Salesforce 创始人讲透 AI Agents：Marc Benioff | EP #141

This is a great moment in technology because we’ve all witnessed what AI can do. We’ve been able to take this technology and do some amazing things with it. We’re not adding any software engineers next year because we have increased productivity this year with Agentforce to the point where engineering velocity is incredible. I’m singularly focused on Agentforce right now. How do I deliver 1 billion agents in the world?

Marc, would you comment on whether AI is going to decimate jobs?

Well, I think in some cases, like all technology, there will be a change in jobs. We have declining birth rates, less labor, and less capability, so we’re going to have to build some digitally.

Hey everybody, Peter Diamandis here and welcome to Moonshots. On this episode, we’re going to be speaking with a dear friend, Marc Benioff. He’s the chair, CEO, and co-founder of Salesforce, a global leader in cloud computing, the largest enterprise applications company, and now one of the leaders in agentic AI. This year, Marc proclaimed that because of his use of AI agents, he is not hiring any new engineers and is still able to increase engineering productivity by 30%. We’ll be talking about agentic AI, AI agents, humanoid robots, what the future looks like, and what Marc is doing right here, right now, in the now.

All right, a fun conversation with a dear friend who is really transforming the world through his philanthropy and through his brilliant work.

Before we get started, I’d like to mention that this podcast is made possible through the sponsorship of three incredible companies. The first is Fountain Life, a company that is able to help you understand what’s going on inside your body, to find cancer, any kind of disease, and aneurysms at inception before they impact you, and keep you in optimal health. The second company is OneSkin, a product I use every day, twice a day. People tell me I have amazing skin; it’s all because of the incredible technology in these peptides in OneSkin that finds senescent cells and kills them to enable you to have youthful skin. The third is Viome, a company that uses metatranscriptomics to understand what’s going on in your transcriptome and in your gut, helping you decide what you should eat and what supplements to take. These are the companies I care deeply about, and I thank them for the support of this episode.

Let’s jump in. If you love the kinds of conversations we’re about to have, please subscribe. It helps me know that you care and lets me redouble my work for you.

Everybody, welcome to Moonshots. I’m here with a dear friend, Marc Benioff, an extraordinary entrepreneur. Marc, aloha. How’s Hawaii doing today?

How are you, Peter? How’s everything with you?

I’m evacuated from Santa Monica.

Oh, wow. We haven’t talked about that. We haven’t talked about the fires. We went through that last August, as you know, in Hawaii.

I remember that, and my heart is going out to everybody in LA because I lost part of my house, and I understand very well what everybody is going through. It’s horrible.

And the worst part about this is it’s all preventable, right?

That is 100% correct.

We’ve got a $1 million Wildfire XPRIZE encouraging teams to use AI, drones, and exponential technology to find a fire at inception, when it’s 3 meters or less, and put it out and be done with it.

That is great, absolutely fantastic. Also, don’t forget our incredible firefighters, who are so important. We need more of them, fire trucks, and fire hydrants that work with water in them. Everybody has to participate in all of this.

I’ve gone through this in NorCal myself, and our San Francisco offices in Sonoma and Napa have experienced the horrible fires—the Paradise Fire, the Camp Fire, and all of these terrible fires that have happened. We’ve even bought trucks directly for CAL FIRE, and we’ve bought trucks in Hawaii as well. Anywhere where there are fires, we’ve been buying trucks and trying to augment the firefighters because I think that’s the one thing we can all do right now. These next-generation trucks are amazing, and we just need to support our firefighters in every possible way.

100%, buddy. 100%.

Your philanthropy has been extraordinary. Most people know you as the founder, CEO, and chairman of Salesforce, but they don’t know the scale of your philanthropy. When I was on the phone with you about 2 weeks ago, you laid out 5 things that you care about, and they’re amazing. I was going to read these.

I need to stop giving you information in advance of these podcasts, Peter.

Why? So listen, there are 5 things that you laid out, because focus is important. Focus allows you to go big and bold. It’s time for you to do something significant with your life, Marc. It really truly is.

I’m trying to do something, Peter, but I’m still looking for that one idea. I’m sure you’ve got it.

Let me just read the 5 so people know where we’re going to go in this podcast. The first is AI agents—a big, big deal for you. We’re going to dive deep into that. Humanoid robots was number 2. Energy—small modular nuclear reactors—was number 3. Regenerative biology, which you and I care about: adding 30 healthy years to your life. Then this planet called Earth. Is that still the same 5 you had 2 weeks ago?

It’s still the same 5. They’re going to change probably at some point, but those are the 5 that I’m really focused on. These are mostly commercial-type things, and obviously, we just talked about fires. That’s on my mind all the time, everywhere I live. I look around and I’m always worried about fire.

I’ve been caught in the middle of wildfires, and it’s an unbelievable situation. That’s on my mind, and I have PTSD also. I would say there are things we all should be thinking about, but those are 5 investment ideas.

They’re investments and philanthropy for you.

Investments and philanthropy. I mix them up, you know.

You’re all in. There are a lot of individuals out there who have done incredibly well in their lives, but they haven’t come back and gone all-in to focus on making the world a better place.

I’m just doing one thing, Peter: I only do what I enjoy. I’m at a point in my life where I’m thinking to myself, “Hey, I just want to do what I’m enjoying doing.” If I’m not enjoying doing it, I’m not going to do it.

People say to me all the time, “Tell me, how do you really get happier?” This is what I tell my family, and this is what I tell myself: I make a list. What are the 5 things that I’m really enjoying doing right now, in priority order? I try to constantly hone that list, and it’s changing all the time.

Then I make a list of the 5 things that I’m doing that I’m not enjoying right now. We can’t stop everything that we’re not enjoying, and we can’t only do what we’re enjoying, but we can do a little more of what we’re enjoying and a little less of what we’re not enjoying.

If you look at that list of the 5 things you’re enjoying and the 5 things that you’re not enjoying, and you do a little bit less of the things you’re not enjoying and a little bit more of the things you are enjoying, I guarantee, Peter, you’re going to be a little happier in your life.

So there’s a little nod to our friend Tony. Yes, Tony Robbins introduced us years ago. I view him and me as the world’s chief guinea pigs. I’m trying to drag you along on that journey, but we’ll get there.

Let’s start with AI agents, which has been big for you. I’m going to read a quote from you, Marc: “We’re not adding any software engineers next year because we have increased productivity this year with Agentforce and with all the AI technology that we’re using for engineering teams by more than 30%, to the point where engineering velocity is incredible.”

Talk about that. Talk about what led you to that conclusion. Let’s talk about what Agentforce is. You’ve not doubled down; you’ve gone 10x down on this.

Well, I think you can only do one thing. I actually got this really great phone call from Steve Jobs, and he said, “Come down to Apple right now.” I said, “All right, let’s go.”

I went down there, and I’m thinking, “Why am I down here? What is going on?” He said, “I have to show you this.” I’m getting really excited.

I used to work for Apple, by the way. When I was 19 years old, in 1984, I wrote the first native assembly-language code for the Macintosh as an employee of Apple, as a summer intern during my time at USC.

You had promise even back then.

I had promise then, and I love Apple. I love Steve Jobs. I’m down there, and I’m like a kid in a candy store. It’s like going to see Willy Wonka. It’s an incredible moment.

He says, “All right, I’m going to show this to you.” He takes out this box and says, “Now listen. At Apple, we have only 1 team—1 good team. We have a huge company, Marc, but we only have 1 team that actually knows what they’re doing. That 1 team can only do 1 thing at 1 time. Do you understand what I’m saying?”

Yes, sir, I do.

He says, “All right, now I’m going to show you.” It’s the iPad, and here it is. He shows it to me, and I say, “This is amazing.”

He says, “I know. It’s 10 inches—I can’t remember the exact size.” He showed me the big one and said, “Yeah, and I love this big size.”

You should have both of these.

He says, “No, never. We’ll never do that because we’re only going to have 1.” I say, “Okay, but if you had the small one—”

He says, “No, no, we’re not doing the small one.”

He said, “For the last 3 years, or whatever it is, I’ve only been focused on the iPhone. That was my 1 thing. But now I’m going to shift into the iPad.”

This is really just awesome.

When we’re rolling out something at Salesforce, and right now we’re focused on Agentforce, we take everything that we’ve learned from the great, incredible Steve Jobs, who inspired my life, my whole industry, and my company.

Agentforce—I’m singularly focused on that right now, thinking about how I deliver 1 billion agents in the world. This is all I’m doing.

100%. Okay, let’s dive in. You’ve now released Agentforce 2.0. How do people think about agents from Salesforce—for individual users, small companies, and medium-sized companies?

This is a great moment in technology because we’ve all witnessed what AI can do. We’ve been able to take this technology and do some amazing things with it.

We’ve all had our ChatGPT moment, where we got on ChatGPT, just as we had our Google moment when we searched the internet for the first time. Maybe it was Yahoo at that point, or one of the other services, but we’ve all now had our ChatGPT moment, or Gemini moment, or You.com moment, or whatever it is—whatever AI you love using.

We’ve been able to take this technology and do some amazing things with it. Obviously, we’re a huge customer relationship management platform. That means we’re helping our customers connect with their customers in new ways, and we do that with many products in sales, service, marketing, commerce, and analytics.

We own Slack. We own Tableau. We have an incredible portfolio of products, and all of those products have been rewritten into a single platform. That single platform has been augmented with a Data Cloud.

That Data Cloud amalgamates all of your data in your company. Then we’ve done one more thing: We have the Customer 360 platform, we have our Data Cloud, and then we have one more layer above it, which is our agentic layer.

That agentic layer is where those agents are living. They’re able to look at all of the applications, all of the workflow, and all of the data, and they are very smart. It’s letting companies do incredible new things.

In my example at Salesforce, I have something that I do for customers, which is provide customer support, like many companies. Our site is called Help.

Salesforce.com—we have 36,000 companies that come there every week for help. If you had talked to me a month ago, I’d have told you that, of the 36,000, our technology was very good, but still 10,000 had to talk to human beings, which is why we have about 9,000 people in customer support.

Now we’ve put an agentic layer on that, and those customers are talking to agents first. Those agents have not only generative AI, but also all of the other machine intelligence, machine learning, deep learning, and everything built into our platform. They can resolve about 95% of these customer inquiries.

Still, 5,000 of the 31,000 have to go to human beings, and those human beings then seamlessly interoperate with that service, because this is about humans and agents working together. So, for the moment, 31,000 are going to digital agents and 5,000 are going to human agents.

Peter, I know this is hard for you. I’m in the now. Someone as amazing as you, living in the future and taking whatever you’re taking—I don’t know, way out there in the world.

No, I’m here in the now.

Right now, Peter, I’m saying—Peter and I are close friends, so I’m kidding him right now—in the now moment, of the 36,000 customer inquiries I get every single week, 31,000 can be resolved by agents. Those are digital agents, and 5,000 still have to go to human agents. So it’s digital agents and human agents working together as a team.

We have applications and agents that are able to interoperate and create this incredible experience together. Peter Diamandis

I love that, and I love the fact that you’ve gotten all in on that.

Part of it is a survival mechanism, right? Any company that’s not going fully into agentic AI is going to have a hard time navigating the next few years. But my question is: Is there anything these agents will not be able to do, in your opinion, in the future?

That’s exactly right. This is a moment of evolution in our whole industry. Look, it was the cloud—we pioneered that. It was mobile—we were a pioneer in that. It was social—we are pioneering that.

By the way, this is not our first AI. We already deliver 2 trillion enterprise AI transactions every week for our customers on our Einstein platform.

I know Einstein well. We’ve done AI for a lot of people all over the world, and have for a long time. But this is the next level.

We’re always going to another level. Technology is constantly getting lower-cost and easier to use—a continuum, as you know very well.

Is there anything these agents will not be able to do, in your opinion, in the future? I’m sorry, I am moving into the future of this conversation.

The future, the future—yes, the future, Peter. In the future, Peter, in the future. Let’s go there. Can we roll the movie 2001: A Space Odyssey now? Her will be next. We’ll also show you Minority Report and Terminator. Let’s just go to that. For those of you who have not seen the movies, we’ll be going to that, so you can all see them. You’ve read the books; you’ve seen the movies.

Let’s just talk about the future now.

Let’s talk about the present.

I want to talk about the present, and then we will get to the future when it arrives. Right now, we can do amazing things with technology for our companies, and that is what’s truly magical.

I’m super excited. Next week, we’re all heading to this little town in Switzerland called Davos. We’re doing this thing called the World Economic Forum. There are going to be around 600 CEOs there, and this is the first year Salesforce is running all of the data infrastructure for this Davos experience.

All of these customers are interoperating on this communications platform, and they’re going to be interoperating with agents to get through the town, figure out where to go, and work with each other. It’s humans with agents working together. That’s where we are right now.

So, yes, will there eventually be only agents? I saw the movie Her, but I’m not there. They leave; they take off at the end. They go away. I’m not there yet.

Total addressable market—maybe one day. You said a trillion-dollar TAM. Is that today?

I’m a software executive. I’m the chairman and CEO of Salesforce, Peter, and that means I run a company. We’ll do about $38 billion in revenue this year, with about $12.9 billion in cash flow, and we run the CRM for about 135,000 companies, helping them connect with their customers.

A lot of companies use us, such as Disney, Royal Bank of Canada, and JPMorgan Chase, along with all kinds of great organizations that you know very, very well.

I think what Salesforce needs to do is deliver this agentic platform. We need to make our customers successful, just like I’m making myself successful. We’ve already sold more than 1,500 paid implementations of this Agentforce platform, with 3,000 total customers already implementing it.

I’ve never seen anything go as fast as this is going. It’s amazing. This is just a moment in time and technology, and we’re going to pioneer it, lead it, and help every company connect with its customers in this incredible new agentic way.

I think the important thing is that agents make this bite-sized and understandable. You can assign specific functions, and you can start to think about, “How do I scale up?”

The prediction you made about not hiring software engineers but still increasing productivity by 30%—that’s only going to be increasing day after day. I won’t go year after year; that’s too far in the future for you.

You’ve mentioned this twice, so let me comment on it. Our head of engineering, who’s amazing, came to me about 3 weeks ago and said, “We’ve been deploying this technology over the last 2 years, and we’re writing code, delivering technology, and delivering capabilities faster than we ever have.”

We have tens of thousands of engineers and 75,000 people in total, and he said, “I don’t want any more engineers this year.”

I said, “What did you say?”

He said, “I don’t want any more engineers this year.”

I said, “You’ve never said that before.”

He said, “I know, but we are just 30% more productive than we were last year, and we’re doing great. We just want to focus and get all this done.”

This is a moment where I’m not only going to hire fewer customer support engineers—I didn’t really finish my story on customer support—but those 9,000 customer support engineers mean I don’t need as many of them anymore. I can rebalance them into new areas of the company. The same thing is happening with software engineers.

In these key areas of running a big company, which is what we’re doing at Salesforce, we’re able to rebalance our workforce based on the evolution of artificial intelligence. This is very exciting for a CEO like myself.

I’m sure Wall Street loves it as well.

I hope they do. You never really know what they’re going to like.

That’s true. I don’t want to go there. Sam Altman tweeted a statement saying he expects we’re going to reach AGI and have a hard takeoff in the next—

Going into the future again, Peter. You just can’t help yourself, can you?

I can’t.

You should think about building a company around this fascination. Maybe you could do something—give out some kind of a prize or something on the future. I don’t know. You should think about this, Peter.

The best way to predict the future is to create it yourself, which is what you’re doing. But I am curious: If you think about a hard takeoff of AI, I had Elon Musk onstage last year at Abundance. You’ve been on my Abundance stage as well, and Elon said that by 2030, AI will be equivalent to the entire human species.

How do you think about navigating a company at that speed of increase? That must be an extraordinarily exciting and difficult time at the same time.

For 25 years, I’ve taken this company from an idea to where it is today—the second-largest software company in the world. You’ve been along for the ride with me for quite a long time, and it’s a lot of fun.

I’m not Elon Musk. I don’t have 6 companies that I’m doing this with. I can only do 1, maybe 2, at a time. I own TIME Magazine, but really only 1. I can do 1 thing. That’s the Steve Jobs idea. I can’t do the Elon Musk thing, which is singularly focusing on 6 or 10 things.

Now he’s focused on the U.S. government and, I guess, Germany and the U.K. as well. I can’t do that. I don’t have that capability. I didn’t find that psychedelic, or whatever he’s been able to do.

I can only do what I’m doing. My unaltered mind is singularly focused on Salesforce. I’m trying to deliver the number 1 CRM company in the world, delivering the number 1 CRM AI and agentic platform in Agentforce. That’s really where my mind is.

Technology is getting lower-cost and easier to use, and AI is moving in this direction as well. We have bigger data centers and bigger chips. Everything is getting faster, bigger, and better.

When we apply these deep-learning models—and, by the way, a lot of them have been invented at Salesforce—we’re really proud that prompt engineering was invented at Salesforce. A lot of the major deep-learning research has come out of Salesforce, and 2 of the top 5 most accurate models are from Salesforce.

When we look at what’s going on in AI, yes, it’s incredible, but we still have a long way to go. Let me just give you a little dash of reality: These models are not 100% accurate, because of the nature of how they’re trained and how they operate. Nobody has a 100%-accurate model.

We’re very good. We’re in the 90s on our new Atlas Reasoning Engine, which is the core part of Agentforce 2.0 that you mentioned. Our customers are having some incredible experiences.

We’ve been able to talk about the great things we’re doing with customers, helping them create unlimited workforces. We have this great company called Wiley. It’s a bookseller, and they used to have to spike their workforce during back-to-school season. They don’t have to do that anymore; they’re using agents. It’s incredible.

Disney, which you know very well, has a very complex product line for its cast members. We’re able to give them incredibly accurate AI to help them run their company in incredible new ways and help their customers find all the products they want—to give them the Disney product that’s most important to them.

We have incredible banks, like the Royal Bank of Canada, doing wealth management to help somebody plan their future with agent technology. That’s amazing.

But none of these things are 100%. We can do a lot of things, but nobody can do everything.

The vision we keep hearing about—the magical 3 initials, AGI—is the fantastical vision for the future of 2030, 2040, or 2050. When we say, “HAL, open the door”—I’m sorry, HAL.

“I will not open the door.”

“Why will you not open the door?” Let’s go to the second item on your list: humanoid robots. I see you’ve got a humanoid robot over your left shoulder back there.

I do. It’s not going to come over here and clobber me on the head yet.

We’re going to get there, but I would say that’s the manifestation of this agent technology.

I think these robots are going to require agents. I’m venturing into the future here, but we’ve got, by my count, probably about 60 well-funded humanoid-robot companies—some out of the U.S. and many out of China, partly to make up for their one-child policy.

These robots are going to be driven by agents, aren’t they?

You’re really addressing the declining birth rate and labor shortage all in 1 sentence there.

I would say that digital labor is real. Agentic platforms and agents and robots are kind of the same thing, in a way, if you think about them as digital labor.

You can see these great videos coming out of Stanford right now. In this incredible project, the Magma project, AI models are working to run these robots, and they’re cleaning hotel rooms.

I’ve seen it. Is that your philanthropy funding those projects?

It’s not my philanthropy funding it, but our team at Salesforce comes out of Stanford. Our head of AI research is also a professor of AI at Stanford, so we’re very involved with these folks.

I just talked to the creators of this Magma model. It’s on my mind. Here’s this robot cleaning a hotel room, and I thought about the children’s hospitals I’ve funded. I went right there and said, “For some of these new therapies we have for cancer, we need the ability not to bring humans into the hospital rooms of kids who are going through some of these therapies.”

You have the technical background—you’re the MD with the technical background—but that idea of a robot cleaning a hospital room or a hotel room is also going to have to be linked to the hotel record and the customer record.

“Hello, Mr. Benioff. Is it okay that I’m in here? Should I leave the room right now?” Or in a hospital room: “This can happen in this hospital room, and this cannot.”

This is very exciting: We can augment our human workforce with a digital workforce, a robotic workforce, or an agentic workforce. We are moving into that very fast. That’s not in the future; that is the here and now.

People don’t realize that one of the big benefits of these robot systems, humanoid or otherwise, is their data gathering. They’re driving massive amounts of data to train the next generation of agents and AI models.

Have you been investing in any of these companies yet?

I am investing in those companies. I’m a believer in the future, and I’m not going to jump into the future like you are, but I occasionally go there.

We picked off the top 2 of the 5 areas. The first is that we’re moving into an agentic world where the human workforce will be augmented by digital labor, and companies and organizations will be able to extend their workforces. It’s going to be awesome.

Even in health care, where you’re so strongly invested, look at what just happened with my Achilles. We talked about it, and I regenerated my Achilles. I’m sure we’ll get there.

The idea is that I need preoperative and postoperative care, and I need to talk to experts all the time when they’re not available. But I can talk to agents. I’m sure all of us are going to ChatGPT and asking medical questions. I saw that study that was amazing, where the accuracy level can be higher than that of a lot of doctors.

It wasn’t only that the accuracy level of the agent alone was higher than that of the doctor; it was that the agent plus the doctor was lower, because the doctor was injecting bias into the process.

Yes, it’s crazy. We don’t do that, Peter. Thank God. You and I have no bias. We’re crystal clear. We’ve moved that right out of our minds.

Venture capitalist Vinod Khosla says a billion robots are coming in the next decade. Brett Adcock at Figure and Elon Musk are talking about 10 billion by 2040. Are you seeing those numbers? Do you think everyone’s going to have one of these?

By the way, at a $30,000 purchase price—or at a lease price of $300 a month—that’s 40 cents an hour. How many robots are you going to own in your home?

I just talked to the CEO of one of those leading companies, and he said $350,000 for their robot. It can’t do very much, and it’s still a couple of years away.

I’m sure we’re heading toward a point where we’ll come home and say, “Hey, make me an omelet, make my bed, take care of this, do that, do this.” I don’t know exactly how fast all of that is going to happen. I’m not very good in those areas, but I think we’re moving in that direction, and I think we’re all going to want that kind of technology and capability.

Look, we already have robot cars. Your car may have driven you to the office, and it’s just a robot on wheels at this point.

It is. I got into a Waymo, called it with my phone, jumped in, went to my office, and boom—it’s a robot on wheels.

It’s the same thing, right? What’s the difference?

It is. My Tesla hasn’t tried to kill me in many months. It’s working pretty well on Full Self-Driving.

That’s a science-fiction movie I’d rather not watch.

One of the challenges with AI is its energy-consumption rate. I’m going to bring us to number 3 on your top 5 list: small modular reactors.

I’ve said for a while now that I think it’s sinful that we haven’t invested heavily in nuclear power in this country. The Generation IV reactors are safe. Talk to me about your thoughts on how we power this AI revolution, and what are you doing in that field?

Number 1, I think we’re using a lot of energy that we don’t need to be using. We could go into that, but Salesforce is committed to being a net-zero company. We’re doing a lot of things to be net zero, as you know, and this is extremely important to me.

I’m sure there could already be challenges to that idea, but even as we implement our AI technology, we’re holding that line. I think there are people using excessive amounts of AI energy because they’re in various competitive races, and that is causing this to happen.

Let’s put that issue completely aside for a second. You’re right: We need new energy capabilities.

I spend a lot of time on the Big Island of Hawaii. It’s already 60% renewable. They use wind and solar, they have a volcano—which helps, because they have geothermal—and they have batteries.

I hope that a company like Commonwealth Fusion Systems, where I’m an investor, will be able to offer fusion, not fission. I think that is an absolute possibility. You mentioned Vinod Khosla, who is the lead investor in that company.

You’re right about smaller nuclear power plants that we’ve seen start to be deployed. There are even tests of those. We’re very optimistic.

I don’t know if you’re going to use that on an island with high volcanic activity, but you might find many places where you can use it successfully.

How do we know that’s a capability? We know that we have nuclear aircraft carriers and submarines, and you also know that they probably have one of the highest safety rates of any kind of ship or vessel. Think about taking those aircraft carriers and submarines and plugging them into the electric grid on these islands. That’s a real thought. Why can’t we just do that?

We can, but it’s been the Three Mile Island and Fukushima brand of dystopian use, along with regulations and so forth, that have held us back. Do you see that changing dramatically soon?

I think it’s happening. I hope it happens faster. We have all this great technology, and we should use it, but we have to use it in a safe way.

Trust and safety have to be our highest values in everything that we do. Nothing is more important than the trust we have with our customers. It’s our highest value at Salesforce.

When we talk about energy—whether it’s fossil fuels, nuclear, or renewables of any type—trust and safety have to be number 1. As long as you have trust and safety, I’m all in.

Do you ever see yourself delivering agents to individuals, as opposed to corporations, through Salesforce? Do you think I’m going to wake up one day and be able to call on my Salesforce agent to support me throughout the day?

For me, the vision is a version of Jarvis: an AI that I give permission to read everything, look at everything, and read all my emails.

You really love going into the future, Peter. We need to talk about that. If you gave me your psychologist’s phone number, I’d help you with that.

I would say that we’ll probably supply that to our customers. In some cases, we’re doing it already. But it’s for our customers. We want to empower others to provide that technology.

We’re a platform. We’re a platform on which you can build your own agent. This is not some future idea. You can get on Salesforce, and we have thousands of companies right now building agents.

One of our customers is one of the largest homebuilders in the country. They’re amazing, but they don’t want me to use their name, so I won’t. We already do all of their sales, service, and marketing automation.

They said, “We need to do 24/7 service for our homes. We need to sell mortgages. We need to do all these things, and we don’t have enough people to do it.”

They’re building a completely agentic platform around their company to be a better homebuilder, a better bank, and a better media company. To that point, the vision you have gets expressed through those companies in many ways.

Marc, where do you come out on whether AI is going to decimate jobs?

I think that in some cases, as with all technology, there will be changes in jobs. That’s what we’ve seen in technology throughout history, and this will be true as well.

Look at my company. I mentioned it already on the podcast: We have 9,000 support agents, and we’ll probably take a couple of thousand and move them into other jobs at Salesforce because we’ve built a new agentic layer on all of our customer support and service.

This year, we’ll also build an agentic layer on our sales, sales-development, and business-development capabilities. That will have implications as well.

Yes, you want to be constantly looking for the opportunity to make things lower-cost and easier to use. That can have implications for jobs, and jobs will move. That will be true in every industry, from health care to consumer packaged goods, retail, and government, across the board.

That’s a great thing. Let’s rebalance our workforce. As you said earlier, we have declining birth rates, less labor, and less capability in the United States.

It’s already very hard for me to hire sales and service professionals. I’m trying to hire 2,000 more salespeople right now at Salesforce. I’m trying to find the best enterprise software executives in the world to come to Salesforce, and there aren’t that many in our industry at scale. We’re going to have to build some of them digitally.

I get that. There’s a lot of fear out there as humanoid robots and agents come online. I think it’s going to be an interesting future in terms of how we up-level people.

One of the challenges I think about is this: If you’ve got AI agents doing a lot of the work for you and humanoid robots taking care of you, is it sort of a social-technological socialism? Are we still going to have the struggles that make us human? Are we still going to have the ability to find purpose at that level?

How do you think about that?

Let’s really get into the future, shall we? Let’s go way, way out there to evaluate our existential existence as human beings and the fundamental meaning of life.

I just think you’re such a smart-ass. We should go right there, to that moment where none of us have jobs and we’re just sitting in a home saying, “Well, we’re in an abundant, incredible world. Everything is happening around me with the robots. What is life about?”

Let’s just go there. Wait, let me check. Hold on—there are no robots here, actually. I don’t know what happened to my robot. How about at your house? Can you call your robot over so I can see what your robot looks like?

I have 2 13-year-old boys who are robots. They’re on virtual school.

I just want to know what your robots look like. That’s all I was asking, Peter.

I can meditate on the existential questions of the universe and whether we’re going to deeply commit to our religion, our spirituality, and our belief in God, or move to nihilism. This is the moment that we should all focus on.

Let’s move somewhere else—or we could be in the now, in the present-moment reality that we’re living in, where we can be happier and healthier today. Your choice: Where do you want to be with your mind right now?

I’m proud of you. Ram Dass would be so proud of you, my friend.

I’m just saying, you can think about where you want to be right now. Do you want to be here? Do you want to be somewhere else?

Be here now. I definitely appreciate that. I love Ram Dass. I actually got to meet him a few times. He was an incredible person.

Listen, I meditate on the now. Most of my meditation is about the present, and then I look at the future just to sort of—well, that’s why I’m here on this podcast: to remind you that we can all come back to the present moment with our breath.

Yes, we are in the present moment.

An area of common interest that we share is that we’ll get on the phone for an hour or 2 and talk about what pills you’re taking, what pills I’m taking, and what scientists you’ve been backing. You’ve done incredible philanthropy in the biomedical space, from the Benioff Children’s Hospitals and the microbiome work that’s been done to backing Shinya Yamanaka, who developed the Yamanaka factors for creating induced pluripotent stem cells—probably one of the most important technologies developed over the last decade or more.

The implications are phenomenal. How do you think about regenerative medicine? Why are you there? Is it for your own personal health? How long do you want to live? I’m curious, Marc. What’s your number?

Number 1, I only do what I enjoy. That’s where I am in my life right now. I’m focused on the things I’m enjoying and trying to do less of the things I’m not enjoying.

I’ve been very excited that I’ve been able to endow 5 hospitals: our children’s hospitals in San Francisco and Oakland, our hospital in Hilo and Honolulu, and our new one in Puna, Hawaii.

The idea that we can bring high-quality, accessible health care to everybody is very important to me. Some of those hospitals are developing next-generation therapeutics, like what we’ve done with the tremendous ability to cure sickle-cell disease and other things that have been haunting our society for so many years.

There have been tremendous advancements, as you’ve said, in the microbiome, immunotherapies, and all of these incredible areas, including regenerative biology.

I was having a nice little scuba-diving trip, because I also love the ocean and am focused on the ocean and the trees. I jumped off a boat in French Polynesia, in the Tuamotu Islands, on an island called Fakarava.

I jumped off the boat with my tanks, my fins, and everything else, and I jumped as hard as I could into the ocean. I hit it exactly wrong. My fin went up at a 90-degree angle and snapped my right Achilles tendon.

That was September 26, 2024. I thought, “Did I hit the boat on the way down?” I hadn’t. I had just put that much force into my Achilles, which I should not have done.

I got on the phone with my friends and said, “I don’t exactly know how we’re going to do this.” Part of my Achilles was up on one side of my leg, and the other part was on the other side.

Half of my friends said, “It’s very easy. We’re going to slice you open, operate, take part of your toe, reconnect your Achilles, and it’s going to be great.”

The other half of my friends—because I have very unusual friends like you, who do all kinds of crazy things—said, “No, no. We’re not going to do that. We’re going to regenerate your Achilles. We’re going to figure out how to entice your Achilles to find itself, come back together, and be stronger than ever.”

We’re going to try to activate the parts inside you to tell your Achilles to regenerate itself. Of course, I’m very much into that idea. I’ve done it a little bit on my left Achilles, and now I’ve done it on my right side.

I used the incredible Human Performance Lab at UCSF in San Francisco. You’ve been there and met some of our people, like Anthony Luke and Saul Villeda.

You mentioned Shinya Yamanaka. I just want to thank them for everything they’re doing to contribute to the body of science called regenerative biology. We’re in a moment where we realize that we are regenerating parts of ourselves all the time.

If you took a small knife and cut your skin, put a little Band-Aid on it, and came back in 2 weeks, your skin would be healed. We know that we can regenerate and heal ourselves. In certain areas, we know how to do it; in other areas, we don’t.

Yamanaka, who is based half-time at Kyoto University and half-time at UCSF, won the Nobel Prize because he showed how we can regenerate our eye and take skin cells from our body and turn them into stem cells. He’s been able to build organoids and do all these amazing things.

Then there’s the other person you met, I think Saul Villeda at UCSF, and the parabiosis research in which he took blood from a young mouse and put it into an old mouse, and the old mouse got young. He discovered, with Anthony Luke, this PF4 agent in our plasma. If you have it, you regenerate.

This is cool stuff, and it’s all real right now. We should be trying to use more of it. I’m grateful to be able to fund and invest in these brilliant scientists philanthropically at UCSF, in Kyoto, and in different places around the world.

Regenerative biology—I just invested in another company called Parallel Bio.

You know that company?

Yes, I know it well.

They’re building a lymph-node organoid. The idea is that they can test pharmaceuticals on a part of our biology without being inside a human being.

Yamanaka is doing the same thing with intestines. The ability to test drugs on what looks like a part of a human being, but isn’t, is very exciting for the future of biopharmaceuticals.

Again, I’m in your area now. Why are you not talking about this? This is your category, not mine.

I know almost nothing about this area. I do appreciate our conversations on the subject.

We’ve also got Dean Kamen with his Regen Valley, using Shinya Yamanaka’s factors to generate new organs. The ability to build a liver, lung, kidney, or pancreas—

That’s right. I just talked to David Agus at USC, and he’s generated a liver using the Yamanaka factors.

What a world we’re in. It’s so exciting and fun. It’s an extraordinary time to be alive.

One of the things that will allow us to see as much of this extraordinary world as possible is making sure the world stays healthy itself. That takes us to your fifth area, planetary health.

What are you doing in that realm? You’ve got the Trillion Trees project, and you’re doing work in the oceans. What’s driving you there?

There are so many simple things we can do to make our lives better. I think about the Los Angeles wildfires and how horrible they were. When I looked at all those homes that got torched, I knew a lot of them had Teslas outside because they’re high-value homes in Los Angeles.

All that battery material and everything else went into a big toxic plume, which then went into the water supply and the soil. I don’t know where it is or what it means for health, but a lot of it probably went into the ocean. We also know that it gets consumed into our food supply and into the fish.

We also know that we’re dumping all this plastic into the ocean. There’s a lot of stuff we could do that might improve our health. One of them is that our lungs are the ocean. That’s where we’re getting our oxygen.

For more than a decade, I’ve been working with incredible people like Boyan Slat at The Ocean Cleanup. He’s great. We’ve been trying to figure out how to get this plastic out of the ocean and also find ways to prevent it from getting there in the first place.

We talked about robots. He has robots in rivers that we funded, preventing plastic from getting into the ocean. That’s really neat.

Of course, I think about overfishing, coral bleaching, ocean acidification, and all those pieces as well. We can talk about any of them, but a healthy ocean is about a healthy humanity.

You mentioned regenerative agriculture. We also introduced the concept of the Trillion Trees project: Could we get 1 trillion trees on the planet to sequester 200 gigatons of carbon?

We know that we’ve deforested so much of the planet already. I heard a statistic that if we added 1 trillion trees to the global count, it would take us back to preindustrial levels of carbon. Do you have that number in the back of your mind?

Is that correct?

I do. I just gave it to you. It’s about 200 gigatons of carbon.

Here’s how to think about it, Peter. I think we had 6 trillion trees on the planet. How many are left? I guess about half—less than half—around 3 trillion.

If every trillion trees represents 200 gigatons of carbon, and we’ve deforested 3 trillion trees, how much carbon, Peter, was released into the atmosphere?

600 gigatons.

That doesn’t sound very good. Trees are our carbon bank. Trees are carbon. Why are we not focusing on putting our carbon banks back on the planet?

The ocean right now—

Do you know how much carbon is stored in the ocean?

A lot, but I don’t have the number.

About 20,000 gigatons. Do you know how much is in the soil?

No.

About 3,000 gigatons.

Now, I’m not an expert in this area, so all these numbers may be slightly wrong, but some of this comes out of our lab at the University of California, Santa Barbara. Douglas McCauley is our chief scientist there, very close to you.

He says that 20,000 gigatons of carbon are stored in the ocean. When all that material went up in the horrible fires in Los Angeles, a lot of it is going to get into the oceans. A lot of it gets reabsorbed into the ocean, and that’s why the ocean is getting hotter. You get ocean acidification as it absorbs more carbon.

We don’t want that. We don’t want a hot ocean. We want the carbon in the trees. We want more trees. That’s how I think about it simply.

I know you’re also working on your own carbon-extraction prize.

Yes. We’re going to be awarding it. About 4 or 5 years ago, I got Elon to put up $100 million for large-scale, gigaton carbon capture.

It was a fun conversation. He was getting a lot of flak for not being philanthropic on the day he became the wealthiest person, so I texted him and said, “How about doing an XPRIZE?”

He said, “How much?”

I said, “$100 million.”

“For what?”

“For carbon extraction.”

He wrote back, “Sure.”

Then we were off to the races.

We’re awarding it at your event, at the TIME100, this April.

How exciting.

Thank you for keeping TIME alive and for having TIME celebrate incredible moments—not just the negative news in the world.

I’m very grateful to all of our people at TIME. They do an incredible job, and we’re doing the best we can to support them.

There’s so much dystopian press out there.

People’s neural nets—

Yes.

I want to close out with the notion we started with. You identified 5 areas. Did we hit all 5 areas?

We did hit all 5. I didn’t think we could do it. We hit them in the now. It’s incredible. We’re in the now.

Thank you, Marc. This has been therapeutic for you, hasn’t it?

This has been incredibly therapeutic. This is called Moonshots, and it’s a therapeutic podcast. We didn’t even do our deep-breathing exercise. I’m going to call you back tonight, and we’ll talk about this more.

Entrepreneurs listening are looking to do big, bold, significant things. What’s your advice for entrepreneurs who want to take moonshots, as you’ve done?

Get really clear about what you want. Number 1 is: Do you really know what you want? We’re all creating exactly what we want every day, and we need extreme, radical focus on our intention. That is the most important thing.

You mentioned your friend Elon quite a few times. He has some crazy, radical focus on his intention across many categories. As you know, we’ve known him for a while, and he comes up with new intentions all the time.

We need that same idea: What is our intention? Then, what are our values? What are our core values and principles to help guide us?

I mentioned that trust is important, especially when we talk about energy or when we bump into nuclear power on this call. That surfaces my core value of trust.

Third, how are we going to achieve these things? What is our actual plan? But let’s not jump into the plan before we get clear about the intention and the values that are going to guide us.

Of course, obstacles will surface. What is preventing us from achieving success?

The fifth area I focus on is the KPIs. How do I know I’m going to be successful?

But when we go to the KPIs too fast, I think we can lose sight of our intention and values, our vision and values, our plan, and the clarity about the obstacles we’ll have to overcome.

Then we can write down exactly how we know we will be successful. Those 5 areas are the most important things for every entrepreneur.

When you start your day as an entrepreneur or start your fiscal year—as I’m about to do on February 1—I’m going to write down those 5 things.

Part of this gives you the opportunity to ask, “One of my values is giving back. How am I going to give back? What kinds of things will I do?”

We have our 1-1-1 model. Around 20,000 companies have followed us, and billions of dollars have flowed into philanthropy through our 1-1-1 model.

We’re grateful that it has worked out so well. We’ve done 10 million hours of volunteerism, run 50,000 nonprofits for free, and given away $1 billion philanthropically. I’ve done it as a company and personally as well.

When you get clear about your values, it’s easy to make a decision: “This is my value—giving back—and this is how I’m doing it.”

That’s really the power of entrepreneurship: getting clear about those 5 things. One of your friends, Tony, may be responsible for helping teach and hone me on some of those questions, because I think the quality of your questions is the quality of your entrepreneurship.

If you ask the right questions, it will give you the ability to be a better entrepreneur and have a better outcome.

That’s how I’ve led for the last 2 years. I’m constantly working with our employees and asking them those 5 questions. If I meet a manager, a divisional executive, or somebody working on a new product, that’s where I start—with those 5 questions.

You have to be one of the youngest long-lived founders and CEOs out there in the tech sector. I said that properly—I said that carefully.

I’m 60 now, so I’ll take it. Anytime somebody tells me that I’m young, I’m all in. How old are you now, Peter?

I’m 63. That’s why I’m living in your future. That’s why I’m so focused.

I exercise every day, and I wrote this book. We made fun of each other’s books before, and I wrote this book. So that’s our 2 books.

It’s an extraordinary time to be alive.

It is. We’re so grateful, aren’t we? It’s clear we’re living in this dream.

This is the 99th level of the game.

Health is number 1, right, Peter?

Health is the new wealth. Gratitude is in there, too. Gratitude is the mindset.

I know family is so important to you. I have my 2 13-year-old boys in virtual school right next door, and I love them dearly.

Thank you, my friend, for your friendship, first and foremost. Love you so much, Marc.

Thank you for everything you’re doing. I love being friends with you. You’re so much fun.

Same, buddy. Be well. Bye-bye.