Unknown

>> Nice video. Arvind Krishna. Thrilled to be here with you.

And not an avatar or a chatbot or anything like that. There's still nothing quite like being there and talking to the carbon based organisms, so thank you. But before we get into AI and quantum markets are tumbling again today, there's a lot going on in the world.

IBM's been through a lot. Are we going to be okay right here? Because there's I mean, the AI trade, IBM's holding up pretty well, but the AI trade kind of falling apart.

There's questions about whether it's worth it. What's your take? >> First, John, it's a pleasure to be here with you in the flesh.

But short answer. Yes, we are going to be okay. A bit more context.

If you just look at the overall demographics, unemployment, productivity, the strength of the US economy that tells me we are going to be fine in the medium and short term. Could there be a little bit of short term uncertainty? Perhaps markets don't like uncertainty.

That leads to what we observed yesterday and today, but I think we're going to be completely fine in the medium and long term. What about this idea that technology and data was bringing the world together right now, this idea of globalism? I mean, it used to be a good word.

Now it almost seems like a lot of people are making it a bad word. Is the world coming apart? IBM does business.

Does business around the world? Is the world coming apart because of data or coming together? Coming together?

Look, geopolitics kind of stayed out of our way for about 80 years. More or less. Yes, almost exactly 80 years.

And now geopolitics is back. That doesn't mean that globalism is dead, but there is going to be we have to worry about the geopolitics. That means that the supply chain routes are going to change a little bit as different economies come into the global economy.

How do you include them? Which block are you part of? So I actually am a firm believer.

I think it goes all the way back to the economists who studied global trade in the 1800s. And I think their perspective was every 10% increase in global trade leads to a 1% increase in local GDP. So if we want to really optimize, even for local, you've got to have global trade.

All right. Call Canada. Um, so so now let's talk about AI from a perspective of so you're five years as CEO this year, I think you were named CEO in January 2020, where it was roses and rainbows, right?

But then as soon as you drive the car off the lot. Right. But you get a pandemic and but IBM's done well under your leadership.

And you had this, this vision that was kind of unconventional at the time, that IBM wasn't going to try to be the one cloud to rule them all, that you were going to be that glue, that helper in the middle to help companies use hybrid cloud to help companies use multiple clouds. How did that set you up for AI today? Look, so 93% of the fortune 500 use our hybrid products.

I think that's a pretty good endorsement of a hybrid strategy. And that means you've got some stuff in the cloud and you've got some stuff on premise. You've got some stuff in one cloud, you've got some stuff in a second cloud.

You've got maybe some SaaS properties, software as a service, which could be on a third cloud or actually in their own cages someplace. And you've got a lot of stuff still right now, 50 to 60%, maybe more, still sits in what are called private data centers. So we held the view that this will be the world.

And actually, as opposed to one cloud ruling them all, it will be multiple clouds. Now you add sovereignty and other countries. Let's not forget 80% of the global GDP sits outside the US.

And now each country, or at least some of those countries will want their own. That's the world that we live in. If that's the world that we live in, then you have to give people the software and the technologies to let them live well in that world.

Automate. Get the productivity, but still be able to live in the environment that I just described. So that's the world we're in.

And you now say, how does it set you up for AI? Only 1% of enterprise data is so far captured. Enterprise, not public data, has been captured in AI models.

How do you unlock all that data? You heard the video briefly mentioned L'Oreal. Well, they're not going to go take all of their 100 150 years now of learned knowledge on cosmetics and which chemicals work on humans and just put it out there.

They want to build something that they can leverage and use, and I think that's a perfect example. Every enterprise we talk to has the same, hey, I got my secret sauce, how do I unlock this? And if you can work in a hybrid environment across private and public, but be able to get them to unlock the value from that data using AI, that's kind of a perfect setup.

Very exciting for companies. What about for people? What about for right?

Because there's a lot of talk about eliminating the need for for people. And that's exciting. As long as the people who they're talking about are not you.

Right. So there's even, you know, Dario Amodei, who runs an AI company, was out over the past day or so saying that in six months or so, maybe I will be writing 90% of the code. So we're not just talking about, you know, replacing white collar jobs in general.

We're talking about code replacing the coders. Is that going to happen at that rate? No.

So I'll be simple. No qualifiers. No.

Okay. So first of all, I think the number is going to be more like 20 to 30% of the code could get written by AI, not 90. Are there some really simple use cases?

Yes, but there's an equally complicated number of ones where it's going to be zero. So you take the blend and I arrive at about 30%. But let me go further.

The naive person might think that that means you don't need 30% of the programmers. Go the other way. If you can do 30% more code with the same number of people, are you going to get more code written or less?

The answer is actually more, because history has shown that the most productive company gains market share, and then you can produce more products, which lets you get more market share. So actually, I think this will turn around to saying if if programmers are that much more productive, whatever, be the number 20, 30, 40, you're going to get even more done so you can get even more market share. So yes, there will be code written by AI.

So I feel great for the programmers after you say that. But now, now do artists do creatives, right. Because there are only so many hours in a day that you have to consume content.

And now we've got Dall-E. We've got, you know, all these different, you know, Midjourney creating imagery. We've got models creating music.

We've got now video, cinematic video, AI driven simulations. Are the creatives going to be okay? Is this a tool in their hand or is this a replacement for them?

It's a tool. Let's be straightforward. I mean, I heard all these debates about calculators when I was entering college.

We had these debates about Photoshop. That was 30 years ago, if I remember. Right.

So I think we got to go and maybe unpeel it one layer. I think one should be very careful about intellectual property, likeness, voices, style of art and so on. And I do think that's an unresolved question.

So park that on one side. Then you can say, but if the quality that everybody produces becomes better using these tools, then even for the consumer, how you're consuming better quality, that means that those who have a great idea, but may not be the best renderer of something, can leverage a tool to get their idea that they can describe rendered well. That will just improve the quality of what we're consuming for everybody.

And by the way, I'm not sure that there's only so many hours in the day. If I've watched the number of hours of video on YouTube or on any of the streaming channels, I think this shows that there is a lot yet more to go, maybe ten times more. All right, there is a lot to watch if you if you binge.

now, take me into the lab. Behind the scenes. You see a lot.

You partner with a lot of companies. IBM's got a lot of smart people working on a lot of things. AI wise.

What are the the breakthroughs, the possibilities that we should be looking forward to over the next, say, year to three, not even 3 to 5, 1 to 3 years. I think that the very large models have done a wonderful job in exposing all of us to the technology, in exciting all of us by what we could get done. What they lack is domain specificity.

What do I mean by that work we're doing with Boeing to figure out how does corrosion occur in airplanes? If you can figure that out, that's both more productive and make makes for safer aircraft. In order to do that, you need to have domain knowledge.

How do we bring domain knowledge into these models so that we can answer these questions? And the second is how do we make them a lot more affordable? You look at some of the data on the current models and where they're going and with reasoning and how they're becoming 100 times more expensive than an already expensive model.

The amount of energy consumption I'll predict for you that five years from now, most of the models are going to be using 1% of the energy of what they're using today. Is that. That's a.

Is that what Deep Sea gave us a preview of? Absolutely. I think Deep Sea gave us a preview that you can live with a much smaller model.

And if a model is about one tenth to 1/15 the size and equally effective, it is 1% of the energy cost. And that's the Chinese model that came out kind of surprised a lot of people. Have you seen enough kind of post the initial shock of where did this model come from?

What was it built on? What can it do that you have some some confidence and clarity around how it actually was built, what it was really running on and what it has communicated to the technology ecosystem. I don't think any of us have complete knowledge of it, and they've been pretty closed mouth about what they did.

However, I think some things are very clear because we can replicate it and we've been on this path ourselves for about two years now. Can you build smaller? I'll use the word distilled models that do what deep Seek does.

Absolutely. I think what Deep Seek did was prove that in a really big way for the globe. Now the question arises still, do you still need some really big models to start from?

And I think that that is what they didn't talk about. And that has led to confusion around, well, do they use a big model or not? Because maybe they kind of like looked at the paper of the student next to them for the test, right.

The distillation. Or maybe the professor who taught them was the big model. I go to the word.

I like the word teacher model that maybe the teacher model was a big model, and they chose to ignore the training that went into the teacher model. But what they did was the student model in some sense is my conjecture. We don't know that for a fact, but at least based on what we've been able to do.

Our conjecture that as the reality. Got some notes on the test ahead of time. However, whatever sort of school high school movie you want to somebody.

Give them a hint. Yeah, a little bit. And maybe you won't always get that.

So down to the industry level in medical, right. I talked to a lot of companies that are using AI to help doctors focus more on the patient and the patient's family in the room versus on taking notes. Um, uses beyond that.

What are you seeing in the areas that are going to make a difference, not just to quality of life, but to existence of life? Look, I think that the application of AI for drug discovery, I think, is a given. Not so much for the trials.

I mean, so I parked the human safety side to one side, but there is an equal amount of time spent in that first five, six, eight, ten years. If you can shorten that down to a few months, I think there will be tremendous, by the way, in the very phase one of clinical trials may be there also. We had a great use case.

I always hesitate about alcohol to use the word ad hoc examples, because it doesn't mean it's a scale yet. We worked with the Cleveland Clinic and they came out with a brand new immunotherapy that they're going to test because it can target certain cancers. And that was done using I discovery.

That's an example where you can speed up the first few years of work. I think that's tremendous. On the safety side, I do think that the human body is very complicated.

So that is why we have such a rigorous protocol. And I think that has served us well, and we probably want to live with that now, even within that. Could it speed up in how you read the reports?

Could it speed up and how you aggregate the information? Could it help you in analyzing? Hey, if we can take maybe half the time out of the second and third phase, that's a great benefit.

But to your point about it being an aid to doctors, to clinicians, to practitioners as we begin to improve. Can it be an aid, not replacement to radiologists? I think these are all wonderful use cases that should come out over the next few years.

You mentioned drug discovery. I can't help but notice that's one of those main areas that people were talking about quantum for. And as all of this conversation around generative AI and adoption, the enterprise and Nvidia GPUs and all this stuff, there's maybe a little less of a spotlight on quantum.

Is quantum still? Does it still, you know, excite you? You get up in the morning thinking about quantum.

John knows me too well. Look, I'm incredibly excited about quantum. There are some who go out there and posit this is 15, 20 years ago.

I'll tell you, it's going to surprise you before this decade is up. That gives us only four more years. Just to be clear.

And I think that we're going to see surprising things about quantum in the next four years materials, maybe carbon sequestration, uh, optimization, pricing, maybe fertilizers and food. And when I say materials, simple molecules, maybe not complicated molecules, simple molecules. But 90% of all pharmaceuticals are simple molecules.

So I mean, but in perhaps you have an infinite well of excitement, just like the hours you have to binge streaming shows. But, I mean, I wonder how quantum ends up interacting with AI, right? Because AI is is sort of guessing at and maybe getting close enough some of the things that quantum was going to have this, this odd level of precision, right, at answering.

Actually, I think that there are two are going to be completely complementary to each other. Let's remember at its very heart, AI is learning from already produced knowledge, literature, graphics and so on. It is not trying to figure out what is going to come.

I use my simplest example. Most people, probably not everybody, but most people either drink a soda or they drink coffee, or they drink tea. The thing in there is the caffeine molecule that gives you energy in a burst of energy.

We know the molecule gives us energy. We have no idea why. I don't think AI is going to tell us this is the shape of the caffeine molecule.

This is how the 160 electrons, if I remember right, 160 electrons in it behave. And how they're this thing. That is what quantum is going to do once we get down to that level of subatomic knowledge.

It tells us how nature behaves. Once we know how nature behaves, then we can learn a lot more, because none of that is in the existing knowledge. That's why I believe the two complement each other as opposed to compete with each other.

How nature behaves is a great mystery. How government behaves is also a great mystery. So we're going to bring together this mystery of quantum and this mystery of government with a special guest who's teaming up with you on quantum.

Let's bring out Illinois Governor J. B. Pritzker.

Governor. Governor, thanks for being here with us. So explain the mystery of how government behaves.

No, I won't hit you with a question that hard. >> I love that as an introduction. A big mystery.

And to explain it all, governor JB Pritzker. But really public private partnership, right? You and IBM are getting together, making investments into the development of quantum.

I think we're at a stage right now as we're watching this Trump administration that just got started a few weeks ago, highly transactional, right in and really scrutinizing research and expenditures on science. How much of that informs your perspective? How different is it when you, as the governor of Illinois, are looking at investing in quantum?

Well, let me back up just for a moment. You know, I started as governor when President Trump was in office in 2019, 2020. And and when I first came into office, it was important to me to make major investments for my state in quantum computing.

We knew this was something that Illinois had a right to win around because of. University of Chicago. University of Illinois.

northwestern. We have five R1 institutions, two national laboratories, all of which have a role in the quantum future. So so we as a state tried to set the predicate to make sure that we would be a winner.

And so we put an investment of $200 million in back in 2019 as a state. I mean, so unusual was that that even the state of California, which is often known as a leader in technology, the the government leaders there, uh, found me at the National Governors Association and said, how did you get that done? We'd like to do that in California.

And they hadn't done it. So, um, so we knew it was the right kind of investment to make. Uh, fast forward again.

Our institutions are really the leaders in this. So just last year, I asked the legislature to put together a, a quantum campus, a support for building a quantum campus in Chicago. We had a perfect location for it.

We had our first major quantum company that was interested in coming company called Psi Quantum, but we also began conversations with folks like DARPA who now have committed to be on the campus. And so to me, I knew that if we're going to as a state, if we're going to leap ahead, we've got to be in the arena that's a little bit ahead of everybody else. And so Illinois is and I believe we will continue to be.

And I should point out that that the gentleman sitting next to me, Arvin, talked to us early on about the possibility of doing something with our quantum campus and actually made the commitment of a national quantum algorithm center, which will be on our quantum campus. So IBM is now a partner with us in Illinois, being a leader in quantum computing. Arvin.

Why Illinois? And talk about the ingredients for an innovation ecosystem. I spent, you know, a decade plus in Silicon Valley.

And it was fascinating to me how research institutions, sometimes government funding, investor presence, all of that. And then, you know, if you got a nice city with some good architecture, that doesn't hurt. California's missing a little bit of that.

San Francisco has a bit. Look, it's to me, it's very simple. So if you believe once in a while you get these really big technological innovations that come about, they can fundamentally improve your GDP.

They can improve the economy and, and and maybe a disproportionate share of the advantages go to the first movers who embrace them. That's sort of the ingredient. Now then you say Illinois, the governor pointed out to those institutions.

But the two national labs there, Fermi and Argonne, they're a big piece of it. If you look at which are the universities, maybe it's not such an accident that if you look at the Manhattan Project, that did happen at the University of Chicago, maybe they still have a lot of the kinds of people who did those experiments. If I look at my alma mater, University of Illinois at Urbana-Champaign, that has a lot of people who do work in these areas, the governor mentioned northwestern, but there's also many other institutions in the area who will play a role.

So you bring together really smart academics. They need access to the infrastructure. You begin to bring together programs like DARPA and other parts of the federally funded R&D, because we've got to unlock the use cases.

So what we want to do is unlock the. Federally funded R&D. I did.

Is that still a thing? Absolutely. I always say, you know, just chill.

There's a lot of ups and downs, but I think federally funded R&D is still very much a thing. I want to add that that I mentioned the Trump administration their first term, because after we made the investment in Illinois of $200 million, we knew that there were Department of Energy grants, Quantum. Department of Energy grants ten of them that were going to be granted around the country, and Illinois one, four of the ten.

And that's in part because we made our own investment, and in part because, frankly, we have the best scientists, we have the best universities for the federal government to be betting on. And yes, I mean, I'm hopeful that the administration now, in its second term, will continue to see quantum as an important investment. We have got to lead in the world in terms of our development of both AI and quantum, because it's going to determine whether our economy continues to be the leader or not in the next century.

Governor, with research funding of major academic institutions under threat now, is there anything that states can do. Send better people to Congress to to, uh. I, I we've got great people from Illinois.

Don't get me wrong. I'm just saying that we've got to we've got we've we've got to make sure that that we're working with our federal partners very consistently. And then I think what we did is the right thing.

We made our own investment, but it's a kind of a matching investment so that the federal government understands that we're not just sort of raising our hand and saying, great, toss some money this way. We're actually making the commitment we have the institutions that are appropriate to to carry that research forward. And so we're a good bet because, you know, we're a long term investor.

Arvind, how are you planning the right time to make the right size investment in something like quantum? In the past, IBM has at times tended to be get ahead of itself in the messaging about certain things. You could you could say that Watson was, we'll say ahead of its time, we'll say Watson was ahead of its time.

How do you get the timing right for quantum with how much you invest? And when you say that things are ready. Actually, I think the governor used the word the National Quantum Algorithmic Center that's going to be based in that rendering.

You saw of the real estate down on the south side of Chicago, on the shores of Lake Michigan. Why am I excited about it? Because it's an example of what John just asked on the funding.

We are investing in making quantum hardware. We'll get a computer ready. Now.

You got to unlock the people who can use it, who can do something with it. If the governor through that center can go bring startups, we'll give them access to the computer. We'll give the scientists at UChicago and UIUC access.

We'll have the folks at Fermi and Argonne figured out use cases that are far beyond what others are thinking about. That is what is going to get it going. Look, in the end, to me, value in technology is not derived from the invention.

Much as we love the invention, value is derived from people using it. We can unlock that. Then we can go get all the rest.

So to your point on, sometimes when we get ahead of ourselves, we get too focused on the invention and not on making sure that there are hundreds and thousands of different companies, private and public, that are using it. So that's what we're trying to get done in this case. And I think we're kind of going at the right pace in this instance.

Okay, governor, where does global talent immigration come in? Very often, these research institutions and these jobs that we're talking about are, you know, you're not looking for just the local talent pool. You want to get into politics.

Now. I didn't say politics. I just asked about immigration.

I asked about talent. Global talent is the right is the right focus. And I just, you know, as you say it.

And as I was listening to Arvind, I have to compliment IBM because they bring the best talent in the world. And being able to partner with a major company like IBM, a company that who's got some of the best scientists in the world focused on this. And I went and visited your your center, your innovation center.

And very, very impressive. And so talent is is where it's at. I mean, you know, as you say, you can build the computer, but you need a whole bunch of people writing software for it.

You need a whole bunch of people figuring out what the right uses are for it. And so the question is, are all of those people in the United States already? I think that's sort of inherent in your.

That's a that's a good way to take the question. And, and I and obviously there are many, many talented people all across the world that they are attracted. They want to be in the places where innovation is happening, no matter what country they're from.

If you're a scientist or you want to be a scientist, the place you want to be is where it's happening and it's happening in the United States. It's happening in Illinois. And so that that that's hugely important to me.

Now, you know, the immigration policies that are necessary for that, frankly, we've all got to focus on the fact that this is a country of mostly immigrants. My family emigrated to this country from Ukraine several generations ago. Um, you know, and by the way, fleeing, um, the pogroms.

So they were refugees, true refugees. There are many people that that and they had nothing. And they didn't come with scientific talent but had the opportunity to succeed here.

And, you know, again, the fact that we have some of the best higher education in the world makes us a hugely attractive place, uh, and for economic development purposes as well. But there are people all over the world, Japan and Europe and, and, you know, all across Asia, Africa and elsewhere, Australia, where you know that talent, we want them to come, they want to come to the United States. We want them to come to the United States.

We want we want people to come here and bring their talent with them and apply that talent. And we want to develop our own talent as well. But you can't develop it as well if you're not bringing the best people from across the world for our people to learn from to.

So we should be an international talent hub and we should have policies that go along with that. Arvind, who are you looking for as IBM right now in a global talent pool? How has the skill set, even the soft skill set shifted in this environment where we're talking about AI?

We're talking about quantum. Look, these shifts are more in the future. If I look right now.

The talent pool that is there in higher education, because that's where we draw the kind of talent for AI and quantum from is still very much there. Are there some a lot of statements are being made about shutting this down, but the ones coming into college are not yet shut down. So we can draw from a global pool.

You can also draw from a global pool when it's remote work that can be outside the country. So that's a second. I'm actually optimistic that for the high skill sets and there is enough noise on both sides of the aisle that you do want to preserve that talent, that we will come to some policies that will let that set go.

That is different than wide scale, national level immigration policy. So that's kind of where I end up on this topic. And I do believe having watched now personally the last 40 years, but reading history, the last 80 years of immigration, that we will find a way to make sure that the correct, talented people are here, which is different than the widespread immigration policy.

Okay. Well, governor, I appreciate you spending this time with us. I know you got a big state to run in some quiet times in America.

Very calm, very calm. Everything's good. Thanks for.

Taking. I get to sleep through it. Yeah.

Thanks for taking some time out with us. Thank you. Appreciate you both.

Governor. Thank you. >> Thank you.

Well, Arvin, now, I think we're going to take some questions. Love to. Yeah, I've got some questions from the audience while those are coming up via Slido.

I want to kind of close out that conversation on quantum. So you say before this decade is done, we're going to see real progress on it. Can you give me a little bit of a teaser?

Sure. So quantum computers for the audience are a little noisy and are a little probabilistic in how they work. You can get them pretty large.

You could get the measure that is used is often qubits or quantum bits. You could. We demonstrated one.

That's a thousand qubits. Big problem is it's got a lot of errors. And it's got what is called a coherence problem.

Meaning it's not very long before it becomes complete and utter noise. You've got to solve those two problems. I believe that we've got to get ten times better, not 100 times better on the error rates, and we got to get ten times longer on the coherence.

Okay, I just saw the toughest question of the day pop up here on Slido. Since we're talking about quantum I'm going to read it. Don't be mad at me, Arvin.

I think you're up for it. Here's the question in one simple sentence like Sesame Street. That would be a word of the day.

Maybe as opposed to a whole sentence, technically, but. Okay. Uh, can you just disappeared?

Can you explain what quantum computing is? Oh, there it is. Right up there.

Absolutely. So normal computing. Think of it as high school algebra on steroids.

It's completely deterministic. You can repeat the same thing if you guys remember high school algebra. Add two numbers.

Square something. You know what I call deterministic functions. That's high school algebra.

Even I, which has got really complicated linear algebra inside at the core of it will come back to high school algebra, where people in college normally lose computing is when it becomes probabilistic. When you start talking about words like find me a minimum energy state, it becomes very complicated math and becomes almost impossible to solve. Normal supercomputers can't do it.

Quantum does it inherently, by actually trying to leverage what are quantum properties of subatomic people? Talk about it being probabilistic. That's maybe one way to look at it.

So it's trying to find and solve these problems, not by trying to deterministically go towards the solution, but by saying, if I look at the problem, this is a state that answers it. Some people try to say it looks across all the possible answers and picks out the right one. That may be an okay way to think about it, but that's not quite accurate.

It's sort of converging on that answer as it operates as a way to describe it. Not quite Sesame Street, but maybe. I was going to say you lost Elmo when you said probabilistic.

Yeah, but Elmo forgives you. Okay, um, we'll move on to another question. High school algebra is like Sesame Street for you, Arvin.

Yes. All right. What will be the impact of quantum computing in the day to day life of people and companies?

Not in tech, not in research. So. So actually, the entire value goes to people.

If there is no value to people, then there is no value that flows back. I'm very, very clear about it. A simple example if you look at the whole financial industry, how do you price if you look at what was happening yesterday and today with the choppiness in the markets, there's a lot of pricing that some people made a lot of money on and some not.

It is based on something called Monte Carlo simulations. I believe quantum is going to let you do them much more in real time than overnight or once a week. That's going to give an advantage to everybody.

Then you say, how do I benefit as a person in the audience. Well, if you get lower interest rates, you benefit. If you can get lower pricing for mortgages, you benefit.

So there is a benefit to everybody. If all of that can sort of calm down and be much more real time. Materials are talked about carbon sequestration, but how about a lighter weight alloy.

So we don't need that much energy for transportation. That's a second I think. Great use case for quantum computing.

So these are some of the examples that will sort of apply for everybody. Well let's bring it closer to home. Somebody's got a kind of personal question here.

What advice would you give to I don't see the question. But it was like, there it is, a first year computer engineering student looking to stay ahead in the space. I would tell you we have a technology called Qiskit.

Others have similar on ways to interact with the quantum computer. You then don't need to understand all of the quantum physics or all of that math that is down there. I would recommend, much like the first year student learns typically Python, Java, one of those languages.

Go learn how to leverage a quantum computer using some of the existing libraries. It will give you some sense of what quantum computers can do. Equally, it will tell you what they cannot do, and that is one way to start to go without having to delve into the things same way computer engineering students learn how to live with transistors and chip design.

Very few of them are getting down into the quantum physics of transistor design, but they kind of still get a sense of how to use them. And that's what I would recommend. Is quantum going to have a ChatGPT moment?

I mean, here's I think the essence of this question from Kartik. Quantum computing lacks clear public use cases, unlike AI, which shows immediate benefits. How do you think companies can improve branding.

To bridge this gap. Give the marketing department at IBM some help here. Actually, uh, personally, I think that branding to bridge the gap is not something I would do to your point, which you made earlier of letting the marketing get ahead of the reality.

Let's first let the tech get ready. Is that three years? Is that four years?

At that moment, how do you define. It's ready? Is the way I would phrase the question?

It's not ready based on what we say. It's based. It's ready based on it's somebody at that point solving problems that are of incredible interest.

That is what I define as getting it ready. So our efforts, whether in Chicago or working with the federal government or with the universities or around the world, is to make people do problems on quantum that actually show the public benefit and the use cases. Once those are big enough so you can do them today, but they're small, so they're not that interesting.

If they get big enough, that will define a movement that will take the world by storm. Then we can begin the marketing, not before that. Here's a question that stretches across all of the topics that we've been talking about AI, quantum, etc.

because energy plays a role in all of it. Natural resources play a role in all of it. What's the role of water and how is IBM addressing the water shortages?

Water access in the areas where these resources need to be available. So first, where where quantum is concerned, it actually uses only about 100 of the energy of AI. So it does not really have all of the same issues.

And it does not have the water issues of liquid cooled data centers. It's got to be very cold, right? It's got to be really, really cold.

But then the amount of energy it uses is very, very tiny. So you cool it down. But it is cooled using actually liquid liquid hydrogen, liquid helium.

And you use metal plates and all kinds of complicated thermochemistry to get it down. But we're talking tens of kilowatts. We're not talking megawatts at the scale we hope to get to maybe megawatts but not gigawatts.

So and that is mostly for the cooling. Running them is much, much simpler. So that is just a sort of statement of fact.

Okay. But if I go back to your broader question on water and energy usage, I firmly believe that we can come up with semiconductor technologies. So that's not even quantum.

That's for AI. There are literally 100th the power consumption of today. And as we do that, then it will reduce the weight on both energy and water that we need in this world.

And just that's one of the reasons that we are working on climate models with NASA, in the hope that as you begin to get better at using AI, then that AI can also be used to mitigate the use or misuse of water in certain places. Where should you grow these crops? Where can you not have a massive impact, etc.

? And so I do believe that it's climate, more broadly speaking, than water, which will play a role going into the future. Well, this question could have come from Oprah, but it didn't.

It came from JC. I like it for a change of pace. What keeps you up at night?

You're so chill. What makes you feel nervous about the future? I think what keeps me up at night is always talent more than anything else.

Do we have all the right talent? Are we offering an environment that is very attractive to people to come work? I kind of say we work on problems that are really meaningful to the world and that are really hard.

We've been talking a lot about quantum. It's a program we've been doing for 15 years. It's not an overnight thing.

We didn't talk about it ten years ago. Why are those people here? Why are they working on it?

They love working with other smart people like themselves. They observe the improvements that we make. I'll give you a shocking statistic.

In some areas we come up with a better technology deep down almost every month. That is what keeps them going. Now, how many areas can we do that?

And that is what is exciting for our people. That is probably what keeps me up at night more than most other things, really. Now I have a hard time picturing you popping your eyes open, worried about the talent, like at 3 a.

m. ? Is that really?

Is that keeping you up? When I wake up at 3 a. m.

and I'll admit there is a once a week, then I will wake up and my eyes are open and I'm not going to go back to sleep. It is probably talent and it is probably the other one I was going to say is some of the state of the geopolitics of our world. Okay.

All right. Good, good. You're like the rest of us.

In that way, I can I feel I feel good about that. All right. What's the most interesting unsolved problem in the AI quantum space that could unlock transformative breakthroughs for technology and society?

Look, I think that AI is going to get better and better. I am one who does not believe that the current generation of AI is going to get us towards what is called artificial general intelligence. All right.

What is art? I know it's part of AGI is that it's not clearly defined, but kind of for the uninitiated, for Elmo, still, what is artificial general intelligence for the. For Elmo.

AGI is when the AI can have all knowledge, be completely reliable, and answer questions beyond those that were answerable by Einstein or Oppenheimer or all the Nobel Prize laureates put together. That's a simpler definition of of I nice. Or as Cookie Monster would say, when I makes the best cookies.

Then I can make the best cookies and the machines that make the cookies. I was a big fan of Sesame Street growing up, so that's just going to be stuck in my head as as we bring this conversation to a close, what's the most important thing for the individuals at South by Southwest to keep in mind at a time, I would say, of high anxiety about multiple things, about the role of technology, the role of business, the role of government, the geopolitics. Look, I think what we all strive for as a society is to improve the quality of life.

And I'm not going to define that. That's because each person has some of their own definitions. How do you improve quality of life?

You've got to improve productivity in society and you've got to be able to give democratized, democratized access to those benefits to everybody. Note I didn't say exactly equal, but democratize. Everybody has to have access to them.

If you want to go after those two things. Technology has proven wonderful, but I'm going to define technology really broadly, whether we think about steel, which gave rise to railways, that let goods get to one place to the other, which I would say was the opening up of the United States or oil, much as we might argue about it. It gave rise to energy, portable energy, which we can use different from the place where it is, which gave rise to automobiles, which gave rise to a whole culture here, what is called ICT.

But we can make it semiconductors for shorthand, which gave rise each one of those five. This country led in having led in it. That gave rise to literally beginning to end, a ten times increase in GDP for the year.

That let us give benefits back to everybody and what we call opportunity, this is what we need to strive for. And if we can lead in this in both AI and quantum, I actually believe it'll give rise to another ten times. But equally, we should have policy, which is why we argue for open innovation, not closed, so that those benefits can be democratized and given back.

So then there's opportunity at least, and everyone benefits. That is what keeps me excited and keeps me going. I guess today's technology is tomorrow's infrastructure.

Arvind thank you. Arvind Krishna, CEO of IBM. Thank you.