Snap Specs, Taste, Midjourney Hardware | Eric Newcomer, Merrill Lutsky, Carter Reum, Swami Sivasubramanian, Thomas Suarez, Mark Gurman, Ryan Daniels, Isaiah Granet

29:54Eric Newcomer:Here we are. This is a big moment for me. I'm glad. Literally, I'm having Ben Smith record on my podcast tomorrow, so it couldn't have been a better T.F. You can get to the bottom of it. I haven't even dug into this, so it's like you're giving me notes. Material. ask him about you gotta ask him the hard questions what happened with the minor scoop around i you have uh i i really i we were saying that when the show kicked off generational run generational run with uh and and uh i when i listen to your show i i'm like taking notes because like you you actually have a really i was calling you like the matador oh yeah you're with like you're with like a bull and you're trying to get them to like kind of run at you but then they run at you and you sidestep it's really good it's really good and it's just i don't know do you feel like you op you occupy like an interesting space because i feel like you're not some like tech optimist pumper insider right like you have you know what six years in black bloomberg serious journalistic chops uh but then you interact with people who come at you with like crazy anti-tech takes and you're able to sort of like play the pro tech guy in some ways.

31:06Like what is going on? It's very confusing sometimes.

31:08Eric Newcomer:You know, I'm someone who came up through journalism. You know, I was working the last guy in through the normal, like I was in my hometown paper, the Macon Telegraph, Sun Sentinel, Tampa Bay Times, New York Times, just like interning my way up, just doing the journalism thing. Thank God Jessica Lesson brought me to work for the information. I was the first employee. So I was like, oh, the Sun Valley thing. Very interesting. Yeah, yeah. Yeah. And so I was working out of her apartment in Dolores Park. Cool. And then I worked at Bloomberg for six years and started to really get Silicon Valley. And so when I launched Newcomer, I did write that I wanted contrarian optimism.

31:44Eric Newcomer:I felt like I started my time as a tech reporter being more negative than most because that was the era of sort of the gadget blogger. And then by the end of my time at Bloomberg, we had sort of the infiltration of, you know, the political reporter who was really gunning for Facebook or whatever. And so it went from everybody's too soft, everyone's too negative. And I really just wanted to write for, like your audience, you know, people in Silicon Valley, people who got it, didn't want to have to explain who Mark Andreessen and Bill Gurley were and just sort of write for the people. And sometimes that's positive, sometimes that's negative.

32:19Eric Newcomer:Puts me in a sort of weird situation sometimes. So I feel like I'm also having trouble putting you in a bucket as like commentary, interviewer, scoop master, scoop athlete. You get some scoops, but it's not the only thing that you do. Like we talk to a lot of people who are just great yappers and they've never gotten a scoop and that's fine. And they have great commentary and they can talk about anything. It's amazing. And then there's other people who are just great interviewers and every week they're delivering like the best guest. And then there's other people who are the scoop athletes. and you never really hear them give a take or their opinion, but they're like pestering everyone in signal all day long, I suppose.

32:57But is that by design? Is that something you see like fragmenting over time as you build the organization? Like what is there a method to the madness? Or is it just like you enjoy all three things? So you bit down.

33:07Eric Newcomer:It is sort of, you know, you do best what you enjoy. I've never heard anyone call me an athlete before, a scoop athlete. Scoop doggy dog. Future Hall of Famer. You know, we're essential reading. I think the best takes are ones where, you know, you go talk to people who know and then you're like, it's not just like I made this up. It's like, yeah, I talked to people. It's going to be right. You know, I want my predictions to come true. I think the best coverage, you know, doesn't just sit down smart today. You know, you look back and you're like, oh, we were saying Google was undervalued when the stock, you know, wasn't getting the credit.

33:40Eric Newcomer:You know, you want to feel like we were right. Like look back and read the record. That's how you build a reputation over time. So would you say you get financial advice? I want people to make money. I wanted to, I was not, you know, that's not the goal, but I want people to read it and they make more money. And so it feels like it's worth subscribing. But yeah, we do, we do whatever we want. You know, it's, it's newcomer. It's diversity in the content from a personality driven, but you do have other folks on staff and then also diversity of revenue streams, right? Ads, subscriptions and events.

34:09Like, was that deliberate on day one? I feel like a lot of people focus, like, I don't know. It's interesting. It's worked, but it feels contrarian in the new media sense because most people will just be like, oh, I want to be like Ben Thompson only subscriptions or TVP ends only ads or someone's really in the conference business. Or they do that way later. It feels like you were able to land three planes simultaneously pretty quickly.

34:34Eric Newcomer:Yeah. I mean, today, event sponsorship is our main revenue stream with newsletter subscriptions being our second largest. I mean, I did the sub stack thing in 2020 in the pandemic. And it was like, this seems like fun. Finally, something in startup world applies to me. I'll jump on that. And I was one of the top tech sub stacks. 2023, I had not jumped on crypto, right? Like I'm sort of a crypto hater. But 2023, I'm like, I don't know, ChachiBT, this seems real. You know, this is actually the tech trend that people have been waiting for, not something that was like getting ginned up. So I had access to a beautiful office in Hayes Valley.

35:14Eric Newcomer:My friends, now their company is weekend. We put on the first Cerebral Valley AI Summit. And it was just sort of like this worked. People were enthusiastic. You know, Ali Goetzi connected with Naveen Rao, Mosaic ML acquired the company for$1.3 billion. You don't get better press than that, you know. Business Insider put in a beautiful headline, you know. And so we were off to the races. We're literally doing our mid-year Cerebral Valley AI Summit London next week. So ramping up for that. So yeah, it's events, subscriptions, trying to, I'm out, I'm out hustling, trying to follow your sponsorship playbook.

35:48Eric Newcomer:So if you want to put in a good word for people, the newcomer all around, like I'd love it. I mean, the, the, the, or the pre-roll or the mid-roll host read ad in your podcast, I think would do very well. And I, and I, I mean, I, I like that there's no ads when you're sparring with somebody, But you did have the mic drop at the end of the Ed Zitrin episode, which I very much enjoyed. And yeah, if you'd snuck it out in there, I would have been like you earned it, you know, would have been good. Anyway, talk about how you're thinking about covering some of the big stories right now. Like SpaceX is dominating the news, but it's an interesting company to cover because Elon's like one of the first like GoDirect CEOs.

36:33He doesn't do a lot of like management interviews. You don't really like do interviews with other people. Like we have somebody from, we have a VP of agentic AI from AWS coming on the show. They're happy to talk up and down the stack. You don't get that access at SpaceX necessarily. How do you think about what your audience wants to hear from you on SpaceX?

36:53Eric Newcomer:I mean, yeah, they control their own social media network. They can publish directly. Apparently they're not doing SEC filings or something. Do you see because they're not doing conference calls? I think they'll do the filings, but they'll just put it on X. It's hilarious. I mean, for the most part, the newsletter really covers earlier stage startups. On our sort of Friday is like a step back. And so we cover SpaceX a lot there. I mean, all this money is about to flow into Silicon Valley. So just everybody getting rich and that being great for the startup ecosystem is definitely one theme. And who's getting rich and who got in early is certainly important to us.

37:30Eric Newcomer:And then, yeah, we can't help ourselves but say this is like madness. I feel like even people buying the shares know that it's sort of a part psychological stock where it's like, well, is there other money that's going to come in that's going to boost this thing up? And as someone who's fairly bullish on Anthropic and OpenAI to some degree, like SpaceX just feels like total lunacy given the sort of insane TAM story that you have to believe. So, yeah, you're getting some of our commentary that it feels somewhat ludicrous, but trying to root it in the S1 and just keeping it in the numbers. It seems valuable.

38:08Can you do some reporting on where the SpaceX employees are going to be doing their upcoming purchasing activity? Homes? Coffee markets? There's variety. No, there's like a variety of communities where. Are you going to say the one you're thinking of? If you look at them, there's only a few homes listed. And I'm like, I know, like I was there's like a there's a car club. This this this. Oh, yeah. Racetrack. And it's called Willow Springs. It's an old racetrack. They're putting a car club there. I know for a fact that a bunch of them are signing up there. I'm a little bit. I'm a little bit nervous that there's that.

38:48I share too many interests with the SpaceX team. Oh, yeah, you're priced out by these. One less, I guess, serious question. Do you think mocking VCs should be illegal? Because there's been a little bit of some drama on the timeline. Or do you think if you major in venture capital in college, they should teach you about the Streisand effect? Yeah, yeah, yeah. That might be another one. No, but there's been some drama on the timeline this week with VC Braggs, and I'm wondering, you know, Silicon Valley has a lot of power in Washington right now, and I'm wondering if now is the time to push to make it just straight up illegal.

39:27Eric Newcomer:And the scourge. Like, it's just, yeah. You know, we tweak venture capitalists. Of course. I guess I'm humorless on this. Yeah, I don't know. BC always needs to be better at taking criticism. Sometimes you're up, sometimes you're down. It's, like, better to be in the mix. In some ways, the SaaS helps you clean up your message a little bit. So I'm always enjoying that. Oh, sorry. S-A-S-S, not S-A-A-S. It's like SaaS. I mean, that does get to something else, which is like, do you get AI fatigue like we do sometimes where we're like, oh, like the Snap Spectacles thing we spent 20 minutes on at the top of the show.

40:10And like, is it the most consequential story, consequential technology? It's an$8 billion company. It's a new consumer product, but it's something that's tangible that at least you can dig into. How do you balance what is tangible and interesting to the reader with where the market cap is actually moving, which might be at the Frontier Labs or the Databricks of the world that's a little more abstract?

40:36Eric Newcomer:I feel so lucky to cover venture capital because you get to cover interesting stuff. That makes a lot of money, right? I could write about oil and gas and it's like you'd fall in dollars, but it's like super boring. So to me, venture is already pretty exciting. Like writing about AI, you get to talk about artificial general intelligence and like whether it's sentient. And it has real like business consequences. So I don't really get tired of the AI story. I mean, certainly AI is dominating, you know, basically all the investment activities. So it becomes a huge focus of ours. So, no, I'm less likely to get dragged into the Snap Spectacle stuff.

41:16Eric Newcomer:I mean, it's always fun to watch for new hardware. But just this question of, you know, are workers getting totally disrupted? Will, you know, change the course of American business? We have these enormous IPOs. So it's one of the more enjoyable times to be a business reporter where the stuff you want to write about, which is a new intelligence being birthed by Silicon Valley, matches, you know, the market caps and what's interesting. And so following the money for the most part has been a super gratifying story. And so we tend to stay pretty focused on that. How are you answering that question of job displacement?

41:51Obviously, there's economic numbers. There's also just sort of reverting back to your own personal experience. Are you seeing competition on Substack from slop farms? Definitely. Tim Ferriss just wrote about this a little bit.

42:05Eric Newcomer:There's one guy clearly ranked above me who's using AI to write the stuff, which is just fascinating. And where does that leave? Have you grappled with what the response is? Because it feels like the scoops are very, very protected or AI resistant. but maybe history or breaking down or analyzing, taking a press release that's already out there, synthesizing some data in history, that's maybe a little bit more commoditized. But how are you grappling with it? Yeah, I mean, I think coming from a perspective, being a person, so transcending just a writing style, but it is literally someone in a publication is accountable, is defensible, but certainly as the writing quality of AI gets better and better, I'm not ruling out the possibility that it's super disruptive.

43:01Eric Newcomer:Yeah, like you said, scoops. I think the sense that the information is coming from real humans, that I'm in the business of knowing what other humans are thinking is fairly defensible. but you know we're using AI to build you know like our ticketing platform for our events one of my reporters was talking about using it just to build his own like story tracker I think we're trying to be creative like I'm I don't want sort of AI written stuff in our stories but I do sort of have existential questions about you know like the fifth sentence in the third paragraph yeah ai puts out a better version than you were going to spend the time to do is it really a disservice to readers that you went in and like punched that up and i you know i think it it's probably good for readers so it's hard to have this super hard line of all human written stuff because we certainly want it for it's a great copy editor you know i pay a human copy editor but still i run uh stuff through claude and it's amazing how much it's catching both you It'll catch million-billion type errors, math, and obviously all sorts of minor typos.

44:14Yeah, I'm imagining that. I feel no sadness about you never writing the line, they declined to comment ever again. It is fine. It can be a Python script. It doesn't even need to be Gen AI. It can just be an if statement because that line is in so many news articles.

44:35Eric Newcomer:We can have ones that's like take out all the perfunctory lines of stories that they feel inclined to put. And you can, you know, read both things or something. When you think about disruption for like a newsletter or media business, however you want to describe it, the only person I think would actually be able to compete, like you weirdly need like a network. You need like trust. You need like you need there's so much like information that I would say informs your work and and other other journalists that that run similar businesses. That's just like not anywhere online. Like it truly only exists in, you know, a small like group chat or places like that.

45:19And so the question is like, I don't think a fully, I don't think AI can like actually just like somebody with AI that is able to build up the trust and the network and the kind of understanding of the space. Yeah, that person is like, if they're willing to just use more AI than you and like they don't care about like the slop allegations or whatever, then maybe they have some type of advantage. But I still feel like more than ever, there's so much information that you need to run a great media business that is not anywhere on the internet. It's not in podcasts. It's not on X. You just have to go actually talk to people.

45:57Eric Newcomer:We play nice with R4Rock. But honestly, it's people in the industry giving it away for free. you know, the desperation for basically content marketing, meaning that some people do a good job of taking sort of their data and using it to get attention. Like you think about, I don't know, I don't know, ramps AI spend data or something. Yeah, like that stuff, you know, used to be sort of the fodder of like a Wall Street Journal type story. And so to some degree, I think everyone being smarter that they need to have actually like great content to get people to pay attention. is one of the vectors of attack.

46:37Eric Newcomer:Not attack on media, but just competitive. Yeah, like walk me through the ramp AI spend is like an example of like going direct in the sense that they publish it themselves. They tweet about it. They go on podcasts about it. The ADP jobs data just gets released on a newswire and then contextualized by the Wall Street Journal. But if ADP all of a sudden hired some really charismatic employee to go on podcasts, that would be like the same effect. And a lot of, you know, the old way of doing it would be you try to embargo it to a couple outlets. And so they felt like they were really getting something.

47:10Eric Newcomer:So then their readers feel that this is a differentiated thing that they're receiving incrementally more reason to subscribe to that publication. And so it used to be a dance that you'd basically be giving the media sort of favor and sort of unique access that we sort of lose over time. I have a question. Do you, 10 years from now, would you, can you imagine having, you know, 10, 20 journalists on staff joining forces, you know, sort of re-bundling? Or do you like being unbundled? I mean, you know, I have, you know, Madeline Renbarger, Tom Doton, Jonathan Weber writing for me now. And they're sort of a core part of what Newcomer is today.

47:51Eric Newcomer:And certainly, you know, with unlimited resources, part of what I want to spend the money on is great journalism, good reporting. It's sort of, it's amazing that, you know, most cynically, like our content marketing for events is profitable, generates subscription revenue, people demonstrate real value. You know, it's like you, the journalism provides sort of a great way to remind people about the amazing events we're putting on. And so I think it's super synergistic. So, yeah, I'd love to keep building up a newsroom, but I don't really believe in a sort of rebundling of sub-stacks. Like part of it, we're all building separate brands.

48:28Eric Newcomer:You know, Newcomer has its own brand proposition. And a lot of the other sub-stacks people like just are sort of, you know, I love Lenny is like crushing it, right? But it's a very different thing focused on product managers. And so if you even look at tech and business sub stacks, we're so different that I've never really been a believer that there's going to be a rebundling in that sense. But some publications like the Free Press or the Ankler or myself get to a level of scale where we become start to look more like publications. Yeah, it certainly didn't really happen on YouTube. Like there are a whole bunch of different creators and Mr.

49:02Beast has a whole team and offshoot channels. But with the Mr. Beast brand and the promise and the editing. but it was never like, oh, Mr. Beast is going to roll up other creators because that's not the way the setup is working. Talk to me about Mafia. I think we agree that we don't need other venture capital firms with game shows.

49:22Eric Newcomer:Don't copy TVP and do a new thing. Yes, and I love that for that reason. But do we need a game show with just journalist competitors, like an American Ninja Warrior with just sub-stackers? What do you think about that pitch? Or Nitro Circus. Nitro Circus would be good. I'm not building my business around it. I'm happy to go on. You would participate? Yeah, yeah. Sign me up. But as you guys know, the challenge of media is sustaining it, right? Sure. Like, I mean, and kudos to Mike Solana at PirateWires. Obviously, he's delivering day in and day out. He gets it. But I do think for, like, show concepts, it's fun to get excited about the idea.

50:01Eric Newcomer:But the business is in the long term. Like, is there variety? Do you have something great? day in and day out. So some of these things, it's like, yeah, when you have, you know, the top people you'd want for a mafia show, it's exciting. But is there like fifth episode with, you know, the C tier founders is much of a sort of much must watch. But that's what's interesting about mafia is that it's not a new company. It's not an enterprise that needs to sustain. It can be just a fun thing, a fun thing that goes as long as it needs to doesn't overstay. It's welcome because it's not this self-perpatic.

50:36Because they have$60 billion in SpaceX coming to hire some camera crews if they need. That is the nature of what's unique about it. I'm glad they did the production.

50:45Eric Newcomer:I haven't watched. I have not watched the full thing. Tom Dotton on my team wrote the piece. But I do think the traders is so exaggerated. And there's a reason for it. I'm not saying that I could host it that way. But there is a reason that Hollywood and the TV business is so exaggerated. And I think they still need to up it even more, the sort of drama of it all. Yeah. Yeah. I saw some people like chirping at the fact that like it only had 100 ,000 views on YouTube. And I'm like, that is insane for tech. That's insane. And like who, like the guests, like who knows these people? There's only a couple.

51:29There's only like a couple thousand founders that matter to a top tier venture capital firm. There's only a couple of founders that a couple thousand founders that even know who these people are. Of course, like some of the contestants were bigger names. But but by and large, these are insiders around the table.

51:48Eric Newcomer:They weren't going to take Founders Fund's money, but now that Trey is great. Yeah, they're in. Yeah, yeah, maybe. How how are you thinking about the evolution of the tier one? It feels like the tier one VCs of when you started Newcomer are not the same today. This is spicy. This is the question of all questions. I mean, I want to say, first of all, I'd put Sequoia, I guess, number one to answer the question. I think I wrote a piece. Well, I'm not even telling you to give me. Name them all? Name them all. Rank them. But you're welcome to. This will be a scoop for us. This will be a scoop for us.

52:35Eric Newcomer:Who do I think are the top tier? No, no. I just mean like to me, there's been a very, very material shift. And no one has like really said it out loud. Like there hasn't been. I think you'd be in a position to actually do this because you have to go talk to the top 100 founders. under$10 billion market cap or below and say like at series a you basically Like make them choose who they would pick every single round done this right? And and there's been there's been some like random there's been some like random kind of like surveys and like Games that you can play to try to get at it But like I think if you actually had it like you were I've seen this idea on Twitter like and people who is the firm or who where's the real switch or who do you think?

53:23Eric Newcomer:The swap would be. I just, I think that there's a very, like, at least when I talk to founders, like founders want to work with not necessarily like other founders, as in like a founder who had an IPO, but they want to work with the founders of the firms. And so there's like a new generation of like, you know, the Thrives and the Green Oaks and things like that, where I see them as like really, really. Thrives is dominating. They would certainly coming up soon on my list. Yeah. And like, and, you know, when you get in a situation like hyper competitive, like Series A, Series B, like those guys are like really sharks.

54:10Right. and like where do they, you know, fit in? The fall off of Arthur Rock, not R-Fur Rock, but the original Arthur Rock needs to be studied. The guy funded Apple, Intel, absolutely insane portfolio, goes out investing with Bernie Madoff. He gets outed for that in 2009. There's a firm that really fell off from the top tier. Did he just get screwed by Bernie or he was? He got screwed. It's not even that bad, but I don't know.

54:41Eric Newcomer:I mean, how much do you think returns are going to drive this, right? In the cool venture story, like getting in Anthropic right now, or obviously making much money on SpaceX, it's like you look at Spark doing the round that people underestimated. That's clearly been really good for them. I mean, Lightspeed dumped a ton of money in. And Menlo, you know, I think it was key to them. So there are firms where I think it's had a big impact on their reputation, but I don't know if that matches. I think that's helpful. I think that's helpful with LPs, but like the reason that it's an interesting question is like, okay, who has the best overall brands, right?

55:18Who has like the most like halo, which firms do founders actually want to work with, right? Who has like the most access to capital? Like all of these things factor in. Who did like, who did the - Elo, like if you - Yeah, yeah. Between these two firms, which one are you picking? Yeah, like people that bet on Anthropic when it wasn't - It's a completely different thing from actually measuring the quality of the venture capital business that is being built. And people still conflate two different goals. One is the sexy seed round to multi-billion dollar exit in a very short amount of time. Cursor is a great example of like, it's just classic venture capital.

55:58You got in sub a billion. You were out at 60 within three years. It's just incredible venture capital.

56:06Eric Newcomer:Right? Incredible venture capital. But then there's a different game where that investment returns$3 billion in three years. And then there's a different one where you put$10 billion in a lab at$500 billion and it doubled. And so you made more money. And at the end of the game, these venture capital firms are businesses. The job is to return cash flow, like to create cash. And so, yeah, oh, this fund had to raise more money. That's not really a critique. That's a critique of like, what's the definition of venture capital? Should we include growth stage investing in that? Should we include your public business in there?

56:41But as a financier, as an entrepreneur of a financial company, all you care about is how much cash are you returning to your investors and to your GPs? And so there's two very different games that are being played. One is for clout and credibility, and the other is just how much raw cash is this business throwing off? What's the value of the firm? Yeah, and I look at it in the most simple view. I think about a series A or a series B and a founder gets 10 term sheets, hypothetically. And then you run that and you ask, you know, 100 founders that, like who comes out in the top three. That to me is like the most interesting because that's like a generally going to be a leading indicator for like who has the most aura, you know, kind of.

57:29The aura list, the newcomer aura list. The newcomer aura list. The newcomer aura list.

57:33Eric Newcomer:We did Founders Choice a while ago, which was these guys. Elo, right? That was more like, you know, it was a little more like founder friendly. Yeah, who do you want to work with? We've published, you know, an LP did a survey of managers. So we tried. There's so many different ways to slice it. Yeah. And yeah, I mean, getting great returns, I think it has some impact on, you know, your heft with founders. Founders like winners. People are making a bunch of money. But I agree that you can't just look at returns and say those are the ones that are going to get picked. There's also an interesting dynamic, particularly with Sequoia, where for a long time they had a reputation as being like not particularly founder friendly.

58:11I mean, it is the genesis of Founders Fund is like the whole PayPal drama around Sequoia. Right. And it's like, what if there was a firm that was founder friendly? And then that went off and happened. And now, you know, obviously there's an entirely new team there. It's a new strategy. but I talk to founders who are like, yeah, they're not friendly. And I'm getting in the gladiatorial arena with them. Like it's a lion and I'm going to ride it like, and they're excited about that. And they're like, yeah, if I don't put up insane returns, they're going to fire me. That's encouraging. Like I need that.

58:41I need to grind harder. And so there's this weird thing where if you're measuring just on like, who will have your back unconditionally, there could be some really terrible venture capitalists out there. Who's not going to help you, but also never going to fire you. and they're just kind of like yeah i gave you their dumb money but so there's a lot more to the shape of like the thing but that adds aura too because like you can be i mean we had andrew reed on the show and he said sequoia won't stab you in the back we'll stab you in the front i was like that is high aura from their lips and like i i mean i'm not in a position to raise money from sequoia right now but like i was like i kind of want to work with andrew like after that that's great like i like that and that is a good good defining line there anyway i don't know yeah i

59:23Eric Newcomer:I think good founders like tough investors. Totally, yeah. But yeah, I agree. Competition and capitalism forced all these venture funds to realize that founders were their main customer and therefore they needed to say they were super founder friendly. And I think Benchmark's doing a good job with the rebuild, but having Travis and Emil running around being like, hey, screw founders over. It's like Ev Randall was drinking beers in college when that happened, okay? let's back off a bed truck for a little bit let's not visit the son the sins of the father on the son right in the bible um on on this list i you know the competing with the midas list or doing some other list or like how valuable are lists in terms of like actual moving the needle for your business like early on at tbpn we did the the medus list which was like a list of ai researchers and it converted a decent amount of email subscribers because we had like an email capture flow.

1:00:24We weren't like directly monetizing it, but it was an interesting experience. We did a couple of those like sort of stunts. But I'm interested to know from your perspective, is that less of a priority? I always hear that like scoops get people to pay. They jump the paywall after a scoop. They got to read it. Whereas an opinion piece might not, even if it's going viral, people might just summarize it.

1:00:47Eric Newcomer:We did a survey with wing of top venture capitalists to rank up and coming companies. And that does really well. It's like great marketing for us. Everybody on the list tweets it out. But we don't really put much behind a paywall because it's sort of meant to be read. And so then it's not a very good subscription driver. It's more just like brand building awareness influence. And so really the challenge of any list is are you going to put it behind the paywall? And for the most of the time, you wouldn't because you want it to get shared. And so I think they're good for brand and like list building, but not great for subscriptions.

1:01:25Eric Newcomer:Whereas putting like documents or behind the paywall, you're going to get this resource of companies that you might invest in or whatever. You know, behind a paywall, you know what you're paying for. And so people are just much more likely to convert. so you know when that LP I mentioned earlier did a survey about venture firms you know we put that behind the paywall I think that did very well so if we're willing to put it behind the paywall they can do well but for the most part the list you're talking about you know we keep in front of the paywall I feel like there's still I feel like a lot of people ascribe the the value in the Midas list and what Forbes does to just the fact that it's a list and everyone wants to be on it and it has like legacy and brand and that's real.

1:02:10But underrated is how incredible the photography is in a Forbes feature. You can hire a great photographer and still not hit that level of like magazine cover look. And if you offer that to someone who's going to win the list. If they understand media and the power of a really high quality photograph and a photo that will live on and sort of immortalize them for maybe a decade, you see a lot of people and their top Google result is still that cover that they did when their company hit unicorn status or something. And you're like, wow, they look great there. And then you meet them in person and they're like, oh, that was a decade ago.

1:02:47You've got some gray hairs now. But offering that as a trade is an interesting proposition.

1:02:54Eric Newcomer:Yeah, and we literally had like two photographers running around our last Rural Valley event. So I couldn't agree with you more. Like getting great photos of these people out of these events has a lot of value. Something that looks different than just the LinkedIn profile photo against the wall, something that has action, background, texture. Like these things are harder to recreate than people think. They take it for granted because they show up everywhere, but they're usually because of a special thing. Nothing's higher status than the Midas list. is the Alami stock photo watermarked Getty Images watermark at Sun Valley.

1:03:29That is the highest status symbol in Silicon Valley, in my opinion. You can't buy it. I just have one more question. Are you going to voluntarily switch to API-based billing for your AI usage? Because you were getting some...

1:03:44Eric Newcomer:I'm corrupt. Yeah, yeah, yeah. You said AI was useful. This is the most narrow call-out. And Zitrot has this view that journalists shouldn't be getting monthly subscriptions because they're falsely subsidized. And we're not really paying what it costs. We need to be paying per use. I don't think I'm that much of a hyper. I think Claude and OpenAI are getting a better deal out of me probably than I'm using. I'm not sure I'm getting this arbitrage. Totally. Yeah, you can run those prompts of like estimate the number of tokens that I've used over the time. And like a lot of people run that and they're like, wow, I've been paying$200 for like 20 bucks worth of tokens.

1:04:25Like they're printing off of me. I think I'm getting had, not the other way around right now. Yeah. Is AI useful? We'll figure it out eventually. We'll figure it out. We'll figure it out together. Thank you so much for taking the time. This is a ton of fun. Let's do it again soon. So happy to do it. Well, I'm glad we saved this for, you know, today. It's been really fun. We'll talk to you soon. Have a great rest of your week. Later. Goodbye. Let me tell you about Codex, since we're talking about it. Codex is the most tasteful workspace for, you know, powerful workspace for getting work done tastefully with AI agents.

1:04:57Whether you're tastefully writing code, tastefully analyzing data, creating tasteful content, or automating tasteful business workflows. Codex helps you move projects forward from start to tasteful. Usually. Usually. Up next, we have Merrill. This is a scoop. Getting a SpaceX employee on the show is difficult. Not many people can do it, but we can because we got Merrill. How are you doing? Rocket man. Doing great. Rocket man. It's good to be back. Dweet, I want to go way back in time. Did you ever have the, I want to work for a rocket company as a kid, I want to be an astronaut? Because you sort of wound up pretty close to it.

1:05:39Oh, yeah. I mean, it was, I don't know, I remember watching like that. Like, do you ever watch that movie, October Sky? Oh, yeah. I love that. That was like, that was like one of my favorite movies growing up. I definitely had like a NASA phase and, uh, yeah, so it's, it's kind of craziness. Somebody on Hacker News the other day commented, like, it's funny how adding stack and trying to add stack tips to GitHub has resulted in the graphite team now building rockets. So here we are. Yeah, it's remarkable. It's almost like a side quest in the long arc of history. You'll be deploying on the moon and Mars beyond.

1:06:13How has it been? What has the last couple of weeks been like? Has it been sleepless? Are you able to continue growing the product, building the product? Or has all the news been infiltrating you? What's the psyche been like over the last couple of weeks? Yeah, I mean, it's been it's honestly been really exciting on the ground at Cursor. Like everybody is everybody's fired up about about like the compute partnership and what that's unlocking for us on the model training side. We also just had our first big conference compile at Fort Mason yesterday. So the marketing team was really working hard to get everything together for that and make that a great event.

1:06:52And yeah, it all kind of led up to yesterday being a crazy news day of the SpaceX announcement. And then the three big feature releases that compile yesterday, including Origin, the Agenda Git Forge that we're working on. So yeah, it's been a frenetic past few weeks, but everybody's really fired up. Did you guys expect the announced, like the sort of second part of the announcement between, you know, the deal between Cursor and SpaceX to get as much like attention as it did? Because I was seeing people that seemingly hadn't seen the first, the announcement of like the first part of the deal. And a lot of people felt like they were processing it for the first time, which kind of makes sense in hindsight.

1:07:34But in our little bubble, I was like, great. Like, obviously they were going to do this deal. Yeah, I sort of had the, I guess I had it in my mind that you already had a SpaceX badge, but that was not the case because this was a partnership and now the deal has been announced. Is that correct? Yeah, exactly. There was the announcement of the option to acquire and then the other, the announcement yesterday was the exercising of that option. And now, I mean, I still don't have a SpaceX badge yet, you know, where we have to wait for the deal to close and then we can actually start working together.

1:08:06But now it's like just antitrust and all the other pieces. That makes sense. Well, congratulations. I do want to talk about origin. Walk me through the thesis there. I mean, there's been plenty of gripes about how much code is getting pushed to the internet generally, how teams are working together, agents are working together. We are in a CPU crunch now. How much of this is just taking a developer workflow that already exists, speeding it up versus reimagining it from first principles? Walk me through the thesis for this product. Yeah, I think the origin really started from a simple realization, which is that every single piece of the software development lifecycle that we have today, every tool, every piece of infrastructure was all designed for a world where humans write every line of code.

1:09:00We've tried to scale that as best we can over the past year or so, but we're really seeing many of those pieces just crumbling under the pressure. First and foremost, just having, from first principles, designing something that can handle the scale of Agenta coding, that's the most important thing. Is it scalable? Is it reliable? Is it performant? Can it handle hundreds of agents working in parallel at once? And that's really what we've designed Origin to be, is the first example of Git infrastructure that is really built for the agentic era. And the other piece of this is that then building on that, if we own the infrastructure, we can start to build it with agents in mind.

1:09:48So it's not just, it doesn't have to be just like the Git objects. objects. It can be, you can start to store agent traces. You can see the whole history of every action the agent has taken, everything it referenced. And you don't have to have, it's kind of silly today that you'll use one agent to write the PR, but then as soon as the PR is opened, like that agent just goes away. It's like, why does that not carry through and help you then, you know, address review comments, resolve CI failures, fix merge conflicts. And really, we want to get to this vision of more like full self-driving PRs that can really bring themselves out to production when in many cases without human intervention.

1:10:29Yeah. Yeah. That's interesting. Yeah. I have a bunch of different things there. I guess the, the question is, yeah. How, like, how important is speed in this context? How important is actually like the volume of code? You see a lot of these, like someone will post like, okay, I just vibe coded this. I pushed this change. Whatever model they're using did really well. They got the job done. And then somebody will go in and look at the PR and be like, 100 ,000 lines of code to change the color of the button? Like this is evidence of slop or this is evidence of poor computer science happening. do you have a stance on the is the answer to slop more slop are we going to get less sloppy more terse in the writing is that something that's coming or do you need to design a system where the new normal is yeah changing the color button the color might be 10 000 lines of code change that's fine let's also store the reasoning trace and the whole and the whole you know back and forth and everything that the agent was doing so that you know yeah it was it was you know 10 years of development time for one small change, it's fine because, you know, hard drives are cheap.

1:11:49Yeah, I certainly don't think that the future is 10 ,000 lines to change a button color. But I do think that the volume, fundamentally just the volume of code and the throughput that we're seeing and all pieces of infrastructure have to handle now is just growing exponentially. And, you know, we've seen now, a lot of teams are now doing like, thousands of pushes per hour. And previously they were doing tens or hundreds. And that's a lot of what we've been trying to make sure Origin is able to handle is that scale. So we've run some simulations recently where we're doing like 80 clones and 22 pushes per second with no downtime.

1:12:32And it's obviously just, this is just like us testing internally, but it shows that like the architecture holds up and it's scalable and it's not, when your agents are ready to work at that speed, origin is not going to be the bottleneck for them. But I do think that we'll then need to, the other thing I think this will enable though is like more agentic review, more back and forth, like more autonomy and more like thinning out of changes and refining of them before a human ever has to look at them. So hopefully we can avoid the case where you have like a 10 ,000 line PR that's doing a meaningless change.

1:13:09Yeah. How is growth overall? I feel like a lot of people watch an acquisition happen or a series of acquisitions happen, and they're sort of like, okay, that's the end of the story. Or they, I mean, it's really easy with Elon Musk to look at it and be like, okay, well, this is an incredibly talented team, but this is data centers in space stuff. This is mass driver on the moon project. Like the main mission is now, you know, not relevant. And so it's less focused. But I imagine that you're just seeing growth in the core business still because teams are still, you know, getting up to speed on pushing code and using all these tools.

1:13:52What has growth been like post-acquisition? What have been the big drivers? What does that tell us about the way just the broad software development market is evolving? Yeah. Yeah, I mean, we've seen, since the acquisition, we've seen continued growth in the Graphite business itself. But really, I think the most exciting thing for us has been seeing the level of interest from companies of all sizes in what we're doing with Origin. So we set a goal for end of July for waitlist signups that we wanted to hit, and we hit that in the last 24 hours since we announced the product. So, it's... That gong is also for the acquisition closing, which is maybe a bigger deal than waitlist time.

1:14:38But honestly, traction on an actual product is probably more relevant than any financial milestone because it means that people are actually interested in this stuff, which is good. Every time we've talked about it at Compile yesterday, on customer calls, it's like, Compile got a huge round of applause. It's so obvious this is a problem that every single company is facing right now, and it's solving this burning need. And that's why we're so excited about what we're doing. And really, thinking to your point around acquisitions, thinking back to the Graphite days, one of the hardest lessons that we learned from building Graphite was that your product is only as good as the infrastructure that it's built on fundamentally.

1:15:21And it might seem obvious, but it's a really hard lesson to learn. And we were always limited by latency, availability, even just UI extensibility. Our customers felt those limits. When our infrastructure was down, we were down too. And it's funny, now there's this whole wave of companies that are going to their favorite agent and saying, copy Graphite's pull request page, make no mistakes, and shipping something out there. And they're really quickly finding out why it's so hard to do what we did. And even then, I think we were never really able to create the level of experience of just totally seamless, high-performance interface that we wanted to because we were limited by the platform that we built on.

1:16:08So with Origin, now with the backing of Cursor and the combined team, we're finally, you know, we're now able to truly realize and accelerate the vision that we had from the beginning of the graphite days. Let's talk about AI doom, job losses. Is there any risk that AI puts DJing out of businesses? Are DJs going away in our AI future? Or will the DJ, the humble DJ, always have a place in our society? What do you think? I think there's always, you know, it's always been about more than. you've actually always been able to just press play on a pre-mixed and yet we still got like actually like a really good white pill for just jobs overall yeah it's like yeah you want to see a human up there and and even the most yeah even the most the most well-known djs like they have like a set that they could just go up there and play and dance around and of course some of them have been accused of that but uh anyway uh great to catch up.

1:17:11Congratulations on the progress. Thank you so much for coming to the show. Yeah, really happy for you and the whole team. What an inspiring story. A true overnight success. That's right. 2020, baby. Overnight success of six years. Yeah. A lot of sleepless nights. But thank you so much for coming on the show. We'll talk to you soon. Great to see you, Merrill. Congrats. Congrats so much. Let me tell you about Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security.

1:17:39And our next guest is here with us. Live in the TBPN Ultra Dome. We have a room from M13. How you doing? What's happening, guys? How you doing? Welcome to the show. We're doing great. I'm great. Give us your review of the World Cup so far. Ooh. Oh, that was definitely the best game in U.S. soccer history. You know, historically, the U.S. has always been competitive, but just could never find the back of the net. So to find the back of the net that many times, that was nothing short of spectacular. I had a – I turned it on. I'm watching for, like, not, like, really focused on the game. I think I was like in the middle of something, but I was like, wow, we cannot get the ball on Paraguay's side, like for the life of us.

1:18:19And I realized like I had the I had the Paraguay's flag look like our jerseys. So we were actually playing quite well. We were we were pressing them the whole time. But I didn't know. And I was like, oh, wow, this is amazing. We're doing great. Yeah, hopefully. Hopefully the U.S. gets excited. Like we hosted the opening night of the World Cup the night before at our house. And I was texting people going, hey, we're hosting this event for U.S. soccer. They're like, what's coming? I was like, the World Cup. The last time I was in the U.S. was only 32 years ago. But yeah, we'll see. Come a long way.

1:18:51Do you track like economic impact from World Cup? Are you reading into it? Some people are saying like that's the reason why we're seeing unemployment so low in a time of AI disruption. It's like there's a lot of hospitality hiring. Do you read into this stuff? I don't do a whole lot. I mean, it's interesting both between the World Cup and the Olympics, right? They always put out these studies. And the question is, are you losing a shit ton of money or who's making the money? Yeah. I can tell you the car service in Seattle, they're definitely making money. They told me it's going to be$2 ,500 for a 30-minute trip when I land for Friday's game.

1:19:24So, you know, there are definitely people making money. Whoa, whoa. Surge pricing? Yeah, exactly. Surge pricing of average proportion. $2 ,500. But it's a great economic lesson. That was what was happening in Vegas for F1. It was the same thing. A couple thousand bucks just to have a car take you from your hotel to the event. It was crazy. Anytime there's huge demand. So you're thinking more long term. Take us through the basic thesis for the fund. Some of your historical investments. Obviously, it seems like you're not thinking quarter to quarter based on, you know, unemployment rate data this month and whether or not the World Cup is causing unemployment or to decline or spike.

1:20:02But what is the thesis for the fund? How long have you been doing this? I mean, we started M13 10 years ago. We just celebrated our 10-year anniversary. Congratulations. Where did the name come from? Everyone thinks we're El Salvadorian gangsters. Okay, yeah. MS-13. Sadly, I'm not that gangster. But brightest cluster of stars, brighter than all the stars that comprise it. It's always been our philosophy. We can connect the dots and create more value.

1:20:25Eric Newcomer:Got it. But, yeah, you know, fast forward 10 years later, you know, we've been the Cedar Series A investor in 17-year-old corn. So sometimes better to be lucky than good. You know, on days like last Friday when we invested in SpaceX at 15 billion and banged out, you know, 150X, you know, we all need more days like that. That's incredible. Thank you for the sound effects. That was well deserved, I think. Yeah, yeah. Yeah. And, you know, I think, you know, even as we at some point we'll start talking about AI, everyone thinks like we haven't seen this before. And I think the key is as a VC, you have to learn how to invest in cycles, right?

1:21:00Like we felt like the world was white hot in 2020 and 2021. Turns out it's even hotter now. Right. You think about a lot of us like M13 and Drees and all of us, we are investing in consumer software, social media, things like ride sharing. That was the implication of the technological breakthrough, which was the iPhone. Right. And so there's always kind of these ripples. And so, yeah, now, you know, we got 35 people, offices, New York, San Francisco, L.A. Always trying to think, what is that next ripple? What is that implication of that technology? How do we invest where the world is bending? And, you know, I think SpaceX is a great example.

1:21:40You invest in outlier people who could create outlier companies. And you hope you take enough shots on goal that sometimes, you know, you get those big outcomes. SpaceX, Ring, Matterport, you're not afraid of hardware. Where'd that come from? What's your thesis on hardware-enabled technology companies going forward? Yeah. I mean, like take AI today, right? Like my housekeeper knows it's going to change the world, right? That is not, that takes no skill. The skill of a VC is to understand where to get the best risk and probability adjusted outcomes, right? And to do that, you got to find the right value layers, right?

1:22:16So yes, foundational models are enormous. They also come at high prices with a lot of competition. The thing I talk about all the time these days is if you think about the last cycle around consumer software, social networks, it was innovators competing with innovators, right? Zuck competing with Evan. John Zimmer competing with Travis. I would argue in this cycle, it's innovators competing with innovators who are now competing with the most well-funded innovators on the planet. Google, Microsoft, Amazon. Well, even before then, you got OpenAI and Anthropik, right? Yeah, yeah, yeah. Then you have the 10 biggest tech companies in the world.

1:22:47Yeah. And I would argue for the first time in history, they actually have the advantage, right? They got the tech, the talent, the data, the capital, and the technical expertise. Google entering AI feels very different than Google entering social networking. For sure, right? Very different. And so when you think about this market, you really got to think long and hard about where do you get that exposure to AI? Where are the value layers, right? We invested in a company called Prepared in the very early days. It sold earlier this year for about$900 million. It was using AI and technology to disrupt 911.

1:23:19Thank you. 9-1-1 call centers, right? They had no competition. Hopefully not disrupt the call centers. Not disrupt the actual practice. It turns out 9-1-1 was still using landlines. The rest of us just weren't calling using a landline. So, you know, simple stuff like that. So in this market, right, it's all about like where, and I think in venture, right, the multi-stage funds play a different game than someone like us, right? We invest out of a$400 million early stage fund. We're playing for alpha, right? And so we're pure play seed Series A specialists. And so you got to be disciplined about how you enter, what the competition is, things like that.

1:23:59So where does the capital come for doing SpaceX at$15 billion? Obviously a fantastic investment, but not in that sweet spot of seed Series A, right? Yeah. Yeah, our first fund, to be fair, was kind of stage agnostic. But I think a lot of it is, look, all the letters are fungible. The way we think about it at M13 is there's only two types of companies in this market. those trying to find product market fit and those who found it who are hyperscalers, right? And so at the end of the day, you're looking for venture scale returns. And then, you know, the thing that we talk about all the time is everyone just says venture is risky and they kind of throw up their arms.

1:24:32That's like kind of true. But if you're trying to colonize Mars, that is riskier than if you're trying to get someone to order a black car on an app, right? So there are different spectrums of risk. And so it's always trying to figure out, is the risk worth the reward? And that's the game of venture. Have you been disappointed in the growth of the L.A. ecosystem? Yeah. I mean, I laugh because I remember people telling me that San Francisco was dead three years ago. Turns out it's a little less dead than expected. The way I describe L.A. is right now it's having a siesta, right? Take a little nap, kind of like in Spain, between 1 and 3 p.m.

1:25:15We can't forget, you know, about$450 billion of enterprise value was created outside of some of the big ones like SpaceX in the last cycle, right? And so you think about this cycle, right now we're still talking about the technology, AI. There's no doubt that's going to come from the technology sitting in San Francisco. But, you know, you guys know other industries. When you think about the second and the third wave, there's no doubt in my mind some of that will come from L.A., right? Nobody understands media, content, creators. It's why you guys are sitting in L.A. You're the best in the game at that, right?

1:25:46And so I think it's too early to count out L.A., just like it was too early three years ago to count out San Francisco. Yes, there were a lot of needles on the street, and it was pretty, right? Everybody was leaving. But, like, yeah, it came roaring back. So I think as you think about what are the new business models we're not even thinking about because of AI, what are those ripples? There's no doubt some of them will be coming. Yeah, it's interesting. I am I'm bullish on the South Bay, El Segundo, South Bay broadly continue to be bearish on on like L.A., actually L.A., the city. Like it's hard for me to imagine.

1:26:26I can't imagine a hundred billion dollar company coming out of Venice, Santa Monica, West Hollywood, places like that, which is it's painful to me, but I think it's the reality. would love to be proven wrong. Hopefully I haven't passed on that company. But yeah, it's interesting that like, it feels like the South Bay is like really working. El Segundo is really working. You know, SpaceX starting in an El Segundo warehouse. Like that's all working, but I feel like the original like LA startup movement can't even claim. Silicon Beach. Yeah, can't claim that. Yeah. I don't think you'll find a$100 billion company coming out of LA.

1:27:04I think you might find some 10 to$20 billion companies And the reason I say that is, you know, the$100 billion company, the trillion dollar companies are going to be foundational models and things like that. Concentration of talent. Yeah. I think you'll find the application layers. Like there's a reason Snap is here. There's a reason Tinder was founded here and stuff like that. I know, but it's painful that like Suno, Suno should have been an L.A. company. That should have been an L.A. company. Yeah. Like, and it's, and it's what, like Boston, New York, maybe they have an SF office. I forget what Mike said.

1:27:40But I think like the dream works of AI, like it has to happen in LA maybe, but it won't be 100 million. Yeah, maybe. But yeah, it's hard to imagine it as like. Everyone's ciphering stuff from LA. Like New York is the home of like the on the street, what do you do for a living TikTok or like the person that's playing the drums and streaming that. Like so many of the, like the internet democratizing content and then the competition from Atlanta, Toronto and, you know, international. There's the big, we were just reading about the director of the Avengers, hugely successful filmmaker, obviously in the Disney portfolio is moving to London because they're doing the next round of Marvel films out there.

1:28:19And so, yeah, we just lost like a whole pool of talent here. By the way, Jordy, I was going to tell you, I'm very jealous that you live in Malibu, since my house burned down in Malibu. Oh, I'm so sorry. I just missed that ocean breeze. In last year or 2018? Last year, yeah. Last year. Brutal. I told my wife the night before we went to bed, she was really panicked. I was like, sweetie, we literally live across the street from the firehouse. I'm like, not far from Nobu. Trust me, our house isn't burning down. Obviously, we were wrong, and we're still the lucky ones. A lot of people were way less lucky.

1:28:49You're in Loslos. Man, I missed that beach breeze. Are you going to rebuild? Yeah, I get to start this fall. Okay. Yeah. So we'll see you in 10 years. Yeah, exactly. Yeah, hopefully the permitting goes faster or something. I don't know. It seems like a rough go also in the Palisades. How are you thinking about consumer investing, like consumer products? Yeah, you know, like two and a half years ago at our LP day, we talked about how we thought for the first time in probably four or five, six years that we thought AI would make the application layer more interesting because we've been doing a lot of the underpinnings, right?

1:29:24We invest in the future money, work, health, commerce. So that is a pretty high swath. But we are doing mostly infrastructure stuff. I think we've been right on that. It's hard to get the application layer around AI. Just things are so frothy and moving so quickly. I do think, I don't know if it will be for us anymore, just given the size of our funds. I do believe the next billion-dollar D2C brands will be built as AI companies that happen to sell a product. The reason guys like us did so many DTC things and then stopped doing it was at one point the math made sense. You could create unicorns of the math.

1:30:01As the cost of Facebook acquisition increased, the math no longer made sense. I think for the first time in history, the math will make sense here for the next two to three years. The reality for M13 is we need generational companies when you operate out of a$400 million fund. So I don't know if it will call our heart. Even if you're doing a true seed and you can do two on 10 and then it's billion dollar exit. Like you're like, OK, it doesn't actually can't return the fund. Yeah, that's a problem. Like most people just don't think about the math adventure. It's great to say, oh, I had this exit or that.

1:30:37But yeah, you need at M13. We always say every check in the base case has to return half the fund. So we need three to five billion dollar companies in the base case. And we have to hope, you know, we have companies like OpenFX doing stable coins or we need to find that next SpaceX and you need to find those decacorns. That's the only way the math works. Sure. What's the team structure like? Two billion under management now, 400 million dollar fund. That feels like enough room for some principals and VPs around the table. But how what's your philosophy to growing the firm? Yeah. So my brother and I founded the firm.

1:31:15Obviously, the way we talk about it is we wanted to build the venture firm that a founder would have built for themselves. So every single person is an operator by background. In fact, we only have one person that's ever worked at a VC. That's either crazy or that's by design. In our model, basically some of those operators now invest for a living. And so that's the typical venture model. We say they're accountable for company picking. Right. And obviously we've historically been good with 17 unicorns at that. But they're trying to find where's the TAM, where's these great founders. We have a large team called our Propulsion Team.

1:31:52They are former operators who actually don't invest for a living. They actually stay as operators. Right. But now instead of operating over a single company, they operate over a portfolio. And, you know, we've backed, I think, 11 previous unicorn founders in the last two funds. These aren't people that need help, but where we invest at the early stages of Cedar Series A, they're not going to have a head of data and a head of talent and a head of product. And so we can use these very experienced operators. We have guys like Carl Alomar, who has built DigitalOcean from pre-revenue to a$5 billion IPO.

1:32:24He now reminds me, I think it's$19 billion today, so he tells me he stopped talking about$5 billion. But having a guy like that be able to fill in the gaps, look around corners for you at that Series A is incredibly valuable. So yeah, so we have about 12 people that deploy the capital, you know, a handful of partners, like you said. So, you know, I think for us, we obsess about our junior talent because those are our rising stars. You know, too many VC firms are a bunch of usually males that have had success and kind of hold on. I think you need the best 22-year-old kid from Stanford who did comp sci.

1:32:59I think you need the best 32-year-old. You know, I'm going to find the next unicorn founder because I probably had dinner with him last week, but I'm probably not going to find the 22-year-old kid who just dropped out of Stanford and did that. It's probably going to come from one of the junior people on the team. And so in everything we do, we always say we didn't want to create a better VC. We kind of wanted to create a different type of VC. And so we take cues from the best of the best and then think about, again, if we were a founder building a VC, how would we build it? How do you think the SpaceX IPO is going to impact the luxury real estate market.

1:33:33Oh, I am so fascinated by this. I love your thoughts, man. Well, I'm just looking, I'm thinking about YC and how there's like, there's a lot of new centimillionaires and there's not a lot of inventory over there. And knowing SpaceX people, they like to be outdoors. A lot of them probably like to ski a little bit. And I think there might be a supply demand mismatch, at least at YC. but we'll see. Maybe we should, your house in Malibu, my house in YC, we try to get a premium for it, get some Anthropic or OpenAI stuff. You can sell your house for stock now. Yeah, seriously. But I mean, I think it's going to be absolutely fascinating, right?

1:34:15Because you think about when OpenAI and Anthropic IPO, a bunch of VCs are going to make a ton of money, but those people's investors are endowments, institutional pensions, things like that. When you think about SpaceX, right, it is every guy in YPO. It is every guy at the country club. It is every guy at Burning Man. Even if they did 150X like we did, they 10X, they 5X. You think about it, there has never been this much value return that was literally 80 % in SPVs. I think it's fascinating because some guy just made 3 million, some guy just made 100 million, some guy just made a billion. Not to mention the team specifically.

1:34:58Yeah, and the team. So when you think about it, like when you fast forward. I saw a whole family with SpaceX IPO shirts. And I was like, I don't know that they're like bankers. I don't know that they were at the IPO on the inside. I think they just like got in early, made some money, and now they're really proud and they're part of like the retail army essentially. But I think it will be good for venture. I mean, I think there's no doubt planes, houses, private schools, country clubs, all of those things, right? And we should talk about coupling our houses together. We could get a premium. You get two for one.

1:35:26I mean, that's an efficient trade.

1:35:27Eric Newcomer:The starter pack. The starter Centillion to Mayor pack. You got Malibu and YC all in one transaction. But I do think it's just good for venture in general, right? Because people have realized these things just take a long time, right? Like we invested a 15 billion and still took 10 years to return capital. So I do think it shows people in success, right? Although these things take a long time, that is a lot of money coming back to the system. Yeah. Any last questions? We got to hop on with ABS in just a second. No, let's do it again soon. This is great. And, yeah, well, feel free to call in from a World Cup game or something like that.

1:36:03All right, you guys going to any of them? I don't have anything scheduled. Nothing scheduled yet. Who knows? I may get FOMO at the last second and try and get in. Yeah, we'll see how it progresses. Is it as crazy as the Knicks tickets? Is it$100 ,000 to go? Or is it more reasonable if you just want to see the game? More reasonable-ish. More reasonable, okay. It can get crazy. I mean, I'm sure if it's like, you know, America. Well, you're telling everybody about it. Yeah, exactly. Now the prices are going to go crazy. Buy and write during the next commercial break. Anyway, thank you so much for coming.

1:36:30This was fantastic. Thanks, man. We'll talk to you soon. Talk soon. And we will tell everyone about Figma agents. Meet the canvas. Your AI agents can now create and modify your Figma files with design system context. Very excited to be partnered with Figma. And we are also excited to have Dr. Swami Siva Subramanian from AWS. He is the VP of Agentic AI. and he is here with us in the TBP and Ultronome. How are you doing? What's going on? I'm great. Hey, thanks for having me. Thanks for hopping on. Why don't we start with a little bit of an introduction since it is the first time you're on the show.

1:37:06Tell us a little bit about how you fit into the AWS organization, what you're working on, and then we can go into Agentic AI. Yeah. I mean, I'm the vice president of Agentic AI. I've been with AWS now for like 20 years. created various things from our database. Hold on. What was it like 20 years ago? I feel like AWS is barely 20 years old. Were you the first person on the team? You and Andy were hanging out and racking servers? Kind of. Actually, literally, AWS would fit in a single conference room, if you can believe it. That's amazing. I mean, I joined as an intern. and my internship project was to build a database called Dynamo.

1:37:52No way. I mean, that turned out to be a bit of a success, so they gave me a job back. That's great. And then I started the database, joined the database team, started a bunch of databases, then moved on to analytics and started AI for AWS. And now that you see all the things with Bedrock and SageMaker, or all the AI tools. Now, as the world is moving more agentic, I kind of spun myself out of what I was doing to focus all in on AI agents. So that's kind of my background. Yeah, I imagine that, you know, the typical database business, the analytics business, all of that is growing, but it's being driven by more agentic AI use.

1:38:37Talk about the new releases, the new launches, and I want to know how do we compare and contrast, continuum from context? How do companies use the right tool for the job here? Yeah. I mean, first of all, we live in some amazing times with AI agents because the art of possible has fundamentally changed. But when you look at it, even where you see actually enterprises are struggling today, first of all, on putting agents to work, they are still stuck in this world of, I call it, the wall gardens. where they are stuck into bottlenecks, where if I go on vacation or even when I go to sleep and wake up, I get signals from like my Outlook to Slack to various other messaging tools.

1:39:21So the time I spent to even catch up used to be like crazy, like at least an hour before I know I get to my job. So why was it that case? It's not like we didn't have AI assistance in every one of these tools, But what was missing was it didn't have the context to fill in the spaces between these various tools. So that's why today actually we launched something called Quick Autonomous Agent. It actually works on your behalf. It gathers all the context across tools. So there is no more walled gardens. But it still has something enterprises love. It has built-in governance and security. That's why the likes of GoDaddy to NBA to everybody is using it.

1:40:04We were showing it live today and the art of possible, even when we demoed it live, people couldn't believe it because now you can actually catch up and generate a PowerPoint deck in a matter of minutes to a detailed dashboard when you talk to your data leak and generate insights. All these things are changing. That is one example of bottleneck we are removing. The second one I talked about was in the world of security. Security is a big area where intersection of how software is getting built and security is changing in a big way because, yes, the rate of change in how you can generate code is becoming very, very cheap.

1:40:47But also the risk on security is also becoming big. So speed and security are almost at a trade-off in every organization. So what we launched today is a service called Continuum. In many ways, the approach behind it is make security not a discrete thing companies do. It's like how antibodies actually fight viruses. Always on continuously fighting is the way we thought about it as well, as an example. Yeah, I want to dig deeper into those continuously running AI agents. I think when people think agentic AI, they go to coding agents because everyone's had the magical experience with those and seen the impact and the spend and the token maxing that happens when you deploy AI agents across a huge workforce of knowledge workers, heavily software engineers.

1:41:40But I'm more interested in the less understood agentic workflows that are happening, maybe deeper and autonomously in a business. And I'm trying to get to a world where I have that case study at the tip of my tongue for every time I book a flight, it runs an agentic workflow to make sure that my seat on the plane is correct. Or I don't know. I don't know how to contextualize, but I know it's happening and I know you're seeing it. So walk me through some of the agentic workflows or agentic AI work that's happening, even if no one shows up to the office. It's a company holiday and there's still tokens being spent.

1:42:20What does that look like? What is the shape of that work? Actually, I mean, that is one of the key things that we are going to start seeing explosion of agents in a big way. Even today, for instance, in your summit, the CIO of Southwest Airlines talked about how they are building agents on top of actually AgentCore, which is our core agent platform service. And they were able to build these amazing new capabilities to do things like crew planning and various others. These are not like where humans are just generating codes, but how that's where they are going. 3M, for instance, are using agents to now be able to do more efficient sales call planning.

1:43:06GoDaddy is able to actually leverage some of our quick agents to automatically reduce something like 15 ,000 hours of manual work that they are able to do. These are like examples of agents, again, not created just for purely creating software, but these are agents running under the hood and they are able to produce value. And this is going to be where the next big agenting workload is going to happen in many ways. And that's what we are excited. What is demand like for model routing these days? AWS is fascinating because you've sort of, maybe it's brilliant by design, but sort of the Switzerland of AI in the sense that there's partnerships with Anthropic OpenAI.

1:43:52There's a whole bunch of other foundation models that are available on AWS. What is demand like? Obviously, different teams have different preferences and cost tradeoffs, and they might say, I want to be with this particular Claude model or this particular GPT model. Other companies might just say, I never want to be caught using a frontier model for something that doesn't need frontier spend. So optimize it for me. Is that something that you're seeing demand for already? Yeah. I mean, first of all, I think for the industry itself, when I started Bedrock, one of the things, we were different compared to everybody else's.

1:44:33We always had this belief no single model will rule the world. This was not a popular belief at that time, but now it seems so obvious. But then, amazing thing is, now that you're in Switzerland, as you actually speak, in Q1 alone, the adoption for our Bedrock Generative AI LLM platform, what we are starting to see is in Q1 alone, the amount of workload, that request rate that we are seeing is greater than all of the previous years combined. That is the amount of growth we are seeing. And it seems to have no signs of slowing down. It's because the rate of actually adoption, especially as people are moving from proof of concepts to production, is just accelerating because they are seeing real outcomes.

1:45:25And that is one of the reasons why so much of innovation is going into things like Bedrock and Agent Core, not just on the model front, but also how we are supporting the things like disaggregated inference and all the HNT capabilities such as launching new context. So these are the capabilities that's going to make actually enterprises move from just cute prototypes that are sitting in where they demo to their execs to actually production to see real value. And this is why I do think there is so much upside in the future. How do you think about AWS as a thought leader for enterprises rolling these features out?

1:46:09I'm just thinking about we're going through this transition of there's a new technology. You need to use it effectively. I feel like a big piece of why AWS was so successful was because it powered Amazon.com. And if it's good enough for Amazon.com, it's good enough for my business or whatever I'm doing. And it feels like now there's an opportunity for AWS to sort of set the agenda on what, it's not token maxing, it's not token minimizing, it's token optimizing, it's ROI maxing or something like that. But how are you communicating to your team, to the teams within AWS about how to use AI effectively?

1:46:47Sometimes you just want to get everyone up to speed, just go and test everything. And then sometimes you want to be, you know, setting a tone that a company that says, hey, look, we don't have a lot of free cash flow here to throw at this new hot technology. If we're going to use something, it has to deliver ROI almost immediately. How are you carrying that through your organization and into the companies that partner with AWS? I mean, it's a great question because if you look at the industry and this history even at Amazon, we first started with everyone should get familiar with these tools. So we were trying to get everyone to adopt so that they know how to use these tools.

1:47:24Now that they have become so familiar, we are now making all the tools available between like KeyRaw or Agenting Code, coding solution to Cloud Code to Quake as our business user productivity solution. Within Amazon itself, tools like Quake is used by hundreds of thousands of people, let alone KeyRoad to many others equally on the software development. Now we are come to a point where we internally, and this is true for our external customers such as Delta and Southwest and many others as well, is we give organizational breakdown. Within Amazon, we say, hey, every VP under that team, how much usage is happening?

1:48:05What is their cost breakdown? And then we roll it out. This is true for Kiro. I'm pretty sure it will be true for Quik. And within Bedrock, which is our Gen AI platform service, we actually integrated with AWS Cost Explorer so that we can actually have people spend. I mean, monitor the spend in terms of where at a per user level, per model level, and a per organization level too. The moment you have visibility, then you know where people are spending. Then you can understand ROI. In fact, the interesting thing, what we saw was that my teams, and there are a few teams in Amazon, I call them frontier teams.

1:48:47These are teams that don't see 50 % productivity improvement. they see 10x to 20x productivity improvement. And when you see their bill on token, it's not actually like$100 ,000 a month. It's actually, even in the Kiro credits, it's something like around$2 ,000 to$3 ,000 a month. That's not that much for 10x to 20x productivity. But I do think this is going to keep going as the world becomes more and more agentic because once they moved to coding to we just launched our release agent because you increase code velocity. Everyone talks about coding, but this is the myth. Suddenly then everything gets stuck because you need to deploy it to real world.

1:49:30And deploying it to real world is a bigger challenge because you can't just increase the volume by 10x here and the deployment is going at 1x. That means everything becomes a technical test. So we made that H &T. And then we made pen testing H &T. So I almost jokingly call it saying, like, hey, you need to write it correct, ship it fast, and constantly keep it modern. And the only way to do it is you need to have all of it be done by agents. And that's exactly how we are now moving towards. That makes sense. Well, thank you for coming on the show. Congratulations on the launches. And hope to talk to you soon.

1:50:07Have a great rest of your week. We'll talk to you soon. Thanks again for having me. Of course. Great to have you. Cheers. Goodbye. Let me tell you about Cisco. Critical infrastructure for the AI era. unlock seamless real-time experiences and new value with Cisco. Our next guest is already in the waiting room, so we'll bring in Thomas Suarez from Raven Renaissance. He's the founder, CEO, and CTO. Put you on your face. Introduce yourself. Tell us about what you're building. I'm very excited for this. Cool. Hey, guys. Yeah, it's great to be here. Yeah, so I'm Tom, the founder of Raven Resonance. We're a team of engineers and designers who have been building and wearing customized AR devices for years and decades.

1:50:47So we're working on Raven Prism, which is the first ambient computer in the form of glasses. Okay. Did you launch this on Monday knowing that the specs announcement was happening? Was that somewhat intentional or just kind of random? So it was actually more of a preview. So we are doing our full launch later this year, but we were going to be at the conference anyway, and we figured it would show some of our hardware and software. Yeah, because it's Augmented World. that's a conference for augmented reality generally. Do you like that term or is that term getting stale at this point? I think it's interesting.

1:51:21Augmented reality is kind of a loaded term in some ways because there have been so many devices over the years. But I think for us the term ambient computing feels really right because it's about technology being there when you need it and gone out of the way when you don't. And I think that's kind of what this device is for. I wore Google Glass for many years And, you know, of course, there's a lot of baggage that comes with that. But there are lots of wearable computers that have tried these sorts of things over years. And I think finally the industry is starting to converge on something that makes a lot of sense for daily wear.

1:51:56Yeah. How are you positioning the product? What are the key stats that you're particularly proud of? Everyone talks about price, weight. Well, maybe even before that, like, I want to know about your product. but like in your view over the last few years as you've been using these products developing them like what is what is the ideal product what does it do what is like should we be thinking we were debating earlier does like do we need to be thinking about like killer use cases like specific things that people do every day or do you know a lot of the specs demos there were some some of the things that were being demoed I was like okay that would be cool like I'd maybe do that once a year, right?

1:52:38Like measuring like a room, like how a piece of furniture would fit in that alone is like, even if it's 10 times better than using like a measuring tape or something like that, that's not going to get me to like, you know, be a daily user of something like this. So like, how have you been thinking about like, what is the ideal product? Because I assume that's what you're building towards. Exactly. Yeah. So as someone who wore this for over 12 years, I can say that are not, not this particular product, but of course, you know, wearables in general. Home traveler. I wish. I wish. There are two main categories of use cases that I think are really relevant.

1:53:14The first is micro-interactions. So this is something where you're in and out of the display within 5 to 10 seconds. So this is things like next navigation direction, a notification that comes in, which sounds really distracting, but it's actually really nice to not have to pull your phone out of your pocket. Changing your music, all that sort of stuff. Even next cooking instruction, whatever. Those are micro-interactions in and out in 5, 10 seconds or less. The other one is reference material. And so this is where you can airplay your phone screen up. You can pin up contextually relevant information.

1:53:42So if you have something that you're working on hands-free or with your hands, you want something that is hands-free that's going to give you that information. And ideally, you can ask an LLM or ask an AI to help you with that task. So those are the two categories of experience that we think about as being really necessary. There's also kind of a third one, which is spatial experiences, which is what we've seen from specs. And I think, I mean, certainly there's a lot of spatial work that we want to do in the future as well. But I think right now, it's kind of like you want to build toward what is the iPod of AR before you build the iPhone.

1:54:17And we really need to get something out that is a nice form factor that's lightweight that people are going to actually want to wear. How do you think about field of view? I'm sort of split on it. I used to be FOV maxer. I used the HoloLens and was like, this is unacceptable. Tried the Apple Vision Pro. It's much better. But then when I think about the Call of Duty HUD, the Meta Ray-Ban displays, I'm like, actually, for a lot of those micro interactions, next turn direction, next cooking instructions, something like that, just put the information right there. It's fine. It's hands-free. I'm actually maybe fine with a smaller FOV if everything's designed around that constraint.

1:54:54from, and it's not trying to be omnipresent. Yeah. So field of view is, is an interesting one because there's a couple of different ways to think of field of view. One is in, you know, the absolute number of degrees diagonal on the, on, on your vision. Um, and typically these devices are, you know, if you, if you go to the really big VR headsets, you're going to be closer to a hundred degrees, um, something like this, this is actually 30 degrees, but that's actually a lot of space that is equivalent to a 16 inch MacBook at arm's length, which is a lot of space, especially if you have a high PPD.

1:55:24So we find that it's actually field of view and pixels per degree. And if you have a high pixels per degree, you have high resolution, which means that if you have a high pixels per degree, but a low FOV, then you actually end up with something that is akin to a laptop in front of you. And so when we think about like this device is Linux based, it allows you to SSH in, you can run pretty much any application that will run on ARM64, which is of course the same architecture as Mac. And so you can run properly as a computer. Okay, world locked versus head locked. I feel like one of the problems the HoloLens was that everything was world locked And you turn your head and the narrow FOV would clip the sphere that you put there and all of a sudden you're looking at a half sphere And it draws your attention as opposed to if it's just head locked Like the screen stays where it's put and it leans in and it does the best that it can with a limited FOV Yeah, so that also it also depends on I think the world lock makes sense for those spatial experiences again you need a large field of view.

1:56:23For something like a head-up display, I don't think world lock makes much sense. You really just need something that, there's all sorts of tricks that you can do in the human-computer interaction. For example, if you have edges on the screen, then it's obvious that you're clipping. But if you allow some padding, any pixels that are black are actually rendered as transparent. So any black pixels around your icons and things like that are going to appear in screen space as transparent, which is the way we think about it. How much is capital a constraint? Because you're competing with companies that are willing to throw billions of dollars at this problem.

1:56:59And so far, I don't necessarily think it's getting them that much closer. It used to be for any hardware company, it was like, oh, you'll need$5 million to do that. And now it's like, oh yeah, no problem. All the tooling is cheap and stuff. And you can outsource the whole supply chain if you need to for a little bit. Also, it's never been easier to raise$5 million. Yeah, exactly. Exactly. Even a big number, oh, we're going to get 100 million. It's like, yeah, there's going to be 20 of those raises this month. So it's interesting when we started this company a lot of investors told us that it was going to cost you know it's going to be over a hundred million dollars to bring this product from idea to design you know early design EVT, DVT, PVT whatever all the validation stages and up to mass production.

1:57:38We're going to be shipping later this summer to the first the first batches and then we're launching later this year officially. But yeah we've done it with under eight figures. and actually there are a lot of smart glasses. We consider it a different category of ambient computing but there are a lot of smart glass companies that will go to an ODM typically in China or somewhere else and they'll take a reference design. That's not what we've done. We're manufacturing in the US. We do source components from all over of course but we do all design in our lab in San Francisco. We build the batches.

1:58:14This unit was actually built in our lab in San Francisco assembled. And we think that's really important, especially for rapid iteration. That is one of the ways that we've been able to, you know, the very small team iterate quickly and get to a product that we're really proud of. How are you thinking of marketing messaging, sort of to bring it back to that idea of like the killer use case, it feels like a lot of these pitches that we see for VR, XR, AR headsets, like they get a little scattered and all of a sudden they're trying to be 25 different things. A lot of the demos don't, if you're not a golfer, do you care about the golf thing?

1:58:52But I imagine that you get some of that flexibility for free by layering on top of AI, because if you're a front door to AI and everyone knows, oh yeah, AI can do knowledge retrieval. It can also write code. It can also generate images. Then you're like, oh, you don't need to explain that. You can just say, this is a better way to interface with any AI front door or any AI system. But how are you thinking about the marketing messaging focusing that in? Or is it more valuable to pitch broad developer platform? Who is the core customer? Yeah, so I think there have been plenty of devices that have tried pitching the developer.

1:59:29This is DevKit, DevKit, DevKit. And that's not really that useful because you get developers who get it and they play with it. They might go on a shelf for the next few years. and really in order for a device like this, even a dev kit to be useful, there has to be something in it, like some killer app. And so people ask us all the time, what's the killer app of AR? You know, there's an interesting question because what was the killer app of the iPhone? The phone, right? It actually is dependent on, yeah, it's true, but it's dependent on. Yeah, well, my take on the iPhone is that when Steve Jobs introduced it, it was a phone, an iPod, and an internet communicator, basically a web browser.

2:00:08And that was the killer app. And then the app store came in the future revisions. And then you got Uber and DoorDash and all the games and stuff. But it was enough on day one that you were like, well, I've been carrying a phone. This is a sleeker looking phone. I'll replace it. And so whenever I go to VR, AR, I'm always like, this should be a replacement for my, my, you know, 24 inch monitor that I plug my laptop into. Or it should be a replacement for a home theater or a 65-inch TV or another screen. Because in many ways, it is just another screen. So are you replacing the smartwatch? Are you replacing the phone?

2:00:45Are you replacing the iPad? Like, you got to sort of laser in instead of just trying to be something that's, like, entirely net new. It's hard. Like, the AirPods were headphones first, you know. The other thing with glasses that's interesting is just how many hundreds of millions or billions of people, like, already rely on on glasses to see yeah so that's a market where yeah it's much less of a you're not actually selling them a new device you're selling them the same thing that they rely on plus a computer built into it which like i think a lot of people like the meta ray-ban display specifically to use for phone calls and for just music like they're just oh yeah i'm going for a walk in the park and i'll just take these and play music on these instead of headphones um and and and it's like then you can get into the, you know, meta Ray-Ban displays and like interfacing with AI and all that other stuff.

2:01:34But I would imagine the majority of the time is spent like taking a couple photos, pretty basic. Everyone has a phone, but this is a little bit easier and listening to calls and listening to music. But yeah. Yeah, there's certainly, I mean, there's a lot of use cases, particularly for technical creatives. So certainly developers, engineers, but also musicians and filmmakers. And there's that kind of niche is what we're going after at first. And then of course it kind of becomes more mass market consumer after that. We also have enterprises that are interested, and we have some enterprise customers as well because we designed for a very strict set of design constraints on consumer and then translates to enterprise.

2:02:08But there's a couple other things that we care really deeply about. Not only Linux architecture, which to our knowledge is the only ones with this type of device doing that, but also all-day power. So we have these guys on the back. This is called Raven Prism, and then these are the Raven Wings. The Raven Wings just come off, And this is actually an expansion port, not only for the hot swap of a battery system, where you can take both batteries off and you can get all day power, but also for future modules. So we're going to release an HDK later this year so people can actually start building their own hardware modules on it.

2:02:42And then the other thing is privacy. I think one of the challenges that this industry has faced is particularly around the camera, but also with just user data in general. I think it's easy for us, you know, we're in San Francisco to remember, oh, yeah, you know, everyone's going to be okay with this. But people are often not okay with all the privacy implications of this. And we're seeing a lot of backlash on certain devices. So a couple of things that we do is we actually have a physical camera cover on here. So you have to kind of take that off in order to use the camera. And it allows you to have this, basically to alert other people around you when the camera is in use.

2:03:20And then there's a lot of advanced cryptography that we do kind of at the lowest levels of the stack as well. and make sure that user data is protected. And we've had a pretty positive response to that so far. How many years until a billion AR or augment? What did you call it? A billion DAUs of glasses. Smart glasses. It's hard to predict. There's an industry saying that AR has been two years away every year for the past 30 years, and it's kind of true. So I don't think I could fully predict it. I think it's reasonable. Meta has been shipping a ton of devices. It's growing the industry, which is really nice.

2:04:02I think it's reasonable to say that we'll at least be in the millions per year, or we are in millions per year, but I think we'll be in the tens of millions, maybe hundreds of millions per year in about four or five years. And then a billion daily active users could be 10 years away. It could be less. It's hard to say. Yeah, but I can tell that you're running, just from the short conversation, that you're running the company to make sure that you are in business with a fantastic product when that moment happens. I was looking at the specs, just looking at the reaction to the specs launch. And Spiegel and Snapchat are in a unique position in that they can kind of just ignore what shareholders think about where they invest.

2:04:50but that being said like it's hard to think about them spending another three and a half billion dollars you know on this effort or or uh just just given how i expect these the product to be received in its current form so i think if you're set up to like you know keep keep burn low really really keep iterating actually using the product all day long which like you must be you must have used a product you must have used like ar glasses more than anyone else on the planet i imagine like I don't know anyone. We know a lot of people that love technology. I don't know anyone that's been sort of like DAU for this long.

2:05:25But I think that's what it's going to take to make a product that people will actually use every day. For sure. Yeah. Yeah. You need the firsthand experience because there's a lot of design decisions that happen for mobile that are not clear for this type of device. And so there's no replacement for just, you know, using it every day. And there's, yeah, it's very important. How on earth can you use eye tracking in those? You really do eye tracking in those? Yeah. So actually, so this device that I'm wearing, this is, these are the same kind of form factors. It's just kind of final materials, right?

2:06:05But there's actually two little cameras in there. That's great. There's one there and there's one there. and those are hooked up to a separate kind of coprocessor, which is on the other side, which is also for privacy so that the main SoC never sees the raw camera images. But that coprocessor, yeah, takes in the data from the eyes and then we have some models that give a gaze vector over to the SoC. That's remarkable. I was convinced that anything with internal eye tracking like that was going to weigh 600 grams, 1 ,000 grams for a really long time. Not anymore. Not anymore. Well, congratulations.

2:06:44We're doubling down on it. I love it. Yeah, we're pretty excited for people to try it because it's quite a magical experience. Yeah, yeah, we're very excited for you. Really great to meet you and come back on whenever you have stuff to talk about. We'll talk to you soon. Yeah, absolutely. Yeah, we're going to be doing a launch in San Francisco in the next few months. Amazing. Congrats. We'll talk to you soon. Congrats on the progress. Have a good one. Thanks. Goodbye. Let me tell you about console.com. Console builds AI agents that automate 70 % of ITHR and finance support, giving employees instant resolution for access requests and password resets.

2:07:16We got the Germinator in the waiting room. Mark Germin, how are you doing? I miss you. What's going on, guys? We miss you. We miss you dearly. Especially when the news is like it's been and everything I read reminds me of you. Oh, yeah? Indeed. That's funny. I see Siri. I see Apple. I see Apple Foundation model. I see everything. I just think the Germinator. Okay. Have you recovered? Was last week exhausting? How are you feeling? No, last week was good. Yeah, it was an interesting conference. I'm glad they were able to keep the keynote time down to under an hour and a half. It was one of the shorter WWDCs.

2:07:56It actually felt that they had taken that five-minute section from WWDC 2024 where they introduced all those new Siri features and then took it and spread it out and milked it for an hour, basically redoing it. And for those not familiar, basically this year's WWDC was all about new Siri features, the big one being personal context. So you could ask Siri, I'll give you an example. Someone asked me for my calendar info for June, July, earlier today. So I said, send an email to so-and-so with my calendar availability for the next two months. And it was able to do that. Other ones could be like, take the note that I wrote about XYZ and text it over to that person, right?

2:08:43These are real world workflow features that I've been using all day for the last week or so, which makes it A, all the more ridiculous that Apple had to delay that stuff two years ago. It took two years to ship. But the most insane part about it is that Apple's responses to these delays all downplayed the features. They said, oh, we spent 180 minutes talking about Apple intelligence and only five out of the 180 minutes were delayed and not shipped. They basically positioned these as, oh, when it was a delay, it's not a big deal. Now that I'm using it, these are actually the only features that matter that they announced two years ago.

2:09:22Yeah, they could have delayed Genmoji. They shouldn't have delayed look up your calendar because that's a functional use that people are going to do. Yes. And just to be clear, the new Siri is really good. Yeah. And 95 % of the people, sorry to say this, are not going to need ChatGPT on their phone because they have Siri AI pre-installed. The way I look at it is that ChatGPT does a bunch of amazing things that Siri AI can't do. But Siri AI also does a lot of amazing things that you'd go to for ChatGPT. So if you're 95 % of the population using AI chatbots and you're using it for looking things up, it's basically the new age Google.

2:10:04It's going to be able to do that. It's going to be able to answer questions for you. It's going to be able to do edits for you. But there's like a ton of pro-level things that ChatGPT does that Siri can't do. So I made a list the other day of things that ChatGPT can't do. big-time research, big-time analysis of multiple different documents and PDFs and comparisons. People use it for tax preparation. People use it for in-depth health things. So there's a lot you're going to want to use it for. So I could see people having both, Siri AI on the pre-installed, and then you're going to subscribe to a ChatGPT or Claude or Gemini or what have you for some of that extra oomph on top of it.

2:10:44But Siri AI is going to be useful for a lot of people and you know, one easy way to say it, just hold on. Look at it. It's like iMovie and Final Cut Pro. Yeah. Yeah. Yeah. Yeah. I mean, we, we, we debated this, like, have you ever Googled anything? And I still Google things pretty regularly. And then I go to LLMs for more pro stuff. And then I go to agentic AI for other things. And like, there's like using a right tool for the job that I think people are, are, uh, processing and absorbing. Um, what, uh, uh, What do you think in hindsight actually delayed these features? Because it does feel like the technology was there a year or two ago.

2:11:20Was it the negotiation, the deal with Google? Like, was this a technology problem or a business problem? It was a mix of both. One was the underlying stack and the underlying foundation models that Apple had two years ago were not very good. And so they needed to go back to the drawing board because they couldn't roll these features out only for them to fail, which is ironic given they rolled out the rest of the Apple Intelligence and new Siri features that hardly worked well and were not impressive. If you've used the new betas, you can tell what has happened here. The under-the-hood models are so much better.

2:11:59Based on what I have heard Internally Apple believes that they're basically Six months behind OpenAI and ChatGPT In terms of the use cases That they have right now So Siri works great Search works great Image Playground which is their image Generation tool The same underlying tech stack that powers Genmoji It's like close to what you're getting from Gemini and close to what you're getting from ChatGPT. So everything basically just works now. It's a great experience. I mean, the first beta is sloppy and it's buggy as you would expect, but they've taken something that was way less than mediocre and they're giving people a baseline now that's actually functional.

2:12:48I would look at it as this way. When you take the iPhone out of the box, it has all these pre-installed apps. They're not great, but they work and you could use your phone based on what comes out of the box. What Apple had with AI was not that. The pre-installed stuff just didn't work. But then you can go to the App Store and you can get a much better messaging app and much better browser, email app, 10 times better versions of all the pre-installed apps. That's the case with AI, whereas now the pre-installed stuff is baseline and it works great, but you can go out and get even better stuff if you want.

2:13:21Yeah. How high are the walls of the walled garden? Like you gave that example of look up your calendar. I use Google Calendar. Send an email. I use Gmail. I'm not in the, I'm not, even though I'm like an iPhone Mac user, I'm not fully in the Apple ecosystem. Is this going to work for me? Or am I going to bump up against different walled gardens as I try and go? Yeah. Another example is like, I don't use Apple Music. I use Spotify. Can I tell? And the shortcut path. I'd be able to tell Siri, like, make me a playlist with these three artists and make a Spotify playlist. I'm guessing no, but. Well, it is too early to tell.

2:14:01It really is going to depend on how broad Apple's developer tools for the new Siri are going to become over time. The first set of tools are there. They're a bit limited. It depends on how closely Spotify and Google and all those other companies are going to want to work with Apple. I would say it's going to be a slow burn. I would say that if it was Google, I would want my applications to work better with Android and Google Pixel so I can get these cool AI features and bring people more into my ecosystem. So I would say even if they opened up the tools to the degree that you guys are explaining those scenarios for, I would say it's not going to work as well as it would if you're tapped into the Apple Mail ecosystem and the other Apple services, which is a shame, but I think over time Apple will get there.

2:14:50And let's not forget, Google is doing very similar stuff right now with Android 17. And a lot of these scenarios I've explained on the iOS side of things are completely available to what you're seeing on the Android side of the world as well with this new Android hub. So if you're deep in the Google ecosystem with Gmail and Chrome, you get a good experience on iOS. But my Pixel friends, they get a great experience. especially with the new Android stuff rolling out. How are you thinking? Wait, first off, were there any surprises? Did anything surprise you? Nothing surprises. Did anything surprise me?

2:15:26Were there any surprises that came out in BGC? I think I was a little surprised that there wasn't a little bit more. It seemed like, you know, well, that's it. I knew everything that was coming in the keynote and beyond. I actually just put a story out about how there's a new iPhone here coming out. in spring just a few minutes ago. So happy to talk about that. But next spring? Spring 27. So I would have expected more out of this dev conference. Okay. Yeah. Let's switch gears to iPhone Air. Seemed like there were some good reviews. I've seen a couple of people carry them, but it felt like it wasn't selling as well as people expected.

2:16:07And so it might just have been a one-off, but there is another one coming. Like, what is the correct story or framing around the iPhone Air? So Apple's engineering data, and this shouldn't come as a shock to anyone, the two biggest complaints for the first-generation phone are the battery life and the camera. And Apple knows this because they came out with that battery attachment on the back, which is crazy. And then they tried to market the single camera on the back of the iPhone Air as being four lenses, whatever that means. So for this second version, they're upping the battery life. And I don't necessarily think that's because of a bigger battery pack, but I think it's the efficiency gains you're going to get from 2 nanometer in this new A20 Pro chip.

2:16:49And then the other big addition will be a second rear camera for ultra wide angle photography, which a lot of people really like on the iPhone 17 and the iPhone 17 Pro and Pro Max models. And so it's going to be a really nice upgrade. And I think at$999, a little bit less expensive than the iPhone Pros, having a little bit better battery life, having that second camera, that's going to be a really compelling offering. And so I would say, you know, 30 ,000 foot view, what they're doing to the iPhone over the last year past and the next few years, what they've got, like this three to four year period of the iPhone, incredibly impressive.

2:17:30They're knocking it out of the park. I mean, the iPhone last year, that launch, the 17 Pro and Pro Max, extremely strong. and the next few years are remarkable as well. This fall, the 18 Pro and Pro Max, the first foldable phone, which I'm calling the iPhone Ultra. I believe that will be the name, but let's see, you know, still a little early. And then next spring, you'll have the regular iPhone 18 and the iPhone Air 2. We'll see if they call it the 18 Air or the Air 2. And then next fall, you'll have the iPhone 20 and iPhone 20 Pro Max those being the 20th anniversary models complete redesign, curved glass slimmer bezels, smaller dynamic island and all that lighter, you'll have cleaner it's going to be nice day by day leaks of like exactly what I was going to do it's incredible and then the second foldable phone so it's a jam packed 2027 2027 is going to be Apple's biggest product year in its history.

2:18:40I'm pumped as a consumer. Will they, do you think, do you think they will ever try to create the conditions to get somebody to own a everyday consumer consumer to try to have multiple phones? Like, will they create, will they create a product like set of features? It's like a digital twin where you have like your iPhone ultra and your iPhone air and you just, yeah.

2:19:07I don't know what I'm going to do as someone who lives and breathes this stuff I don't know what I'm going to do because I can already tell you right now I want the new iPhone Air I want the iPhone foldable and I want the iPhone 20 so it's like I don't know what I'm going to do what a first world problem to have we're talking about not even two phones but now I want three phones but it feels like for them I don't think you're going to be the only one and it's always been interesting to me you know people complain about iPhone pricing but there are not that many products in the world that are as like that have as much utility as an iPhone that you use as much as the iPhone like if the iPhone was$10 ,000 there would be a big market for them right and so I feel like Apple don't give them any ideas but I just feel like over time it's like how do we make more money it's like well you make a feature so that you can have, yeah, this is the iPhone I use on the weekend.

2:20:05This is the iPhone that I use during the work week or whatever. They should do that. They'll probably get a lot of heat for that, depending on how they articulate it. Hopefully, they don't hire you as the marketing person for that. They're just going to come across. Well, I'm just giving them business ideas. I think I'm giving them whatever it is,$1 ,500 a year for iPhones. I don't even know the pricing. but they could probably get more. They will. I mean, I would expect the prices to go up on all of these things. The foldable phone, obviously, is going to be$2 ,000 plus. The other thing to note is in iOS 27, they actually have a feature where you can stick for not.

2:20:45Wait, they think they can price the ultra higher than specs? They're out of their mind. I think they'll be higher than specs. But they added a new feature. Who, Snap? No, I'm joking. No, he's joking. Oh, we can get into that. Specs are very highly priced, yeah. But iPhone or iOS 27 has a feature where you can synchronize a single phone number and carry your account between two iPhones. So it does have that on the phone number side of things, but you're not going to be able to, you know, pick up a, make a feature to have the weekend stuff and during the week stuff and the night stuff. That'd be cool though.

2:21:20I don't know, screw it. Maybe I'll get the foldable two and the 20 at the same time in a year from now. So we'll see. You're welcome, Apple. Does the foldable eat into iPad or is it purely additive? I think it's purely additive. I think that Apple has positioned the iPad itself at this point as a companion device to the iPhone and the Mac. I think they've done a great job at that. So I don't think the foldable is going to take away iPad sales, particularly because of the price delta is so high. There will be a new OLED iPad mini coming out soon, which I think a lot of the iPad mini fanboys are going to be pretty pumped about.

2:21:53That's great. What about what you got for the Apple Vision Pro fanboys like myself? What are we doing in AR VR glasses? We just talked to the founder of Raven Renaissance. He's making augmented reality glasses, obviously specs launched as well. Is there any movement there or anything from WWDC on the software side that might actually hint at or translate into wearables? Vision Pro added this new feature where it uses the external cameras to give you a sense of your external environment and object identification that's going to be a feature at the heart of new AirPods coming out at the end of next year, it's a project called B798, where they have cameras on the AirPods that can help you with navigation and seeing the world around you cloud processing, feeding that data into Siri AI, same functionality will come to the smart glasses from Apple non-AR, right, no displays Those will be rolled out at the very end of 2027, hopefully in time for the holiday season.

2:22:55But they're a little behind on those. Augmented reality glasses from Apple, I would expect much later in the decade. New Vision Pro. So the Vision Pro is on ice. You're not going to see another new Vision Pro redesign. but what you are going to see is a completely rethought, revamped from ground up mixed reality headset from Apple. So. Okay. I'm listening. Okay. I'm thinking 2029, maybe end of 2028, but it's going to be a long time. Where's the Vision Air? We got the Vision Pro. They killed it. Just make it lighter. They killed it. Well, this is going to be a lighter. This is going to be a much lighter end to end retooling, revamping of the headset.

2:23:42It's not a Vision Pro 2. It's a start from scratch type of product. Give me the Vision Ultra. Make it 25 pounds on my face. Yeah, it's going to be, you know, it's interesting. When the Vision Pro launched and they were looking for different ways to take this product and, you know, give it a market, this is going to sound surprising to you. But obviously, on one hand, they were looking at how do we make this thing cheaper and lighter. One of the directions they were actually also looking at, how do you make this thing even higher end? How do you make this thing even more expensive? How do you make this thing even better from a technical chop standpoint?

2:24:20How do you reduce the latency? Should we do a model that's specific for medical grade, for enterprise, for these high value targets that maybe would be willing to pay$5 ,000 for this machine, a business class version of the Vision Pro. Ultimately, they're not moving in this direction right now. It's all about that end-to-end retool. Where does Apple sit on enterprise versions of their products? Like the Mac Pro, is that the last time they had something dedicated? Like Mac Studio, do they think of that as like an enterprise-grade product? The MacBook Pro is an enterprise product. I feel like MacBook Pro is extremely consumer, but I don't know.

2:25:02How did they think about like actually segmenting out that market? It's been a while. They had the XServe servers 15 years ago. You can make the argument like the Mac Pro, Mac Studio, what have you. But the thinking in Apple is make these baseline products for consumers that consumers love and make them so enjoyable that you're going to want to use them at work too and then provide the software tools that allow you to do that. And if you remember, 2008, the iPhone renaissance starting then, a lot of credit goes to the price, cutting the price down to$1.99. A lot of credit goes to the App Store.

2:25:45Some of the credit goes to 3G. Some of the credit goes to geographic distribution. But the one thing that people leave off when they talk about how did the iPhone go from here to here is 2008 is when Apple added enterprise support for the iPhone. And so getting these things in businesses is really, really important. For the Vision Pro, just like HoloLens, Magic Leap, and other AR, VR companies before them, they had no choice but to go all in on the enterprise, or not go all in, go a little bit deeper into the enterprise than they normally would like with their products, because that was their biggest market.

2:26:19Today, all of Apple's other products, watches, phones, iPads, Macs, they have really big footprints in the enterprise, but the enterprise is still only a sliver of the overall market. 90 % of their sales are consumer. Vision Pro, the split is a lot bigger in favor of enterprise. Any update on the lamp? The lamp? Oh, the tabletop robot. I still expect that to come 2028. We're going to see the smart home display. I know, I know, I know. Smart home display probably as early as the end of this year. That's going to be an amazing product. And I think it's going to be a blockbuster. And I think it's going to be a gigantic seller.

2:27:05Just imagine having this thing on your desk, your wall, your counter. You can walk up to it. It could recognize you. It could pull up the stuff for, you know, you and let's pretend Jordy and John live together. Jordy walks up to it. All of Jordy's stuff comes up. John walks up to it. All of John's stuff comes up. It has this personalized Siri. It can grab all that stuff from the device. It has FaceTime. It has intercom. It has smartphone control. Control your music. You can watch video on it. I think it's going to be a really great holiday gift when this thing finally comes out. How will it control the music?

2:27:38Are they going to finally make a run of matching Sonos, which is in a very, very tough spot as a$1.7 billion company? Are they going to try and make a run? Is that it? $1.7 billion? $1.7 billion down. What's Peloton? Peloton. Let's see. Peloton market cap 2.4. So bigger. It's crazy because like Aura is like a$10 billion valuation, right? I mean, it's amazing. Yeah, but I mean, Sonos has frustrated a lot of people with the software. They rolled a bunch of stuff back. But in terms of actually offering a consumer, prosumer level experience with, you know, Wi-Fi connected speakers that are available in more than just one SKU or two SKUs, you know?

2:28:23Like, yes, you can go all in on the Apple speakers, but then you just have a ton of really small speakers around. Yeah, I found out I have speakers built inside of my walls, and it gave me insane. They're not Sonos, thankfully, but I was like, it gave me insane Sonos PTSD because I'm like, imagine you have an issue. Well, you take the aux from that and you plug that into whatever you want, some sort of receiver, but Apple doesn't make that receiver right now. You know, Apple had such an opportunity with the HomePod in speakers, and they completely blew it because of AI, because they didn't create an app ecosystem around it because they made it exclusive to their customer base, no Android support.

2:29:00The list goes on. Pretty disappointing. They haven't given the HomePod a meaningful update in six years since the HomePod mini launched. They did a little minor refresh to the HomePod in 23. But I thought when the first HomePod launched in 2018 that they were going to be able to really go all in on speakers, different form factors, different versions. but smart home is one of two pillars for Apple right now AI wearables and then AI smart home devices and so you'll see the smart home hub you'll see one other thing in home security you'll see that lamp as you call it you'll see a refresh of the home what if the lamp and the home security combine and you have killer lamps sounds like a nightmare And then you combine it with AI.

2:29:53Wait, are there any Apple products with motors in them? No, this will be the first one with an artificial arm. And it's a motorized Apple product that can move around in space. And so if you think about it, this is a piece of the legacy of the Apple car because this project was part of the Apple car. AI and mechanics and part of that division. And it was one of the things that got left over. In fact, the person who ran the entire Apple Car project when it shut down, Kevin Lynch, this was the guy who ran software for the Apple Watch for a very long time and helped build that product from the ground up.

2:30:38His project now is the lamp. The lamp. Lamp man. It's not quite a lamp. It's a big deal going from running the Apple Watch to the Apple Car to the Apple Lamp. Do you believe the conspiracy theories that suggest that the Ferrari Luce is just a rebadged Apple car? No, no, no, no, no, no, no. The Apple car looks like a new age VW bus. Oh. You know, I think the Luce would have been successful as$150 ,000 Apple car. Yeah. But as a$700 ,000 Ferrari, I think they missed the mark. Well, who knows? You might need to buy one to get access to other cars. Ferrari is a very complicated business model. You never know how it'll do.

2:31:22It's funny. I saw people go on chat GPT after the Lucho was announced. They typed in make me an electric Ferrari. They're like, I made a better design in 15 seconds. Not hard to do. Any predictions for specs? It feels like a normal company would launch this product and then probably shutter the program. I wouldn't be surprised if Spiegel triples down, but it just feels like I cannot see the market for a$2 ,000 pair of smart glasses that still feel like they're not going to add a tremendous amount of value to my daily life. I like Evan Spiegel. I like some of the stuff Snap is doing. I think the design of the glasses actually looks pretty cool.

2:32:18I haven't tried the latest prototype yet, but I think they're well-designed and a nice product. The question is if there's going to be a killer app for it, and right now it doesn't seem to be the case. I think the price actually isn't terrible. You know, the vision when we started talking about these things half a decade ago was they need to get to smartphone prices. Smartphone prices are now$2 ,000. But if you can get this thing down to a grand, which I don't know how that's ever going to happen. You're talking about something fairly compelling. Yeah. Meta Ray-Ban displays are at$800. Do you think they're losing a bunch of money on that?

2:32:59I think they're losing money on them. I don't think people are buying them. Basically, meta, what they've done is they've created a three-tier system. Tier A is the non-display glasses that look exactly like normal glasses and don't have a display. And those have been extraordinarily popular because of the Luxottica brands pushing those. Tier B has been the meta display, which is this little HUD, little heads-up display you have in there, and you use the wristband controller. The third tier is the true AR glasses. I don't think that mid-tier with the HUD is going to last. The price is relatively solid, but I just don't see that as something people are going to be compelled by.

2:33:43You want dual display. You want binocular displays. You want immersiveness. Then on the other hand, if you tried the Vision Pro, you're spoiled. And so it's like, really, do you want to use any of these things? Because did I see a photo of you on a plane wearing a Vision Pro the other day? Oh, yeah. I watch whole movies. Which one of it? It was one of you two. Yeah, it was me. It was you. John, John. I'll watch the entire Lawrence of Arabia from start to finish in a Vision Pro. Your hair is so similar, I can't tell who's who anymore. But that's me on a plane. I love that thing for movie watching.

2:34:16It's amazing. If you've installed Vision OS 27, have you installed it yet? No. Better? The new series is pretty good on there. And then they have this new environment. I don't remember where it is in the world. It might be Iceland. It's really remarkable. And then you can set any of your panoramas now, the good ones at least, as your environment. So a few little bells and whistles tweaks in VisionOS 27. That's good. I don't think it's the – yeah, it's not the software. I'll be on a long plane flight on Sunday. I'll take a panorama of the Ultradome. And then when I'm on that plane, I'll be able to sit here, re-watch our interview on that screen there.

2:34:54That's funny. In full immersion. I'll be sitting there like this. Mid-journey hardware release. Any predictions? Have you followed it at all? Yeah, I've been looking at this closely. We're going to be covering it tonight. I believe the launch is 6 p.m. The head of hardware engineering is someone who came from the Vision Pro team. He was a manager of the Vision Pro hardware team. It seems like they've been working on this for a couple years. I would file this immediately in the category of products that may never launch or may never be bought by anyone. I wonder what we'll sell more. Yeah, experimental.

2:35:28The specs for this product, probably the specs, to be quite honest with you. I think the last thing someone sent me before I came on with you guys was that the product is going to be large in size. so I don't know what that necessarily means and I guess that would rule out a headset of some sort smart kegerator who knows

2:35:47my head went to local compute something for prosumer creators type of thing I don't think that's interesting I think you need something image something visual not another Mac studio last question about wwdc uh the parental controls the focus on uh it feels like apple solving some of the like societal anxiety around like our phones bad for kids our phones bad for people all this yeah it is existential um like what do you think they were driving there was that message received because obviously ai is everything everyone talks about but how important are those features like What were your takeaways from how, like why they chose to focus on that so much?

2:36:38How important are those features? And then will that actually move the needle on like young device adoption? A couple of things. Regulators around the world are going absolutely nuts about this stuff. Smartphone addiction, social media, et cetera. And Apple needs to address all this stuff before it becomes their next EU related problem or EU like problem. The other thing is AI. And not that Apple has had great AI on the phone, but obviously the fears of AI taking over things. And then the other thing I would say is they launched ScreenTime, I believe, in 2018, which is their parental control system.

2:37:17It's just been an unmitigated disaster that hasn't worked properly for like five or six years. And so they really needed to revamp it. But it was bug-driven? because I've noticed sometimes I'll go into screen time and it'll be like thewashingtonpost.com is 23 hours and 59 minutes. And I'm like, I had one tab open and it was just like churning and sending something. Super buggy kids being able to just do whatever the hell they want on their phones, even if they're not supposed to, just a bunch of issues. And so they needed to do an end-to-end revamp. And it's smart timing given they've finally got the AI in place.

2:37:53And then you've got a lot of people who are like, I don't want digital devices anymore. I want an MP3 player. Bring the iPod back, right? Get this screen up. Paul Merlucky is out there banging the gong about retro, futuristic devices, the game board. It's a real thing. It's a real thing. Would Apple ever do a re-release of an iPod? I don't think so. No, that's not their style. That's Paul Merlucky's domain. I think this is – well, you guys know how this feature works, right? Like you turn it on and then over time you can expand the feature set, who people can contact. applications you have. So you can really, you know, turn the iPhone into a dumb phone now and then let it get smart over time.

2:38:33It's actually a remarkable concept. And I think if executed well, it's pretty compelling and all the other phone makers are going to copy it. Interesting. Happy birthday, son. You're 17. You get to unlock the calculator app now. Exactly. Enjoy it. Don't go too crazy with those numbers. Who knows? Thank you so much for taking the time to come great to see you great thank you guys see you next time appreciate it soon have a great rest of your day goodbye let me tell you about the new york stock exchange want to change the world raise capital at the new york stock exchange just do it stop making excuses every founder that we talk to that raises capital at nicey loves it they love it they say it's incredible yeah nps is through the roof just go do it nps is legitimately When you're raising that Series D or F, why not just head over the New York Stock Exchange?

2:39:27We got Ryan Daniels from Crosby, the co-founder and CEO in the waiting room. Let's bring him into the TBP and Ultradome. Ryan, how are you doing? Good to see you again. What's the latest? Good to see you guys. Welcome back. Thank you very much. You're bench maxing. Bench max. Always got a benchmark. Tell us about benchmarks. What are you launching? Why is it important? Break it down for us. so today we uh published a benchmark it's the first benchmark that we know of that just looks at it's it's simple in theory but a negotiation between two uh lawyers like as far as we can get to completion and uh i think the main takeaway is there's a lot of art to negotiating and the way that lawyers negotiate contracts is really complex because there's so many moving variables and this is what we do at crosby we try to get those close faster and so for the first time we can actually measure, you know, how do agents do on the first red line?

2:40:20That's fine. We kind of know that already, but how do they go over and over through the process where the judgment of a lawyer really takes hold and they just try to figure out who is their counterparty and how can they get them to yes. Next step, you got to train a voice model to be able to yell on the phone at the other client at the counterparty. Let them know that you're not doing another turn of the forms. The battle of the forms is over. Now, obviously there's a lot that goes into this. talk about the data. Are you getting good redline data, ground truth data from firms? Are they happy to partner with you on this?

2:40:51Like, what is the value prop to actually get the correct data? Are you doing it yourself through a data firm? Like, what's the strategy to get the ground truth here? Because it is, it's not the same as like a math problem that you can just verify with lean. Yeah, that's, that's like, that's the key point of it. And, and like, you're totally right. Like, Like, so we worked with Mike R1 really closely on this with Ali and his team. They've been amazing. They have like a really great legal focus that they're building out. And, you know, we have our own captive law firm here. That's been our kind of shtick from the beginning was in order to get the full benefits of what agents can do in law, start your own law firm, sell end-to-end legal work.

2:41:31And as agents get better, our margins get better. And we worked with Ali. And the hard thing is, like, there is no right answer. and we basically had large groups of real lawyers negotiating all with the same positions for the same parties and seeing like how much of those lawyers agree with one another on the right way to respond to those other lawyers and so we could just see the overlap and as the negotiation went longer and longer we saw a ton of deviation between those lawyers which is fascinating like they're all very senior they're all very experienced this is a real negotiation but at the beginning the lawyers kind of more or less did the same thing and we just saw the way the agents we then built a rubric based on how those lawyers actually worked and tried to create some sort of ground truth for the best way to negotiate.

2:42:13But to your point, I don't know, maybe if we just did everything in all caps, the candidate would just accept it and we could yell through the margin. So you have to try it. That's where you pay the big bucks if you got big lungs and can scream. It seems like all the models did very similarly. 50%, 50.5 for GPT, 5.5. Weird numerology there. 5.5 gets a 50.5. Gemini 3.5 Flash got a 45.1, Opus 4.8 got 44.4, and Fable 5 got 47.3. That seems surprising. I mean, it seems like the window for testing was low, or is that something like Fable is more tuned to cybersecurity, bio, those sort of reinforced learnings, those feedback loops, and maybe it hasn't been applied to legal yet, but what was your interpretation?

2:43:04We put a big asterisk on Fable because we got one testing run through and then we lost access. Wow. So we still would want to do a few more. Sure. If and when we get access back. You know, I think it's not surprising. I think what was most interesting to us were a few things. One, in the first red line, like the first read of a document, and agents really did poorly. They scored poorly there because they couldn't service those first issues. As the new issue went on, we knew the issues. Models did quite a bit better. But the biggest thing was models are really likely to just say yes and accept some term because you want to keep the deal going and that was part of the instruction.

2:43:46And that's just not good lawyer. A great lawyer knows how to make you feel like you're getting your side but still push the deal forward and protect the client. And that's where the real judgment came in. And so, you know, in our mind, we have a lot of work to do because we want our agents to be able to take the work off our lawyers' hands, to be able to think through something and say, hey, like, we push back on this for this reason and really get the counterparty there. So, you know, I think the models did better than we thought they would, but there's a lot of work to do. We're having a little bit of interference.

2:44:15I have one more question. Go for it. Do you have anything you need to get in? I just want to know about the value of the harness because it seems like all of these models are sort of neck and neck, you know, within a few percentage. What happens if you wrap them in a very bespoke harness that your team builds? Can you get to 70 percent, 90 percent? Like, where do you think the upper bound is on this benchmark? Yeah, that's a great question. And like, you know, everything's published. It's all open source. So, like, the harness that we use, everything can all be replicated. Yeah. there's a lot of ways to push.

2:44:48I think we've invested a lot in the harness that we wrote specifically that's really good at editing Word documents. That seems to be a really hard thing. And so there's ways to push there. I think ultimately more and more bespoke models that are really post-trained on just making judgment calls, stating their reasoning will be an area. So both of those things help. But at the end of the day, there's a ton of intelligence in just the closed models, and I think it's not being fully exploited. So, yeah, all areas to explore. Those are the right questions. How has your team experimented with token maxing?

2:45:22Like I met you guys are heavily funded. You're building an AI law firm. I imagine you're not telling the team, like, you know, be conservative with tokens, especially because you can just like learn faster. And it's kind of the whole point of the business. But where does token maxing make sense with legal? Have you already found an equilibrium? What is your overall kind of like ethos around it? I feel like as a company, we're not successful enough to be like worried about token costs yet. We're still in a place where we can just be like putting as much money. And I think more and more law firms like Kirkland just allocated half a billion dollars to building their own AI stuff are just trying to figure out, you know, they're charging.

2:46:06They're charging huge amounts of money for the labor spend of legal. So they have a lot of room to throw money at tokens. and so long as you're like basically spending less than the equivalent of a thousand dollars an hour on tokens which is even for this still hard to do like you're still coming out with a good margin so like we you know we're spending a ton you know we did a lot of work with fable we got early access to it and and uh we spent a lot of money with it and like you know and so i think there's still room to go um because ultimately it's either the model or the human and the models are still cheaper than you know hiring um expensive lawyers for some of the tasks that we don't think we need lawyers for.

2:46:40Yeah, that makes sense. Well, thank you so much for taking the time to come. Yeah, sorry for the technical difficulties, but yeah, great, great to get the update. And we will talk to you soon. Have a good rest of your day. Goodbye. See you, Ryan. Let me tell you about public.com, investing for those who take it seriously. Stocks, options, bonds, crypto, treasury, and more with great customer service. Our next guest, Isaiah from Bland, is the co-founder and CEO. Welcome to the show. How are you doing? What's up, dude? I'm doing good. This is like the coolest thing I get to do for this funder is announcement.

2:47:12Fantastic. We'll kick it off. I know. I know. It's way too. Tell us about the fundraise. How much did you raise? This is way overdue. How much did you raise? We raised$50 million. Hit it again. Hit it again. Because it should have happened. There we go. Congratulations. Now tell us, what do you do? Long overdue. Yeah. So you guys know when you call basically any big company in the U.S. and you have to deal with one of those really annoying bots. We replace that with AI. But our specialty, we like to say, is like, we want to do the calls where if something goes wrong, someone sues or dies. So we do the really complex, hard-to-handle calls, and then we do that without open AI.

2:47:53I don't know if I'm allowed to say that on here. And drop it to anyone else. So you're training your own models. Yeah, correct. And, okay, so more specifically, Like, are you working like 911 calls? Is that, I imagine that's not your focus, right? Or maybe it is, but I know there's some companies that just focus on that. Are you thinking like hospitals? Like what kind of, give us some more maybe concrete examples. Hospitals, airlines, and yes, 911 calls eventually. So we're actually working on a product offering for that right now. I can't say too much about that, but I do have this weird feeling that like poetically I'm going to die by a botched AI 911 call one day.

2:48:40Uh, and that's kind of my mission. Don't man it. Don't yeah. Yeah. Yeah. You gotta, you're a time traveler going back in time. I gotta warm myself. You gotta, yeah, you gotta, you gotta save yourself. Um, what, what, how, uh, describe this market, like describe voice, describe the voice AI market, right? Like you guys have, uh, some competition from the frontier labs, but it seems like less competition than maybe one would think, or maybe it's like almost certainly has to be less competitive than coding agents, right? Where you have thousands of companies and the labs and the hyperscalers all competing.

2:49:24Yeah, and you'd be surprised. So like two years ago, we were one of the first companies to do this. This was like well before the labs. And I have a little binder of screenshots of different VCs that I'm sure talk about their voice AI pieces today who said, no one will be talking to AI two years from today. And by the way, very unpopular if you go into a fundraise reminding them that they said that. They don't like that. I don't recommend that. Chained gloves for sure. On that point of like, no one will be talking to AI, it seems like, are you actually increasing the length of phone calls? Like I'm seeing like the typical call last 30 to 45 minutes.

2:50:05That feels like too long. Honestly, that feels like a long time. Are people asking your models to write Python for them or something and they're transcribing it one, one character at a time? Like what's going on? It'd be a little roundabout, but yeah, we, we do calls with like 90 year old patients doing 45 minute long remote patient monitoring calls. So if you want them to put on the blood pressure device, right, and you want to figure out if they're having a heart attack, you have to care about their grandchild first. That is an immovable object you have to get through. These are lonely, lonely people.

2:50:40Oh, that's interesting. So you actually have to somewhat like put guardrails on, obviously, but not the card rails can't be too narrow to the point where you're a phone tree because you actually do have to open up the customer, open up the user on the other end of the line and start actually having a conversation before you can do a more complex interaction. Interesting. You would not believe. Yeah. Do companies have like specific ideas about that, preconceived notions, whether they should or shouldn't do that, whether that's table stakes? What does that look like? So we work with a couple of really big banks.

2:51:13And one of them had a minute and a half long disclosure we had to read at the start of every call. I know, thank you. I worked that in there. So if you're a big bank watching this, please disregard everything else I've said and just reach out. But we had to have a minute and a half long disclosure. So if you wanted to have a human being talk to RAI, you had to sit through a minute and a half of straight reading out different things to them. And then we get on calls with them and they'd say, it's varying, right? Not every bank's like this. We've worked with some really big ones that are less painful, but they're like, yeah, how do we know the AI is not going to say something racist?

2:51:52But we're pretty sure that it won't. And you have to go through this really intensive testing process. And so what a lot of people don't realize is that these call centers and these really big ones that we're dealing with, every business has tried to get rid of them every possible way they can. They only exist because there's a gap between what the business wants to offer and what it can't offer. And so when you hear about like AI receptionists or you hear about e-commerce, where's my order calls? That's not what we're doing. We're doing the hard part of the$250 billion a year spend in the U.S.

2:52:25And that is really early. That is like astoundingly early. I don't think people realize how early it is. Makes sense. uh what is i mean you're talking to big banks like what is the actual uh go-to-market motion funnel is it all sales driven or are there inbound do you have an inbound team or is there any like uh sort of like self-serve option at this point um we have 350 000 self-serve users but most of it is actually inbound leads so we do viral marketing every once in a while uh you guys might have seen this but we had a billboard this was back before it was like cool to do billboards that said still hiring humans, question mark, and it had a phone number on it.

2:53:06Not stop hiring humans. That is a different company that we're not happy with that. You guys were the first. We were the first. And I actually got a chance while the founder of that other company was getting a haircut to go up to him and tell him what I thought. Because I found out we were the same. Oh, haircut. Haircut's crazy. Very captive. He's locked in. Wow. Captive audience. Sort of a captive audience, right? the poor people did not know what was going on and why we were so upset at each other. Yeah, because you invented billboards. So for him to come in and not only run a billboard, but to use similar language, it's just, you know.

2:53:42Same graphics, same everything. I was like, he's like, oh, I'm a big fan of your billboards. I'm like, yeah, I know.

2:53:50Ted Turner. But yeah, so we bought the rights to Soulja Boy's voice, for example. So you all know Soulja Boy, like crank that, right? kissing through the phone. So we've done a lot of viral marketing and we've managed to get CIOs, great people to come in and actually come to us. We do a lot of outbound too, but primarily. Wait, so does Soldier Boy help you win in the enterprise? Like you would not believe because he was really popular around the time that most of the CIOs were like teenagers or young adults. Yeah, yeah. He probably crushes it like AWS reinvent in Vegas. Are there any of your customers actively using Soldier Boy's voice for their customer service?

2:54:35Or is that just like kind of a big deal? I'm going to need to read a disclosure to you for a minute and a half, but I'm bringing in Soldier Boy to do it. I think that's a winning formula. The retention there would probably be really quite good. Why not? Yes, actually. There's a lot of restaurants because apparently people really love it. But we have his name published under a different name on the platform. So we have Soldier Boy's official voice, and then we have a name that's hidden for the other one. The other voice is very popular. So there are a few enterprises that if you stumble upon it and you call, you might not realize it, but you're actually talking to Soldier Boy.

2:55:13The one and only. Talk about working with Paul Lieberstein. Yeah, I never get to meet the celebrities.

2:55:27It's always the marketing team. He was incredibly nice to work with. Very pleasant. Ramp had Kevin or Brian. And I have to be very careful about the names and not mixing them up. But we were like, who's the most boring person that we could possibly think of? Because the company's name is Bland. So it kind of comes with the territory. And everyone was like, well, Toby from The Office. Paul, I think, was humored by that. Took it very graciously. And we ended up having a lot of really great interaction with that. And it sort of goes along with our marketing, which is fun and engaging, but not maybe disruptive to the brand where you're like, these guys, I wouldn't trust them to handle my 911 calls.

2:56:13Imagine calling 911 and Soulja Boy picks up. Soulja Boy. That would be worrisome. I gotta say. But good lead, Jen. Well, thank you so much for coming on the show. Who did the round? So it was our insiders plus Dell and HubSpot. So we actually went very corporate on this round and we were excited to have them. So you guys will be hearing more thoughts on new models and stuff. Great to finally meet you. Thank you for having me. We'll talk to you soon. Cheers. Have a good one. In market news, more Fed officials have signaled a rate increase as the next move. The central bank held interest rates steady as Kevin Warsh's first meeting its chairman.

2:57:01But nine of 19 Federal Reserve officials penciled in at least one rate increase by year's end, up from none in March. And so the market is selling off down about a percent today based on that. There's also a G7 meeting that's happening between AI leaders. Donald Trump was seated next to Sam Altman and Demis Fassabas from Google's DeepMind at a summit in France on Wednesday discussing AI export controls and all things AI. The AI leaders are huddling at lunch with the G7 in France. And I'm sure there'll be more things to come out of that. There's still more back and forth on export controls. The Fable 5 is still embargoed in some way, and they're working through that.

2:57:51There's been a little bit more reporting, but not much major movement there. And then also, U.S. has held off on blacklisting China's DeepSeek, more than 100 firms and more than 100 firms that are deemed security risks. That was part of the Fable 5 rollout was that there was a South Korean telecom company, according to the reporting, that had access to one of the most advanced models for philanthropic. And that firm, that South Korean telecom company, had potentially ties to China in some way. And so the U.S. government was skeptical of that South Korean telecom company. And so there was a debate over that and whether or not that crossed a bright line.

2:58:40and so from the Washington Post, Anthropic later disclosed that the list had ballooned, the list of companies that would be getting the most advanced models had ballooned in roughly 50 additional entities had already received access. Senior officials began to consider using export controls to claw back the technology after the company did not identify new recipients for days. When Anthropic finally turned over names, the administration discovered that one recipient was a South Korean telecommunications company. The administration suspected, alleged, of having ties to China, officials said. And so that is the key of the dust up there.

2:59:15But we'll be continuing to follow that. In other news, much more tragically, many of you will have seen this by now, but Joshua Bayer, who is the CEO and founder of Capital Factory, was in a plane crash in Laredo. I guess early, early this morning, coming back from Mexico to Austin, I unfortunately never got to meet Joshua, but only heard tremendous things about him. And he really was an important figure in the Austin startup community. So sending our prayers to Joshua's family and friends. And yeah, really, really tragic. Rest in peace. Well, we should close on a positive note in some way. There's a whole discussion over SpaceX potentially using the high share price, the incredible valuation to acquire more companies, create a roll-up.

3:00:11There's a piece in the Financial Times we can run through another day. But Bill Ackman shares a Hall of Fame opening sentence. One of the things that makes SpaceX so valuable is how valuable it is. a tautological value argument. Of course, what he's actually getting at is that while the stock price is so high, that serves as a currency for acquisition. And when you're a public company, you can acquire companies very easily with your public stock. And so there's a very interesting window. Ben Thompson wrote about it on the back of the cursor acquisition closing or being announced that the option has been exercised.

3:00:54But it is a very interesting debate. We touched on it a little bit yesterday. Is there gonna be an acquisition spree? Will there be a roll up? Will SpaceX buy NeoCloud assets, energy assets, chip assets, like TerraFab has been talking. What's involved in that? I mean, they're trading at what, 10 times Intel at this point or something? I don't know, what is Intel market cap? But there's so much that they could do. They, well, it's not 10 times Intel. Intel's a$600 billion company. So that would be a big one. But there's a lot in the supply chain. There's a lot in the AI world and the space world that they could partner up with if that's the direction that they want to go.

3:01:34But it would be a very different direction. And so everyone will be watching it very, very closely. Anyway, thank you for tuning in to TBPN today. Leave us five stars on Apple Podcasts and Spotify. Sign up for our newsletter at tbpn.com. and we will see you tomorrow. And hello to Keith, who seemingly just joined the chat. Hey, Keith. You're a couple hours late, but happy to have you here. Three hours late, but we're happy to have you here. Come back tomorrow. I hope you all have a tremendous afternoon. Thank you to Cooper for always coming in at the market close as well. I appreciate it. I appreciate it every day.

3:02:09And have a wonderful afternoon and evening, folks. We'll see you tomorrow. Goodbye.