← Intelligence Squared
Intelligence Squared
Demis Hassabis and Sebastian Mallaby on The Quest for Artificial General Intelligence (Part Two)
00:33:03
Loading summary
Transcript62 lines
- [00:00]
Podcast Host or Advertiser
Springtime is my catalyst to switch out the major players in my closet and take stock of what I have and haven't been wearing over the last year. It's a great time to get a bit more intentional about what you're wearing day to day and if I'm getting rid of anything, I want to make sure that I'm replacing it with quality pieces and I've been turning to Quince for that so often recently. Their clothes are made really well and price even better, so it makes shopping for and wearing their pieces simple. Quince uses premium materials like organic cotton and ultra soft denim and their lightweight linen pants, dresses and tops start at just $30. They've also got incredible accessories. I just picked up a cognac Italian leather sling bag which is a huge upgrade from the crossbody that I have been using. The leather itself is really beautiful and I also love the gold hardware it comes with. I think it's just such a sleek bag. Everything at Quint's is priced 50 to 80% less than similar brands because they work directly with ethical factories and out the middlemen. So you're paying for the quality and craftsmanship of the products, but not a brand markup. Refresh your everyday with luxury you'll actually use. Head to quints.com intelligence for free shipping on your order and 365 day returns. Now available in Canada too. That's Q-U-I-N-C-E.com intelligence for free shipping and 365 day returns. Quince.com intelligence
- [01:26]
Progressive Insurance Advertiser
insurance isn't one size fits all, and shopping for it shouldn't feel like squeezing into something that just doesn't fit. That's why drivers have enjoyed Progressive's Name
- [01:35]
Shopify Advertiser
your price tool for years.
- [01:36]
Progressive Insurance Advertiser
With the name your price tool, you tell them what you want to pay and they show you options that fit your budget enough. Hunting for discounts, trying to calculate rates and tinkering with coverages. Maybe you're picking out your very first policy, or maybe you're just looking for something that works better for you and your family. Either way, they make it simple to see your options. No guesswork, no surprises. Ready to see how easy and fun shopping for car insurance can be? Visit progressive.com and give the name your price tool a try. Take the stress out of shopping and find coverage that fits your life on your terms. Progressive Casualty Insurance Company and affiliates Price and coverage match limited by state law
- [02:19]
Mia Sorrenti
welcome to Intelligence Squared, where great minds meet. I'm producer Mia Sorrenti. In this episode, we return for part two of our recent live event with CEO and co founder of Google DeepMind and Nobel laureate Demis Hassabis and journalist and author of the Infinity Machine, Sebastian Malaby. Hassabis and Mallaby joined us recently at Friends House in London to discuss the rise of artificial intelligence and the pursuit of AGI. They were in conversation with co anchor of opening trade on Bloomberg TV, Tom MacKenzie. If you haven't heard part one, we recommend jumping back an episode to catch up. But now let's return to the conversation live at Friends house in London.
- [02:56]
Moderator or Interviewer
We're living in a moment where we're thinking about AI in warfare, whether it's in Ukraine or in Iran. We're thinking about maybe some of us anthropic and mythos and cybersecurity. The clear positive is that is that guiding light that you've had throughout the scientific discovery and specifically AlphaFold2, AlphaFold3, isomorphic labs, is that where when we think about the risks, benefits, balance for AI and AGI, is it very clear for you that it's medical advances and biology that some of the clearest wins that we're going to see?
- [03:29]
Demis Hassabis
So I spent my whole life working on AGI because I believe it's the ultimate tool for science and broadly construed so science and medicine, I think the number one thing we can apply AI to is curing terrible diseases. So number two on my list would be helping with the environment and finding new renewable energy sources. You can see that's what I care about because that's what I personally done with my time other than building the core technology, things like AlphaFold, isomorphic labs and all of our. We've had a Science Group, an AI for science group for nearly 10 years now. Some of our rivals are starting to set up science teams to use the AI, but we've had one for 10 years including that's where AlphaFold came from. And so we work on fusion, we work on weather prediction, we work on climate models, material science. So there's almost no area of science that I think AI can't help with. And I think if we do that right, in the next 10 years we'll have almost a second renaissance, a new golden era of scientific discovery. And ultimately the reason I started on this journey is I wanted to understand the nature of reality, all the big questions in the world, nature of consciousness, what is time, all of these things which are incredible deep mysteries and it's amazing, I find it amazing that we all just go about our daily lives without thinking too much about this sort of Distracted with what we do. But for me, it just stops me dead in my tracks all the time. And this is what I think about late in the night, in the small hours of the night is how is it we don't even know what time is? This is crazy. And we're just going around doing things. It's crazy to me. And then of course, well, what do we do about that? Is the question. Because I guess I was sort of born with this insatiable curiosity and also almost haunting thing about like this is a deep mystery here in the universe. It's like an incredible puzzle. That's how I feel about the universe. How are we going to solve it? And then I guess that's why I got to AI early is like, I tell you what we need, we need some help. Even the best scientists. I used to read all the biographies of the top scientists, all my friends favorite heroes from Feynman to Turing, and they're all really smart. But there's a limit to how much we can comprehend, I think with just unaided, with our human minds, even with scientific method and other things. And so it was obvious to me that AI, it was a two factor thing. Not only would it be an amazing tool for science, but it would be incredible scientific artifact in of itself as an amazing engineering artifact and possibly help us to understand our own minds better, which is also one of the deep mysteries. So that's why I would have done this no matter what. Today it's become the biggest thing it is, but for me it's the most fascinating question and the most fascinating technology there is out there. But it's dual purpose. And we also knew that at the beginning of AGI, which is why we were worried about safety from the very beginning of DeepMind in 2010, and we still are now because it's general, so it can be used for anything. And I think it's down to us in the industry to, to try and push it and advance more quickly the areas that are gonna be great for society like medicine and energy.
- [06:42]
Moderator or Interviewer
Where are we right now on the balance of risks, do you think?
- [06:46]
Demis Hassabis
Well, look, I think right now one of the issues with this ferocious commercial battle. So look, the competitive part of me loves that. Of course it's like great, it's the kind of top competition there is in the world. But in the back, the problem, I can't. So part of me I like that I lean into that. That's what helps me lead in this situation. But at the back of my mind I've got this gnawing feeling that there's something much more important, much bigger than the commercial race, which is overall getting AGI safely over the line for humanity and to make sure that the benefits fully outweigh the risks. And I'm going to try. We're only one actor in this now. There's five or six leaders, and there's China as well, and the Chinese labs. And so I think in the next few years, the story's still to be written like how this is going to go. But I do think there needs to be more cooperation and coordination at an international level, ideally, although that's very hard with the geopolitics as it is today, around safety topics and debates around the benefits versus the risks.
- [07:54]
Moderator or Interviewer
And Sebastian, having spent time with Demis in the research for this book and the process of putting it together, do you come away feeling more reassured or less reassured? When we talk about governance, we've got companies like Anthropic who are deciding themselves not to put their latest model out broadly into the market because their concerns about cybersecurity. That's a decision they're making. It's not a regulatory decision, it's not a government decision. Do you come away from your conversations more reassured or less reassured?
- [08:21]
Sebastian Mallaby
Well, I think the good thing is that as Demis said, he was thinking about AI safety very seriously from the time he founded DeepMind. So it goes back to the very origins of his quest. And that's reassuring. And it's reassuring that he's tried various things within DeepMind and Google to try to have safety principles or guardrails or
- [08:43]
Sebastian Mallaby (continued)
oversight committees and so forth.
- [08:45]
Sebastian Mallaby
But I think the big conclusion that Demis reached and that I talk about quite a lot in my book, is
- [08:51]
Sebastian Mallaby (continued)
that in the end, if there are many actors, as Demis was saying, five
- [08:56]
Sebastian Mallaby
or six, plus China, you have a race dynamic. And this becomes bigger than an individual AI leader can control. And to some extent, I wrote an
- [09:10]
Sebastian Mallaby (continued)
earlier book which was a biography of Alan Greenspan. It was called the man who Knew. And one of the journalists who interviewed me about the book said, well, you could have used the same title for Demis Asabis, because this is the man who Knew. He knew from the beginning that this could be dangerous.
- [09:27]
Sebastian Mallaby
But it doesn't necessarily mean just because you know it that you can control it. And so what I'm getting at is that when there's a race and there's a collective action problem, you need the government to step in and coordinate a joint set of safety principles, a joint sense of only rolling out Models which are thoroughly tested and they're going to be safe, and no one lab can do that. It's up to the government. And furthermore, actually, it's up to some sort of discussion between different governments, because just one wouldn't be enough. Let's say America decides to be safe, which under the current political circumstances might be unlikely. But anyway, put that aside. Let's say America tries to be safe and then China isn't safe, for the sake of argument, then the world isn't safe. That didn't solve anything. Right. So we need governments to come together and agree on some safety guidelines. And I think what's going on with Mythos, which is this anthropic model that apparently can find vulnerabilities to cyber attacks which haven't been discovered despite thousands, if not millions of human tests and experiments with these operating systems or search engines or whatever, you put Mythos on them. And bugs, flaws, vulnerabilities are identified that were invisible to human discovery. Now, that's not totally surprising in the sense that with AlphaGo, there were strategies in GO which hadn't been discovered by humans ever, despite the fact that humans have been playing GO in evolving strategies for centuries and centuries. So once again, you have an infinity machine that comes along and unscrambles the problem and does find bugs. But the proposed solution from Anthropic, which is let's roll it out to just six or seven companies at first and let them harden their software systems against cyber attack, that's not sustainable. I don't see how one individual company can determine the nature of the oligopoly, which will be privileged to get the early release of the model, and that the world is just cool with that indefinitely. So we need governments to come in.
- [11:52]
Moderator or Interviewer
And what comes out very clearly from the book, actually is Demis and the team's insistence from a very early stage on those safety protocols and how, frankly, that moment's lit attention with the parent company. I want to turn it over to questions now from the audience. So Louise has sent in a question, and this one's for Demis. Out of the universe of possible scientific problems, how do you choose which to tackle and I guess in which order?
- [12:16]
Demis Hassabis
Well, part of it is what I just described about the types of problems that are suitable for these types of methods, so that we try to make sure that whichever problem we pick, it fits that rubric. But that's still quite a lot of choices. So then the next level about that is to pick problems that have the most impact if they were to be Cracked, because every single problem, even if you have a general system, it takes a lot of special casing and domain knowledge and interdisciplinary research to actually crack the problem. So like AlphaFold, you know, took four or five years of working specifically on that hybrid system and it needed a lot of specialists as well as general machine learning techniques. And so we have this notion of this idea of a root node problem. So if you think of the tree of all knowledge or all possible knowledge, and it's enormous, and then there are certain problems where if you unlock them, a whole new branch of, it's like a root node, whole new branch of avenue of discovery or research or applications is unlocked. And those are the things that I want to spend time with, at least a deep mind doing. Other people can also take our work and go further down the leafs of these branches and solve the really important practical problems. But I think our job as being very close to the core technology is to go after really meaningful root node problems that unlock things for maybe millions of researchers downstream. And I think AlphaFold is a perfect example of that, where protein folding, it was a beautiful grand challenge in of itself. These exquisite proteins that everything in life depends on, they're like little nano machines and what's the shape of them? But then if you crack that, not only would it be an amazing scientific discovery, it would unlock drug discovery and biological understanding to a new level. And indeed that's what's happened when we put Alphafold out into the wild to the community and open sourced it and all the 200 million protein structures which we folded in one year because AlphaFold was fast enough to do that. 3 million researchers around the world are now using Alphafold. And someone was telling me, a friend of mine in the pharma industry, that any, almost any drug discovered from now on will have probably made use of Alphafold somewhere along the lines. And that's the kind of impact I always dreamed of having when I set out on this at the beginning. So I want to. There aren't, you know, picking the right problem is an art in itself. And it's a sort of a taste question we sometimes call it in science, which is basically your intuition telling you this is, this is the right time to tackle it, there's enough data and this is how we couch the problem, and here would be the downstream benefits from it. And then that justifies putting in potentially a big effort into trying to solve that ourselves, because obviously there's so much choice we could have, which is an amazing situation to be in for somebody who loves many different subjects. It's like the perfect tool because it allows us to go into lots and lots of domains.
- [15:20]
Moderator or Interviewer
So another question, potentially on, potentially an accelerant to that journey for you, Dennis. How will quantum computing affect the journey to AGI?
- [15:31]
Demis Hassabis
Yes, well, I mean, look, quantum computing is also a really hot area and we're lucky that actually within Google and Alphabet there is a quantum computing team that's one of the best, if not the best in the world. And I know the leader of that very well, Hartmut Nevin. And we have a kind of long standing bet for the last 10 years of like, which one's going to be first and because they'll obviously be able to help each other. So if quantum computing comes first, you could potentially, potentially train very quickly an AI system, maybe very powerful one on a quantum system. Also, if AGI came first, one of the first things we'd probably get it to do was finish off the designs of the quantum computers. So I suspect AGI is going to come first, but they're both making good progress.
- [16:16]
Moderator or Interviewer
One of the things that's come up in this conversation is your love of gaming, and that's in the book as well. Whether it's chess, but also table football, apparently you were pretty, pretty vicious in terms of making sure you beat everyone in the, in the DeepMind office. And you had some, you had some accolades for that. So there's a question from Ryan. How does your background, and of course you famously, of course set up your own gaming company and, and designed plenty of games as well. How does your background in games influence your, your approach to building AI for Google?
- [16:46]
Demis Hassabis
I think it's affected my entire journey to here. So when I think through it, and this is one of the things that, you know, talking to Sebastian made me reflect on was, you know, I think I've used it in at least three ways, games, and I think there'll be a fourth way. So one was probably in the early days training my own mind with games. So through chess and the other games that I played, I think no doubt it helped form the way, I think. So that's step one. Then it got me into my love of games and computers came together with game design, video game design, and also AI, because I was basically funding my early AI research through the guise of making AI for games. All the games I wrote, video games I wrote professionally have AI as a key component of the gameplay. And it was another thing that convinced me we were onto something and how magical it would be to pursue AI to the fullest. So it Also introduced me to being an entrepreneur and also big engineering projects. And then finally with DeepMind, the beginning of DeepMind. We use games, as I mentioned earlier, as the perfect testing ground in the early days when no one knew if deep learning was going to work or reinforcement learning, we all know that now, but back in 2010, no one had any idea if that could scale or be. In fact, no one had combined them together either. We did all of that in the early days of DeepMind and proved it out on computer games like Atari, then Go, which was the pinnacle of games. AI for the last 50 years, thought to be uncrackable. And then real time strategy games like AlphaStar. So maybe the final chapter for me in games will be when I, you know, when we deliver AGI or safely over the line is maybe get AI to use AI to help design new, incredible new types of games that, you know, we can barely imagine today. I think there's a lot of cool things to be done there.
- [18:35]
Sebastian Mallaby
Could I test an idea on Demis, which is that I think another one you missed out is that the strike team structure that you use at DeepMind, where there's a sort of very clever combination of two approaches to blue sky science. When you're founding a company that doesn't try to take existing science and turn it into a product, instead you actually have to invent the science first. You need to give the scientists a lot of freedom and they have to do blue sky thinking. You have to go off and write papers and they do their thing. But then you have to marry that bottoms up approach with strike teams where you see, or actually Demis sees, that some area of the fundamental science is maturing to the point where if you pushed it hard enough for a couple of years, you would get an amazing result like AlphaGo or AlphaFold. And that's a much more top down structure. And that comes, I think it's fair to say, from games.
- [19:31]
Demis Hassabis
That's a good point. And actually there's a lot more analogies between game design and game production and doing blue sky science than you might imagine. So especially in engineering science like AI in games, we sort of perfected this idea of marrying cutting edge engineering, 3D graphics, AI itself, but even like the hardware like GPUs were invented for gaming, don't forget. Right. And then of course become synonymous with AI now and, but marrying that with a lot of creativity and design and amorphous things like is this going to be fun? So you're trying to do this cutting edge engineering, but you're trying to accommodate this very creative process and combine them. And that takes very special skills and very special people who are able to cross those boundaries, some of whom I see in the room here, who I've worked with many years on these projects. And it's if you skew your lens a little bit, I felt that could work for science, even though I think no one had brought that over to scientific endeavors. And in a company where instead of the game designers and the game artists and the creative side of that, you replace that with the research scientists who of course it's different. It's a completely different skill set. They're technical, they're maybe not artistic. But what separates a good scientist from a great scientist is not their technical skill, it's their creative capability, in my opinion, and their ability. This thing about taste or to sniff out what the right approach is or whether to push harder on something versus something else. And it struck me doing my PhD, this was very similar. The best scientists I learned with and was working with reminded me a lot of the best game designers. And then the engineering is pretty similar. It's cutting edge engineering in both sides. So actually the whole approach of project managing AAA games I more or less lifted and brought across and translated to the early days of DeepMind.
- [21:32]
Moderator or Interviewer
With that success again of AlphaFold2 and that idea of taste, I mean maybe that's something that AGI cracks for us in terms of what actually underpins that.
- [21:46]
Shopify Advertiser
Starting or growing your own business can be intimidating and lonely at times. Your to do list may feel endless with new tasks and lists can easily begin to overrun your so finding the right tool that not only helps you out but simplifies everything as a built in business partner can be a game changer for millions of businesses. That tool is Shopify. Shopify is the commerce platform behind millions of businesses around the world and 10% of all e commerce in the US from household names like Gymshark, Rare Beauty and Heinz to brands just getting started. Shopify has hundreds of ready to use
- [22:23]
Progressive Insurance Advertiser
templates that can help you build a
- [22:25]
Shopify Advertiser
beautiful online store that matches your brand style. And you can tackle all the important tasks in one place. From inventory to payments to analytics and more. No need to save multiple websites or try to figure out what platform is hosting the tool that you need. And if people haven't heard about your brand, you can get the word out like you have a marketing team behind you with easy to run email and social media campaigns to reach customers wherever they're scrolling or strolling. Start your business today with the industry's best business partner, Shopify and start hearing. Sign up for your $1 per month trial today at shopify.com realm go to shopify.com realm that's shopify.com realm
- [23:12]
Home Depot Advertiser
it's time to bring on the blooms at the Home Depot with Spring Garden Deals. Find savings on hanging baskets and flowers to brighten your backyard or any space that needs instant color. Then get everything you need to plant and protect them, with low prices guaranteed on soil and mulch. Dig into Spring garden deals for four days at the Home Depot, now through May 10. Exquisite supply. See homedepot.com pricematch for details.
- [23:41]
Microsoft Advertiser
Study and play come together on a Windows 11 PC and for a limited time, college students get the best of both worlds. Get the Unreal College Deal everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 Premium and a year of Xbox Game Pass ultimate with a custom color Xbox wireless controller. Learn more@windows.com studentoffer while supplies last ends June 30 turns at aka mscollegepc
- [24:12]
Red Bull Advertiser
ready to soundtrack your summer with Red Bull Summer All Day Play? You choose a playlist that fits your summer vibe the best. Are you a festival fanatic, a deep end dj, a road dog, or a trail mixer? Just add a song to your chosen playlist and put your summer on track. Red Bull Summer All Day Play. Red Bull gives you wings. Visit RedBull.com BrightSummerAhead to learn more. See you this summer
- [24:39]
Indeed Advertiser
when you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills certifications. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a $75 sponsored job credit@ Indeed.com podcast. That's Indeed.com podcast terms and conditions apply. Need a hiring hero? This is a job for Indeed. Sponsored Jobs.
- [25:13]
Moderator or Interviewer
A question from Asha. What is the hardest scientific problem standing between us and AGI right now?
- [25:21]
Demis Hassabis
Well, there's two schools of thought on that and we're pushing both as hard as possible. So one is that school of thought is that we can just scale up our existing methods and maybe make some incremental advances to what we know today. Maybe generate some synthetic data to increase the amount of data they can learn from and this will sort of rise all boats, right? And that's been the history of the last two, three years. I would say that Scaling laws they're called, they work. And we're still getting amazing returns from building 10 times bigger systems every year, pretty much. And that's what's obviously fueling all the compute build out and the infrastructure and all the things that you've all read about. Now the question is, and I think it's 50 50, is there also one or two things missing still? Big breakthroughs that are still needed to get us all the way to AGI. And we can see today, even with the best leading models, whether it's Gemini or Claude or GPT, that they're incredible in many dimensions, but they also have some basic flaws. Still, in my opinion, an AGI system shouldn't be this kind of jagged intelligence. It should be consistently good across the board. I think that's one of the hallmarks of generality. So I think that's missing. Also they don't do continual learning. We train them in the labs and then we put them out in the world. And yes, they can memorize things and have context windows, but they don't really learn in the way we learn out in the world. And that's critical. I think for these systems to be able to take on tasks and really useful tasks in the world, they've got to be able to adapt to what they find in the specific context that they're in. And there's other things like long term planning isn't solved. So there's a bunch of things where you could imagine either new capabilities are needed, like the level of AlphaGo or Transformers, these big leaps of architecture, or maybe it'll be like language where in the end we had the architecture which was transformers in the end, but we needed to do it at massive scale. And so I just approach it with the scientific method, which is this is an empirical question. We can debate it all day, but nobody knows. So this is one advantage of being relatively a large research organization at Google DeepMind is we can bet on both sort of equally hard. So there's a scaling is roughly half the people at Google DeepMind work on the scaling side of things in Gemini and then the other half work on Frontier Blue sky sort of new research ideas. So that whatever the answer is to that, I want us to be first to it. And I think it's 5050 in the sense that I don't think anyone knows what the answer is. But one thing we do know is the current systems we're building will be a key component of the final AGI system. Of that I'm like 99% sure it's the only question is, will they be the only component or would they be merely a key component?
- [28:17]
Moderator or Interviewer
They don't become redundant. They are essential to the.
- [28:20]
Demis Hassabis
Yes. I can't imagine a way, a world in which they would be like, oh, this was a completely wrong track. I think we've gone too far. I think they've proven out too much and they've done. It's not. They're too amazing. So I think scientifically, I know there are some, some of my academic colleagues think otherwise, but I feel like that's, that's been proven out already.
- [28:37]
Moderator or Interviewer
And in terms of the architecture, you talk about Transformers and you talk about that's 5050 in terms of the two different approaches. Do you sense you and the team getting closer to an architecture that you're feeling more confident could be that pivot point?
- [28:48]
Demis Hassabis
Yeah, I mean it wouldn't be so much like, I mean there could be something out there that's even better than Transformers, but I think that's unlikely. But I'm talking about additional things on top of the system, maybe with the thinking perhaps bringing in more AlphaGo ideas around planning Monte Carlo tree search. So we're trying out lots of these ideas and I would say we have a lot of promising results, but we've got to see how that unfolds over the next year.
- [29:11]
Moderator or Interviewer
A question for both of you from Paul. So I'll start Sebastian with you. What skills will be essential in the future? What should parents encourage their children to learn to prepare for future careers?
- [29:22]
Sebastian Mallaby
Well, look, I mean, high level, just goal setting, not necessarily high level, in fact, but we don't really want the machines to set goals for us. So whether you are doing a voluntary organization in your neighborhood or you're setting up some project, you know, whatever, at a sports club or having an idea for a play you would like to put on, I mean all those things of initiating a project, we would really rather have humans do most of that, I think. And then the second thing I think is sort of doubling down on the human, which is a slightly abstract, loosey goosey kind of term. But if I think about my work, I write, but I don't just write. It's very important to go and talk to human sources, interact with them, learn from them, understand them deeply. And I feel that's probably something that the machine is not going to take away from me. Even if at some point the next
- [30:19]
Sebastian Mallaby (continued)
book, the book after that, finally these chatbots will be able to do long
- [30:24]
Sebastian Mallaby
form writing, which they definitely can't do now,
- [30:28]
Sebastian Mallaby (continued)
even if the writing part, my role was diminished. I think I would repurpose myself towards the more human, human part, which is human to human interaction with the people I write about. Maybe human to human interaction with you guys. I think if we had a couple of servers on the stage and the servers were generating voices, you might not have showed up. So it's sort of. And chess is the great example of this where in 1997 Deep Blue, the computer system of IBM, defeated Garry Kasparov. And ever since then humans have been second rate at chess relative to machines. But since then, more humans play chess, more humans follow chess professionals and they're watching the human professionals, they're not obsessing about which machine system is best because humans relate to other humans. And I think that's what we should keep in mind when we design our own Futures.
- [31:22]
Moderator or Interviewer
Sorry AlphaGo, what's your take on that question?
- [31:28]
Demis Hassabis
I agree with everything Sebastian said. I think those types of activities and skills will become ever more important in the next few years. I think your question was specifically about the youth of today. And what I would say there is. And what I've been saying at universities and schools is, you know, I think you've got to lean in to the new tools and become almost superpowered with them. Sort of like my generation did with computers in the first place, right. And then there was the Internet generation and then so on. And these, there'll be incredible new opportunities opened up by these skills. It's going, these tools, it's going to disrupt a lot of things, that's for sure. But that actually brings opportunities. If you're sort of thinking about it in an entrepreneurial way, I think it may actually need to change education too. Like I feel like maybe some education is in the room. Like I think we should be really reconsidering education from the ground up. My view of it would be that you want to maybe invert the classroom so that the classroom becomes more about collaboration and project based and creative problem solving. And actually, you know, how do you think differently and come up with new ideas, the taste thing, work with others almost like you might get in supervisions at Oxbridge, the supervision part. And then you do the rote learning actually outside of the class where you do it with your AI systems and it's personalized to you. And actually if you think about it, it's a bit crazy what we do today with teachers because every teacher and I have a lot of friends who are teachers who they have to prepare the course materials for their particular class for the math gcse and there's probably thousands of teachers around the country doing that, almost the same material. Wouldn't it be better if you just had one world class person who is brilliant at lecturing doing that wrote material and then the teachers could focus on something probably much more fun for them, which would be encouraging the kids to think and dream and create. Maybe more like, maybe a Montessori school works like today, but actually pulled all the way through the whole of the school system. I don't know if anyone will be ambitious enough to try that. Maybe someone like Singapore will or something like that. But I do think that it would be crazy not to rethink that. And obviously as part of it you would introduce how do you make use of these amazing tools in a way that enhances your brain and enhances your creativity, not diminishes it. And just like with every technological advance, I think there are good and bad ways of using it. I think there's the same with the Internet, the same with mobile and the same with computers in the first place. And I really believe in. What's amazing about us as a species is our brains are general intelligences. And that means we're infinitely adaptable. And I think that's the history of humanity. If you look at where we're sitting today, our brains were involved for hunter gathering and look what we've managed to do with our minds is build modern civilization. I always say when I go over on my too frequent transatlantic flights on the 747, I'm just, I think we all just do it and it's sort of, it's just a normal thing. But I just find it amazing when I look out the window, you're seeing Earth below you and this plane that's so reliable, this massive chunk of metal flying through the sky. I mean, how did we do it? This is incredible. So I believe we'll continue to go that way.
- [34:46]
Moderator or Interviewer
And I think that is a fantastic note to end on. There's a million other questions we in the audience could ask you. We've run out of time, but I think it's a fantastic note to end on. Demis Hassabis, Sebastian Mallaby. Thank you very much indeed. Sebastian's going to be signing his book.
- [35:03]
Mia Sorrenti
Thanks for listening to Intelligence Squared. This episode was produced by Margarita Valparto and it was edited by Mark Roberts. For ad free episodes and full length recordings. You can become a member@intelligencesquared.com membership and if you'd like to join us in person at future live events, you can Find tickets and see our full program over@intelligencesquared.com attend. You've been listening to Intelligence Squared. Thanks for joining us.
- [35:44]
Stitch Fix Advertiser
Stitch Fix Stop shopping. Get styled. Not today sweatpants. Somebody's wearing jeans that fit.
- [35:52]
BetterHelp Advertiser
No photos please. I'm just a regular dad who happens to have a stylist.
- [35:56]
Microsoft Advertiser
I really look my best when someone
- [35:58]
Moderator or Interviewer
else makes the decisions.
- [35:59]
Demis Hassabis
Hey, we can all see you two way mirrors.
- [36:02]
Stitch Fix Advertiser
Just share your size, style and budget and your stylist sends personalized looks right to your door. Stitch Fix get started today@stitch fix.com I want to hug you. I'm going to hug you. I'm coming in for a hug.
- [36:14]
Blinds.com Advertiser
The right window treatments change everything. Your sleep, your privacy, the way every room looks and feels. @blinds.com, we've spent 30 years making it surprisingly simple to get exactly what your home needs. We've covered over 25 million windows and have 50,000 five star reviews to prove we deliver. Whether you DIY it or want a pro to handle everything from measure to install, we have you covered. Real design professionals, free samples, zero pressure right now. Get up to 45% off site wide. Plus get a free professional measure@blinds.com rules and restrictions apply.
- [36:45]
Sweetgreen Advertiser
There's a new way to Sweetgreen Meet Graps handheld Hearty and made for life on the market. With bold, chef crafted flavors, fresh ingredients and over 40 grams of protein, they're built to satisfy without slowing you down. Try wraps today in the app or@order.sweetgreen.com available at all participating locations.
- [37:14]
BetterHelp Advertiser
When I found out I was going to be a parent, I immediately felt a lot of anxiety and worry. So I went on to BetterHelp to try to look for a therapist to help me with that.
- [37:23]
Mia Sorrenti
My relationship with my family and with my boyfriend and with myself were suffering. I really needed help.
- [37:29]
Demis Hassabis
I was ruminating a lot. Really getting those thoughts out to a
- [37:33]
Shopify Advertiser
therapist and getting feedback was just life changing.
- [37:37]
Moderator or Interviewer
Discover what better help online therapy can do for you.
- [37:40]
Home Depot Advertiser
Visit betterhelp.com today.