other buckets.

2:31:45And I think that you'll see more of, I think with enough scale in financial services, but also true for any industry, everyone gets into everyone's territory. And so the only thing that matters, again, is sort of what companies are going to rise above the others in terms of product velocity and product quality. And that's just what we're narrowly focused on. There's a question from the chat. Can you explain how external market making works on Calci? that's been a for some reason a hot topic recently but you know market makers are part of any financial market you kind of need them um to basically have liquidity in markets and actually calci and prediction markets have less customer to market maker flow than traditional markets if you look at options for example it's like the vast majority is you know geordie to a market maker like citadel whereas on cash actually the vast majority is you know geordie versus john And then some of it goes to market makers.

2:32:37And it's an open, transparent order book where everyone's competing on price. And we have actually a separate company called Cash Trading that trades on the exchange. But they're a very small percentage of any liquid markets. Really, their function has been, for new markets, a little bit weird. Yeah, liquid markets. Yeah, that makes sense. Because if it's some really, really niche thing, who's going to put in the first$500? Exactly. You take the risk. And they're not very profitable. It's actually, they're really focused on providing a good customer experience that we bootstrap markets. rather than any meaningful part of the business model today.

2:33:10And if we took it out, it'd be actually a worse experience. So I think it's definitely not positive for the ecosystem. But it's a bit like Uber. You know, when the adverse interests got impacted, the taxis, they were coming up with all these reasons, right? Like, you know, about all these kind of random reasons. But I don't think there's much truth to it. Yeah, so I'm sure you can't comment on any specific lawsuit. There's a number of them. But what has been the, I think there's quite a lot of prediction markets experts that have looked at some recent lawsuits against prediction markets and said, they clearly don't understand how this works.

2:33:47Can you comment at all on some kind of misunderstandings broadly? Yeah. So what Cauchy has done is first regulate prediction markets as a financial instrument under this agency called the CFTC. People have been hearing more about the CFC recently because it also regulates crypto. And that's one of the main financial agencies. There's the SEC that does stocks and CFC that does commodities. And then we did the same thing with elections. And now we did the same thing with sports. And the way that it works is like financial markets, those are regulated at the federal level. And so the law around these markets is just federal.

2:34:24They kind of report to a federal government and federal regulator, not a state government and state regulator. And there's a bunch of reasons for that. But, you know, it's kind of how the constitution was formed, which is some stuff makes sense at the federal level and some stuff is more local and makes sense at the state level. And we are one of these things that fall under the federal level. And federal law preempts state law. So if you are OK on the federal side, state law doesn't really kind of apply to that exchange. And that's why we have one regulator, which is the CFTC, our federal regulator.

2:34:52And again, I think it's normal. When something so disruptive happens to an industry, the people that are adversely impacted are going to come after it and come up with all sorts of arguments for why it shouldn't exist or why Airbnb was terrible and all these different things. But at the end of the day, the thing that drives it long-term is, is this a great product and our consumers loving it and using it? And the answer is yes in this case.

2:35:16Off of the success of Calci and Polymarket, there's been a bunch of net new prediction market startups that are created. Is there a possibility that this market ends up having these sort of niche, maybe more like vertical marketplaces? Or do you think that the platforms with the greatest liquidity and the deepest liquidity will ultimately just absorb those sub-markets?

2:35:46it depends on how narrowly we define prediction markets versus broadly like i really think of prediction markets as kind of just like a next general like like expansion of financial markets to touch anything calci means everything in arabic but really if like if markets kind of progressively grew over time calci what we did is just like kind of widen that set dramatically over what it could touch. So I could see some, you know, startups innovating on like specific verticals over time and doing reasonably well, but there is real concentration of liquidity and concentration of volume that happens in these types of markets that is hard, I would say, to battle with.

2:36:24And so I think at least from that aspect, like I think the cards are probably mostly shuffled already. That makes sense. Last question. There was a viral clip of you talking about Donald Trump Jr. Do you have anything more to share on his involvement? Because I was watching that and I was like, yeah, it's kind of like, hey, where are we going with this thing? It seems like politicians have a deep insight on how campaigns use these prediction markets, but can you share anything more about his involvement in the company? Yeah, I mean, look, first of all, that clip is a clip and you know how these clips are taken.

2:37:03Who needs context? Yeah, it's like, we don't need context. Anyways, look, I think that we have done, like one of the main products that took us mainstream was an election market. And that brings a lot of attention from politicians on both sides of the aisle. And you see it, you know, Trump at the time was using his prediction markets on during the election. And actually, Mamdani more recently was using his Calci odds pretty consistently during his election. And so in some ways, like you're going to see a lot more, like prediction markets are going to touch, financial markets are going to touch the news and they're going to touch the political process because they bring more truth to all of the above, all of these categories.

2:37:41And in some ways, it's good that we get more and more, I would say, politicians involved and engage with these markets. The one thing I'll say about this, and again, it's in the same bucket as the other things that we discussed, where there's industry dissidents that are against prediction markets that find all these different reasons for why prediction markets might be bad. But the thing that happened is not this administration necessarily, even though this administration is pro-innovation, is we won that lawsuit on the election market, which has really redefined what the landscape, what the boundaries of what the financial market is.

2:38:14And that lawsuit was won, you know, is in the Court of Appeals in D.C. It's a very progressive panel. It was a panel of Democratic judges where we won 3-0. So people want to make it out to be a partisan issue, even though I don't think truth needs to be a partisan issue. It's just, you know, these markets work. People love them, and they generate a lot of insight out of them. And I think that will win the day at the end of the day. Last question from my side. How does the CFTC view when a market participant has some type of alpha or non-public information and they're betting on a market based on that information?

2:38:56From my view, as somebody who gets data from, we work with Polymarket, we use Polymarket data on the show. if somebody has inside information and they're trading on that information, it actually makes the markets more accurate. So in some ways, as a user who's just like viewing markets, it's I want people that have inside information on global events to be trading so that the markets actually better reflect reality. But what is like the CFTC's view on that type of activity? Because like things get thrown around all the time, insider trading this or that, but I don't actually know what the actual law says.

2:39:38Yeah, that's actually a great question. It's a point of debate in this land. But I think there's some distinction. So CalShare is a regulated exchange. So everything we do, in some ways, a lot of the laws and the rules are very similar to what you would expect in a New York Stock Exchange and some of the traditional financial markets. The question of insider trading is interesting because what you just said could also apply to the stock market. Right? Like if you want to accurately price a stock, maybe we should let it start a trading happen. And the reason why it's actually not allowed is because it makes the game unfair.

2:40:11It makes the market unfair. And if the market is unfair, liquidity dries up. People just stop participating. Right? And that's why you have to have reasonable rules of the road where people can reasonably expect to be treated fairly in this marketplace. where there's no kind of asymmetric or structural advantage for one participant versus the other. And we take the similar approach here. So if you actually have insider information, which is information that like, you're not supposed to reveal to the public, you're not supposed to trade on it because trading on it is a way to reveal it to the public.

2:40:44And so that makes kind of more balanced, more fair marketplace. And I think we're very focused on that. But it's a very interesting question. It's one the industry is battling with, But we take a hard stance on insider trading. Yeah, because if somebody goes and they go and they vote in a local election and they see like, okay, I talked to somebody there and they said they were voting this way. And I talked to another person and they also said they were voting this way. And then somebody trades on that information. Like, is it actually like, is that, you know, how do you define that type of activity, right?

2:41:18It's like anybody could go down to the polling. anyone could go down to the polls and kind of like or voting center and just see ask the same question right so anyways well I was going to say it's the same as the stock market right if you go and sit in front of Walmart and count everybody that's going in and out and then you know during the day and forecast their sales from that that's actually fair game now if you call your cousin at Walmart and ask them for information they have internally that they're not supposed to reveal to the public that's inside of trading and I think we have a very similar line here yeah yeah yeah that makes that makes sense um uh very cool well super helpful um and yeah congrats to the whole team it's a pretty massive milestone huge and uh yeah great great getting the update thanks so much for taking so much guys thanks for having me we will talk to you soon talk soon have a good one adquick.com out of home advertising made easy measurable plan buy and measure out of home with precision our next guest is matt mullenweg from automatic He is in the Restream waiting room.

2:42:21Let's bring him in. They are parent company of WordPress.com, Tumblr. Welcome to the show, Matt. How are you doing? I think we have you on a hot mic. We might have you on a hot mic, hopefully not. Welcome to the show. About to come on the screen? Yes. All right, well, a little drum roll. You're all about to experience TBPN for the first time. So Matt, our audience is live. We're streaming on them. They're streaming on us. How are you doing? We're streaming with each other. We'll clap too. Fantastic. How are you doing? Good to meet you. Howdy, howdy. Thank you so much. I know this is a little nontraditional.

2:42:59So we're kind of like two hours into our big annual address, the State of the Word. It's kind of like our State of the Union speech. But thank you so much for allowing us to connect. A lot of folks in the room have never heard or seen TVPN before. Fantastic. So this will bring a lot of new folks into your world. and I'm excited for some of your world to learn about WordPress. Yeah, give me the state of the word. And then also I want your personal word of the year. We've been debating what the word of the year should be over here. Ooh. So state of the word, and I'll say the state of the word is strong.

2:43:32Okay. There we go. That's good. Let's hit the gong. We're hitting the gong for that one.

2:43:39A strong state of the word. Congratulations. We actually just did a live release of WordPress 6.9. So WordPress does major releases three times per year. We were able to do it right here on stage. We had a little button that we pushed. We got to get it going next time. I love it. That was pretty fun. Don't worry, I didn't just ship it again. But, you know, one of the things about WordPress is it's not just built by one company, but it's a community of WordPress 6.9, over 900 contributors from all over the world. Different countries, different languages, different companies, all coming together.

2:44:14And so that was pretty exciting. My word of the year, and actually a theme we were just talking about, is I'm going to choose freedom. Freedom. Powerful. So as technology starts to influence more and more of our lives, how we travel, who we date, the things we learn, the news we're exposed to, the sort of freedoms that are embedded in an open source license. I like to refer to open source licenses sort of like a bill of rights for software. Gives you an inalienable rights that no company or person can take away from you. And that freedom and agency, I think, is really, really important and something that I think, you know, as technologists or builders, that we should try to embed into everything that we do.

2:44:52Give us an update on Beeper. I was super fascinated by that product. I love walled gardens. I also love tearing down the walls of gardens. It seems like a good shot across the bow of the iMessage walled garden. How's the progress going there? Are you using the service personally daily? Are we going to see a lot of growth there? Well, obviously, I'm using it daily. So I would think of it not as like replacing a wall of garden, but more like allowing your gardens to come together. So I'm sure you, like me, I have friends on lots of different networks. And some of them always love to use WhatsApp.

2:45:34And some of them always love to use, you know, Instagram or LinkedIn DM sometimes. I even get some interesting stuff there. And I hate it when I miss these messages, you know, because, you know, checking all the different apps sometimes, or the notifications, I might miss something. So think of it not unlike how email clients, you know, can bring in lots of different email accounts. Beeper takes all the different networks where your friends already are and maps them together. Now the plus and minus is that it's not going to replace the networks. Like I still keep all the different sort of specialized messaging apps because like, for example, if someone sends you an Instagram story, when you click on that, you're going want to load Instagram, for example.

2:46:11So I think of it as complementary and hopefully even increasing the usage in a very small way right now. It's pretty nascent. But in the future, think of it as like sort of a different interface. So you might still have like the dedicated apps, but then having this all in one inbox that you can sort of manage everything, tag people, have folders, and does cool features like scheduled messaging across all platforms, or even just like weird heuristics that are pretty simple to do. But like, show me all the, don't just show me unread, but show me all the people I've messaged that haven't messaged me back yet.

2:46:41Oh, yeah. Sure, sure. We've talked to some young hackers, some startups who are building, you know, sort of Beeper competitors. And their whole value prop is like, we figured out a way to get it into the iMessage ecosystem. Do you think that we need a new regulation there or some sort of law change or some result to actually open up iMessage? message or do you think that with enough tricky hacking, it can be done? Well, technically, it's not hard. Well, it is hard. It's very possible to reverse engineer these networks. However, as we saw with sort of a previous iteration of Beeper, if the network really, really doesn't want you to do that, it's probably not good to pick a fight with a trillion dollar company.

2:47:30Yeah. So perhaps these things might happen through open source or something, but as a commercial company, I think ultimately you have to be somewhat respectful and try to complement these networks. So how Beeper works today is we don't support iMessage on the mobile or Android. In theory we could, but Apple has indicated that's something they don't want. We do support on the Mac OS client. We have a way to integrate with sort of iMessage using some APIs that are available on Mac OS. And so on macOS, we can bring in your iMessage. Got it. But again, I'm building this for the long term. Yeah. And we are a commercial company as well.

2:48:05Sure. So, you know, we want to work with the networks. And, you know, perhaps there can be regulations like the European DMA or things that can encourage interoperability. But ultimately, I think that the sort of people who run these networks have to see a longer-term benefit for them. and for things like some of the other networks I mentioned that Beeper works with, I think their business model and everything, the increased usage is really useful for them. I think for today, Apple's business model, particularly in the US, kind of the lock-in effect to the device business, which is of course where they make a lot of money from iMessage, probably indicates that, and less forced to, I doubt, they will adopt sort of iMessage interoperability.

2:48:54But who knows? Sort of like they used Lightning for a while and eventually got USB-C and all of our lives got better. Who knows what will happen in the future? Talk about links on the internet. I feel like we're at a point in time where social media platforms are trying to keep users in their own applications so that they can monetize them to the fullest extent. Meanwhile, you have LLMs, which are ultimately doing a lot of the same thing. They're taking content from all over the internet, trying to keep users in the individual applications. It feels like WordPress in many ways is making moves to kind of like almost fight back against that.

2:49:34I might have that incorrect, but I feel like it's important if you're running a business independently online, it's great to have people like on your own website so you can develop a deep relationship with them. But what is your view on that? We're very much anchored around X as a business. Obviously, X has had issues with links or, you know, chosen to demote them in the algorithm over the last couple of years. But give us kind of the state of the union on links. That's a broad one. Well, I will say X is actually a great example. And I've talked to Nikita about this. So they now, they've shifted some of the balancing of links.

2:50:12And they now have this really nice in sort of app browser. So you've probably noticed that now. That when you load a link, you actually still have the ability to like and reblog and everything. and I think that's kind of the future. So I do think that you can have things that are complementary because so much of the great content and everything is more on this open web. It doesn't have to be fully embedded in the app, but that is sort of a technological change. So I would say I actually point to X as someplace where I think things are going in the right direction, although I do agree that sort of time when links got really de-boosted and everyone had to do it as a reply was kind of weird and sucked.

2:50:49So for WordPress publishers, We support so many different types of websites, and different types of websites, I think, might have different motivations. So, for example, a popular plugin for WordPress is called WooCommerce. It's an e-commerce plugin. It actually runs on about 8.9 % of all websites in the world are now running this e-commerce plugin. You can think of it like an open source Shopify. And if you're selling something, a merchant, you just want to sell the product. You might not necessarily care that someone comes to your website to buy it. So some of the new things that are happening in partnership with OpenAI and others where we're allowing products to actually be like browsed and bought inside of the LLM are pretty exciting.

2:51:28I also think that the incentives of these open source chatbots in particular are very complementary to the open web. So, for example, like if you're on Amazon, Amazon really wants you to say or eBay or Etsy or something like that. They want you to say in their marketplace on their system. But when you think of how Google works and sort of the growth of Google in the open web, you know, they have their search pages, but they also would link out. And that was a whole part of their business model and how they grew. We're seeing that with the chatbots as well. And in fact, something I talked about a little bit earlier is that the traffic from bots, both from them crawling, but also user-initiated actions, is exploding and has already surpassed sort of human traffic.

2:52:09And it'll be interesting to see where that goes in the future. So, you know, there's never a better time, I think, to invest in having a domain, but also invest in publishing. And, you know, just like you might have a direct relationship. Like, for example, I suppose I could get like a chat GPT to summarize today's TVPN episode, but it's more exciting to watch it. I think that creators developing a direct relationship and brand is going to be part of the future as well. Very, very cool. Well, there's so many more things that I want to add, but I know you're in the midst of your own presentation.

2:52:39So thank you for tuning in. Come back on soon. And thank you for having us. The view is spectacular as well. It's a pleasure to meet you. I'd love to come down and hang out when I'm in LA next. So thank you so much. Thanks so much. We'll talk to you soon. Great chatting. Have a good rest of your day. See you. Bye. That was a first. Hsleep.com. Exceptional sleep without exception. Fall asleep faster. Sleep deeper. Wake up energized. That's our sponsor. 94 last night, John. I think I smoked you again. I lost my phone. Lost your phone. Well, we have Jason Freed in the restream waiting room. Let's bring him in.

2:53:17There he is. Jason, how are you doing? Good to see you. Good, and you? Congratulations. Massive news today. Break it down for us. What's up? Was there big news today? I missed the news. What was the news? No, you're news. You're just calling out everybody. You are the news. Trello, name and names. He's name and names. A lot of people don't do that. A lot of people say, oh, the competitors, the best in class solutions, the Gartner hype cycle. No, you called them out. you put them on the map. We had some fun, yes. We launched a new product today called Fizzy, which is kind of a fresh take on Kanban, an old idea.

2:53:47Obviously, it's been around for a long time. But we have a different spin on things, different take on things, and felt like it was time to do something new and kind of bring it back to the basics and also add some fun and color and vibrancy, which is missing in the software industry. I feel like the people might be colorful in a sense, but the products are very much the same. And so we wanted to do something different. And that was what we did today. Why new name? Why not a new tab in an existing product? Right. Well, Basecamp, which is our biggest product, has Kanban in it. We call it card table there.

2:54:21But the thing is, is that Basecamp is very popular, but let's say there's 100 ,000 accounts, right? About 100 ,000 companies use it. It's a small number in the end. And there's a lot of people who can use something like fizzy that are not going to use base camp base camp is a much bigger system it's for bigger projects and there's a lot of small things that people need to do and organize and track and so building a small standalone thing just feels like it makes more sense frankly for for this kind of thing so do you i mean do you have an idea of like who is the target market startups individuals like like you use this to plan your thanksgiving dinner yeah i I mean, the target market is me and us.

2:54:59Basically, we build things for ourselves. I don't think about who we're making things for because we're making things for us always. And the idea is that, you know, actually, let me just say this. I find the best products in the world are made by the person who's making them for themselves. That's been my experience, like enthusiast products. And then other people find them and other people discover them. And you find out that you're like other people and other people are like you and they kind of dig it, you know? And so we're always thinking about ourselves. I said to someone this morning that I feel like TVPN is that way, where sometimes when I'm driving home, I want to watch TVPN, but I'm like, we just made it.

2:55:38I just lived it. I should probably watch something else. So I never watch the show myself. Sometimes you just call me and say, hey, we haven't done enough podcasts today. Yeah, then we just make it. We do a podcast on the fly. Then we just talk about tech news more. I'd love to know about the actual process for building the product. Who was staffed on the team? How many people? What time period? When did you start? Do you have a designer, developer? What's the prompt? I imagine you just used one prompt for this? One prompt was all you needed. Yeah. So what's interesting is we actually also open source this.

2:56:15So this is fully open source. It's a SaaS product and fully open source. So you can run it yourself for free, which means you can go into GitHub, actually, and look back at the very first commit about 18 months ago and see everything we did along the way. All the changes we made, all the dead ends, all the starts and stops, exactly who was involved on our team over time. And it's changed. So we had typically we have two designers, one or two designers on something. Then there's other people who chime in here and there, who jump in here and there. Different programmers jump in at different times.

2:56:46But it's fully documented, which is very rare. You'll almost never, ever see this in commercial soft. Basically, almost never. Sometimes, but almost never. Especially going back to day one. What ends up happening is you can do this thing where you can basically, on launch day, you can clear the log, basically. And then from that point on, people can see what you're doing. But we opened it up from day one about 18 months ago. So it's actually all in there. The team size in total, probably about six people worked on it here and there over 18 months. But for the most part, it's usually two or three people working on something at a time.

2:57:19How do you think about pricing these? Yeah, I feel like as in 37 Signals fashion, pricing will be opinionated. So I'm excited to hear how you guys approach this one. You know, we don't really, well, we have a price, but I don't know if it's the right price. Never do. It's 20 bucks a month, unlimited users, unlimited usage. One price, no chart, no table, no contact us. Just a price tag. Like if you went in and bought a pair of jeans or peanut butter, it'd be like, how much is it? Talk to the sales rep. They're going to look you up and down. They're going to say, well, how much should this person pay?

2:57:55Right. What watch are you wearing? All the things, right? So it's$20, but we give you 1 ,000 cards for free. So there's no time limit on the trial. You get 1 ,000 cards for free. And if you never use a card as like a to-do item or something. If you never use them up, it's free forever. and you can also run it for free if you want to run it yourself. It's open source. Yeah. So we're basically just serving as a host. If you want to just turn it on, sign up and be going, we'll host it for 20 bucks. Currently, look, this is an introductory price. We could change the price six months from now. If we do, we'll let people lock in where they were.

2:58:28We're not going to change prices on them, but we might raise it. I don't even know what we'll do, but we want to pick a number that was fair. The other thing I want to do is I want to price this more like an accessory. This is not the only tool. The software industry is interesting because it thinks that whatever it makes, it's the only thing anyone ever needs. The thing is, is people need a lot of different things. And so Fizzy is not going to be the only thing you have. It might be one of the many things you might use. And so we kind of price it that way. It's like an accessory,$20 a month, kind of a no-brainer, unlimited users, cancel any time, no upfront anything.

2:59:00And it just feels like that's the right place to start. We'll see it where we end up, but that feels good for now. If I pay you to host it, where is it hosted? China. No, so it's hosted. We have a few different data centers. So it's not, well, it's in our hardware. Yeah, what I'm getting at is like, it would be easy to just throw this on AWS, but like you're the one company that doesn't just do that. That's right. So we have a data center in Chicago. We have one in Amsterdam. We have one in North Carolina. So we have in a few different spots and it's all on our hardware and other people's data centers.

2:59:39We rent space and data centers. That said, again, you can also, if you just don't trust us, don't want us to do it. You can put on your own stuff, including like a simple droplet, like a digital ocean something, whatever you can find that will host something basic will work for this as well. I mean, you actually can host it in Alibaba Cloud if you want. It's open source. That's the whole point of open source. I could put it on. I hope someone does. There's an AI company that recently had a code red. Oh, yeah. Have you ever had a code red ever once? Not like that. not like a competitive pressure code red let's make sure we kind of focus on this competitor but we've like screwed up and have all hands on deck to fix something.

3:00:22I mean there was a moment I think probably Did you ever learn the like did you ever get overly fixated on a competitor and sort of like learn that because there's like that's like YC like just law right like don't overly focus on competitors like you're probably not going to die as a company because you're a competitor. You die because of, I think they say like indigestion or something like that. Right. Most wounds are self-inflicted. Yeah, exactly. But sometimes you have to actually have a lesson be fully ingrained. You have to learn it the hard way. I'm curious if that was the case. I think there was one time way back when we used to have a product.

3:01:06We still have a new product now called Campfire. But way back in 2006, we launched Campfire, which is a real-time chat, group chat. And back then, we could not shove this down people's throats. Like, nobody understood group chat for a business. It just was very, very hard to sell and to move and was a very small product for us. And then Slack came out, and I saw it. I remember, oh, shit. Like, yeah, they nailed it. Like, we just, yeah. It was crazy because them nailing it was IRC. Like, I used IRC back in the day. and the hashtag channels, like everything, like all the primitives had been like battle tested in IRC.

3:01:41The other thing is Slack doesn't feel like that outside of the world even from a, I'm sure you have opinions on Slack's like design, but it doesn't even feel that, like you guys probably could see that and be like, oh, that's like, like the design was opinionated and, you know, fun. It felt fun. Slack felt fun. I mean, IRC, of course, is very geeky and whatever, but yeah, the fundamentals were there, but Slack had a wonderful onboarding experience. It felt fun. They had great integrations. They just kind of like totally leapfrogged us in that world. And that was like fine, but it did. It was the first time I felt like I felt that sort of nervousness in my stomach.

3:02:21Now I didn't feel it against our business because Basecamp is a very different kind of product and it was fine, but it was Campfire specifically because I was frustrated. I was trying to figure out how to make it better. And then I saw them come out like, oh shit. Like, Yeah, that. That's how you do it. So that was one time. But I just don't think there's any reason to focus on competitors. You can't control them. You don't know what they're going to do. You don't know if they're going to be around in three months or three years. You don't have the same economics as they do. So it doesn't really make sense.

3:02:51Like, for example, I'll take Hey, our email service, hey.com. We have 40 some odd thousand paying customers for Hey, right? Which is, if you were Gmail, it'd be an absolute abject failure to only have 40 ,000 paying customers. That would have been shut down years ago. In seconds, right? But for us, it's a multi-million dollar business because we have 60 people here. So for us, it's a great business. So I can't go, well, Gmail is killing us. They're not killing us. They're doing their thing. We're doing our thing. So I think you've got to, in my opinion, the only person you actually compete with are your own economics.

3:03:26That's not a person, but the only thing you compete with are your own economics. If you can make it work, if you can make it viable, you're fine. If you can't, then you can't. Your cost. You compete with your costs. You're competing with your costs, yeah. Every business needs an AI note taker. What are your opinions on AI note takers? If they join the call, are you admitting them or are you letting them stay? I'm pretty harsh. I always let them sit out in the cold. I never let them in. We don't have meetings. I couldn't even invite one in if I wanted to. We don't do that. But I will say I have been in a few calls recently that other people have set up, and there's been like an AI transcript, and it has been quite handy.

3:04:02It's really pretty impressive when it works really well. Strangely, Apple can't seem to get voicemail transcriptions to work at all. I mean, Apple is just struggling with all the basics on transcription, even just talking to your phone and like, Whisper works. It works in the ChatGPT app. It works everywhere else. Apple just has not implemented it properly. And it's not crazy AI God. Like it's literally just take the words that I'm saying and write them down verbatim. And that's a huge benefit because if you're in a business call, sometimes I just want to search the actual transcript. I don't even need you to summarize it or put action items or go do things for me.

3:04:41Not agentic, none of that. just actually write down exactly what I said so that when I say, you know, uh, you know, we had, you know, when I say AWS or whatever, I can go search for when that happened in the transcript. And a lot of, a lot of companies just haven't been able to implement that. It's been weird. I agree. I think, I think frankly, that is one of the best use cases of, I don't even know. It's not even AI though. It's just, it's transcript transcription software is very, very handy. And I think like, this is the thing, like it's, it's transcription software has been around for a long time.

3:05:10It's gotten better and better and better, but it's not like AI. In other ways, it has been AI for 20 years. It's been the original AI in many ways. It's like throw a bunch of data at it and try and estimate what things are. Even like OCR, these similar things, they're not AGI. They're AI in the sense they're narrow. It's the recommendation algorithm on YouTube or TikTok or in Netflix or this specific, you take a picture of a receipt Does it understand the text in there, even if it's kind of a dark photo? Yes. That's specific, narrow AI. And that's great. But we need to actually get those things working on our phones.

3:05:48I agree. I agree. We left AI out of busy, by the way. We made a conscious effort. We actually had some in for a while and pulled it out and had it back in and pulled it out. I'm just like, I want to remove stupid from this. I don't want to add intelligence. I'm going to remove stupid from the software. So it's just so straightforward that it just works. and you don't even feel like, God, I wish I had AI for this or for that. So V1, no AI. We'll see what happens down the road. Again, it's open source. So the really interesting thing with Fizzy is that there is a world where you can just actually sit back and do nothing on AI.

3:06:20And if AI is real and valuable to your users, they will get it stuffed down their throats via their OS, via their browser, because Atlas is going to be trying to jump off. Perplexity Comet is going to be trying puppeteering their Fizzy. and the rest of the system that they're using, whether it's their phone or their laptop or their desktop, it's going to bring the AI to bear with computer use. And so you might never have to build it. I think so. In fact, this is actually interesting really quick. Recently, OpenAI added a Basecamp connector to ChatGPT. And we didn't even do anything. So they did all the work.

3:06:57And they just sent us an email saying, hey, we're launching this Basecamp connector in a few weeks. Great. I'm like, this is fantastic. We don't have an MCP server. They just did it. And so I just think, to your point, I think more and more of that's going to happen, which is it's going to be available on the OS or someone else is going to do it or whatever. And to spend all this time to build it into the product specifically, I just don't feel like it's the right, the best use of initial. An initial V1 should be focused on the product itself and not the other things that it could possibly do.

3:07:25Again, later on, maybe there's stuff that comes in. Maybe people, via the open source version, submit some PRs that have some AI stuff. We'll see where it goes, but we didn't need it for V1. Yeah. There's just always a question of where the AI lives. Like, do you need to go and pre-train your own model to answer questions? Or if you set up a good knowledge base, will you just get sucked into the next pre-train automatically? And you can just go to ChatGPT and ask about you, and you'll be there. Anyway, Jordan. I want to keep hanging out for an hour, but we do have to wrap the show because we're going to look at a studio.

3:07:57We have one last question from David Senra. Yeah, one last question from our mutual friend, David Senra. Can we get a wrist check? What watch are you rocking on launch day? I might be the only person who coordinates their watch with their software. It's possible. So I'm wearing today, I'm just wearing a, I'll take it off, because I don't know how to quite hold it up otherwise. This is just a vintage Heuer from 1974, which is a birth year watch. Let's see, hang on. Whoa. Oh, birth year watch. Hang on, hang on. I love the, yeah. Hang on, let me see. The focus is hard, but. There you go, there you go, there you go.

3:08:31There we go. I love the orange. Thank you. It's colorful. Fizzy is colorful. Fizzy is full of color. It's the most colorful watch I own for the most colorful product we've ever made. I knew you. I literally said I knew you were going to match. I knew it was going to be intentional. This is so good. It's a little sad. A little bit. It's a little bit sad. It's fun. It's fun. This is joy. It's joy. This is joy. I don't have like a green. Why were you guys wearing a green jacket a few days ago? What was that about? It was Shopify. Black Friday, we were celebrating commerce online. And so we wore solid green suits.

3:09:07No, it's the Shopify color. Shopify's signature color is green. Oh, I didn't even know that. Yeah, Shopify's green. Yeah. Okay. It is a little confusing because we use a dark green in our brand theme. And so it actually paired up pretty nicely. We also have yellow suits for when there's big ramp news. We will wear solid yellow suits. I think I've seen that before, too. Yeah, you've probably seen those. Those are fun. I've seen that. Hard to miss. Nice. Well, Jason, open invite to the studio. We'd love to hang out for like a full hour. Everybody. Just loving it. Everyone's having a great time. Everybody.

3:09:37Let's do it sometime. I'd love to. I think I got an email about that. So we'll figure that out. Amazing. Awesome. Appreciate it. All right. Thanks for having me on today. Thanks for the whole team on the line. Talk to you soon. Very exciting. Thank you. Talk to you soon. Bye. Getbezel.com. Shop over 26 ,500 luxury watches. You're not going to believe it, but this was actually the next ad read up. Fully authenticated in-house by Bezel's team of experts. and we got to close out the show. So I'm going to tell you about Wanderer.com, book of Wanderer with inspiring views, hotel-grade amenities, dreamy beds, top tier cleaning, 24-7 concierge service.

3:10:05There are so many more posts that I want to get to. But we will get to them. There's a new Arena Mag out. You got to go to Arena Mag. Check it out. We are featured in this Arena Mag issue 006, the three martini lunch. We had Julia on the show, of course, to talk about it. But now it's in print. There's a lot else going on. And we will be back tomorrow. Yes. Sorry to cut it off. Thank goodness. We're having a lot of fun. I would be in a very bad place if we weren't podcasting tomorrow, but fortunately we are, so we'll see you tomorrow. Leave us five stars on Apple Podcasts and Spotify. Thank you for hanging out.

3:10:37Goodbye. Have a great evening.