1 00:00:00,050 --> 00:00:07,980 Welcome to the Oxford Martin School and to this lecture this evening by Paul Clark. 2 00:00:07,980 --> 00:00:18,990 Yesterday, we had the budget from the government, which is clearly very excited about research and innovation, about infrastructure. 3 00:00:18,990 --> 00:00:24,940 And then layered on top of all of that, we have the corona virus sweeping around the world. 4 00:00:24,940 --> 00:00:38,190 So given all of those things going on to have a tech guru of the Jumah, which Dominic Cummings would obviously get very excited about, 5 00:00:38,190 --> 00:00:50,100 and someone who knows more about logistics and how to keep food on supermarket shelves and showing up to people's houses seems to be extremely timely. 6 00:00:50,100 --> 00:00:59,280 So it's great to have Paul Clark here this evening. Paul is responsible for an incredible operation. 7 00:00:59,280 --> 00:01:05,040 Nineteen hundred engineers in Accardo. 8 00:01:05,040 --> 00:01:21,880 And not only is that transforming retailing, but it's also doing its best to shake up and bust incumbents and disrupt all sorts of other sectors. 9 00:01:21,880 --> 00:01:26,020 Originally, Paul read physics at St. John's here. 10 00:01:26,020 --> 00:01:31,810 He was the duty of so Chris Goldsmith, who's in the audience, 11 00:01:31,810 --> 00:01:44,980 and then went on to do a variety of different software engineering, tech, start-ups and consultancy roles. 12 00:01:44,980 --> 00:01:57,660 And has become an increasingly visible and influential figure in the whole field of tech and logistics. 13 00:01:57,660 --> 00:02:04,450 He's on the government's elite Council on the Robotics Growth Partnership. 14 00:02:04,450 --> 00:02:12,100 He's chair of the CBI Innovation Council and he's on the National Food Strategy Advisory Board. 15 00:02:12,100 --> 00:02:19,420 But the other remarkable thing about Paul is that he's a hands on guy. 16 00:02:19,420 --> 00:02:31,180 The first apps which were created, Accardo, he jointly wrote a list as his past time designing circuit boards. 17 00:02:31,180 --> 00:02:38,650 And he continues to write software as well as during all of the other things that you've just heard about. 18 00:02:38,650 --> 00:02:50,720 So it's a great pleasure to have you here this evening, Paul, in the Oxford Martin School. 19 00:02:50,720 --> 00:02:58,340 Thank you, Jim. So I just kept vaguely set up here. So, yes, it's definitely raised the stakes for me to have my executer here. 20 00:02:58,340 --> 00:03:05,660 And that was definitely not something I was expecting. But anyway, delighted. But unless it takes very little to get me back to Oxford. 21 00:03:05,660 --> 00:03:10,210 So it's it's very exciting to be here. And. No, no. 22 00:03:10,210 --> 00:03:17,270 Sorry. Go away. So the robots are taking over. 23 00:03:17,270 --> 00:03:20,600 Yes. I can't really complain about that. All right. 24 00:03:20,600 --> 00:03:24,020 So, yes, I'm going to talk a little bit about what it says. 25 00:03:24,020 --> 00:03:29,180 Here are recipes for exploring the transformation of food. 26 00:03:29,180 --> 00:03:33,200 But I'm definitely going to stray off onto some other big topics, too, 27 00:03:33,200 --> 00:03:39,650 that are increasingly what I spend my time on because I I'm no longer involved in the 28 00:03:39,650 --> 00:03:45,170 operational side of the businesses and explain I'm off much more on the futures kind of stuff. 29 00:03:45,170 --> 00:03:51,320 So I want to start by talking about the kind of the global context that really underpins 30 00:03:51,320 --> 00:03:56,660 why we need to do significant transformation when it comes to food production. 31 00:03:56,660 --> 00:04:07,040 And of course, we have to start with climate change, which is now irrevocably bolted into the worlds like Geist. 32 00:04:07,040 --> 00:04:15,260 And obviously, food production is at the heart of many of the UN's sustainable development goals. 33 00:04:15,260 --> 00:04:20,160 But, of course, the world population is continuing to grow and there are many more mouths to feed. 34 00:04:20,160 --> 00:04:31,200 And unfortunately, in regions where food is most scarce and we have to also be mindful the fact that the the unintended sort of consequences, 35 00:04:31,200 --> 00:04:34,970 if you like, of agriculture is having on the environment, 36 00:04:34,970 --> 00:04:39,360 whether it be greenhouse gases or biodiversity, soil erosion and flooding, 37 00:04:39,360 --> 00:04:46,550 the kind of stuff that Jim worries about and tries to solve is becoming increasingly well understood. 38 00:04:46,550 --> 00:04:49,580 And depending on who you want to listen to, you know, 39 00:04:49,580 --> 00:04:58,790 these are problems that we either need to solve by 2050 in terms of net zero or perhaps 2030 or maybe some other date. 40 00:04:58,790 --> 00:05:03,230 And in, you know, kind of post Brexit world that we now live in. 41 00:05:03,230 --> 00:05:09,620 The implications for food security, food prices, food import tariffs and food standards are upon us. 42 00:05:09,620 --> 00:05:12,710 And that's, of course, before the crisis that we're now living through, 43 00:05:12,710 --> 00:05:17,720 where food security couldn't be even you know, it's been elevated to another whole level. 44 00:05:17,720 --> 00:05:27,020 And we find ourselves at Acardo slightly in the middle of that game in terms of how will people who can't leave their homes get their food. 45 00:05:27,020 --> 00:05:36,200 And there's a growing demand from from consumers to understand where their food comes from and indeed to reduce the food miles to get it to them. 46 00:05:36,200 --> 00:05:45,260 There's a growing focus on food packaging and food waste, whether it be from regulators, suppliers, retailers or the consumers themselves. 47 00:05:45,260 --> 00:05:50,570 And the NHS is shifting its focus from treatment to prevention. 48 00:05:50,570 --> 00:05:55,580 Food has a significant part to play in our general health and wellbeing, as we know. 49 00:05:55,580 --> 00:06:03,410 And we've seen the impact of things like the shark attacks on products that the FMC G's manufacturer on, on consumer behaviour. 50 00:06:03,410 --> 00:06:05,570 And as one of the many things unfortunate, 51 00:06:05,570 --> 00:06:11,660 I can say this now that my mother used to say to me as a child that I dismissed at the time as a complete nonsense, 52 00:06:11,660 --> 00:06:16,170 but which of course turned out to be true. You are what you eat. 53 00:06:16,170 --> 00:06:23,990 And so she was a little bit ahead of her time on that one. So what are the challenges we face in responding to these issues? 54 00:06:23,990 --> 00:06:33,530 Well, food is a complex, a properly complex, interconnected ecosystem with lots and lots of moving parts, not all of which we fully understand yet. 55 00:06:33,530 --> 00:06:37,250 And we need to be careful where and how we squeeze that toothpaste tube, 56 00:06:37,250 --> 00:06:45,380 because there are real risks of all sorts of unintended consequences as we try to optimise different parts of that system. 57 00:06:45,380 --> 00:06:49,640 And we have to face the fact that it's going to be very hard to transform 58 00:06:49,640 --> 00:06:54,740 traditional agriculture in terms of the impact that it has on on the environment. 59 00:06:54,740 --> 00:07:02,450 The effects of climate change are likely to place more more pressures on conventional agriculture in terms of more extreme weather events, 60 00:07:02,450 --> 00:07:04,900 flooding, droughts, storms and so on. 61 00:07:04,900 --> 00:07:13,220 And furthermore, traditional agriculture is frankly very inefficient at turning photons from the sun into calories and protein, 62 00:07:13,220 --> 00:07:22,850 as we know, and that we then want to consume. And if and if you haven't come across it, I strongly recommend this blogsite by Professor Sir Ian Boyd, 63 00:07:22,850 --> 00:07:33,430 who was previously the CSA at different chief scientific adviser, and he's one of the UK's leading climate scientists. 64 00:07:33,430 --> 00:07:40,370 And in his twin September twenty nineteen blog post that you'll find on their Securing the food of the future. 65 00:07:40,370 --> 00:07:46,880 He shares three really interesting facts or statistics that to meet both demand and net zero targets. 66 00:07:46,880 --> 00:07:53,670 Efficiency in traditional agriculture needs to. Piece by a factor of five to 10 times by 2050. 67 00:07:53,670 --> 00:08:00,450 So that's a 500 percent to a thousand percent increase as compared with the current trend of efficiency gains, 68 00:08:00,450 --> 00:08:03,330 which is typically a few percent each year. 69 00:08:03,330 --> 00:08:11,730 And then if you consider a plant that is subsequently eaten by an animal that we then eat of the sun's energy landing on that plant, 70 00:08:11,730 --> 00:08:18,450 only one to two parts per 10000 will actually be converted to calories within the animal meat. 71 00:08:18,450 --> 00:08:27,750 And finally, roughly 80 percent of UK agricultural output is produced by 20 percent of farmers on 50 percent of the land. 72 00:08:27,750 --> 00:08:34,830 And that means that about half of the UK's agricultural land has very low food productivity. 73 00:08:34,830 --> 00:08:41,030 And so if we start talking about, if you like, the system that gets the food to us, 74 00:08:41,030 --> 00:08:46,680 there's already a lot of food that is shipped to us from far away, as we know. 75 00:08:46,680 --> 00:08:54,330 And here we've listed just some of it. And that comes in these huge kind of containers or or reefers. 76 00:08:54,330 --> 00:09:00,660 And one of the challenges is lowering, increasing lead, competing demands from consumers. 77 00:09:00,660 --> 00:09:05,040 On the one hand, they want variety and availability, but without seasonality. 78 00:09:05,040 --> 00:09:10,830 But on the other hand, they don't want to think that they're in any way contributing to any of these factors. 79 00:09:10,830 --> 00:09:13,830 I'm not going to read through them. You can read them there. 80 00:09:13,830 --> 00:09:22,170 And and that, of course, means that globalisation of food production is somewhat at odds with current consumer trends. 81 00:09:22,170 --> 00:09:26,790 So we're going to need some alternative recipes, if you like, for food production. 82 00:09:26,790 --> 00:09:31,110 We're gonna need find more efficient ways to convert photons into calories and protein. 83 00:09:31,110 --> 00:09:36,810 We're gonna need to minimise those in and unintended consequences of food production and consumption on the environment. 84 00:09:36,810 --> 00:09:41,400 We're gonna need to view the world food production machine as a closed holistic 85 00:09:41,400 --> 00:09:46,260 ecosystem and work out how the UK is food production machine sits within that. 86 00:09:46,260 --> 00:09:51,690 As part of that holistic eco system, we need to consider things like how do we move food around and keep it fresh, 87 00:09:51,690 --> 00:09:56,250 including issues such as food, miles, waste packaging and circular economy models. 88 00:09:56,250 --> 00:09:59,970 How can we better educate consumers and transform their behaviours, 89 00:09:59,970 --> 00:10:07,860 including issues such as quality versus quantity of food, alternative sources of protein, waste, incentives and packaging? 90 00:10:07,860 --> 00:10:09,870 How do we change our relationship with food? 91 00:10:09,870 --> 00:10:18,690 Its impact on our health, including issues such as food, as medicine, personalisation of food, obesity and treatment versus wellness. 92 00:10:18,690 --> 00:10:27,960 We need to build a holistic, validated and fully costed model of our existing food system, a food production system both to help us then optimise it, 93 00:10:27,960 --> 00:10:34,680 but also so that we can do like four like comparisons with alternative food production models. 94 00:10:34,680 --> 00:10:41,700 Of course, we need to continue to look for ways to use science, technology and automation to help transform traditional agriculture. 95 00:10:41,700 --> 00:10:48,150 In terms of its efficiency and impact on the environment, whilst being mindful of those limitations that Professor Enjoy talks about. 96 00:10:48,150 --> 00:10:54,000 But at the same time, we need to explore new forms of farming that offer the possibility of delivering the efficiency 97 00:10:54,000 --> 00:10:58,860 and sustainable get sustainability gains that we require to meet those net zero targets, 98 00:10:58,860 --> 00:11:06,240 including technologies such as vertical farming, insect protein, aquaculture, algae production and others. 99 00:11:06,240 --> 00:11:11,820 And we need to move food production from an agricultural process to a manufacturing process with 100 00:11:11,820 --> 00:11:18,600 improved sustainability whilst preserving or hopefully improving factors such as quality and nutrition. 101 00:11:18,600 --> 00:11:24,960 And all of this is is very much at the heart of the national food strategy that's under development. 102 00:11:24,960 --> 00:11:29,970 So before I talk about some of the puzzle recipes we're engaged in around food production, 103 00:11:29,970 --> 00:11:35,580 I want to introduce you to some of the disruptive ingredients, if you like, that we cook with at. 104 00:11:35,580 --> 00:11:41,100 So as has been said already by Jim Acardo is a world leader in smart logistics. 105 00:11:41,100 --> 00:11:46,140 I used to introduce Acardo as the world's largest online only grocery business, but we aren't that anymore. 106 00:11:46,140 --> 00:11:48,540 We still own 50 percent of one of those. 107 00:11:48,540 --> 00:11:54,750 It's the other half being owned by Marks Spencers, but that as of last year, we've divested that into a joint venture. 108 00:11:54,750 --> 00:12:01,740 But what that means is that it's left what was always there, which is this technology company at our core, owns that. 109 00:12:01,740 --> 00:12:04,080 That's not our core. But we'll get to that. 110 00:12:04,080 --> 00:12:10,230 So our founding vision was to use a huge amount of technology and automation to do online grocery scalable, sustainably and profitably. 111 00:12:10,230 --> 00:12:16,680 And we've done that. And it was also part of that vision that once we had evolved decisions for ourselves, 112 00:12:16,680 --> 00:12:20,460 that we would make it available to other retailers around the world and were 113 00:12:20,460 --> 00:12:25,290 busy doing that at the moment with our platforms and the our business model. 114 00:12:25,290 --> 00:12:30,000 Are these huge automated warehouses that the largest, most sophisticated, their kind in the world? 115 00:12:30,000 --> 00:12:35,400 And as a disruptor, you know, there was no template for them. We had to build them from scratch. 116 00:12:35,400 --> 00:12:37,980 We had to create much of the unblind technology ourselves. 117 00:12:37,980 --> 00:12:45,800 We've always written all the software ourselves, often throwing away the software that came with that hardware and then more recently. 118 00:12:45,800 --> 00:12:49,370 And so and then buying in the hardware, often stealing it from. 119 00:12:49,370 --> 00:12:54,290 All sorts of different industries and adapting it for purposes for which it was never intended. 120 00:12:54,290 --> 00:13:01,340 But then more recently, we started to build the hardware too, including the swarms of thousands of robots that now assemble our customers orders. 121 00:13:01,340 --> 00:13:06,830 And the way to think about this, if you haven't seen it before and film and if you have, is think of it like a giant chessboard. 122 00:13:06,830 --> 00:13:12,290 And on that chessboard, there are robots that a bit like rooks can move along rows and columns. 123 00:13:12,290 --> 00:13:18,290 And under every chess square is a stack of storage bins containing groceries up to 21 bins deep. 124 00:13:18,290 --> 00:13:23,540 And the robots, they can stop on a square. They lower a kind of a grab, a bit like a fairground grab of one, 125 00:13:23,540 --> 00:13:27,410 actually pick something up and they latch onto the top a bin in the stack and they bring it 126 00:13:27,410 --> 00:13:31,460 up into the body of the robot and then they can move to another square and drop it off. 127 00:13:31,460 --> 00:13:38,690 And if that's all they could do, they get very bored and frustrated so they can also bring those bins of cucumbers or sausages, whatever it is, 128 00:13:38,690 --> 00:13:42,800 to a pig station where a robot or another story, 129 00:13:42,800 --> 00:13:51,110 a human or another kind of robot picks groceries from a bin into a customer order or to decant stations where products are coming in from. 130 00:13:51,110 --> 00:13:56,510 Suppliers on pallets are decanted into the bins and then the robots put them away in the grid. 131 00:13:56,510 --> 00:14:01,460 And we call these hives. And in our second generation warehouses, there are typically two of them, 132 00:14:01,460 --> 00:14:08,300 one for ambient products and one for chilled products, frozen products we do in a different way at the moment. 133 00:14:08,300 --> 00:14:14,000 And the robots, they collaborate with one another and be really careful if they're soffer engineers in the room, 134 00:14:14,000 --> 00:14:19,910 because it's not what Soffer engineers call collaboration, because it's not they don't talk to each other. 135 00:14:19,910 --> 00:14:25,140 There is a hive mind that kind of runs the whole process and orchestrates the robots. 136 00:14:25,140 --> 00:14:33,500 And but that pseudo collaborative behaviour means that we can pick a typical 50 item order in about five minutes. 137 00:14:33,500 --> 00:14:39,020 And that's really, really important, given that there is this recipe for immediacy that's going on where everybody wants that order, 138 00:14:39,020 --> 00:14:43,580 you know, within hours or minutes rather than days. 139 00:14:43,580 --> 00:14:48,980 And if a robot, say, wants to get to a bin that's fourth down in the stack, 140 00:14:48,980 --> 00:14:54,170 it just gets three of its friends to move the top three bins out of the way and then it grabs the one that it wants. 141 00:14:54,170 --> 00:15:00,860 And in our latest warehouse, to use this technology, which is in southeast London, which went live in June 2018. 142 00:15:00,860 --> 00:15:05,600 Each of these grids is about the size of three football pitches. 143 00:15:05,600 --> 00:15:10,530 And when it's fully ramped up, it's not at the moment which is turning up the dial at the moment. 144 00:15:10,530 --> 00:15:14,420 It love about 3500 robots roaming around on top of those grids. 145 00:15:14,420 --> 00:15:22,490 And more interestingly, we've got to build 30 of these warehouses around the world over the next three years for our Acardo smart platform customers. 146 00:15:22,490 --> 00:15:26,810 And that's if we don't sign any more deals, which we certainly intend to do. 147 00:15:26,810 --> 00:15:33,770 So the Holy Grail, though, of all of robotics in online grocery is actually the picking and the packing of the customer orders. 148 00:15:33,770 --> 00:15:41,420 And that means for us, picking 55000 different products with multiple different form factors and packing them in the correct sequence. 149 00:15:41,420 --> 00:15:45,050 And in the correct 3D orientation into carrier bags. 150 00:15:45,050 --> 00:15:52,220 So it's unlike welding, you know, the chassis of a of a car and a production line or spray painting it for us. 151 00:15:52,220 --> 00:15:57,830 It's very much about making smart decisions on the fly and dealing with the unpredictable. 152 00:15:57,830 --> 00:16:03,470 So it's all about the kind of the grippers, the machine learning and the vision systems rather than the robotics. 153 00:16:03,470 --> 00:16:10,160 And one of the many robotics research projects that we have are engaged in is about developing these next generation grippers. 154 00:16:10,160 --> 00:16:14,140 This was the Horizon 2020 one called Soma, which finished earlier. 155 00:16:14,140 --> 00:16:20,210 So sometime last year. And it's about building grippers that have capabilities, similar human hands. 156 00:16:20,210 --> 00:16:23,750 And with, you know, it or not, you know, since you were babies, 157 00:16:23,750 --> 00:16:29,690 you've been learning strategies for how to pick stuff up as part of learning how to interact with the world around you. 158 00:16:29,690 --> 00:16:36,590 And if you were gonna pick up a wine bottle here. But if you're going to pick up a wine bottle and put it into a wine rack, you know, 159 00:16:36,590 --> 00:16:43,520 you you would instinctively grab it by the neck because you know that if you grabbed it halfway down, your your hand would be in the way. 160 00:16:43,520 --> 00:16:49,160 I mean, maybe you weren't doing this as babies, but but the point is, you would also automatically and, you know, 161 00:16:49,160 --> 00:16:55,880 cantilever it by putting perhaps two fingers under the neck because you know, that to turn it through 90 degrees, you're gonna have to support it. 162 00:16:55,880 --> 00:17:03,740 Now, nobody taught you how to do that. You know, you either learnt by observing your parents doing it maybe as part of your misspent youth, 163 00:17:03,740 --> 00:17:08,810 but or you you experimented and came up with that strategy yourself. 164 00:17:08,810 --> 00:17:12,260 But, of course, robots have no idea about those strategies. 165 00:17:12,260 --> 00:17:20,470 So they have to either be taught how to do it using learning by demonstration or which is one of the basket of technologies, the AI family or you, 166 00:17:20,470 --> 00:17:25,190 they have to be given the chance to learn those tragedies themselves, either in living labs, 167 00:17:25,190 --> 00:17:33,260 which I'll come and talk about, or using simulation simulated models, if you like, of that physical environment. 168 00:17:33,260 --> 00:17:39,920 But what they don't have time is to learn the strategies on the fly when they're actually picking a real order because that would take too long. 169 00:17:39,920 --> 00:17:43,760 So they have to sort of acquire them in advance and put them on the shelf. 170 00:17:43,760 --> 00:17:49,120 And then when they're faced with a particular situation, they can go, oh, I need strategy number thirty 37. 171 00:17:49,120 --> 00:17:56,790 Taken off the shelf and they apply it. The particular problem that's in front of them. It's one of our other Horizon 2020 funding projects. 172 00:17:56,790 --> 00:18:02,370 Is that still ongoing? Finishes soon. Is about building a humanoid maintenance robot. 173 00:18:02,370 --> 00:18:05,340 And that's not to what to do, 174 00:18:05,340 --> 00:18:11,610 but learns by observing human engineers that work and then uses inference to work out how to collaborate with them on their tasks. 175 00:18:11,610 --> 00:18:19,020 And we've been gently torturing this robot in one of our labs in Hatfield to see what it can do, 176 00:18:19,020 --> 00:18:21,840 a bit like kind of subjecting it to darker type challenges. And. 177 00:18:21,840 --> 00:18:26,940 And the reason why we want to build robots like that is that is, of course, what we learn along the way, which is transferable. 178 00:18:26,940 --> 00:18:32,550 But it's also the fact that we're going to build so many of these warehouses around the world that automating their maintenance is important to us, 179 00:18:32,550 --> 00:18:39,930 but also ultimately automating their construction to the next ingredient is, of course, A.I., a machine learning. 180 00:18:39,930 --> 00:18:44,890 And, you know, A.I. lets you do the really exciting things with other disruptive technologies. 181 00:18:44,890 --> 00:18:49,170 You know, without A.I., a robot is occupied just a pile of electromechanical experiments. 182 00:18:49,170 --> 00:18:56,490 And it is the smart glue in the industry found for zero family of of of, if you like, component technologies. 183 00:18:56,490 --> 00:19:02,010 It can discover or at least unearth knowledge that's been hidden to humans because of factors such as complex in bias. 184 00:19:02,010 --> 00:19:06,060 But most importantly, it's a recursive technology. Shush. 185 00:19:06,060 --> 00:19:11,460 Sorry, you can use one generation of A.I. to help train the next more advanced one. 186 00:19:11,460 --> 00:19:15,330 And that means that if you can call it an intelligence and let me be very careful about that, 187 00:19:15,330 --> 00:19:21,480 but it will certainly evolve its fourth version of intelligence faster than human intelligence. 188 00:19:21,480 --> 00:19:26,400 So when it comes to disruptive technologies in our terms, it is in the Tolkien sense, 189 00:19:26,400 --> 00:19:30,840 the one to rule them all, at least until quantum computing comes along and knocks it off that perch. 190 00:19:30,840 --> 00:19:36,540 But and applications of AI machine learning they completely pervade are our platform. 191 00:19:36,540 --> 00:19:39,870 So things like natural language processing, you know, 192 00:19:39,870 --> 00:19:45,480 customers placing orders with voice and allowing people to talk to robots and things like that, image recognition, 193 00:19:45,480 --> 00:19:51,390 smart machines that we build that route, optimising the routes that our fans drive on a daily basis, 194 00:19:51,390 --> 00:19:54,930 providing monitoring and oversight across the platform many more. 195 00:19:54,930 --> 00:19:59,250 And we're on this kind of slightly crazy journey, if you like, to build an Acardo brain. 196 00:19:59,250 --> 00:20:06,690 So a general A.I. that understands our business model, that understands our data, taxonomies, that has oversight of our N10 platform. 197 00:20:06,690 --> 00:20:11,580 And I either can answer questions about the business that can help us work more efficiently and make less mistakes. 198 00:20:11,580 --> 00:20:16,680 That can act like the third JRA on an aircraft spotting the onset of problems that are about to impactors us that can 199 00:20:16,680 --> 00:20:23,520 help coordinate our suppliers and that can help our customers shop faster with less friction and greater delight. 200 00:20:23,520 --> 00:20:29,250 But obviously, for many of you in the room, we'll know that kind of general A.I. is completely beyond the current state of the art. 201 00:20:29,250 --> 00:20:35,640 So what we're really doing is focussing on building, if you like, the individual pieces that puzzle the individual neurones of that brain. 202 00:20:35,640 --> 00:20:42,210 And then we'll join them up over time as A.I. Technologies, computing power continue to advance. 203 00:20:42,210 --> 00:20:45,720 And we're really not so fast on the current outcomes. They're great, 204 00:20:45,720 --> 00:20:51,120 but we're really far in fussed about is the learnings and the competencies that we 205 00:20:51,120 --> 00:20:56,580 will acquire along what is now firmly for us and A.I. and robotics first journey. 206 00:20:56,580 --> 00:21:02,340 And the next ingredient is the digital twin. 207 00:21:02,340 --> 00:21:04,710 Now I'm gonna explain what system twins are in. 208 00:21:04,710 --> 00:21:12,870 A second friend, he doesn't know, but for me they are the Cinderella of this disruptive family of technologies because they're often overlooked. 209 00:21:12,870 --> 00:21:19,560 But with A.I. and robotics as the big ugly sisters and digital twins combined with robotics let you, 210 00:21:19,560 --> 00:21:25,560 do you risk the physical by learning earlier and faster within virtual worlds and digital twins? 211 00:21:25,560 --> 00:21:32,310 When you couple them with machine learning, they can. You can use them to explore and optimise those virtual worlds. 212 00:21:32,310 --> 00:21:35,610 But unfortunately, the terms is between has been completely hijacked. 213 00:21:35,610 --> 00:21:39,960 And lots and lots of people talk about digital twins that aren't ready, digital twins. But for us, 214 00:21:39,960 --> 00:21:47,910 what differentiates a true digital twin from other digital models or simulations is the fact that your conjoining these digital and physical worlds. 215 00:21:47,910 --> 00:21:54,780 So let's have a real example. Imagine you wanted to optimise the traffic flow around a city like Oxford, which definitely needs it. 216 00:21:54,780 --> 00:22:02,760 The road network and the traffic on it is your physical twin. And you'd probably put cameras in census and junction on the junctions and collect all 217 00:22:02,760 --> 00:22:08,310 the data about the people and bicyclists and dogs and everything else moving around. 218 00:22:08,310 --> 00:22:13,020 And then you'd feed those data into a digital model or simulation of the traffic network. 219 00:22:13,020 --> 00:22:16,590 And then you could use that model to optimise the topology of the road network, 220 00:22:16,590 --> 00:22:22,260 maybe for future things you had built or the timing of the traffic lights and things like that. 221 00:22:22,260 --> 00:22:29,220 And you could then feed those outcomes back into the physical twin and actually update, if you like, the timing of the real traffic lights. 222 00:22:29,220 --> 00:22:32,580 And that's the conjoining of the circle that makes the digital twin. 223 00:22:32,580 --> 00:22:39,750 And another example of that for us is the digital twin that we have that looks after these swarms of robots. 224 00:22:39,750 --> 00:22:49,510 Each of these robots that fly around at about four metres a second, they generate about 5000 data points a thousand times a second and. 225 00:22:49,510 --> 00:22:56,290 That's about a gigabyte of data per robot per day, or about four terabytes of data per sworn per day. 226 00:22:56,290 --> 00:22:57,640 And that's just one warehouse. 227 00:22:57,640 --> 00:23:05,050 And there's no way that human engineers staring at screens could oversee, let alone optimise, such a complex system in near real time. 228 00:23:05,050 --> 00:23:11,260 It's completely beyond human scale. And so what we do is we stream all that data is the cloud and we build a healthcare 229 00:23:11,260 --> 00:23:16,120 system up in the cloud that looks after the swarm by doing predictive maintenance, 230 00:23:16,120 --> 00:23:18,820 trying to spot the onset of problems before they become problems. 231 00:23:18,820 --> 00:23:24,700 And it's a bit like the what we can do with patient data and wearables and sensors in the home, 232 00:23:24,700 --> 00:23:30,580 maybe for humans where we can do remote medicine, if you like, with a similar system. 233 00:23:30,580 --> 00:23:37,780 But what's special about a swarm of identical robots is if any of them does get sick or needs a rest or charge of batteries or have a service, 234 00:23:37,780 --> 00:23:42,220 it can come off into the pits and then any of the others can take its place. 235 00:23:42,220 --> 00:23:48,310 But as well as feeding into the healthcare system, we also feed all that data into our digital twin of the hive. 236 00:23:48,310 --> 00:23:52,120 And that optimises, if you like, the behaviours of us of the swarm. 237 00:23:52,120 --> 00:23:59,500 And then we update the real parameters of the control system of the real time control system that manages the hive. 238 00:23:59,500 --> 00:24:04,750 And this is, if you guessed already is a fish is a video of that digital twin. 239 00:24:04,750 --> 00:24:10,480 And we build simulations of all sorts of parts of our business, like demand forecasting what people are gonna buy in a van, 240 00:24:10,480 --> 00:24:14,020 rooting, you know, the routes that are thousands of fans drive each day. 241 00:24:14,020 --> 00:24:17,800 And what we're really on a journey is to build an end to end flight simulation, 242 00:24:17,800 --> 00:24:23,140 if you like, of our whole e-commerce fulfilment and logistics platform. 243 00:24:23,140 --> 00:24:26,020 And learning with digital twins is powerful, but it has its limits. 244 00:24:26,020 --> 00:24:34,090 It's really quite hard at the moment to simulate more abstract concepts such as public adoption, ethics and privacy concerns in a digital twin. 245 00:24:34,090 --> 00:24:41,620 And this is where living labs come in. And unlike technology, demonstrator's or sandpits, living labs or all about learning by doing. 246 00:24:41,620 --> 00:24:46,240 In the real world environment, delivering real services to real customers day in, day out. 247 00:24:46,240 --> 00:24:49,690 And why that's important is because customers keep you honest. 248 00:24:49,690 --> 00:24:55,570 They give you feedback. Sometimes good, sometimes bad. And they help drive pace and agility. 249 00:24:55,570 --> 00:25:02,290 And that living lab provides a physical environment that is representative of the real world. 250 00:25:02,290 --> 00:25:07,090 But it's initially more constrained to allow you to learn faster and with less risk. 251 00:25:07,090 --> 00:25:12,190 And you can then loosen those constraints over time as your capabilities and competence grow. 252 00:25:12,190 --> 00:25:17,680 And we use living labs to let our robots to learn how to pick those 55000 in products. 253 00:25:17,680 --> 00:25:24,760 But also the whole first warehouse in Hatfield was a living lab because over its kind of 15 years, 254 00:25:24,760 --> 00:25:28,480 when it was being rapidly, innovation is now in a maintenance mode. 255 00:25:28,480 --> 00:25:35,020 It was delivering millions of customer orders, but at the same time, we were ruthlessly inventing new mouse traps, 256 00:25:35,020 --> 00:25:43,060 ripping out the old one and putting the new one in whilst it was kind of in-flight, which is easy to do, say, but not so easy to do. 257 00:25:43,060 --> 00:25:49,300 But the UK needs living labs to accelerate learning towards the future of mobility and smart cities and things like that, 258 00:25:49,300 --> 00:25:54,070 complex environments where we're not going to do it by theory alone. 259 00:25:54,070 --> 00:25:59,110 And we need to learn about how the interplay of different technologies, changes to regulatory models, 260 00:25:59,110 --> 00:26:03,940 stakeholder management, safety, new business models, public adoption and many others will all fit together. 261 00:26:03,940 --> 00:26:06,680 And these are complex challenges to overcome. 262 00:26:06,680 --> 00:26:12,190 And at the moment, I think we believe there's far too much kind of trying to swim the Atlantic by standing around swimming pools. 263 00:26:12,190 --> 00:26:15,250 And so we're embarking on turning the Hatfield Business Park, 264 00:26:15,250 --> 00:26:22,390 where we're based into a living lab to explore the intersection of autonomous vehicles, drones, robots, smart infrastructure and smart services. 265 00:26:22,390 --> 00:26:27,760 And we formed a consortium are now in the midst of our final funding application for that. 266 00:26:27,760 --> 00:26:32,350 But this living lab will just be the first step in a much bigger vision. 267 00:26:32,350 --> 00:26:40,480 Firstly, to turn Hertfordshire into a county scale living lab and then to encourage other business parks and counties around the UK, including Oxford, 268 00:26:40,480 --> 00:26:47,950 to replicate what we've done and what they are doing already in many cases, to create a family of these interconnected living labs. 269 00:26:47,950 --> 00:26:54,220 And we believe that then we might seriously move the dial in terms of things like the future mobility. 270 00:26:54,220 --> 00:27:01,060 And we as a company operate at this intersection of A.I. machine learning, robotics, digital twins, living labs. 271 00:27:01,060 --> 00:27:06,370 We build smart mobile machines which are plugged into the world around them with the Internet of things that stream their data, 272 00:27:06,370 --> 00:27:14,530 exhaust the cloud, and then we use digital twins and living labs to help design and optimise those complex systems and. 273 00:27:14,530 --> 00:27:19,120 With these living lamps along the way, I'm gonna get the food and say we. 274 00:27:19,120 --> 00:27:25,270 We will evolve and test the frameworks and standards for crowdsourcing digital twins as a step towards assembling. 275 00:27:25,270 --> 00:27:31,330 These are the national scale. And this all leads to the idea of creating a national digital twin of the UK, 276 00:27:31,330 --> 00:27:36,280 which is a truly ambitious vision and some might dismiss as unachievable. 277 00:27:36,280 --> 00:27:39,410 But there are people as mad as us who believe in this. 278 00:27:39,410 --> 00:27:46,000 So the centre of digital built Britain, part of Cambridge and University, and the National Infrastructure Commission. 279 00:27:46,000 --> 00:27:54,790 Others do share this vision, and it's all about how you would crowdsource them and glue them together using kind of common frameworks and whatever. 280 00:27:54,790 --> 00:27:56,320 And it's a task that you'll never finish. 281 00:27:56,320 --> 00:28:02,290 You're never gonna build a national twin because there's always going to be more things and higher levels of fidelity that you might as a model. 282 00:28:02,290 --> 00:28:07,270 But the good news is that you don't need to get to the end because each of the digital twins in its own right is useful. 283 00:28:07,270 --> 00:28:11,980 But when you start gluing together, you get combinatorial effects. So, for example, 284 00:28:11,980 --> 00:28:16,330 having additional twin of the UK rail network would be cool and useful for optimising 285 00:28:16,330 --> 00:28:20,440 timetables and minimising delays and risk the impact on maintenance and things like that. 286 00:28:20,440 --> 00:28:25,450 But if you then start combining that with digital trains, with road network airports and seaports, 287 00:28:25,450 --> 00:28:29,320 now you start to get insights across the whole transportation system. 288 00:28:29,320 --> 00:28:35,020 Then you add additional train to the energy network and you might work out where you should put all the E.V. charging points and things like that. 289 00:28:35,020 --> 00:28:38,720 So I think I've said enough participants. 290 00:28:38,720 --> 00:28:44,250 I messed up. Right. Let's carry on. So. 291 00:28:44,250 --> 00:28:51,630 One last bit before we go on to recipe's, which is let's just imagine you are going to erect a building like the Shard. 292 00:28:51,630 --> 00:28:55,950 You know, it's a highly complex building with lots of new technologies. How would you go about that? 293 00:28:55,950 --> 00:29:02,100 Well, obviously, it's you know, what you do is you clear the ground. You get a group of builders down to the site. 294 00:29:02,100 --> 00:29:06,120 You show a picture of the building you want and you say get stuck in. 295 00:29:06,120 --> 00:29:11,880 And now whenever I say that, you know, a true group like you and they don't smile, 296 00:29:11,880 --> 00:29:15,680 it's quite worrying because then I think, oh, my God, that's how you do think you build a shop. 297 00:29:15,680 --> 00:29:19,710 But anyway, you're smiling. So that's okay. But the you wouldn't do that. 298 00:29:19,710 --> 00:29:24,900 It would be an absolute unmitigated disaster. So what would you do? Well, the first thing you might do is build digital twin of the building. 299 00:29:24,900 --> 00:29:29,490 I certainly would. And but the other thing you would definitely do is you create a set of layer drawings. 300 00:29:29,490 --> 00:29:35,400 So architectural engineering, h fact, electrical, plumbing, networking, building management and so on. 301 00:29:35,400 --> 00:29:38,370 And those buildings, boops serve many purposes, 302 00:29:38,370 --> 00:29:43,380 things like transforming the high level vision of the architect into something that can actually be built, 303 00:29:43,380 --> 00:29:48,540 helping drive greater collaboration and communication across stakeholders and contractors, 304 00:29:48,540 --> 00:29:51,900 enforcing building standards and best practise managing complexity. 305 00:29:51,900 --> 00:29:55,680 Because you can turn the layers on and off and just see the ones you're interested in. 306 00:29:55,680 --> 00:29:59,760 And after the building's been built, it acts as a document of. 307 00:29:59,760 --> 00:30:03,270 They act as a document of record aid. The maintenance. 308 00:30:03,270 --> 00:30:11,220 Well, the challenge, I think, is that we're not just trying to build a smart building without a digital twin and a set of ladder drawings. 309 00:30:11,220 --> 00:30:15,300 I believe we're trying to build and operate a smart country without them. 310 00:30:15,300 --> 00:30:21,810 So I believe we need a set of Ladds semantic maps of the UK, including of the food system, 311 00:30:21,810 --> 00:30:25,800 to enable people to understand the bigger vision and how the different pieces fit together, 312 00:30:25,800 --> 00:30:30,060 to enable people to navigate around the institutions and government, 313 00:30:30,060 --> 00:30:35,490 which are often impenetrable from those on the outside to enable people to understand the mesh network of stakeholders, 314 00:30:35,490 --> 00:30:43,890 relationships, technologies and so on, in order to foster better collaboration and foster innovation, to enforce standards and interfaces, 315 00:30:43,890 --> 00:30:50,550 to enable a diverse population of contributors to crowdsource ideas for completing and improving the maps. 316 00:30:50,550 --> 00:30:56,130 And I believe that would even foster a new form of distributed democracy by enabling communities to take ownership of 317 00:30:56,130 --> 00:31:02,370 completing their parts of the map to help people avoid creating the same non-competing technological building blocks. 318 00:31:02,370 --> 00:31:07,950 If you like the common Lego, where the unhealthy diversity at the moment is not going to be good for us. 319 00:31:07,950 --> 00:31:13,620 I'll come back to that to capture via a change control of versioning the historical evolution of the maps as a document of 320 00:31:13,620 --> 00:31:21,660 record and then to provide methods to safely and control me access and share information across and within the semantic maps. 321 00:31:21,660 --> 00:31:25,860 Because I can assure you, if you go and talk to people like she says que as I have about this, 322 00:31:25,860 --> 00:31:29,490 they get quite twitchy about what this in the wrong hands would do. 323 00:31:29,490 --> 00:31:33,990 Anyway, the base layers would be the kind of geospatial layer. 324 00:31:33,990 --> 00:31:38,430 So, you know, environments, buildings, roads, the physical assets where things live. 325 00:31:38,430 --> 00:31:45,550 And that's what organisations like the Geospatial Commission, part of the Cabinet Office and the audit and all that survey are focussed on building. 326 00:31:45,550 --> 00:31:51,630 And then you might have some properly locked down layers for things like the security services, defence, police and so on. 327 00:31:51,630 --> 00:31:55,560 And then as you move up through the layers, you might have layers for institutions such as national government, 328 00:31:55,560 --> 00:32:00,630 local government regulators, NHS still restrictive but visible to more people. 329 00:32:00,630 --> 00:32:02,400 Then there might be layers for more general purposes, 330 00:32:02,400 --> 00:32:07,770 including fostering collaboration and communication between a technology company such as Accardo. 331 00:32:07,770 --> 00:32:12,750 And then as we move up through the layers, we get to more abstract concepts such as relationships, stakeholders, influences. 332 00:32:12,750 --> 00:32:16,380 And this is where we're moving from. Knapp's to semantic knowledge Groff's. 333 00:32:16,380 --> 00:32:22,890 And eventually we get to layers that might be publicly accessible where regions, communities and even interview systems could decorate. 334 00:32:22,890 --> 00:32:28,920 So, for example, the Highlands and Islands of Scotland might want to add rather different information to their part of the map than say, 335 00:32:28,920 --> 00:32:39,300 you know, a city like Oxford. So those are some of the disruptive ingredients that are important to rest in the recipes that were cooking. 336 00:32:39,300 --> 00:32:47,070 But I want to talk a little bit about how how we kind of glue those together in terms of how we approach in innovation. 337 00:32:47,070 --> 00:32:53,040 We've always been this kind of strange blend of a technology business, a retailer and a and a disruptive innovator. 338 00:32:53,040 --> 00:32:57,540 But we've recently now become a platform business, too. 339 00:32:57,540 --> 00:33:02,130 And it's really that innovation factory, which is where the exciting things happen. 340 00:33:02,130 --> 00:33:07,500 I'm going to skip a little bit here because we're going to run out of time otherwise. 341 00:33:07,500 --> 00:33:12,870 And when you're doing a kind of military campaign, there are those areas you want to capture and occupy with your own forces, 342 00:33:12,870 --> 00:33:16,860 the areas where you need to collaborate with your allies in order to make advances in 343 00:33:16,860 --> 00:33:20,250 the areas where you need to stay in touch with what's going on with intelligence, 344 00:33:20,250 --> 00:33:26,910 but where you're not going to be directly involved and the areas where you currently occupy that you can't occupy, which you may need to give up. 345 00:33:26,910 --> 00:33:35,040 Well, how do we like to think about our disruptive landscape, if you like, in a similar kind of way to that campaign? 346 00:33:35,040 --> 00:33:37,080 And in terms of how we map things out, 347 00:33:37,080 --> 00:33:42,420 so the top right hand quadrant is are the things that are truly differentiating where we're going to do it ourselves. 348 00:33:42,420 --> 00:33:47,220 We're going to do the R&D. File the pants, we're going to own the intellectual property and monetise it. 349 00:33:47,220 --> 00:33:51,840 Then the top left hand quadrant is still stuff that's very important, but not quite as strategic. 350 00:33:51,840 --> 00:33:56,370 Where maybe we don't have the skill sets, do it completely ourselves, where we're going to collaborate, 351 00:33:56,370 --> 00:34:00,270 we're gonna find partners, we're going to do joint ventures and we're gonna share the ancestral property. 352 00:34:00,270 --> 00:34:04,600 Bottom left hand corner is where things are now moving towards commodity. 353 00:34:04,600 --> 00:34:08,250 But there's still not you know, vanilla products won't do it for us here. 354 00:34:08,250 --> 00:34:16,260 We want to work with suppliers, get early access to technologies, but also influence the development pipeline of the products that are being produced. 355 00:34:16,260 --> 00:34:20,340 And then the bottom right hand quadrant is where things have now moved to commodity. 356 00:34:20,340 --> 00:34:27,660 They are truly not differentiating. If we ever built them in the past, we should stop building them and we should just buy them off the shelf. 357 00:34:27,660 --> 00:34:33,900 And our research activities are a bit like sending out scouts to explore that top right hand quadrant. 358 00:34:33,900 --> 00:34:39,780 And the patents that we filed are a bit like defensive bunkers to hold land until the main forces arrive. 359 00:34:39,780 --> 00:34:45,510 And sticking with the military analogy are advanced. Research teams are like special forces. 360 00:34:45,510 --> 00:34:51,330 They are multidisciplinary, they are self-sufficient. They operate with different rules of engagement. 361 00:34:51,330 --> 00:34:54,330 They can go behind enemy lines and blow stuff up. 362 00:34:54,330 --> 00:35:02,680 And once we've mapped our destructive landscape, the next step is to identify the technologies and competencies that will be required to implement it. 363 00:35:02,680 --> 00:35:07,470 And this is a bit like, you know, having a complex kind of Lego model that you want to build. 364 00:35:07,470 --> 00:35:10,890 But what you could doing is trying to work out what are all the different shapes of Lego pieces 365 00:35:10,890 --> 00:35:16,620 that you'd need to build it and how many of each of those pieces you'll need to constructed. 366 00:35:16,620 --> 00:35:23,190 And one way to view our history is that really over the last eight years, we've been on this kind of crazy shopping trip of innovation assets. 367 00:35:23,190 --> 00:35:26,610 So data, intellectual property, knowhow, technologies and competencies, 368 00:35:26,610 --> 00:35:31,110 we have to have used online grocery as the vehicle for that trip and all of that, 369 00:35:31,110 --> 00:35:35,550 the competencies that the really valuable part, because although they take time to acquire, 370 00:35:35,550 --> 00:35:39,000 they're a source of competitive advantage, they're not easily displaced. 371 00:35:39,000 --> 00:35:42,930 And once you've acquired them, they can be used all over the places, all over the place. 372 00:35:42,930 --> 00:35:49,410 And some of our key core competencies are things like simulation, A.I., machine learning, optimisation, motion control, 373 00:35:49,410 --> 00:35:55,770 understanding how things move in, search of things, cloud computing, robotics and disruptive innovation itself. 374 00:35:55,770 --> 00:36:02,820 And the reality is that those innovation assets know almost nothing about groceries because they don't need to, 375 00:36:02,820 --> 00:36:07,290 because groceries in our terms are just atoms with certain properties. 376 00:36:07,290 --> 00:36:11,610 They're actually quite troublesome atoms because they have to be kept to the right temperature and, you know, 377 00:36:11,610 --> 00:36:16,830 they'll crush one another if you pack them in the wrong order and they have short shelf life and things like that. 378 00:36:16,830 --> 00:36:24,430 So what that means is if you can do if you can banish food atoms, you can actually manage a bunch of other atoms, too. 379 00:36:24,430 --> 00:36:32,100 And and that's what we're now engaged with. And I used to run a car technology that does the heavy lifting in terms of building our platforms. 380 00:36:32,100 --> 00:36:39,390 But in 80, April 18, 2018, I handed that over and to go off to create a new division called The Office, the CTO Opto. 381 00:36:39,390 --> 00:36:45,150 And our Rippert remit is very much around research, advanced research, IP strategy, 382 00:36:45,150 --> 00:36:48,870 relationships with government, lots going on there at the moment, rhetoricians, the universities, 383 00:36:48,870 --> 00:36:53,370 whether it be collaborative research or internship programmes or Horizon 2020, whatever it is, 384 00:36:53,370 --> 00:36:57,000 stuff with schools and digital literacy, technological future of the business. 385 00:36:57,000 --> 00:37:01,320 But it's also about pursuing these spinner applications. 386 00:37:01,320 --> 00:37:09,090 And here's a slide from a recent are out results announcement that shows some of the spinner applications that we're working on and on here. 387 00:37:09,090 --> 00:37:12,600 You can see things like vertical farming, which I'm about to get to, car parking, 388 00:37:12,600 --> 00:37:17,820 baggage handling parcels, siltation, container ports, rail yards and many others. 389 00:37:17,820 --> 00:37:20,790 And what they have in common is this concept of atoms. 390 00:37:20,790 --> 00:37:26,790 And indeed, a CARDOZ mission is a kind of tag line that you see on our stuff is changing the way the world shops. 391 00:37:26,790 --> 00:37:34,290 But Ottos mission, as it says here, is even more exciting, which is changing the way the world stores sorts assembles moves and sells atoms. 392 00:37:34,290 --> 00:37:39,060 And so we have quite a lot to do. And the first thing we do when we create these spin outs is the far patterns. 393 00:37:39,060 --> 00:37:42,740 And next thing we do is build simulations which eventually will become digital twins. 394 00:37:42,740 --> 00:37:46,790 And here is a very early version of a video for car parking. 395 00:37:46,790 --> 00:37:51,000 I'm afraid it is early one because the later ones have intellectual property that I can't show you. 396 00:37:51,000 --> 00:37:56,370 But you get the idea from this. And these car parks are going to be important within things like smart cities, 397 00:37:56,370 --> 00:38:03,150 not just for conventional cars, but for the storing, car sharing and autonomous vehicles, 398 00:38:03,150 --> 00:38:07,200 including delivery vehicles of different formats so that they can be stored, 399 00:38:07,200 --> 00:38:12,660 charged and serviced and then dispatched on demand close to where they're needed without clogging up our streets. 400 00:38:12,660 --> 00:38:16,590 And this design turns out to be more efficient and cheaper and even more environmentally 401 00:38:16,590 --> 00:38:20,670 friendly because you don't have all the nasty concrete than conventional carhops. 402 00:38:20,670 --> 00:38:26,010 And as you can see from that list, most of these applications are quite far online grocery, 403 00:38:26,010 --> 00:38:34,590 but some of them do have synergies that core business and and the obvious one there is vertical farming as well. 404 00:38:34,590 --> 00:38:43,580 So let's talk about vertical farming. Okay. So people have been trying to do this kind of non conventional farming. 405 00:38:43,580 --> 00:38:53,220 Well, you know, the Aztecs know PanAm and greenhouses have been around for a while, but they all relied on free solar energy. 406 00:38:53,220 --> 00:38:59,940 And of course, people have played with doing vertical farms for certain kind of high value cash crops for a while now. 407 00:38:59,940 --> 00:39:04,260 And indeed, this is quite an interesting one, which we ourselves may get, too. 408 00:39:04,260 --> 00:39:07,350 But but the fact is, we're not going to feed the world by, you know, 409 00:39:07,350 --> 00:39:13,460 sticking farms on top of rooftops in Manhattan, whether people start doing this or even on the side of buildings. 410 00:39:13,460 --> 00:39:18,810 I'm afraid that's just not going to move the dial. So we need to find other ways to do this. 411 00:39:18,810 --> 00:39:27,120 So obviously, there is a revolution going on in terms of lighting at the moment in terms of the cost of that coming down rapidly. 412 00:39:27,120 --> 00:39:34,170 And we we why vertical farming is interesting is because it will help us address food security issues. 413 00:39:34,170 --> 00:39:42,480 It will enable it allow us to be breed plants for taste rather than for pest or drought resistance. 414 00:39:42,480 --> 00:39:49,330 It means that you can pick the plants when they're ripe rather than trying to write them in transport by having multiple cycles. 415 00:39:49,330 --> 00:39:56,400 You know, there is certainly no concept of seasonality. You can use about five percent of the water that you can with commercial agriculture, 416 00:39:56,400 --> 00:40:00,090 significantly lower food, miles, if you build them close to the customer, 417 00:40:00,090 --> 00:40:05,580 which will get to better than organic in terms of food quality demand matched, 418 00:40:05,580 --> 00:40:12,070 which means that you only can grow and bring two, if you like, fruition, what is actually going to be consumed. 419 00:40:12,070 --> 00:40:18,090 And and those go back to those kind of customer trends that I mentioned or desires that I mentioned earlier. 420 00:40:18,090 --> 00:40:24,090 No soil erosion, pesticides or water pollution growing up to up. 421 00:40:24,090 --> 00:40:34,170 If I'm growing these systems using machine learning to optimise them end to end and finally very, very dense and highly automated. 422 00:40:34,170 --> 00:40:39,730 And if you ask yourself, where do you find plenty of sun to generate cheap electricity? 423 00:40:39,730 --> 00:40:42,300 A shortage of water, because, of course, these are closed systems, 424 00:40:42,300 --> 00:40:47,310 which is why you can use so much less water and cheap land that nobody wants to use for anything else. 425 00:40:47,310 --> 00:40:52,800 And of course, the answer is in deserts are now increasingly people are starting to build large vertical farms in desert, 426 00:40:52,800 --> 00:41:01,170 as well as things like large algae production facilities. But the fact is, the world is not going to be fed using leafy greens and basil, 427 00:41:01,170 --> 00:41:04,680 which is kind of like most of the stuff that people grow at the moment, the vertical farms. 428 00:41:04,680 --> 00:41:10,350 So we need new varieties that to be bred suitable for vertical farms. 429 00:41:10,350 --> 00:41:14,790 We need these also for things like exploring not just food that we eat, but food as medicine. 430 00:41:14,790 --> 00:41:21,240 So nutraceuticals and pharmaceuticals and everything that we haven't yet exploited, if you like plants for. 431 00:41:21,240 --> 00:41:29,940 We want to explore the whole area of personalisation of food. If you if we can grow you a. a lycopene enhanced tomato, 432 00:41:29,940 --> 00:41:34,620 you might be able to take eat that rather than taking statins if you've got a particular heart condition. 433 00:41:34,620 --> 00:41:44,130 So lots of exciting things there. Plant based proteins, growing insects, aquaculture and potentially other living things in these dense systems, 434 00:41:44,130 --> 00:41:49,110 exploring new ways to convert green electricity, if you like, into calories, which is the whole purpose of this. 435 00:41:49,110 --> 00:41:54,480 But obviously, vertical farming is only gonna be one part of this new Frudakis ecosystem. 436 00:41:54,480 --> 00:41:59,280 So we have to find all those other ways of doing protein and algae and things like that. 437 00:41:59,280 --> 00:42:05,370 Now, if you remember that grid that I showed you earlier. Hold that in your mind, because I'm going to give you four ingredients. 438 00:42:05,370 --> 00:42:15,560 Here are a few. Angry is kind of it. But now imagine that each of those little cells, each of those boxes was its own separate experiment. 439 00:42:15,560 --> 00:42:22,650 Okay. And imagined that each one is orchestrated as all that whole family of of experiments is orchestrated and 440 00:42:22,650 --> 00:42:31,050 monitored using machine learning and A.I. and and that inside those little boxes you're doing experiments for, 441 00:42:31,050 --> 00:42:34,800 if you like, for creating new kinds of digital growth recipes. 442 00:42:34,800 --> 00:42:41,310 But you're also doing accelerated breeding of new varieties because we have to produce lots of Iraqis that are suitable for farming. 443 00:42:41,310 --> 00:42:45,780 But, you know, we don't necessarily look like the varieties do in the fields at the moment. 444 00:42:45,780 --> 00:42:48,360 And then using maybe even one day, you know, 445 00:42:48,360 --> 00:42:54,810 digital twins of the plants themselves to help steer where we might do the experiments were a little bit way off that at the moment, 446 00:42:54,810 --> 00:43:00,840 generating different growing environments on demand, including potentially ones that don't exist yet. 447 00:43:00,840 --> 00:43:07,110 So whether it be, you know, new environments that will exist because of climate change or if you're Elon Musk, 448 00:43:07,110 --> 00:43:14,520 maybe working out how you're going to do terraforming on Mars by creating that kind of condition and and much more exciting than any of that, 449 00:43:14,520 --> 00:43:24,090 just imagine you could recreate the conditions that were that led to a particular variety or vintage of of wine. 450 00:43:24,090 --> 00:43:27,870 Now, you could do vintage wine as a service. I mean, how good would that be? 451 00:43:27,870 --> 00:43:34,640 But anyway. And. Now, I want to talk about when you start putting all these ingredients together, which is part of our vision. 452 00:43:34,640 --> 00:43:40,040 So the first ingredient is the idea of the vertical farm of different sizes. 453 00:43:40,040 --> 00:43:44,360 You know, these like our warehouse, it come into four sizes. These vertical farms will be the sizes. 454 00:43:44,360 --> 00:43:47,270 Some will be near the customers. Move further out. 455 00:43:47,270 --> 00:43:55,310 If you then glue that onto one of our kind of distribution and logistics machines, that is very, very good at getting the stuff to the customer. 456 00:43:55,310 --> 00:44:03,170 That's that's cool. Then if you add the third ingredient, which is food preparation, but using robotics. 457 00:44:03,170 --> 00:44:05,480 So think about dark kitchens here. 458 00:44:05,480 --> 00:44:13,010 And a company we invested in last year could carry curity is involved in that is building robotics, but very much focussed on on food preparation, 459 00:44:13,010 --> 00:44:20,810 whether it be things like, you know, chilled food, personalised chilled food or recipe boxes or salads or whatever. 460 00:44:20,810 --> 00:44:26,990 And then the last ingredient is one called Acardo Zoome. So we launched this last year. 461 00:44:26,990 --> 00:44:32,670 If you live in West London at the moment, it's the only one that exists at the moment. But we're going to photocopy it soon. 462 00:44:32,670 --> 00:44:38,570 It. You could order from a range of 15000 products. 463 00:44:38,570 --> 00:44:42,200 Bigger than really anybody else who's doing kind of immediacy at the moment. 464 00:44:42,200 --> 00:44:48,050 And we advertise that you can get it in one hour from clicking on your phone app to actually drive your kitchen table. 465 00:44:48,050 --> 00:44:55,820 But it's typically sub half an hour or even some 20 minutes. And indeed, that company, Carrer Curie, when they were doing their launch event, 466 00:44:55,820 --> 00:45:00,110 they told me it got to 10, 30 at night and they ran out of vodka and nibbles. 467 00:45:00,110 --> 00:45:07,460 I mean, that is a proper first world problem. And they they they decided, oh, we're are unless West London, let's use Acardo Zoom's. 468 00:45:07,460 --> 00:45:10,820 They started ordering for carnivals and they came within 20 minutes. 469 00:45:10,820 --> 00:45:15,680 And that's great. And I think they then just carried on ordering folk animals until they really know what they were doing. 470 00:45:15,680 --> 00:45:19,520 But anyway, the. But anyway, so that's all very exciting. 471 00:45:19,520 --> 00:45:25,760 So that's the idea. If you put all that together, vertical farm next to a warehouse, but close to the customer. 472 00:45:25,760 --> 00:45:30,740 Now you've got an ad with automated kind of robotics to prepare the food. 473 00:45:30,740 --> 00:45:37,130 You've now got an extraordinary integrated food machine. And you can go from plant to kitchen table in perhaps two hours. 474 00:45:37,130 --> 00:45:44,420 And unless you live on a farm, that is freshness that you will never know as well as less food, miles and water and all the other benefits. 475 00:45:44,420 --> 00:45:46,910 I went through it again. But you know what? 476 00:45:46,910 --> 00:45:55,080 We need to think way bigger than this, because the transformer racial impact, the technologies like the ones I've talked about robotics, 477 00:45:55,080 --> 00:45:59,960 AI, quantum computing can have on our lives is going to be completely non-linear. It's going to. 478 00:45:59,960 --> 00:46:04,100 And that means that our preparation for that world needs to be non-linear, too. 479 00:46:04,100 --> 00:46:12,080 But one of the challenges I think we really struggle with as individuals, as institutions, as a country, and you as a species is, 480 00:46:12,080 --> 00:46:20,390 as we're talking with Chris before we started, there is insufficient, really big, really long term, really disruptive thinking going on. 481 00:46:20,390 --> 00:46:25,730 And we're not going to solve problems like climate change within a five year democratic term. 482 00:46:25,730 --> 00:46:30,800 And to support more of that big, long term disruptive thinking, we're going to need to structure ourselves to do it. 483 00:46:30,800 --> 00:46:35,930 We're going to need create processes and tools to support it. We're going to need to teach our children how to do it. 484 00:46:35,930 --> 00:46:41,330 And we're going to have to nurture it when it happens. But we're also going to need a compelling vision to drive it, 485 00:46:41,330 --> 00:46:47,900 a holistic vision for a truly smart, equitable, prosperous and sustainable U.K. might look like. 486 00:46:47,900 --> 00:46:55,700 And here, I believe that government has a role not to create that vision, but brackets, God forbid, close brackets, 487 00:46:55,700 --> 00:47:02,130 but rather to convene a diverse group of stakeholders to do that, putting up the Christmas tree for others to decorate, 488 00:47:02,130 --> 00:47:08,300 so to speak, because to signify to successfully compete with some much bigger countries around the world, 489 00:47:08,300 --> 00:47:13,460 we're gonna have to find ways to play the innovation game smarter, more selectively and with greater leverage. 490 00:47:13,460 --> 00:47:18,620 Or, as I like to often put it, we are gonna have to find the asymmetric warfare model of innovation. 491 00:47:18,620 --> 00:47:27,350 And if you've ever sat, as I have in government department waiting areas and watch some of those videos that are playing on the screens, 492 00:47:27,350 --> 00:47:33,920 you would absolutely believe that everything is great. Nothing is broken and that the UK could conquer the world. 493 00:47:33,920 --> 00:47:38,780 Well, as a country, I think we've really got to stop drinking so much of our own Kool-Aid. 494 00:47:38,780 --> 00:47:45,110 We have to get real about what is in the UK is top right hand quadrant like the one I showed you earlier. 495 00:47:45,110 --> 00:47:48,500 And what is the elemental Lego required to implement it? 496 00:47:48,500 --> 00:47:54,050 And then we need to invest heavily in creating all those Lego pieces, whether they be technologies, 497 00:47:54,050 --> 00:48:01,550 institutions, skills, research, programmes, competencies or whatever, because we do not need diversity. 498 00:48:01,550 --> 00:48:07,370 When I was a child, there were five different shaped Lego pieces. Now you go into a toy store and you buy fairy castle thing. 499 00:48:07,370 --> 00:48:12,560 It's got 50 different Lego shapes, you know, and it's great for building fairy castles and it's beautifully pink. 500 00:48:12,560 --> 00:48:15,900 But actually, when you combine that with all the other Terminator's sets, nothing. 501 00:48:15,900 --> 00:48:19,190 A number of Lego pieces just explodes and we just can't afford that. 502 00:48:19,190 --> 00:48:26,690 As a country and the amazing thing is wherever I go. 503 00:48:26,690 --> 00:48:32,600 Around, I find the same recurring patterns at play, whether it be in the airline world, 504 00:48:32,600 --> 00:48:38,600 in extreme environments, the Future Flight Programme, the National Food Strategy and others like fractals. 505 00:48:38,600 --> 00:48:40,460 These these are recurring patterns, though, 506 00:48:40,460 --> 00:48:46,730 things like the need for synthetic environments to speed up innovation and optimise complex systems of systems, 507 00:48:46,730 --> 00:48:52,600 the need for common standards to be able to crowdsource those those models together and build those digital twins. 508 00:48:52,600 --> 00:48:59,430 I was talking about the need for more sophisticated models for sharing assets such as data and digital twin. 509 00:48:59,430 --> 00:49:10,700 So including the passports to do that because free data, open data is great, but it's a massive subset, if you like, of the data. 510 00:49:10,700 --> 00:49:17,540 And you won't get companies like Acardo and others sharing their data unless they know who is going to be shared for and for what purposes. 511 00:49:17,540 --> 00:49:23,360 So it won't be weaponized against them. And if it's not just data for data is needed for doing together digital twins and other assets 512 00:49:23,360 --> 00:49:28,010 to the need for living labs to accelerate learning across lots of different problem domains. 513 00:49:28,010 --> 00:49:32,720 The need for better maps to understand these complex landscapes, because wherever you go, 514 00:49:32,720 --> 00:49:37,220 you find people don't have maps of their areas and they don't really understand how their world fits together, 515 00:49:37,220 --> 00:49:41,060 let alone how their world fits within all the other worlds. 516 00:49:41,060 --> 00:49:48,560 And we need to work out, as I said already, what the elemental Lego is that's needed to implement the four divisions associated with those landscapes. 517 00:49:48,560 --> 00:49:54,020 And we need to think beyond the current challenges and focus on what I would say is the first derivative for the mathematicians. 518 00:49:54,020 --> 00:49:58,520 That room. It's not, for instance, just about where how we deploy robots to solve problems. 519 00:49:58,520 --> 00:50:05,630 It's about how do we create robots that can build and repair robots, how can we build robots that can build factories that build robots and so on. 520 00:50:05,630 --> 00:50:11,270 So it because that's what creates scalability by design, and that's central to what we do. 521 00:50:11,270 --> 00:50:17,300 But I think it's absolute key for the UK to I'm conscious, I'm eating into question time. 522 00:50:17,300 --> 00:50:21,410 So I'm but I'm gonna give you one more two more big ideas. 523 00:50:21,410 --> 00:50:25,250 Okay. So. Right. So I'll leave you with three big ideas, actually. 524 00:50:25,250 --> 00:50:30,440 The first is. Yeah. But by to get one free. So we are a retail aren't we. 525 00:50:30,440 --> 00:50:37,580 So then why the first idea is about how do we move atoms around the UK more efficient and sustainable and scalable ways. 526 00:50:37,580 --> 00:50:42,350 And the fact is that many supply chains are fragmented and inefficient, 527 00:50:42,350 --> 00:50:48,510 including the last mile with everybody kind of doing their own thing with very little coordination and collaboration. 528 00:50:48,510 --> 00:50:54,470 And this is true about freight in general, but it's definitely true about food, which is an extreme case. 529 00:50:54,470 --> 00:50:57,230 And there's a little point about transforming food production. 530 00:50:57,230 --> 00:51:02,510 If we can't get that food to where it's going to be processed, stored, sold and ultimately consumed. 531 00:51:02,510 --> 00:51:04,310 So it's time for another analogy here. 532 00:51:04,310 --> 00:51:10,150 For those of you in the room who use public cloud, and I'm afraid all of you do if you don't know it because your phones do, 533 00:51:10,150 --> 00:51:13,880 you have the advantages of public cloud over on premise? Infrastructure are numerous, 534 00:51:13,880 --> 00:51:18,410 and they include things like foster innovation funded by using their income from lots of customers to drive 535 00:51:18,410 --> 00:51:24,920 your R&D common stance and interfaces shared middleware that no company could afford to build for itself. 536 00:51:24,920 --> 00:51:27,860 Higher levels of security, resilience and robustness. 537 00:51:27,860 --> 00:51:32,960 The ability for users to buy capacity on demand, which means they can scale it up and get it down as they need. 538 00:51:32,960 --> 00:51:39,410 And that leads to faster experimentation and greater scalability is enabling users to start 539 00:51:39,410 --> 00:51:45,350 focussing on where they add real value rather than building operating their own infrastructure. 540 00:51:45,350 --> 00:51:49,940 And so to drive greatest sustainability and efficiency into our supply chains. 541 00:51:49,940 --> 00:51:54,560 What about creating a public internet of freight for the UK? What would that do? 542 00:51:54,560 --> 00:51:58,670 Well, it could provide a common set of data standards, interfaces, governance, 543 00:51:58,670 --> 00:52:03,830 protocols, metrics, analytics and s allays a common set of smart middleware services. 544 00:52:03,830 --> 00:52:13,160 An ecosystem that suppliers, processes, delivery firms, retailers and others could connect to in order to provide and receive services and data. 545 00:52:13,160 --> 00:52:19,910 Operational economies of scale and greater security. Resilience, robustness, improved utilisation of scarce resources, whether that be energy, 546 00:52:19,910 --> 00:52:25,790 land time, transportation, network bandwidth and the rest reduce pollution, waste and congestion. 547 00:52:25,790 --> 00:52:31,470 A single point of integration with post Brexit measures for goods entering leaving the country. 548 00:52:31,470 --> 00:52:38,220 I think extra on steroids, the ability to provide common regulation, quality stands, 549 00:52:38,220 --> 00:52:43,090 analytics and monitoring support for new, cross-sector and complet across competitor models. 550 00:52:43,090 --> 00:52:47,180 If Accardo is delivering a pizza to a hospital, there might be a bag of plasma. 551 00:52:47,180 --> 00:52:49,460 They'd like us to pick up a bit in our refrigerated van. 552 00:52:49,460 --> 00:52:54,260 Well, those are two worlds that will never find each other at the moment because they're completely separated. 553 00:52:54,260 --> 00:52:58,310 Whereas if you actually had a system that is managing that as a holistic model, 554 00:52:58,310 --> 00:53:03,440 we will find those intersections and we need to a high fidelity simulation or digital spin, 555 00:53:03,440 --> 00:53:09,810 of course, of that end to end Internet, afraid to help optimise it and and manage the complexity. 556 00:53:09,810 --> 00:53:14,630 And it isn't just about a digital intranet. Afraid that's where we might start. It's about physical ones, too. 557 00:53:14,630 --> 00:53:19,790 So things like rapid transit networks for freight, not least to get more stuff off our roads. 558 00:53:19,790 --> 00:53:24,050 And here you need to think about things like Hyperloop, which is obviously a must thing. 559 00:53:24,050 --> 00:53:28,600 But for, you know. People. But in vacuum. 560 00:53:28,600 --> 00:53:34,090 But think about the same thing about tubes on the ground with freight moving stuff around. 561 00:53:34,090 --> 00:53:43,000 And how to make better use of our railways and so forth. And the second big idea is about Catalyst's, because as a scientist, the room. 562 00:53:43,000 --> 00:53:48,100 Well, no, you know, Catholics either make reactions occur that weren't naturally occur or they speed up other ones. 563 00:53:48,100 --> 00:53:52,480 And I believe there are a set of national Catholics that will help the UK work smarter, 564 00:53:52,480 --> 00:53:56,380 faster, more efficiently, safer, more sustainably with greater competitive advantage. 565 00:53:56,380 --> 00:54:03,270 And what might be a list on that? Be on that list. National Catholics. Well, a national digital twin and a family Samiti maps the UK, 566 00:54:03,270 --> 00:54:06,970 a national network of living labs, to accelerate learning across lots of proper domains. 567 00:54:06,970 --> 00:54:10,720 A national network of maker labs to balance digital and physical worlds because 568 00:54:10,720 --> 00:54:14,560 the most exciting stuff happens at the intersection that two of those two worlds. 569 00:54:14,560 --> 00:54:21,370 Which is kind of the world that we live in. But it's also to foster the next generation of inventors and tinkerers. 570 00:54:21,370 --> 00:54:26,800 About last but not least, I'm going to leave you with the most important national catalyst, the more. 571 00:54:26,800 --> 00:54:28,390 Which is education. 572 00:54:28,390 --> 00:54:38,590 And I'm mindful of the fact that I'm in a, you know, educational institution and that probably what I'm about to say will ruffle some feathers. 573 00:54:38,590 --> 00:54:42,940 But I'm going to say it anyway when it comes to the uncertainty that we now face 574 00:54:42,940 --> 00:54:46,930 regarding the impact of these disruptive technologies that I've been talking about. 575 00:54:46,930 --> 00:54:52,030 One of the very few insurance policies we have for that future world is education. 576 00:54:52,030 --> 00:54:55,630 But as an employer, as a father, and indeed as a citizen, 577 00:54:55,630 --> 00:55:02,500 I would assert that our current education system is not fit for purpose in terms of preparing the next generation for that smarter, 578 00:55:02,500 --> 00:55:09,340 more automated world that they will inherit. I believe it's broken now, but the cracks are just going to open up wide very soon. 579 00:55:09,340 --> 00:55:14,050 And this is a very big topic. And I look, I'm just going to skim the top of it to finish. 580 00:55:14,050 --> 00:55:16,540 But many of the skills and techniques, I believe, 581 00:55:16,540 --> 00:55:24,010 that we're currently teaching our children are going to become as devalued in the years to come as the encyclopaedia has been by the World Wide Web. 582 00:55:24,010 --> 00:55:32,040 And instead, we need to focus on teaching enduring metter skills, such as learning how to learn collaboration, creative thinking, 583 00:55:32,040 --> 00:55:40,270 problem solving, intersectional thinking, mind mapping, design thinking, goal setting, emotional intelligence, innovation and entrepreneurship. 584 00:55:40,270 --> 00:55:45,550 Because if education is meant to be about preparing the next generation for the smarter foci, 585 00:55:45,550 --> 00:55:50,110 for their future life, and instilling a love of learning because it's going to be all about lifelong learning, 586 00:55:50,110 --> 00:55:55,540 then I believe we're failing at the moment in terms of the structure and curriculum of our current education system, 587 00:55:55,540 --> 00:56:03,850 because the relentless focus on exams, tests and the regurgitation of BARG schemes is consuming almost all of the education oxygen, 588 00:56:03,850 --> 00:56:08,380 leaving teachers with very little time for spontaneity or for just sharing their love of 589 00:56:08,380 --> 00:56:14,410 a subject and and pursuing the curiosity of their students to see where it might lead. 590 00:56:14,410 --> 00:56:19,450 And if we allow education to switch our students off the joy of learning, then we will do them an uncomfortable to service. 591 00:56:19,450 --> 00:56:22,810 But on the other hand, if we are unable our children to leave school, 592 00:56:22,810 --> 00:56:30,310 having learnt how to learn full of curiosity and with a holistic set of future proof skills and with the joy of ongoing learning, 593 00:56:30,310 --> 00:56:33,250 then I believe they will be well equipped for their life ahead. 594 00:56:33,250 --> 00:56:40,000 So I believe we need to completely rethink our education system from the ground up and in so doing, future proof that online curriculum. 595 00:56:40,000 --> 00:56:48,820 So I'm sure some of you are sitting there politely thinking an Internet afraid, a set of national Catalyst's rethink our education system. 596 00:56:48,820 --> 00:56:53,860 He's completely nuts. It's definitely time to sell my Acardo shares. Well, maybe I am. 597 00:56:53,860 --> 00:56:56,200 But then you know what? You do have to dream big. 598 00:56:56,200 --> 00:57:02,570 And I can assure you that many people thought that Google Street View was completely unachievable a man before it just simply got done. 599 00:57:02,570 --> 00:57:08,410 So I inadvertently use that government phrase that I write. 600 00:57:08,410 --> 00:57:13,270 So I want to finish by setting you a challenge, because I believe great leadership is about many things, 601 00:57:13,270 --> 00:57:17,290 but it's definitely about inspiring people to do things they had no idea they could achieve. 602 00:57:17,290 --> 00:57:25,030 And I believe we need to inspire our citizens with a great vision and bigger thinking in order to power our nation to do even more amazing things, 603 00:57:25,030 --> 00:57:29,890 including transforming our food system and including solving climate change and the rest. 604 00:57:29,890 --> 00:57:36,280 So as leaders, technologists, employees, parents and citizens, I believe we need to challenge ourselves to think really big, 605 00:57:36,280 --> 00:57:43,810 really long term and really disruptively about our future and then to challenge our businesses, institutions and governments to do the same. 606 00:57:43,810 --> 00:58:09,358 Thank you very much indeed.