We got Py in the waiting room, let's bring him in. Let's bring him in. I wish he was here in person. He can do the wander. Oh, he got the Ramp on. There we go. Ramp! Hello, brothers! Hello, welcome to the show. Good to have you here.
Oh my God, it's great to be here. Long time.
First of many, first of many, first of many. Uh yeah, break it down, what happened today? Why are you wearing a Ramp hat? Take a victory lap, take a victory lap.
No, I always wear a Ramp hat. I've heard that TVPN is just unbiased journalism, and so I've wanted to join you for an unbiased conversation.
Unbiased. Yeah. Guests and host of this podcast may hold securities. Ramp is the presenting sponsor of the show. Um, anyways, great to have you on, Py. I wanted to go back to March 21st, 2022, when you — this was like probably your second time covering when they announced their $8.1 billion valuation. And was that the back-to-back round, the double round, or whatever? I mean, every single round for a minute there, you basically would read a Ramp fundraising announcement, you'd close The Information, and then you'd get a notification on your phone that said, 'Scoop: Ramp is raising another round.'
Yeah, it was wild. That was the $8 billion — that actually might have been the third piece. I think the first time I wrote about them was December 2020. I think they were — we didn't talk about a valuation, but I think they were valued at $300 million at the time. So like something just pathetically like seed and Series A at that point. That's like a personal net worth type number. And then the next time I read about them was when they had raised back-to-back rounds. I think it was like $1 billion and then $1.6 billion. And I spent a lot of the time in these pieces, one just being like, you know, praising Ramp obviously, and then two, kind of justifying that if you're growing this fast, these valuations kind of end up catching up. I think there was a period in between — they took a down round, they took their medicine in 2023. But now they just kind of continue to grow fast, and we're back way above that $8.1 billion valuation. Yeah, thinking about the Ramp down round — it was directly tied to the market, not business performance, which is why they're back above that last financing. They were taking that down round while still putting up ridiculous numbers. And also, a down round — I mean, Facebook did a down round at one point. But down rounds can mean so many different things. I saw some down rounds during that year that were brutal. Like pay-to-play, total cram down — it's $100 million on a $50 million pre, so the new investors are taking two-thirds of the company. And then there's a 10% option pool for the executive team or something like that that's going to get rolled out. And literally anyone who touched the company for the first half decade that they were running it is just gone, wiped, like didn't exist, never existed. You got paid W2 wages. I hope you put some in Nvidia stock because you got nothing now. That happened all the time and it was terrifying, but you know, it happens.
Yeah, I was on the other side of a few of those. But I remember reading about the Ramp piece, the Ramp down round when it happened.
Yeah, and that's when I really knew. In 2020, 2021, you can be impressed by a company, but a lot of people are growing fast, a lot of valuations are going up. When you read a story about a company doing a down round in this terrible market, and it's pretty much gushing, and all of the commentary around it on Twitter about them doing a down round was gushing commentary on how smart it was for them to take the hit and keep going — that's when you know you got a company that's pretty special on your hands.
Yeah, so you talking about Ramp safety?
Yeah, so this is John's — so yeah, so Py, like obviously Ramp's moving really quickly. We're worried about a future where Ramp is like closing the books so fast that you go to ask Ramp to close the books and it thinks, like, if we didn't have any humans, we would never need to close the books, so it just kills everyone. Ramp AI. Yeah, kind of like a Ramp doomer scenario. Like, should we pause Ramp? I've seen some people throwing out open letters, maybe trying to lobby the government to slow down Ramp. What do you think? What's your take?
I mean, I think it's a real risk. I think pausing doesn't work though, right? Then the Chinese are going to start, you know, the faster. So you can't do that. The thing that I get worried about is that you run out of atoms in the universe at some point. If you're an investor and you're thinking, can they continue to double every year, at some point you run out of atoms in the universe. And so that to me is kind of one of the bear cases, I guess. Yeah, I think Ramp generally, what they're doing is they're building a machine god that lives in your CFO's Chrome tab. But ultimately it will be a benevolent machine god that lives in your CFO's Chrome tab, focused on saving you time and money, and it knows its lane. So in a post-Ramp society, kind of like a post-AGI, post-Ramp society, I think humans will be doing fantastically. I think it'll be great, at least the CFO suite will. They're going to get all their time back to think strategically about the business, about life. So that's one of the things — my dad is a CPA, and so when I write about Ramp, I think about my father. How many basketball games he could have been at of mine had Ramp existed at the time. The guy was a business consultant, he was traveling all the time. He'd fly to Germany for the day and fly back. And imagine, Ramp's AI just does that for him, and we're hanging out, he knows me better, I know him better.
And it just suddenly gets super dark. He's like, 'My dad never made it to games in the first 15 years.'
Actually, he did. It was pretty incredible. He would fly back, take the early flight back to make it to my stupid game. So good on him. I just mean he'd be a lot less tired had Ramp existed, you know.
I'll give a personal anecdote. I got to hang out with my kids this weekend because we're on Ramp and the books close themselves. Pivoting a little bit — John and I were talking this morning. There's too many top signals to count right now. Everywhere you look, you got Trumpcoin, you got Cardano going in the reserve, you got OnlyFans competitors doing something we don't know. And we got humanoid robot companies with one customer raising at $40 billion. What is BMW's valuation, by the way?
Yeah, that's a good question. Oh my God, if it's less than $40 billion... 53 billion euros? Okay, so what's that in USD? We're mogging the Euro these days. But yeah, it's probably $50 billion. So yeah, there's definitely a bull case to be made of why Figure should be worth as much as their customer, their first customer.
But anyways, how do you think as an investor right now — we're clearly living through one of the greatest times in history. We just had a private spacecraft land on the moon, Elon's launching rockets every day, there's austerity in the government which is potentially bearish for GDP short-term but probably good for the dollar long-term. How do you think as an investor and even a writer and somebody who's called a lot of trends early — what do you think the next five to ten years?
Oh, the next five to ten years. What I've found really fascinating is that a lot of the stuff that I'm not investing in has been super overvalued, and then a lot of the stuff that I've been investing in has been undervalued and underappreciated. And I think it's going to be worth a lot more in the next year or two. So that's kind of my take. I'm not a Figure investor. Big fan of their videos, and actually my daughter loves Figure videos and was really excited to see two of them together. I don't know. I think there's a lot of probably overhype in AI. There's probably going to be a couple of $10 trillion AI companies that are going to pay for the whole thing. Humanoids scare the hell out of me as an investor because it's this really impossibly hard thing to do with a lot of really smart people working on it. If you pull off this impossible thing, then you're competing against these 10 other companies. Like, who wins? Big consumer surplus event. But there's human labor doing things. There are some people I've heard who don't just podcast and write who do physical labor. Human salary and wages — it's like $45 trillion, it's like half of the world's GDP. If you underwrite robots as some small percentage — I think Figure already got credit for 0.1% of that GDP, which is pretty amazing. But there will be massive, massive, massive companies built here, and I think there's also going to be a lot of people who lose a lot of money. We could talk about what's happening in venture. I think the big funds are actually super well positioned. There's going to be a lot of these emerging manager funds who try to compete who I just think get kind of wiped out.
Do you have a take on General Catalyst going public?
That's a good one. I don't have a good take on that.
I mean, it's really — one of my favorite pieces of yours over the year, I'm not going to know the title, but you had a really great deep dive on this idea that as our tools become better and more efficient, a company that would have taken 100 people to build five years ago now only takes 20. And I think we're really seeing that right now in AI, where you have companies like Cursor — a big one — with 20 people doing $100 million of ARR. It's been amazing to see that play out. Take a victory lap there, maybe double-click in and share some more insight. But I feel like you basically called that and did it in a very eloquent way. But I expected it to be something like maybe five years from now that would be the case, but it feels like it's that way today.
So there's a few different ways that goes. There's the Russell conjugation where the person who's doing something different than you is doing it bad and you're doing it right. And I literally think about this when I see all the charts of these companies making more money faster than Ramp. I'm like, that's actually way too fast and totally unsustainable. Ramp, however, is actually growing perfectly. That really actually goes through my head — doing it just right. But they're not, and I think that actually scares me.
Yeah, the bigger growth rates scare you. Your growth rate is perfect. Growing faster is unsustainable, slower is disaster. You're growing at the perfect rate. The Goldilocks effect. The Goldilocks growth rate.
I mean, that's Eric and Kareem just really nailing that. But I do think there is something about that that really does scare me — that fast growth and how easy it's become. So 'you can build more with fewer people' took me to two different things. One, it just made me scared — this was back in 2022 — really scared about software in general, because you can go out and build with five people a company that grows really, really fast. And two, I think the other point of that essay was because we have better tools and because you can do more with fewer people, startups should be able to actually start fresh and compete with these companies that you never thought you'd be able to compete with. I'm writing a follow-up on Primer right now, which is taking on K through 12 education, taking on public schools — not a group of incumbents that you thought you'd ever be able to take on. And I think they actually can. You can put education on a Moore's Law-like cost curve and make education better for less and less money. So that's pretty amazing. Obviously Anduril is starting to approach Primer's valuation. That I think is the really exciting part of it to me — less that five people can build things really quickly, and more that 100 people can build things to take on these really old, big companies.
What's your AI stack? You've been a little bit critical of AI's writing ability, which I think is fair. We also have Cougan Benchmark, which we do about once a week, where we get the latest model to try to tell us a joke, and it's pretty hilarious when you read it. Like, you're a standup and it just bombs intentionally, and it's funny to watch someone bomb intentionally. So it's good in that regard. But how have you been — what's your AI stack, and what's actually useful?
I mean, I still use Claude and ChatGPT and Perplexity to just kind of like — if I don't understand something, I'll throw it in. Or to make sure that I've explained something the right way, or to get feedback. It's basically a better Google search and a little safety blanket, since I don't have an editor, so somebody to tell me that this isn't the worst thing I've ever written. 4.5 is actually a little bit better. It stopped doing the thing that killed me that I've tweeted about a couple times recently — the em dash. It's not just a blank em dash, it's like a totally different blank. And 4.5 has stopped doing that, which is really good, because I would see that in people's writing and it really made me hurt.
What was the PG one? 'Delve'?
'Delve.' But you know what's funny? When I read older stuff that totally predates AI, I find 'delve' in it all the time. I think PG might have been wrong. I thought there was some weird round-tripping thing where the RLHF team in Kenya that OpenAI was paying to do the RLHF stuff had learned English in a very proper way. And so 'delve' was one of the words that was in the vocab, the course that they all took. And so they thought that it was a much more popular phrase than Americans think it is. Americans use it, but just not all the time. So it's kind of weird. But you'd see these aberrations all the time, where you'd ask it to write a tweet and it would use hashtags. And it's like, that's kind of what a tweet is, but not really, not today. It's not a banger. No one uses hashtags. Or if you do use a hashtag, it's in this funny, meta, ironic way.
Anyway, so yeah, give me some other hot takes on —
I literally just asked it today because I was trying to make my own joke for this podcast. The joke that ended up becoming the fact that with this unbiased conversation about Ramp, there's that one phrase — like, 'no bias, no something, no conflict, no interest, no mercy.' No conflict, no interest — that's what I was looking for. So I said, 'No, nobody, no dot dot dot,' and it was very confidently, 'No bias, no bet.' No — 'The phrase you're likely thinking of is no skin in the game, no opinion.' Oh my God. I still got it. Not a paperclip yet.
What's going on on Substack these days? They've got some social activity that's happening that seems to be picking up. You're still writing on Substack, correct?
I'm still writing on Substack. I've seen a lot of conversation about people going over to Substack — you can use links there, it's a friendlier conversation. I have not done that at all. My people are on X.
