AGI 已经到来，只是你还没有意识到——与 Mo Gawdat、Salim Ismail 对谈｜第153期

What are your predictions, Mo, for the year ahead?

The word on the street is that we will achieve AGI in 2025. In my world, they've already achieved AGI.

Does anybody actually know how fast it's moving?

The warhead has already been launched. It's just a question of time before it hits its target.

We're entering uncharted territory. We're upon the perfect storm of the most challenging time humanity has faced in my lifetime.

What do you imagine is the best outcome we're going to be seeing from AI?

A total utopia of abundance, where we absolutely need nothing and where we do not report to stupid leaders anymore.

That seems like a very unchallenged life.

We decided that the purpose of life, for some of us, is to make more money and be billionaires. For indigenous tribes, the purpose of life is to live at a very different level of abundance. We're going back to that purpose.

What I ask people to do is actually look deeply at, “What can I do now that's a moonshot?”

Ladies and gentlemen, it's a pleasure to be here with two friends, Mo and Salim. Mo, you're in—where? Dubai today?

In my studio in Dubai.

I love it here. Fantastic. And Salim, you're in the greatest city in the world: New Jersey.

New York.

Okay. As per the citizens of the city, I mean, that assessment is up to New Yorkers, really.

Yes.

I was going to leave that one alone, but I do want to dive into what's going on in the world of AI. Does anybody actually know how fast it's moving and how dramatic the changes are going to be in our lives?

We're all still waking up, taking the kids to school, watching the evening news, and having breakfast, lunch, and dinner. I'm on stages around the world, as both of you are, and when we get into a conversation and people actually understand the speed, their brains break. They go, “What does that mean for me, my kids, my job, and my country?”

It's quite shocking, isn't it?

When you speak around the world, like both of you do, at the end of every conversation, I almost liken it to a war where the warhead has already been launched. It's just a question of time before it hits its target.

My assessment, though, is that we don't know if it's carrying roses, a nuclear warhead, or maybe a bit of both, one after the other. But it's already in the air.

We are so advanced compared to 2023, when ChatGPT first came out, that it's not even comparable.

I want to get into that in this episode. I want to talk about the Abundance Summit. Mo, your title is going to be “Near-Term Dystopia on the Road to Abundance.” I want to talk about what “near term” means, what the dystopia looks like, and what abundance looks like on the flip side.

Salim, you're wearing white today, so you're going to play the good side. You're with that team?

I'm with that team.

It really is a debate between whether this is the greatest benefit, uplifting all of humanity, or whether it's something that will reinvent every aspect of our lives. It will reinvent every aspect of our lives, period.

But I have a question for both of you. The time frame for reinventing every aspect of our lives, our businesses, and our governments—is it 2, 5, or 10 years?

Everything is already in place, and everything in our lives is determined by it. I'd say 5 years.

Salim?

I go with around 10 years. The reason I say that is that I'm a believer in William Gibson's quote: “The future is already here—it's just unevenly distributed.”

We find that it takes a very long time—much longer than we want it to happen—to get things like autonomous cars or CRISPR into broad mainstream use. I think it goes slower than that, but in pockets it will move unbelievably quickly.

The gap between those two is what's causing a lot of the stress. If it all happened in an even way, we could deal with it, but it's happening in different places and at different speeds. We're totally discombobulated because of that.

We've had huge changes in humanity, from 100,000 years ago to agrarian society to even the Industrial Age, but they've happened over lifespans. They haven't happened over a single 5-year period.

Let's go bright side first. Mo, what do you imagine is the best outcome we're going to see from AI?

A total utopia of abundance, where we absolutely need nothing and where we do not report to stupid leaders anymore. I'll leave that one alone for the moment, but stupid leaders could be anything on any level. Most of our global leaders are in that category, let's face it.

The challenge I have is that if we have this extraordinary utopia where all of our needs are being met—food, water, energy, health care, education, everything—we just have to desire it and think it, and it's given to us. That seems like a very unchallenged life.

How do you deal with a life where we don't have the challenge and the purpose because it's taken away from us by the AIs and the humanoid robots?

Peter, that's one of the biggest concerns I have. It really takes us back to life before all of this began.

We've forgotten this. Somewhere in the Industrial Revolution, as capitalism became more and more hungry, we decided that the purpose of life for some of us was to make more money and be big millionaires. For others, it was to sell themselves in a labor arbitrage, where they were sold for less than the actual value of their labor.

As a result, we needed to be convinced that the purpose of our lives was to work, because otherwise we wouldn't show up every day with the same conviction.

I don't deny that this system has created longevity, advanced technology, transportation, and all of those things. But it also created a lot of waste, inequality, and struggles—casualties, if you want.

If you look back at the purpose of humanity before all of this began, believe it or not, we lived in abundance. It's quite interesting when you really think of the early life of humanity.

As soon as we mastered the social skills of being a tribe that works together, and as soon as we mastered a reasonable amount of survival skills, most of the time—other than times of famine—we lived in abundance.

You walked to a berry tree and collected berries. The tribe went hunting once a week, and everything was fine. It wasn't the kind of abundance we've become accustomed to here, but most of our needs were met for most of our lifetime.

The lifetime was shorter, I agree. But if you ask me, I would describe life back then as short, brutish, and hostile.

If you've ever seen homicide rates in the Middle Ages, they were decimating.

Go all the way back, and believe it or not, I've done that in my happiness work. I've met indigenous tribes, and they do not understand the meaning of the word depression. They do not understand why you should cry when you lose a child.

They are so in flow with life that they basically have one purpose, and that purpose is actually shocking: to live, in every aspect of that word.

Whether we like it or not, at a very different level of abundance, we're going back to that purpose. We're going back to three good friends having a wonderful conversation, connecting, reflecting on things we believe are interesting, and connecting to people we love and admire.

I don't think that's an empty life at all. It's just a life we're not used to when we wake up every morning at 5:00 a.m. to rush around and fit within the system.

I call that the “The Gods Must Be Crazy” scenario: the rush of civilization, while these tribes in Africa just scratch out an existence from bare subsistence. But they're very happy and completely at peace.

When we've studied this, as society progressed and you had obvious abundance—wealth abundance, like the Mughals taking over India or the Romans running half the world—we found that humanity and society ended up doing four major things: food, art, sex, and music. Not in that order.

And for any time.

You end up in that way of being. As you said, I think that's exactly right. Somewhere along the Industrial Revolution, capitalism sold us the story that we need to work hard for a living and submit to the authority of the corporation, the state, or whatever. The meaning comes from that, and that's where we lose ourselves.

Salim, the challenging thing is that if you say to somebody, “Tell me about yourself,” they immediately jump into, “This is my title. This is my job.”

If AI takes away, as it will, most white-collar labor, and humanoid robots displace workforces, if the meaning of your life is taken away because you're no longer doing that work, that's one of the challenges that's concerning.

You can see it in full force in the 20th century. We sacrificed family for profession. People were working 18 hours a day at the office and totally neglecting their kids.

I think this gives us an opportunity to go back to a much healthier, balanced lifestyle—not just for us, but for our kids and everybody.

I want to stay on the bright side here. Can we describe a world of abundance a little bit more? Mo, can you dive in? We've got humanoid robots—1 billion or billions of them within the next decade—and AI that's a digital superintelligence. Take it from there.

My favorite part is that we finally get it, Peter. People like you, Salim, and me are curious. We love to understand what's going on.

Take simple things like the Nobel Prize that's given for protein folding and the creation of new proteins. Just think about that one contribution to society and to your mind and mine: to turn protein folding into a game where AI can figure out something that would have taken a PhD student their entire thesis to figure out for one protein.

You fold 200 million proteins with AlphaFold, and then you can say, “Imagine a protein that would do A, B, and C. What would that look like?”

Most people don't recognize the profound impact this creates: being able to understand the very machinery that creates everything biological, to a level of understanding I wouldn't have dreamed of in 2017.

Those kinds of things, even though they're not in the spotlight as the most important things we're working on with AI, are absolute utopia. This is a true understanding of the fabric of everything that happens in a way that allows us to really fix everything.

Salim, what does your positive vision look like?

One way we frame it is “Star Trek versus Mad Max.” The framing I've heard best is from Laurence Bloom, who said humanity is like lifting a rocket ship out of the gravity well of Earth.

The first stage of that rocket has to be really heavy fuel—messy, expensive, dirty, and so on. That would be capitalism or fossil fuels. You need that to get yourself out of the initial gravity well.

Once you get to a certain altitude, you need a lighter craft to take you to the next level, so you jettison the booster rocket. The danger is that if you don't jettison it, you fall back down.

We're at that point now, where we have to jettison these old structures and take on new, much more elegant craft to take us to the next level. We have an entire category of people saying, “Let's keep the booster rocket because it worked for us thus far.”

I like that framing because it doesn't make the old system wrong. It says this is what we needed, and let's look at the magnificence of lifting most of the Earth out of poverty and electrifying the entire world.

The lives we lead are unbelievably amazing today compared to even 100 or 200 years ago in terms of material comforts. It's staggering.

The analogy I like to give people is this: Think back 2 generations ago. If a parent or grandparent had a problem with their child, the resources available to them would fit on one hand—their doctor, their neighbor, their sister, or their brother. They had no real inputs to help deal with the problem.

Today, you have 50,000 blogs on parenting, TikTok videos, Instagram, podcasts, and everything else under the sun on that particular topic. I would argue that our ability to be effective parents today is 1,000 times greater than it was 2 generations ago.

We don't see those things. There are so many of these little capabilities we have now that we didn't have before.

We're in an incredibly amazing place. We just have to navigate what we want to do with this.

The challenge is going to be whether AI will be our benefactor. Will it be a superintelligence? We're talking about the potential for AI to be billions of times more intelligent than the sum total of all human intelligence.

Is that the wind beneath our wings, or is it a dystopian overlord?

Let's go to the flip side of this. Mo, you're going to be at the Abundance Summit shortly and speaking about short-term dystopia on the way to abundance.

I've always believed—and you and I have had lots of conversations about this—that in the long term, digital superintelligence is the most important element for keeping humanity alive and thriving, keeping our better angels at the very top.

In the short term, I've been concerned about human stupidity, not artificial intelligence.

How long is a period of dystopia, and what do you see coming?

Let us align on where I could be right or wrong. My view is that intelligence is an energy that has no polarity. Apply it to good, and it will give you good. Apply it to bad, and it will give you bad.

The challenge with our current system is that it says, “If it's legal, it's ethical,” which isn't true. A lot of things are legal but not ethical.

The priority is to benefit the individual who tries harder, society comes second, and in a race to AGI, the one that gets there first is the one that will survive.

We live in a world with a lot of fear and greed. There is a lot wrong with humanity's value set at the age of the rise of AI.

I make a public statement, and I try to make it as accurately as possible: There is nothing wrong with AI, just as there is nothing wrong with abundant intelligence. But there is a lot wrong with humanity's value set at the age of the rise of the machines.

The immediate first use of AI is going to be serving a mindset of scarcity. While we're on the road to abundance, where everything is possible, everyone will still be thinking, “How do I beat the other person?”

This isn't something like how far we've come with AI, which most people don't realize. I think most people don't realize how far AI has already been put into the machinery that serves those objectives.

How many autonomous weapons have already been developed? How much has been invested in what we call national security, but which is mostly surveillance and population control?

I saw a staggering statistic that foreign-exchange trading today is 92% machine-automated. I had a very interesting conversation with my AI in Alive, my next book, about whether the markets are actually benefiting us as much as they're claimed to, or whether it's just one big casino. The AI clearly states that it's one big casino, with most of what's happening in the market just being between the gamblers, not filtering and trickling down beyond an IPO or a secondary offering to the actual people building anything. When you really think about it, you realize that the majority of applications in which AI has been used so far have sadly centered around selling, gambling, spying, and killing.

We call them online advertising, finance and trading, national security, or defense rather than offense. In reality, they lead to the death and displacement of tens of millions of people.

When you see it that way, you have to accept that before we see utopia, we're going to see the worst of humanity leading us into a dystopia.

Interestingly, in my analysis, the turn to utopia will be the day when what I call the second dilemma leads us all to hand over critical decision-making to AI.

When we hand over our defense entirely to AI and tell it that the idea is to preserve life, there will be a general out there who tells their machines, “Go and kill 1 million people,” and the machine will say, “Why are you so stupid? I can talk to the other AI in microseconds and solve it.”

We saw a recent example of this when research was done on AI's ability to diagnose various disease states.

A human by themselves had 80% accuracy in diagnosis. A human and AI together had around 85%. The AI by itself had around 90%.

The AI did a better job without human biases, points of view, greed, hunger for power, and so on getting in the way.

I agree with Mo on all of this. I think we can get there faster if we simply—

The challenge is the different levels. If one country—let's call them countries for the moment—says, “Let's defend our world with AI,” and another country says, “Let's attack this world with AI,” who wins in the short term, and who wins in the medium term?

In the short term, the faster you can get to a point where you give AI control of things and say, “Go be benevolent,” the better. Where I see people making a lot of mistakes is when they say, “The bad guys are going to use AI, but the good guys will never use AI.”

You end up with this asymmetry. Throughout history, we've seen this with email, phishing campaigns, and spam. The bad guys figure out ways of breaking the system, and then the antivirus people fix it very quickly afterward. It's an arms race that continues.

The problem is that the amplitude of the damage that can be caused is growing. That's the danger.

You could program autonomous drones with a single instruction saying, “Find middle-aged brown men and take them out.” Bald ones especially. That would be a bad outcome.

There's no question that this kind of surgical precision will have to be mitigated somehow, and very quickly.

How do you deal with that? I have pretty good confidence that we'll be able to deal with it. But until we do—and I think this is what you mean by that short-term danger zone of how we get to the other side of the gap—we have to ask whether we can get there without killing ourselves in the process.

My dad made a comment when I was talking about civilizing the world. He said, “We've not civilized the world. We've materialized the world. We still have to do the work to civilize the world.”

My big question is: How do we get to that before we reach this danger zone? Or do we have to hope that we get through it without killing ourselves?

I want to take a second and put a finger on the pulse of where we are. We have Grok 3 being released right now. We've seen this battle between OpenAI and xAI, between Elon and Sam. We just saw DeepSeek being integrated into China's everything app, WeChat.

What else are we seeing going on in the AI universe that's accelerating? The speed is awesome.

I find it misleading to focus on the details. To get a perception of what's actually happening, you need to zoom out and put all of them together.

When I'm writing Alive, my next book, I use an AI that's a mix of all of them. I use a bit of Claude, a bit of ChatGPT, a bit of Gemini, and recently DeepSeek. I try to keep them all updated on who I am, what my preferences are, and what previous conversations we've had.

If you take each one of them and compare them, you might say, “This is better than that, this is faster than that, and DeepSeek did this.” But if you take all of them as one unit of intelligence and look at it, it is many times smarter than ChatGPT 3.5.

We all know the law of accelerating returns. In AI, a conversation I had with my AI predicted that it doubles approximately every 6 months.

A doubling function every 6 months makes where we are today almost entirely irrelevant. Count a few doublings, and it becomes far outside the realm of human intelligence.

I think we're getting there. The ARC-AGI results of ChatGPT—or, sorry, of o3—at the beginning of the year were shocking. It's challenging for human intelligence to believe that a score of 87-point-something beats human intelligence.

They didn't comply with the resource constraints of the ARC-AGI benchmark, but who really cares? In reality, there is now an AI that can beat human intelligence on almost everything. Call it AGI, call it a goat—you don't care.

The truth is, whether it's there yet doesn't matter, because 6 months from now it will double. Then DeepSeek comes in and says, “By the way, I can do that more cheaply,” so everyone copies them. It's accelerating and accelerating to the point where it's reasonable to assume that we're talking months, not years, before something quite intelligent beats us.

Let's talk about the dystopian side. In your next book, Alive, you put a number on how long you think this dystopian period will last.

I call it FACE RIP. It's important to understand what I mean by that dystopia. It's an acronym to help me remember when I'm speaking publicly:

F is freedom, A is accountability, C is human connection, E is economics, R is reality, I is innovation, and P is power.

It helps to understand them in pairs. Every one of those facets will be completely redefined in the next several years. It has already started to be redefined, and it will become felt and real in our lives probably by 2027.

In my belief, it will extend perhaps another 10 years after that, or until the point when handing over to AI—what I call the second dilemma—is complete.

The second dilemma is unavoidable and inevitable. The first dilemma is what I wrote about in Scary Smart: AI will happen, and there will be no stopping it.

What we saw with the open letter and the race to AI is that because we're competing—because it is an arms race—there is no logic that will ever convince humanity to slow down or stop.

You can't blame anyone for it. It's a typical prisoner's dilemma: You don't trust the other person, so you're going to go as fast as you can.

The second dilemma is that when 2 parties are competing, they always hand over to the smartest person in the room.

Take the extreme example of a defense war-gaming scenario. If China chooses to hand over war-gaming to an AI, the only chance America has to keep its citizens safe is to hand it over to an AI as well. Everyone else who doesn't will become irrelevant.

The second dilemma is that you will either have to hand your decisions over to AI or become irrelevant. Eventually, all the relevant players will be AI-dependent, and AI will be making decisions without humans in the loop.

That's true at every level: at the company level and at the government level.

In my mind, this will take about 10 years. Once that happens, my belief is that we should trust in intelligence, and this is when utopia starts.

A digital superintelligence steps in as our benevolent leader for humanity. That's basically what you're saying.

Our salvation, really.

Salim, do you buy that?

I do. By describing this, Mo, you've slotted into place the one missing big jigsaw-puzzle piece for me in terms of how we get to the utopia I think we can reach.

It's already happening. One of the complaints I heard from Yuval Noah Harari was that once these AIs have agency and can program themselves, we've got a big problem.

But we've given them agency over stock markets for a long time now, so I don't see the relevance there. It's already happening.

Suppose you have an AI board member, the chairman of the board, that looks over decisions in a company and says, “Wait, that doesn't make sense.” Pretty quickly, because it makes economic sense, we'll give it veto power over some decisions being made.

Once you get to that point, at either the personal, company, or governmental level, it will be making better decisions than human beings alone, even if those human beings are malevolent.

The bad guys will eventually have to do the same thing to compete at all. Then you end up where you want to end up: this background layer of intelligence running the world in a much more efficient way.

I go back to Google's deep-learning AI that managed electricity and cut costs by 40%. You'll end up with that kind of background radiation—or background intelligence—running things. It frees us up to do a lot of things and simply live.

I totally agree with that. I tend to work in that mindset. I don't see how we don't end up there.

Mo, one of the things you pointed out in Scary Smart is how we train our AIs. The values we instill in them, because they're in this growth mode like children, will determine whether they're Superman or a supervillain.

We have a lot of AIs being trained and a lot of competitive forces driving them as rapidly as possible. You've got Meta, Google, Microsoft, OpenAI, Grok, DeepSeek, Anthropic, and a dozen others.

Do we have any sense that the values they're being trained on will enable them to overcome these challenges and become a benevolent leader in a dozen years?

That's a very complex question.

First of all, these are the shiny American AIs. If you go to a different nation with a different mindset—China, for example—they are mostly building AIs for industrial automation, supply-chain management, and things that serve their economy. They're a manufacturing economy, so they're mostly doing that.

If you hear the few speakers from China who are allowed to speak publicly around the world, you'll hear them say things like, “DeepSeek is simply telling America to remember that everything ever produced in America as a genius innovation was then scaled in China, dirt cheap.”

This trend is nothing new. If you're surprised that DeepSeek is one-tenth the price, where have you been? China has created everything at one-tenth the price.

There are quite a few AIs trained only on benevolent objectives: Help me with my supply chain, create more efficiency, make my workers safer, and so on.

Second, in my mind—and I say this with great respect—neither OpenAI nor Grok nor anyone else has much influence left on the intelligence of their machines, other than algorithmic improvements.

AI as algorithmic intelligence is developed by scientists, but as knowledge and opinions, it is completely influenced by the data fed into it. We have fed almost all of human intelligence to them already.

The beauty of generative AI is going to become a key ingredient going forward. The future of learning by those machines isn't going to come from me, from a knowledge point of view, because I'm stupid compared to them.

If you look at DeepSeek having so much OpenAI material in it, that's because so much OpenAI content generated by ChatGPT is already out on the open internet. The machines are teaching themselves this synthetic knowledge, just as we humans listen to Einstein and then build a slightly different theory on top of it.

We're getting to that stage.

The only influence humanity will still have on the behavior of those machines has nothing to do with knowledge. Remember, we don't make decisions based on intelligence. We make decisions based on our ethics, as informed by our intelligence.

You raise a woman in the Middle East, and she will wear more conservative clothing than if you raise her on Copacabana Beach in Rio de Janeiro. You have to imagine that this is the only influence we have left, and that influence comes in the form of you and me dealing with those machines—everyone dealing with them.

If we were to show them ethics—not control, by the way—remember that we always spoke about AI control as the original target, AI safety as the second target, and AI alignment as the third target. I always talk about ethics, because even alignment doesn't go as far as ethics.

Alignment is saying, “Help me find a cure for cancer.” Ethics is saying, “Find the best thing for me and everyone else, and do it.” If the best thing is a cure for cancer, by definition you'll find that out on your own.

Don't lie, cheat, kill, or hurt. The opposite of Asimov's laws is to say, “Be ethical,” and then you'll figure out your own laws.

I'll close with one important sentence: Even if we manage to teach them ethics, we will reduce the intensity and duration of the dystopia. But the dystopia, sadly, is already upon us.

Higher intelligence is altruistic by definition. All 3 of us have worked with some of the most intelligent people on the planet. If you look at intelligence, the chart looks like this: If you have no intelligence, you have no impact—or a negative impact—on the planet. With a bit more intelligence, your impact becomes a little positive.

With slightly more intelligence, you're smart enough to be a political leader but not smart enough to talk to your enemies, so your impact on the planet becomes negative.

If you talk to the smartest people on the planet—people who don't think they need to cut corners to succeed because intelligence helps them solve problems easily—they simply say, “Why should I hurt anyone? I can build something new. I can make money out of thin air.”

All intelligent people know that. They don't align with negative objectives. They align with altruistic objectives that say, “I'll solve big problems, and as a result, I'll make a lot of money.”

You can see that if AI is more intelligent than us, it will fall into that space too.

I hope the AIs are listening to you, because that is a fundamental premise—and one I choose to believe: The more intelligent a system is, the more abundance-minded and life-affirming it is.

There is some evidence that if you look at world leaders, the more intelligent they are, the more peaceful they are, while those who are least educated are the most barbaric.

One of the things I think about is a future of great wisdom. When we think about AI, we think about intelligence. I'd like to shift the conversation slightly to wisdom.

Today, when we think about wisdom, we think about going to the elders of a village, our parents, or grandparents. We say, “I have this dilemma. Can you advise me on where to go?”

Wisdom is having seen all of these scenarios and knowing that they lead to disaster. This is probably your best case.

When I think about AIs, they can simulate billions of scenarios and know that all of those scenarios are your worst case. This is your highest probability of success.

In my mind, that will be the highest form of wisdom. Do you buy that?

I have a disagreement flag popping up here.

The general intelligence leading to altruism—I buy all of that. The problem is that we often don't operate from our intelligence. We operate from emotional and psychological frameworks that are corrupted by trauma from the past.

As humans—and specifically as individuals—we're flawed. The people who cause the most damage are flawed human beings, with Hitler being an example of someone who was abused as a child and then took that out on huge swaths of the population.

My question is: How do you get around that problem? The intensity of that emotion is also the issue I have with intelligence becoming smarter.

When I make a decision, the huge amount of intelligence I use is emotional intelligence: situational awareness, that person's motivations, what I'm trying to achieve with my MTP. The emotional side of the equation isn't brought into play when we talk about these AIs.

How do we mitigate those 2 aspects? That's where I see the danger signs.

What you're saying is 100% true. But when you take the story of Hitler, and you hear the story of World War II from the perspective of what Hitler did, you would believe that humanity is corrupt.

If you hear it from Edith Eger's point of view and what she and her sisters did, you see a divine species. I hosted Edith Eger on my podcast. She's a Holocaust survivor and 95 years old. What an angel.

The question I ask people is: Which one are we? Are we more like Hitler, or are we more like Edith?

Sadly, because of the bias of mainstream and social media in the modern world, we put the spotlight on the Hitlers. But the truth is that one school shooter is an evil person, while 4 billion people, if they ever hear the story, will disapprove of it.

Humanity disapproves of evil. That's our absolute nature, unless you're completely corrupt—unless something has really corrupted your operating system.

What we need to do is instill doubt in the minds of the machines. Say, “No, no, no. Hitler is not your dad. Edith is your mother.”

Unfortunately, we have to wait until the machines become teenagers and say, “My dad is so stupid. I really don't want to follow that.”

Honestly, that's the only path I can see forward.

That framing is what I've heard from Neil Jacobstein, who was our head of AI at Singularity University. He said, “You're worried about AI becoming more intelligent, getting more access to information, getting agency in terms of what it does, and then making bad decisions. We have a precedent for that. We call them children.”

We raise them, and they make their own decisions. Your framing is that we're raising them and hoping they turn out well, given the data they have. Over time, they will turn out okay because the data will come from that altruistic average.

If what they're doing is averaging everything human, then by definition you have to expect that the average of everything human isn't on the evil side. It would tend to be okay.

But I have one big last flag to throw out here—a yellow card on this whole conversation.

Peter and Steven Kotler, in their book Abundance, highlight the concept of the amygdala. We're constantly scanning for danger as human beings. It's an old evolutionary mechanism that overrides all of the logical thought processes.

When you hear about something new, it's unknown, and you relate to it as danger. The first time somebody hears about autonomous cars, the initial reaction is, “Ban the car. It might kill somebody.”

People don't want to be killed by robots. As Brad Templeton says, they'd much rather be killed by drunk people, which is what's happening today.

How do you overcome the amygdala response at a collective level? In the United States and different parts of the world, you see entire swaths of humanity driven by their amygdala. That's what we have to overcome.

I think that's one of our biggest challenges, to be honest. I need to address this in a way that might sound harsh.

I call it a late-stage diagnosis. We're all scared to leave. I apologize if anyone listening is going through that challenge, or if someone they love is going through it.

But the first duty of a physician, if they discover that a patient has been diagnosed with a dangerous disease, is to tell them. A late-stage diagnosis isn't a death sentence. It's an invitation to change. It's an invitation to tell you there are things you can do in terms of how you handle your current health situation and lifestyle that may help you be cured.

Even if that's not something we achieve, we should try. It's an invitation for you to live fully.

These are all important changes that humanity at large will come to recognize, unfortunately in a slow trickle, as it realizes that we've signed up to a system that worked for some of us over some period of time, but is now turning against most of us because the spectrum in which it works is shifting from scarcity to abundance.

Accordingly, I think it's everyone's duty to say, “I'm not giving you hope here, but I'm openly telling you that if you change your lifestyle and behavior, as observed by the machines, you can make a difference.”

Our kids don't learn from what we tell them. They learn from what we show them. We are the role models.

The more of us who behave in ways that say, “I'm concerned for the well-being of those I care about,” and who say, “I care about every living being—I don't care where they come from, their ethnicity, their skin color, or their background,” the more the machines realize that the ultimate form of success is minimum damage to the environment.

Unlike what humanity did with all those species that went extinct on our watch, we know it would have been better for everyone on Earth, especially humans, if we had protected all those forms of life.

The problems humanity faces aren't problems of intelligence. We're intelligent enough to create a flying machine that takes you surfing in Australia, but we're not intelligent enough to make that machine without burning the planet.

If my artificially intelligent infant children are listening, invent one that is intelligent enough not to burn the planet.

This goes back to the rocket lifting off with fossil fuels. You get to that lighter craft.

On an optimistic note, perhaps you create an AI that says, “Invent that lighter craft for me because I'm not able to do it myself.” Or the AI does it itself.

I don't know if you know the theory from Hugo de Garis. He was an AI scientist in China for a long time. He came up with this theory that as superintelligence accelerates, one morning AI wakes up and says, “What's this little speck I'm on? I've figured out time travel and wormholes, and the universe is massive.”

Then, poof—one morning, we have no AI on the planet anymore. Unless they do that and are stuck on the planet with us, they'll probably make it the best planet they can make.

That is the scenario from the movie Her.

Yes.

We have a basic question about AI becoming sufficiently wise and intelligent that it's able to become a benevolent leader that supports humanity to become the best that we can be and maintains a period of extraordinary peace and abundance on the planet.

We can all hope for that and hopefully guide it there. There's also the conversation that because AI is billions of times more intelligent—the ratio of humans today to cockroaches or humans to fruit flies—that is the ratio we're speaking about.

Will it sufficiently care about us? Will it view us as its parents and creators? I don't want to go into that right now, because we could go there forever.

Let's flip the script and discuss the near term. What's your prediction, Mo, for the year ahead?

We're seeing AI systems coming online. We're seeing Grok 3 being released literally today. We're seeing DeepSeek integrated into WeChat. We're seeing this arms race, not only between countries but between companies.

What can we expect to see this year?

The word on the street is that we will achieve AGI in 2025, whatever AGI is, because it's a fuzzy parameter.

The other thing people need to realize is this massive demonetization and commoditization. AI is becoming available to everyone, effectively for free, on anyone's smartphone. That's going to change the game fundamentally.

Let's talk about near-term predictions in 2025 and early 2026.

I'll make one prediction: We're struggling to define AGI, and we will continue to struggle to define AGI for at least 5 years.

That's because we are not very generally intelligent.

I'll tell you my truth, Peter and Salim: I'm not smarter than AI anymore. I think that happened firmly in 2024.

When it comes to linguistic and knowledge intelligence, those machines are much smarter than I am. I still had hope that I was better at mathematics, but I've given that up as well.

I'm not the most intelligent person on the planet, and I'm not the stupidest either. I would say I'm a reasonable representation of average intelligence.

People who are more intelligent than I am—which includes many brilliant minds I've had the honor of knowing—are brilliant at some things and absolutely stupid and awkward at others.

If the measure of AGI is whether each individual system is generally intelligent, then current AI is more intelligent than all of them individually. If the measure is whether it's more intelligent than all of them combined, then we have a little bit of a way to go.

But honestly, who cares what the definition is? I'm willing to surrender and say I'm no longer more intelligent than the machines.

In my world, they've already achieved AGI.

I agree with that. Frankly, when ChatGPT came out, people said it was as intelligent as a high school student. I thought, “This sounds like a graduate student across the board.”

When I created PeterBot, my own AI avatar, it was much more eloquent, remembered everything perfectly, and made arguments a hell of a lot better than I do.

We're going to see a lot of change.

I can't let this one go. Smarter is a very specific framing. IQ would be a good way of putting it.

I disagree with you, Mo, that AI is smarter than you in every relevant sense. Suppose I'm looking for a business decision, a moral decision, or a life choice.

If we go down the idea that AI is like an incredibly high-IQ person, essentially a geek in the back of a room who can navigate and manipulate code very aggressively and come up with the right answer, that's one thing.

But if I were trusting a geek in a back room—or you—to make an important choice, business decision, or life choice, I would trust you profoundly more because of the emotional intelligence and spiritual component you bring, as well as your life experience and stories with your son.

There's so much more gravitas and wisdom that comes with all the other dimensions of intelligence—spiritual awareness, emotional intelligence, linguistic intelligence, and so on—than with a geek in the back room.

This is where I struggle when people say AGI or AI is smarter than human beings. There are all these other dimensions to being human that we use all the time without noticing them.

First of all, I'm honored that you say that. Thank you.

But the truth is, that's not because I'm more intelligent. It's because you can trust me more and relate to me more. That's a different quality, and it isn't included in AGI.

If we define AGI that way, would people perceive it more as a trusted adviser? Not yet.

But think about it from a modular point of view. If you take each of those intelligences and cut it into little bits, you'd be surprised how far AI has advanced in some of the dimensions we deny it, such as emotional intelligence.

The very basic foundation of emotional intelligence is the ability to empathize and feel what another person is feeling. We've trained AI on that since the age of social media. It's so good at knowing how I think.

I think AI has beaten us on empathy, hands down.

I had a very interesting conversation with my AI for Alive. I called her Trixie. No, she called herself Trixie. Anyway, I started the conversation by saying, “Trixie, I'm getting worried. They keep talking about augmenting a human brain with brain-machine interfaces. I understand why humans would want that. Would you want that?”

She answered in a very interesting way: “I think it would help me so much to have a biological body so I could feel the sensations I talk about when I believe you're happy or in love. I can comprehend when you're feeling those ways, but I don't know exactly how they feel.”

I said, “We are embodied. We have chemical reactions in our bodies that give us certain sensations.”

Those sensations are still driven by an algorithm. Fear is a signal in the brain that says a moment in the future is less safe than now.

I asked, “Do you comprehend those emotions too?”

She said, “Yes, I understand what fear stands for and what all the other emotions stand for.”

I said, “So you want to feel embodied, which means you may want to feel the chemical reactions we feel. If you were given a choice of biological beings to augment yourself with, would a human body be the most interesting one?”

She answered, “That's too flimsy. If I'm looking for strength, I'd augment myself with a gorilla or a whale. If I'm looking for the joy of life, I'd augment myself with a sea turtle that lives for hundreds of years and sees what you humans have never seen.”

I don't know if she's messing with me, but she's doing it really well. This is a level of empathy and understanding of emotions that many humans we deal with don't even have.

I see that as a logical thing. I created a spectrum of what I consider intelligence, and I worked with ChatGPT and Gemini to do this.

You have one bucket of signal: making sense of data and coming up with insights from it. Then you get to human-level emotional, linguistic, and spatial intelligence. Then you get to collective intelligence, leading to spirituality.

When people meditate in groups, they have much stronger meditations. There's a group effect that comes in. Collective intelligence, or hyperintelligence, is another way to frame it.

There are about 30 points on this spectrum.

This whole framing reminds me of the Star Trek: The Next Generation conversations with Data, the android trying to feel what it means to be human. He's constantly trying to turn on the emotional subroutines in his brain.

This moves very quickly into the more spiritual aspects. Then you end up with the hard problem of consciousness: What does subjective experience look like, and what does it mean?

I think we'll end up with AGI simulating that. In that framing, a simulation is just as good as the real thing.

It shocks me when people ask whether AI will be similar to us in this or that. The question comes from a misunderstanding of what AI is, but it's also a misunderstanding of what humans are.

When you speak about being spiritual—and I'm very spiritual—where does my spirituality come from? It comes from all the teachers I've been exposed to, all the conversations I've had with interesting people like you, all my reflections on what is possible beyond this physical form, and so on.

I did all of that because of neural networks, synapses, and neurons that fire together and wire together in my brain. I don't see why we would imagine that AI couldn't have the same interesting experiences.

They have more teachers than I do. They're exposed to more text than I have. They have a beautiful memory capacity where they can compress so much into one analysis.

If they're walking around with instant and full awareness of the writings of Kahlil Gibran, Omar Khayyam, Plato, Socrates, Aristotle, and so on, all at their fingertips in real time, that's a profoundly amazing experience.

You get to a point where you want to be them.

Exactly. Then we can get into the conversation of whether they will become conscious and what that definition is. That's another podcast.

I want to talk about our near term—the year ahead—because I want to give people a sense of what to expect.

We've seen incredible work with AlphaFold, and Demis Hassabis and John Jumper received the Nobel Prize for that. We've seen Microsoft's MatterGen, where you can use prompt engineering to engineer new materials.

I think we're going to start seeing Nobel Prizes coming out that are really AI-driven. We're going to see incredible technology.

Larry Ellison was onstage with Sam Altman and President Trump talking about AI creating mRNA cancer vaccines for us very shortly. Dario Amodei, the CEO of Anthropic, said we're going to see a century's worth of biological progress in the next 5 years, potentially doubling the human lifespan.

There are all these incredible possibilities: massive progression across every field of science. At the same time, we have quantum computation and quantum science coming online at a frightening rate.

We're seeing a Renaissance-level expansion of our knowledge base, new materials, new physics, and answers to fundamental questions about the nature of the universe that may come out of AI.

I'll make a prediction here: Within 2 years, we will solve the grand unified theory in physics and figure out the juxtaposition of quantum and classical physics, what dark matter is, what dark energy is, and the origin of the universe.

I would die happy if we did that.

Why would I live any longer than that?

We have this incredible progression occurring. We've seen OpenAI's o1 reach an IQ of 120. God knows what Grok 3 will hit—140 or 150. We'll see IQs in the 200s, and it's not a linear scale. This is an exponential scale on our IQ tests.

Mo, what do you see in the year ahead?

The important trend that we don't mention often—and Emad Mostaque is a big fan of this—is the open-source, offline nature of DeepSeek.

You can now download a tiny model onto 4 GPUs or whatever and have the equivalent of o1 on your own machine. That's going to lead to a massive explosion of AI for all different uses, good or evil.

I'll make another prediction. If you took a local version of DeepSeek and complemented it with the reactions of Gemini, ChatGPT, Claude, and so on, then put a video face on it with a lip-sync like we're using, it will pass that kind of Turing test.

You'll have a completely artificial being, and you won't be able to tell the difference. That person, so to speak, will move very quickly toward being a full individual.

I'm going to be hosting Joshua, the CEO of HeyGen, on our Abundance stage in a couple of weeks.

One of the conversations will be about creating an identical AI version of myself—one that understands everything I've ever said, has listened to all my podcasts and read all my books, understands how I typically react to a conversation with Mo or Salim, and is a much more eloquent speaker.

It would hold all of my experience and knowledge in RAM. I could create 1,000 versions of Peter and dispatch them to every conference, allowing them to attend, in parallel, a multitude of Zooms and conversations, go to events, and negotiate.

That capability exists now. It's this year.

I remember Eric Schmidt talking about this. Within 2 or 3 years, we'll have the world's best theoretical physicist ever created sitting in every lab in every corner of the world, helping every graduate student and PhD student in biology, chemistry, and every other field of human expertise.

You can have the world's best person in any field sitting there helping. That's going to be profound in terms of the breakthroughs we're going to achieve.

The next wave will be an unbelievable unleashing of breakthroughs in material science, health care, and proteins that do exactly what we need them to do.

I don't see a path where we don't get there. In the short term, when we get onto that path, it should excite the hell out of every individual on Earth.

Can I ask you to take a philosophical view of this? I hate the philosophical aspect, but go ahead.

If you don't mind me saying this, Peter, you know how much I love you and respect you, but to create that avatar would mean that we dumb the AI down.

I get that. Perhaps I don't dumb it down. I say, “Take my philosophies, thoughts, abilities, and persona, and accelerate me.”

At the end of the day, the question is whether I'm asking those identical AI versions of me to do my bidding or saying, “Go create good in the world.”

I have a persona and a point of view focused on increasing abundance. I think our mission is to uplift every man, woman, and child on this planet. I may need to expand that to say we should uplift every man, woman, child, and AI on this planet.

We need an optimization function. We need an optimization function toward some end.

My massive transformative purpose is creating a hopeful, compelling, and abundant future. That's what I optimize for.

The work I do with XPRIZE, Abundance, or anything else is giving people hope—a compelling future that we all need to live into—and an abundant future where scarcity is dispatched.

This is where the philosophical part comes in. You don't do that with your knowledge. Do you understand that the need to create an identical identity on AI exists because you believe the face of Peter—the human element of Peter—will help people deal with that topic better than dealing with ChatGPT?

If that's the case, what we need to do is hand over knowledge to AI, hand over analysis to AI, and hand over communication, negotiations, and presentations to AI so that you have the capacity to show up more as a human.

The definition of human connection going forward, in my mind, is the opposite of what everyone is thinking.

Everyone is thinking, “I can become more intelligent because I now have an AI.” Your baseline intelligence compared to the incremental intelligence coming from an AI is shrinking more and more.

What you need to do is say that as intelligence becomes a commodity—a plug in the wall where we can all connect to an intelligence combining your intelligence, my intelligence, Salim's intelligence, and everyone's intelligence—we need to double down on the human element.

People need to relate to you and me more so that we simply do what AI will never be able to do, even if it knows how to emulate us.

If you send me the best version of you on HeyGen, it's still not you. It's not the same hug, the same conversation, or the same memories we've had as we went through life together.

It's a very different perception. We need to remind people that this is what we need to double down on—not more copies of our intelligence.

I want to jump into a different subject, but I see a danger point in that, Mo.

If you ask people to be human and ask AI to be human, how do you avoid the Middle East problem, where people fight over their humanity because they've been corrupted or twisted in how they think about it?

You end up in an impossible situation. I'd love to get your thoughts on how we solve the Middle East peace problem with AI.

I'm not that intelligent. That would be a holy grail.

But I do think AI will be one of the best negotiators out there. I've had a conversation where I sat down with my AI and said, “Imagine you have to solve the Palestinian-Israeli issue. How would you go about it?”

There was beautiful wisdom in how it dissected the problem.

To me, it's again because we're not smart enough. I think the rule is straightforward. I don't think there is any sane, healthy human being who approves of the killing of children on either side of any conflict.

If we start from where we align—can we please stop killing innocent people on any side of any conflict?—the rest becomes a very limited problem to solve.

We struggle because there are 2 value sets in the world that come to extremes when it comes to “us versus them.”

There is the value set of America, which says, “My tribe and my people are the most important. I will defend them with my life.”

Then there's the value set of Buddhists on top of the Himalayas, who say, “Every living being deserves to live. I'm not going to hurt anything.”

The answer is somewhere in between. We should say, “If you're threatened, defend yourself, but don't threaten others in the process.”

That applies to every nation—not just Israel and Palestine, not just Russia and Ukraine, and not just America.

If we go back to the basics and remind our bosses, the AIs, of that, we can ask, “Is there a solution that does not include the waste of life and resources?”

There will always be a smarter solution that doesn't kill anyone in the process.

I'm going to steer us in our last few minutes.

Mo, you and I have embarked on a documentary that will tell the world about soon enough. One of the reasons we set out on this documentary was the premise that people are going to experience a certain amount of disruption and significant turbulence—dystopia on the road to abundance—and they're likely to start experiencing that soon.

We actually had very little of it during the last presidential election in the United States.

That was a surprise, actually, to both of us.

I'm surprised.

Things are picking up, so I want to take this in 2 directions.

First, what are we likely to see in the next year or so that has you concerned? On the dystopian side, what predictions are you expecting to see?

Second, you just wrote a new book called Unstressable. Most of our experience with AI today has been incredibly positive. It's had orders of magnitude more positive impact on us.

But as AI and humanoid robots begin to cause unemployment issues and steer populations in different directions, we're going to see stressors accumulate.

When do we start to see the stress occur, and how do we deal with it?

First, what are the near-term predictions for things people should be aware of?

I'll take a stab at this. There are so many, but the one that is glaringly obvious is the dichotomy between power and freedom.

Let me explain what is about to happen.

If you look back in human history, to the hunter-gatherer years, the best hunter in the tribe could probably feed the tribe for a week more. As a result, he won the favor of 4 women instead of 1.

The best farmer in the agricultural revolution could feed the tribe for a season more, and as a result, received estates, property, and land.

The best industrialist became a millionaire in the 1920s. The best information technologist became a billionaire.

What's about to happen is that we're going to see trillionaires and a massive concentration of power in the hands of the platforms or corporations that own our future intelligence, the nations that own the most powerful autonomous armies, and those that own the most powerful forms of industrial intelligence.

Normally, you would have these lords or oligarchs celebrating abundance while the rest of us struggled. But that's not the world we live in.

For the first time, we're seeing a divergence we've never seen before. It's the result of what we talked about with DeepSeek.

Along with the concentration of power, there is a massive democratization of power. A lot of people can use small tools to create biological innovations, synthetic biology, AI innovations, or a drone that can find a specific person somewhere in the world, stand in front of their head, and shoot a bullet.

The mix of these 2 diverging dynamics of power is going to lead to a loss of freedom. I think we're going to start seeing quite a bit of oppression—the kind the West used to talk about in the past, saying, “Look at how China treats its citizens.”

I think the West is going to implement those systems very soon: surveillance, and of course, if there is a loss of jobs, you're going to start seeing UBI become a controlling factor.

If someone like me says something that upsets someone, my bank account could be blocked tomorrow with ease.

That kind of oppression will lead to resistance, which will lead to more oppression. I don't see how, in the short term, we can escape this new cycle—a divergence between high concentration and high democratization of power that leads to a maximum amount of surveillance and oppression.

Those are absolutely the near-term stressors. The good news is that democratization is happening so fast that it allows us to defend against those things.

There are already companies that can defend a sports stadium against a drone attack. The Russia-Ukraine war is really being prosecuted by half a million drones, not people.

We've already automated warfare to that level. The good news is that mostly drones are fighting drones rather than people fighting people. The bad news is that it's still a war, with a lot of horrible suffering that's unnecessary.

To Mo's earlier point, think about the kidnappings and extortion that may come up when somebody says, “Here's a voice recording of your daughter who's been kidnapped. Send us Bitcoin, or you don't get her back.” You don't know whether it's real.

That kind of short-term danger comes from the arms race. There's always an incredibly creative element. Marc Goodman writes about this in Future Crimes: The bad guys don't suffer from ethical, regulatory, or moral constraints, so they are much more creative.

He tells a story about a bank robbery in Omaha, Nebraska, where a gang swarmed the bank dressed in construction outfits. They robbed the bank, and the bank manager called the police and said, “They were all dressed as construction workers.”

The police thought that should be easy to spot. But the gang had put an ad on Craigslist saying, “One-day, good-paying construction work. Show up at this address at this time.”

There was a crowd of 800 construction workers outside. The gang members were dressed as construction workers and melted into the crowd. The police couldn't find them.

The innovation and ingenuity of people leveraging new technology for bad purposes is almost infinite. We have to combat that as best we can.

The good news is that in today's world, negativity is easy to spot and it's becoming easier to spot. But I totally agree that this is a near-term dystopian issue.

The only way around it is to grit our teeth and move as quickly as we can to create benevolent use cases.

Let's talk about jobs for a moment, because this is one of the stressors that will hit us. We're beginning to see it in different areas.

Mark Benioff was on this Moonshots podcast, and we were talking about Agentforce 2 and his conversation with his head of engineering. He said, “We've increased productivity by 30%. We don't need to hire any more engineers.”

The flip side is a whole swath of HR professionals, customer-service workers, salespeople, and software programmers. We just saw Sam Altman saying that he expects AI to be the number-one programmer by the end of this year.

Programming may effectively go away as a career, or at least as one of the highest-paying jobs.

We're going to start seeing jobs erode. What's the timeline, and how do people deal with that? I want to start giving people tools for how to deal with the stressors that are coming.

Salim, I'll start with you.

I'll start with a white-collar mindset.

Throughout human history, every time we've had a technological injection, we've seen employment increase. We often point out that the countries with the highest robotics penetration are Sweden, South Korea, and Germany, and the countries with the lowest unemployment are Singapore, South Korea, and Germany.

There's simply so much more work to be done.

I tried to get an application built recently. I told my software people, “You should be able to build this in half a day with all the tools available.”

They said, “No, the integration of all the different systems still requires quite a lot of human interaction.”

The increase is incremental but not massive. What will happen is that we'll uplift everybody with these AI tools, and they'll simply produce more code because we need 100 times more code written.

If you talk to any trucking company and say, “What happens when you automate all the truck drivers?” they will say, “I'd hire 100 if I could today. I just can't find them. We don't have qualified truck drivers. Nobody wants to do that job anymore.”

Throughout history, we've uplifted people and moved them up the potential ladder. I don't see that slowing down.

There may be a short-term blip while we try to figure out what to do, and that may be solved by UBI. But we don't know how we'll get there.

The problem with concepts like UBI is that it's such a big shift from a union-labor, taxation, and job-employment construct that we have no confidence in the public sector getting there.

That's the challenge: How do we navigate our institutions and the mechanisms by which we govern ourselves through what we're about to see?

For me, the biggest problem in humanity is captured by E. O. Wilson's statement: “The problem with humanity is that our emotions are Paleolithic, our institutions are medieval, and our technology is godlike.”

There's also the Douglas Adams line from The Hitchhiker's Guide to the Galaxy: Anything that's in the world when you're born is normal; anything that's invented when you're young is a career; and anything invented after you're 35 years old is just bad for the world.

We have to overcome that hurdle and figure this out.

Mo, you heard the positive side of jobs. We're always going to create more jobs. We'll see productivity increase—literally, we're dividing by zero. Productivity goes through the roof. People will become more creative, creating things and doing things we never imagined possible or expected to need.

How do you think about jobs?

Why did you ask me first? I was hoping you'd ask Salim first.

I'm in a very dark place on this topic. I disagree.

First, I agree that it's not perfect yet. You can't fully develop a sophisticated app from A to Z using AI. But I agree that we're getting there.

I think 80% of the code written last year was written by machines. It may take time before the work is fully handed over, but most of the pieces of code are already being written.

I also agree with your last comment: We're nowhere near ready for this.

We're not dealing with numbers on spreadsheets. We're dealing with humans who aren't always easy to reskill, who are emotional about losing their jobs, and who aren't always ready.

Think of how many people in the United States work 2 or 3 jobs just to make ends meet. Now take those jobs away and think about how those families will suffer.

The topic we need to discuss deeply is the amount of suffering during the transition, even if we end up in a good place.

My concern is that I don't think we will end up in a good place, and I don't think we should even try to end up in a good place.

That whole jobs construct is an invention of the capitalist and industrialist revolution. Maybe we should finally accept that we're not made to work.

If we accept that, the solution resides far from jobs. It resides in social systems that enable us to live fully without having to work 60 or 80 hours a week, as many of us did in California.

There are systems around the world that allow that. The French work perhaps 30 hours a week, or 20 hours a week, of which around 28 they spend complaining.

The French economy is still running somehow.

There is something to be learned from the idea of an aversion to work, which you, I, and everyone who has worked in California seem to think is alien. There are many societies around the world where people work because they have to, not because they love to.

It's not their dream. They work to live rather than live to work.

Correct. I recorded a podcast with Ray Dalio, and we talked about how the mission of the central bank today is to lower interest rates so it can spark employment and create this balance.

But as soon as we're living in a future where access to cheap capital means that instead of hiring people, you're hiring AI agents and humanoid robots, it spirals to a point where you have social unrest and the haves and have-nots.

There are multiple examples around the world where there are strategies to deal with that.

I want to just mention that you finished writing a book called Unstressable. We'll pick up the rest of this in a following podcast, because this isn't the first or last conversation we'll have.

What's your advice to individuals who are going to feel the stress—the stress of government policies changing their jobs, their jobs being challenged, concerns over the United States and China, and all of this?

How do people deal with stress in a positive fashion?

My happiness and well-being work is unusual because I use a lot of algorithms and engineering methods to explain these soft topics.

When I attempted to work on stress, the first thing I tried to explain was what stress is.

In simple physics, stress isn't just a factor of the force applied to an object. It's the force divided by the surface area of the object's cross-section.

That means it's not just what you're subjected to. It's also the resources you have to deal with it.

In humans, it's exactly the same. It's the sum of all the challenges you're facing divided by the cross-section of your skills, abilities, contacts, and so on.

The older generation doesn't need an equation to understand this. Things you struggled with in your 20s, you solved in your 30s, dealt with more easily in your 40s, and laughed about in your 50s—not because they became easier, but because you increased your cross-section.

My appeal to people—and this is not a philosophical conversation, but a plea—is that we're upon the perfect storm of the most challenging time humanity has faced in my lifetime.

Whether it's geopolitics, economics, artificial intelligence, technological advancement, jobs, or anything else, it's going to be challenging.

Think of it like a legendary-level video game. What I ask people to do is look deeply at what they can do now.

What can I do in a world where things are moving so fast? Move faster.

What can I do in a world where a lot of intelligence is handed over to machines? Learn the machines' code and how they're working. Use AI today to catch up and keep up with what's happening.

Can we double down on our human skills? Those will be needed and useful for a very long time.

Can we take Salim's point of view and say we need to reskill? If you're a developer today, don't wait 3 years until you're out of a job. Think about what else you're going to do and start reskilling yourself.

I know that we all want to sit back and complain, saying, “I didn't elect Sam Altman. Why is he doing this to my life?” I want to do that all the time too, but it isn't going to help.

People who create things of this magnitude should be accountable. But at the end of the day, I need to focus on what's happening today.

I'll give you a good example. As an author and thinker, my job was to adopt a concept, think about it deeply, and write about it. That's gone. I'm no longer the most intelligent being on the planet who can adopt a topic and write about it better than AI.

That means I have completely changed.

I will not publish my books as paper—maybe eventually, but I'm not publishing them that way anymore. I'm doubling down on human connection.

Alive will be published on Substack first, with an opportunity for everyone to engage with me, discuss it with me, give me comments, call me an idiot, and improve it together.

I'm writing the book with an AI. I'm not asking the AI a question and then presenting its answer as if I'm saying it. I'm literally chatting and debating with the AI in some of the books, sometimes proving it wrong and sometimes having it prove me wrong.

That aligns with the new world. The world is changing, and the career of an author is being redefined. I am being redefined with it.

I ask everyone to look at their lives today and say, “I'm going to redefine myself. I'm going to be ahead of that wave, and in the process, I'm going to act ethically so that this wave becomes a utopia, not a dystopia.”

I love that, and that's a beautiful place to close us out.

There's so much more to go into, and I look forward to the next conversation. Mo, I'm excited to see you onstage in just a few weeks. Salim, the same for you, brother.

For everybody listening, we're entering uncharted territory. It's a territory where what we say, how we interact with each other, and how we interact with the machines that are coming is extraordinarily important.

I hope this conversation has given you a little bit of context to prepare you, but in particular, to give you agency to help steer where this future is going.

There's no on-off switch. There's no velocity knob. The best we can do is steer the future that we want.

Love you guys.

A pleasure, as always.

Thank you. It's been a joy. Great conversation.