So I'm curious, how do you think about the cost of compute versus cost of labor?

I will say that as a maxim, history is not necessarily a good predictor of what the future is going to look like because, you know, you have to update your priors.

Things change.

With a lot of technologies that fall within the realm of things that are automated, with some time, the cost of running something automated increases.

versus the cost of human labor always inverts.

So the cost always starts high and ends at zero, right, or negative relative to what it would cost for a human.

That was true with the printing press.

Are you nudging your companies to like experiment? From my experience,

But there's a more positive version of this AI future which says, yes, some of the more simple or administrative tasks get done by technology and that frees us all up to do the more challenging stuff and we can create more and be more ambitious.

I cannot emphasize this enough.

All of that abundance that you're hearing about comes at a cost.

It just does.

So fine, we've got all of these lawyers who no longer have to toil through reading things manually, doing all this stuff.

So that's great.

So what does that organizational process look like inside of the organization? We don't know yet.

So we moved on to what we're all supposed to do when faced with such an array of disorientating circumstances.

So what have we lost in the last better part of a decade, not acting on what you suggested and coming to it at this point?

It's really tough to allow a technology to develop and at the same time dictate how it should be developed.

But there are examples of how to do this in other fields, one of which is biotechnology.

So there is a process in place.

You can't just go into a lab and tinker with genetics, totally on your own.

You can't do that.

There are rules in place that allow innovation to happen and basically regulate the end product.

What do you hope our listeners take from this moment? How should average citizens be thinking about what's going on and what agency they have in all of this?

So what have we lost in the last, you know, better part of a decade not acting on what you suggested and, and coming to it at this point?

It's really tough to allow- A technology to develop and at the same time dictate how it should be developed, but there are examples of how to do this in other fields, one of which is biotechnology.

So, you know, there, there is a process in place.

You can't just go into a lab and tinker [laughs] with genetics totally, uh, on your own.

Like, you, you can't do that.

Um, there are rules in place that allow innovation to happen and basically regulate the end product.

We just never did that, um, when it comes to artificial intelligence.

What do you hope, you know, our listeners take from this moment? You know, how should, like, average citizens be thinking about what's going on and what agency they have in all of this?

Hmm

... combination of, um, signals that sort of come together over time.

There's a whole data model behind it.

So these are just individual, "Here is a thing that we have noticed and the trajectory that we see it headed in." And at the moment, you know, we've got a thousand plus persistent trends that are evolving as they're emerging that we're tracking because we have to track that for the-- for as building blocks for other inputs to our work.

A convergence is the intersection of multiple forces and trends and uncertainties that intersect and create the sort of net new.

Uh, the combined impact is greater than any one of those things alone.

And their system-level changes.

How do you reconcile some of this, like, big thinking with, like, "We gotta be on the ground, like, doing the work"? Are you nudging your companies to, like, experiment?

And it feels like AI has just become this accelerant that people don't know what to do with.

Let's click back a little bit before we go forward.

So first of all, you mentioned the printing press.

And this is this influential moment in time when if you go back to the Gutenberg press, at that moment in time, you would have had one or more, probably several people literally putting – Tiles together into a press so that you could much more quickly than a human scribe would have been able to do mass produce, mass being small scale at that stage, leaflets and pamphlets and books and whatever it might have been.

And that did fundamentally shift the course of humanity because it meant that there was much more supply in the marketplace of printed materials.

And that galvanized education in new ways and religion and the places where people would learn how to read.

So in a way that printing press became a platform for additional innovation.

But if there's a fundamental shift to the operating system of society, how do we even square that? How do we even consider what that will mean?

What, to you, does a good future with AI look versus a bad one? And where are we now?

Yeah, look, a good...

Where are we right now? Depends on what country you're in.

In my country, we are not in a good place.

What would good AI look like? First of all, having a...

Agency that is charged with reviewing the data and the algorithms, making sure that what's being developed is in the best interest of the public, and having some system to enforce.

So think of like the IAEA, which regulates...

And there's entire conversations around is grades, is going to a university in the end going to be necessary, yes or no? What do we actually need to know? What is the most important of three skills,

I mean, if open AI goes belly up, I know there'll be a huge financial.

Well, there's a couple of things to consider.

First of all, everybody's making an assumption that the way that AI is growing and the power needs that it has today is what will be required in the future.

And we know that's not true.

We know that the existence of different types of machines, quantum machines, for example, enable AI in different ways that use less power.

We know there are other types of compute that aren't here yet, but that are coming.

And so it's not the same thing as standing up a steel refinery, knowing that you're going to need steel for the next 60 years.

And obviously open AI has some vested interest in making that happen.

this week? Yeah.

Um, it depends on who you ask.

From my point of view, uh, there's a big war happening that the United States launched so there's you know, that and China and Iran historically have relations whereas we, we don't.

So I think that there was some amount of, uh, let's see if we can figure some of this out.

Um, Ch- China is very, very in- where this is a little bit of, um, a little bit of 1914, uh, where we had all of these hotspots around the world starting to, to heat up because there were these political, um, political problems all over the place.

They're...

So it feels a little bit like that right now and, um, you could, you could look at this as like she may, she- China may see this as an opening to take Taiwan by force, which they've certainly talked about.

And obviously, OpenAI has some vested interest in making that happen.

... not everybody was here the last time, so.

It's a, um, terrific organization that brings together American leaders and Japanese leaders, um, who actually live in other places around the world besides just Japan and the United States.

And the point is to establish a network and relationships for the purpose of bringing the two countries more firmly together over time.

Um, and so...

And there are some...

I, I've...

I guess this is going almost my 10th year.

And obviously, OpenAI has some vested interest in making that happen.

I wanna hear why.

You know, everybody who's probably listening to this is either a leader or on track to be a leader in some kind of organization, and undoubtedly you have noticed that sometimes organizations stay on the same path when it's pretty clear that that path is going nowhere, and the world around them is changing, and what they need to do is either adapt to that change or reinvent for that change because that change is inevitable.

There were so many fundamental changes happening that it seemed inevitable that those changes would coalesce and s- sort of form new things, and it reminded me of Joseph Schumpeter, and then I found this book, which I read.

It's actually part two of a much bigger tome.

It's Capitalism, Can Capitalism Survive? It was written in the '40s, so it's, it's, it's heavy, but the third chapter is called "The Process of Creative Destruction." Basically, what he describes is through this convergence of different forces, typically technology-focused, but others as well, the old structures are destroyed and something new hits and exists in its place.

That is always happening in some way, but it's really intensified over the past few years.

In our case, as an organization, it just did not make sense to continue publishing a gigantic static report.

Right.

Is that fair? And where do we go from here? And what does that mean for tech and business in the next handful of years?

We just need time.

This is the thing.

Are we going to get to a place where We have tools that perform, like right now, there's a whole bunch of tools.

Most of the tools perform between, you know, like they can do the same level as a person or, right? So the definition of artificial narrow intelligence was, can you do a singular task or a narrowly focused task at the same level or better than a person, right? And the answer for a long time in lots of places has been yes.

I mean, that's GPS.

That's your anti-lock brake system in your car.

Where should we go next? What, what area within that, you know, do you think needs to be top of mind, you know, maybe for business leaders to better get a sense of what's coming in this convergence landscape?

Who's winning?

Right.

So, again, I think at the moment China is winning, which is a controversial view.

But let me tell you why I think this way.

China has spent years investing in infrastructure.

The United States has really had sort of a freewheeling innovation focus.

And the challenge is we've got a handful of companies that are making investments.

what's at stake, because we all like to talk about a race, but many of us forget what the winning of it really means and why many in Washington and more broadly in Silicon Valley and everywhere are worried about China leading when it comes to artificial general intelligence or even superintelligence.

What do you think about the realities of that?

Well, I think at the moment what we have is predictability, where all we had before was growing uncertainty.

So this could ease some short-term supply chain fears.

So that definitely helps companies like NVIDIA and AMD and other U.S. semiconductor toolmakers who had been frozen out of Chinese markets.

But I also don't think that NVIDIA is going to be allowed to sell just whatever it wants.

That said, given Nvidia's new $5 trillion valuation, which also happened this week, this is really, really important, not just for investors, but for our entire economy.

I think something like 40,000 companies use Nvidia GPUs for AI and for accelerated computing.

The idea being that they can actually go it alone.

Mm-hmm.

And I had a professor, um, who was very adamant that we, we did not use this advanced calculator technology.... uh, (laughs) you know, and I think everybody's been s- through some version of that.

Um, a calculator is just a brick if you don't know basic math, uh, and in a way, AI is also kind of a, a brick if you don't know what to do with it.

But it is becoming easier to know less and be able to sort of do more.

So, this is really up to everybody and it kind of relates to the vibe coding conversation earlier.

You know, one of the key pieces of missing information is not what do you put into the prompt window, but as a company, how comfortable are we dumping in the corpus of our own data so that people can build stuff with it? I'm much, much more concerned about the lack of depth about what happens when you sell or license, um, your data or you allow employees to dump your data into a, a corpus or to make their own corpus and then throw, like, a LLaMA on top of it, a, an open source tool, um, what the potential consequences of that in terms of lost revenue in the future or lost ability to make revenue or now all of your stuff is out there for anybody to use.

Those are the kind of fundamental thinking questions that I think we s- should be spending more time with.

So if AI adoption is inevitable, I'd like each of you... what is the smartest way for leadership to shape it rather than chase it? Amy and then Sayash and Rajiv.

Yes.

So, just as the Cold War was heating up, um, there was a man named Herman Kahn who had been hired by the Air Force to help predict the aftermath of a full-blown nuclear war, because at that point, the United States was building up a gigantic stockpile of warheads, as was the Soviet Union. So, it was a very dangerous, horrific time. Um, at that point, Herman Kahn was trying to explain, "Look, there are probably many different futures depending on the decisions that get made," but military strategists only saw two. We continue to buil- everybody's continuing to feel anxious 'cause we have all of these warheads, or total annihilation. There was no in-between. Herman Kahn did not want total annihilation, but also couldn't get these people to change their minds. So instead, he borrowed from Hollywood. He knew that the data alone you could use to predict, but it wasn't enough to influence the decisions that a leader might make. So instead, he came up with the idea of a scenario, a scene, and started telling stories about what might happen in the aftermath of an attack, and these stories were so visceral, they weren't emotional, but they evoked emotion. They were visceral, they were detailed. He described every child's lunch, their milk containers would explain exactly how much radiation was in it, and- and what that might be like, and his approach was to describe not total annihilation, but a world in which we all survived. So, the way that I like to describe what I do, because there is a predictive piece of it, and th- there is a heavy amount of quantitative and qualitative research, but none of that matters if we don't influence how people make their decisions. The future is not aspirational, it's not optimistic or dystopian. You have to take a pragmatic approach because the future arrives through the decisions that we make in the present, and oftentimes, those decisions are the result of how people feel in a moment, which could be influenced by 100 things that have nothing at all to do with the decision that they're actually making.

So I, uh, when, when I began to think about this conversation, I called my husband, Jack Soll. And Jack is a psychologist who works in collective trauma and collective resilience worldwide. And basically, he said... I said, "I just came out of this presentation, and I'm, I, I'm really going through this rollercoaster, and I'm wondering, does she actually think about the emotional response of the people who are listening to what she's saying? She's taking us on the..." I thought it was brilliant, eh? The, the, let's be really clear. And then he said, "Look, when people are traumatized or when they experience threat, their imagination is constricted. And so instead of just talking about future, try to think about the preferred future," which I think is what you are actually aiming for, and I think is what Frederik is also alluding to. And a preferred future is a future in which you participate. But in your scenarios, if you work with people who are themselves numb and not allowing themselves to experience the consequences of their actions, how can they actually make decisions?

So, in my world, the concept of a preferred future came from Herman Kahn. So that was the, the man that I mentioned at the beginning. This is not, uh, necessarily optimistic. It is, again, a little bit more pragmatic. So given what we can know to be true today, the data that we have access to, we can't just imagine a total optimistic utopian future that is very likely implausible. Um, so while it may feel good to do that, in practical terms, we're not gonna get there. So a preferred future is, given what we know to be true, what we can control, acknowledging what we can't, what is our best possible outcome at that, at this moment? And I think that's the piece of this that people, um, certainly in business and government, but I think everyday people get wrong. There is a enormous difference between aspiration and action, and it's good to create a vision, you know, for what you want the future to become. But it has to be rooted in reality, and rooting things in reality sometimes means acknowledging pain or discomfort or the facts as they exist that may not align with your worldview. That's a very tough thing to do. And when we advise people to rehearse futures, reh- by, by which I mean almost, you know, play out a movie of how things, the, the many, many different ways things could turn out, that starts with asking the question, "What if?" And it may seem like a simple thing to do, to ask, "What if?" But asking, "What if?" for real is a radical act. Um, doing that on your own is difficult. Doing that in a team of your peers, or in a, if you're, you know, with a lot of other people, asking, "What if?" and having a real conversation, that is a hard, hard thing to do. But it is necessary if you want to get to your preferred version of the future, given where we all are. And that's true whether that's your personal life, you know, or your business, or your government, you're working together as a group, whatever it might be.