Tech'ed Up

The AI Ripple Effect • Dex Hunter-Torricke (DeepMind)

February 22, 2024 Niki Christoff
Tech'ed Up
The AI Ripple Effect • Dex Hunter-Torricke (DeepMind)
Show Notes Transcript

Head of Global Communications and Marketing at DeepMind, Dex Hunter-Torricke, joins Niki in the studio to inject some optimism and hope into the doom n’ gloom AI conversations dominating the airwaves. He highlights exponentially innovative developments at the Google subsidiary like AlhpaFold, scientific discoveries of new crystals and minerals, and AI-improved weather forecasting - all of which feeds his optimism for the future of tech. 

“AI is just opening up such extraordinary new possibilities for us in this century. So, I think it's actually hard, once you understand where the technology is going, not to be an optimist.” -Dex Hunter-Torricke 

[music plays] 

Niki: I'm Niki Christoff and welcome to Tech’ed Up. 

Today in the studio, I'm talking with Dex Hunter-Torricke. He and I are longtime former Google colleagues. He's worked at the United Nations, at Facebook, at the Oversight Board for Meta. He's worked with Elon Musk, and he's currently at DeepMind, Google's AI subsidiary.

Today, we're asking Dex to tell us some positive things about AI and about tech in general.

Dex: Great to be here, Niki. 

Niki: You're visiting DC from London. I really appreciate you coming into the studio. You said when you got here that you thought I had a cool apartment, but it turns out this is not my cool apartment. 

Dex: I mean, it's a very cool studio. [chuckles] 

Niki: It is. We have a kegerator and a Godzilla poster. But no, it's not. I just realized we've not seen each other in person in like 10 years.

Dex: Yeah, crazy. 

Niki: It's crazy, but we see each other on the interwebs. 

Dex: The magic of social media. 

Niki: The magic of social media. We met at Google when you were speechwriting for Eric Schmidt. And I was doing his press at the time. We sat next to each other.

Dex: Yeah. 

Niki: And it was a little bit of the golden age of Google, I think, which we can talk about if we want. 

Dex: [lauhgs] I think we're still in the golden age! 

Niki: Okay. Exactly! This is why I brought you on. You've worked directly, especially as a speechwriter, with Mark Zuckerberg. You've worked with Elon Musk. And you're an expert communicator .

One of the things I want you to communicate today is the positives of tech because I think I've become, and by I, I mean Washington, D. C. is a little too cynical. [Dex: laughs]

And I want to get back to the, the optimism about tech because I think we both, when we fell into tech and both of us fell in from other non-tech worlds, we really were optimists and I think a little bit too much cynicism has come through.

You did a TED talk in 2018 and you talked about a lot of positive things in, in tech and trends. A lot has happened since then - you remain optimistic?

Dex: I do. I mean, if you think about technology and like the role it's played throughout history, it's always been this motor of human progress. And we're now in an age where, almost every day, technology is giving us these extraordinary new possibilities and a whole bunch of different fields. And I think,  AI in particular, and - [interrupts self] of course, this is what brought me to DeepMind and all the work i'm doing now. I think AI is just opening up such extraordinary new possibilities for us in this century.

So, I think it's actually hard once you understand where the technology is going, not to be an optimist. 

Niki: And so, I think the reason maybe people aren't is partly because of a communications issue. [Dex: Sure] So, the negative potential externalities, job loss, the destruction of humankind, the destruction of democracy, or whatever, right?

These are the things getting play in the media and they capture people's imagination, fear cells. And perhaps there's a communications issue around what we should be optimistic about. So maybe we start with job displacement. I've heard you talk about how this is just standard operating procedure as we get new technology.

Do you have thoughts on that? 

Dex: Yeah, I mean, I should say all of these new opportunities being created by the technology will obviously also come with massive challenges. And these are things that have to be taken seriously. We have to manage them. Previous generations of technology, of course, have led to changes in, in jobs, transformation of industries.

And AI will also, no doubt, have that kind of impact in a whole bunch of different ways. But, also, as those new technologies come online, we also expect that we'll see entirely new industries, which we can't even predict today. Think about the number of jobs you've done in your career, Niki, or the number of jobs I've done which - there's no way our parents would have been able to predict that these would be things we would do. 

Niki: I mean, my parents definitely didn't think I could be a bullshitter for a living. 

[both laugh]

Dex: And definitely, like, if you'd been a kid and said,”I'm gonna go work on the future of information infrastructure or how to connect the world.” People would have, they would have looked at you with their eyes glazed over. And when we look at the future of AI, it's going to be a lot like that.

It's going to be talking about a future that doesn't exist yet, which is a very, very hard thing to do. So, I think you're right. There's a communications challenge there but absolutely when we, when we look at some of the implications for sustainability or energy or medicine, the impact on resources, like, actually there's a bunch of things which should give us a lot of hope.

Niki: I think it's interesting that you talk about completely new jobs that, for example, to your point, when I was a kid in the 70’s we didn't have anything like what we have now.

They couldn't have possibly imagined that the functions we've worked in and around would exist. And I think one of the issues with AI is we talk about computers or robots replacing humans in their current professions, but those professions will be either gone completely and replaced by totally new things we're not even imagining, or supplemented.

And I think this is actually a really good point because people say there'll be more efficiency and job creation, but it's really abstract. And I think part of the issue is we don't know what they are, so we can't possibly describe what we don't know, but inevitably there will be more jobs.

Dex: Sure. And the sort of nature of human ingenuity, right, is you do more with more. As we have access to those new tools, actually, I think humans have a pretty good track record of actually saying, “Oh, there's even more applications and problems that we want to go out and solve as a civilization.”

So, I think, it sometimes does get connected, I think, with how you view the world and whether you do have that sort of more optimistic view of human nature. I think the other thing, right, going back to whether these are new jobs or they're transformed jobs, certainly there'll be lots of roles which exist, which already exist, but they will evolve because of the technology.

People will find new ways to do things that have existed previously,and things like generative AI will add interesting new tools and things that create a lot of value, but the need for those roles will be there for a long, long time. 

Niki: So, one of the things we were talking about before we started taping is how there's a little too much conversation potentially around generative AI and ChatGPT and these tools that we're using.

And you mentioned that you'd love to talk about the science of AI, like the real, the positive, use cases that are really concrete and can help people live better lives based on science. 

Dex: Yeah, absolutely. I mean, don't get me wrong, generative AI and all the things we're doing, at this moment in time. Incredibly exciting. I mean, it's a whole new moment.  

Y’know, a whole new wave of productivity - people using these tools to enhance their creativity. I think actually there's lots of exciting applications which we haven't even even seen yet there, but when it comes to science as well, we're talking about using AI and technology to solve these fundamental problems. In some cases things that are sort of age-old problems for humans.

And how do we live healthier lives? How do we protect our ecosystems? At this moment where the climate is changing very, very rapidly. How do we deliver the future of energy production for the world? Those are all areas where we're seeing, y’know, extraordinary breakthroughs right now.

So at DeepMind, for example, one area that I'm incredibly excited about is materials discovery. We just published research at the end of 2023, where a team had discovered more than 2 million new crystals. And actually being able to identify millions of new types of materials actually might help us then develop new breakthrough technologies in areas like battery technology, or much more efficient solar technology.

So, there you start to see knock-on impacts on other realms of policy, other technologies that might fundamentally transform our quality of life. 

Niki: I think that's really interesting. So first of all, when you said the thing, this is, the cynic in me just always comes out. But when you said that, it reminded me when we met, and you were again working for Eric Schmidt, who was the president and CEO of Google at the time, and I was his PR person.

So your job was to be innovative and creative and come up with all the fun stuff [laughs] That was what you were doing. 

Dex: That was your job too! 

Niki: No, my job was to say no to things that were too innovative! I wonder if you remember this, but Eric wanted to talk about how one day we would all take our password like a vitamin every morning [Dex: laughs] and it would change and I was like, “I never want to hear it again.”

[both laugh]

“Nobody wants that.” 

But when you were talking about discovering minerals, it makes me think of how you've got these billionaires doing asteroid mining. It just feels so abstract and kooky. And I think when you talk about. It's happening now. You've found new crystals. There are new materials. 

We are really dependent on problematic nation-states for lithium. And we're held somewhat hostage by geopolitics and industrial policy when it comes to chips. And so, the idea that we could find not just domestic mining, but new things altogether, [Dex: mm-hmm]that is a really concrete example that, to be honest, I'd not heard about until just this minute.

Dex: I mean, AlphaFold, another perfect example, right? One of the signature DeepMind achievements, now something that's very big at Google DeepMind, after decades of study, scientists had predicted the structures of fewer than 200,000 proteins, which are, of course, the microscopic building blocks of life, key to understanding a whole bunch of different realms of science where we want to make progress for, for the human race.

And, DeepMind was able to go and build a technology, AlphaFold, which was able to successfully predict the structure of more than 200 million proteins and make that data freely available to the world. And you've got scientists and labs all over the world using that data now, using that knowledge to go and advance things like ways to overcome antibiotic resistance, to figure out how to create enzymes that can eat plastic pollution. 

Niki: I love that one, by the way! 

Dex: Yeah. So, like, you, you see, like, this is just like one example, right? Of something that's happening now.

And we're starting to see progress being driven around the world. And actually going back to the time when we worked together, right? I remember vividly, like having this conversation with colleagues on our team about “What is innovation?’ It's not just what's new, and often people use it sort of interchangeably with just what is new, what's novel.

And innovation, of course, is about how do you have inventions and discoveries where, after you've achieved them, you get new waves of invention and discovery around those things. And, that way, the thing that you've discovered ends up spreading through society. It goes rippling out and creating all this new value.

We're in this moment right now where the ripples are spreading incredibly rapidly. I mean, actually I think it's sometimes challenging working in the AI space just to keep up with the pace at which we're seeing all these extraordinary breakthroughs. 

Niki: I love that you said that, too. So first of all, this has just been an injection of optimism because I sit here being really cynical and I, I said recently, “It's just iterative.” But what you've just said is, “It's not iterative. It's actually exponentially innovative.”

Dex: That's what I think. Yeah. I mean, of course. Lots of parts of this sort of grand puzzle of discovery, they do involve a lot of iteration. Great discoveries take time. And you look at the impact of individual scientists and researchers and institutions. When history books are written about these things one day, often you start to see that it's very hard to attribute credit just to, like, one person or one set of people. 

Actually these moments that seem to come out of almost nowhere, actually, they don't come out of nowhere. They come through the decades of work that precedes that, sometimes centuries of work, but, I think now we are seeing individual researchers and scientists and different teams really getting access to tools that are so extraordinarily powerful that you can both have that sort of massive, almost exponential impact, but at the same time you are contributing to something much, much bigger.

And the, these two waves of innovation are unfolding at the same time. 

Niki: And when you think about the challenges we have from not having enough food around the world, climate change. Not just climate change, but adapting to climate change, having models so you can figure out how people can move out of the most at-risk areas.

I say this as someone whose mother lives in South Florida, [Dex: Right] which is like an area where people are not moving out, but they probably should [chuckling] if they paid more attention. But yes, having climate models that are more accurate, storm models. Being able to have more accurate scans, body scans, [Dex: right] so it's not a human looking at two different scans, it's actually a, a perfect comparison.

That to me seems very close, probably, and maybe it's iterative, but then it will become so much bigger than just that. 

Dex: That's right. We had a moment last year, actually, where in the same week we were announcing we had Lyria, which was the world's most sophisticated music generation model at the same time, we were also announcing we had a technology called Graphcast, which is the world's most accurate 10 day weather forecasting system.

So, actually, like, you see how these moments are sort of all coming together. The weather forecasting stuff is actually really cool. Particularly as a Brit, where we all obsess over the weather. It's like our national pastime. [laughs]

Niki: I mean, it It's funny because I've been saying it for a while. I no longer believe in weather forecasting because it just seems like it's not that accurate. So, when I was a kid, you could call on the telephone, you could call the weather and it was always right, like three or four days out. And I'm not saying that weather forecasting has gotten worse. Weather's become more unpredictable. 

So, anyway, I'm in favor of this. [Dex: yeah!] 

And I think that's a real life application that can help people in their day-to-day life. And maybe if we start to tie AI to really practical things like this that improve quality of life, rather than saying abstract things like efficiency will improve, which nobody knows what that means, that might be helpful at elevating a more positive discourse around this, because I do fear that we're going to overreact to the unknown hypothetical externalities that are negative.

Dex: Yeah, absolutely. I mean this is something obviously, Niki, you have a unique vantage point on as well, having spent your career in technology. There's all sorts of progress that I think folks working in the industry can feel really excited about. We feel really inspired. We have that foresight of where it's going to be used and the value it will create. 

And at the same time, these things are disruptive. And Silicon Valley, often in “ye olden days”, disruption was just this thing you celebrated. And I think with time and maturity, people understood disruption is actually disruptive.

And it's a thing that's profoundly unsettling for a lot of people, and that is a communications challenge. How do we manage this and explain what the effects really are like. 

And that's not just because we want to spin any of this stuff. It's an essential part of preparing societies to be able to manage these issues. And actually to be able to come up with solutions in time to, to mitigate some of the challenges. 

Niki: And this  is where the earnest part of me comes out. Because having worked at Google, I worked there for eight years, and we literally sat next to each other [Dex: yeah] when we were both in Silicon Valley.

I believe that Google genuinely tries to do the right thing when it comes to technology, and I don't think it's just spin to try to get ahead of some of these disruptions and learning the lessons that move fast and break things, [which was not a Google saying], it's not helpful for society and that people have lost trust and so to rebuild trust in these tools.

I think we absolutely have to, and I do genuinely believe most of the big tech companies, Microsoft as well, like, are good actors, and are trying to do the right thing.

Dex: Yeah, absolutely. Ultimately, why do people go and work in tech and choose to work on these really, really hard problems?

I think for a lot of folks it's because they want to create something of value which really makes society better for the next generation. There's lots of other places all of us could go and do things, but this stuff is really, really meaningful. And, for me, it's personally really meaningful.

I started my career at the UN, as you mentioned, and I, later, ended up in places like the Meta Oversight Board, and these are all places which were dealing with very thorny societal and ethical issues often. And, of course, AI will present all sorts of challenges when it comes to how do we develop the technology in a societally responsible way, and in a way that's safe for people.

And I definitely would never want to work for an institution that didn't take those issues really, really seriously. And I think at Google folks are really, really motivated to do the right thing. Here at Google DeepMind, of course, safety and responsibility has always been this very, very big part of the company and how the organization has evolved.

From developing a set of AI principles very early on that went on to inspire how Google was developing their AI principles all the way through to the work that's going on today,  y’know looking to respond to the challenges of generative AI looking ahead to the challenges of sort of AGI technical risks, it's something that is a really serious, big part of the work that we're doing here.

We have this sort of mantra at Google about bold and responsible when it comes to that  AI future and you can't have one without the other, really.

Niki: That's right. And that is an evolution in the time that you and I've been working in tech. 

The other thing I think people forget, and I put myself in this category, but how magical things are that now we take for granted, things that are absolutely magical, that help our safety, that help convenience, that make us more connected.

It's easy to get down about social media. You mentioned being on the Meta Oversight Board. So, just for anyone who doesn't know what that is, it was essentially the external body that would make hard calls on content or high-profile users. [chuckling]

Dex: That's right. Yeah. It was, it was all the sort of most challenging, sometimes controversial issues around freedom of expression and human rights. And in some, some cases, really, really significant with the implications for how different communities would be engaging with social media and using that to amplify their voices on really, really hard topics. So, yeah, I mean, that was, that was the very, very, sharp edge of the spear [Niki: Right] of content moderation.

Niki: Well, not to put too fine a point on it, but whether or not President Trump would be allowed to stay on Facebook. [Dex: mm-hmm] And then, making that decision based, or sort of overseeing what Facebook slash Meta had made a decision on. So, I think actually this plays into everything else you've talked about today, which is: We are practicing as an industry and learning to be more deliberate, right?

At first, you just put stuff out and see what happens. Then you realize the tough calls, and the ethical calls, and then you realize, ethics is different based on where you live and what group you are. And so then, you have to really put in place processes and principles, and that's what, hopefully, it seems like people are doing ahead of time with AI, or at least at the same time that we're developing the tech.

Dex: Yeah, very much so. I think the, the last decade, decade and a half of societies around the world responding to the information revolution and social media and smartphones and so on. It was obviously really, really hard. I mean, we dealt in our careers with a lot of the challenges that were created by that moment.

And now with AI , and the moment that we're going through with AI, I think a lot of leaders and institutions around the world have obviously woken up very, very quickly to the potential massive applications for that, both for good and ones that will create more challenges for society, and they want to be ahead of the curve on this. And that's why you've just seen the huge volume of conversations that are unfolding now about this technology, the efforts to try and build consensus across different industries, across different communities, and we had moments last year like the UK AI Safety Summit where Google DeepMind was playing a big part. And this year I think we'll see a lot more energy as well. 

Niki: Yes, and you're in Washington partly to meet with reporters and columnists and folks writing about this and to tell some of the positive stories about what DeepMind's doing.

Dex: Yeah, absolutely. Look, as somebody who used to work in the aerospace industry, my colleagues I'm sure are thoroughly sick of me talking about what did it look like to be communicating, for example, the story of the space race when that moment was kicking off many, many decades ago and how we tell the story is really, really important for shaping how do people ultimately perceive the value of the things that we're doing, for supporting these big societal moments and transitions. 

So, I think it's something I'm really, really passionate about: helping folks to figure out just what the heck is going on with this moment.

What are the things that we should be focusing on? What are the good use cases of AI? What are the things that we should be very, very attuned to in terms of the issues that we have to mitigate and manage against? 

Niki: I will say there's sort of the lobbying class and the consultant class in D.C. which I'm in, and then there's the reporters who are really even more, sometimes taking even more pessimistic views, so it's good to be in town and telling those stories.

 Also, you are a very seasoned speaker. This is how we got reconnected. So, we sat next to each other a hundred million years ago, [Dex: laughs] and then we've seen each other on the internet and on LinkedIn ever since.

I was talking to someone about great speakers, people who talk to conferences, conventions, on panels specifically about science and technology and sort of futuristic visions. And he said, “ You know who you should really look up? You should really look up Dex.” And I thought, “I know Dex! I'm going to reach out to him.”  

And so, as you're speaking around the world this year, and you're talking about DeepMind and AI, what are the things you're most focused on?

What do audiences really want to hear more about? 

Dex: I think most people are still just trying to figure out “What the heck does this mean for like my work and my life?” You can get into those bigger conversations about the sort of macroeconomic impacts and geopolitics and all these things, they're all really important topics.

But, lots of people are just trying to understand, like, “When am I going to be able to use this stuff? What is it going to do in my role? And what do I need to do now to prepare myself for that future and to take advantage of the moment?” And I think AI Is something that lots of people are just really excited about, but they really don't actually know where are the entry points to this.

When smartphones came along, we all had mobile phones already. It wasn't that massive a leap to also figure out, “Okay, I'm going to be rocking this handset, which then also gives me access to the internet, which I'm familiar with.” And of course, there were lots of innovations that came along, and, you have the app economy that arose. But it was something which I think built on actually decades of, like, understanding. 

AI is something which offers us tools that are just so dramatically different, and in some case the applications are just so much more powerful. And so, I think people are still trying to figure that out. I spend a lot of time unpacking that, the scientific applications. I think that that is something people are hungry to understand, as well, because we're all hungry for hope. 

Niki: That is so true. We are hungry for hope. 

Dex: Yeah, we we want to live in a world right where one day, some of the diseases that, right now, they are the things that cause immense suffering in the lives of the people we care about, these are things that are defeated and that technology can help us have a better quality of life.

We want to live in a world where there's no more climate change. And we're assured that our children, our grandchildren will have something good to inherit in this world. And these are the things I try to get into, right? 

And then I think everyone also naturally has questions about the safety and responsibility element. After the last couple of decades in technology and all the challenges and all the controversies, people do want to understand, “Is this going to be something that I should feel comfortable with?” I think sometimes folks in the AI field often forget that we've had many decades of cultural and science fiction tropes about just the evils of AI.

Everyone jumps straight to the Terminator. [Niki: Right] And, of course it wouldn't be particularly sexy for making a movie or a Hollywood production where AI is just really, really good and it does things that are, like, quietly valuable. 

Niki: Right, that created more efficiency and productivity.[laughs]

Dex: Exactly. So, I think, actually being able to get into some of that and to figure out what is real here. I think that's what I spend most of my time on. 

Niki: I think it's super helpful. This is why I needed you to come in because we have so many policy people and so many DC types and having someone come also just from London and who's in a slightly different angle on things to come in and talk about the optimism of it.

I am actually genuinely quite optimistic about AI. 

Partly because unlike social media, which is fraught because it's connecting humans, this is connecting data in ways that are going to, it's going to help humans. It's not just humans interacting with each other, which I think there's good things and bad things about this, but with data in particular, there are so many, like you said, with the space race, it's so exponentially innovative and we'll have so many other applications that it is really exciting.

And I needed somebody to come in and sell hope to me. So, thank you.

Dex: Excellent!  Well, I'm always down for that. 

Niki: Thank you, Dex.