Loading summary
Transcript204 lines
- [00:00]
A
Foreign.
- [00:07]
B
Welcome to Trends with Friends. We are in December and we're calling this one Trends with Friends code red. You can see the enthusiasm in my voice with with me today is my regular guest who's we're gonna take a victory lap today. Michael's been on Trends with Friends. I don't know, together for a year, year or two. Talking Apple, Google, kind of winning the AI race, you know, putting up, sticking with it, helping me stick with Google and Apple through the last year while OpenAI got all the press.
- [00:45]
And he's ex Goldman, as everybody knows. Looking good. Michael too. Looking good. Welcome back. It's been a month since we've done Trends with Friends but I haven't had much to talk about other than degenerate. And then Roy Rubin is in la.
- [01:04]
You know, we're not going to give out his exact location. He can and Calabasas somewhere. And he is joining us, Roy is and his other partner, Roy. I have two Ethan's that work for me. It's Roy and Roy R Squared ventures that focus on e commerce, mostly Israel and United States, I would say. Right Roy?
- [01:26]
A
That's right, that's right, yes.
- [01:27]
B
And the founder of Magento, which is still going strong I think independent or owned by ebay. Who's it owned by today?
- [01:34]
A
Owned by Adobe today and actually next year will be 20 years since we started this project. So you know, I'm very proud.
- [01:42]
B
I don't know if Mike knows the story. Quick story on Magento which is like one of the first big open source. Mike Willis was head of research at Goldman Sachs. Mike, do you remember Magento back in there? It was after you left Goldman Sachs.
- [01:52]
C
I think it might have been after. Yeah.
- [01:54]
B
But ebay public. Michael took ebay public at Goldman Sachs, so.
- [01:58]
A
Oh nice. Okay, there's a connection there.
- [02:01]
B
And eToys.
- [02:02]
A
Oh wow.
- [02:03]
C
EToys as well.
- [02:04]
B
Yes.
- [02:04]
C
And don't forget Webvan since you're and webvan doing the great hits.
- [02:07]
A
Michael's famous five years old at that point. So yes, I remember all that quickly.
- [02:12]
B
On Magento because this is important to the discussion as we talk about agentic commerce that OpenAI is doing. You're the expert here. Roy writes a lot on LinkedIn so we'll share some links. Michael obviously has his my favorite AI newsletter full of links every day, almost relentless. Roy's starting to write more which is always a good sign when the smartest people, you know, are writing. But right quickly on Magento and then R Squared and then we'll get into it.
- [02:37]
A
Yeah. So Magento was An open source e commerce platform.
- [02:43]
B
Still is.
- [02:44]
A
Still. Well, still is, but now, you know, much more cloud. So certainly Adobe acquired it. It's known as Adobe Commerce now. Started almost 20 years ago. We came out of an era where open source was kind of the hot thing in the market. Developed what was then very revolutionary in terms of having a platform that was extensible.
- [03:08]
And open and allowed anybody to download it and kind of do what they want with it. Developed a great enterprise business. Sold it in 2011 to eBay, PayPal and then it was spun out in 2016 out of eBay. It was part of ebay Enterprise by that time. Spun out, acquired by Primera Capital and I rejoined the board at that at that time and we sold it to Adobe in 2018 the second time. But it powered, you know, hundreds of millions of dollars over the years of, you know, of goods. The funny thing about, about Magento, I still talk to a lot of entrepreneurs today that Magenta was built pre cloud and what we deliver to our customers was a zip file. So that's how old I am when I still remember that we had to compress into zip files and send it to our customers because we never took care of hosting infrastructure any of the jazz that we know of today. It was very different back then.
- [04:03]
B
So since then it's a way to get me to stop talking the. And since then we become friends as I am with my and we've ridden around the world thousands of miles together. You're a beast.
- [04:14]
A
I mean you're a photographer essentially. So you just have to tell me when you're out again and I'll.
- [04:17]
B
Roy, Mike. Roy could be right. I. I can't. I can't fucking drink water while I'm riding because I'll fall over. Roy will like do booking, make reservations, flight reservations. We'll be doing 20 miles an hour on his phone. He uses an Android. Mike. He uses an Android. Okay.
- [04:35]
C
I love androids.
- [04:36]
B
An Android man who went chat GPT. We'll get into this and now is back I think thinking Gemini and so at the high and now doing investing out of you're raising your second or third fund.
- [04:49]
A
Our third fund at this point.
- [04:50]
B
Yeah. Yeah. And born in Israel, served in the military or born in the States.
- [04:56]
A
Went back to born in Israel but lived in the US most of my life. Grew up in the US and then spent five years in Israel from 18 to 23. Went to the military and and then came back here when I was 23.
- [05:06]
B
Yeah. So we've been to Israel and then we're going to Israel in March, Mike, we may drag you on this trip in March.
- [05:11]
C
Yeah, I'd love to come.
- [05:12]
B
Okay, yeah, that's a great idea. We got to get Mike on this trip. Okay, so here we are. It was an OpenAI world, even though Gemini was probably not scared. We can get into all the possible tactics, but now I think there's this Wall Street Journal's covering it. Michael, you've been covering up this code red. So Gemini 3 comes out. Michael, you always thought Google could win the long game on this. Tell us what's going through. I mean, I got to think Sam knew this too, because he's not. Maybe he's delusional, but maybe he's not. And so tell me how this flipped so quickly, even though everybody, like guys like you, knew the end game. And even Fred Wilson, right? Like, even Fred Wilson. So quickly, before I die, Fred, this is how great the markets are right now. Forget. Sorry, just how great technology is right now. Forgetting politics, forgetting markets, forgetting anything. You got the cloud, you got the smartphone. It'd be fun to have low interest rates, but we still have relatively low interest rates. We've got stupidity around us, and we a lot of distrust, but you've got drones, you've got semiconductors, you've got AI, you've got the cloud, which intense development still around the cloud. You still have the social network world. You have, you know, stable coins, you know, and so tell me how. What's going through everybody's mind now that, like, there's a chink in the chat GPT armor? Why is it a. Is it a Code Red? And. And just give us your opinion, then we'll go to Roy.
- [06:52]
C
Well, it's reported to be a code red within OpenAI, where Sam Altman apparently put out this internal memoir. That was the reporting. The irony here is that, you know, OpenAI just turned three. ChatGPT just turned three. Last week, November 30, 2022, was the. Was the anniversary, the birthday, as it were, of ChatGPT, and December 2022 was when Google reportedly declared a code red in 2022. So it's three years later, Google now, under Sundar Pichai, who everyone said, you know, was not a wartime CEO on AI, etcetera, and all of his peers at Max7 were ferociously building, investing, investing gazillions, etc. Google has been playing the long game. And that's why two years ago, I said, hey, Google's going to win this round with ChatGPT. I put a detailed post, I put some stuff up in the last few days on this. And it's not a, it's, it's, it's, it's, it's a, it's a thing, It's a thing you could say because it's a movie we've seen before. We saw Microsoft versus Netscape in the mid-90s when I led Goldman's Internet effort. Netscape basically, you know, took over the world.
- [08:05]
B
It was your entry. It was for guys like me who weren't on the Internet yet. It opened up me from DOS pre Apple, it opened up me from like truly to the Internet as a 30 year old person.
- [08:18]
C
And for three, four years everyone thought Netscape would take over the world. It would just grow, grow, grow. Take browser share and it would be the thing. Microsoft saw it a year later I had Bill Gates on stage and the Netscape folks and Bill basically he announced on stage they were going to make their browser free.
- [08:36]
B
Did he make an announcement? Was that stage on Epstein island?
- [08:39]
C
Just one moment it was adjacent. We could see it in the horizon.
- [08:44]
B
Goldman was involved. It could have been just on the dock. Okay. But no, no, anyways, I'm kidding. So keep going. So, so Gates is up there.
- [08:53]
C
So the point I'm making is that three years is an interesting time, right? It's when you look at the world of kids and you know, toddlers, you know, when they go to kindergarten, they realize they're not the only one. And, and, and in the AI world Google has, you know, now all of a sudden all of these companies are really competing and this last two or three weeks, Google seems to have come from nowhere. It was not nowhere. They've been working on it for three years, actually years longer. Google, you know, actually came up with the, their researchers came up with the transformer paper in 2017 that gave us the T in GPT and, and then Google bought the, the brought back the lead author of that paper about a year and a half ago for $2 billion. His company character AI. So they've been building, they've been doing this under the thing and they've got a full vertical stack. And this last week they announced that their Google TPUs, which is their AI chips are going to be, which they've used for all of Google's work for all of these years that they do a great job for doing AI. They're built directly by TSMC and for the first time they're going to be available via Google Cloud to people like Meta OpenAI is going to use them lots of other customers. So GPUs you know, they're going to basically compete with Nvidia and have 5, 10 million TPUs being used by non Google customers over the next couple of years worth tens of billions of dollars to Google. So that's on the, on the chip side and on the model side they came out with Gemini 3 and Gemini 3 just totally beat everyone's expectations relative to the best from OpenAI which was GPT 5.1.
- [10:36]
And Sam and his team have seen this coming. They've been trying to do a big upgrade on GPT. They've had a lot of research issues. Remember they've also been under the direct laser fire of Meta over there, Zuckerberg, who's basically spending 10 billion a year hiring every researcher on the planet earth, half of them from China. And OpenAI saw a lot of damage, lost a lot of people. So I'm not making excuses for them, but it's a real world, this is a global competition. And yes, Sam does have challenges to make sure he continues his poster boy lead on models in the next two or three years. Very important because he's already announced over a trillion dollars of infrastructure cloud deals. Now that he's separating from Microsoft, he's got to somehow finance his own data centers. So you know, he is a toddler competing with the big boys and, and he's got a big challenge. So that's why it's, it's, I, I, it's not unexpected that he was candid enough to call a code red within his company three years after Google called the code red. And you know, this, this exercise will continue. I don't think they're over, they're, they're, They've got over 800 million people using GPT. Google has 650 million by their numbers. You know, my view is that over the next couple of years if both companies do their job right on the consumer side, they'll each have a billion plus out of, you know, 8 billion people on the planet, 5 billion on the Internet. Everyone wants to use Internet agents reasoning, et cetera.
- [12:11]
B
Because unlike like ICQ or chat products or aol, like the more you use something theoretically, the more you use ChatGPT, the better it gets theoretically. And there's a lot there. Does it matter about UI at some point and switching or does it UI matters?
- [12:30]
C
Absolutely. You mentioned of ICQ which again ages us.
- [12:34]
B
This is, you know, chat that became Netscape chat product.
- [12:39]
C
But these were network businesses in the context of AI models. Yes. The more people you have use them, if you have permission to Use their data. Your models get better both from a training and an inference point of view. And the more data you have, the better it is. It's not different than what Google has done for 25 years in search. The more people use Google search and there's over 10 billion people searching every day. Google search gets them way better than anybody else because they know what people are looking for and they can optimize it for search. So search is just a, a different version of AI in that way. It's machine learning, it's algorithms, et cetera. So, yes, more users, more data, more, the better the product. So it is existential for OpenAI to have to the lead position billions of users and they've got to figure out a way to continue to be ahead of the curve and figure out how.
- [13:30]
B
To raise money just to spend it on debtors and how to charge a customer. And that's where I want to come to the customer side.
- [13:37]
C
Correct.
- [13:38]
B
I'm the last guy to use a product investor, right? Because I'm like, I'm not a tech person, so I'm not going to be in the line of scrimmage with all this stuff. And so I'm friends with people like you and Roy, who's riding his bike, ordering Ubers and, and showing me Chat GPT way back at the beginning as a Google person, saying, oh my God, how you're never gonna, I think it was you, Roy. I'm never gonna use search again. Even though you were an Android user or you were, you were paying for Chat GPT and you're an E commerce background and you recently wrote this incredible piece, you have to explain it to us on LinkedIn about.
- [14:19]
When platforms start building out. So the code red to you means what and where are you in the usage and what you see around you.
- [14:27]
A
So first of all, I love that as a consumer, I love that companies have code reds because we benefit as consumers. Let them code red this all, all day long, right? Because we're going to be the beneficiaries of it. So, so, so, you know, so I love it.
- [14:43]
You know, I think my behavior has changed. I used to be very deep into Chat GPT. I was. Loved the interface. I love the experience. And I have to tell you, if I have to look back over this past maybe six months, certainly in the last few weeks, the AI mode, which I didn't really understand at first, really sticks to you when you can do a quick search as you've always done in the last, you know, wait a.
- [15:10]
B
Minute, you're Saying the Gemini mode.
- [15:12]
C
Yeah, Gemini AI mode. It's the AI search.
- [15:16]
B
Keep going.
- [15:16]
A
But like, are you finding yourself also just leaning on it a lot more than you've done maybe in the past year or. Sure, right.
- [15:25]
B
Six months ago with Google, like, it's a very sticky experience for me when I search, but keep going.
- [15:30]
A
So, you know, before this podcast, I opened up Chat GPT and I actually looked when it was the last time I used it. Now this is something I used every single day, multiple, multiple times I met.
- [15:40]
B
Some of my smartest friends were like that, Howard, you're. I just better be living on Chat GPT. And I'm like, why, like, Google works Like, you're.
- [15:51]
A
It's been days since I've used the Chat GPT app on my Mac days. And I. This used to be an hourly occurrence, you know, maybe even more so my behavior.
- [16:02]
B
That's why it's code red. That's why it's a code.
- [16:04]
C
And no. And it's coming through in the data. Chat JPT is down literally 6% in. In their weekly users.
- [16:11]
B
But I think it's worth numbers in the last month. It's worse than the 6%. Right. You have the weight of Google with their 10 billion again, I think what makes this different is we were starting with an installed bait. Like, you know, people throw around the 700, 800 million number, of course, based on the installed base. When finally a killer app came out, you know, with an app store and with the way social works, we were waiting for a killer app because we thought crypto was a killer app. We thought the next social networking app was the killer app. There was no killer app. You know what the killer app was? Robinhood Degeneracy. I wouldn't have predicted that was the actual killer app of Web2 Coinbase, Robinhood, now Polymarket, Kalshi. So there was never the next killer app. It became an interface in technology. So when everybody threw around the 7,800 million numbers, it wasn't. That's like. That's like 30 million people using ICQ back in the day relative to how many people are on the Internet, of course it should be 7,800 million. And based on how much free press and actually the product was good and how crave people were for something new, that made sense to me. Right. So I think this goes to what you're saying, Mike. It did take, you know, three, four years. So. So you don't use it. And why are you. How are you using it?
- [17:26]
A
I got to tell you, I just got a Renewal notice for, for, for Chad, the work group edition, whatever the, the business edition. And my partner coincidentally emailed me and said hey, should we try Gemini? And I've been using Gemini. I told him, listen, we, it's part of a workspace product. Go for it. We already have it.
- [17:42]
B
Yeah.
- [17:42]
A
And I'm not going to renew ChatGPT. So I think the revenue.
- [17:46]
B
There you go. That's my worry. It's just like all these other ones that have these shoot up revenues.
- [17:52]
A
Yeah.
- [17:53]
B
Nothing worse than having. Forgetting their cost and their, their cost that they weren't making money yet off that revenue. Nothing more dangerous than having the wrong revenue. And the irony is that there's nothing worse. I've done this with stock twits, you know, because I started stock twits and I'm, I was a business person, not a tech person. All I thought about was revenue, not growth, right? And guess what? It was really bad revenue. Even though we had good revenue the first three years. And then you, you realize you tricked yourself in like focusing on the wrong thing, but keep going. Mike and Rick.
- [18:32]
C
No, I just wanted. Just because you brought up revenue. This is just recent data. Look, in three years OpenAI has become the most successful subscription driven business on the planet, right? Much better than AOL, which we've been talking about. They owe almost 15 to 20 billion in revenues. Over 5% of those are paying $20 a month customers. And they were just in the process of building their ad business, you know, to counter Google because that's the only other way to do it on the consumer side. And they postponed it in this red. Code red. He said we're going to refocus on making sure our model can compete with Gemini 3 and Anthropics Claude and so on.
- [19:12]
B
And here's an anecdote about why I always fucking distrusted Sam. And the fact is we all knew it was going to be an ad model, right? Or at least I knew like great that they built a subscription business, but it was not going to be big enough. Right. And ads will work. My problem is always, and this is the blue link problem that everybody complains about Google. If you're going to be an AI product and you're truly going to do ads and if the best, let's go into drugs. If the best product for people to search, we already know what the cheapest product is to go order drugs. Cost plus Drugs Mark Cuban. So theoretically, if you search anything that is related to like buy this generic drug, there only should be one link, right? Like so in A true. In a true world, the ad models fly. And guess what? That's why they won't like to me. How do you try if open AI, if AI is supposed to be better and it's not serving the answer that you complain was flawed about Google, then Jesus Christ. Like the whole thing's gonna come tumbling down, meaning anytime someone searches AI, I gotta like cost and a drug problem. It's Mark Cuban's product, plain and simple because it's marked up like Costco. So it's like the whole thing's wrong. So we'll get into the agenta. So they paused ads. Obviously ads is gonna be a business for them at some point, but it's flawed. Just from what I said, like I don't trust them, right? I already, I already trust or distrust Google enough that you're not going to get me switched. Open AI, Their big problem is this slowdown in revenue and the fact that they're not going to be able to work on new revenue models because they got to catch up to Google and they don't have YouTube maps, they don't have Android, they don't have Gmail.
- [21:03]
A
Well, they don't have a business model.
- [21:04]
B
They don't have Chrome.
- [21:07]
They don't have.
- [21:07]
A
A business model outside of subscriptions, right? And we know subscription revenue is going to decline, right? Because, you know, because we're seeing obviously the, you know, the competitive landscape. So what is their next act? And it's unclear, is it apps? I think that's where they're making their bets.
- [21:23]
B
So let's talk about apps, because this is where you said, like, tell me what happens when you start doing apps. Because this is like near and.
- [21:28]
A
Well, if you're, if you're an infrastructure play, if you're a platform player, you kind of want to leave some holes in the ecosystem for others to plug into you, right? Because you want to create opportunities for others to create, you know, some tailwinds for distribution of the platform. That's always been the case.
- [21:43]
Here. They're starting incredibly early in trying to figure out what the point solutions, what the revenue centers are going to be like. When you look back into my, and I actually did some research before this podcast. Microsoft, Apple, right? Infrastructure operating system, Google Search engine, right? All these platforms, it took them years to figure out what an app business, right, should look like in order for them to create some additional growth engines for their core business. OpenAI is starting much, much sooner, right? They're out of the gate. If you think about this, trying to figure out what they build in conjunction with the core subscription business, with the core software business that they have. And it feels like they're not leaving a lot of room for others to go and figure that out.
- [22:33]
Which, you know, in, in my opinion is, is, is a signal for, hey, we need to figure out ourselves what the revenue opportunity may look moving forward, where the big growth engines are going to come from versus trying to figure out what the or, or letting the ecosystem build on top of us. And then we can be acquisitive, we can maybe duplicate. We can, we can, you know, we can really begin to double down on what we see works. So they're willing to take the risk and figuring out what these apps may look like in order to create these, these additional lines of business. Because I think they see a real concern for their core fundamental business. There's some pressure there.
- [23:16]
C
So what, what is that? What is the.
- [23:18]
B
So Code Red is good for consumer.
- [23:22]
But in this world, a Code Red could just be like lead to more code reds for them. Right? Because if the revenue starts declining, the money gets harder to raise.
- [23:34]
The lawsuits are still going with Elon. What is the rest of the world thinking? The Gro, the Nvidias, like, you're so close to all this. Like, and I want to get. And then I want to come back to Shopify and E Commerce and all this. But because Royal, I gotta have him chime in here because E Commerce still. How does it all fit into all this? But what it. I mean, Jensen came out and said, oh, congrats Google, you know, blah, blah, blah. Like, it's a $4 trillion company. And it was like kind of like a weird. The Internet took it as like a weird thing to do. A press release saying, you know, Google's great, but like, why do you need to say that? I've. I've been, I don't know, tech, but I'm an anti Nvidia guy. I think he's doing the right thing by shareholders. He's got to do the right thing by shareholders. Tim Cook had to go kiss the ring and bring like whatever he did a key to the world. Trump bow and kiss the ring. They ought to sit there at dinner. So I'm trying not to judge, but man, Jensen really was doing the circuit, which he had to do. And he was really cowtowing to Trump as he could and as he probably should for the shareholders. But it just feels so gross. And Google's really not had to do anything. I don't even know when the last time I saw Sergey or Larry was. And now they're number two and four in the world in wealth. And so the clamps seem to be really coming down and. And then all of a sudden Apple then. So Apple is back at all time highs and they haven't even been in this battle and recognize on $4 trillion market cap quietly this year. They haven't been on any podcast. Tim Cook did have to go kiss the ring. Nobody likes that he's thinking about retiring. The stock's still. Who's going to replace him yet the stocks at all time high. You Mike, you and I have been on this forever. Your reasoning was. I'm trying to remember all the reasons you still liked Apple. But the real reason looks to be like these guys all go kill themselves. We're still the delivery mechanism for half the other than Android whatever it is. Half the. The and you go into an Apple store and it's packed.
- [25:52]
And then the other clamp is Fred Wilson who is an Android guy and I'll flip back and forth and he just bought the Solana phone on Android. Right. Like he's, he's a. He's back writing. He's so excited about all this stuff. Right as crypto meets AI meets security meets nuclear and energy. Like Fred's back writing just like you guys like all the bet ohms back writing all the all you know, Michael Burry's out writing in, in finance. Everybody. Michael Lewis is back out to him. Everybody's back in the game because there's so much going on. As you said, code reds are good kind of for everybody when the, when the stakes are this high. Like how does the. How is Apple laughing right now and what's in the video?
- [26:32]
C
Well Apple remember they, they signed you know, OpenAI as a partner about a year and a half ago.
- [26:38]
B
Yeah. Is that the killer move? Just give OpenAI like the perception that they were going to.
- [26:43]
C
It was a, it was the right move at the time because OpenAI is the leader. But look at the news right now. They've. It's. It's all but certain that they're going to work use Gemini, a proprietary version for Gemini from Google to power Siri and that's going to be the big thing next year. So that accelerates Google's distribution and they.
- [27:01]
B
Didn'T have to spend on centers. They just built more stores to sell.
- [27:05]
C
Their product have high quality and they're using Oracle and then we're using everybody else to do that. It's. You're totally right. And what I'm getting at is Apple continues to be in a very strong position. My view on Apple has been, look.
- [27:20]
B
I think it's stronger than ever.
- [27:21]
C
It's, it's, it's beyond stronger than ever.
- [27:23]
B
Maybe not the most innovative company anymore, but it's stronger than ever.
- [27:26]
C
It's, it's way innovative. They don't get enough credit for innovative because the media is focused on things like large language models. You know, there's a boatload of innovation that goes on underneath if you, you know, no one really thinks about, you know, they got six ecosystems, operating systems, the key apps across all of them. They control three to four, $5 billion, a billion users in notes, in messages, etc. They work across every imaginable platform from a computer to a smartphone to a wearable, the AirPods, etc. In seamless way. There is Android company that matches it.
- [28:06]
B
So does Google.
- [28:07]
C
Sorry, so does Google.
- [28:08]
B
Does.
- [28:08]
C
But they're not, they're not as Google as amazing Google is from an Android point of view. Howard Their Pixel phones are less than 2% of Android phones around the planet. Google is a, Android is a distributed strategy where hundreds of OEMs, mostly in China and India make lots of phones. Most of them don't ever talk to any Google service.
- [28:32]
B
Okay.
- [28:32]
C
Google makes no money from Android for the most part. They make money on search, etc. So it's a very different strategy is what I'm trying to say.
- [28:40]
B
Yeah, and Apple is still a nice little. It's a nice little have.
- [28:45]
C
It's a huge have. It's a huge have. Especially in the world of AI. They will be able to distribute their Gemini.
- [28:54]
B
Fred Wilson said he's like, it sits in the corner. He got his 95 year old mother, he wrote this post saying, you know, and Fred hasn't talked about AI. Like he's investing in AI and he's not. Hasn't been writing for a couple years and I'm lucky I get to sit down with him and then all of a sudden he's writing again. Two weeks, he didn't even tell me he was going to start and then he just wrote a post about him getting his mom hooked up with agents on Gemini. On Gemini because it hooked in with Chrome and it was easy to show her what to do. And I'm like, are you kidding? Like forgetting the code red. Fred Wilson and Roy Rubin. I got to talk to Brian Nergaard who was like Howard chatgpt, if you're not on chat, I'm sure he's flipped, but I haven't. I mean the Data points are insane. You were always on this Mike. Um, I think Om probably was, you know, anti OpenAI, but like he's more media. And then when I heard Roy write this post about app. When I saw Roy write this post about apps and then tell me I'm thinking of switching as an early OpenAI user, that's all I needed to know.
- [29:54]
A
But let me ask you a question, right? These models are so great, right? And they're going to surpass each other, right? There's going to be a continuous evolution in movement. Do we really care as consumers? I like the point where it's all great. It's all fantastic.
- [30:08]
C
Good enough.
- [30:09]
B
I don't care that my son uses Open and I don't care if he switches. Like I don't care. I don't care if he uses so Fire Chime. I don't care if he uses Robin Hood or.
- [30:18]
A
Right. This is why the, you know, where.
- [30:22]
B
We got to have a business model, Roy, in this, in this type of environment, you do have to have a business model.
- [30:26]
A
What I'm saying, Google and Apple are, are the edge, right? If we use AI on the edge, right? Where it's easy for us to do it through our phones or through our Google, you know, search behavior, that's what matters for scale, right? That's where the winners are going to be. I find it hard to think that we're going to recondition ourselves as we've done this last two, three years, to use new apps, new modes of interacting with AI. In a ChatGPT case, for example, long term, if we can get AI at the edge, it's all we need.
- [31:00]
C
That was the core reason why I said Google wins because they're already there. You know.
- [31:09]
Have to buy a new computer or a new device to be able to download. It was just on, you know, it was in the same browser. She probably didn't even know she was using Gemini. It was just a Chrome browser, which by the way, 70% of the people on the planet use other than Safari. And so it was easy. That's what I did with my mom when she was around a couple of years ago. Yeah, you go to the thing that's there, it's good enough. To your point, Roy, the models will continue to get better, but no one will have an advantage given the monies that's being spent, research is being done that a dramatic advantage over one versus the other for any length of time. So they're good enough for now. And so then the advantage shifts to all these other variables. Distribution, you know, presence, the, the way where the user habits are. You've got billions of people who already used to typing. You know, my sister doesn't know you know how to use Google. She basically types into her Safari browser and because Google pays Apple $20 billion a year, she thinks she's using Google when she's typing in her queries into the top address bar of Safari. Okay? And, and, and, and, and, and, and then in the next month or two she'll be doing the same thing, asking big long questions. She'll be using Gemini. She won't have an idea.
- [32:30]
A
Well, that's why, that's why OpenAI released their @ browser because they said listen, we're not going to change beh people are going to use browsers. We're not changing.
- [32:38]
C
That. No one's going to use.
- [32:39]
B
OpenAI. Like I.
- [32:41]
A
Sat. Have you heard.
- [32:42]
B
About. Everybody was saying perplexing and all these.
- [32:45]
C
Brow. I've written extensively about AI browsers. Cool about the AI browser But the problem, Roy, is, you know, with a new browser you have to download it. Atlas is a great browser. I've tried it. Comet from perplexity. There's six or seven of them. The problem is most people, you know have not downloaded a new browser in a decade or more, okay? If you're on Apple, you're using Safari browser. If you're on anything else or even on Apple or whatever, you're using Chrome. End of story. That's.
- [33:15]
A
It. I downloaded Atlas. I played with it for like 48 hours, okay? And it overtook my Google search experience, which just pissed me off. And I installed it because I said I still need Google search. I still need to be able to search for something. And when you take that away.
- [33:31]
B
From me, game set, match point, okay? Like Microsoft, I'm not even talking about them because they're great, they have cash flow and I can't, I can't, I don't know. I own it because I index and in Facebook I guess I own because I've indexed. God bless. Maybe the glasses are an edge case that they can sneak up. I don't think it's big enough but they can sneak up. You know, it was a smart, some smart moves he made that I think Apple could have made. But Apple should have bought luxottica and owned all the like I think Apple should have just bought luxottica and just owned all the brands around your eye instead of building their own glass. But like forget that. It's. I don't think it's big enough and it's just too independent. I'm just, you know, in the world of Warby Parker, I just think it's going to be hard. Right. And Apple just has that distribution at the stores.
- [34:15]
A
Right?
- [34:15]
B
Right. Wherever city you go to, that's important. You're five minutes from an Apple store, and if you're 20 minutes, you'll still make the trek if you need something. So they're just going to continue on that McDonald's road. They own the real estate, they own the trust. People want to work there. The products are good. People are addicted. But the more important thing that Google have always said is YouTube, if we get down to it, video. If we look at Netflix and the streaming wars and the content wars, Netflix is now the largest streaming platform. Bigger than Netflix, bigger than everybody. Every podcast, this will be on YouTube. You can't win that. So they have the edge and Apple has the edge and then they have the meat because they have your attention, because the true attention is on YouTube. YouTube's on your TV now. It's not about Twitter or Facebook, it's YouTube. So they own the two most important.
- [35:15]
A
Things. The.
- [35:15]
B
Edge. Not they own the edge, but they're on the edge. They have the stack and they have Gmail and they have maps and they have the biggest one of all, YouTube, which is just going to keep compounding, period, end of story, and the search that goes with YouTube. So I just think it's just getting going for Google at this point. People don't want to believe because.
- [35:37]
A
It'S. There's leaps and bounds.
- [35:39]
B
Ahead. Yeah, what's.
- [35:40]
A
That? I think they're leaps and bounds.
- [35:42]
B
Ahead.
- [35:44]
For keeping me in this. I think Apple's now leaps and bounds ahead by not going into debt to do all this stuff. Think about these companies that took debt on that we never thought that would have to take debt on. We never thought Oracle would have to take debt on again. We never thought Facebook would have to. We always were talking about their cash. What are they going to do with their cash? All of a sudden they're taking debt at 5, 6% or 4% instead of 0% when they could have done it six years ago. But there was no use for it. There was no use for it six years ago. So I don't know. There is like a serious problem underneath Google and Apple that a lot of people have fallen into the trap of like Mag seven and I think even other than Elon, who has space and, you know, I hate to say it like he's. He's back on top, but I don't think Grok has a chance. Is anybody using.
- [36:30]
A
Grok? I use Grok through x.com tweets when I want a clarification for something. I think they're embedding of Grok inside of.
- [36:40]
B
X. So it's a good use case on Twitter. I guess I don't do it.
- [36:42]
C
That way, but that's a few hundred million people on.
- [36:45]
A
Twitter. Yeah, I love it there because.
- [36:48]
C
It'S. Early adopters like us is what I'm talking about. It's.
- [36:50]
B
Not. And you love it because what. It really does explain things for.
- [36:54]
A
You. Yeah. I can ask immediate questions. I can ask for references. I can really learn, you know. You know, at times when there's a question that, that pops up and it's integrated into the, into the tool. It's a, you know, it's a brilliant strategy. The question is, is it going to change my behavior now to go to Grok independently? I think the answer is no. I'm. I'm honestly shifting a lot more towards Google. I never saw that coming. And it's fascinating to see, you know, to see that.
- [37:23]
B
Evolution. Okay. I think we've given enough. I think we've given enough to this one. And then. So I want to end it with. Go ahead. Mike and I want to talk E commerce and we'll let.
- [37:32]
C
Everybody. Yeah, shifting to E commerce because the other one, my other E commerce.
- [37:35]
B
Is other than the edge is.
- [37:38]
C
Gone. It's the.
- [37:38]
B
Third. The edge is gone.
- [37:42]
There's Chinese phones. Potentially. Tesla's the edge with their car. So, like, Tesla's going to be relevant. I'm, I'm actually long Tesla right now just because of the chart, but also because.
- [37:54]
Like, they're enough in the game and he's just, he's just tough enough and smart enough and mean enough and angry.
- [38:02]
A
Enough. And so Howard, I just wrote about this also, and he's out of.
- [38:06]
B
Politics, basically, and he's up for grabs. You know, I heard about.
- [38:10]
A
This. We have a new Y that I got the, that I.
- [38:12]
B
Bought. You probably love it. You probably love.
- [38:15]
A
It. Let me tell you. Let me tell you what's happened. Right. My wife didn't even want it. I just bought it for the F. For the fsd. Just to try it and see and you know, and see how it is. We have five cars in the house. It's the only car getting used. We're fighting over it.
- [38:25]
B
Now. Yeah. Said the same thing. Like, Fred hates.
- [38:29]
A
Elon. The FSD is a game changer. It changed. It's one of those things that, that, you know, change your life. It's that moment that, you know, it's rare. You know, technology has done this a few times. Mobile.
- [38:42]
B
Phones.
- [38:45]
In New York, he went full FSD because he goes to play golf on the weekend and he has a bad neck. And you know, he trusts it. Like, I mean, he's still in the driver's seat. You're still in the driver's seat. But like, I'm going to get my son. I go to my son. I go, Max, I'm not an Elon fam. But you got to go get. You got your first bank account. You have a job. Just go get a $200 a month Tesla and like lower your costs there. And you know, want to have to say that. And Michael, you and I have been kind of bearish on Tesla. It's kind of not done anything still. No, you're bearish. But I'm, I'm, I don't want to talk about financials. But like when I hear Fred and I hear Roy Rubin and in the original Tesla, bulls are still bullish. So there's new, it's like kind of like there's a new group for, for Tesla. There's finally a new group. And while everybody's battling on AI and space, the Tesla doesn't have as much competition as I said. And Roy, you went to the Mercedes and it was a beautiful.
- [39:42]
A
Car. It's collecting by expensive Mercedes. I don't want to touch it. I don't care about the leather quality. Get driving to be fully automated. That's the new luxury. It's, it's, it's, you know, and that's what I argue that, you know, the new luxury is time. Right. If you can get my time back, I'll pay anything for it. Yeah, right. I don't care about the quality of the build. I care about my.
- [40:01]
B
Time. Okay, so that was another big unlock is that a lot of people are the f. That's finally starting to work. Obviously Google has Waymo, so we're not even. I forgot about.
- [40:10]
A
Waymo.
- [40:11]
B
Right. Like we're talking about all this shit and I forgot about to throw in the. Google owns.
- [40:15]
A
Waymo.
- [40:16]
C
Yeah. And that one without steering wheels. Without steering wheels. That's. Without. That's Waymo with no person in the driver's.
- [40:23]
B
Seat.
- [40:26]
And I'm going to give Google one more credit here because they're all evil. I know they're all taking my data and I'm a goner at some point with what they could come up with me. I tell you a story. My sister, my daughter Rachel's talking to ChatGPT about, you know, they're getting therapy or whatever. They think they're getting therapy. Go, Rachel. You're killing me. Everything you're telling that chatgpt about your dad is going to come back to haunt me, right? Like, you know, it's not about you getting therapy. It's about someone unlocking what your dad was like at some point in a, you know, she's doing it in a harmless way, I'm sure, but like, I don't trust Sam or any of these server companies or Nvidia, the China, they're all going to cut. I'm going to give Google the most important props as a human being into someone with hopefully a soul is remember the Baidu deal? They spun it out. They're not in China on purpose, right? They've done all this without having to fucking kowtow to China at all. Have they had to go in there and beg for anything? No. So they're. They're a $4 trillion company ignoring China. That. I mean, I'm sure they're in China, but they're getting, They've never gone in and begged to sell into China. They spun out of Baidu, you know, so, so kudos to them. In this world of everybody kissing the ring because it's table stakes now. Kiss the ring, get some orange fucking goo on your, on your lips. But Google never went in there to kiss the ring in that way. Jensen's got his mouth open and begging for China, right? And Google hasn't had to do that. So, so, so anyway, so E.
- [42:05]
C
Commerce, so let's, let's just want to. On the E commerce side, you know, two years ago when I made the Google call, within a week I also made the call on Amazon that Amazon was also the least recognized company this week. Amazon just had their big developer conference called Re Invent. I've written about it. It's amazing. They are finally, they're the largest cloud server company. They're bigger than Azure, they're bigger than Google.
- [42:29]
B
Cloud. Probably a buy in a world where you believe in.
- [42:31]
C
AI. What I'm.
- [42:36]
Everything that Roy was talking about in terms of agents and E commerce, et cetera, number one on the AWS side. AWS is now totally open for business with deep agentic products and services. They have their own chips, Amazon does, that'll compete with TPUs and Nvidia chips, et cetera. They are in a very good, strong position. Again, we were talking about distribution. They touch almost every major Business large, medium, small, around the world that wants to do anything with AI, they're going to use just like Internet people used AWS 20 years ago. AWS is now 20 years old to get on the Internet with a credit card to do business and do a website and so on. That's what's about to happen with AI in a massive way. And Amazon AW is the way for most people, most regular businesses to get online and use AI agents and everything. And, and, and they just announced this ton of detail. I won't nerd you guys out. Number two, I just want to add on the Amazon because of Roy's experience with Magento is that Amazon on the E commerce side of the business is taking a hard line against agents from perplexity. They did a cease and desist against perplexity. They're going to, they're going to take the hard stance against OpenAI. Even Google that hey, so they'll.
- [43:58]
B
Be the ones, not the content people that win the OpenAI thing. It's interesting the evil, you know.
- [44:02]
C
The enemy of anyone's on the E commerce side because the E commerce companies basically are saying use Chat GPT. They just did a big partnership with Walmart. You know you can buy Walmart through ChatGPT but Amazon's saying the ah, we're not going to let you without permission. You can't send your agents into Amazon.com and buy all this crap for all your users. We want to either get a piece of the action just like Apple wants 30% of the app Store sales or we want to work with you that we can upsell because once you lose that customer control, as Roy knows on the, on the Shopify or whatever in terms of the customer piece of it, you're losing a lot of other revenue opportunities. And so they don't want to give it away to the AI companies. And this is what is being known as the dead Internet or the broken Internet piece of it. Just like the AI is upsetting the content business where I don't have to go to newspapers, New York Times, whatever. It just summarizes everything for me and tells me what's news. Same thing's happening with E commerce and Amazon is on the front line of the max Sevens basically saying we are the big gorilla here. We want to basically you have to go through us. If you do it without our permission, we will sue. Sue the bejesus out of you. Which is what they're doing to Perplexity. Okay, yeah, it's a big thing. Yeah, the selectity is a Canary in the coal mine. Because what perplexity is, they're very hustling and I really like them, but they're not the big company. They'll be part of something else in a year or two.
- [45:42]
But in the meantime, the problem that is being identified here is something Amazon is basically, it's a really a shot across the bow for OpenAI, for Google, for anybody else who wants to have their AI becoming the key way to shop for.
- [45:57]
B
People. So does Anthropic become bigger? Do they. So, so, Roy, I don't know if you thought about this. How does.
- [46:02]
C
This. They just make rumblings that they're going to do an IPO in 2026. Anthropic is all E commerce. Their bread is buttered on the enterprise side, so they're not playing in the E commerce game. Their core thing is basically businesses using Anthropic for coding and for a whole bunch of other enterprise services. And Anthropic, remember Amazon is an investor in Anthropic and so is Google. So is Google and so is Google and now so is Microsoft and so is Nvidia. Those two just announced billions of investment in Anthropic, which is why Anthropic is comfortable enough this week to say, to leak out that, yeah, we're talking to bankers in 2026 to go public at a 300 billion 3,350 billion valuation compared to 500 billion for OpenAI. Remember, anthropic was 20 billion. I remember that a year and a half ago. I mean they've come a long way. And you talk about revenue models. Anthropic's model is all enterprise APIs through.
- [47:04]
B
Amazon. They have a better business.
- [47:05]
C
Model.
- [47:07]
They have much better margins than OpenAI wants to do everything right, which is their opportunity because they're chatgpt. So and so there are three ways to make money on the Internet. One is we talked about subscriptions. Number two is ads. Number three is E commerce, which is what we're talking.
- [47:25]
B
About. I think there's a commerce because I think ads bleed into prediction markets and it's a new form of.
- [47:30]
C
Media. I view that as.
- [47:32]
B
Transactional. Yeah, it's.
- [47:34]
C
Transactions. Yeah, because you're paying, it's going.
- [47:36]
B
To replace a lot of ads. That's why I'm bullish on prediction market because it's actually not betting as much as it is a better model for media to, you know, like to see the price. Like, like, as I tell my daughter, it's like instead of yelling on the Internet. Right. So we'll End with prediction, where we got to get to E Commerce and we're going along because you guys are great and I get. I don't get to talk to you together. But.
- [47:59]
You know, my daughter's like, duh. Like all these kids who are arguing, I'm like, forget arguing, guys. It's gotten us nowhere. What Kalship.
- [48:08]
C
In.
- [48:08]
B
In. In. In Robinhood did, right, is insane. But, like, what Kaushi and polymarket are, is a new form of news, right? It's not perfect yet, but it's gonna. You know, he was on 60 Minutes for Christ's sake. They have like, no customers. Poly market, they got $15 billion valuation. You know, full disclosure, I'm a small personal investor in an spv.
- [48:32]
But they, they can't. They can't say they're not a tech company because the valuation will come down. But all they are is a media overlay, a yes no button instead of a stupid takeover ad, right? And all that revenue is going to cnn. Just got to deal with the prediction markets, right? Because it's media, right? Like, if you're talking about the election, show me the odds. Don't give me a meter at the top of your site. I just want to hover and see the actual odds. And I don't have to go to Twitter. If I see the odds, I don't have to complain. So I think this is a huge. Again, this is not this show. This is a huge disruptor to social media, right? It may help. You could go place your bet and of course you can still go to social media and blow your brains out and yell and be angry, right? But I think what people don't realize what I'm saying. Like, while you're using Grok, I get it. I'm on front page of Polymarket and Kyle should. To see what the real news is and what's really happening, right? Like, I don't need to hear Trump call somebody a name anymore. He is the president. I can't. But I can go to Call Sheet or polymarker and see what are the odds of something. Now, they could be wrong. Like, it's imperfect, right? And we know that. And the markets aren't deep and they're going to be wrong a lot. But go place your bet, then go argue. Like, the point is, why argue with what already is happening with people's money on the bet? So, so, so that could come.
- [49:52]
A
A few words on commerce because I have to wrap up in a couple.
- [49:54]
B
Minutes. Go ahead. So. So where does Magento eBay Shopify fit in here because obviously you know.
- [50:01]
A
Here'S, here's my take, right and whatnot.
- [50:03]
B
Right? This is what, not what everybody's using. So there's like four major like channels for forgetting YouTube but like the four major channels to sell shit.
- [50:11]
A
Right? So, so, so commerce is, you know, as, as, as Michael, as Michael alluded to, right. It's a very complicated experience, right? There's, there's upsells, there's cross sells. There's a lot of dynamic actions that take place in order to really optimize the, you know, the funnel, the conversion and the sell through. And you're not getting any of that on these front end, you know, chat GPT like interfaces, right. You're getting a transaction so you're leaving a lot to be desired, right? So as a consumer myself.
- [50:44]
When I do check out and I've tried Both, both the OpenAI ChatGPT version and the perplexity version, there's a lot to be desired, right? It's a very transactional one to one type of an, you know, of an approach. It's not, you know, this, this past 20, 25 years of evolution in commerce kind of stopped, you know, when you, when you, when you think about what they're able to deliver right now. So there has to be parity at some point. The experience has to be materially different from these chat GPT type engines where you're getting a lot of what you're used to get by going directly to the websites. And I feel like it's an experiment now. It doesn't really drive a lot of volume. I think that merchants are going to prefer for folks to be landing on their websites where they can optimize the funnel, where they can optimize the conversion. They're not getting any of that opportunity to do so by by basically uploading their catalog into ChatGPT. So I think, I think it's all noise. I, I don't know how much of it is real. I, I suspect there's not much.
- [51:46]
B
Happening there but that's the keyboard noise. It's affecting Shopify. It hasn't broken out to all time highs. Ebay was trying to and hasn't Amazon's, you know, lagged and whatnot is quietly I see, you know, all these kids are using whatnot for e commerce and obviously there's.
- [52:01]
A
TikTok. If I were one of the.
- [52:03]
B
Vengeance.
- [52:06]
The ultimate thing is it just stays as it is. E commerce people like people want.
- [52:11]
A
You to, you know, unless, unless you build a very interesting consumer experience where you can start to bring in a lot of those core technologies and allow third parties to build on top of your infrastructure. You know, it boggles my mind that, that, that ChatGPT hasn't opened up the infrastructure for commerce beyond just product listings. You know, why can't I, as a technology provider actually provide services for those that sell on top of ChatGPT? Right. Almost like. But this is where it gets competitive with Shopify. Right? Because at some point they're going to have to build an ecosystem not only of merchants, but also of technology partners that can create these upsells, cross sells these complex personalization.
- [52:54]
C
Products.
- [52:54]
A
Yeah. On top of it, it's a competing platform. Right. Right now they're all playing nice because they all need the news. They want the stock to be top of mind, you know, and certainly get the mind share. But, but ultimately I don't think this is going to work and this is going to be effective. The numbers are going to fall behind in terms of performance by selling through these, these interfaces, they're not going to be as good. So if I were them, what I would do is I would open up the platform and start to build a competing commerce platform. If you've got demand through your user base and you can build a, you know, interesting experience, I, I think there's a story here that can resonate. I would be excited about.
- [53:37]
B
That. But as you're looking at your fun, you're looking for people that are thinking through.
- [53:41]
A
That. Well, listen, these are all, these are all walled gardens. This is my problem right now, right. We're basically, you know, and I've written about this too. We're, we're, we're living in the Gemini world, the, the open AI world, the cloud world, the Perplexity world, you know, by the way, are those worlds going to talk to each other? Because I could see a world where that's all gated, right? So you'll have different types of experiences, different types of truths right across each of these worlds. So it's very interesting to see how this develops.
- [54:13]
But I think there's a lot to be desired and I just think it's all noise at this point. Maybe I'm a naysayer when it comes to this. I'd love to see that grow and become relevant. But I think that they're not really getting it right.
- [54:26]
B
Now. So Shopify, Magento are still fine because there's no, like. Because the noise is just noise. They may not be. Yeah.
- [54:35]
A
Okay. Yeah, that's my sense. Listen, this could change.
- [54:37]
B
Right? But I don't think, but it's. I, I believe.
- [54:40]
A
You. Yeah, Michael.
- [54:41]
C
You. A lot of experimentation, I agree with it. But you know, all the technologies underlying are changing dramatically. This agent stuff is changing like at 5, 10x a year, I mean, the technologies are changing. So it's like nothing is settled. So just look at the next two or three years as experimentation shop. Everyone's going to have huge opportunities to innovate, whether you're on the model side or the commerce side, and cut deals. In the meantime, they'll sue each other. We saw this in, you know, in earlier cycles. We're far from a settled landscape. The most important thing though is how do regular people, you know, Fred's mom, et cetera, what are their behaviors? And their behaviors are, you know, where the real opportunity is. And that's where, whether you're OpenAI or Amazon, they need to be there. And Amazon, luckily is just like Apple. They already have the majority of the world's E commerce transactions on anything and everything and behaviors and habits. So that, guess what, they're going to be able to deploy a lot of things that you're talking about, Roy, on capabilities.
- [55:46]
As aggressively as any startup out there or as an OpenAI. So it's a very interesting environment for the next two or three years. In the meantime, I still think Amazon is viewed as an underdog in AI, just like Apple, which I think is incorrect. Just like Google was.
- [56:07]
Three months ago as an underdog and they're not now. And there's a code red at OpenAI, which no one thought was going to be a.
- [56:15]
B
Possibility. All right, very helpful. I got a pop too. What a thrill. We'll talk about Agentic next.
- [56:22]
C
Time. Great to talk to you.
- [56:24]
B
Guys. All right, gentlemen, thank you, thank you, thank you. Have a great week, everybody. Very soon.