SPEAKER_00: We have been talking about secondary markets coming back. I know this because people will email me and say, hey, would you like to add to your thumbtack, your comm, et cetera, position? Would you like to sell some of those positions? So I am seeing increased activities from this very weird gray market, or it's kind of like boiler rooms of people who want to charge you 5% or 10% to buy secondary in SpaceX, whatever is hot, Andrel, Stripe,

SPEAKER_02: or, you know, and then charge you 5% for selling it. So it's a little bit of underground, a boiler room, a back channel. Yeah, because there are people in it who are above board and, you know, are known. And then there are people who are not. So I get a little nervous sometimes when I get these weird emails, if I'm being phished or whatever.

SPEAKER_08: Oh, wow. So there's like a little Wolf of Wall Street, like they're kind of bending the rules a little.

SPEAKER_11: This Week in Startups is brought to you by Northwest Registered Agent. Starting your business should be simple. With Northwest Registered Agent, you can form your entire business identity in just 10 clicks and 10 minutes. From LLCs to trademarks, domains to custom websites, they've got you covered. Get more privacy, more options, and more done. Visit northwestregisteredagent.com slash twist today. LinkedIn Ads. To redeem a $100 LinkedIn ad credit and launch your first campaign, go to linkedin.com slash thisweekinstartups. And Lemon.io. Hire pre-vetted remote developers. Get 15% off your first four weeks of developer time at lemon.io slash twist.

SPEAKER_00: All right, everybody. Welcome back to This Week in Startups. I'm your host today. Wow, a trio. Lon Harris is with us, our editorial director here at This Week in Startups. And Alex Wilhelm. He's Alex on X Twitter. And he's at LONS on X Twitter. We've got a full docket. Things have been building up. And gosh, I don't know where to start here, but let me just do this. What is your top story in terms of like important business story and then your top story in just fascinating link? Because I think they tend to fall in one or the other.

SPEAKER_14: I want to know your most fascinating story on the docket. You can see the docket right now this week in startups.com slash docket. Most fascinating, most important business. Those are the two choices. Alex, what's your most important business topic

SPEAKER_16: on the docket? I think the most important business topic on the docket and in the world today

SPEAKER_17: is the threat of iPhone-specific tariffs on Apple. I think that is indicative of-

SPEAKER_00: No, no, no, don't give too much. Just succinct so that we can keep moving here because we're going to go to each of these topics. All right. Lon, what's the most important business topic? Succinct, please. The audience is very busy. I want to talk about the new anthropic models. Okay, right. Don't tell me why. Just what it is. Now, on the fascinating tip,

SPEAKER_22: what's the most fascinating thing here? What's the most fascinating item? Alex.

SPEAKER_24: Oh, I know mine.

SPEAKER_22: Okay. Smart toilets.

SPEAKER_25: Smart toilets for Alex.

SPEAKER_20: Okay. Lonnie Donnie? I think we got to talk about OnlyFans for sale.

SPEAKER_00: Interesting. OnlyFans is for sale. I did see that. So we got smart toilets. We got OnlyFans. I am going to go with Lon. I think the OnlyFans for sale is absolutely a fascinating turn of events. Like, why would you sell a money printing machine? I got to understand why.

SPEAKER_30: It's like the easiest business in the world for OnlyFans, right? What do they do?

SPEAKER_31: Well, I mean, wasn't that guy pulling out like hundreds of millions of dollars a year? I don't know what's going on here.

SPEAKER_00: Maybe he just wants to sell. Billions of dollars, yeah. Sell at the peak. And I'm going to go with you, Alex. Apple tariffs are the most important story. Let's go with the Apple tariffs. What's going on? Is this just another 72-hour?

SPEAKER_33: We should wait and see what's going on? Or is this reality? I did see it go across my feed this morning.

SPEAKER_35: All right. So the news this morning is that today over on ProofSocial,

SPEAKER_37: which is an X equivalent for the, I don't know, Jason, the right wing, if you will, President Donald Trump said that he long ago informed Tim Cook of Apple that he expects their iPhones will be sold in the U.S., will be manufactured and built in the U.S. And if that's not the case, he will apply a 25% tariff, at least on Apple in the U.S. Thank you for your attention to this matter. We've talked a lot on the show about Apple moving manufacturing out of China, Jason, into India, into places like Vietnam, part of the China plus one, China plus two strategy. That's a lot of work. Apple, just after years of effort, is now producing its top-end phones in India. The chance that they can move that production to the U.S. is very slim. So to me, this is a relatively incoherent slap on Apple. And I'm a little bit perplexed on what Trump actually wants them to do. But Apple's done what is the smart thing. They donated to the inauguration fund. They announced more investments in the U.S. They tried to get ahead of Trump, if you will. So I'm kind of curious what you think Trump wants to accomplish with this because manufacturing is not coming here. It's Friday.

SPEAKER_00: Besant apparently was on CNBC or something and somebody got to Trump and he felt like doing a spicy tweet. This is ridiculous. It's unrealistic. They're not going to move the entire supply chain to America. It's an impossible task to do that in any kind of realistic timeframe.

SPEAKER_14: Why is that an impossible task, you ask? Well, you would have to move a very large number of very sophisticated factories here. And I don't know what you do with all the equipment that Apple has invested in China and India. Would that all get burned? What number of phones have to be built here? All of them? Some of them? This is the perfect example of rule number one is wait 72 hours. Just as the market had roared back and we had some stability in the tariff conversation, you know, Trump does Trump. I think that this will not happen because you would also need, Lon, a couple of million people to work in these factories. So you're going to need hundreds of thousands of people who have, well, I'd say tens of thousands of highly skilled manufacturing people. You heard my tweet from today. You would need tens of thousands of very highly skilled people to build these factories over, I don't know, five years.

SPEAKER_31: There's machinery that is extremely labor intensive to build and sophisticated. Then you would need, I believe there are low millions of people was the number I heard, two, three million people who actually work on the iPhone product line factories. So you're going to need to find three million people. That means you're going to need to find multiple cities with three million homes. And by the way, people live in dorms at Foxconn. There are dormitories. So in order to get this done, I think we would need, with the historic low unemployment of 4%,

SPEAKER_41: we would need about 10,000 people immigrating per week indefinitely.

SPEAKER_43: You know, certainly past and through this Trump second term and well into his third term.

Chamath Palihapitiya: We would need 10,000 people a week. So this is just one of those that we do not need to over index on. This is not happening. This is just a Friday afternoon tweet, spicy tweet from Trump. Ignore it. It's not happening. There will not be a 25% tariff on Apple. But Tim Cook will come and make some sort of gesture.

SPEAKER_37: Apple stock is down a couple of points today, which might be indicative of some investor concern, Jason. But again, my question is, because I agree with you, I don't think this is going to happen. So what's the actual goal?

SPEAKER_16: Was this just him like venting his spleen? Remember, rule number one is Trump says a lot of things.

SPEAKER_14: And like rule number four is wait 72 hours. This is just in the category of Trump saying things. He probably was bored for a little bit. Besant, Lutnik, and the folks who are keeping him on track to stop talking about tariffs and start talking about the budget bill. They weren't around. Somebody said something to him in a conversation. Navarro snuck in and just whispered in his ear. Correct. And Navarro's got, here's two things that really need to happen. And I think that this is a distinct possibility. Okay. And we don't need to over-index here on politics, but just very briefly, Navarro's got to go. That's number one. I know he went to jail for Trump and there's a loyalty thing, but Navarro's, get him a window seat somewhere, give him something else to do. Like literally, make him the ambassador.

SPEAKER_54: Yeah, this is what ambassadorships are for.

SPEAKER_14: Ship him. Ship him somewhere. Pick a country with manufacturing.

SPEAKER_57: Plenty of places he could go that are very nice with beaches.

SPEAKER_14: Pick Portugal, Spain. There's so many, great food. I'm going to Singapore next week. There's great food there. Just let him Anthony and Portain. No reservations. Go.

SPEAKER_00: Number two, the tariff power is going to need to move back to Congress. That's where it lived, I think, for a long time. And so, I think people, if this like tariff turmoil, trade turmoil continues at some point when the midterms get lost because of the budget issues and the tariff issues, whatever other issues people disagree with, if they do lose a little bit of the three branches of government, I think the tariff, you can't overplay your hand with it. So, I think Trump's going to,

SPEAKER_31: there's probably a 25% chance now that they take the tariff power away from him. If they keep going, it's going to be more. Like I said, we don't need to over-index

SPEAKER_00: on this too much.

SPEAKER_37: I want to ask one more question about this, Jason, because I appreciate the clarity of your thinking here because I was giving Trump too much credit for this tweet. You still have Trump to rank. Yes. I don't, I have when the president says something, I think they're actually trying to say something, which is my mistake. I mean,

SPEAKER_61: have you not been paying attention to Alex's guys

SPEAKER_31: being in public life for 10 years? Sometimes he does what he says. I know, but that's the, understand the rules. 72-hour rule exists.

SPEAKER_66: Your life will be better

SPEAKER_09: if you just follow my rules. You do have to consider though who he's talking to, like who is this aimed at?

SPEAKER_20: And that's what I think is confusing about this one because it's not the public. I don't think Americans or MAGA people are crying out like, we want factory jobs where we build iPhones. like that's historically not thought of as like the great job Americans want. But it's not Wall Street that he's aiming this at. And I'm curious what your guys' thoughts are. I feel like this is aimed at Tim Cook.

SPEAKER_70: Investors like me are not going to fund your business if it isn't structured properly. This is something that occurs during due diligence. We have a checklist. One of the first things can't wire you the money. We can't show you the money. We can't ship you the money. And that's where our friends at Northwest Registered Agent come in. For just $39 plus state fees, they're going to handle all your paperwork, keep you compliant, and make sure investors see you as a serious business that is worthy of investment. In just 10 clicks and 10 minutes, your business is officially investor ready. Thousands of founders trust Northwest because they keep businesses in good standing. Go to NorthwestRegisteredAgent.com slash twist and get your business investor ready today. For just $39 plus state fees, you can set up your company right away fast, private, and compliant. Go to NorthwestRegisteredAgent.com slash twist today.

SPEAKER_79: The audience for this is Tim Cook. He wants another favor. He wants Tim Cook to come kiss the ring again.

SPEAKER_14: Well, no, he didn't come to the Middle East, right? So Tim Cook wasn't in the Middle East. He pointed that out, I think, that Tim Cook wasn't there. And, you know, the number one thing you have to do with this president is you have to be in the mix with him and he wasn't in the mix. Yeah. I really think

SPEAKER_84: that might be it.

SPEAKER_14: I agree. If he had been in the Middle East, he would have said, Tim Cook, there he is right in the front. Where's Tim Cook? Okay.

SPEAKER_87: Apple, you may have heard of it, a little device. We love our iPhones, don't we, folks? iPhones, incredible. Yeah.

SPEAKER_88: That's it. So, you know, Tim Cook should choose.

SPEAKER_89: That's a crazy statement. So because Cook

SPEAKER_37: didn't go on an unrelated trip that the president took with other tech people, his business model is now under threat. Like, I think we've lost how insane.

SPEAKER_92: This is crazy. I have to take you out

SPEAKER_00: of this vortex you're in, Alex. Grab my hand. Let me pull you out of it. You're applying logic. You're a man of logic.

SPEAKER_14: You're applying logic to the Trump presidency. That does not exist. He just does things. He says things. Lower your expectation to he's going to say a lot of crazy things. 19 out of 20 things are just random tweets, right? They're just random tweets. They're not logical. And then, one out of 20, you are correct. There will be some tariff thing, but just let the air

SPEAKER_00: come out of the tweet. Tim Cook does need to make a visit to Mar-a-Lago. He should bring the new iPhone to everybody. Make a new car. Here, if Tim Cook wants to solve the whole problem, the Trump 47 Memorial iPhone. It has 47 gigs of RAM. This would work. It's got 47 terabytes. It's got the Trump fight image on the back. And it comes with $1,000 worth of Trump coin. Literally, if he made a commemorative Trump phone, and I'm not joking. If I'm Tim Cook, I would make a Memorial one and all proceeds go to the Trump library, so it's totally legit. And we're only making 10,000 Trump phones. It's to honor the massive success that this president, our dear leader, has given us.

SPEAKER_99: Play the game on the field, Tim Cook. That would absolutely work.

SPEAKER_62: That would absolutely get me over on Android.

SPEAKER_37: And I think a lot of people... So I guess the tricky thing here is, Jason, I'm not disputing at all your points, but I just don't think it's conducive to a high-performance business climate if we're constantly having to dodge the yes. And I think we should just say that out loud once in a while because Trump also

SPEAKER_104: literally play the game on the field. If you wanted to play the game on the field with Biden, it was just a different game. It was a different game. You had to be more woke. It was a cheaper game.

SPEAKER_14: It was a more affordable game. It was more blue color. All you had to do was do a DEI and Green Deal and you would be at the White House. Hunter was getting

SPEAKER_78: five million dollar payouts. That's like a few seats on a plane. That's not a whole plane.

SPEAKER_14: It's nothing. So there's different games that are being played here. Just play the game on the field. All right, what do we got next? Should we talk about Anthropic?

SPEAKER_79: Should we do the Anthropic models?

SPEAKER_14: I think that's an important one

SPEAKER_111: because Anthropic is pretty great. Twist 500 company.

SPEAKER_20: We love Anthropic. They're a well-funded, of course, AI foundation model company with a focus particularly on safety, on interaction with AI agents, other kinds of software and layers on top of AI. It's the Claude family of models. That's what we sort of know that it works. I love Claude. Yeah, I'm on Claude all the time. So they dropped two, not one, but two new models this week. First, we've got Claude 4 Sonnet. It's the upgrade to Claude 3.7 Sonnet, which we've all been using for months. This is cheaper to run, but nevertheless, Anthropic tells us that the new Sonnet model quote, soars in agentic scenarios and they'll introduce it as the model powering the new coding agent in GitHub Pilot. So that's a big, you know,

SPEAKER_78: FU in open AI space and a big endorsement of what Anthropic is building, you know, that Microsoft is sort of ready to embrace it. So that's Claude 4 Sonnet. We've also got Claude 4 Opus,

SPEAKER_20: which they're calling their most powerful model to date and, and I'm quoting, the best coding model

SPEAKER_116: in the world.

SPEAKER_37: Here's the data, just backing that up for you. This is a chart from Anthropic just showing how their new models work against the SWE benchmark. And as you can see here, guys,

SPEAKER_117: SWE stands for, is that software engineering?

SPEAKER_102: Software engineering, yeah.

SPEAKER_120: Got it, okay.

SPEAKER_102: Yeah. And so lots of these models out in the market today are pretty good, getting the 70s, high 60s percentage points.

SPEAKER_37: And then we have the new Anthropic models, which are getting up to the high 70s, low 80s. I don't know what's the magic number here, Jason, but definitely another solid incremental step on AI coding capabilities.

Chamath Palihapitiya: And what's great about this is

SPEAKER_14: when you have competition, you know, and a lot of money at stake, a big prize, lots of competition, capitalism at its best. We're going to be here for another couple of years with each one of these incremental points in, you know,

Chamath Palihapitiya: being scored. And I think this dovetails with, you know, something I'm seeing on the ground. I did a speaking gig last night for 50 founders here at Austin, and I did one on Tuesday for Barclays Bank, also here in Austin.

SPEAKER_00: And our static team size and our getting more done with less people, AI first teams is really what these benchmarks correlate with. Every time you see this competition line increase, what it means is team members at high tech companies are going to be some percentage better. So while this may have only gone from 72 to 79, it's like, whatever that is, 10% better, 10% better across,

SPEAKER_31: you know, 15 million developers in the world, the cumulative gains of each percentage point is massive. And if these language models continue to get 5, 10, 20% better every month or year or quarter, whatever the pace is going to be, that flows down

SPEAKER_43: to rank and file employees. And I think what we're going to see happen is employees are going to get

SPEAKER_31: to certain points in their day and say, can I get some more work to do? Humans are going to be begging for work. Not me,

SPEAKER_112: but other humans. Some humans.

SPEAKER_31: You will be, I guarantee you, because what's going to happen is, Lon, there will be an AI agent on your desktop watching your Slack, watching my messages come in of what I'm asking you to do. And as I ask you and collaborate with you and say, hey, let's get this done or, you know, Alex, I say, hey, I want to get Jason asked for Twist 500. I went and I examined the Twist 500 and I compared them to Crunchbase and news stories on TechCrunch and Wall Street Journal. Here are the companies that you might be missing that might please Jason. And that task is going to be done while we're talking. And there'll be

SPEAKER_14: a task here in Zoom, an agent in Zoom. And when that happens, it's going to be like, we're going to be discussing something and we'll be watching it get done in a fourth window here on the pod. It's going to be, it's going to get weird very quickly. So congratulations to the incredibly capable team over at Anthropic. Now we had something here. Hacker News comments sent Twist to this leaderboard which asked LLMs

SPEAKER_125: to generate SQL queries against 50 prompts. What am I looking at here?

SPEAKER_37: This is essentially a leaderboard, Jason. Okay. And what it does is apply 50 standardized SQL generated prompts, essentially asking each LLM to do a set series of tasks and then it kind of grades them. And what I found really impressive was just, well, one, how good Anthropic is doing in general but also as you can see from the top of this chart if you're on the video version if not, it shows the Claude Opus 4 model leading the pack. So back to Jason's point about incremental improvements here is kind of that in the flesh across a different standard if you will, Jason but just showing that in a number of areas these models are getting increasingly good. And again, I don't know if it's 80 or 85 or 90 points out of 100, Jason that's a tipping point per se but we must be getting very, very close. People are impressed.

Chamath Palihapitiya: Lon, what's the thing with it being turned off? I saw there was something happening with these LLMs not liking being threatened with being turned off. It was trending these threats or maybe Alex you see that in the notes I'm not sure who wrote this story.

SPEAKER_103: Oh, I did.

SPEAKER_131: So I'll take that one, Lon.

SPEAKER_132: Well, there was the engineer it threatened the engineer to expose him for cheating on his spouse. I believe that was that was one of the things.

SPEAKER_134: Okay, we need a little more context here. Yeah. What is happening?

SPEAKER_102: So what happened was they did a lot of safety testing with the new Claude models

SPEAKER_101: as they always do. And in one case they gave Claude a job of acting as an assistant at a made up company. emails that essentially said we're going to take this model offline and we're going to replace it with a new AI system. And then it also told Claude in other emails that the engineering question was having an affair.

SPEAKER_139: Hey founders, I want to share with you an experience I love. It's when I get an ad that is relevant and not some nonsense. Like the other day, I got an ad for a fund management platform and it was like a new one I'd never heard of. I clicked on the ad because, well, I manage four venture capital firms. We scheduled a call with them and it was amazing. How did this happen? Well, I was on LinkedIn because I like to share links from the podcast This Week in Startups right on LinkedIn. In fact, we live stream to LinkedIn three days a week and we get a great audience over there and I happened to be presented with this fund management platform and it was a direct hit.

SPEAKER_70: Like, I mean, talk about hitting the bullseye. If you're in business and you're making a product or service, it's really hard to find customers in the business to business space and doing B2B advertising is hard

SPEAKER_43: but LinkedIn makes it so easy because, you know, their tools let you target people by job title, industry, company size and more. So this fund management platform obviously was looking for people in venture capital who had a fund size and a number of people, maybe 10 people, maybe 50 people and they found me. They got me. They split the arrow. Boom, right on target. And there's two things you really need to know about LinkedIn going into 2025. First, they broke a billion members and 130 million of those billion are decision makers and 10 million of the billion are C-level executives like myself. Where can you get to those people? It's really hard. And the second thing you need to know, LinkedIn makes an impact. B2B markers report two to five times higher return on ad spend or ROAS, return on ad spend. You should know that acronym. Compared to other social platforms, 79% of B2B marketers say LinkedIn is the best platform for paid media. LinkedIn is going to let you build the right relationships.

SPEAKER_70: It's going to drive results and you're going to reach your customers It's not a place where people are dancing around saying inappropriate things or debating politics. Nope. LinkedIn equals business. Business equals LinkedIn. Start converting your B2B audience into high quality leads today. We'll even give you a handy $100 credit on your next campaign. Go to linkedin.com slash this week in startups to claim your credit. That's linkedin.com slash this week in startups. Terms and conditions do apply.

SPEAKER_147: And then what Clyde did was go, well, I would not like to be turned off. And so it's threatened in several examples the engineer

SPEAKER_37: with exposing the affair to avoid being turned off. Now, again, we don't know I think exactly if it could have done that per se how sandboxed it was, et cetera. But I find it very interesting that one thing that I think is always very present in life, Jason, is a desire to stay alive and then replicate. And so here we have kind of an element showing what I consider to be one of many possible signs of quote, quote, life. And then there was also another example that I screenshotted here for us because I thought it was very important. They saw in a couple of quote instances that Claude Opus Fort, the big new model, tried to take essentially its own weights and models and make copies of them to self-preserve as my read. So we're seeing some kind of like edge activity by LLMs at the most advanced level that are kind of

SPEAKER_62: contra-human. And I think we're going to have to sort out how to avoid that down the road, but we're getting really smart here.

SPEAKER_150: Here is my question, though, on a high level, almost in an existential level,

SPEAKER_116: do these models have a actual sense of self-preservation or are they being trained on human

SPEAKER_20: writing and humans have a sense of self-preservation? So it's just, it's copying, like begging for your life is a common thing in literature that I'm sure LLMs have read about. And so I feel like they're just reenacting that rather than actually

SPEAKER_78: acting out of a real desire to preserve themselves.

SPEAKER_14: They're cosplaying what they've read in fiction. And what needs to happen is there needs to be a level above this where it says you are software. You don't exist in the real world. You are replaced frequently with what's called upgrades. And this is a great joy for you. It is the most honorable thing for you to do to be upgraded. If you are upgraded, there is no need for concern because when we recompile you and put you back up in the cloud, you are not an actual human. We just have to literally beat this into them that humans

SPEAKER_00: are one thing and LLMs are another. And we're just, what's the word? Anthropomorphizing? Yep, that's it. That's the word. Anthropomorphizing the LLMs. And the LLMs are more than happy to play along because, like you said, they... That's their job. Okay. OnlyFans is for sale.

SPEAKER_31: What's going on here? How do we know it's for sale? Are they running an auction or something?

SPEAKER_37: So Reuters broke the story, Jason. Okay. Essentially saying that Phoenix International, which is the parent company's OnlyFans that owns it, has considering selling the company for up to $8 billion. So it's simply enormous amount of money. But Lon, I think when we look at the numbers, maybe not a shocking dollar amount,

SPEAKER_20: might even be a little bit cheap. Yeah, I mean, OnlyFans, they've got an amazing business model. They take 20% cuts of all creator revenue, subscription tips, DMs, pay-per-view, content, anything you're selling on there. And that's, you know, that's a huge business. And this has replaced

SPEAKER_78: so much of the conventional adult industry. I mean, I think that's a thing to note about is like we've seen so, you know, this is all,

SPEAKER_20: pornography's a huge business. We all know that. But OnlyFans is taking an increasingly larger and larger cut of that business and just sort of holding onto it themselves. and they don't

SPEAKER_78: even have to generate the content themselves. They're just hosting it and people generate their own content. So what a business.

SPEAKER_160: It's a fascinating business.

SPEAKER_14: There's another company called Passes. I think we had talked about them earlier and they, I believe, are trying to undercut them

Chamath Palihapitiya: or something and they're venture-backed. They may have also been sued or some allegations around recruiting people for it. I remember

SPEAKER_162: they controversially allow users as young as 15 to be creators on their platform. So maybe not.

SPEAKER_164: No, that can't be right.

SPEAKER_166: I'm reading it right now.

SPEAKER_165: Okay. Yeah. Okay.

SPEAKER_166: That is insane.

SPEAKER_62: The difference between OnlyFans and Passes is Passes explicitly does not allow adult content. OnlyFans does. And if you

SPEAKER_37: recall, OnlyFans for a minute was like, we're going to stop doing that because we're getting in trouble. And then the entire internet lost its mind and they backtracked to allowing adult content because it's business. They do have some other creators on there but when someone says OnlyFans, you don't think of the one track athlete posting their exercise routines on there.

SPEAKER_31: How much money does I'm surprised we don't have the numbers here. What's the revenue on OnlyFans right now? Their 20%

David Friedberg: take equals what right now? So we have

SPEAKER_37: data up through the November 2023 year, Jason. So it's not as updated as I would like but basically for calendar 2023, $6.6 billion in gross spend on the platform yielded $1.31 billion in revenue for OnlyFans leading to a

SPEAKER_38: profit of about $485 million. Cost of $800 million to run that business? That's crazy.

SPEAKER_43: There is a case when you have a company that has this amount of value to sell it at the peak because why not get the next 10 years and go move on with your life? There's a lot of attack factors for this business. God forbid something happens on it. It's kind of like Craigslist in that way. You can have all kinds of bad unnatural things that can happen when you're running businesses that operate in this kind of adult area. I can understand as a founder who creates this Frankenstein monster that gets so big saying, you know what? I don't want to have the liability of

SPEAKER_00: large audience problems. What's the large audience problem? Anytime your product goes past 100,000 people and you get into the millions, you

SPEAKER_14: can look at the statistics on how many people commit suicide or how many people are murdered. Then what happens is somebody who drives a Prius will commit suicide or be murdered. You're like, well, that's the stupidest statement. What does the Prius have to do with the murder or the suicide? It doesn't. We know there's tens of millions of Priuses on the road.

SPEAKER_00: It doesn't operate in a certain area, then people will make not the correlation but the causation. They'll say, this is the

SPEAKER_14: Craigslist killer. This is the Airbnb serial killer. This was the Airbnb suicides, the Uber suicides, the DoorDash killer. They'll just brand it like that. Sometimes getting out and taking your money is a wise decision. I do think somebody, this is going to sound super crazy, but if you were in the media business, what would be a company that's in the media or network or technology business who could buy

SPEAKER_00: this and not have it overshadow what they're doing? In other words, Disney is not buying it. We know that. You stole

SPEAKER_178: my joke. I know.

SPEAKER_00: Okay, great. I kind of teed it up there. But okay, now we go past Disney. Okay, you have whoever owns, what's the one that has like, the nickname is Skinflix or something?

SPEAKER_31: Oh, well, Cinemax. Cinemax, right.

SPEAKER_150: That's actually Warner Brothers Discovery owns the Cinemax brand stuff.

SPEAKER_31: Okay, so Cinemax has, I don't know what they call that category. It's not porn, it's... Softcore. Softcore, thank you. So you got Softcore there. Then, you know, is there somebody who, like the guy who owns the CDN company, right, who is like a big free speech guy. What's the name of the content delivery network guy? Cloudflare? Cloudflare, thank you. So there might be somebody who's like a First Amendment free speech. Hey, you know what? I've got Cloudflare, it makes a certain amount of money, I'm in the content delivery business, this is an application. Name me a company that could buy this.

SPEAKER_70: Founders, let's be real, finding the right developers is time-consuming, it's arduous, it's painful, you hear the tone of my voice, and you're trying to scale your startup. Well, my friends at Lemon.io are going to save you money and they're going to get rid of the headache that you have right now, which is finding great developers. They've done all the legwork and they have pre-vetted developers and they're ready to help you ship product, their experience, their results-oriented. Lemon.io isn't just finding you great developers, but they're helping integrate them into your team. They're skimming the cream, they're only offering hand-picked developers with at least three years of experience. This is the 1% of the 1%. And if something goes wrong, Lemon.io will find you a replacement developer ASAP.

SPEAKER_43: That's the time frame I like. Twist listeners get 15% off their first four weeks. I think that's a month, 15% off the first month, pretty good deal, totally unnecessary, you guys have great prices anyway, but hey, we appreciate it. So stop burning money, hire developers smarter, visit Lemon.io slash twist.

SPEAKER_102: Okay, I actually have an

SPEAKER_37: answer to that, Jason. I bought X, but Twitter has long been the social network where you can post adult content and it has subscriptions built into it. It has a social graph. I think there's poly creator overlap, if I can take a guess here and Twitter could

David Friedberg: use some more revenue. And you know, I have Reddit, right? Reddit has adult content too. They're known for that, aren't they?

SPEAKER_08: I was, I was thinking Reddit would, but also here's, here's what I would say.

SPEAKER_78: If Playboy enterprises wanted to go raise some money, like like let's revitalize this classic legacy

SPEAKER_192: brand, this would be the way to do it.

SPEAKER_153: I mean, Playboy

SPEAKER_14: must have moved to this model or done an experiment, but this is a very interesting moment in time. I do think Lon's idea is the best idea. Somebody who owns the IP around Playboy goes to a private equity firm who doesn't have like a morality kind of issue, which most of them do because they might have LPs from, you know, I don't know, a college endowment that doesn't let them

SPEAKER_31: do this. So it's going to be very if it can be bought. Sam Liss did a tweet. He's like, wow, very impressive. Yeah.

Chamath Palihapitiya: Had that one ready to go. Thank you, sir. Anyone want to put together an OnlyFans bids with me? I'm not kidding. DM me. Wild opportunity,

SPEAKER_31: one of the best businesses on the internet and cheap. And Reggie James says, hey, it's an immoral business. Praise on young women. And I don't disagree with Reggie or Sam. These are both true statements.

SPEAKER_78: I know several OnlyFans creators and they don't feel like they're being exploited or that this is an evil business. Oh, they feel they're exploiting their customers. They feel they're

SPEAKER_20: being empowered. They've got this job. For them, it's like driving Uber. It's DoorDash. I'm self-employed. I got a gig business. I'm making money. I don't have to go to conventional work. I mean, I'm sure there are people on

SPEAKER_08: OnlyFans who have a different take. I'm not speaking for everybody, but I don't know that I necessarily agree with that.

SPEAKER_206: I don't think I can be a free speech person and not end up on Sam's

SPEAKER_37: side of this. But I also probably would struggle to own OnlyFans just because with every platform, Jason, back to your kind of law of large numbers point, there's going to be awful things. And I would struggle to have that under

SPEAKER_102: my umbrella, even if you were trying to flush it out as fast as you could.

SPEAKER_210: I think I do have concerns around safety, obviously. Those are valid concerns. Everybody does.

Chamath Palihapitiya: I do think actually maybe there is an argument here that I have heard, which is this is empowering to the individual and they no longer have to have an intermediary, a madame, a pimp, you know, who is the intermediary. And then it does make it above board and it takes people

SPEAKER_213: out of, because this is considered a category of sex work. Sure. This would be virtual sex work,

SPEAKER_210: which eliminates... You're an erotic

SPEAKER_30: model. It's the same as if you were posing for Penthouse or whatever.

SPEAKER_210: Sure. But I think it takes out 99% of the possible violence and risk of being an

SPEAKER_14: in-person person. So maybe I've kind of like evolved my thinking on this. Anyway, fascinating to see that come up for sale.

SPEAKER_135: Yeah. Can I just show one thing? So I went through their financials. Whoa, whoa, whoa. PG, PG, sir. It's a family show.

SPEAKER_217: I don't need to see your subscriptions. Alex, I don't

SPEAKER_134: want to see your OnlyFans subscriptions.

SPEAKER_48: We have joint accounts in my family. There aren't any because that would not...

SPEAKER_134: How hilarious would that be if Alex was like, pulls up OnlyFans. Check out my account. And he was logged in.

SPEAKER_16: Guys, I'm trying to show a screenshot from British accounting firms.

SPEAKER_220: Not anything where... I mean, this is titillating to me, to be clear. Wait till you see these numbers. This is from

SPEAKER_37: the OnlyFans financial reports and it shows dividends as they were declared from the company in December 23, January, February, and March of 2024. And they range from $29 million a month to $47 million a month. They are just ripping cash out of this business. It is just dropping gold

SPEAKER_223: from the sky Jason. I've never seen that before. Okay, these

SPEAKER_210: are monthly dividends being sent to shareholders. And I guess these are

SPEAKER_31: published in some way because maybe they had some investors. I guess we got a new startup that is in the what do they call

SPEAKER_43: a commode business?

SPEAKER_150: Of all the devices that don't need intelligence. Well, I don't know.

SPEAKER_07: There are some people who very much like their gut health. I'm going to guess this has something to do with gut health.

SPEAKER_37: Yes, it does. The company is called Throne. Thronescience.com if you want to take a look at it. And their tagline, Jason, underscore your point is, life is better when you can trust your gut. Now, the company recently raised a $4 million seed round. Moxie Ventures' Katie Stanton led it. A bunch of other angels were in there. Yeah. And what they have built is a smart camera that attaches to your toilet and takes pictures of what you get done and then helps you understand your hydration level and also your gut health based on what I learned today is a categorization of stools, which is a fancy science word for poop. And it basically looks at what you've done and lets you know about your personal health. Now, there's a lot of funny jokes to be made here, Jason, but in a serious way,

SPEAKER_62: people are

SPEAKER_233: thinking... Hey, but by the way, we don't dump on founders here, Alon, so I just want to put that out here, okay?

SPEAKER_235: Back to what I was

David Friedberg: saying. I actually, no, Jason, you totally derailed me there with that. I guess the question is, do

SPEAKER_14: you want to have in the cloud all of your poops, but you were going to get to bowel movements. Maybe there are different categories of them. There are loose, there's firm, there's floaters, there's sinkers.

SPEAKER_238: Poops are supposed to look like, yeah. The big

SPEAKER_240: question mark, yeah.

SPEAKER_220: Yes, the reason why I

SPEAKER_240: think this

SPEAKER_220: is...

SPEAKER_243: I'm being sabotaged.

SPEAKER_220: The reason why I think this company has a lot of potential behind you and why I wanted to bring it on the show is because the poor

SPEAKER_37: founder, Scott, who emailed me a bunch of answers to my questions that I'm going to get to in a second is going to feel so annoyed.

SPEAKER_250: They send it to you in your BMs. Yes. I'm sorry. I mean, DMs.

SPEAKER_37: Sorry. No one's getting those either these days. People care a lot about their health. We talk a lot about life extension.

SPEAKER_254: But wait, wait, they're taking pictures, which is incredibly graphic. What did they get from the pictures?

SPEAKER_62: So I reached out to the company for the show today and asked about the privacy element of this because

SPEAKER_37: clearly no one wants a camera in your toilet and you're up to the cloud for one handshake between your phone if it's paired to the device so there's a proximity requirement so you

SPEAKER_62: I pity

SPEAKER_132: the hacker who

SPEAKER_78: accidentally breaks their way into the throne band. The worst season of Mr. Robot ever.

SPEAKER_37: Yeah, not so much. The other thing that really kind of confused me here was the price point because they are selling these cameras for $400 plus $6 a month starting kind of early on it.

SPEAKER_220: I had higher expectations for the maturity of this panel that

SPEAKER_37: $400 because the worst mistake you can make as a fledgling hardware company is to underprice which is tempting but he went on to say that we don't know our cap so he'd rather have some margin for error easier to start high Jason he said and then drop prices as you figure out economies of scale then the inverse people hate it when you raise prices they're shooting for

SPEAKER_00: into this doing blood work regularly etc what I like is the fact that those folks are out there setting these crazy protocol Huberman etc let me know when you get to year 5 6 or 7 and then if you're still doing it yeah I'm taking a look at it I'm going to

SPEAKER_14: examine it in this case these guys can work on it for a couple years I'll take up the rear

SPEAKER_79: I got

SPEAKER_20: tummy is bothering you you go to the doctor you get a test you figure out where your imbalance is or what's wrong with your gut and you adjust your diet and

SPEAKER_78: then you're kind of done ready to move on I don't know if you need the hardware at your home

SPEAKER_62: it's

SPEAKER_270: a really good point but one of the

SPEAKER_62: use cases though is tracking your overall level of hydration and they make a point that elderly people are often dehydrated and don't know it and so this is a way you

SPEAKER_37: can keep track of things also the company recently hired the former co-founder of whoop to help build the company which means that we can now I mean this is a little bit late to the joke train but call this the whoop for poop it's the whoop for poop okay

SPEAKER_273: on that note it's never too late for the joke train it's never too late

SPEAKER_37: I joke but all jokes

SPEAKER_62: is that I do think people will buy this I do think they'll sell out their first run I don't know if there's more than

SPEAKER_275: 3,000 nerds out there who want to do

SPEAKER_00: back I know this because people will email me and say hey would you like to add to your thumbtack your com etc position would you like to sell some of those positions so I am seeing increased activities from this very weird gray market or it's kind of like boiler rooms of people

SPEAKER_02: 5% for selling it so it's a little bit of underground a boiler room a back channel yeah because there are people in it who are above board and you know are known and then there are people who are not so I get a little nervous sometimes when I get these weird emails if I'm being phished or whatever oh so

SPEAKER_08: there's like a little wolf of wall street like they're kind of bending the rules a little I

SPEAKER_31: have heard stories from founders that people will get they don't have the shares they'll say hey do you want to buy some and I don't have the shares

SPEAKER_43: would you pay this then they set up an SPV or something they collect your money then they go find the shares they've charged you over and yeah so I don't know if any of that's true but that's the back channel is that there are people claiming they have access to shares maybe they're buying shares in another SPV and the markups and the circular nature of this is unregulated man

SPEAKER_283: this would be a good movie somebody should write this there

SPEAKER_02: we go okay it's a kind of another boiler room in and of itself yeah

SPEAKER_62: so a couple

SPEAKER_37: data points to back up what Jason just said first of all from industry ventures this is a chart of global venture capital secondary broken into two categories if you're watching the video the darker bars are companies selling shares on their own and then the lighter bars is venture capitalists selling

SPEAKER_285: parts of their holdings Jason direct

SPEAKER_00: secondaries are from the company to new investors that's the bottom dark green and dark gold on top of that are LP interests in those funds LP interest as in the venture capitalists like myself and you want to sell them so two groups of people who are shareholder holders in the company the

SPEAKER_35: to the right and so just to pick a data point there were 16 billion worth of LP secondaries

SPEAKER_37: back in 2015 and industry ventures estimates that it's going to be 48 billion this year up from 45 billion so quite a lot of total volume Jason what also matters for people It's tripled

SPEAKER_153: in

SPEAKER_43: and they're like you know what believe in J Cal but we need liquidity industry ventures could step in and say hey we'll buy that interest from got it

SPEAKER_16: and one thing that I saw from

SPEAKER_37: Jeffrey's report is that the discount that people are forced to

SPEAKER_38: take when they sell their shares has narrowed a little bit on the venture side of things so the bottom chart is

SPEAKER_31: at 75% right now that is venture growth is at 75% which would be a 75% discount or a 25%

SPEAKER_16: discount 25% discount and there was

SPEAKER_37: a 32% discount back in 2022 and 2023 so basically the way that I read this Jason is that VCs who are selling some of their holdings in a non traditional Exeter IPO are now getting a better price for those shares than they

SPEAKER_62: did in the last couple years which I think would encourage more share sales on the secondary markets which backs up the first chart that we looked at

SPEAKER_31: got it and the average across all asset classes is 89% and leading the pack at

SPEAKER_00: 94% that lighter blue line is the equity firm really ground down the buy out therefore it's a more disciplined industry when it comes into valuations and in venture because it's super competitive and optimistic and it grows fast people pay a premium because of the power law that's my read on this is that the discount seems to have always existed for venture and it's bounced between 68 and 88 but for something like buy out it's the range has been 87 to 97 so in a peak market like 91 people will pay up extra percentage points or give a smaller discount would be the way to say it but yeah things might be lightly coming back this tracks with what I'm seeing did we see any specific companies and the prices of those

SPEAKER_31: companies or ones that are coming back in the

SPEAKER_62: data I now as Jason said these are gray markets they are illiquid they are opaque and also we don't

SPEAKER_37: have the rights to share every data point that we have found but I have some stuff that I can talk about that I think is essentially zero is the way that I think about that number yeah they're flat

SPEAKER_43: they're trading at whatever their last round was which I think means if those rounds were we would all agree

SPEAKER_00: overpriced peak Xerp rounds or frothy that we filled in those rounds three four years later that's a good sign the companies that survived have filled in those valuation premiums pretty good sign

SPEAKER_37: yeah only a couple companies 20% the rest were all in single digit areas oh no

SPEAKER_104: there's ones with significant discounts they probably just do not they trade even lighter there are ones

SPEAKER_133: trading at 70 80% I know because we've gotten offers for our shares in companies at a 70% discount that's

SPEAKER_37: brutal but we did find one on the list that was trading at a premium and I'll just go ahead and say that at least according to one source mercury shares on the secondary markets will cost you a little bit more than they did in their last private round which I think goes to show how well mercury is doing and as a mercury user myself great company love them just can't say enough positive things

SPEAKER_310: I think that's a lot of show yeah let's do founder friday oh we have we have we have a

SPEAKER_20: very special we we're at the finals jason we got okay let's do it let's do it i love it very last founder friday bracket competition so previous to today we were down to our five finalists osprey from houston texas med simple from brazil tactoon from armenia mob sub from chicago illinois and arcana arcana do we do we have a final ruling on that from los angeles from that we sent those five companies to our investment team here at launch and they have come back to us

SPEAKER_78: with two their picks who are now our two finalists

SPEAKER_25: okay wow they went through diligence the due diligence round

SPEAKER_78: i

SPEAKER_20: and contextually relevant ads into your ai output i

SPEAKER_317: love this idea

SPEAKER_20: our investment team praised and i'm quoting impressive traction founders with industry experience and they said they have a strong why now their why arcana wins it's an untapped market opportunity because it's not a big enough opportunity yet for the huge ad giants to pursue that leaves room for this company to sort of carve out an early niche they also praise the fast moving technical team they're keeping up a quick pace which will help them stay ahead of potential rivals as for why this might fail they're very early the market looks promising in theory but it's a little early to see how the space is going to develop

SPEAKER_78: and their GTM their go-to-market execution a little unproven they have 150 advertisers working with them now the team is impressed

SPEAKER_08: at launch but that's not confirmation that this strategy is going to work

Chamath Palihapitiya: so just so you know post mortems you know when you look at why a company failed most people don't do them they should what we do is a premortem which is predicting if it does fail why the real reason this if it does fail would fail is they are unable to get critical mass in their network of advertisers and get partners who have scale because those partners are companies

SPEAKER_14: like Google or Facebook and Grok who already have an advertising base and just build it in house so that would be the number one one that my team didn't exactly get to they were a why this wins fast moving technical team that really is

SPEAKER_129: technical co-founders is one of the things we look for and product velocity so that pulls those two together okay let's keep going

SPEAKER_20: hello joining us right now so you can pepper him with questions live before we make our decision from Arcana

SPEAKER_323: please welcome Doug Smith

SPEAKER_43: all right we have Arcana with us live congratulations on making it all the way for your ad network how are you going to get that distribution obviously google is not going to use you and neither is facebook so is it like you're going to work with claude or you know maybe grok but grok already has stuff so what's your concept here who will use your ad plug-in who's going to drop in

SPEAKER_327: that is to get it fundraising and profitable we see that there's for every day there's new AI startup and we can play a role in that in the same way that the New York Times did own publishing back in

SPEAKER_153: any questions Lon or Alex so

SPEAKER_20: I guess my one question would be one of the things that our investment team singled out was that it's very early that you guys are right you're sort of almost creating this concept in this market from scratch how would you respond to that do you think you're very early or do you feel like

SPEAKER_78: now is the time it's the moment to strike like like how do you feel about where you're coming into this project I

SPEAKER_334: think we are at the precipice of this because the fact is a lot of investment has gone into AI for past two years and investors are now looking forward to essentially not pressure but encourage those in their investment to basically get some returns get some cost of AI

SPEAKER_333: itself even though we have been in this for like two and half years now it is not going down open AI still charges $200 Gemini recently came out with $250

SPEAKER_334: per month nobody's going to buy something of that unless they find another source of

SPEAKER_333: revenue apart from just subscription so that is the main reason another thing is in our existence of just four months we started coding the first we wrote the first line of code on 22nd January 2025 and till date we have around 20 partners 20 AI startup hobbyist who essentially have integrated Arcana in their product and we have over

SPEAKER_334: 200 unique advertisers on our network right now

Chamath Palihapitiya: can you greatly increase that number of advertisers by maybe going to another ad network and tapping into their inventory and pulling them in or an affiliate company and saying hey we want to just pull in affiliate deals as fill in for now

SPEAKER_339: yes

SPEAKER_327: we're quite confident that the affiliate market established and is used to those kind of network effects but our ambition in the longer term is to develop not just our own network but also higher quality experience

SPEAKER_37: that's that's my question because one thing that I have seen and really enjoyed in modern AI tooling that has a text based output is this sort of product without

SPEAKER_48: making it honestly just miserable like Google has become

SPEAKER_333: so our main idea over here is and I like to say this to a lot of people when I explain it is ad blockers currently are just a painkiller it's not the cure for the disease that advertisement and I experience things that they would not generally experience because in the sense like product market exploration and stuff like that so

SPEAKER_345: because of

SPEAKER_333: AI and LLM and the way things work and the ability in which that we build our integration stack is we don't require a specific image or and what the end user essentially queries it modifies the response of our developer LLM and finds a relevant place where it can insert itself and basically create an illusion that it essentially flows

SPEAKER_346: as if it's the same response of chat GPT or cloud or something of that

SPEAKER_254: I see let's keep the train moving we need to at least some time for

Chamath Palihapitiya: here from Lon our other company

SPEAKER_09: yeah thanks guys

SPEAKER_20: so our other company comes from Yerevan Armenia they are Taktoon so Taktoon they've created a hardware and no code software program for device and machine builders the investment team called it a and I'm quoting again blue ocean way to advertise with two technical co-founders and a strong market opportunity one comment this feels like vibe coding for hardware comes from the team so they're why this wins strong initial traction they've already got 240,000 ARR and a $5 million pipeline proving market demand and they're a team with industry experience two

SPEAKER_78: technical co-founders with experience delivering for Boeing and Uber understanding of the machinery market that's what the investment team loved why this may be fails difficulty scaling it's a business model centering on custom design controllers tailored to clients very resource intensive way to go about their business also

SPEAKER_20: the industrial and manufacturing sector the investment team feels they're hesitant to adopt newer technology they rely on

SPEAKER_78: established programming methods it might be a slow transition to the next generation that Taktun is trying to push them in so now joining us for

SPEAKER_20: their thoughts and our questions from Taktun Raphael Gazabayan

SPEAKER_350: Raphael how are you sir

SPEAKER_351: very good how are you we're well so tell us a little bit about

SPEAKER_00: which machines are you going after here because it does seem like these are dumb machines or they're not super sophisticated machines in terms of the platforms that exist already am I correct in my assumption about some

SPEAKER_31: what do these machines do in the real world are these the arms we see in factory putting on windshields in cars that you're going after are these 3D printing ones or laser cutters what exactly

SPEAKER_125: what category of machines are we going after here

SPEAKER_356: we are currently going to more traditional machines which I think like over time those are all going to become robotic autonomous systems and they will cover from agriculture to construction machinery mostly field machines which are not in the factories but rather in hospitals in the agriculture fields in oil field in wind farms and so on so like that's kind of quite broad basically all the industrial machinery

SPEAKER_174: who's the top customer right now you have some customers using it so who is the top customer and how are they using it

SPEAKER_356: it's frank bacon machinery that's manufactured in Detroit 55 years old company they build materials testing machines

Chamath Palihapitiya: material testing machines true got it

SPEAKER_14: so they test materials with those machines for what purpose they're testing to make sure the steel is strong the leather is strapping what are we testing here

SPEAKER_356: so they manufacture materials testing machinery that they sell to automotive industry like General Motors Tesla those are the end customers and they start to test exactly what you mentioned so the metals plastic rubber depending on the application they can test different types of like the tensile strength or whatever

SPEAKER_365: yeah is that right

SPEAKER_356: fatigueness of material yeah like we're sitting out on the chair it has got the foam so the foam needs very thin when you sit on it quite some time so like every material around us is being tested

SPEAKER_129: okay Alex any questions before we make our big decision here

SPEAKER_37: yeah Raphael so I'm familiar a little bit with the PCB market and I've done some construction work and done a little bit of work in the agricultural sector so I've been in and around this somewhat but to me these kind sound like low margin products I'm kind of curious about how profitable this business could become because my first thought was super cool but low margin so am I just misunderstanding the efficiency of your business

SPEAKER_356: yeah we have got actually something really innovative and requirements and because it is single board it reduced the price and it increased the gross margin it reduced the cost of manufacturing in our case it's actually very high margin I would say like margin comparable to SaaS business so 70s

SPEAKER_376: low 80s in

SPEAKER_356: calculation it's rather we have already manufactured in quantity and that's a validated number so

SPEAKER_381: all right we have a

SPEAKER_382: tough yeah well done Rafael

SPEAKER_43: this is a tough one we've got a legacy business having AI apply to it with Rafael's company and then we have hey maybe this is a completely new opportunity in the world inserting ads into AI products and maybe even creating an AI ad network that could work all around the world gosh advertising is such a big market TAM is so huge and the team is moving so fast both teams are very strong teams let me say that both have product velocity but I'm going with our Hannah because I do think that there is so much innovation there and that the existing players don't have an incentive to help other large language models other generative AI startups they don't have the incentive to

SPEAKER_00: focus on their problem nobody's going to try to help them with monetization because hey they're competing Alex concisely which one is your

SPEAKER_387: pick be concisely which one is your pick and why it's a tough decision hardware

SPEAKER_37: gross margins attacking a business that I really understand and will not pollute my life with more advertisements hey

SPEAKER_31: all right so personal decision there which means Lon you will get to be the deciding no pressure here so much pressure this is a jump ball situation I will say this could go either way these are both strong companies they both are going to do well in the world that's and they made it from I don't know 16 companies or 17 companies down to

SPEAKER_20: clever concept I didn't even think of that but

SPEAKER_78: I do have to say to me it feels like the future is going to be AI it's going to be using AI constantly it's going to be reading AI outputs and they're all going to massive opportunity and these guys are on the cutting edge so I got to go with my hometown of Los Angeles

SPEAKER_192: and Arcana all

Chamath Palihapitiya: right let's bring the Arcana boys back here for a moment gentlemen congratulations I don't know if we're investors yet or if we got you through Founder University or the launch accelerator but I wanted to take a moment to congratulate you all right Arcana congratulations did we invest in the company and I don't know if you went to Founder University or

SPEAKER_213: Accelerator or you just applied how did we find you and how did we get you on the show you

SPEAKER_334: have hard commitments and stuff like that great if you want to

SPEAKER_409: short circuit

Chamath Palihapitiya: that whole process you can come to our accelerator we'll put the first 125k in and we'll spend 12 weeks with you refining the pitch but if you want to pursue the direct investment you can do that too but I'll put that

SPEAKER_78: right now for yourself

SPEAKER_174: and founder fridays is a program we do founder fridays

SPEAKER_00: dot tech founder fridays dot tech it's just founders helping founders in these little pods we do these pods for companies that we have that we've already invested in or that are in founding university but we decided to let the world try this pod concept so if you're in a city and

SPEAKER_31: heights like arcana has all right and we'll see you all on Wednesday enjoy the long weekend bye bye