Introduction

These are my slides, Abundance in an Era of Penury, the Energy Transition, the Workforce Transition, AI Value Shift.

What do I do in the era of AI? Policymakers come up to me, senior executives, partners, even government ministers. What do I do?

And as we move along, what I'm going to basically say is, I'm missing a slide here, but that's okay.

Abundance Meets Penury in the AI Era

Reasoning at scale and collapsing marginal costs

We're basically in a moment where I feel incredibly optimistic. We're finally at this moment where we can get reasoning at scale, the costs are collapsing. In simple economic terms, the marginal cost of a lot of the things that we do is going towards zero.

And yet, we are in a time of penury, of deficit, of unevenness. Not everyone, sir, is going to win.

And in fact, this transition, what do we do about AI, is brace yourself. It's going to take nine years for the economy to transition and for us to figure out what to do. do.

Accelerating scientific discovery and shorter design cycles

The thing that gets me very excited is the shortening of the time frame to do this, the scientific discovery.

The things that used to take 15 years to build are getting shortened in terms of their design cycle, whether it's a tokamak fusion reactor design in the geometry or it's a proteomic design of a certain type of amino acid or something else that's in your industry.

Because of the way in which we use tools and we problem solve and we think about bicycles for the mind if you will, we are starting to see that

engineers, scientists who work in wet labs understand how to use these tools if they're purpose -driven and they have a goal in mind and incorporate that into the design cycle.

It's very quick to see in for example in China and the guy's actually not Chinese he's Latvian started a company and it effectively has labs in Boston and in Montreal and delivers drugs that are FDA approved prove phase 2a faster than any of the other big pharmas.

How he uses AI to iterate. That's another one of the themes from earlier today.

So the scientific discovery gives me lots of reasons for optimisms. And we're moving from that lab to society faster than ever before.

The amount of days it takes for us to be able to use things is so much faster.

Designing for 8 billion people—and 100-year lives

I want to take this perspective, but we're not trying to help a certain number of users or a certain country. This is not for Singapore, it's for the 8 billion people on the planet.

And if that's your optimization function, you have to start thinking about what are the larger things that AI is going to have to tackle and solve while not disenfranchising all 8 billion.

The other thing you have to take as a given is that we are moving incredibly quickly towards 100 -year lives. The young woman born today in Singapore is going to live to 105 with a 50 % probability and 100 with a 50 % probability in probably 70 nations.

We are already in this phase where people will live significantly longer and we have to design for everybody on this planet living a hundred years abundance in a time where there's just going to be just not enough to go around.

The catch: unevenness, short-termism, and the J-curve transition

So this optimism that I have is not a given and there are constraints.

So the catch is there's unevenness everywhere. William Gibson's famous quote, which is the future is already here, it just is not evenly distributed.

And in fact what What we're having right now is that we have too many people taking a short -term view. Because uncertainty, massive uncertainty, what does it do? It makes you think about here and now and local.

It shortens your time horizon. And unfortunately, it makes the domestic more important. That's what becomes the more important win.

And truthfully, when economists talk about, oh, but it's all going to be fine later when we have more jobs and more people being able to use these tools, that transition is painful. And that transition might last an entire decade or a generation. That's what's called a J curve, and that transition is going to be difficult to swallow.

When the state of New York says it doesn't want any extra data centers, it's because of that cost associated with it in the interim period of time.

There's a lot of friction in the system. I'm going to skip past this in the interest of time.

The Two Rate-Limiting Steps: Energy and Workforce

But effectively, the two rate -limiting steps, things that we have to think about to solve, are energy and workforce.

force.

So if someone says to me, hey, what do you do about AI as a policymaker or as a CEO or a C -suite or a Microsoft customer? Basically, what I would say is in the context of where we are today, it doesn't seem as important in the trilemma of energy.

Now, let me just explain, take a step back.

1The energy trilemma (affordable, reliable, green)

I was on the board of EMA here in Singapore. What is the energy trilemma? We want affordable energy that's always available.

That's green. Okay. We want all three of those things all at once and it's a trilemma because you can't have it all.

What has happened in the last year, or since Trump got elected? Well, the green part, let's put that aside.

Straits of Hormuz, a lot of other reasons. As a result of that, our targets around plus two have already gone up to plus 3 .5.

The thought that we had two years ago that we're going to have to supply green electrons in order to build tokens is not necessarily true, even though that constraint is absolutely true there. So we have to think in terms of plus 3 .5 is already kind of locked in, and the risks associated with that, these tail risks are huge.

Deployment timelines, tail risks, and why nuclear matters

The earliest deployments of some of the solutions that we're gonna see are going to be in the early 2030s.

I will talk about that transition. It will take some time for us to solve cancer, for us to think about ways in which we could create fusion, for us to be able to talk to animals.

It's gonna happen very soon, but it'll be in the early 2030s.

And then finally, I talked about SMRs, but it's truly nuclear. If that the certain administration has done right, it's funding nuclear.

He has funded 11 projects which are going to push his country towards nuclear advances that you couldn't get in normal times.

You can guess what country I'm talking about, but it could be China, it could be Russia, it could also be Canada, South Korea.

These countries understand the importance of these technologies to basically be a solve for all of the issues associated with energy.

I didn't necessarily think it was such a complete solve I helped set the 2050 strategy for Singapore there are gonna require all four of the different valves but if you're thinking about future of energy you have to start thinking that

is my country is my is my city is my neighborhood prepared for nuclear

Workforce transition: upskilling and usage as the quick wins

because it is one of the only panaceas that we really have the workforce the two quick solves I sell people is upskilling and usage I'm going to talk about both of them but probably a little bit more from a Microsoft study

perspective and by the way I retired from Microsoft days ago it was July I was the head of corporate development doing deals like M &A deals and investments for Asia all across all of our businesses I did go to Stanford but I was an MBA not a PhD and the thing that makes me most famous is my research

at MIT around pull the goalie it's a hockey or analytics problem back to this this. Satya Nadella talks about AI as a bicycle for the mind.

Look at these two gentlemen. They're incredibly smart guys on these incredible tricycles, unicycles, and bicycles being able to take investing or co -childing to a next level. I do it too.

One of the things that I did with particularly this talk is I took a Word document which was published in a research think tank and I asked it to help me turn it into slides. That was the purpose.

Then I said, well, what's the audience? And I rewrote big chunks of it. And then finally at the

the end I had a critique and said, what would you do to improve it? I'm going to tell you what it said at the end of the presentation.

Codifiable work, automation, and who wins the transition

Codifiable work is automatable work.

Philip Hughes is going to lose 9 million BPO jobs in the next few years.

It's happening. It's real.

And the winners are the one who do two things. They re -skill first and they're incredibly fast in terms of usage.

Now re -skilling is hard, but what does it really mean? Put the tools in front of you and tell people to go do it. Not the courses, but just make the the access to the tools available and get rid of the stigma.

And in so many other countries, there's a stigma, oh, I don't want to use AI, et cetera. Not here in Singapore, because you're being recommended to play with some of these tools, not just in places like this.

The winners are lapping the laggards. What I saw at Microsoft is people coming in our room, and you could tell whether or not they were the winner, that was one lap ahead, or the laggard that would be all out of business.

And it was usage, it was skilling, It was attitude, but it was also at scale, taking some of the things that they did so well and applying small little experiments that are thoughtful about how AI could augment and add to their teams. It was clear that FWD was figuring out how to do policy management with agents earlier than any other insurance company. Do you think Access is going to do that fast?

And this is Access, apparently, its old training center. Do you think that other companies like JP Morgan are figuring out how to use AI faster than the smaller banks? Yes, indeed. We see this every day.

Divergence at Company and Country Level

The winners are lapping the laggards. It also happens at a country level. Singapore is way ahead in terms of some of the other countries.

Indonesia is actually catching up, but it's doing quite well. But divergence or basically this idea that you're not as good at AI shows up as capital flight. People leave the country. I'm originally Canadian.

Canadian, they leave Canada to go and build their company somewhere else. There's fiscal stress. There's all kinds of ways in which your economy is less dynamic.

This happens in your household. This happens in your community. If you have a community that's like this and

think, oh, I don't want to use AI, it will unfortunately lead to a less dynamic, a less adaptive environment.

Variable geometry cooperation and fit-for-purpose procurement

Mark Carney said, we need variable geometry cooperation. Basically, it says, I'm going to build a coalition around a concrete outcome and then I'm going to dissolve it.

Whatever that might be. We do this in business development in large corporations, but we need to start doing this at a policy level as well.

So if you're talking to your friend at GIC or you're talking to your friend in Malaysian think tank, they are starting to think about how do they walk this tight rope between US and China.

They build alliances for specific purposes and allow them to assemble to achieve a goal and then they break them down. So variable geometry applies all the way through

because you can't get there on your own. But you need streamlined procurement, procurement, fit for purpose risk, cutting the red tape, keeping the safeguards.

You can read the rest of this, but basically I'm going to go back to nuclear.

Public–private partnerships for energy systems

Nuclear fission and fusion are great archetypes for how you could build a public -private partnership, work on grid permitting, figure out the technology when it comes to line and transmission, maybe even work on different ways in which you can actually decentralize the grid.

Time shift, place shift electrons, there's all kinds of complexity to figuring it out, but it is not something that you do as one one individual company or one individual government.

It's going to require variable geometry. It is not the strongest that survive, whether it's a company or a nation. It's the one that's most adaptable to change. Not Darwin, but Megenson, a contemporary of his, said that.

There are no unknowable miracles required for this to happen. The things that we need in order for us to lead with science, to train up our people by giving them the tools that we talked about for the first hour,

to give them the permission to do so there was a usage slide it's coming I know where it is it's at the end I'm gonna do you mind if I jump ahead to it okay nope is an older version okay I'm gonna have to give it to you orally

Usage Changes Attitudes

usage okay we played a quiz game here do you guys understand the American left in the American right right the left the right which side uses AI the most guess just say left or right. Which side is more favorable towards AI? Left or right? Anyone paying attention?

Okay, the answer is the side that's most favorable is the one that uses it the most and that applies to every country.

Singapore leans towards usage and therefore if you use it even once at work a week or once a O'Day, your favorability rating towards AI jumps incredibly high. Look at the three of us.

Unfortunately, the American left doesn't use it, and therefore they're fearful of the AI. Even in one country, you see this massive divide, which we don't really see in any other country between the left and the right. Usage is the key.

If you start using it, if you start tinkering with that if you start with purpose, that's a lesson for everything in life, have ikigai, then you will get more comfortable with what it is that that particular tool, assistant, agent, species, call it what you will, will help you in your achievement of such goal. If you don't, you kind of get left behind.

And at a national level, and I'm going to deliver this to the country called Tajikistan in a few weeks, to the minister, at a national level, if you don't harness this, you're not going to succeed.

From Executive Assistant to AI Chief of Staff

So I'll end with the story of basically, you know, when I was putting this together, I said, okay, so this, I have open claws, we're organizing everything, told me where to show up, et cetera, et cetera, told me a little bit about the audience, and it's a good executive assistant, and it interviewed me, and it got to know me, and has some sort of history with regards to who I am, and even maybe understands a little bit more about my voice.

So at the end, I said, said, okay, let's critique this. What else could I put in to this particular presentation that will be useful to this audience?

And it said, and I don't have this verbatim, it's in my head, it said, you're missing the most obvious thing. Your AI chief of staff, you know, figured this out, did this, did this, this, this, this, this, and you're not even talking about it to this MindStone group?

I thought that was really interesting, so I'm going to tell you guys about that. This is what it said I should be talking about. But I think the most interesting thing was that it actually promoted itself from being my executive assistant to my chief of staff on the basis of how well it knows me and the tasks that it thought it was doing.

Conclusion: How We’ll Be Judged

At the end of the day, I just want to end on this, which is I think that we will be judged by this next generation. Some of you in this room are in this next generation. They're going to say, how did we handle this transition? Did we do a good job?

Because that utopia is coming. Maybe it's 2035. Maybe it's 2040.

some of these problems will get solved we figured out go we're gonna figure out proteomics biology is right around the corner then energy but right now those people who do not have jobs who are disenfranchised are not in the nations that are evenly distributed how will we be judged by the those 8 billion of the 8 billion because that is what we could be doing right now to make sure that all

the things that we do with regards to AI really lead to abundance soon whether Whether it's cheaper medicines, affordable desalinated water, reliable clean energy, all the things that we need in order to build our community from the ground up.

Let's start small with using AI from a practical and energy and health perspective to help everyone.

And that's as good a place for me to end as any. Thank you very much, guys.