LIVE FROM NEW YORK: Ramp's Valuation Reaches $22.5B, Figma Going Public, The Personal Superintelligence Debate | Eric Glyman, Avi Schiffmann

You're watching TVPN.

Today is Wednesday, July 30, 2025. We are live from Manhattan, New York.

That's right.

The fortress of finance, the capital of capital. The temple of technology. That's right, the temple of technology with us to Manhattan. We're here live from New York City for a few reasons.

What are we wearing?

We're wearing yellow suits to celebrate.

That's right.

Fundraise ramp raise their series E2 today.

Congratulations.

As one does, we will have Eric Lyman joining the show in just a little bit. In about 50 minutes, 52 minutes, he will be joining.

Can't wait.

First we got to talk about Mark Zuckerberg. That's right. Been completely natified. Nat Friedman.

They natified my boy.

Already. Already having an impact over at Meta. Will depue has the post. He says, lmao. Zuck has been natified. Plain Text Times New Roman now going to hit mainstream corporate media like a bomb. And this is from Mark Zuckerberg's announcement of his. A little more into his strategy and product strategy with super intelligence for everyone. For everyone.

Potentially going to be personal.

It's going to be personal thing. So he says personal superintelligence in an H1 clearly HTML tag in. This is all in plain Times New Roman text. You can see post. He says, over the last few months we have begun to see glimpses of our AI systems improving themselves. The improvement is slow for now, but undeniable. Developing super intelligence is now in sight. And that's kind of the breaking news. This is like when we talk to DJ Co over at Neuralink and they said we are starting to see glimpses of superhuman powers from Neuralink patients. And we kind of were so in the midst of the conversation that we kind of missed that scoop.

But reminds me of the Buzz Lightyear meme where it's like everyone is saying superintelligence is now in sight.

Yes, yes, yes. Yeah. I mean the question is like, it immediately, it immediately asks like, okay, what does that mean? Because super intelligence, broadly, in some ways it means like higher IQ than any human. But also this idea of, of it, it's self improving, like that is the definition of superintelligence in many realms. And so the question is like, what was this glimpse? Is this because you can make the argument that using AI to write code in an at a foundation model lab is super intelligence even if there's still a human in the loop, you have to take the human completely out of the loop. And then what is the prompt? Because if the prompt is Just is just go fix this one little bug and it does it and it does it agentically and no one touches it. That's different from the prompt being like improve the system or never having to.

Prompt it at all.

Yes, yeah, yeah, exactly. Because. Because you've, you've given it this plan of just let's.

It's agentically, proactively. It just wants to improve itself.

Yeah. And then there's also kind of like.

How nobody has to tell you to bench press on Mondays.

And. Yeah, and the flip side is like you can like some of the algorithms for reinforcement learning feel like they are self improvement because you are not laying down a specific. There are emergent qualities where like when they trained GPT 4.5, it doesn't feel like that was RL'd on four Chan Green texts, and yet it gained the ability to do that and make them funny. And it wasn't necessarily a, like a stated goal. And so you could imagine that that's an example of this. But all of these are kind of squishy definitions. But it's certainly a, a good rallying cry that we're on the cusp of this. And, and many of the foundation model lab founders and leaders have been kind of like teasing this and discussing how much have we seen here. But I think everyone wants more concrete evidence.

Zuck put out a video on, I'm sure it was on Instagram and threads and Facebook and everywhere else, but the video took an interesting angle. He basically says, he basically was like, other players in the industry want to automate all valuable work. And his message was effectively, we want to give super intelligence to the people. And so Zuck is, yeah, he's finding his wedge into the market.

And he kind of did that with Instagram and Facebook in some ways in the sense that like he gave everyone a printing press, he gave everyone a TV show or a distribution, potentially it's in the algorithm. You still have to win. But everyone has the ability to publish now, thanks to, and YouTube and a lot of other services, of course. But you know, the modern social media networks enable anyone at basically zero cost to stand up an operation that previously had very high fixed costs in the sense of like you needed to own a TV channel or put wires in the ground or buy a radio tower to distribute your content. And now you can just click a button and go live or something like that.

Yeah. And the messaging was broadly around AI to help you improve yourself, level up, learn things like that. So it's a very optimistic message, far different than what Some other executives in the industry have shared, which is all white collar.

Like the most extreme, where it's like we're not even going to give anyone this product. Like we are going to have full control over the super intelligence. We're building the same thing, but we're distributing it. Zuck wants to make it free for 5 billion people. And Ilya, so far, you know, he hasn't really talked about this too much, but the vibe has been no one will have access to it. It will, it will be completely self contained. And so a dramatic shift from, yeah, anthropic.

Anthropic as well, was saying, you know, all white collar work or 50% of white collar work being replaced.

Yeah, they kind of walked that back. Short timeline, but yeah. So the white collar work thing is interesting because Zuck does address that a little bit. So let's just read through a little bit more of his. He says it seems clear that in the coming years I will improve all of our existing systems and enable the creation and discovery of new things that aren't imaginable today. But it's an open question what we, what we will direct Super Intelligence towards. And this is the question on like white collar work. I suppose in some ways this will be a new era for humanity, but in others it's just a continuation of historical trends. As recently as 200 years ago, this reminds me of that guy saying like 300 years ago we were living in caves. But this is actually true. 200 years ago, 90% of farmer, 90% of people, of humans were farmers. Growing food to survive. Advances in technology have steadily freed much of humanity to focus on less subs, less on subsistence. Subsistence, subsistence farming. Farming for yourself. Growing food just to keep yourself alive. You need food, so you grow the food and then you immediately eat it. And then you go back to working on the food and more on shoes. At each step, people have used our newfound productivity to achieve more than was previously possible. Pushing the frontiers of science and health as well as spending more time time on creativity, culture, relationships and enjoying life. I am extremely optimistic that superintelligence will help humanity accelerate our pace of progress. But perhaps even more important, that super intelligence has the potential to begin a new era of personal empowerment where every where people will have greater agency to improve the world in the directions they choose. As profound as the abundance produced by AI1 may1 day be an even more meaningful impact on our lives will likely come from everyone having a personal super intelligence that helps you achieve your goals. Create whatever you want to see in the world, experience any adventure, be a better friend to you, those you care about, and grow to become the person you aspire to be. Meta's vision is to bring personal superintelligence to everyone. We believe in putting this power in people's hands, to direct it towards what they value in their own lives. Yeah, super intelligence for the people. This is distinct from others in the industry, not naming names, who believe superintelligence should be directed centrally towards automating all valuable work, and then humanity will live on a dole of its output. At Meta, we believe that people pursuing their individual aspirations is always how we have made progress.

You know a job that was impacted by technology? What? Knocker uppers? These were human alarm clocks that would walk the streets of London with either a long stick or a device to shoot frozen peas at windows to wake up workers in the morning. So before.

Real thing.

Yeah, this was a real thing before clocks, alarm clocks specifically were widely used. You would effectively pay somebody a small amount to come up to your window in the morning and hit it with a long stick or shoot it with peas. So this was a job that was impacted by, again, alarm clocks becoming widespread.

How'd the rooster industrial complex fare with that?

Yeah, the roosters were rocked by the original alarm clock. The original alarm clock. If I didn't have neighbors, I think I'd get a rooster. Have you ever, like. Have you ever. Have you ever spent. Yeah, they were called knocker uppers. Have you ever spent. Have you ever, like, stayed at a.

Farm chicken and accidentally.

Accidentally got a rooster? Yeah, that happens.

It was fine. It wasn't that big of a deal. Yeah, I woke up pretty early.

No, it's. I mean, it's nice as an adult. Like, somewhere around 5am it starts making noise. It's kind of a nice noise.

I get up earlier, I wake the rooster up.

John wakes up the rooster. You walk out and you're in your yellow. In your ramp suit.

Pees at the rooster in your ramp suit.

Just. Yeah, very good. Well, speaking of white collar work, the Fed holds rates steady, but two officials are backing a cut. There's dissent at the Fed, highlighting a fraying consensus among policymakers who are debating the effects that tariffs will have on the economy and inflation. So the Journal just put this out a few minutes ago. The Federal Reserve held rates steady for the fifth straight meeting Wednesday, but are faced with rare dissents from two officials seeking an immediate cut.

These are officials within the Fed.

Yes. The decision followed.

Obviously, they're facing.

There's another official who's also in agreement with them.

Venture capitalists.

Venture capitalists. I was thinking about that the other day. It is interesting that if you look at real estate and private equity, VCs always get memed for benefiting from low rates. But in many ways VCs can continue to deploy into companies because they're just betting on that company's prospects over a very long period of time. If you're trying to deploy into real estate and traditional private equity, when rates are high, things are way, way less workable. So I think it's, I think big real estate and big PE are really trying to shift the ZIRP blame on the venture capital class and I won't stand for it.

Yeah, I mean there were a lot of companies that, like startups that got hurt during the interest rate rise. Some fintechs were sensitive to rates, but many weren't. I mean you think about banking products and if they're like all of a sudden we went from offering our customers 2% interest on their deposits to 6% or 5% and that just allows us to take, you know, a fraction of a larger, you know, interest rate payout every month that actually improved.

Their economic mercury is the best example of this. Every, every time the.

Were they profitable?

Yeah, yeah. They're printing money. They still are. And every time the Fed would hold a meeting, their revenue would basically go up immediately with no, they didn't need to hire new people, it's just default based on the fed funds rate.

So that kind of like, you know, it hurts some of the fintechs. It helped some of the fintechs.

Yeah.

And then a lot of just the normal software companies were like completely unaffected, sort of loosely. Broadly.

Loosely affected because maybe VCs couldn't raise their next fund.

Yeah. But it was like very, as opposed to like we know examples of people who started real estate roll up plays, bought a bunch of facilities, put got variable interest loans on them and then.

And are now producing like 1% yield.

Exactly. And then when the interest rates went up, their mortgage payments on all those assets went through the roof.

Yeah.

And all of a sudden they were upside down all of their economics. So.

Yeah.

So very rough.

Yeah. If you went heavy into SF real estate in 2018, 2019, 2020 on variable rate loans, you just have gotten absolutely even.

The meme of like buy a boring business, that's the example I'm thinking of, is, is okay, you bought a, you bought a boring business that cash flows, you know, a million dollars a year, but you bought it such that you have, you know, okay, 500,000 of that cash flow was going to service the debt that you used to buy the asset. And then all of a sudden your interest payment on that loan doubles. Then you're not making any net profit all of a sudden. And that really puts you at it.

So Fed officials maintain their benchmark policy rate in a range between 4.25% and 4.5% as they weighed how importers, retailers and consumers will split the costs of bigger gong hits. Sorry, cost of higher duties on imports. The outcome of hand to hand combat over who will bear the burden of tariffs. Figures to shape the trajectory of inflation and hiring later this year, which could resolve whether and when the central bank resumes cuts in the months ahead. So hinting at a September rate cut, or at least that's how people are kind of like reading through the lines here.

Yeah. I mean, also, we blew out the GDP numbers, right? It was like 2.2 was the expectation that came in at 3.

3.

Amazing.

Quick hit. A lot of people are. A lot of people are. Yeah, we got a. We have a new gong here in the great city of New York.

Okay, I want to go back to the personal super intelligence thing for, for a little bit with Zuck. So this is distinct from others in the industry who believe superintelligence should be directed centrally towards automating all valuable work. It feels like maybe a shot at SSI or anthropic or OpenAI. But I feel like I've noticed that Sam is. Yeah, maybe they're turning down Zuck's billion dollar offers, which will get.

Allegedly.

But the, the interesting thing there is that I feel like Sam Altman's been sort of pivoting his rhetoric around this. And he said something similar recently where I think it was in the soft singularity. He said, in that blog post, he said something similar to as recently as 200 years ago, 90% of people were farmers. And his point was, if you went back in time and you showed a subsistence farmer in the year 1800, hey, there are these people like, let me take you on a tour of the modern economy. Here's someone whose job is email. Is email.

Yeah, that, that, that, that's, that, that, that like 100 years from now we're going to look back and be like nine, you know, like, why were they doing email? Everybody that went to school became an email professional.

It's like, yeah, yeah, yeah. Or like calendaring or phone calling.

I mean, wait, they had to use the computer. They had to type. They had to, like, physically.

Why didn't the robots just do that?

Y.

You need key. If it's a, if it's a machine, why do you need to hit.

Yeah. And it's scary because it's like, okay, well like if I'm not doing that, am I doing anything?

Like my email job?

Yeah, yeah, yeah. My email job is directly tied to.

To my self worth. Yeah.

And my income. I'm a subsistence emailer. Yeah. And it's the same thing if you went back and said to somebody in 1800, hey, you're growing food, but in the future you won't, you won't have a job as a farmer. They'd be like, well then I'm screwed.

Right.

Like I'm, I'm done for.

Yeah.

It's like, no, you'll have an email job. You'll be able to, you'll be able to do email all day. But we need to debate more about like what this actual product looks like because we were talking about that earlier. We know that when you, they're charging you 20 bucks or $200 a month because it's expensive. And we were talking to George Hotz about this and he was saying like, they're losing money on me just because the inference cost.

Yeah. Claude came out yesterday and said one person was costing them tens of thousands of dollars a month on their $200 a month plan.

Exactly.

There was also. So Sam is projecting to spend 35 billion on inference and 55 billion on training models between 2025 and 2027.

So that I'm less worried about the training model. I see that as capex. I'm more thinking about the total inference cost right now across all the major labs. Like it's got to be in the billions. It's probably not in the hundreds of billions. If you look at how much OpenAI and Anthropic have raised and how much they're burning, a lot of that goes towards salaries. But if you just look at the actual inference cost, it feels like Mark Zuckerberg could eat that cost for a long time. Yeah. And then if he wants to keep it free, there's the very logical path, which is ads. He already has the most robust, most targeted ad network and he can bring all those ads in and start monetizing those users. And so the more you're using, the more ads you see it's perfectly aligned. It's an auction. And so someone who's doing low, low economically valuable work will see like cheaper ads. It's perfectly price discriminated. So I don't see Meta as having problems like scaling inference and keeping it free. The questions going back and forth.

Yeah, yeah. The questions I have are that. That wasn't exactly answered here. Although reading between the lines, it's maybe less important. It sounds like a shift in strategy, but doesn't. When I read that announcement, I don't think, oh, they're going to go heavy into open source. Like that doesn't scream. That doesn't, that doesn't scream. We want to be the number one open source model. They might still, but this feels more like an application play to me.

Yes.

The areas that you can imagine Zuck wants to dominate in, are in create the, the action of creation. So content creation heavily ties into the meta ecosystem already. Their business is dependent on, you know, billions of people globally creating content all day long, some professionally, some recreationally, but all of those people he wants to serve there. The knowledge kind of learning. I think he's signaling here that he cares about. He clearly understands it. ChatGPT has incredible PMF around, around just like learning, understanding the world, etc. And I could see him playing there. The areas that I'm interested to track going forward are, are how much, how much, how much resources go into this companionship. Right. That's another area that's like driving up user minutes on ChatGPT. Right. And we've talked about it before, that can ultimately threaten the time that, you know, if people start spending two hours a day with ChatGPT, that's potentially two hours a day that they're, you know, maybe, maybe they have it on audio mode and they're still using Instagram, but it's, you know, potentially threatening user minutes on his platforms and then the big one, probably a long shot, but just search. Right. I'm sure he's always want, you know, search multitrillion, dollar market. Very, very.

He was going to launch a search engine for a while. Yeah, like there was where you like the search bar in Facebook was getting so powerful that you could search the web never became a serious Google competitor, but was a. It was a.

The other thing here to track is. The other thing here to track is threads. We gotta like do a proper like new deep dive on threads. I don't think when it launched we didn't even have the show, but it sort of launched, peaked in the charts, fell off completely. But I think it's building back up. There's actually quite a lot of activity there and he's gonna have confidence to see like, Okay, I have WhatsApp, I have Instagram, I have all these different platforms that I can drive users to have the best ad engine in the work in the world if I want to. If I want MSL to create these like new consumer applications, I'm going to have the ability to get those in front of billions and billions of people. So it is a massive unfair advantage on the just like training and inference and like capex side and then massive distribution advantage. OpenAI has benefited from, you know, tremendous organic virality. Everything from just people telling their friends to these studio Ghibli moments.

Yep.

Etc. But, but yeah, it'll be interesting to see. Okay, where does he really want to compete? Does it look like something like Chat gbt? That Chat GPT again is this like incredible tool that allows you to do all these, you know, anything you want. You can study with it, it can keep you company, it can help you with your homework, it can help you write emails, it'll transcribe a meeting, it'll, It'll make an S1 for you. Right. And so that, those are the big things. It's like, okay, we have a much better sense now of where the team is going to be focused. Again, it doesn't read like open source to me, but it's still. There's so many different ways that you can kind of like chop up the, the category and the opportunity.

I have a rebuttal, but first I'm going to tell you about Vanta. That's automate compliance and security and trust with vanta. Go to vanta.com to check it out. So my rebuttal is that the, so the question that I have, I agree with a lot of what you said. I guess the question is I have this personal super intelligence. Do I want it in a meta property app right now or do we need a new app? And then if I need a new app. That's really hard.

That's what I was saying. Well, really hard. But at the same time how many ways I, I re downloaded threads because I kept getting notifications and it would tell me something to the effect like it would give me three words out of like an interesting, you know, post and I could only see it if I re download edited it.

Right.

And so I'm just wondering about if.

I, if I'm, if I'm interacting. If I go to Instagram to do my homework, I'm going to get sucked into reels, right? And it's like I want to go to the library to study and ChatGPT feels like a library and it's been very difficult even if you have massive distribution to actually scale share of queries. Even, even, I mean, Threads, they have.

What, 7% right now?

So threads, I think is a special case potentially because of the political dynamic of Elon Musk. Whereas with, there's no political dynamic between ChatGPT and Gemini, Google has massive distribution. They've been able to get to 450 million DAUs or MAUs, but they haven't been able to break over like 10% of queries. Share of queries, Something like that. And so I wonder if, you know, you launch a consumer app for studying and your personal superintelligence, like the logical way to distribute Meta's AI projects and products is through their existing platforms.

Yeah. And so, yeah, and you can imagine that you're in the Instagram tab, you have a boring picture, it makes a Ghibli version of it.

Even if you're in a. What like imagine you're in a study group with a couple people in the same class in a WhatsApp group chat. And then there's. There's a super intelligent person, there's a super intelligent agent or AI in the chat with you and you can talk to it and you can answer it. Right.

Well, so, and to be clear, they have the Meta AI app. Sounds like you're not a DAU yet. But. But it's number three on productivity. I imagine Zach wants to get that to number one. Okay, you can imagine the other form factor here is obviously glasses. Right. He's just invested billions into owning a piece of luxottica. I imagine as part of that investment he gets. It's going to probably make it so that other AI assistants can't get access to.

That's actually maybe the better. Like personal super intelligence, like tutor.

No, and he says that in the video. He says that in the video. A lot of, A lot of. A lot of what he says in the video.

Glasses to read a book. If you have, if you're, you know, farsighted, you put on the glasses and so you put on the glasses to study. And then it's. It's seeing the textbook that you're interacting with. It's seeing your screen and you do. And it. But, but specifically, like the Meta Ray Bans don't have the ability to show you an endless stream of reels. And that's what I'm most worried about is that if, if.

No, they do though. Do you remember that? Do you remember that?

Orion does, but not the matter.

I know, but that's where we're going.

Yeah, I don't know, maybe, maybe you.

Don'T think they want to merge Orion and Ray Bans in the future?

I don't know. I just think that like it, it would be like, I mean it's already.

You can already do products.

If I choose two different products for an educational like environment and I have one that's ad free and distraction free and the other one is like, do you want to see this? Amazingly viral.

I'm not, yeah, I'm not betting, I'm not that, I'm not betting that American students will be using Instagram to study with them at all. But, but yeah, it does feel like, you know, Zuck's vision is the, the, this, this is. He's making a big bet on, on, you know, Vision VR AR being the future computing platform. He talks about this saying, you know, you're going to be in your kitchen looking at your fridge and it's going to be like, hey, did you know that you could make this kind of quesadilla? I see you have the ingredients, I'll give you a recipe and then you just like make it and guides you through it. And then again, like you said, you're reading a textbook and it's popping up. You can ask questions in real time to the AI based on the information that you're seeing or you're studying or even the interesting data point. I think it was yesterday they came out and they're saying they're letting applicants use AI in coding interviews.

Yeah, we. Which is total Cluly victory.

Total Cluley victory. Credit to Roy Lee there for making.

Cheating controversial branding, but directionally correct.

Yeah, for sure.

Well, let's go into the potentially bigger bottleneck for superintelligence at msl. But first let me tell you about Linear. Meet the system for modern software development. LINEAR is a purpose built tool for planning and building products. So interesting comment. This comes from rehearsal. An interesting comment from former Meta employee. Energy is the biggest bottleneck right now. Even if Meta wants to spend 100 to $150 billion on capex for AI infrastructure, they can't. It's not just Nvidia, it's not TSMC. Transformers, power equipment, cooling equipment and the availability of power is all limited right now. Schneider Electric is completely booked until 2030. Even if you have the money, you can't spend it. And this is from. Oh, this is a screenshot from like a GLG caller. Tigus call or something. He won't. So is he serious? It must be Alpha Sense or something like that. So they're Talking about Meta Capex and they talk to this anonymous expert who is clearly a former meta employee says, yeah, there's no way you can spend $150 billion right now. If we look at any of the transformers, not the, not the machine learning architecture, but the actual physical energy transformers or any power equipment or cooling equipment or even construction or availability of power, there's no way it's going to happen. You talk about, if you talk about Schneider who supplies a lot of the power equipment, they're completely full until 2030. They're not negotiating on price. It's basically you're lucky to get an order. Everyone talks about Nvidia, but there are many other things beyond Nvidia for this thing. Even if you have the money, you can't spend it.

And Josh Steinman quoted this post and said, if you're an American engineer and want to lead the technical aspect of building a transformation transformer company, my DMs are open. So apparently there is an American. Yeah, a new American Transformer.

Remember I've been asking, I asked the transformer co founder architecture, core weave and I also asked some semi analysis folks like we've all learned about Nvidia, then we all learned about tsmc, then we all learned about asml, then we learned about SK Hynix, the memory supplier. Like we've learned the entire supply chain of AI. And I was like, what's the next company we're all going to need to learn about? And it sounds like it's Schneider Electric. I've never heard of this company before for today. But we gotta, we gotta dig in and understand. Do they have competitors? Is there an opportunity? What's going on? I'm very curious to know how electric works. But continuing on Zucks, this is a.

French company, French company founded in 1863.

Founder mode.

So they're. I don't think they're still gonna be hard to be in founder mode if you got started back in the 1800s, but they did about $36 billion in 2023. So massive, massive company market cap 137 billion euros. Okay so I don't know, that's like What, a billion USD?

50 bucks. Anyway, will Depew says the bigger story is not that Zuck is giving out $400 million offers, is that people are turning them down. What that might, what might that mean? Of course will works at OpenAI. Not conflicted at all. Zuck is targeting mirrors, lab thinking machines with offers between 200 to 500 million made to a quarter of their team. And one over $1 billion and not a single person has taken the offer. And there's, there's some people going back and forth on like, what that means. Maybe you just believe in the mission and you already have 1% of the company and you think it's going to be worth a couple hundred billion. So you're like, well, just wait and deliver on this and I think we're going to win. You could also not want to go work at a big company. There are a variety.

Also, if you've worked in AI for the last four years as a top researcher and you're not worth at least 50 million by now, like, what are you doing?

Yeah, a lot of the thinking machines folks, like, the reason people speak so highly of them is that they were at other foundation labs, great labs, early on, and probably got big stock grants that have grown a ton, but they've.

Been able to get liquidity against. And this is.

Yeah.

I mean, there's so much information warfare happening right now. Right. There's so many different incentives for people to say, oh, it was this offer and they turned it down. Or there was no offer made. Or it was, you know, somebody from Meta was chiming in and saying, like, we made a few offers, one was big. But there's a lot of misinformation here.

The real problem is, is that there.

Aren'T bull market and misinformation.

The real problem is that there aren't, there aren't enough luxury goods. Like if, if Zuck really wants to make people bite at a billion dollar offer, like, what's the difference between making 100 million a year and making a billion? It's like, what are you gonna buy? Yeah, maybe a boat. But he should start a new watch company, a higher end. Richard Mille.

That's right.

Five million dollar entry point.

I think, I think you're onto something.

A hypercar company that starts at $75.

Million should just buy the top five.

Somehow I don't think this stuff appeals to AI researchers. Yeah, you don't see a lot of them.

They're immune. They're immune.

Flex culture.

Yeah.

Yeah. I mean, also, if they're super AGI pilled, like, they just are, like, what's the value of money? In a couple of years, who knows? We're going to be in a completely different economy. There's a whole bunch of different. There's a whole bunch of different reasons, but maybe they like the team.

Yeah. The other thing is, I think many people will be uniquely motivated by winning and challenges and, and then and then having a, you know, do you want to be one of the top five people at Thinking Machines or do you want to be the number 30, the 35th, you know, researcher at a bigger organization? A lot of people want to have that.

There's also been so much like memetic warfare around this where, like, if you take the offer, you're instantly branded as a mercenary forever. And you can never be seen as a missionary.

I was getting, I was getting some notes from. From a few people about. So we released the Metis list, a collection of top AI. A loose. A loose collection of top AI research yesterday. It's just a bunch of cool people. Alan Turing people were pissed that he was included. They said he fell off, was underrated. But. But yeah, there was some. I was getting some interesting notes around a handful of people in the top 25 saying, you know, this person was kicked off of every team that they've ever worked on. Which is funny. They got to be way lower. They shouldn't even be on the list.

So anyway, it was all in good fun. We love everyone on the list and many more, and we're going to expand it and add to it and have more fun with it.

Hopefully eventually everybody will be a researcher. And the other thing is, Elon came out yesterday right after we published the Metis list, saying they don't have any researchers, everybody's engineers. I think it's a good framing. I think it's very possible that some researchers want to self identify as engineers. And if you're an engineer, I think it's cool to all be in the same bucket. And so I think that, I think that can make a lot of sense for Xai's culture.

Well, whether you're a researcher or an engineer, you got to get on graphite. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. It is an interesting debate about are we in an age of research or engineering at this point? And it feels like Elon is the master of engineering and is rarely trying to come up with.

And infrastructure.

Yeah, infrastructure. He's rarely trying to come up with the completely novel new science, but he believes that Grok at a certain point will be able to do that. But if you see how fast he's.

Raised, he's going for the greatest feats of engineering.

Engineering. And that's the culture is like engineering.

He's like, I started a research. I started a research nonprofit almost a decade ago called OpenAI. I'm over research.

I'm over research.

I'M doing engineering now.

Anyway, let's read through Ben Thompson's analysis of the Figma S1, the Figma OS and Figma's AI AI potential finally course in Manhattan for Figma's IPO tomorrow. We're very excited about that.

We will be live from Nice, the New York Stock Exchange.

Yes. Cannot wait. So Bloomberg has a story here. Figma's initial public offering is approaching 40 times oversubscribed, according to people familiar with the matter. Is that good the design and collaboration software? We'll have to. We'll have to ask everyone tomorrow if it's good.

Did they get the same revenue multiple oversubscribed?

They are could be the year the year's most in demand listing the company is guiding prospective investors at its IPO, which could raise as much as 1.2 billion is expected to price above the marked the marketed range, the people said. The San Francisco based firm and some Investors are offering 36.9 million shares at 30 to $32 per share after increasing the range Monday from $25 to $28 each order taking is expected to end at noon on Tuesday in New York, Bloomberg News has reported. Oversubscription refers to when the number of orders for shares in the IPO exceeds shares that were originally offered.

And so a lot of demand for figma shares, which makes sense. It has been one of the most in demand secondary helps you think bigger and build faster. That's right.

Design and development teams build great products together. What's not to like?

Nothing like mixing an ad into at least this is Ben Thompson's analysis. Right. So Ben Thompson is not sponsored by figma.

No. We had a friend of the show come on and break down the S1 when the, when the numbers dropped and it was kind of like the dream case for SaaS. Right?

Coming out and immediately being the top, you know, from a purely metrics based look, the top SaaS company in public.

SaaS company and of course a lot of people point to the rule of 40. You add the growth rate to the burn rate, how much money you're losing. If you're losing a lot of money but you're growing really fast, that's fine if you're, if you're making a ton of money but growing slower. Either way they should add up to at least 40% higher the better. And if you add figmas up they are, they are in the 70%. So well beyond 40. 40% according to the rule of 40. Which is of course exciting to the venture capitalists who popularized the rule of 40.

They basically created profitability.

So, he says, I think you can make the case that an $18 billion valuation is in fact too low. Indeed, I think you could make the company make the case that the company is worth more than 20 billion. Adobe agreed to buy it for 20 billion back in 2022. That, of course, is despite the rise of generative AI, which at first glance appear to be a threat to Figma. Because you just go to ChatGPT and say, hey, build me a collaborative software tool and maybe it just makes mistakes, don't make mistakes. But he says, I would go in the other direction. Generative, Generative AI and the opportunity it resents, it represents is exactly why Figma should go public, as Field strongly suggests, and get ready to be acquisitive, get.

Ready to get ready to get bought.

Do some deals we love. We love a Dylan Field deal.

The Figma OS Ben says the reason why I'm optimistic about Figma comes from my belief that Figma is not simply design software. Rather, it's something much more fundamental and valuable, an operating system with a collaboration built in. Here is how I explain Figma back when Adobe tried to acquire them. Forgive the long excerpt, he says, but it really is the point as far as my optimism about Figma is concerned, he says here. It is important to note that Figma's most prominent victim to date has not been Adobe, but rather Sketch. The explosion in apps led to a growing market for product and design, product design and development, as designers needed to lay out a whole host of different app views and developers needed to know what exactly the designers wanted them to build. Photoshop and Illustrator, which were primarily focused on the creation and editing of single files, were the completely wrong tools for the job. Sketch, which was released in 2010 as a Mac OS app and is still Mac OS only, although it now includes the ability to view files on the web completely, took the category by storm. Adobe eventually responded with Adobe XD, but not until 2015, and the product is still in third place. Sketch, though, is now only in second.

Really quickly, let me tell you about adquick.com, out of home advertising made easy and measurable. Continue.

Jordy Chris, the CEO of Ad Quick, is in New York as well. We're going to try to meet up with him in a little bit. Sketch was not a disruptive product. It was like Adobe's products, a desktop app. It just delivered a solution to an emerging product category that Adobe was slow to respond to, and that was like the designing of web and mobile Products. Figma though, was something completely new. The company, which was founded in 2012, made a bet on the browser and spent a full four years building V1 of the product this included. Which is just so funny because normally if you talk to a founder building a software company and they said yeah, I'm three years in and I haven't shipped an mvp, you would just immediately write the company off because like you know, why haven't you shipped? In this case it was necessary.

They quit your job, raised $10 million vibreal launch video slop like product in two seconds. That's the flow and then a ton of bugs and then sort it out and then maybe it works, but it's a good thing.

Don't worry about securing personally identifiable information at all. Just get to the top of the charts and figure it out later. So this included writing the editor in C, cross compiling it to JavaScript using the ASM JS subset that let it achieve desktop like control of memory and performance performance and building its own rendering engine from scratch using WebGL, which had only been released in 2011. WebGL was the why now for the business and Dylan was very early to recognizing that it was the epitome of leveraging new technologies to compete on a new vector, which in this case was collaboration. And you know, anybody, anybody under the age of 25 like probably doesn't remember the pre pre Figma era but like having different files and sending them back and forth and like sending comments by email, it's just like it. Even, even personally I only briefly was was working without having access to Figma. So Ben says, think about my brief description for this product design and development segment. Multiple app views created by design and implemented by engineering imply at least two different people working on a project, a designer and a developer. In reality, the huge number of views and increased functionality in apps meant that there was an increasingly large teams on both sides. Sketch definitely made it easier to design an app in one place, but it did nothing to make it easier for teams to work on the app together. Figma, because it was native to the web, effectively got collaboration for free from the beginning.

Yeah, this is a crazy screenshot. I don't know if we can show this, but on Figma's website the company compares itself to Adobe xd, but I use Photoshop and Illustrator. Don't stop. We believe you should use the best tool for the job. If you're photo editing, choose Adobe Photoshop. If you're doing detailed illustrations, Adobe Illustrator is a no brainer. But if you're looking for the best tool for UX design. Figma is here for you. Plus you can easily import images and SVG code into Figma and have your cake and eat it too. And so putting the competitors products right there. So and recommending Eventually Photoshop and Adobe did launch a competitor called XD& XD wound up is getting killed by Figma. But no one ever subscribed to Creative Cloud for xd. You just used it because it was here, says Ben Thompson. Here's the problem though. To the extent that Adobe's products were being used as part of figma's workflow was the extent to which Adobe was slowly but surely losing control of its long term relevance. Figma could for example, start investing much more heavily into its photo editing or illustration capabilities, rendering Adobe's products increasingly unnecessary. What is more threatening is figma's cross platform capabilities. Third party developers can build plugins for figma that make it possible to easily add the sort of editing capabilities. So regulators blocked the Adobe deal and Ben Thompson says he thinks it was correct. The regulators got it correct that time and based on the run that Figma has been on ever since, it seems like it could be a win for them as well. Yeah, and, and it's interesting because a.

Lot insane roller coaster for the team. Yeah, and, and Figma I think during the acquisition process clearly didn't have the insane urgency to launch new products because they thought they were going to, you know, Adobe and would be part of an organization with a bunch of different individual apps and, and watching how they reacted after it was blocked and then came and just started shipping like crazy.

And so the FTC has gone after some, some mergers that have made no sense. Like they, I believe they blocked like Roomba. The Roomba one. Yeah, that one didn't make sense. The worst example I can think of is Meta bought a VR fitness app and they said that Meta was trying to create a monopoly in the VR fitness market, which is just not a market at all. It doesn't even exist and I don't know anyone that does VR fitness and it's very rare and so it totally makes sense to let that team get an exit and go hang out at Meta and try and work on this until creating the market, creating market and then as soon as it takes off there's gonna be a ton of competitors immediately. And so it seems like totally fine to let that one go through. And then there are, there are other sides where, where like truly there are like two players and they're merging and there's going to be a lot of, a lot of competitive dynamics that come from that. Anyway, let me tell you about Bezel. Pre owned luxury watches authenticated in house. Bezel is the top marketplace in the world for authenticated luxury watches. Yeah.

And again, just more context. So Amazon, Amazon's acquisition of iRobot, the maker of Roomba, was blocked due to antitrust concerns, which was primarily from the European Commission. They were worried about a monopoly in home robotics, I guess, and the CEO ended up just resigning.

What's interesting is like the robotic vacuum market is extremely competitive because of a lot of the Chinese companies, UFI and there's a few others. Like, it is definitely not as much of like, it feels like a lot less of a sticky product because yes, it maps your home and it's supposed to learn, but really if you move to a new place and you get a different robot vacuum cleaner, like you're pretty much good after a day. You're not like, oh, I can't get out of the Roomba ecosystem. Whereas like, even Sonos is stickier than that. And so I never had a problem with that.

And IROBOT shares are down 95% over the last.

Looking over there, we'd actually got blocked at the. Then they just got cooked.

Yeah, they actually went public July 31, 2020.

No way.

Stock popped to $133 a share by early 2021. It's now sitting at 428 million, 130 million. So a little bit of value destruction from the European.

Imagine how frustrating that'd be. You're like, I think I'm getting out to like the most logical acquirer possible. So figma position and the AI potential Consider the position figma has in the software design process. First, figma enables collaboration, their original differentiator from Adobe and Sketch. Second, figma creates one canvas for app creation, everything based on the same URL instead of files that are passed around with increasingly absurd file name appendages. You know, final, final, final, v3, v4, v1, final, final. Third, Figma serves as a source of truth for an app's name design. These capabilities seamlessly translate into AI enabled workflows. And so.

And these capabilities seamlessly translate to AI enabled workflows. Collaboration today between a host of humans could very well mean coordination and orchestration of AI. Critically, figma isn't assuming or relying on an AI doing everything. Granted, that is a theoretical threat to the company. But the problem with AI only development processes is that they are extremely difficult to debug or Decompose. Figma gives a natural canvas to add in AI.

Yeah. So Ben Thompson kind of closes out talking about figma as the OS of design being important. What operating systems have is coordination and orchestration abilities. One canvas where all the work happens.

The source of truth.

And then for now, for now, figma gets its nascent AI gates the capabilities behind a full seat license, I. E. The most expensive plans. They also have usage based model. But they. Ben Thompson is advocating for figma to move towards some sort of usage based or task based monetization model. This is what selling workforce.

Yeah, Salesforce selling the end product.

Exactly. As opposed to based. And so that will be a business model transition. And he says that the transition will be tricky and it's arguably the best argument against figma going public. Because if all of a sudden you shift to hey, we don't charge per seat, we charge per finished design or finished website or shipped website, then all of a sudden you could see that you wind up with a much better business model. Four quarters out, but you have two rough quarters and the public markets don't like that. And people always talk about that being one of the reasons. One of the drawbacks of being public is that is that you have a lot of investors that are a little bit more short term oriented. But he says, I think however that it is notable that Field mentioned M and A dur in his argument for going public. Figma would do well to start acquiring AI companies to plug into every aspect of its stack. Particularly now when AI capability is nascent and not yet a threat to figma as a whole. Above, I was skeptical about AI replacing figma, but that is a skepticism that is relevant for the short term. Figma's task is to replace itself before anyone else has the chance in the long term. So.

And I am very excited for tomorrow. Before we have our guest, should we talk about friend.com we should.

We're. Well, we're going to give to call them later. 30 minutes. Quickly. Let me tell you about Adeo customer relationship magic. Adio is the AI native CRM that builds scales and grows your company to the next level. And we have Eric Gliman in the studio. Correct.

Get over here. Go Eric.

Congratulations.

Looking sharp.

Oh my God. Great, great.

I love these.

Have a seat here.

Great to see you.

Great to see you. Good.

You got a good. You got a Celsius there.

Fantastic.

Welcome to the show.

Thank, thanks.

Our humble New York temple of technology.

Yeah, give us the update. What's the news today? Break it down.

Great jacket by the way, you got little touches of yellow in there.

It's not yellow enough for my taste, but there is a little bit of yellow.

By the way, our tailor designed these to celebrate and, and really symbolize saving time and money. And so I think they did the job. I think they did the job.

I, I understood the assignment.

We'll get you. We'll get you. We'll get you one too.

We'll definitely get you.

We got to do this.

We got to go. Man on the street in, in like Times Square. What, what do you do for a living? What, what business do you run? Do you want to save time and money? Let's get you on Ramp right now. We'll introduce you to the CEO ramp.com. phenomenal. Yeah, but break it down for us. What happened today?

Yeah, well, first, I mean, we might need a bigger gong.

I know.

You have no idea. We're going to the ends of America to get an American made gong. And we're looking for something that will actually be floor to ceiling. You know, it'll be, you know, something like 20 by 20. So working on that. I look forward to hitting it for the next one, the Series E3.

Yeah, bring it down.

Phenomenal, guys. Thanks for being here. Here in New York, the capital of capital, it really is technology here at TVPN. No, it's. We raised $500 million at a 22 and a half billion dollar valuation.

There we go. There we go.

Has a nice ring to it.

I'm really appreciate all the. And I'm losing track, but it seems like Ramp has not been able to get a fundraise public without it leaking, you know, five days previously to the press. But this is official. I'm glad to make it official here.

It is official. It's the right valuation. And look, early on in my career I got this advice that the biggest battle a startup faces is people not giving a shit, people not caring. And so you know what, we're happy that people care even if things come out a couple days before, but we feel super lucky. And so that's the big news of the day. But I think so much what we focused on today, today especially in the raise, is what we're going to do with the capital. Between the last raise, which we were proud to hit the gong for here six weeks ago. Six weeks ago, some of it actually came from, actually from the product. We launched our first AI agent for controllers and the data was like nothing I'd ever seen before.

So that catalyzed this 100%. Yeah, that's my big question. Y, E2 why not? F. And then also it feels like you're like the king of like a bunch of fundraisers in a short amount of time. There's clearly some benefits to that, but why not?

I got to give some credit to the business, John. I mean Eric's, Eric's a great CEO, but the business is really, you know.

Well, I, we haven't talked about this before. Like, we'll share it. Like right now July is pacing, we're pacing to do 20, 20% more purchase volume in July than June. Like it's, it's been one month. Like, like the business is growing extraordinarily quickly.

What's that, what's that meme with the guy on Hot Ones where he's.

No, no, no, guys, don't, don't talk.

We need to it's have Celsius.

But no, no, no, that's insane.

The business is itself is, I remember.

I'm sure in, back in 2020 you'd be happily sending, you know, investors like a, you know, we grew 20 month over month, you know, but on like, you know, whatever 1% of the scale. Yeah. So, so really wild.

Yeah, I, I, I, I guess like the bigger like meta question is just like, like there's a world where you stack a bunch of these wins into the mega deck and then just do like one massive fundraise every 18 months. There's this meme of like fundraising is a distraction. The CEO should be out there working with customers, improving the product. Obviously you're doing that, but how do you do these back to back fundraisers without them being distracting? I know you have a big team now, but like what goes into the decision to do like, I don't know if they're even smaller, but just like more frequent fundraisers because that feels like something that's unique to your company.

Yeah, yeah, for sure. So one, it's sort of a funny thing, but people, I actually didn't understand this before in the startup world. Like what is a series A or a B or an E2.

Really?

Somebody who did a 55 million dollar seed round.

Yeah.

And then, and then we pressed them.

We were like, is this for marketing? Like, yeah, you know, you're still gonna get the expect like the expectation will be that you perform at like the, at the series B level based on.

That company, it's gonna feel like a series B company.

Like I almost think like calling things by letters is totally silly. All it actually meant was we had a set of documents that we use Whenever financing happened, there's hundreds of pages of documents that go back and forth between company and investor. Yeah, calling it an E2 just meant that we used the exact same documents.

Save time and money.

Save money on lawyers. No, it was really just that. And so look, it was very efficient to get the financing done. And I also believe a lot in this concept of marking to market of in real time, what are companies ultimate. And you know, when you look at the fundraise from about a month, month and a half ago, I think is when it was announced, done earlier, you know, we raised, I think in Exchange for about $200 million, approximately 1% more shares came onto the capital. So it was a very small amount of dilution. With this raise of 500 million at 22 and a half, it's like 2 ish percent. See, these are not that dilutive.

Yeah, it's not completely restructuring the company, but, but, but it's giving you a more modern mark.

Adding to the war chest.

Adding to the war chest. And I think it's a good thing because one, you know, allows people to invest at the right point in time. But two, because you're, you're not raising far ahead of where metrics are. It means that the company's not that distracted, you know, for, for the actual organization. Like we've just been obsessed with going, yeah, and you're not in a position.

Where everyone's like, oh, like at this valuation, like Ramp doesn't just need an Act 2, but they need an Act 3, 4 and 5. Like they gotta, you know, create super intelligence and all this crazy stuff because like, like it's like based on the actual progress of the company.

100%.

Right.

Makes sense. I want to talk about the product, I want to talk about the AI agent thing. The interesting, my experience with Ramp was that it was in many ways an AI agent before AI agents was a buzzword. And I know that obviously when you scan a receipt, it goes through an OCR process and the data, the text is extracted from the image. And then I would assume that you were an early adopter of LLMs to help clean that up. And that was in many ways a more like a super magical AI agent experience. But I'm not prompting it, it's just something that's happening in the background. Ramp, it's invisible. And you're using some, you know, deterministic code, some probably probabilistic code, some older machine learning algorithms for ocr, some newer lms to clean up the data and tag things and say, this is A restaurant, that's a flight and you know, do all the magic that helps categorize expenses. The question is like, it feels like a big shift to put the, the, the, the, the agent in the hands of the user. And so you said that you're seeing like incredible adoption. What are people actually using for? What's the feedback? And, and I mean like, yeah, what else are the learnings of like where that goes? Do people want that or do people ultimately want to be able to prompt a few times and then have things baked into the core behind the scenes workflows?

Yeah, so I feel like, so let's talk a little bit like, what are like AI assistants? What are AI agents? What's the distinction? Because I actually think people are like pretty, pretty sloppy to. And so I feel, well, to be.

Clear, if you want to raise money right now, you just build SaaS and call it an agent. Boom.

But let's talk. We have a lot of people who are building, who are listening to this podcast. So I think first for, you know, assistants or co pilots, the idea is a person prompts, the copilot, will review it will make a recommendation, but then ultimately it's on the human to, to go do the work. I get a recommendation, I input it back and forth. This is maybe similar to an experience that you might have on a chatgpt or that kind of a thing. An agent is a bit different. What you do is you give it instructions and access to tools and you say under certain conditions, if you see an outcome that I've instructed you to drive, just go do it. Right.

Forever, essentially forever.

And I'm going to modify things from time to time. But it's a little bit more similar to, let's say you hire an intern or a new hire, you teach them how to do the job, you watch a bunch of times and then they go on and you periodically check the work. And so that's a bit of the distinction. And one of the, if you think about what we're actually optimizing for, one of the core metrics is we look at what was the amount of work that a ramp customer was able to do in a single, let's say in an hour or in a minute. And so as we look over the years, that's gone up a lot. If, if you joined in 2023, you know, mathematically we can show that you are categorizing three times more transactions, pulling in three times as many receipts, whatever the metric is, as you did just two years ago, for every minute you're spending on the site. So it's all to say, like, every minute is counting for three times more. I think that ratio should be 30 times more. And if you look at what happened. So we launched, we made this announcement about three, four weeks ago for this first AI agent. And we had early adopters notion webflow Quora. The stats just absolutely shocked us when we looked at this. Functionally, these companies were able to get 15 times more work done, transactions categorized. But the more interesting part is these agents were way more accurate. These agents, all day, they know the expense policy better than any manager at the company. And so it was 99% accurate. It got a lot more tasks done. And suddenly, like for many people who've worked at large companies, you spend like an hour just reviewing expenses. And you're like, why am I spending an hour reviewing like someone's Uber or their purchase? And that's just gone. And you're focusing on actually interesting exceptions. And so where I think this kind of adds up to is for a lot of years, people had to be forced, forced to like, learn how does software work and how do I think like software? How do I learn to use Salesforce, how do I learn to use Google, how do I learn to use Excel? Now we're teaching these software programs to think like people. And that means you can teach software functionally the same way you would an intern to go and figure out what expenses are in and out of policy. And that very avid software that doesn't take breaks, doesn't take holidays, that is obsessed over every deal of your expense policy can catch.

When your policy online, if you, if you need to talk to it at. You're on a, on a work trip and you need a. Yeah or yeah. Therapy. But I was going to say, like, you don't. If you can avoid messaging somebody on the weekend to ask if something's in or out of policy, it's like it benefits everybody.

Yeah, yeah, that's exactly right. No, and, and I think you bring up like another, another point of like a lot of work is sort of if this and that, like, if you get a per. If you get a, like a, you want to buy a new piece of software and that vendor sends you like a quote, then procurement negotiates it. Once procurement's negotiated, then legal takes a look at it. Once legal says it's fine, then finance says, is there an invoice? And we go and pay it out and then it goes into accounting. But if you have lots of digital agents functionally, all the time online and responsive These things can occur simultaneously. And so it means actually instead of waiting days to, to go and onboard a new vendor, it can be done in minutes. And so it just means it's much easier to do work. And I think the big picture of like why are we here at all, it's actually just to save time and money to make it easier for more companies to operate. And that's what this stuff adds up to.

It's really nice when you don't have to update your mission statement. I know as new technology. I mean we've talked about Google's original mission statement statement which is loosely like you know, organize and, and make the world's information like valuable. Right. And it's like okay, well generative AI like does that very efficiently. Agentic workflows applied to finance saves people time and money. It's the same, same mission.

I want to take a post about a post from Keith Raboy. He said five out of six board meetings he had in the last couple months were about how to design a company's organization for AI. I'm interested in and we've been debating this in the context of meta super intelligence. Is artificial intelligence an app like Facebook, Instagram, WhatsApp or is it like their finance team, like their infrastructure team, like their database team that sits across all of the apps as like a service layer. And I'm interested to know like you have a product now and I imagine that there is a team that is working on building that product and that product has AI in the name. But then I imagine that every single person from designers to engineers to you know, the ops people are using artificial intelligence in their day to day. And I'm wondering if you're seeing a new like cross functional team develop or how you're thinking through that trade off of like horizontal, vertical versus vertical, maybe both teams in just terms of running a high performance organization.

I love his post and I think it's actually like I'll put it this way, I don't think most people have really come to terms with the reality that computers can see and hear and think now.

It's crazy.

It's so different and if you're just going and repeating the same organizational structure that was implied place when anything were possible. Like it's a little bit of like are you, are you awake? Like, like this actually is the right question and one I think people should be spending a lot more time in, you know, in earnest.

We need more business, we need some new business, you know, management textbooks. It's true for sure. It It's a joke, but there's no, there's no, like, enduring. Yeah.

So if you don't, if you have a whole piece of your org chart that doesn't have any AI on it, like you're going to ship that in the product and then people are gonna be like, wait, this part, this piece of the product feels very manual and like web 1.0 almost.

Yeah, the. I mean, one. One thing that happened yesterday was Zuck saying, like, basically saying you can now use CO AI in your coding interviews, which sounds insane. Like, wouldn't you want to still hire for people that. That might just, just naturally be good at engineering and then get better? But he's basically making the statement of, no, we're so committed to this new paradigm that, yeah, you should use every tool you have available to do the best possible work.

I think it's totally right. And I'll admit something, even for people who are interviewing soon, there are some job functions at Ramp where it is actually not possible to pass an interview or coding interview without using an LLM. And that's intentional. It's actually intentional to figure out, are you curious about that? Are you using these tools or not? It's not a question of iq. I think there's brilliant people who, you know, don't need an LLM to think. But I think that the capabilities of being able to access an extraordinary amount of compute as well as specialized, you know, knowledge. Why wouldn't you?

And yeah, it's like, just because Scott Wu could do a problem in his head doesn't mean he shouldn't use a calculator.

Yeah, yeah, yeah, yeah, it is. One of the old, like, coding challenges was this thing called Fizzbuzz where you try and write out numbers from 0 to 1, and if it's, if it's odd or if it's divisible by three, you write Fizz. If it's divisible by seven, you write Buzz. And then if it's visible by both, you write Fizzbuzz. And it's like a very basic coding challenge, but it always assumes that you have access to Python. And it's not saying, okay, you have to go and like, fab your own semiconductor to do it. And so it makes sense that, like, you know, we are in the age of LLMs. Like, that's going to be a tool.

That don't give Eric ideas. Kareem might start requiring the engineering team. You're going to have to be able to set up your own semi fab in order to write it out in.

Binary with a pen and paper engineering.

With your own silicon.

Yeah. One interesting thing we've talked a lot about this move 37 concept, Lisa doll, the Google DeepMind challenge, they beat Lisa doll and in move 37 it was like this moment of true inspiration where AlphaGo put a piece on the board that was so unexpected, Lisa Doll stood up, went outside, smoked a circle cigarette and was like, had his mind blown. Right. And we've had moments like that. Maybe the Studio Ghibli moment is this like we've had passing the touring test. But there's also been plenty of areas that feel more intractable in terms of artificial intelligence. And it feels like if you are like with the latest IMO problem 6 was this combinatorics problem that I don't understand but neither does OpenAI or Google apparently. So you know, we're basically the same. But my, but my question is like, is like with the AI Agents product, are there, are there things that you see CFOs, modern CFOs do, like what is the humanities last exam, the intractable problem, the most elegant, beautiful thing that a CFO can do that feels like, oh, this is going to be something that's happening and requires like the human touch for a very, very long time.

I think my immediate thought, give you, give you some time to think is it feels like AI today is really good at doing the work that like a management consultant could do. So you can like tell them like I run this type of business, like make me a marketing plan or come up with five ideas for marketing strategies in Q4 and it will create like a great report, it'll look great, it'll have some, maybe some decent ideas and yet it still feels like to do truly exceptional work. Like we like Monday, like the three of us spent like a couple hours talking about some, some campaign and like debating it and all these different points and it's like we could have had AI in the mix there but you could imagine to be tossing in ideas that, that were kind of maybe random and maybe something helpful. But it was still like you just like people needed to just like sit there and just like hash it out. And it feels like that's, that's like you know, do the work to like do analysis, create reports when it comes to actual knowledge work. But anyways, I don't know what your thought is.

I love this line of questioning and I think it's going to get super interesting, especially as AI is going from a primarily academic tool. And also in some ways I feel like it's jumped into software engineering in part because software engineers know what it is to code and they're working with AI and so, so they have good taste and there's these great feedback loops and so it's happening there sooner but it's starting to make it into general business, logic, finance, everything. I think a good place to start if you just want to decompose. The problem is like, what is the job of finance actually? What is it? And I think if you were to just really simplifying it down, it's in my view some subset of you're going to define a set of rules that govern who is able to spend what, under what circumstances, what are the controls you have in place. Once activity has occurred, you spent money, things are coming in. How do you record that activity accurately, close your books and then based on what occurred, how do you goal, seek to more profitability, more growth, more cash flow, whatever thing you're going to go try to go do. Yeah, that's about it. And if you really kind of simplify it and it's just a set of rules and algorithms, these systems, one, they're incredibly complicated to do each of those things. And so I actually think like brilliant people have over the last 50, 100 years been reduced to like, I don't know, in some cases like tabulators just calculating things or like downloading the same spreadsheet over and over or matching the receipts and doing like very low level tasks and only doing a little bit, you know, after you finally look.

A great example is like when any, any like, like TV show about business. Yeah, it's like, like you look at like the office, all of this. Was any real work done? Like you didn't, was there ever? You don't, you don't really. Do you remember a time when they were maybe once or twice they called a customer or they did a customer meeting, but outside of that it's like thousands of hours of just like, yeah, doing. They're doing business, but they're not. Like, are they really getting anything done?

And they.

So I think, I think the point you're making is like how much of the work that we do is, should, should be the work that we're doing.

You nailed it. Like, I think so much of work is very mechanical and it's, you know, we want finance to be forward thinking and strategic, but the reality of a lot of finance is looking backward.

I think that's, yeah, I think that the strategic thing is really key. Like, like Tyler Cowen would say that like Relationships are very lindy and very going to stick around for a long time, trust building, that type of stuff. But Ben Thompson dumped some sort of, you know, 10Q into ChatGPT and said like make me a Ben Thompson style article. And it did some of it, but it didn't have. Yeah, yeah, it didn't have like the actual strategic insights and the storytelling that can actually like marshal the proper resources. And I feel like that's in many ways the Jabba cfo. It's risk taking, it's deciding the right strategy for the situation and it seems like the frontier models are not quite there yet.

You nailed it. And I think that when I think about the next set of years with the fundraise we tried to lay out of like we're seeing glimmers of like this whole industry is going to get rewritten. What do we think the industry looks like over the next three to five years? What is the org like look like? And we laid it out in the post. But I think the biggest transformation is like really, really. I think most of the finance function today unfortunately is very backward looking. But I think as automation agentic workflows are coming to the forefront and platforms like Ramp actually more people can be looking forward, can be spending time, you know, on like for Ben Thompson like what are the really interesting breakthrough theories and how do you go and make things better? And when I think of like what is that last exam ultimately for a finance organization is, you know, I think there are companies that have taken these tests. So I think one of the interesting studies I think about is Amazon in its early days there. They were very early and they would send email follow ups to customers. You just bought this. And the old way of doing email marketing to customers was someone would write like it's if you've seen like the Trader Joe's newsletter where someone writes like these poetic things about the new ingredients and cool stuff, come try this. And whatever. Amazon did that in the 90s. There was an editorial team talking about the great new products that were coming online on Amazon.com and then there was this personalization team. It's like I don't know, P12N or whatever it is. And they said we're going to a b test it. This personalization team. You're going to get 10% of the send volume and your job is to just go and try to find these weird connections. These move 37s of like you just bought this. You may be interested in this product. And for a long time humans crushed the personalization team. There was something kind of in the example of how these marketers thought that was better get better and better and better till one day the personalization team, the volume started winning and it kept winning. And there is no team that is sending like newsletters around. And so, you know, it turns out algorithmically, on single products, you might be able to go and have a better algorithmic recommendations. I still think for quite a while there was a role for creative CFOs. And thinking about these unexpected strategies, it's sort of, in some ways, like, I don't know, make it about you guys for a second. The media industry has been at it for a long time. You know, there's live streaming, you know, TV has been around for, for a while. No one saw this. And suddenly, like, this is one of the most, I think, important things happening, you know, in, in broadcasting. It's your show.

Thank you.

You guys saw it.

Yeah.

And I think there's room for people of great taste who are great marketers, who are great leaders, CFOs. And, and so no, I'm, I'm, I'm bullish actually. When you have time to think about the interesting work.

100%. Yeah. Like, there's one world where you see the expense policy is like a, a set of if statements. And there's a different world where you start thinking about expenses as like bets and risk taking and actual like what is the net present value of buying that particular thing? Will this pay off for the business? Does this actually increase shareholder value when you buy even just those pencils versus those pens? Like, that's really also.

I mean, I mean, yeah, there's so much, you know, finance historically is, you know, looking backwards and you can make decisions based on that. Be like, well, we spent this on this. We spent. We shouldn't do that again. But there's also so much you can bring in of like even helping understand, projecting out. Did you know, if you adopted this tool and then you grow headcount 20% or 30% or 100%, you know, this is going to be in a year. If you keep this existing, you know, plan contract, you're going to be spending, you know, some crazy multiple of what you're paying today. So. And there's so much.

So what's next for the company? Is the job finished?

Yeah, everybody, everybody in the chat's asking, job finished?

Is the job finished?

The job is not finished. Guys, I'm still.

We're still.

We can't believe it. There's so much to do.

Let's go Job not finished. Confirmed.

No, guys, this, this is, I love this thing. No, it's like, it's like we're, we're pleased, we're happy. But like, this is the crazy thing about our industry is that any business that spends money could be a ramp customer. And we've come a long way.

They will, they will be, they must.

Be, they must be, shoulds and must. I know we're working hard on it, but guys, most, most people, you know, pour one out for our friends, like expenses, closing the books. It is the worst hour of their month. It doesn't have to be. We're going to change that.

Fantastic.

Incredible.

Well, thanks so much for stopping by.

Thanks a ton. A ton of fun.

See you later.

We'll see you later.

And please recommend your tailor to me. I got to get one of these things.

Well, we'll put you in touch. We'll come to the fitting. Awesome. Great to see you, Eric.

We got to hop on with Avi Schiffman in a few minutes. Let's read through some posts though. Interesting fact about the IPO market.

We needed a polymarket on whether Eric would say jobs not finished. Although it probably would have been sitting at 100%.

Jeff Richards says the last five technology IPOs are up 142% on average. The last 20 are up 93. 3%. Performing well beyond lockup windows last 12 months. Prevailing narrative, there is no IPL market. Smart companies prepared anyhow and we're ready to take advantage of favorable conditions. And it's Chime, Voyager, Voyager, Circle Mountain Hinge Health, Etoro, Core Weave. Tons of companies on here Service Titan all got out and are sitting at pretty solid offer to current percent gains. There's a few that are down, but by and large the IPO window is wide open.

It's wide open. And on that note, there was some news yesterday that an ultra massive asteroid is on a possible collision course with the moon made to trigger destructive meteor shower on Earth. And Andy from 2 cents says it's in 2032. Enough time for an IPO lock in. So really quickly, let me tell you.

About numeral hq.com sales tax on autopilot. Spend less than 5 minutes per month on sales tax compliance and we can give Avi Schiffman a call.

Give him a ring.

Let's give him a ring. And this.

We haven't done many of these before. These sort of FaceTime call ins, but we're gonna give it a shot.

I'm just gonna call him on the phone. I'm not even FaceTiming it. How you doing my man? Hear us.

Happy launch day.

Splendid.

I can hear you.

Happy.

Very, very bold to to launch on the day of a ramp fundraise. But you're breaking through.

Congratulations.

Seen a lot of posts.

Thank you.

You've been following the story for a long time. Very excited that you got these out into the world. Break it down for us. Are you sharing any numbers or has there been any feedback? I saw a bunch of posts that you were reposting. What's the feedback been so far? What did you actually ship? Is this the final product? This developer preview? Give us kind of the high level overview and then we'll dig in.

Sure.

Yeah. Am I live right now?

You're live.

You're live.

Oh, dope. Yeah, actually, I mean just one, one minute ago, right before you called a friend of mine that I sent it to, he's. He works for Arc boats and everyone there loves it so much they just called me and just. They just ordered a whole bunch of them.

So there we go. You're selling enterprise software. Yeah. Yeah.

I was gonna ask are people abusing this yet to. To do work? Instead of. They're like, yeah, he's pitching me a friend, but really I just want an intern.

I just want to sell stuff.

Yeah, I think some people wanted to like help them with productivity and like that's kind of the con of talking about it on Twitter. But yeah, I have no idea.

I.

The cool thing about friend is I can't read people's chats because I don't have their friend. It's kind of like a, like a living Yugi key in that way.

Interesting. Interesting. Is it. Is it on device or just like end to end encrypted or secure somehow or.

I mean it's kind of just. I look forward to seeing people try and crack it but basically it's basically encrypted and sorts on the dev. There's like a key in each device that like I don't have access to.

Got it.

Because I don't have the actual friend. But anyways.

How much do people pay for these?

Some people are $129.

What you say?

I was asking how much people paid.

Oh, 99 for the pre orders but now they're for sale for 129. Dude, I would never ship a developer.

These are.

This is the final ready to go. Ready for. Ready for the daytime. Yeah. What's next? Do you immediately go into getting feedback v2 software hardware? What are you excited about improving? I'm sure you're getting a lot of feedback already.

Yeah.

No, no V2s anytime soon. I think the product is the best I could possibly do with the current state of tech. The real marketing push will come sometime soon when our next movie drops, which will be a proper feature film that we're submitting to film festivals.

No way. You're in a whole movie.

Okay. This is a new meta.

Yeah.

The cinematic video is dead minutes.

Yeah. This is why I didn't want to.

Drop another video today. Like, that's boring.

That's amazing. The meta moves. But the. The.

The new feature film, I promise you, will be the greatest cultural technology thing of the 2000s. I think it's going to be amazing.

Bold statement.

Can you tell us anything about who's attached? I know. Was it Sandwich Video that worked on the previous film?

If you had sandwiched you a feature film, it would cost like $50 million.

I'd be surprised if you guys knew them. But, you know, All Gas, no Brakes are channel five.

I know, of course, Callahan, deeply. Yes.

He interviewed the former president.

They split off a while ago. All the sex scandals. But. But the production crew behind All Gas no Bricks has been following the whole project.

Very cool.

And it's. It's going to be kind of like that Kanye documentary from a couple of years ago.

Okay. Yeah, it's gonna be.

It's gonna be great. When's is it? The August. September. When are we. When are we talking?

I have no idea. Probably not this year because I want to submit it to film festivals first and win them. The closest one is the Berlin Film Festival and the awards for that will come out in February.

Do you think the price of, you know how the price of like a cinematic two minute, like intro video has dropped a lot because, like, you know, cinema cameras are getting cheaper. Sony has the FX3. Like, you know, you can just use free editing software, AI tools. Like, do you think the price of making a feature film has dropped such that, like, more companies will be doing this in the future?

I mean, probably, but I think like, ultimately you just got to like, care enough to make something good. I mean, for the Friend launch video from last year, I location scouted it, I casted it, I helped write the script. I was there on set. Like, you just have to be involved.

Yeah.

And I guess be. I don't know, just be good enough to make something go. I thought the Waves video was probably the best second to Friends video I've seen.

Which one was the Waves?

Oh, yeah, like the glasses.

It was a nice video. It Wasn't just like, you know, flat. It wasn't annoying.

I completely agree. Yeah, it was great. What. Take me on a tour of the other hardware devices. It sounds like you're a bit of a fan of waves. What else is going on in the world that's interesting. What are you. What do you think's a dead end on the tech tree?

Well, I am still disappointed that there's not really any interesting consumer stuff. I still find polymarket to be pretty much the only like genuine mainstream consumer thing of the last billion years. That's, that's worthwhile.

Yeah, that's real.

Yeah. You're partnered with them?

Yeah, we're partnered with them.

Novel.

It's, it's, it's cultural. It's cultural. It's cool. Shame is cool. I mean, it's, it's, it's new. It's, it's what I want to do.

It's just so hard to break through a consumer because it's either like it's hardware and it's a billion dollars to do a big run or, or it's like, oh, it's anything tied to social networking and like there's a super durable monopoly. I mean, I guess ChatGPT broke through in the consumer context really, really massively. But you know, it instantly just turned into like a Mag 7 company.

They're doomed.

I feel like, okay, for starter. I really. I'd be very surprised if the whole Johnny I've thing actually ships.

Okay.

I just, I don't think that's gonna ship. That's maybe my take.

Not sure.

I mean, it's just there, they're just. Look, they're just too old to understand how to go towards, let's say the new. Like the main target audience that I'm very interested in is Gen A because they don't say the word AI companion. You know, I named it Friend because they just say they're talking to their friends. They know how to intuitively talk to an AI, but also view it as like a platonic companion. It doesn't have to be a sex bot or an assistant, which I think is really only old people that need some kind of utilitarian value from.

What about, what about GPT psychosis? Is it real? Is it a risk for the friend? Is it a risk. You worried about it? You worried about someone getting 7 million comps deep in French?

If you truly fall to that? I mean, I think that's your problem.

Well, I would hope that you at least build some guardrails in just to be safe. Because, you know, why not?

Yeah, just make sure the friends check on.

We're using Gemini 2.5, which I think is a great model and Google does a good job allowing it to human person.

Do me a favor and remove those. What is it? It's not SCP Horror copy Creepypastas. Just remove those from the training set, please.

Sorry for that guy.

Yeah, yeah. Anyway.

Yeah, well, we'll see. You know, the friend used to be a little bit more psychotic. When the battery died, your friend actually was truly dead.

That's wild Tamagotchi style. What about. Do you ever play Fallout?

Sure.

Yeah.

What about a pip? Boy, is that a good idea. I was pitching that to people. If you're on Shark Tank, would you invest in that if you had, if you were shark.

I. I do think that like, like single feature consumer electronics will be popular again. Like, I don't understand why everyone's got this boner of trying to build the iPhone up AI when the iPhone ever came out without the ipod. You know, I tried to just do one single feature.

Totally.

You know, do it well end to end.

Yep.

I think there'll be more, more products like that. People want like interesting new form factors and to kind of disperse from their phone a little bit. They're not replace it.

People.

People want to scroll their TikTok. So that's just not going away anytime soon.

Yeah, yeah, the phone's kind of end.

Of heat is forever.

People are going to be on their phones. But then, you know, Apple Watch sort of broke through VR headsets.

The Apple Watch did not break through. It's the most dorky product of all time. I would never even want to watch.

I see. I would say. I would say probably half. Half our guests wear. Wear an Apple Watch.

So you're like, you're like $100 billion in revenue a year. Didn't break through. It's a failure. Wind it down.

AirPods. AirPods.

I hear you. You. I hear you. And I agree with like, the aesthetic criticism, but certainly like the idea of like a computer on your wrist is like somewhat popping.

Sounds like you're wearing one right now.

I'm not.

I'm not.

I am a proud, proud Brad ambassador for Bezel. I appreciate mechanical watches.

I appreciate Holy Trinity.

Appreciate on Constantin. They would never think to put a battery in that thing.

Yeah, I mean, personally, I don't want to be tied to time like that.

But sure, sure shots for fired at the watch community.

Wait, so you don't want. Okay, no, yeah. No time whatsoever.

What is. What does Gen Alpha. Think about AirPods. Are they. Are they always in or is it. Is it. Or AirPods?

I think some people doing interviews with them in and they're like, yeah, they're on transparency mode and it's like. That's the crazy.

I don't know, I don't even wear headphones really. It's just like, just got a good speaker system.

What about like analog cameras? You carry like a film camera or like a Q3, something like that?

Sure, dude. I mean I just bought this same camcorder they used in 28 days later, that zombie film.

No way.

Highly recommend it.

Yeah, yeah. We, we brought a VHS camcorder with us or mini DV on this trip to do a little.

You think.

It'S exactly the mini.

Mini dv.

Yeah, yeah, yeah, yeah.

What's your read? What do you think Zuck is going to do in the AI companion space?

Good question.

AI friendship.

I mean I think it'll just make the concept like you don't buy a dog because you have the low self esteem and I don't think AI companionship, which right now people think is just a product for like lonely people or people with low self esteem, etc. I think it will just be like a new kind of companion that people will have in their lives.

Makes sense. Anything else, Jordy?

No, this is great. Congrats on shipping. I believe you're sending. I don't know if you already sent.

One friend dot com.

I don't know if you already sent one to us, but happy to, to purchase and support and we'll, we'll bring it into a show. We'll bring it into a show. It's gonna be a high stakes, high stakes, high stakes demo. You're gonna. The, the friend will be with us live for, for three hours, so.

Okay, well you might have to wait.

A couple years then.

No, I mean it's okay for something I got you.

We'll ship it out to you this month.

Yeah, that'd be awesome. Yeah. Well, we, we do a bunch of product demos with our intern Tyler as well. We can demo it in a smaller setting. A segment on the show maybe.

That would be fun. I hope you enjoy the, the whole thing. I think people, people yearn for opinionated hardware again and I think there's not a single user facing detail that at least I did not choose.

I love it.

If you like it or don't like it.

I love it.

This is something I chose.

Yeah. I'm so sick of Christmas and the holidays coming around and it's like, I can't get you music because you already have Spotify and you don't want a CD anymore. I can't get you a book because everyone has audible. I can't get you a DVD set because everyone has itunes. And like, just the gift. I feel like the gift economy has been nerfed by the proliferation of the Internet and you're like kind of bringing it back and it's like, even if somebody wants an Apple watch, like, they'll probably just buy that themselves. It's pretty rare to like make sure a good gift. Whereas this is something that's unexpected just by virtue of the fact that you're a new company. And so I feel like if I gifted this to a few people, it would be a surprise and delight. Maybe they don't love it forever, but it's something that's new and it's different and it's fresh and I love that.

Well, maybe for Christmas time also, John. I mean, I don't even have Internet at my house. I just listen to finals and.

Okay, yeah, but yeah, you have to admit you're in the minority there.

I don't know. I think I'm starting a new wave.

Okay, well, expect, expect all 600 hours of TVPN pressed on vinyl under your Christmas tree in December 25th.

You should actually do that. That would be great.

Expect it. If you have a, if you have a record player, you're going to be listening to our sultry reading. Ben Thompson on vinyl. Be great. Anyway, okay, we'll talk to you soon, Alvi. Congrats.

Thanks.

Bye. Thank you.

Anyway, opinionated young man in tech.

We love it.

Always a fun conversation.

We love it. Also get on FIN AI. FIN AI, the number one AI agent for customer service. The best performing AI agent for customer service. Delivering higher quality customer experiences and higher resolution agents.

Bake off. Champion of the world.

And there was an interesting Australian rocket launch that happened.

It was very interesting.

It did not go very well. The rocket Eris, launched by Gilmour Space Technologies, was designed to carry small satellites to orbit. It goes up, it kind of slides. It went for 14 seconds and it was, it was a little bit of a crazy slide to the side, to the left.

Sad to see. Sad to see, but you have to. You gotta fail. Apparently in aerospace, in rocket science, you got to fail a lot to. Even Elon is failing.

We were watching this video with somebody and was like, was anyone on board? Obviously not. We send satellites to space. We don't send humans to space.

Yeah, I don't Think we should laugh at the misfortune or the failure. But there was a very viral and funny post where someone was saying, you know, imagining the Australians saying oh no.

Very funny. But yeah, I mean, I mean it is, it is very fortunate that way. We're in the era of like, you know, autonomous rockets from day one and so you can blow them up a lot.

Rockets are tough too. It's hard to hide a catastrophic failure. You know, you can launch your, your AI, you can launch your agentic workflow and it can fail. But when you launch a rocket, people are going to find out.

Oh, this is interesting. Avi mentioned Polymarket. There's a, there's a trader on Polymarket who's up $750,000 and just can't seem to lose. They have $2.4 million on rates going unchanged today. They'll profit another hundred thousand dollars if they're right. Better trader than Pelosi. And wow, they really went all in.

And they were right.

And they were right. So they made another 100k. What, what an impressive trader. Crazy. Yeah. It feels like a lot of the people in Poly Market have like scaled up to being like, it's more like a hedge fund and it's like their professional job. Like the famous example of the person who called the Trump thing. Right. It was like he was paying for his own polling data just to, just to make that bet.

In other news, Palo Alto Network acquired Cyber ARC for 25 billion. Palo Alto, good name for a company.

Palo Alto Networks. Sanford Road company. Probably didn't lose money.

Great name. My family moved. My dad's family moved to Palo Alto I think in 1906. Yeah, so way back. But they bought Cyber ARC for $25 billion. Cyber ARC is an Israeli identity security provider and cyber. Sorry. Palo Alto NETWORKS is down 6.24 today, so never know with these things. Sure.

In other acquisition news, Sam Ross shares that Numerl has acquired SpendRuby as part of this acquisition. We're excited to welcome the entire rubycard team to numerl. This marks our first acquisition and it's been a long time coming. They first met during y Combinator's winter 2023 batch. They became fast friends given our shared experience building and supporting e commerce brands. Turned out that I wasn't the only one who had dealt with the nightmare that is sales tax. The team will join the engineering team to help accelerate our global offering including integrations and industry leading tax engine. Together we'll solve one of the hardest problems for growing businesses. Sales tax compliance. I agree. I've experienced myself and I use numeral for my E Commerce company. Joe Weisenthal was talking about gdp. Also talking about the odds of the Fed rate cut, which we already covered. The odds have dropped substantially, as they probably should given that the American consumer is undefeated and the American economy continues to be on a tear. Speaking of Apple products, Freya Lobo says, one of my longest standing hobbies is thinking about whether I should get an iPad. It's not about the iPad. It's about imagining an alternative life I might live as someone who has an iPad. Love to wonder about it.

52K apparently the struck a chord. This is apparently many people dream of an iPad life.

Yeah. Yeah. When the latest iOS came out and everyone was like, okay, the liquid glass is a little crazy. We don't know if we like this. It's cooking Tyler's phone. It's not maybe the best update yet. They'll get it there but it's amazing on iPad. And I was like I want an iPad.

I want it.

But then I actually thought about my day to day of doing the show and writing and screenshot stuff and I was like, I think I would hate it.

Yeah. My dad's retired and when I see him use an iPad, I go, what a simple, what a simple but nice life. You know, if you can just leisurely tap around, maybe read an email, fire off a thumbs up emoji, you don't have to do much.

Well, whether you use an iPad or an iPhone or a Mac or some other device to access the Internet, you got to go to public.com investing for those take it seriously. Another question about Yu Chen Jin says.

Not a single person from Thinky or Anthropic took Zuck's offer. Why? Some good reasons here, some of which we covered already. Rich mission greater than mansion. They've seen AGI hate Big Corp life. Upside is higher, value of money is diminishing. Post AGI no faith in Meta or Zuck and people are voting def not.

6 because a Lambo is a Lambo. Pre or post AGI fact check true.

Experts confirm this is true. Kosh over on the Ramp team. I don't know if this was intentional. Ramp was valued at 22 and a half billion 225. 225. And he says bench 225 a little easier this morning. Feeling absolutely jacked. So shout out to numerology is really good. Mike Isaac says, pretty clever. Someone who claims to have scraped public listening data from a number of public figures, politicians Celebrities, journalists spun up their alleged playlist and made it into a site. Thankfully, mine isn't too embarrassing. But others and people were talking about Palmer Lucky's.

He listens to all the hits. He likes pop. And he was like, this is mine, and I'm proud of it. And everyone's like, this is amazing. I got to listen to Palmer's list. But PanamaPlaylist.com, if you want to check it out. Also, like, sort of an odd data leak. Like, I. Like this was this. This happened during the election. The Venmo transactions were public, and so they'd find, like, J.D. vance's Venmo, and then they'd see, like, J.D. vance paid someone else. So they're connected. So if you're out there, and this is funny.

So are you on your stuff? So you're on this list. You're on the Panama playlist. You're. There's a playlist here from you called Immortal.

Oh, yeah.

And it's the Four Seasons.

Yes.

Concerto.

Yes.

Hero, A Family of the Year. Jesus Walks. I'm on. I'm on the list as well. I don't even know what this playlist is.

I'm proud of my playlist. You're free to go listen to them if you'd like. I should just make them all public. I love.

Let's See, Keith Raboy, Deadmau5.

A lot of good EDM lesso Spotify playlist. It is funny because I like. I didn't expect it to be public, but obviously I don't.

Sean. Sean McGuire's listening to some Apex twin. Muramasa.

Yeah.

I don't know. This is so random. I don't use public play.

I mean, I feel like we got to go.

Dalian is listening to some Credence Clear Water.

That's cool. I like that.

Gary Tan is listening to Crumb.

Okay.

It's an interesting list here.

You got to. We got to figure out to really curate that so when people land on it.

We're working on a playlist for a little something tomorrow.

Okay, cool. We'll have to put the Four Seasons on there.

That's right.

Aldi. One of the greatest.

Something like that.

Speaking of tomorrow, inspired by oomph. Are you the 2003 Dom Perignon where the heat wave killed 5,600 people and made the best graves of all time? Have you heard this lore about Dom Perignon? Apparently there was a crazy heat wave in 2003, and now it's a. It's a beloved vintage of wine. I had no idea.

I Did not know that.

I need to, I need to learn more about that. Anyway, Percy Jackson was right. SF is the new Rome, says Lucas. Both SF and Rome have exactly seven hills. Temperature, temperate, sunny sea climate. Status as a pilgrimage site where immigrants working to become citizens center of a productive valley. See military, industrial, business, power. And look at this photo. I, I, it's not going to translate on the screen, but this, I had, I had no idea that the palace of Fine Arts could be shot with the, with the, Is this one better? I don't know. I'll hold it up. I don't know. We're there. Go to four. Yeah, camera four. There we go.

Put it for.

Ah, I'm fighting the air conditioner conditioning, but the palace of Fine Arts. Incredible image. Incredible image. And it does look very Roman. So.

San Francisco.

San Francisco. Undefeated.

Undefeated.

Oh, Luke Ferretor. So there was a piece in Bloomberg about Luke Ferrator.

Lots of debate. Was it a hit piece? Analysis, yeah.

Profile. We're going back and forth. Spore said the Luke V. Farador article is hardly a headpiece. And some of y' all are wildly overreacting. We have some of the overreactions here. We got Katherine Boyle saying, when I worked for the Washington Post, there was an understanding that the family of elected officials was generally off limits unless they put themselves in the arena. A hit piece on a low level government employee that goes after his mother by name is so appalling. Why would anyone in tech talk to Bloomberg after this? Sean McGuire says, Luke Verator is an American hero. He says it's a hit piece. This is a hit piece is a badge of honor. Par for the course. Luke ver is still 23. We got to get him out of the government. We got to get him into Silicon.

Valley, we got to get him into sas.

Term sheet. We got to get him a term sheet. That's the key.

I mean, think about how many, think about how many people have probably approached him and said, I will give you $50 million to start a company and you can immediately take 5 million of secondary. Like I'm sure, I'm sure some, some offer in that range has been made, but it is worthwhile having him in the government working on important problems. I think you had it, you had a framing which was, you can look at some of the cuts that Doge has been making and say if you believe that the government is efficient and that the money is being spent well, then you're going to be really upset. If you believe that there's a lot of potentially a lot of waste, and that there's money just going to people that shouldn't be getting it or going to causes that might be harmful, then maybe it's positive. And so it's naturally like an incredibly politically charged topic.

It's completely reasonable. If someone says to you, America is spending $100 million to help the less fortunate, you'd be like, well, I don't want to take that away. Of course it feels good to help the less fortunate. But if you dig in and you find out that it's like, well, that's what it said, but in fact, it was all going into someone's pocket who's super rich, you'd be like, wait, why are we doing this? This, this makes no sense. And so the, the, the stated names of every program are always with the best, with the best intentions. But, I mean, you know this from running a business. You, you, you spend money on a lot of different projects. Some of the projects go off the rails and you need to cut them. And so the idea that, like, cutting spending is super controversial. This is something that was done in the Obama administration, like, it should not be as politically charged as it is, but unfortunately, we are in a very, very politically charged environment, so. Casey Hanmer was particularly upset about this article. He said, disappointed but not surprised by the tone of the article. The authors tried desperately to smear Luke Fairtor with their poorly engineered invective, but instead only showed their limitation of their vision and the mediocre cowardice of their sources who couldn't even air their own uninformed opinions under their own name. Not a single mention of the ballooning deficit, not a single mention of collapsing state capacity implied that USAID was cut because Luke misread some database. And not that Doge found that USAID had illegally ignored an executive order by continuing to pour money into unaccountable foreign NGOs, some of which was illegally coming back into partisan US organizations. It is astonishingly irresponsible to paint a target on Luke like this. A number of sources made claims about his motives, which is defamatory. An article like this brings dishonor upon the profession. The other, the other side of this is that, is that I think this isn't Luke's first rodeo. Even though he's only 23, I think he is holding up well. And I'm. And I'm. And I think he is. He's. He's. He knows that this is, this is par for the course.

And so, yeah, and it's, it's making a large Number of people think that he might be bad or.

Oh, sorry.

Well, no, like, you know, this could be an example where. Where a million people think that he's bad, he should be removed.

And then.

And then 50,000 people.

Yeah, they're crazy.

Think that, like, I will support this person throughout my entire life.

All it takes is, like, one crazy person to be really, really destructive. So rough. Anyway, Spore has the other side of the argument. Says the Luke Farador article is hardly a hit piece. Some of y' all are wildly overreacting. You don't just. You don't. Just like that a technologist is being scrutinized for working in an extremely important governmental position. Position. Sorry. But the Fourth Estate has no reason to kowtow to Silicon Valley. And I kind of agree with this. Like, you know, I'm a taxpayer. You work for the government, just like I would expect. Like to know where the government funds are going and what, you know, is this particular government program efficient or not? I also would like to know who the people are. And so I do think this is of the public interest. And I think it's okay for reporters to dig into what Doge is doing, because even if you're pro Doge, you want to make sure that it's working effectively. And so that's. So that's reasonable, I think Spore continues and says, I super support the new media strategy of going direct and telling these publications to f up when they try and slander you. But then you don't really get to complain when they're not characterizing you how you want. That's not how it works. For what it's work. For what it's worth, I read the piece. It's a. It is a too long portrait of work, Luke. It is extremely long. It takes over an hour if you listen to the audio version.

Yep.

His history and experience and spends a ton of time on some of his most impressive accomplishments. I agree with that. It is very, very. Yeah.

And they were nice enough to make him look very cool in the picture.

I thought so. Although the. The red coloring does make it look a little like Darth Vader esque, then it gets a bit critical. And yes, it does include some stupid jabs. But again, if you're going to work in the government in a highly public and highly controversial role. Yeah. People are going to look into you. They're going to find things they don't like and talk about it. That's the game. Obviously, the publication does not like Doge. They don't like Elon or Trump. And so yeah, a glowing piece was never going to happen. But if you genuinely read this piece and think they slandered him and dragged his name through the mud, you and I are two different are reading two different articles. Oh, but hey, he's just a kid. Well, one, don't diminish the guy. He's incredibly talented, smart and accomplished. So and also, I don't know how old a person is. If they're operating a key government position, the public has a right to understand who they are. And so I thought that was a good, more nuanced take from a tech insider about the piece. And maximum Max Meyer chimes in with a rebuttal and says his work in government is fair game. The main thrust of the criticism was the bizarre focus on elements like this shared by Mike and the extreme bias in the subtitle suggesting he's kind of some sort of evil sellout. The bias of writers is fair game too. Oh, so yeah, you can criticize the writers too hard to call him a.

Sellout considering he's raised he hasn't raised.

Money for a SaaS company yet. He did sell out, in my opinion. Going to work. Going to work for the government, working for a nonprofit. That's the selling out. You got to stay in the arena, build some SaaS.

Yep.

Anyway, post from Ara Kharazian. No one is talking about this, but the bull case for AI and economic growth. AI spend is up 60% in construction and 35 times in manufacturing. Faster planning and smarter automation will transform.

Let's give it up for big construction and little construction for adopting.

Interesting that they're adopting AI faster and more aggressively than professional services, technology, finance, retail.

This is this is growth rate though. So that doesn't mean they're yes, but doesn't mean they're indexed growth.

Sure.

So it doesn't mean they're they're actually spending more. It just means that they're they went from zero, they were not adopting it at all and now they're adopting it like crazy.

Yeah, I mean I guess they, I guess they all they all kind of started at 100, what, two years ago? Anyway, we should talk to Ara about that more and dig in.

Relevant post Primary Topic at five of six board meetings this week we covered this with Eric, but he Keith says How to Redesign a startup Org in the Age of AI.

Oh, we should close this this extremely important lesson from Peter Thiel. Or maybe we should do the a few more posts. But David Senra said don't divide your attention focusing on one thing, one thing yields increasing returns for each unit of of effort. At a micro level, an extra hour of focus on the current project has a much higher return than an hour on something new. Or worse, five minutes each on 12 new things. Before you ever do something, you should understand the opportunity cost versus existing things. Don't rationalize that something you, you want to do is complementary when it's not. At a macro level, understanding that applied effort has a convex output curve is a very useful discipline when considering new market areas. This convexity, this convexity means that the opportunity cost of transferring resources from existing projects to new ones is high unless the new area is incredibly valuable. Anything we can do to extend an existing convex curve is worth so much more. And there was someone in the comments saying, well, Elon is the, is the, is the counter example to this. And of course, Elon and Peter Thiel were business partners for a number of years and continue to work together in many capacities.

So with that reading, should Timothy Chalamet be partnering with Lucid Motors to promote their new car?

I think so.

I think this looks like a car straight out of Arrakis. So yeah. And think, I think it works. Yeah.

And I think, I think this is very much, this actually builds his brand as a, as an actor. Like, I see this and I see, yes, it's an ad for this car, but it's also an ad for Timothee Chalamet. So he's partnering with Lucid Motors. Bold, discerning, unapologetically original. And, and, and you know, his, his primary thing is point a camera at me. And so this is an extension kind.

Of the main thing.

It is, it is the main thing. And this does not feel, if he was starting a car company, I might be like, okay, well that's probably going to pull you away from the next, really killing it in the next Dune sequel. But I was, but I was debating someone with this about Elon. I think we mentioned this on another show, but people say, oh, Elon has so many different things. He was the CEO of just one company at a time for 13 years straight. So he starts zip too in 1995. He doesn't become the CEO of Tesla until 2008. And at that point he was already the CEO of SpaceX. And so he was effectively like a one CEO person. He did company man different companies back to back because one sold and he couldn't run it anymore. He was ousted. But, but he was basically just focused on one thing for 13 years straight. And so I think a lot of people will say like well Elon can do it. And it's like yeah, yeah. I mean he exited a company for over $1 billion very early in the dot com boom and then you know, also spent 13 years focusing on things. So yeah, maybe if you're at that stage of your career, maybe that makes sense. But be very, very careful.

In other news, Kate Clark has a scoop. Iconic Capital will lead Anthropic's next mega round valuing the OpenAI rival at $170 billion. Hit the gong. John Qia and GIC are also in talks to invest. This is pretty wild. The they're adding $120 billion to their valuation since the last financing Iconic, it was only at 50, sorry, 110. So they were at like somewhere around 60.

That's crazy. That's so it's such a big gap.

And I think Anthropic is expecting to be run rating at around 10 billion by the end of the year. So seems like a reasonable multiple given the growth rate and given they are the 170 ton gorilla in the code gen market right now, this will be.

At the top of the request that SBF sends to the President. Begging for a pardon.

Yeah, you were saying this. So FTX owned 13%.

They effectively led I think their series C or something. They put $500 million into the business.

Off the balance sheet.

13% of the ownership of Anthropic and if they'd held on to that it would be worth 20 or 30 billion dollars. Of course they went through bankruptcy and.

Had to restructure and whoever bought that position is doing very nicely.

Exactly. So a bit of a liquidity crunch going on and a lesson there about don't, don't get over your skis. If you're effectively a bank, maybe set up a separate vehicle for your. For your private market illiquid investing because you could get caught.

That's right.

Anyway, thank you for watching today's stream. We will see you tomorrow live from the New York Stock Exchange.

I cannot wait. I wish we could wear these tomorrow. I think we're gonna, I think it's a little. We've got some other suits for tomorrow but thank you for tuning in. Unique show today. We're figuring out how to do the show on the road. Thank you to our production team. Chad's on the, on the production team. That. What's that? Yeah, yeah, yeah. Ben, Michael Scott. People, people know, they know the names. But anyways, thank you for having us and we will see you tomorrow.

See you tomorrow. Cheers goodbye.