LIVE CLIPS
EpisodeĀ 10-16-2025
Martinique, Mauritania. Yeah. Well, let me tell you about Vantage Automate compliance. Manage risk, prove trust continuously. Vantage Trust Management platform takes the manual work out of your security compliance process and replaces with continuous automation, whether you're. Pursuing framework or managing. Did you see this Consumer Reports report? Paris didn't.
Product and it's powered by the different foundation models, but the. The consumer doesn't really care. This is a great question because, number one, you may know that Salesforce also has a huge research team. We invented prompt engineering. The first prompt right here at Salesforce Research. Commercialized by others, of course, but hit the gong for the first prompt. Hit the gong. Gong. Thank you for your service. Thank you. Thank you. And listen, look, some customers want to use our models. Some customers.
Season and essentially just got paid for double dipping into an existing content. Yeah, that's pretty remarkable. Last question and then we'd love to have you back on. This was an awesome conversation. How, how excited do you think the average app developer, business owner in general should be about OpenAI launching ads? Because I know a number of businesses over the years that like we're basically be born born because a new ads platform like I have friends, company with hundreds of millions of dollars of revenue. They've told me like we would not be a big company if we didn't, if we didn't get to hyper scale on Facebook in the early days. And So I assume OpenAI will want to make the ads product cheap and performant early on to just drive this huge influx of volume. But I'm curious what you think the opportunity will be early. Everyone should be excited about OpenAI launching an ads platform. This is beneficial broadly for the economy. Ads are the driver of the Internet economy. Look, but for Facebook ads, but for conversion, optimized Facebook ads, the DTC category won't exist. E commerce much smaller than it was. There'll be far fewer small businesses in this country. Ads is a growth engine not just for E comm or not just for tech bros. For everything. It is the beating heart of the economy. Everyone should be excited about this. Look, I remember there was Chamath was complaining at one point that like I remember the exact number but he's like, you know, I invest in these companies and 70% of the money goes to Facebook ads. And I wrote a piece saying you should be thanking Facebook because if Facebook wasn't there to absorb this ad spend, those companies wouldn't exist for you to invest in. Right. This is great. This is. Yeah. Also the companies wouldn't spend the money if it wasn't driving results. It's not like you just, you don't, you don't. If a camp. If you spend 20, you know, 20 grand on a campaign and the CAC is too just high, it doesn't make sense. You just turn it off like that's the beautiful thing. Exactly, exactly. So I mean, I think, you know, there's a reason to be very optimistic broadly about the new opportunity that will be engendered by just by LLM empowered content engagement. Right. Or AI empowered content engagement. It doesn't have to be just LLMs. But I think like where people get mixed up is they think, well, you know, these ads need to be different. There's some different way to sort of integrate ads. Well no, the formats need to be unique and they need to feel native. But ads are ads, right? If you drive performance and you get conversion feedback and you get the feedback loop going and you deliver more value than you were paid, right? There's compounding there and it grows the economy. So, so I'm super excited. I call this commerce at the limit because I think what all and a lot of this technology is not being applied to the consumer facing side of things. It's getting applied to the back end, right? So that's what annoyed me about Facebook's last earnings or measures lost earnings because people are like, oh look at all these big projects. Gem Andromeda Lattice drove 3% improvement to click through rate or 2%. You know how massive that is at that scale, how meaningful. But the other thing people ignore is that compounds every time I spend a dollar and I get more back, what do I invest the next time? Not a dollar, right. I invest what I got back. So it's more and it compounds over time. I'm going to make more money and it's going to grow over time. So that 2 or 3% this quarter will also exist the next time that money's recycled. Right. And also those products are not static, they're not frozen in time. They will also continue to improve. So the thing is like these products and this is all on the back end, this is not seen by consumers but this is where a lot of the investment is going. And so people talk about the capex like the capex is delivering real returns. What are you complaining about? And the thing is like that is that's driving real returns but it's driving more spend, it's driving compounding spend and that's where a lot of the investments going. So I think there's, there's reasons to be excited across a number of service areas here and it's not just the consumer facing side. This is amazing. Thank you so much. I feel like I just got. Let's get you back on schedule next.
My job right now is to sell you, to come to Dreamforce, and also to look at our products, to cor. Sell you our core values. That's my job right now, right? That is my job. Yeah. And how quickly. How quickly can you clock if somebody's going to be great at sales? Does it take you, like, 20 seconds? 30 seconds? A minute? We've all seen. We just had a great conversation about Mr. Beast. Yeah. Would you say, on a scale of 1 to 10. The 10. One of the best communicators, most aggressive salespeople in the world. Where would you put them? One or ten? Ten. Ten. Yeah. And boom. That is important. And I think that that is. And you see, like, all of a sudden you're watching this thing. One of the games is happening. He's buried somebody underground for two years or something. And either digging the person up. Are you still alive? Whatever. And then just as they break, they go. And one more thing. I've got a chocolate bar to sell you. And this is the best chocolate bar you have ever tasted. Let me have the guys coming out of the ground now. Hey, will you try this chocolate? What do you think? Look, I've been underground for two years. This is the best chocolate I have ever had. I mean, he has to be one of the great communicators, but also one of the greatest salesmen I have ever seen. Is that valuable in the age of AI? 100%, absolutely. 100%, absolutely. All right, where do you get. Last question. Where do you get your energy? It's off the charts. It's incredible. I'm getting it from.
Making your. You're the final. Your viewers. You're the final Boss of Enterprise SaaS. I absolutely love it. What makes. Why, why do you believe that sales is still a great career path for young people? I personally believe it is. I believe it is the most valuable skill in the world. And. But I think a lot of people see all these AI agent for sales rep companies getting funded and they, they might think I don't want to go down that path because I don't want to go down the sales path because I think that's going to get automated away. Why do you. What's your kind of worldview there? Well, it's a great question. I mean, you know, at one level there's between 20 and 100 million people we actually counted them that we have not been able to call back since we started Salesforce 26 years ago. Between 20 and a hundred million people. We didn't call them back not because we didn't love them, not because we didn't like them. We didn't have a people. And you know, those are. That idea to be able to call everybody back. You know, how do you do it? Like you can't call every listener back who's already contacting you. You know that. So one thing is, yeah, you're gonna have a SDR or sales development representative. The ability to call everybody back. Okay. The ability to qualify to have a conversation but listen face to face sales or face to face communication like we're doing right here. Like, by the way, I'm. I don't. I'm not an AI. As far as I know, I'm here. I'm not a biological computer running an LLM. At least I hope I'm not. I feel like an AI would grab a water bottle and hit it like that. Very suspicious. What I want to say, I wanted to say this to you. I also just hired somewhere between three and 5,000 more salespeople. Wow. Okay. I'm growing my sales force. It'll be more than almost. It's. I think I'm going to try to get to 20,000 account executives this year. That doesn't include systems engineers, managers, the infrastructure sales engineers. Yeah. We have 80,000 employees at Salesforce. 80,000. And about a quarter of them are just people who are trained in our product who are designed to help you. We want to make sure that they can help you be your success. That is why at the end of the day I think sales leaders listening, asking good questions, empathizing with the customer, connecting deeply with the customer, having fun which is super important for us and for you. I know. I watch everything that you guys do. You love enjoying and having a good time. We want to do that, too. Isn't that what sales is really all about? And tonight, you know, look at the conference is happening, but across the street is the St. Regis Hotel. And last night I was there, and I'm coming down. It's a long day. It's a long two days. I haven't slept in two weeks. Coming down the escalator. Look at the bar is filled. And the bar is filled. And what is happening in the bar, do you think? And it's not our people who are in the bar. I looked. It was all the customers talking to each other and connecting, going more deeply, you know, having that human touch, you know, because, look, AI, we love AI, okay? But AI, it's not the same. It's not. AI doesn't have a soul. It's not that human connectivity. It's not, you know, at our depth, AI is not born. It's, you know, AI is made. So, you know, we. There's a humanity that still needs to be nurtured. Yeah. People want to buy from people, and. Grown people want to buy from people. Thank you. I mean, that is what it's all about. And that's why I think sales actually is. And you can see right here, my job right now is to sell you, to come to Dreamforce and also to look at our products to core. Sell you our core values. That's my job right now. Right? That is my job. Yeah. And how quickly. How quickly can you clock if somebody's gonna be great at sales? Does it take you, like, 20 seconds? 30 seconds? Minute. We'Ve all seen. We just had a great conversation about Mr. Beast. Yeah. Would you say on a scale of 1 to 10? The 10. One of the best communicators, most aggressive.
Value at the end of the day. And if they're not getting value, they'll. They'll churn or they'll find another. So. But how do you think about this conversation around seat based versus kind of value based pricing? What a great question. Look at. I just subscribed to Chat GPT. It's a seat based model, you know, that you probably pay 20 or 200. I just subscribed. Is that. What did you use that by. Did you use that, by the way, to. To make that picture with. With Sam? It kind of that. That picture you guys. Had, Brock. Because we were wearing like conference necklaces. Okay. And I said, take it off. And Chat GPT. Goes, oh no, no, no, no, no, no copyright infringement. And then. And then Gro said, great, I'll do that. And do you want me to put in the photo of the podcasters with you? I said, I'd like to do that also. But you might have seat based pricing like those are. That's a good example. Right? You might have all you can eat.
Your Japanese viewers are going to go up, say, we are X. We are X. Which is Yoshiki official. Yeah. Or whatever it is. Yes. We're. This. This is so fun. I. We got it. We got to do this more often. We were joking on the show earlier. We were referencing a post, you know, the whole, like, nothing ever happens investment meme. Right. Which is like. Which somebody was highlighting that OpenAI runs on both Salesforce and Slack. Thank God. Wanted to ask you. Does not. Does nothing. Does nothing ever happen? Nothing else matters. Nothing else matters. There you go. Yeah. You're still unforgiven. By the way, a super intelligence would use Salesforce.
What are you. Count us in for the next one. We'll be there. We'll be there. We'll be there. Next time. Super bowl of SA. Why? Why? Why? I mean, this is our 23rd Dream Force. You have 23 of them. I want an explanation. I want to know why. No, you. You guys are the number one podcast in the world. I want to know why you're not at Dreamforce. And I want to know your music. Ch. I want to know your music choice. I do love Metallica. I grew up on Metallica. Yes.
Modern software development, streamline issues, projects and product roadmaps. People were having a lot of conversations surrounding OpenAI's numbers from the Financial Times. Of course, they have 800 million weekly active users. Five percent of those are paying 40 million 13 billion in AU ARR, which implies a $325 annual ARPU, or $27 a month per paying user. And this post went pretty viral. Google would turn 800 million users into 32 billion of revenue. Is that just based on their current revenue and their current user base? Because you have to think like Google's like so much more heavily monetized that the gap between 13 billion and 32 billion. It's not as much as I feel like it should be based on how young ChatGPT is. They don't, like, you can't actually advertise on it yet. Like they don't have an ads product. And so it's pretty remarkable that they're monetizing at the rate that they are already. That was my takeaway from this because Google's been grinding the ads product for 20 years now. Yeah, yeah. And I think what Malte was meaning to say is that he follows it up. He says the tweet was meant to say, you can run this into a profitable business. Sure. I expect value per user is going to be higher than Google. Oh yeah, because OpenAI is making 13 billion in revenue, but burning like 20. Right. And so you add those together, you get about 33 billion in revenue to fully offset the burn, which is kind of like. Exactly. Google's monetization rate. And so if you monetize the time and the products, like, similarly, you get to break even pretty quickly. Yeah. And. OpenAI is projecting to get to 100 billion of revenue. The projections are crazy. We should go through those. Doom slide says extremely polite thread by epoch. OpenAI's projection.
Also, it looks like another era in. German maybe or something, I don't know. Did you hear that yesterday Paxos mistakenly minted 300 trillion of their stablecoins? I did not follow that at all. Exactly what happened? So stablecoin issuers like Paxos and Circle and Tether, they mint new stablecoins? They had an internal error apparently that and it sounds like it was a fat finger. They accidentally minted 300 trillion in their PyUSD. Their stablecoin. Did this crash their market or something? I wonder what happened here. No, I think they were Paxos immediately identified the air and burned the excess. It's resolved. Yeah, they said that there's no security breach, customer funds are safe. They've addressed the root cause. But how fat of a finger do you have to add? Like I imagine like 4x? Maybe the dev like fell asleep. It does feel like a cat on the keyboard. Yeah, it doesn't exactly. Yeah. Joey, slight rounding error. Don't ask about the dollar backing. Yeah, so people are basically pushing back and be like, okay, you were able to just create $300 trillion on chain. What was it backed by? Or is there some sort of disconnect? They bought 300 trillion. Glad they glad they figured it out. Maybe they bought 300 trillion of treasuries, you know. Well, before we move on, let me tell you about profound. Get your brand mentioned in chat GPT reach millions of consumers who are using AI to discover new products and brands.
Cancer and free education. And I don't want to have to make that call. Yeah, Tyler. I think that this development updates me in two important ways. One is that it's a vindication of Rune's concept of text is the universal interface. If you look at the actual model, it's an LLM. And basically, as I understand it, the data that went in is basically for every cell. They have just a bag of words that are just like every gene that's expressed, ranked from most expressed to least expressed. And so it's basically just like a text representation of a cell. Which is interesting because I think when a lot of people thought about, like, creating a model of the human body, you just go to like a 3D model or something. I don't know. I don't certainly go just a text. Yeah. I mean, I don't know if it's. Is it literally text or it's like tokens. Right. And then it's like, okay, DNA is already in tokens. It's just a string of numbers. Yeah, yeah. It's not that different. Yeah, I guess. Yeah. I mean, I don't think this, like, this is definitely vindication. I don't think this, like, was that big of, like, an update though. Right? This is. Anyone who's, like, sufficiently AGI pilled is like, yeah, obviously, scale things up, we're going to get big. Well, that's number two. So number one is that you don't need some sort of new data primitive. You can actually just use tokens. You can use text to understand a cell and then find a connection between biological systems based on just a pure textual representation of the cell and the interaction. So I think that's an interesting vindication for, like, text is a universal interface, not just in coding or text or knowledge retrieval or any of these other things. It's also applying to bio. You can take bio, transform it into text, and then do interesting things. What do you think? Yeah. If you look at AlphaFold, AlphaFold was not that crazy.
Of finance, the capital of capital. Big news. We covered it a little bit yesterday, but big news out of Google. Sometimes when something just hits the timeline and we're live and we're doing the show, it's hard to process how significant something is for sure. And I mean, what a funny. What a funny turn of events. In the meantime, while OpenAI is fighting for their life in the timeline against allegations of moving into adult content, Google is saying, hey, we cured cancer. We've done it. We cured cancer. Which way? Western lab. Yeah. Yes, exactly. They didn't actually cure cancer, but they made some progress, and it's definitely updating some people on what AI can do in bio, what AI can do in. In cancer research generally. It's very complex. I'm not an expert in bio or any of this stuff really. So I needed to use a call of duty metaphor. And so I read the piece. It comes from Sundar Pichai. He says, time is money, save both easy use, corporate cards, bill payments, accounting, and a whole lot more all in one place. Go to ram.com. just kidding, he said. An exciting milestone for AI and science. Our C2S scale 27B foundation model, built with Yale and based on gamma, generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells. Now, that is not in people. It's not in mice, it's not in rats. It's just in cells. And this is just one potential link between a drug and cancer cells. But it's very exciting, very promising. And so I was trying to figure out the appropriate analogy, and what I landed on was, of course, Call of Duty. And so in Call of Duty, there are a bunch of players on the map. Some of them are on your team, some of them are on the other. The opposing team. Exactly, the opposing team. You can think of them as cancer cells. When cells develop cancer, when tumors exist, they are on the enemy team. You gotta hunt them down. You gotta find them. The people that are hunting them down, those are the killer T cells. That's your immune system. Ideally, you want all the tumors, all the cancer cells to be really obvious to your immune system. So your immune system can go around and get a bunch of headshots. Double kills, 360. No scopes, of course. Of course. But it's hard because a lot of these cancer cells, a lot of these tumors, they exist in what Google puts them as. Like, it says that they're cold tumors. Basically, they're invisible to the body's immune system. You want to turn them hot, you Want them to light up on the mini map? How do you do that? You got to pop the UAV. And that's exactly how this works. So this particular C2S, which is cell to sentence model, ran a bunch of different correlations between different drugs and their effect on cells and found a drug that does just that. It's a conditional amplifier, meaning that it boosts the immune signal and it basically turns these cold tumors hot. And now this, this is not new. There's a lot of drugs that basically act as UAVs for the immune system that allow the immune system to target cancerous cells or tumors more effectively. But why this is special, why everyone's obsessed with this, why everyone's white pelling so hard, is because they used AI to do this. And of course this was in conjunction with scientists and then it was verified in a lab setting, in the real world setting. But it's still very exciting. Now, just to put it in perspective, Google found one drug that was correlated with this effect, which is great. There's something like 600 maybe FDA approved drug cancer indication pairs in the US alone, and only about 5% of drugs that even get submitted to the FDA actually get approved. And so the total number of discoveries like this, if you just think about it as a drug that helps fight cancer or helps identify cancer, if you think about the pool of those humans have discovered probably tens of thousands, maybe hundreds of thousands of those, narrowed those down, sent those to the FDA, the FDA has approved maybe 5% of those and we get 600 that are on the market. And then of course, some of those are more effective than others, some of those have less severe side effects and so they become more popular. Some of them are just good at advertising, probably, or good at the cost benefit ratio. So they're very effective to a lot of people and it's really cheap. That probably is a blockbuster cancer drug, something that's really, really, really effective, but only in a very odd population. It's really expensive, probably less of a blockbuster cancer drug. But what matters here is the slope more than the Y intercept. And so right now, if you just look at the scoreboard in your Call of Duty world, AI got one point on the board and humanity has like tens of thousands of these discoveries. Basically, if you that's a very rough estimate. But it's like, it's a lot. We've discovered a lot of cancer. And this is for one very real reason, which is that people have been desperately hoping that AI systems would be able to discover novel cures for cancer. Right. This has been something that if you look at anybody at any of the big labs over the years, they probably have at least one sound bite where they're talking about. Maybe it'll be potential. My question is, does OpenAI really have time to even compete on this front? Right. It's very easy for Google, with hundreds of billions of dollars of revenue, to dedicate resources.