NCI: AI Bubble, AGI Hype, and Blip 3.0

Speaker1: [00:00:00] Hi, how are you? Welcome to tonight's event. Lumida Wealth here. Nicolina Delgadillo, I work for Lumida and along Ram. We've been hosting this almost once a month, focusing on building community. This is, I think, our first time with Michael and Ram. They've been h- b- they've been talking about so many s- subjects.

Sorry, I'm a little nervous. I don't know where my mind is. They've been doing a lot of podcasts together for a very long time now, and we thought that this was a chance to go for them to see them live. So I wanted to introduce Ram founder, CEO of Lumida Wealth, and- Woo! There we go. Go ahead. 

Speaker2: All right, thanks everyone for joining here.

We're really excited to host Michael. Let me give you some background on Michael and I first met. It was at Howard Lindzon's it was a venture event, and Michael and I were at the [00:01:00] buffet. We struck up a conversation. This is in '23 spring. Good memory. So it's a few months after the ChatGPT moment.

Speaker3: That's right. 

Speaker2: And so Michael's a very humble guy, so after some prying, I discovered that he was the lead internet analyst at Goldman Sachs, and that was before Reg FD, so you could be on the banking side and the research side. You could be both sides of the wall advising and doing research. And so- We had 

Speaker3: some, some rules 

Speaker2: We had some rules, and so we sat down and I essentially s- just stuck next to Michael f- and then the, a crowd started forming because he has one of those longest running daily blogs- ... on all things AI, and he goes deep. He knows technology. Knows the form factor. He's advising NVIDIA, OpenAI, so he has deep connectivity in Silicon Valley.

So then after some time, a crowd starts forming, 'cause we're going deep on grid computing, edge computing, the cloud, we're in the value chain, and it became a mini conference unto itself. And then finally, like seven hours goes by [00:02:00] and it's almost midnight, the whole thing is cleared out and we're still chatting.

And then, but we have still so much to discuss, I escorted you back to your hotel room. I had to disclose to my wife, "I just walked a man to his hotel room here. We're just talking the entire evening here." But it was a real joy, and you've been the most frequent guest on the Lumida Non-Consensus Investing Podcast.

That's fun. People appreciate your thoughtful and brilliant insights on really all topics including AI, defense tech, which we'll focus on tonight but also just the big trends. So we- we're gonna focus 30 minutes here, kinda rapid fire topics and then we're gonna adjourn, go upstairs, enjoy some food and the rest of the evening.

All right? So let's get started. So we'll start off with Anthropic. Over the weekend, the CEO of Anthropic Dario, published a, a missive that we need to slow down AI development. That was followed by a message from Elon Musk and also his chief rival, Sam Altman. But seems very coordinated. Everyone wants to slow [00:03:00] down now.

Speaker3: And Sir Demis- 

Speaker2: That's right ... from 

Speaker3: Google. 

Speaker2: That's right. And now of course the exception being Mark Zuckerberg, says no, we gotta go. Of course, he's got the funding to make it happen. And then someone left Anthropic after six weeks of being there, saying, "Hey, if we don't slow down now, the risks are catastrophic to humanity."

I think that's kinda nonsense. I imagine your view is similar. What's your take on this slowdown? Why is this discussion around slowdown happening as opposed to accelerating to the future? 

Speaker3: No, I think it's a very topical item. I let me just zoom back a teeny bit. Sure. Because and by the way, that evening you talk about it was marvelous because you know, ra- Ram's being humble because he's extraordinarily deep on and horizontal on a lot of subjects, so we had a lot of fun talking.

The-- I've been tracking technology waves now for, close on to 40 plus years, starting with the PC. I joined Goldman Sachs in 1982, and that was the birth of the PC. That year was the PC was the man of the year on Time Magazine for the first time. [00:04:00] And obviously, Bill Gates, Steve Jobs created, two iconic companies.

Goldman Sachs, we took we took Microsoft public, I believe in '86, Dell in '88. I can't believe they're now, 25, 30 years later, these huge AI companies now, obviously. But what I'm getting at is that if you go to the PC and then the internet, where I had the opportunity to set up the internet research effort starting in '93 at the firm and that, that had its run and, same kind of echoes.

Is it a bubble? Is it overdone? What are the issues? And then obviously, mobile was the big thing, and now cloud and now AI. So a lot of waves. What is extraordinary about this one, and I've said, and I talk about this daily on, on the, on, on a Substack, and I've been saying ever since I've been doing it, I've been doing it now daily for almost three and a half plus years.

And I said when I started that, everything I've done in my career has been a dress rehearsal for what this is, whatever this is, and this is a multi-decade thing. I know people get very caught up, especially the founders, "Oh my [00:05:00] God, we're at the AGI, we're at the singularity." There's all these There, the, one of the critical things that's extraordinarily different about this w- tech wave is the cult nature of this.

This the, the foundation of the AI wave in 2014, '15, there's an iconic all night event that is very famous now, Larry Page of Google and Elon Musk. I've written extensively about it. They were deep into night, just like our evening, seven hours, they were up all night.

And they started as deep friends, and they ended up becoming rivals. Elon basically saying, "AI's gonna kill everybody," and so this is not new. And Larry saying, "This is just next step of evolution of humans, you're a spec- speciest," et cetera. And OpenAI got founded in that muck as well- and there's a lot of convoluted history. I won't belabor that. But my point being, what we're seeing today with so with Dario and Anthropic, et cetera, is just a continuation of [00:06:00] that soap opera. I call it a soap opera because unlike most investment curves, there is a a, a religiosity about this that I've never seen in any of the tech waves.

And this is deep, not just at the founders of the Anthropics and the OpenAIs that we're talking about, certainly Elon the Google guys the, the two co-founders, but it's deep and there's a, a, a decade plus history of what the, what's been called the access-- accelerate-- accel versus decel kind of stuff.

And obviously the, Anthropic and OpenAI were one company, they're siblings. They're, they broke up five, six years ago, and now they're both heading toward trillion dollar plus IPOs. And what I'm getting at today what happened this weekend, I call it the Blip 3.0. How many people remember the blip?

What I... What was the blip? Which was Sam Altman getting fired from his own company by his board three years ago, getting rehired over three days, over the weekend. It was a very intense w- [00:07:00] weekend for the entire industry. And very dramatic. I... we all watched it. I was- involved in some of that stuff from a couple of vantage points. Then we had the Blip 2.0 a few months ago when Anthropic, Dario, said, "Oh my God, this our latest models, Mythos, are very s- dangerous from a cybersecurity point of view." Took it to the government freaked out, and there was a blip, what I call Blip 2.0.

So three, three week it last-- two and a half weeks it lasted. Where the government said, "Everyone slow down. We gotta figure out what this is." And and now with this paper on Saturday by Dario, with all the founders jumping in- 

... 

Speaker3: And and saying, "Yeah we're with you," is the Blip 3.0.

And this, I think, could last far longer than three weeks. Because the, the core issues are two of these companies obviously are going for mega AI IPOs. So the... You can take- Yes ... a very skeptical view that these guys are talking their own book. 

Speaker1: Yeah. 

Speaker3: They're trying to get everyone excited about the dangers of AI, because then you can [00:08:00] get the government to create all these regulatory capture type of things.

It also allows them to potentially all collectively slow down a teeny bit on the gazillions of dollars they need to invest ahead of their IPOs. So the m- the metrics look better. The Anthropic, it got leaked. Because while Dario's talking about the world is gonna collapse, they're also pedal to the metal on, on their financings.

Dario's about to do his IPO. You... We'll see the S-1 drop any day now. They're, they've already trial ballooned $2 trillion valuations, eighty percent margins. We're gonna talk about that. Talk about that. And OpenAI which used to be the Coke versus Anthropic's Pepsi, is now definitely the Pepsi on its way to potentially being a Dr Pepper.

Who knows? We'll see. And he's Sam is at least twenty billion dollars in recurring revenues behind Anthropic. And and he took advantage of this brouha or the blip, what I call the Blip 3.0, and said, "Oh we are not gonna do our IPO this year. We'll do it next year, [00:09:00] but it's not the right time."

And then literally the day after they trial balloon, they're gonna do another raise about at a valuation of one point two billion on eight fifty, was their last private, Anthropic's about nine fifty. I'm going into the weeds of this. is that this is literally, I described it in one of my posts as literally both of these companies are at the frontier.

They're leading models on the planet. The engineering is extraordinary. The technical- The best talent in the world is at Anthropic. Yeah ... the, the technical the, what they're doing technically and building the the compute is extraordinary. I've never seen this in 40 years. I took the first data center companies public in '90s.

Now we're basically adding a lot more zeros. It's a whole different game. So you, it takes away your breath in terms of the scale of this. But both companies are Zooming down F1 cars on the autobahn with both the feet on the gas- 

Speaker2: Yeah ... 

Speaker3: and on the brake pedal at the same time. How- And that's what's going on

Speaker2: [00:10:00] how can they actually break? The game theory is to grow revenue. Now, of course, Anthropic, to your point, they're gonna go public soon. They gotta shift from revenue growth to a, a raise, beat model, and earnings and free cash flow and margins will start to matter. And so I'm more on the cynical side of this.

I think it's very silly to buy the Anthropic, IPO. They reported that 80% margins, that excludes payouts to distribution partners. 

Speaker3: It's, you gotta look at the numbers, how it's calculated. 

‍