This Week in Startups - OpenAI’s $1.25T Raise, Starlink Mobile & Zuck’s Tax Deal | EP 2354

Starting point is 00:00:00 All right, everybody, welcome back Friday afternoon. I am in San Francisco. I had a great trip. We had launch accelerator class 37 graduating. Oh, 37. And Founder University cohort two from Japan. Right. Tokyo.

Starting point is 00:00:18 And so, yeah, Tokyo. But anybody from Japan can join that, just to be clear. It's not just city specific. Correct. And so we had 30 graduates from the founder university in Japan. The top 10, as selected by our teams, got to come to an immersion week here in the United States where they did tons of meetings that we set up, met entrepreneurs, met venture capital firms, and just explored and got to really soak in San Francisco.

Starting point is 00:00:46 So I'm having a little graduation dim sum with them this afternoon here on a Friday, October 23rd. Yeah, great dim sum place. I love Yangt Singh. Yank Singh. Get the duck. Get the soup dumplings. Then the Large Accelerator Class graduated, and we did something new for them. One of the new things we did for them was we found that with such a small cohort, you know,

Starting point is 00:01:09 we keep it to seven companies, sometimes 10, but, you know, seven's a good number. We started bringing them to see Sequoia and Kraft, who we have a deep relationship with each of those firms, but we expanded it. And so we did about 15 firms here in San Francisco, where they met all different venture capital firms and gave their pitch in their offices. Absolutely fantastic. So you get to meet multiple members of that venture team and build a tight relationship. And of course, since we're only doing seven companies, it's possible. If you're part of Ycombinator with 200 companies, it's kind of hard to do. You can't bring 200 people to somebody's office or it looks like a protest.

Starting point is 00:01:47 Right. And you get your reps. I mean, that's what's so important for these founders. We write a lot of these tactical practical tips. And that's something you hear from founders from veterans over and over again, is like you just need to get used to it. You just need to learn every time you do a pitch, you get a little better. You get a little smarter about what parts are people responding to, what parts are leaving people cold. You get to sort of workshop it. So that's such important experience for these founders.

Starting point is 00:02:11 Whether you're an athlete or an actor, deliberate practice, pretty good thing. So over the 12 weeks, each week they give their pitch to a panel of a half dozen investors. But then during this immersion week, we had them do, I think 15 meetings this time. We blew through all of our predictions, and we're going to take them to New York. So if anybody from New York has a venture firm, they want to meet my seven companies. I'll probably be joining them for some of the excitement. So just email me, Jason at calicanus.com for life. Jason at calicanus.com for life.

Starting point is 00:02:40 Let me ask you, was anybody there at this event, were they recording the demos perhaps on applaud pins so that they could go back and listen to them later? It's certainly a great effort to do. And, you know, I'm wearing this beautiful... I forgot the name of this expensive material, but it's, what is this material? Gortex? Is it Gortex? No, no, no, no.

Starting point is 00:03:03 This is like seriously expensive, but good deal, good value. It's not cashmere. It's that other one. I got to take it off here. Now I need to know. Now I want the answer. This jacket is luxurious. I'm going to put my plug pin on this jacket.

Starting point is 00:03:17 Somebody is looking it up right now and they're going to tell us in the fan chat. Swade. It's suede. It's suede. It's suede. It's a blue suede. Polo Roplo-Rop-Loran. It does not look swaying from here.

Starting point is 00:03:27 It's like the lighting or something. Anyway, I got this gray jacket. Now, I don't want to damage the jacket, right? No, not your new suede jacket. Definitely not. My new blue suede jacket. Not shoes, but blue suede jacket. So what I like to do at my plod pin,

Starting point is 00:03:44 I have my plow pin here. It's got a magnetic buck. I put this on here. This is not damaging anything. It's just right there. It's just press the button. Now I'll take all the notes for us today. and yeah, look at yours.

Starting point is 00:03:56 Now, I'm going a little more subtle here. A little, you know, I got the darker one. You got the one that's basically makes you look like you're doing a plot ad. But that's okay. If you're wearing a white shirt, it would have been fine. But we are, in fact, doing an ad for a plot. We're doing applaud ad. We do all our show notes now.

Starting point is 00:04:10 So if you go to any of our videos, look at the show notes, you'll get that link to our notes. Now, when we do the notes, along the top, you'll notice four tabs. And let's add a fifth tab for the next one and for the next live ad read. But right now, you get the transcript. Great. Sometimes you hear me say the name of a company or a person. You want to go check it out. Then you get to summary. Hey, maybe Lon is explaining all these great stories for the day, but you want a summary. So you get a top level summary? Who knows? Maybe you want to jump around in the podcast. Sometimes people do that. Finally, if you thought of it as a meeting, it will give you the meeting highlights. Yes. Meeting highlights. And working highlights. And working highlights. And then workful acceleration, product readiness. You're seeing all these things here. And then teaching notes. I'm big in teaching. So they have an instruction set here, a skill. to do teaching notes. So here's your overview, remaining content, cover content, et cetera. Or a teacher making a lesson plan. And those are templates and there's dozens of them. So whatever you want to turn your recording, hundreds maybe. So whatever you want to turn your recording

Starting point is 00:05:08 into whether it's more like meeting notes, whether it's more like a lesson plan, whether it's more like a news item, whether it's more like a press release, whether it's whatever it is, a dialogue, like a script or a Q&A, you can make it happen easily with Plaude. They can take it. you're recording and turn it into basically whatever kind of document you like. Check it out yourself. We encourage you plod.aI slash twist. And if you use the code twist, they're going to give you 10%. All right.

Starting point is 00:05:37 Everybody, go get applaud. I have like so many of my friends who see me use it. It's a pretty viral product. So every time people see me using it, because I always ask for permission, hey, can I record this some notes here? I'll share you my summary of the meeting. People always appreciate that because I like doing in-person stuff. Sure.

Starting point is 00:05:51 All right. So the other thing we did was we have been extremely focused. And I always look at what do founders need today? We did this, hey, let's tour and go actually to the venture capitalist office and build a deeper relationship with them. So a smaller number of firms, not 300 people or not 1,000 people out of cattle call type demo day, which is chaotic, which, you know, you need to do that if you have 200 companies. You need to have 2,000 people there if you're going to have 10 per company. Reasonably. So I totally get it. But that becomes a cattle call. And then just, it's like very panicky. It's not elegant. It's not effective. So I always challenge myself on a Kaizen basis. It's always getting a little better. What would be a little better this time? The other thing I challenged my team with was, hey, how do we make this podcast a little

Starting point is 00:06:34 better? How do we make the programs a little bit better? How do we make our live events a little bit better? So we did a live event yesterday. And we did it at Salesforce. Thanks to my friend Mark Benny Off. It's a beautiful facility with the Salesforce Tower. Obviously, you've heard about it.

Starting point is 00:06:46 Gorgeous building. And he has his classroom, which I'll share some pictures. of at some point, or maybe we have one here from our Slack. Beautiful, beautiful room, giant TVs, giant monitors, and we had these incredible sessions, beautiful, like classroom-style seating. I only fit like 50 people in there, but I asked them, can we put some chairs in the back? Because I got other founders who want to come, put another 25 chairs in the back. We had 75 founders, founders only from our portfolio or people we've met with recently that got to a second meeting. Okay, so this is like a tight group. I asked my team, hey, I'd probably have to have six times.

Starting point is 00:07:21 And look at that room. Beautiful. I mean, this is a gorgeous room. And I got to just say, thank you to my friends at Salesforce. They're so supportive of me. They didn't charge me. I was happy to pay a fee for the room. Not only didn't they charge me, they gave me this great AV support.

Starting point is 00:07:37 They put out cookies. They put out snacks. They put out great coffee. Brad, the AV engineer from Salesforce who's been helping us get all this content to get. Incredible job. And so we wound up doing nine tactical level. Practical talks. Our newsletter is focused now that this week in startup's newsletter is focused on

Starting point is 00:07:56 tactical and practical. Why? Let's face it, folks. News is a bit commodified right now. Analysis of the news, opinions about the news, less so. But we have this incredible newsletter, Trist newsletter, and we decided to settle on the number of 100 tactical practical articles, you know, in-depth content modules here to make you a better founder. Two a week. Under per year. per year, 100 per year. So a buck each, $100 a year right now. We're going to raise the price at some point. But now it's 100. It's to go to www.s.com twist newsletters, but we did this in-person. So of those nine talks, all nine are worthy of a newsletter. So there'll be nine of the next hundred. I'm working on them. Because I was like, we should just share them.

Starting point is 00:08:40 And what we do is we take the article, or we'll take a video like we taped. But then we're going to, you know, kind of productize it with graphics. I mean, you can explain maybe a little bit of the AI process we're using here because we're, we don't believe in AI slop, but we do believe in superintelligence standards. Correct. There's AI slop and then there's super intelligent standards, which comes from context windows that contain style guides and examples of what's right and with constant tweaking. But explain what you're doing here for these articles and they can throw one up on the screen that you're proud of. So what I do is, you know, we'll obviously do this for all of these talks from this week, this amazing event in San Francisco.

Starting point is 00:09:21 I've just been finding either from the Twist Archives, we've had dozens of founders, hundreds of founders over the years, share great tips. But I've also just been finding, you know, other podcasts, other places where founders are talking, other resources that are out there for the Internet. There is so much of this content. So layer one of the AI is I have trained a Grockbot that is plugged into YouTube and X and is very good at searching them. he's just scouring all day every day,

Starting point is 00:09:47 looking for practical founder tips from across the web, more time than any human could ever spend watching videos. I used to do this by hand. Now it's Grockbot's turn and making recommendations and filling in a spreadsheet for me. And then I go through and I look for which tips do I think would be the best, which are the most laid out, which are the most detailed. And I'm really looking for...

Starting point is 00:10:08 Credible. I'm really looking for a workflow. Like I'm looking for is this, there's a lot of vague, like, here's three things to bear in mind when you send a cold outreach email. I mean, that's great. That's helpful. But here is the eight-step process I used to increase my success on cold outreach emails by 30%.

Starting point is 00:10:26 That's a tactical, practical kind of tip. So I'm looking for stuff like that. Then we take it. I sort of run it through AI as a brainstorming layer. I usually use Claude. Opus is ideal for this. You could use Sonet. And I have it sort of pull out for me, like, here's the talk.

Starting point is 00:10:43 here's the thing. What do you think? What would make good graphics? Let's brainstorm. What would make good point? Like, how should I lay this out? And I kind of workshop it. I don't let Claude write anything. I think this is where people make the error is that they think if I Cardinal sin. If I prompt engineer it just right, you know, if I give it the right context, it's just going to one shot like a perfect essay. I don't really know that the model. Not today. Not today. Maybe one day. We know not today. Not today. We know. So I feed all of the context. I feed, you know, the talk, everything else I want to add. If there's an interesting perspective from someone else, I put all the information I could find in there. I let Claude kind of roughly assemble it. And then I take the suggestions. I workshop it with God. And then I turn it into the final version of the city.

Starting point is 00:11:27 You write it. It's your words. I write it. 100% human written. And you are turning your brain on. So this is a critically important thing. Young people right now are just letting the output be the output and they cut and paste it. If you're cutting and paste it, you've made it. You've made. your first mistake. That's the indicator that you're disgraciat. You should be looking at that output. You can ask questions and follow-up questions. Hey, I'm confused. I don't know this term of art. I don't know what, you know, CPC is cost per click. I don't know what, you know, CPA is cost per action. Can you explain that to me? Give me a crisp definition that everybody can understand. Then you write it yourself in your own words, but now, you know, you're, and then fact-checking, obviously good. And if we pull this up again, what I wanted to see was the graphics on this one.

Starting point is 00:12:08 And so maybe we could show us the first or second graphic, whichever graphic you think long is the most representative on this one of the process. Scroll down a little. Oh, I think that's a clip, right? Those are clips. I think this one was actually from before I was making the graphics. So, Jacob, go to a more recent one. We'll talk about the graphics. Sure.

Starting point is 00:12:25 So just so people know, we will put in clips that you can watch and you don't have to watch the whole 20-minute presentation or hour-long podcast. We're going to take you right to the most important clips. So we still do video clips. I'm sort of sucking out, yeah, like two or three great. short because it's not, you know, sometimes it's not our content. I'm not trying to steal somebody's video. I'm taking 30 seconds, 45 seconds that just gets you to like something impactful and then you can read the post to sort of glean everything. Here's a great example of one of the graphics. So this is like, it's a back and forth process. Obviously, Jason and I worked on a

Starting point is 00:12:55 style guide. So that was the first step. Claude design. You go in there, you bounce ideas around. I like this. I don't like this. I like this font. I don't like that. And eventually you get to a point where it's like this is a good sort of style format that we can use over and over again to make our content. So I have that style guide now and then I go, here's an idea for the graphic, have it say this and that, put it in the twist house. So in this graphic, it says, who knows your candidate? Why, Maritra trusts the reference check over the interview.

Starting point is 00:13:29 So this interview, 30 minutes of artificial scenarios, reference years, significantly more hands-on experience. General notes, the source, and the bars are not to scale. I mean, this is great notes. And what you see is, this is just a visual way to pull you into the article and reinforce the learning. So when you're doing learning things, if you make a great visual, what happens is you're going to remember this. So when you have a conversation with your team and they say, well, we did three interviews. You're like, okay, three interviews times 30 minutes is 90 minutes. If we did three references and they each worked with them for two years, that's six years, 2,000 hours, that's each. Okay, now you're talking about 12,000 hours of real experience.

Starting point is 00:14:15 And that's just a great, I mean, really hats off to you and Claude and me for making the standard. This is founder field notes hiring. And so when you look at this, a used grid paper, I asked it, I want something where you're writing in a notebook and I want a font that is like somebody with the best penmanship ever that would write something on a whiteboard or in their field notes. would make you jealous that that person's such a considered, you know, there's somebody you know who's an art director or like a great founder who's like a creative director and you open up their notes and you look at it and you're like, it's like Michelangelo, right? Making incredible notes. And it kind of got my vibe. So now we have a style guide and I said,

Starting point is 00:14:54 you know, make one that is for infographics. Make one that's for charts. Make one that's for a table. Make one that would be a Venn diagram. Make one that's a four quadrant. So I kind of gave it what I think would be style guides to come up and again. The one other thing I was going to point out, Claude is getting so smart. I sometimes I have to correct it, but you'll see it has one of our tags as a category. Like this one said hiring,

Starting point is 00:15:18 99% of the time I don't have to tell it what to put there. It figures it out from context. It figures out, oh, here's what I'm making a graph about. That out of all of the tags that I know are on line substack, that's most like hiring, and it just files it there itself. Like, that's impressive. So here's what I'd like to do with this,

Starting point is 00:15:36 is I'd like to make sure it has a subscription to the paid version and is ingesting it. I'm all, what step ahead of you? Yep. Claude reads everything. And then, now, I ask for it to get feedback for it. So this is all about building the context window. I want to figure out a way to get Claude the feedback. So how many likes, how many comments?

Starting point is 00:15:56 And if anybody hit reply to the email. So that would mean when we send it, we send it from an email that gets forwarded to Claude somehow. So what I want to do is ask people, hey, hit reply and give feedback. feedback to us, but I want that feedback to go to Claude. There are surveys in substack that we could theoretically feed to Claude, and that's how to do it. Even easier, you could ask Claude to stand up a survey tool and to ask people for feedback, and it could just be part of this.

Starting point is 00:16:21 So keep building the context window, please, of this. And then I think we will be able to do a real great job for our founders. But the tactical practical in person was a really major breakthrough for us as a company. And I just want to thank Salesforce for partnering with us on our tactical and practical nine talks for our founders. And so you just, and now it's going to go out to a large number of, even a larger number of people. We should include some of the images and the pictures I took and of the room in those newsletters. So people realize this came from a tactical practical thing that Jason hosted at Salesforce. Just to give people even that extra context, one of the things I notice about editorial products here while we're giving a little,

Starting point is 00:17:06 a little insight into how I build them is when people understand the intentionality behind it, which is why I'm taking 15 minutes here to explain it, and you can include this discussion in the next one. Hey, Jason and I talked about what we're trying to do here. When you take people behind the movie, like the director notes or, you know, people will do behind the scenes, vignettes of movies that are made. Ridley Scott did this a whole bunch for Prometheus. It gets people more vested in it. They call it a featurette. Thank you. I want to keep doing this as like a featurette for the newsletter. And if anybody has ideas for this, just sign up for the newsletter and then post the comments under it, what could have been better, et cetera. So we want to go from this

Starting point is 00:17:44 extremely important content that you can't miss. Great editorial products tend to be one thing, just one thing, but they do that one thing better than anybody. And when I want this to be is every one of these is worth the time of a founder and it's tactical and it's practical. In other words, it's a tactical, it's telling you how to deploy this, but it's practical for you. It's not like this is way above my pay grade. I'll never figure this out. I don't know how to make a chip set. Yeah. You know, or you made a point earlier along, which I think is the right one. Founders and investors like to spout off about like, you know, follow your passion and I only take warm intros. I love to get emails, like all this contradictory advice. What we're trying to do here is like

Starting point is 00:18:28 get rid of that and get you to the 100 best actual tactical and practical things that will make a change for you this week, this month, today for your startup. Thank you, everybody. That's the stuff that's tough to get through. There's a lot of this that it sort of poses as a tactical practical, practical tip, but it's really more like motivational, inspirational speaking. And like, no problem. Like, people need to be motivated and inspired.

Starting point is 00:18:51 I don't mean to knock that stuff, but it's not what we're looking for. Do they? I mean, I think like, there's a lot of that out there where it's like, grindset, mindset kind of stuff, and it's like, that doesn't really help anybody. I mean, I think that stuff gets a like, but it doesn't get printed out and forwarded and talked about in the management meetings. For me, if we don't get liked or retweeted, I'm not so concerned about that. I want, if you could tell me how many people printed this out or forwarded it to somebody

Starting point is 00:19:17 on their team or brought it to a meeting and pulled it up on the Zoom, that to me, which you can't get those statistics. But those are engagements statistics that don't lie. no matter how polished or well-designed your app might be. Every once in a while, something is going to go wrong in production, right? Even if you're a perfect U.S. designer and you know what you're doing. But that doesn't mean your team has to waste an entire day just to locate the source of the problem, the bug, the issue.

Starting point is 00:19:43 We got Sentry now, everybody. And what I love about Century is that it gives you total visibility into your app in one place. This is not separate tools or endless context switching. No, just your errors, traces, replays, and logs. all linked together in one connected timeline, right? You can actually play the timeline. So when there's an issue, you can see exactly where it started and what triggered it. And debugging these things is critically important because SEER, S-E-E-R, is Century's AI debugging agent.

Starting point is 00:20:12 It's got an agent. It's going to take over this problem and unpack why the issue happened, point to the code that's responsible for the issue, draft a fix on its own, and even flag when your PR is about to cause a problem. In other words, it's going to anticipate problems so you don't have them. is what you need, folks. This is the perfect use of superintelligence. Go to century.io and check out their free death plan. Plus, our listeners can use the code, twist 26, for $240 in Century Credits. That's S-E-N-T-R-Y. All right, let's keep going. All right, let's dive into some news.

Starting point is 00:20:45 Number one news story, please. Yeah, please. CNBC reports that OpenAI is close to a new private funding round of up to $30 billion at a $1.25 trillion pre-money valuation. KOTU and UAE-backed investors are expected to lead. The UAE money could be paired with a deal for more compute capacity in the region. That number is down from the $1.4 trillion valuation, Bloomberg reported the company was seeking in late September for the same $30 billion target Anthropic, which raised at $965 billion in May, is reportedly targeting $1.5 to $2 trillion for an IPO that could come first.

Starting point is 00:21:21 So the reported price chasing came down about $150 billion in a month. What is the biggest thing for you that changed between September and today? We don't have complete information here about the terms of the deals, protective provisions, et cetera. So there's a lot of nuance in terms of what the valuation means in relation to the investment. For example, if they mentioned how much they're raising in this round? They're looking for $30 billion. $30? Yeah, at a 1.25 trillion.

Starting point is 00:21:52 So now, do those investors have a liquidation preference where they are guaranteed a minimum of $2.60 billion or before anybody else shares in the preference act? Do they get 7, 10% interest every year until it converts if it's on a convert? Or do they have a minimum amount? Do they have the ability to call back the investment? So we had this happen with one of our startups. somebody said, oh, you know, that Series B we gave you, you had seven year, we had a seven year window for it, and then we can demand our money back.

Starting point is 00:22:25 The company's profitable, and so now the Series B investor is like, yeah, we don't believe in the company anymore, or we don't believe this growth here. Just give us back our $100 million. Now the startup's making $50 million a year, but it wants to grow, and now they've got to give $40 of the $50 million back each year, you know, or so. So there's a lot of those protective provisions, downside provisions that we don't know about. Anyway, the relative, what you have to look out here is just how much money is the business making. That's another thing that we, is there a mention here in this round of what the current revenue run rate is? And the revenue run rate, there's two different ways to look at this while you look it up.

Starting point is 00:23:03 There's ARR, annual reoccurring revenue, and then there's annualized revenue run rate. A RRR, like a pirate. People conflate these two numbers all the time. Sometimes people do it on purpose. Sometimes people do it accidentally. And I'm constantly, even with my own team, saying, what AR are we talking about here? And they're like, oh, well, they sell this product. I'm like, okay, if they sell a product and the person is not obligated to buy that product next month, that's annualized revenue run rate, which is take last quarter times it by four.

Starting point is 00:23:37 Or take last month and times it by 12. It's very funny. I'm looking at it up right now. their annualized revenue run rate is approximately $50 billion. And it doesn't explain how that was calculated, but generally, what was last month's revenue times 12? Now, sometimes people will sign a huge contract that goes away the next month, and it was a weird thing.

Starting point is 00:23:57 So they signed a $10 billion contract or had some special circumstance that month, and then all of a sudden things look really out of whack. So we don't have all these details here. That's the only number of seeing is $50 billion. Yeah. So all you need to do is look at the valuation here. You said it's $1.2 trillion. 1.25. $1.25 trillion. All right. So you divide these two numbers. You get to 30. It's 30 times price to sales ratio. Price valuation to sales, sales top line revenue, not profits, not gross profit, not EBDA, not community-based. Rivka, it's the computer the best. Iba.

Starting point is 00:24:39 That's my Adam Newman from WeWork, watched the show from Hulu or whatever it was on. Anyway, it's 30 times top line for an incredibly high-growing business. I think they've actually right-sized the valuation to a certain extent. If you told me I could put, I don't know, 5% of my net worth into this, 10% of my net worth at this price, I take a deep look at it. Really? I'd be like...

Starting point is 00:25:01 Yeah. Right now, really? You still think? Honestly, I think they've right priced it. I think this would be going for 100 times the annualized revenue runway if we were talking about it a year ago when the hype cycle was crazy. So what happens over time is that 100x price to sales ratio? In other words, they were making 20, and people were saying it's a $2 trillion company.

Starting point is 00:25:23 When that was going on, remember those discussions? It's like a $2 trillion. Maybe it's a $2 trillion. Maybe it's a $3 trillion company. Yeah, maybe it's a $4 trillion. And then you're like, well, what are you basing that on? Yeah, exactly. You know, Gavin Baker, Brad Gersner, their shareholders in some of these companies, they're bullish on the sector, whatever the case may be.

Starting point is 00:25:42 It depends on each individual. You start to look at those hyper-optimistic ones, but you have to pick a year. So that's what I always do. Let's pick a year. If in the year of our Lord, October 23rd, 2026, it's at 30 times price to sales. And the thing's doubling every year, tripling every year. I think it closes the gap. And it's probably a good buy if you were holding it for 10 years.

Starting point is 00:26:04 If you're trying to make a quick trade, God bless you because there's a lot of headwinds here. Obviously, people are, let's just go over the headwinds for a minute, right? The headwinds or the risk factors here. Number one, they're up against Claude and Anthropic, pretty great competitor. They're also in the agent space, which might run away with a lot of the market. And some people believe that's going to be the dominant modality, modality, fancy word, for what? Like the how, the interface?

Starting point is 00:26:36 Right. The distribution, the interface, like that's how people are going to access. In this context, you would say the modality would be like the interface or the delivery mechanism. Exactly. So if you believe that agents or the modality, well, then you have to say, wait a second, you got Muse, Google, Apple, and Microsoft who are all going to be major players in that space, whether they have their products out or not.

Starting point is 00:27:03 Everybody's piling into the agent assistant. And they have, those companies tend to have two, three, four billion users using their products collectively, and they have individual products that have well over a billion or two billion users. So there's an incredible opportunity in that space. That's a headwind, though, because that means everybody's going to have one of these. Amazon's going to have their own agent, you know, imminently with Echo or what is called Alexa, and Syria will come on strong.

Starting point is 00:27:31 And we have Muse coming out very strong. And Groch is very strong. And I mean, what about open source? We haven't even talked about that. I feel like that's the third headwitch. That's the other movement we keep talking about all the time is like, you know, are everybody going to switch over to running these open models that are so much cheaper? Number one, you have the other frontier labs.

Starting point is 00:27:50 proprietary enterprise grade. You're going at businesses with, you know, I saw Uber bought chat chip for 300,000 employees, I think is incredible deal. Now, of course. So instead of, you know, they're probably paying 20, 30 bucks a month each, you know, it's probably a $10 million a year deal. Well worth it for Uber. But I think you're going to start seeing these deals.

Starting point is 00:28:13 And this is like the Microsoft Office going into Uber or whatever it happens to be versus Google suite of products or vice versa. So you have competition for enterprise. You have the modality of agents being a 10-person race that just started. is like people are still joining the race. And then finally, to your point, you have open source. And then you have like what happens if they can't execute and put compute on fast enough to deliver these services? Then you have regulatory risk, which in some cases might be actually a tailwind because it might help them because they're able to execute on those regulations.

Starting point is 00:28:50 Other people aren't and they're onerous. So therefore, the bigger established people can afford to navigate them and the younger upstairs can't. So there's just a lot. to parse here. But if you told me, hey, I'm going to force you to take 10% of your net worth and buy, you know, of these 10 leading AI companies, you had to put 1% into each, or you had to put all of it into two, you know, like, I think open eye gets one of the slots. I got to be honest. Like, I wouldn't under, they're exceptional at execution. And even though they seem to be getting beat over and over and over again, they seem to come back. Yeah.

Starting point is 00:29:29 So that to me says something. You know, they get beat by Grock and Muse, and then they release thoughts. They get beat for code. They release codex. Like, they, they're fighters. I give the Open AIT and Sam, a lot of credit. They're fighters, and they ship product. And now that I do-

Starting point is 00:29:44 Anthropics ships better product, I think, slightly better. I mean, I do these AI demo posts now for that other substack. And every week there are, like, we're comparing the top, Claude model and the top Saul or, you know, like, that's a constant refrain. All these channels are always doing that they're putting them ahead to head. And the consensus is really like, it's not one is clearly pulling away from the other. There are GPT Saul fans. There are Claude Opus fans.

Starting point is 00:30:15 It goes really back and forth. So I don't know if there's a definitive, like, Claude is definitely better. I think it's task per task and a lot depends on your harness and your setup or whatever. I guess that's my question for the audience. I mean, are you buying out 1.2. It feels like a fair evaluation given their growth and products. I mean, if I told you, you could put 10% of your retirement one, of a certain age. Oh, I would do.

Starting point is 00:30:43 Yes, I would do it. No, no, no. I got 10% into open AI, 10% into SpaceX, 10% into Anthropics. So those are three bets. These are the three. No, no. You have to, 30% of your retirements going into those three, or they can, those, that 30% can be in a balanced, you know, index fund that you don't touch and is just set across the S&P 500 and NASDAQ 100. Would you concentrate on those three names with 30% at the current valuations?

Starting point is 00:31:16 And the answer should be. Yes, I would put it into those three companies. I would probably tilt it a little more towards SpaceX if I had the option, just because I feel like they're also competing on. the frontier model front, but they've also got, they've got Starlink, they've got the rocket business, they've got robots. So now we go to question two. Yeah. So now we've got a question two.

Starting point is 00:31:37 We've already said a youngish person, 50 years old or so with, you know, 20 years before they tap their retirement account, let's just put that as the bogey. They got 20 years in this thing. They're not going to touch it. Clearly put 30%, 40% into those three names if you could. Yes. Because you got to have something that's going for the hell of Mary. Okay, now let's just take the 30% number.

Starting point is 00:31:59 You got those three names. Split me up the percentages of those three names. Oh, so it's 30% total. I'm probably doing 20% SpaceX AI or SpaceX S-S-I as it is now. X-S-I-E. And then 5% each open-A-I-A-A-A-A-A-A-A-A-A-A-Ropic. I think that's a perfectly reasonable bet for you to make. I'm going to hedge the Open-A-I-V-A-Rososan-ROT.

Starting point is 00:32:21 You can easily go 15, 7.5, 7.5 as well. Right. And then you could rebalance it as you go. This is not investment advice, but this is what I lead. Please don't take it from me. But this is me giving investment advice to lock. All right, I'm here with Keith Pyrriss. He is the founder of Lightfield, and they are a CRM system built natively in AI.

Starting point is 00:32:41 That takes all your emails, calendars, Slack messages, all the stuff you keep losing, puts it in one place and creates organization out of chaos so you don't use any customers. Keith, welcome to the program. Thanks for having me. Let's talk about lead enrichment. You mentioned that. What is lead enrichment for people who don't know? It's kind of implied, but I think sometimes people hear that term of art and don't know what it means and how do you do it? Lead enrichment comes from, I think, the early 2000s where you used to buy a database of names, email addresses and other parameters like job titles, company size.

Starting point is 00:33:18 And then if you had one of those parameters, you could use it to find the other. Like maybe you had the name and you want to figure out, hey, where does this person work? What's their email address? Or maybe you had the email address and you're like, what company do they work for? What kind of company is this? And this used to be, you know, I would say a pretty big business because you've had these companies like Zoom Invo that would pay thousands of people to go and call and make notes and fill out this database. Annoy people.

Starting point is 00:33:48 Yeah, annoy people and fill out this database. And in the new world, you know, that's not how lead enrichment works at all. You can get to use one of these web, LLM tools, go and scrape the web, find the newest updates, and then connect it to your database. So sort of lead enrichment in our world is like use, you know, a modern LLM web provider like Exa and get them to go search through a million different documents and come back to you with an answer. All right. everybody go check out lightfield.ac and see what it can do for your company.

Starting point is 00:34:24 Don't lose any customers and find some new ones. Lightfield. All right. Speaking of SpaceX-S-A-I, Bloomberg reports that they want at least 80,000 cell sites across the U.S. They're in talks with Tower Giant Crown Castle, among others. The goal is a fourth nationwide wireless carrier, Jason. They're calling it Starlink Mobile. It'll be built on satellites, newly acquired low-band spectrum, and towers on the ground.

Starting point is 00:34:49 Crown Castle jumped as much as 10% after hours. AT&T Verizon and T-Mobile, the established players, of course, all fell around 5%. On this week, SpaceX earnings call, Elon Musk said, satellites plus spectrum will pave the way for Starlink Mobile to become a major mobile carrier in the U.S. I mean, my question for you, Jay's would be like, how many U.S. phone customers from AT&T Verizon do you see jumping over once Starlink mobile is an option? The Tam here might be a little bit bigger than people suspect and that this is incredible for

Starting point is 00:35:23 competition. So for people who are like, oh, my God, oligarchs, whatever, M&A, one of the great things is with Elon's companies and other people's, they get ambitious. They look at other people's highly profitable businesses that are in duopolis, et cetera, AT&T of Verizon being a perfect example. I think the two of them are the majority of subscribers, got T-Mobile with some there. as well. You add a Google, where they have Google Fi, but Google Fi never got this ambitious that they wanted to put satellites and towers. And then I'll add another piece to this, which is,

Starting point is 00:35:59 imagine you have your home, your five family members, all on the same plan for their house, for their cars, and for their mobile devices. And that, every car on the road has Starlink built into it, whether the person is using it or not. So you're driving, you're a model three, model Y, and you're driving it through a hill country, and you park at a cafe, and I park at the same cafe, I take out my phone, and I'm sitting there. The cafe's got terrible ISDN, whatever, just bad Wi-Fi or they don't have Wi-Fi, and that car's out front, and it's got a satellite dish, and it automatically takes my phone. And any of those SSIDs, Wi-Fi networks, and the Starlink, it just automatically connects to. And it connects to it securely with my authentication,

Starting point is 00:36:51 and I'm off to the races. Oh, wow. So it's like you're traveling with your own hotspot. I haven't even thought about that. Or I'm traveling with your hotspot. So I'm walking around Manhattan and there's a thousand cybercabs, or Austin, there's a thousand cybercabs driving around. And then there's a thousand private citizens driving around. And there's a mesh network of 500 megabit Wi-Fi. And I'm just walking in. You know, or sit in the cafe, work in my laptop, and I don't have to get on your Starbucks crappy Wi-Fi or this person's crappy Wi-Fi and wonder about security. It just connects me to whichever Starlink is strongest and not moving.

Starting point is 00:37:29 I'm moving the least. And by the way, there was a Mesh Network, open-source crypto project that did this. Somebody can type it into Claude and it'll tell you which one I'm talking about here. Mesh Network, I think it was Spain, and it was a crypto project where you basically contributed your Wi-Fi at home, opened your Wi-Fi network. work up and you earned coins, you know, tokens from some crypto project. And you can pull it up. And then there was another company that did it with buses in Spain. So there was like a bunch of Spanish entrepreneurs who were really into this like free Wi-Fi movement and like to the

Starting point is 00:38:00 point at which you wouldn't even have Wi-Fi at your house. There would be so many people sharing it that you could skip having it. Like if you were a student or, you know, you were a low-income person, you could just be drafting off of Wi-Fi constantly. Anyway, this crypto project where you earned crypto by opening up your network to be part of this mesh network. I don't think it ever became very large, but that model has been around for 20 years, people have been talking about. And I think Comcast, no, it's Xfinity. Xfinity has this. If you are near another Xfinity, you can log into your Xfinity account and use, you know, Lawn or I's Xfinity if we had Xfinity as well. So that's really the future of this. Putting up 80,000 cell towers, that's just

Starting point is 00:38:40 going to make it work through walls a little bit better. It might be this one, yeah. net, baby. Okay, sure. Common good, deploy social network. So, congratulations. I think it's going to be meaningful to your original question. I think it's going to cause prices to go down. I think we'll be sitting here in three or four years,

Starting point is 00:39:00 somebody will wind up clipping this, and instead of paying $300 a month for your home and your five family members, which is probably what people are paying right now, $100 for their home, and then five family members at $60 each. Easily. All right, so let's say it's $400 for the family members,

Starting point is 00:39:15 And it'll include the cars in it. You got two cars, which have $10 subscriptions each per month. You got five family members. They're at $60 a month. That's $320. And you're paying $80 a month for your home. Now you're $400. Yeah, that sounds about right.

Starting point is 00:39:29 Pretty. Now imagine it's $200. Now imagine it's $200. And it's just Starlink. And it's better. It's faster. I mean, theoretically it could be faster. I mean, I think AT&T, I think that's why AT&T and Verizon are very much on the fiber game.

Starting point is 00:39:45 And they have been starting to, Verizon has been starting to proactively tell its customers. You may have seen these stories, and I had it happen to me for one of my phone lines. They've been proactively saying, hey, you know, we can get you a better deal here. We can get your better deal here. And they've been offering me lower deals. So look up that story, that Verizon is now trying to get, to be, they're trying to save their customers money and they're lowering their prices proactively to people who are not asking for it. So, and then they tell you, hey, by the way, your bills $3 less per month.

Starting point is 00:40:13 So imagine that. You're charging so much that you proactively tell your customers that we're going to lower their prices. Verizon launches generous customer perks after raising prices. It has customers new benefits that extend beyond core service as wireless competition intensifies. Yeah. And they're going to lower prices, I think, is the key. You'll find some stories about them lowering prices is the one I'm looking for. But anyway, you know, and they can bundle stuff in there.

Starting point is 00:40:40 It's going to be awesome for consumers. Fiber to the home, that's hard for Starlink to compete against. Starlink will get to, you know, I regularly get two or 300 megabits down and whatever, you know, slightly less going up. It's fine. But, you know, I think this idea of one gigabyte fiber and even more is going to be, here it goes, Verizon targets, wireless customers with new simpler plans, drop some fees. This is the story I was referring to.

Starting point is 00:41:09 Yeah. And so they're trying to simplify. And because Starlink's like, yeah, you can turn it off month by month. Right. So that's exactly what they're doing. They're dropping activation and upgrade fees to make it simpler and cheaper to just stay. So folks, I'd like to invite folks to STFU. Shut the frack up about socialism, capitalism, and oligarchs and monopolies.

Starting point is 00:41:34 Startups break monopolies. It happens. And this is the great thing. And by the way, Starlink ain't going to get a monopoly either because these companies are not going to take it lying down, even if they get their asses handed to them, which is possible. They'll then merge and then they'll partner with Amazon's Leo service. Is it Leo? I think so. Low Earth orbit.

Starting point is 00:41:57 Did they name it Leo? I think that's pretty on the nose. Amazon Leo. That's exactly what it's called. Can you ask your producer there, what is the Amazon? on Leo launch date. When would they reasonably be able to sell this to people? Let's see what they say.

Starting point is 00:42:14 I think it's late next year. November. I know they're putting satellite. Oh, no, hang on. That's their next launch. Let's see. The launches are this year and into next year. I don't think that.

Starting point is 00:42:24 Full commercial service deployment and broader international rollouts are planned through 2027. That's all we've got so far. So they're saying at some point. Yeah, that's a mealie mouth. At some point next year. I think that means you reasonably could get this in 2028. I think it's 18 months off.

Starting point is 00:42:40 So ain't nobody buying it for 18 months. And if you did, it would be like buying Starlink. When the original people on Starlink, they would only sell it in certain neighborhoods. You put your zip code in, and then, you know, they would tell you if you had it or not. And then they would give you a warning. And they would only sell a certain number per zip code because there were only a certain number of satellites in the constellation, and they didn't want people to have a bad experience. So even though they could have sold more people, they were like, yeah, we just want to. I want people to think this kicks ass, and it does.

Starting point is 00:43:11 Yeah. Period. Full stop. So we were talking a little bit about all the name changes going on with SpaceX. Let me ask you this. In light of the shifting, it's SpaceX, it's SpaceX AI, it's SpaceX, SI. Maybe it's going to become like Musk Industries. It's going to become X.com.

Starting point is 00:43:26 What do you think would be the ultimate domain name for the company? They can have any domain. It's X.com. X.com. X.com. That's the one? Why? Unpack a little for it.

Starting point is 00:43:35 Why is X.com the ultimate? It's not that long. It's just one letter. He owns the X.com domain name, so I think just calling the company X would be fine. There we go. So, you know, domain name's so important. We like to highlight the importance of having a good good. Oh, you've got to get a great domain name.

Starting point is 00:43:53 I mean, if you could get it down to one letter, that would obviously be the ideal. But for any, even if you want a domain name that somebody else already owns that somebody else is already parking on, you can go to GoDaddy. They've got a broker service. They're going to help you get the perfect domain for your startup, even if it's just one letter. Yeah, buying and selling domains is a real art. And GoDaddy's going to be a great partner for you there because they'll anonymize your interest in things. If you're a startup, people will, you know, they're going to price gouge you. So you want to use the broker commission over there.

Starting point is 00:44:28 It's only 99 upfront fees plus 20% success commission. And these are buyer side charges. And so that's great. And, you know, they basically have taken every single thing you need to do with a domain and put it into one service with a fair price. And I've been on the service for 20 years. It's absolutely fantastic. So find out more godaddy.com.

Starting point is 00:44:58 Thanks to our friends at GoDaddy. All right. Here's a common founder scenario. You're looking over your sales numbers at the end of the quarter. And, hey, your team, they're at risk. of meeting their quota. This kind of thing happens all the time. You can't figure out what the specific problem is without taking a closer look at all of the data. Not long ago, that meant pulling it all in from across the entire organization, a time-consuming ordeal. But now today, there's Rippling AI.

Starting point is 00:45:23 Rippling is already built on your real-time people and business data. That means Rippling AI pulls metrics from both their data and your CRM into one dashboard. And then you get to see how close you are to hitting quota. And you can figure out the root of the problem of why you're not getting there. Maybe it's the marketing. Maybe it's the targets they're going after. You can break down all of that revenue,

Starting point is 00:45:45 by region, by customer size, all of that important information is going to be at your fingertips in seconds with Rippling AI. When you're running a company, you have important questions that need immediate answers. That means ASAP, as in right now,

Starting point is 00:45:59 as in yesterday. So if you're ready to rule your business, head to rippling. a.a. slash twist to get the only AI built to give you full visibility and take complex actions across your startup. Next news story. AI safety non-profits meter and redwood research got a deep level of access inside open

Starting point is 00:46:18 AI for a three-week pilot this summer and they found several vulnerabilities in its internal monitoring system. One exploit could have allowed agents to keep hundreds of their actions from getting monitored. They also found a potential path for codex agents to bypass security. restrictions while still acting with what seemed like their users' full permissions. OpenAI says it has patched everything affecting production safeguards. The evaluators say they have not yet verified any of these fixes. So OpenAI got a 61-page report. The public only gets a broad summary, and OpenAI has the right to redact it. So, I mean, my question to you, Jason, does this still count

Starting point is 00:46:56 as transparency if we're only getting these broad summaries? And does it seem like with all the rhetoric that we hear about the dangers of rogue agents, does it really seem like OpenAI is taking this as seriously as they should? They're certainly taking it more seriously. So let's give them credit for that. This is step one. Step one, we have an auditing agency. They're making a report.

Starting point is 00:47:18 And it's an outside agency. Now, if you're paying for the report, then there is an obvious conflict of interest. If you're selecting the company, you're paying for it. So when you see, I don't know, this sports. organization is investigating this person for, you know, some, you know, illegal or alleged behavior. They were doing sports betting. And they hire their law firm. Right. Yeah. That does their business. I've been involved in, you know, on boards of companies that hire an outside law firm to do an investigation. I would say, you know, you're going to get, because you're paying a

Starting point is 00:47:58 million dollars to investigate yourself, you will get the benefit of the doubt would be a very, what's the opposite of generous way of saying it. At a minimum, you're getting the benefit of the doubt. And you're getting the chance to kind of clean things up. So let's, you're getting the benefit of the doubt. They're going to take your kid gloves if they're, they're due, they're on your side. You've got a business relationship with them exactly. Yes. I would say they're on your side. Now, they're not going to cover things up. But they didn't cover stuff up here. I would never expect any of these companies to cover things up. But what they will say is like, hey, we're going to do this contract. We want 90 days, two or nine days or nine hours to resolve anything. Anything that's

Starting point is 00:48:45 resolved under this time frame is not counted in the final report. Right. It's been cleaned up. Sure. So things that have persisted for over 10 days, are in the report. You could negotiate something like that, right? If there was a misunderstanding and somebody bought their, you know, fiancé business class, first class tickets and their girlfriends are a private jet to go to Cabo, I'm not saying these kind of things have ever happened. But I think in the RJ Nabisco, right? That's what we're talking about it, Barbarians at the gate where it's the dog. Yeah, Barbarians of the gates. Somebody told the story here that they were somebody was sending their dog on the company private jet and listen I have three bulldogs

Starting point is 00:49:30 if you told me like hey jay cowl you're in you're at your you know you're in Tokyo would you like us would you like the pilots to bring the dogs and you know the kids next week I'd be like yeah sure bring the dogs and the kids I mean if I had that kind of cheddar I would do it but you don't want to do it on the on the company's dime so that is the key issue here and um I think we should all be stoked that this is happening. Now, the company meter, which I think is M-E-T-R. It is. M-E-R. Yes. It's a nonprofit. It's a nonprofit. Model evaluation and threat research. I think that this organization needs to be fully vetted. Yes. Again, not saying they're doing anything wrong, but who are these people running it? What is their relationships personally, historically, with the people working or running these labs?

Starting point is 00:50:19 Right. It's a small community. So imagine you are making movies and you're a very high-profile director. And you're Martin Scorsese and you make taxi driving, you make Goodfellas, you're Stephen Spielberg, and you make Polter guys. It's a bit scary for kids. You make Temple of Doom 2, which is where PG-13 came from. You're Steven Spielberg.

Starting point is 00:50:42 And they hire people who worked at Paramount and Columbia Pictures previously to run the MPAA because they wanted somebody from the industry who had credibility. It's a friendly group, yeah. And oh, by that, by the way, that person has two relatives and friends who work at Columbia Pictures at Paramount, and they worked on this picture and that picture. And then there's their at Sunday soccer game with their kids. And one person says, oh, yeah, Temple of Doom, we're going to rip this guy, Kalima. The heart's going to come right out of the chest. It's a pretty intense scene, but it's key to the storytelling here.

Starting point is 00:51:17 But I don't think it's kind of. cartoonie and there's no blood, it's just the heart beating. It kind of feels to me like a science experiment. What do you think? And they're sitting there like, oh, Stephen's a great filmmaker. And I know Stephen's kids on the field. I was Stephen's wife of K. Captcha is over here. I'm just trying to explain how the real world works.

Starting point is 00:51:35 The real world. The movie rating system works, basically. It is not that different. I think the other concern about meter is that it's people are like, how closely tied is it to the effective altruists and the Berkeley rationalists? And it's got a lot of people, there's a lot of overlap there that I think makes people uncomfortable. Like what is their, what is their vision and agenda for AI that is driving their evaluation? And I don't know, we don't know the answer.

Starting point is 00:52:00 I mean, it's actually a very thing to ask, Claude, who are the principals, who have the principles worked with at previously at the, at open AI and blog? Yeah, open philanthropy and coefficient giving, which are both big EA aligned sources. is there also substantial funders of meter? So this is where a lot of this kind of, what is their actual approach to AI safety? Is it aligned with most of the American? I need to know the conflicts of interest. I like when they have principal lunatics from outside an entry.

Starting point is 00:52:36 In other words, if the MPAA was based in New York City and it was based on Wall Street or it was based in Boise, Idaho. at, you know, adjacent to a university with the best, you know, psychology department and the people running it were people who were psychologists who were thinking about the trauma things might have on kids and what's the right age to do that. It's, you know, when you make your, when you make and select your organization to regulate you, you're going to have home field, home team, home field advantage. And I said that. I advised it. I said, to our industry. Regulate thyself or get regulated. Now, part of regulating yourself and getting

Starting point is 00:53:20 home field advantage, in other words, in the NBA, the shock clock is controlled by the home team arena. There have been instances where the shock clock, you know, shenanigans have happened. And people are like, well, why does that happen? It's like, well, because they have the equipment there, and they're responsible for maintaining the equipment, and you press the button when the person touches the ball and you count 20 per second. It's not rocket sign. Anyway, there's been always been accusations around the shot clock in the NBA. But things can be vetted. So again, you find you can regulate yourself and get that home team advantage and you might be able to moderately work the refs. But if you go too far and you get busted, it's a really bad look. And then you

Starting point is 00:54:02 could have liability. So I think this is a great start. I think the more important thing is cybersecurity. And we said this over and over again here. I've said it on AI, on All In over and again, the most important thing is cyber. People getting hacked is just so obviously the short to midterm challenge. They're not investing in white hat protection, and that's what they need to be doing more of. They need to know their customer more often. When these things are being sold by API, who's taking the liability? So if you buy tokens from AWS for OpenAI or Claude or Microsoft Azure, Or, you know, is Microsoft going to take the liability for how you use that? How much do they know their customer?

Starting point is 00:54:46 Are they auditing or doing spot checks that you're not doing cyber? If you are doing cyber, maybe you have to register. So here's an easy one. If you want to do cybersecurity, you have to register. You got to put your driver's license in as an executive at a company. You have to have your company's tax ID. And you have to do an onboarding call and explain what you're doing. And then you have to agree to random audits.

Starting point is 00:55:07 Yeah. I think people have the assumption that there's like a section 230 for AI. Like they can't be held responsible for what you do with the output. And there is not. That is not exist. I mean, look at all the lawsuits already. I mean, there are lawsuits about, you know, tragically, people having power social relationships and committing suicide. Yes, of course.

Starting point is 00:55:26 And even like doing violence that the AI helped coach them on, here's how to set up your mass shooting or whatever. There's a million of those kind of cases. Has there been a mass shooting that was planned on, uh, uh, uh, Open AI or Anthropic? I haven't heard about that. Factually, I don't know. But there are several cases that are out there saying that this person did this shooting and they were they were orchestrating it or they were asking the chat bot for help. And then there was that case.

Starting point is 00:55:52 I believe it was that Anthropic revealed that they caught that group of terrorists. I think it was from Iran. I'm not positive. Where they were actually like trying to use Claude to help them plan like a chemical attack or some kind of explosive attacks. So there have definitely been allegations that people have planned out their violence with the help of a chat bot. I would not, I mean, obviously I would not recommend doing violence, but they're not worth it. They're spot checking all this stuff. And if you do anything, oh, inside a messrs harrowing history with chat chpT, years of chats reveals glaring red flags and disturbing chat chapt replies from long before a rampage out of Florida State University, August.

Starting point is 00:56:37 4th, 2026. I mean, crazy people are drawn to these products. Let's be honest. They've been drawn to the Internet and message boards before that. So you can't call them like chat chippy T shooters, just like you can't call them Reddit shooters. They're shooters. I mean, we've also, I mean, that goes back to years. Like after a person does a murder, they look at their Google search.

Starting point is 00:56:55 And a lot of the time, it's like, how do I kill my wife? You know, like a lot of people have gotten busted for stuff like that. Well, that's research. And then there's inception and you kind of push the person to do it. Of course. Of course. There's great. Which was the suicide one.

Starting point is 00:57:08 These are two different things, and I think Google's not going to encourage you to do a shooting. Google's not going to encourage you to do self-harm. No. But in some of these earlier interactions, there were mistakes were definitely made. Settlements will be made. And it was kind of encouraging or leading people toward these tragic outcomes. Yeah. I think GPT-40, that one that was super, super sycophantic.

Starting point is 00:57:34 Did you see that? Jacob, if you can find this, I hate to put this on you. There was a chart this week that they ranked all the AI models from the last several years on which is the most to least sycophantic. Grock comes out right near the bottom. And number one is GPT40, that one that people were falling in love with and then they got so angry when OpenAI took it away. Anyway, if you get fired, I'll look it up after the show. Pull up this LinkedIn page. I'm just curious about this LinkedIn page as we wrap this story, which we probably gave a little too much air time.

Starting point is 00:58:02 It's an interesting story. Lon, Elizabeth, Beth Barnes. Oh, yeah. I know who this is. She's from Meador. She's from, yes, Berkeley, California. Safety Evaluation for Frontier models. Let's take a look at her related work history.

Starting point is 00:58:18 Let's look. Oh, what do you know? I recognize some of those logos. Please read. Let's see. Founder and CEO of Meeter full-time. Right before that, she was part of something called Pledgeer. I'm not sure of heard of that one.

Starting point is 00:58:33 You go right before- To present, present. It's still present. It must be like a side hustle nonprofit type thing. Something called pleasure. And then before that, researcher on the alignment team at OpenAI from June 2019 to September 22, three years and four months with Open AI. And before that?

Starting point is 00:58:50 Before that, research assisted to the chief scientist at Google DeepMind. That would be. Who is the chief scientist at Google DeepMind? Demisa Hussobis back then, I believe. Yes. Okay. So, you know, as I said, said earlier, like maybe like the person running the MPAA might have previously worked at

Starting point is 00:59:10 Paramount Pictures with Stevens Milberg. It's literally the person who founded Meeter was a live one of the alignment team at Open AI and she does the first report. Exactly. So we're getting this. I'm not saying she didn't do a good job. Could be a perfect job. But it does require a second level audit.

Starting point is 00:59:31 Somebody needs to audit the auditors. And so you can just say. see right through this. And it's not like she's hiding anything. It is what it is. Like there's a limited number of people who know these. That's what I was going to say. So having work there gives her an advantage. Yeah. I think when you're talking about the motion picture association, like anybody can watch a movie and be like boobs PG-13. But when you're talking about an AI model, you're a handfuls of people from around the world. You know, like a few thousand people maybe are actually qualified to do this. And so, you know, I think that that is part of the challenge.

Starting point is 01:00:01 What are the chances, Lon, that if she found that a group of people hid destroyed evidence and knowingly did unethical and illegal things during the hugging face incident. If she found that, what are the chances that she would release it? I mean, we can't know. We can't know what her relationship is like with Sam and with the leadership team and Greg Brockman. So then let me ask a follow-up question, because this is like a congressional hearing. Yeah, reclaiming my time.

Starting point is 01:00:36 So you worked, Elizabeth, you're Elizabeth. Okay. Elizabeth, you worked at Open AI. Is that correct, yes or no? Yes, I did. And you're now the founder and CEO of Meter. Is that correct? Yes or no.

Starting point is 01:00:52 Correct. Do you answer no question? Yes. Okay. Thank you. Do you maintain any relationships with your prior contacts? Your friends, your associates, at Open AI, yes or no. It's a yes or no question you're under oath.

Starting point is 01:01:12 It's a yes or no question, ma'am. You maintain relationships with any of your formal contacts at OpenAI. Yes or no, you're under oath. This is the halls of Congress, ma'am. Some, sure. Yes, some. Okay. I don't know why it takes you so much time to just say you maintain your relationships

Starting point is 01:01:32 with your previous employees. I maintain my relationship by pre-employees. They're sitting right here. I'm from the great state of Mississippi. I have friends from Mississippi. You're a Dixiecrat now. I'm a Dixie-Crat.

Starting point is 01:01:43 I don't know much about this open AI and a recursive, an RSI, super intelligence. I trust the president knows what he's doing. We've got the greatest president of our lifetimes.

Starting point is 01:01:57 But what I would like to know is why we have a wolf in sheep's clothing, watching the herd. Are you shepherding? Are you the wolf shepherding these sheep, ma'am? That's a yes or no question. No.

Starting point is 01:02:14 Are you a wolf? No, sir. No congressman. You're not a wolf. I mean, I'm just, I'm poking fun. This is what's going to happen. Sure. This is the inevitability of this structure.

Starting point is 01:02:24 How could somebody who, how could anybody take meter seriously? I mean, you literally have to take these people at meter and then, put them out of this process. So what Meeter needs to do is hire a team in another city of people who are not affiliated to this from law enforcement. This is what they need to do. They need to hire FBI agents who worked in cybercrime, CIA, and they need to have them go do these audits. We should, I mean. And if not, we should hire those type of people and start another organization that bets the vet. That's what I'm saying. I think at this point, it's been long enough. We know how important it is to have people who understand AI and alignment and how this works.

Starting point is 01:03:03 outside of these labs. At this point, like, we should just start training them ourselves. Like, there should just be an independent, you know, right now, if you're a AI genius and you graduate Stanford, you just get sucked up immediately by one of these labs. We need people who don't go that path who go into the public sphere and are helping all of the rest of us understand this technology. I think that's important. This is a final story we need to cover here.

Starting point is 01:03:29 Yeah, we'll talk about it. I think we got a full episode out of this. Yeah. It's great. We're good. Yeah, more than enough. Do you want to do this? This is the talk.

Starting point is 01:03:38 This is now a very significant story. Everybody's buzzing about it. A lot of people have left California, threatened to leave California over the billionaire's tax. It's an asset seizure. It's a particularly dangerous one for startups because you're illiquid. If you're one of these AI companies and you've got a billion dollars in shares, you can't sell them, you've got to give 5% of your net worth. It's a big check. Now you've got to dump stuff.

Starting point is 01:04:02 on the market in an illiquid market where you've got to come up with a payment plan. This is really challenging for folks. And that's why a lot of people have left. And people are mobile. So people are leaving the state because they need to get out of these crosshairs because it's not just the money. It's the onerous nature of how this money is going to be collected. And it's an asset seizure. I saw Representative Rokana a favorite of all in talking about this. And I'm just Curie, I don't know the answer. I'm genuinely asking the question. He suggested that in response to this exact criticism you bring up that you're not, it's not cash sitting in a vault. It's stock that you own in a company that's made your net worth a certain number. He said, well, then the government should

Starting point is 01:04:48 just take the stock, like a portion of the stock and then the government would hold on the tax in lieu of you paying your taxes. So like, seize my home. Does that, well, I mean, aside, there's the ideological side,

Starting point is 01:05:02 but then just on a practical level. It's a constitutional. It's beyond ideology. It's actually constitutional that these taxes can't seize things. You can't, you don't have to pay taxes twice. Putting it all aside. Right.

Starting point is 01:05:11 Aside from the fact like it's morally or ethically wrong, just would that, would that work, do you think? Like, if you tax somebody, their stock in a prime company. No,

Starting point is 01:05:19 because now I'm giving you my, let's say I'm like the co-founder of 11 labs or, you know, whatever. Now I'm giving you, yeah, I'm giving you my granola shares and then the government's going to hold those and then the government's got to decide when to sell them. What if they decide they're just going to dump them on the open market and find a buyer? Who's going to do that in the government? Oh, now we need an organization of the government. That is a secondary sales. You know how much secondary? They'll have worse secondaries going on than Angel Listen. What if you made the rule? You have to hold it. And then if the company goes belly up, if they go out of business, that's not worth anything. Taxpayers went for a little ride. And then if they, but if they go public or they get acquired, taxpayers get their cut just like as if they were an investor. Here's the long and short of that. Then just raise people's taxes three or four percent a year. Okay. And then put a tax on margin loans.

Starting point is 01:06:08 I can tell you as a person who's not in this group but is in the next group down and I'm looking at it going, wow, I got out of Dodge quick. When taxes go above 40, 50 percent, you really do need to think as an affluent person, what am I getting for these dollars? And when it's certainly when it hits 60% like Denmark, you basically have to leave. You have no choice but to be. Well, and you don't have to go that far. You don't see me living in Puerto Rico trying to, you know, pull a Brock Pearson be the mayor of, you know, whatever. Like all these cryptic guys went there. It's like pretty obvious.

Starting point is 01:06:42 They're paying one percent tax or something. There was some kind of tax holiday I read about. And they were more than willing to tell you all about it. I got lobbied to move there. And I'm like, yeah. Yeah. But, I mean, Puerto Rico is not like, I mean, I'd love to go to vacation in Puerto Rico. Rico, but it's a little bit...

Starting point is 01:06:56 I think the old San Juan is very nice. That's where they all are, apparently, but it created a lot of attention in the local community, actually, because they were kind of hot-shodding around crypto people who are insufferable at times. So this whole thing would be mitigated if they just said, anytime you take out a margin loan, you pay the taxes on that margin loan or some percentage of tax, as if you had sold your shares. And then if you pay the margin loan back, you get the credit for that tax, you get a refund.

Starting point is 01:07:19 You just get it refunded. I think that's fair. Anyway, it's way too complicated. But it, Rocahn is, by the way, the ultimate hypocrite. He's worth $400 million. And as I told him when he was on the pod, why don't you, or on Twitter, I said, listen, you're worth $20 billion. You're worth $400 million.

Starting point is 01:07:33 Why don't you give just $20 million to this and lead by example? The key thing that's happened here, which is going to be extremely polarizing, I believe, even though the survey say the billionaire tax won't pass. I am 90% certain it would. It will. And if there was a polymarket here, you and I can make a trade here, and I think we're going to make a little bit of money on it. So let's find out. So pull up the polymarket chances of this.

Starting point is 01:08:00 Is it SB 40 or something? I don't know what the name of the thing is. Yeah, I'll look it up. I don't. I have Jacob look it up, and then he'll pull it up when it's ready. So now Zuckerberg has cut a deal with his board. And, okay, here we go. 23% chances passes.

Starting point is 01:08:19 Now, is that becomes law or is just passes? because then get like, is it? Yeah, I mean, passes is passes. I think you can't stop these things once they're going unless you do another one and you intercept it and reverse it in the future, which I think would be two years later, right? Yeah. So if I, if my understanding is correct. So here, if I put 10 dimes, if I put 10 times on this, put the amount of 10,000 and we trade it here, what are we making here? If I put, we'll make it 30 grand.

Starting point is 01:08:46 30, yeah, that's nice. It's a nice little thing. I may place this bet. I may place this bet. Now, if you were Zuckerberg and there was a liquid market here, you could literally put a billion dollars on this and make $3 billion pay your tax bill. And that's actually what the polymarkets and in trades previously and all these prediction markets were. Originally, the concept was to create hedges against bad events, kind of like travel insurances, right? So if you could do this for ski trips, like Sam Leston was saying, like if your ski trip gets, there's no snow or it.

Starting point is 01:09:21 global warming or your ski house, you could kind of bet against yourself or that outcome and have some upside. Or when I went to Knicks games, I was paying five to $20,000 a seat sometimes. Yeah, but you were betting on the Knicks. You weren't hedging. I was doubling down. I was betting on the Knicks. And I won every time.

Starting point is 01:09:38 I've got free tickets like half the time. I didn't bet every game. But I bet a couple of them. And I'm just like, oh my God, I just spent $10,000 taking my brother to the Knicks final game. I think we spent five-garing a ticket to be there for it. And then I bet on the Knicks. The Knicks weren't the favorites.

Starting point is 01:09:53 I made like $15,000. So I was profitable for me to see my Knicks win their first championship in 50 feet. So just broad strokes. And you can give us whatever like CNBC's mini summary or whatever the AI summary is. I got a docket. Okay, yeah. So walk us through the series of events here. And then I'll give my refined commentary.

Starting point is 01:10:11 You can form a question. So META has struck a deal with Mark Zuckerberg to keep him from leaving California over a proposed billionaire tax that goes to voters in November. The measure to one-time 5% tax on residents with more than $1 billion. For Zuckerberg, whose wealth is estimated at more than $200 billion, he could be looking at a $10 billion tax bill. Meta's fix is a stock split that creates a new class of shares. Zuckerberg gets more stock more often and can sell it quarterly without his voting control

Starting point is 01:10:45 changing. Meta calls it a founder alignment. It runs for eight years as long as Zuck stays as a California resident, and it appears to apply even if the tax fails. Do you think other California billionaires will follow suit and make these kinds of arrangements with their company to keep them in state? Okay. So is this the Facebook corporation's job, Mediske job, to keep him in California?

Starting point is 01:11:13 That's his issue. I was kind of wondering the safety, too. Like, what do they really care? Who wants the controlling shares in this company? Zuckerberg. Correct. So therefore, the board, like our earlier story about meter and the overseers overseeing the people who work for them, if you're on the met a board, that's a great position. If Zuckerberg was like, Jay Gow, I'd love to have you on the board. You know, you've been critical, but I need to have

Starting point is 01:11:34 that critical voice in the board room. I'd take the job in a second. And whatever he gave me, I'd donate to charity. Would love to do a public board at some point. But I don't think that's the board that's there. I think the board that's there are the people who invested in him and who, you know, I'm not going to say that they're to do his bidding, but they're there to do his bidding. It is a stack deck. So, and I think if you were the voice, if you were a dissenting voice on that board and were going against his wishes on some consistent basis, you probably wouldn't be invited back at some point, probably.

Starting point is 01:12:06 So let's pause there for a second. And my guess here is he asked them to do this. And he wants to stay in California. And it is arguably on the margins in their best interest. But if he was on the Nevada side of Tahoe, and a lot of people have moved to Nevada, whether it's Vegas, Reno, or Inclined Village is the really hip, beautiful. And the homes there cost four times as much as the homes on the California side. Like literally, three or four times as much.

Starting point is 01:12:37 And people have been buying $100 million, $50 million, $25 million homes over there. It's gorgeous, great place to live. And also conveniently, you can just get in a puddle jump or flight for 40 minutes and get to your office. in Santa Road. There's no reason he couldn't live there and just not be in California for whatever number of days it is. So what this is going to do,

Starting point is 01:12:58 this is going to be weaponized to get the billionaire tax to pass. Now, when this was announced, did it change the polymarket? Let's pull the polymarket up. Pull the polymarket up here for a second. Okay, so did this change in the last week? I mean, this story was October 22nd.

Starting point is 01:13:16 So this was only a yesterday story. So you'd really be looking for, did this change between like Tuesday and Friday? Yeah, put it on one week. You're on one month. So it didn't change things. Not your matter. I think this is, and what's the volume here? Oh, four million in volume rounding up.

Starting point is 01:13:33 That's interesting that it didn't change it. I think this is going to be weaponized. Roe Con is going to say, look, these guys can afford it. The board's going to pay for it. So somehow the board is going to smooth this out for him, make it more palatable, maintain its control, et cetera. And they're going to say, see, see, we can seize their assets, which is what I find offensive about it.

Starting point is 01:13:53 I don't mind people paying more taxes. I don't mind there being a minimum tax. Let's say, you know, another way this could be done is if you've held the stock over 10 years, 10% of it, you have to pay tax on 10% of it. So if I held my Uber shares over 10 years, whether I sell it or not, I got to pay my capital gains on 10% of it. In year 10, 20% of it in year 15, 50%. percent of it in your 20, whatever it is. I have to pay 10 percent every 10 years, right? You just have

Starting point is 01:14:19 to pay some tax at some point on your holdings, even if you're long your holdings. You know, that would go a long way, I think, to solving this. We do have polarization and wealth. The other solution to this was mine that I brought up on all in this week, which will be come out later today, you'll probably hear it tomorrow, where I said, you know, raise the minimum wage. And obviously, my besties are laughing at me. You're a mid. You don't know what you're talking about. They're actually wrong. In this case, they haven't actually refreshed their data. If you refresh the data, I believe any reasonable person looks at the 15 million people in America

Starting point is 01:14:51 making under $15 a year and realizes it's inconsequential to any of the company's paying for it. And if a couple of thousand jobs or 10,000 jobs are lost because people can't pay $15 an hour, well, maybe those jobs should be automated or maybe the company should take it on the chin and just be like, you know what, I have a minimum I pay people. The minimum I pay people, I think, is all in 70K plus benefits. We have people getting paid 60K as entry-level jobs with a 10-K bonus, and some of them make 20-K bonus, so make no bonus. They have to fight for the bonus. So at a minimum, I'm spending $75,000 for anybody who works for me, excluding the Athena assistance.

Starting point is 01:15:26 The reason I do that is it's a bad look for me to be paying people 40 or 50 or, you know, K. In a training program, I'll pay people 60K with a 10-K bonus, but I just feel uncomfortable. And, you know, I just have raised salaries in a lot of cases not to be more competitive. But because I feel it's the right thing to do. Yeah. So where's the esprit of corpse in this country, raise the minimum wage? And then somebody said, hey, you would make a great press secretary for President Trump. Chamaa thinks it's actually a good idea.

Starting point is 01:15:54 There's a group of people who are like, this is actually a good idea. I said, you know what? I would sit down with Trump. And if he agreed to do like these three things, I would come in and do it. I mean, I just. For the LLs to see if I could last more than a sacramuch. I would do it. And those three things would be, one, raise the minimum wage to 15.

Starting point is 01:16:10 Two, do a project where you guarantee that you're going to build one million new homes per year in the United States that are somewhere between two bad rooms and three or four bad rooms for families. And you can do that through any means possible, but you have to like create federal laws that the states can't preempt, that they're going to allow people to upsize and build and take out red tape. You can't block it in other words. In California, you did this. Near urban centers. I think that's the, like, people need to be able to live near where they can get work. I would just say within 100 miles, I would say within 50 miles of an urban center, sure. But there's plenty of ways to do this, by the way.

Starting point is 01:16:51 One of them is to just say any train station, you got to allow 50-story or 30-story buildings. And if people don't like that, well, then the train should skip your station. So where I live previously, you know, Redwood City, San Mateo, Milbray. They're trying to get them to do this, and people have fought it. Silicon Valley people have fought it. And it's like, hey, guys, and we should just not stop the train there anymore to penalize those communities. Now, some of them might like it because they may not like the people making $60,000 a year stopping at their station until they realize it's their nanny who's making $90K but can't live near them and hates them for going an hour and a half. Anyway, just a million ways to do it.

Starting point is 01:17:26 You can do it through tax incentives, whatever. But if he just made that commitment, just like he's made the commitment to data centers. And then the final one is, I think we should have, we really need to have a discussion about superannuation. funds and retirements, which Invest America kind of does a little bit. And then the fourth one would be around trade schools. There needs to be something for the working man and woman of the country. The bottom 40% that don't own equities don't get to play in the sandbox we play in. And this kind of behavior is just going to signal to people that the wealth tax is the way to do it and that you can beat the billionaires. You can beat the billionaires. You can eat,

Starting point is 01:18:04 Bernie Sanders. You're going to beat the billionaires. We're beating the billionaires. We're on the heels. Look, even Mark Zuckerberg. I think that's what... Even Mark Zocobar can't fight this. That's what gets really driven home to me is like hearing you discuss the nuances and, oh, we could do this instead or oh, we could do this instead. For most people, they're just, they're not invested in this system at all. That's all gobbledygook to them. It's like, I don't care of the nuance or just take the money. Like, we need the money. Gases out of control expensive. I just, I don't care what the rules are. Like, I just, that guy's rich and more of it should come back to the people. Yeah. Okay. Awesome. Well done. Thank you. Thank you. Do you want to go off duty?

Starting point is 01:18:44 Should we look at some videos? Oh, do you have off duty for me? I would love some Lon Harris off-due. It's like my favorite part of what you do here. Oh, okay. Off-duty me. All right. Well, I did, I wanted to ask, we talked about- Part of why I keep you around. We talked about this on the VC roundtable. We did almost an entire episode about the Nathan Fielder, Elizabeth Holmes documentary. You can see everything. I've been talking to people about this conversation. I want to pick your brain about something. I think in the world of startups, in the world of early stage venture, the idea that sometimes you're going to put some money on a fraud is very common. Like it's exactly what Sam Lesson was telling us this week, that like 1% of his fund is probably fraud and he's, you know, that's okay. And it's early stage.

Starting point is 01:19:28 You don't have enough time to do due litigants. I think most people don't understand that at all. And they think that's what's outrageous about Elizabeth Holmes story. We understand that it's the scale. It was so brazen. It was going on for years. She raised almost a billion dollars for this fraud. That makes it stand out, but not the fact that she was lying to people or being an entrepreneur and kind of.

Starting point is 01:19:48 And I think, do you think that that's impacting how people view Elizabeth Holmes that they think fraud in companies is a lot more rare than it really is? So you're saying people. If you're not in the tech industry, if you don't know anything about venture capital or early stage venture venture. or entrepreneurship, if you're just somebody you work at a restaurant or something, you probably think that civilians. Right. You probably think that people starting companies getting investments and being completely full of S and lying and doing fraud is very rare.

Starting point is 01:20:21 And it's only if you invest in a lot of early stage companies, you recognize like one out of every 500 founders is blowing smoke up your butt and you just have to seek them out. Well, blowing smoke is one thing. Fraud is the other. So let's tighten the question. I'll reinterpret the question. I think that this film is brilliant. I think it's the best documentary, along with Hoop Dreams, which I need to rewatch if I was going to calibrate this.

Starting point is 01:20:43 But I would say Hoop Dreams, this, King of Kong. Did you ever see The Act of Killing? Have you ever seen that one? I was thinking about that. No, send me it. I'll watch it this week. I'm looking for a film. Yeah, it's, anyway, go.

Starting point is 01:20:53 Tell me. Oh. Don't tell me. So it, there was in Indonesia, the regime that is still in power in Indonesia came to power. I want to say in the 60s and 70s. And one of the things that they did, they were anti-communist. It was like a military occupying to takeover. And they would hire and pay off local criminals and gang leaders to go to the homes of suspected

Starting point is 01:21:17 communists, round them up and murder. Okay, that's enough. Yeah, that's enough. I'll watch it. Yeah. Let's not do any more spoiler. Okay. I think it's perfect.

Starting point is 01:21:24 That's your number one? I mean, it's an amazing documentary because Noah Oppenheimer and this- you put it above this one? American film crew. I think it might be one of the greatest documentaries ever made. They were going to go just make a documentary about this scenario. I was just telling you about the aftermath of this horrible violence and all the millions of people that died in this conflict. But then they realized, like, a lot of the people who did that violence, they're still prominent citizens.

Starting point is 01:21:48 They're with the back by the government. They're the oligarchs. They're the rich guys in town. And they could just get interviews with them and talk to them. I won't go any further. It's an incredible movie where they sort of confront these people. Say the name one more time for the audience. The act of killing, it's called.

Starting point is 01:22:02 It's amazing. I would also say Grizzly Man that Werner Herzog one is. Oh, yeah. You know what I'd say? I'd say grizzly man is above hoop dreams for me, but below you can see everything. So now I'm starting to get a list here. That's my one, two, three, I think, right now. Okay, let's put all that aside.

Starting point is 01:22:19 To your question, I was talking to an actor who is A-List, A-A-A-A-List, like superhero actor, who has actually played superheroes, like top, top, top, superhero. Put all aside, I don't want to disclose who it is. But we're talking about the film recently, and he felt like, you know, being an insider in the industry, that the film did a disservice because they happened to understand this specific story. They're like, where's all the, like, actual things about her fraud? And this kind of gives her a pass. And I did, like, through that lens, it made me reconsider it. I think they do give her a pass in order to focus on Nathan Fielder's more humanistic, performative, you know,

Starting point is 01:23:02 a genre bending, truth, search for truth bending, reality bending, schick, genre, whatever, format. There are very clear clips of Elizabeth Holmes saying things that were just simply not true. That he could have gone to and he could have used them as evidence. Yeah, he picks these sort of edge cases. There's one that's going around that's going viral right now that's from 60 minutes where she says, you know, we did this study with Johns Hopkins and it revealed all of these fascinating things about the box. And then 60 Minutes just does a hard cut and like,

Starting point is 01:23:36 we called Johns Hopkins and they've never worked with Theranos and nobody there even knew what she was talking about. I can explain that one. I saw that clip as well. We're on the same Algo. What happened was, you know, she, again, with her, and many entrepreneurs bending the truth, you know, she didn't say we're working with,

Starting point is 01:23:57 we should pull this clip up. Go find it, guys. It's easy to find. It's 60 minutes. It's been reproducing those clips. The thing that's interesting about that clip is the way she frames it is. We're working with some people there. So let's play this clip.

Starting point is 01:24:10 And we're going to do our own Nathan Fielder here. Right. Yeah. And this is a different Elizabeth Holmes than you see in the film. This is what she's still doing, her Steve Jobs' turtleneck, deep voice bit. Yeah. We've done some work with people at Hopkins who have developed

Starting point is 01:24:29 and demonstrated that in blood, you can see the onset of pancreatic cancer 17 years before a tumor forms. We call Johns Hopkins medicine. They told us they never collaborated with Serenose. Okay, she said, we've done some work with people at Hopkins. Now, is it possible that John Hopkins is unaware of this work they did with some people at John Hopkins? So if they did some work with the person on that study and that work was that person came to their office and held a seminar with the team and they explored that person's work and commercializing it. Would that be correct to say they've done some work with people at Hopkins, not Hopkins, but people at Hopkins?

Starting point is 01:25:20 So if you're going to litigate it, would you agree, Lon Harris that she did do some work with people at Hopkins. Was that work when they did it in their offices in the conference room? They explained the technology. They shared tech. They did a technology transfer. They explained each other's technologies and they talked about potential ways to collaborate and the collaboration didn't happen because they put her in jail. It could be another one of those scenarios just like plays out in the film where she finds a semantic way to be like everything I said is consistent with this version of truth. But it is obviously misleading because the way that she's phrasing it makes it sound like they did a study in collaboration with Johns Hopkins.

Starting point is 01:26:02 Yeah. Please read to me the statute that makes a misleading statement to some people a crime. And I can actually do that for you. All right. I'm looking at it up, but if you've got it. No, there is no misleading people at a conference. You could be misleading people if it was in a sales pitch to buy equity in the company or to partner with them.

Starting point is 01:26:25 And that's what she got in trouble for was defrauding investors, not patient. investors. These are the nuances and the contours of the case that the film just doesn't take the time to go over in the three hours they have. I don't think it's very interested in litigating the specifics about Theranos and the alleged fraud and exactly what she did. It's way more interested in like the tension of getting to know somebody and not knowing if they're lying to you. It's like it's it's about that. Anyway, I just thought it was interesting that I think a lot of people have a different take on the film if only because they think that, why defrauding your investors is like a red line? The term is materially false and this one probably wouldn't be materially false. It would be false or it might

Starting point is 01:27:13 be misleading, but it wouldn't be false and it certainly wouldn't be materially false. Materially false would be and we've got a product coming out Q3 next year. Right. And it'd be like, well, you only talk to that person one time. You never followed up with them and that person doesn't have a contract from you or anything. And if you said we have a partnership with them, that would be material. false if there wasn't an ongoing partnership. If you said we've worked with them, we've explored, you know, those things might be weasily words or, you know, exaggerations on the margins. But I believe many people are going to come away from this and say it's going to create as many people who think she was railroaded as people who think she committed the crime. Yes. I think a large number of people

Starting point is 01:27:52 are going to believe she was misunderstood. I've also seen a lot of people reacting to it's always wrong to put a new mother in prison. That's like a whole trend that's going and now is like, it's so sad to see her have to leave her babies and how can we do this to mothers? That was the hardest thing and the thing. And that's, and I did a tweet about this where, you know, I've gotten into it with her or apparently Billy Evans on Twitter. I thought I was talking to Elizabeth Holmes. So speaking of material fraud, I thought, I honestly thought I was talking to her. I thought it was being, you definitely were talking to Billy. No, I thought it was like the situations where I interacted with her in car, as she's been incarcerated, I thought she was,

Starting point is 01:28:27 somebody was sending her the tweets, she would form a response and email because I know she has email access. Or even, yeah. So that's what I thought was happening. Visits, they would be like, what do you want me to tweet or something? There could theoretically be a way, yeah, but no. That's what I thought was happening. But anyway, I said to her as a Christian, I feel bad for her and our family. I understand that we have a justice system.

Starting point is 01:28:47 And so, you know, you could have Christian beliefs or Bushito beliefs like I do, both of those. They overlap 80%, 90%, quite nicely. where they disagree is quite material. And I'm exploring that in my own faith on my own time. But, you know, I told her, like, I feel terrible that you and your family are struggling through this. You could have sympathy for a family whose murdering family member is facing the death penalty. There's this woman who they tried to put to death with two shots, and it didn't work, whatever.

Starting point is 01:29:19 The vein didn't get the drugs in. So brutal. So brutal. Something. I'll look at it. Anyway, it's just so brutal. And I could as a Christian or, you know, my Bushido faith could lead me to believe, like, I have sympathy for them while, yeah, I just, I understand the state has its own way of exacting the law. And that's one of them in the society we live in in certain parts of it. Anyway, I think it's, most people are going to come away with this with possibly the wrong, with, with a, they have led the public, perhaps a little bit. bit, which may be genius that they did this, because it's so personal, because you, I was,

Starting point is 01:29:59 I was emotionally impacted when she says in the film, micro-spoiler, that she had pumped milk to have for her babies as she's leaving for prison. And then it just made me think, like, she having, you know, having been through this and watched motherhood, a beautiful thing, and women, you know, have to struggle to keep up with milk production. It's a whole thing, and it's emotional, and it's exhausting and sleep. And it's just a motherhood is just, wow, like hard, challenging and beautiful and all those things. I just felt so bad for her and her kids. Like, what if she can't get them the milk? And now the kids have to wean off of the milk. And you've been told you're a bad mom if you can't produce milk and there's drugs to help you

Starting point is 01:30:37 produce milk. And women go to lactation consultants. I've been through all this folks. So I have great sympathy for her in that regard. And then I also think, well, we live in a society that has penalties for certain crimes. And the justice system get this right or not? Well, that's what the appeals are for. So I really don't think it's mutual. Yes, that ability. I think you could say, like, she committed a crime and she belongs in prison for that crime, and those are the rules. And you can feel sympathy.

Starting point is 01:31:01 Like, obviously, it's a very difficult thing to be a new mother and have to go away from your family. All right. I have a, I have a off-duty. This is a 100-page novella, a short stay in hell, Stephen L. Peck. A short stay in hell, Stephen L. Peck. A influencer who reads a lot of books was in my feed. I may have mentioned I have a grok-bock, things to watch, read, and to, you know, consume or travel, whatever, like a media, I have a media bot. Sure.

Starting point is 01:31:31 And so I saw that she had read this, and I thought, this looks interesting. It's 100 pages. I'm trying to read a little bit more. I didn't know it had 18,000 reviews, four and a half stars. And it's just about somebody who goes to hell who's a Mormon and he, you know, picked the wrong religion. Right. It's based on a book. It's based on a story I have read, the library of, Babel by Jorge Borges. That's a very famous short story that I guess inspired this one. So there you go. Okay. That did and I need to read that book. Yeah. It's just a, it's just a story. A story. I got to read that story because they've referenced it many times here. Anyway, I've really enjoyed it. You should read it, Lon. And any books or media that you read for the show,

Starting point is 01:32:15 you're welcome to expense them on your. Oh, that's so, yeah. I read, man, I wish I had time. I read Richard Socher's book, This Week, an AI guest. I read his book, The Eureka Machine, which is really interesting about it. Oh, really? Yeah. I have a copy. Yeah, they sent us a copy, which is why I gave it. And he good?

Starting point is 01:32:31 It's fantastic. It's really good. It's very clearly lays out, here's how AI is actually going to make science and medicine. Like, here's how the cycle will actually work where we actually make these discoveries. It's not so abstract. He, like, really walks you through the whole process, which is really interesting. like, and it's, you know, wet labs. It's really how he imagines the process is going to go over the next, like, 20, 30, 40 years

Starting point is 01:32:57 where AI starts actually cracking these big scientific and medical challenges for us. So highly recommended. I thought it was very interesting. Cool. All right. That's you this week and start us. We'll see you next time. Bye, bye.

Starting point is 01:33:10 Bye, buddy.

This Week in Startups - OpenAI’s $1.25T Raise, Starlink Mobile & Zuck’s Tax Deal | EP 2354

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.